| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Institute of Neurobiology and Department of Anatomy, University of Puerto Rico Medical Sciences Campus, San Juan, Puerto Rico; 2Zentrum für Molekulare Biomedizin, Friedrich-Schiller-Universität Jena, Jena, Germany; and 3Fishberg Department of Neuroscience, Mount Sinai School of Medicine, New York, New York
Submitted 5 September 2006; accepted in final form 31 January 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
; Marder and Calabrese 1996; Pearson 1993; Stein et al. 1997). The patterned cycling of the CPG ensures a basic stereotyped activity that then, however, must be modified to meet varying behavioral demands. This modification is often achieved by neuromodulation (Harris-Warrick and Marder 1991; Katz 1999; Marder and Thirumalai 2002; Nässel 1996; Nusbaum and Beenhakker 2002). Typically, the modulation is complex. Most CPGs are regulated by multiple modulators, each with a different but overlapping, sometimes even apparently redundant, constellation of effects on the intrinsic properties of the individual neurons and their synaptic interconnections (Brezina and Weiss 1997; Harris-Warrick et al. 1998; Katz 1999; Marder and Thirumalai 2002; Nusbaum et al. 2001; Skiebe 2001). The same modulators typically also act on the muscles that the CPG drives (Calabrese 1989; Hooper et al. 1999; Kobayashi and Muneoka 1990; Meyrand and Marder 1991; Weiss et al. 1992; Worden 1998). In general terms, the complexity of the modulation likely reflects the complexity of the CPG–effector network, with multiple interacting components and feedback loops, whose activity must be controlled in an integrated manner. For example, recent experimental and theoretical studies have emphasized the need to tune the properties of the muscle effectors coordinately with the activity of the CPG (Brezina et al. 2000, 2003a,b, 2005; Chiel and Beer 1997). However, the detailed logic of the modulatory architecture is not completely understood in any modulated system.
This paper is part of our work to approach this question in an exceptionally simple CPG–effector network: the cardiac system of the blue crab Callinectes sapidus (Fig. 1; for reviews see Cooke 1988, 2002; Hagiwara 1961; additional references in Fort et al. 2004). The crab heartbeat is neurogenic, driven by a rhythmic motor program generated by a simple CPG, the cardiac ganglion (CG) that is embedded within the heart itself. The CG contains only nine neurons: four local premotor interneurons that are thought to act as pacemakers and five motor neurons that send bursts of spikes out of the ganglion to contract the single-chamber heart. Through mutual electrical coupling, all of the motor neurons fire in precise synchrony (Fort et al. 2004; Tazaki and Cooke 1979). Recent studies suggest that feedback from the heart muscle can, in turn, modify parameters of the motor program produced by the CG (Mahadevan et al. 2004; Sakurai and Wilkens 2003). Altogether, this is perhaps the simplest complete CPG–effector system that yet allows the mutual interactions of the center and periphery and the integration of the system into a coherently acting whole to be studied.
|
; Cooke and Sullivan 1982). Anatomically, there are two major ways in which the modulators reach the heart. First, neurosecretory neurons in the CNS project to peripheral neurohemal structures, including the pericardial organs (POs) located within the pericardial sinus, where they release the modulator into the bloodstream in a neurohormonal fashion (Fig. 1; Cooke and Sullivan 1982). Second, regulatory neurons in the CNS project further, through the dorsal nerve (Fig. 1), directly into the heart where they innervate the neurons of the CG and possibly the myocardium (Alexandrowicz 1932; Maynard 1960). In our previous work (Fort et al. 2004), we examined the anatomy and physiology of one modulator, dopamine, in the Callinectes cardiac system. We found that a single pair of modulatory CNS neurons very likely delivers dopamine to the heart in both ways, as a neurohormone released from the POs and through direct innervation of the heart. Correspondingly, we found multiple effects of dopamine on the physiology of the system (Fort et al. 2004; see DISCUSSION).
Dopamine, however, is not the only modulator of the cardiac system and, indeed, probably never acts on it alone. Here, we study another important modulator of the system—crustacean cardioactive peptide (CCAP). CCAP which was initially purified from the pericardial organs of the shore crab Carcinus maenas (Stangier et al. 1987). It is an amidated nonapeptide with a cyclic structure (PFCNAFTGC-NH2) that has no structural similarity to any other known peptide. After its original identification in the crab, CCAP was identified in additional crustaceans (Stangier and Keller 1990) and insects (Cheung et al. 1992; Lehman et al. 1993; Stangier et al. 1989). In addition to modulating the heart in a number of crustacean species (McGaw et al. 1995; Saver and Wilkens 1998; Saver et al. 1999; Stangier 1991; Stangier and Keller 1990; Stangier et al. 1987; Wilkens 1995; Wilkens and Mercier 1993; Yazawa and Kuwasawa 1992), CCAP was found to act on several arthropod visceral muscles (Donini et al. 2001, 2002; Groome and Lehman 1995; Stangier and Keller 1990), the crayfish retina (Gaus and Stieve 1992), and the crab stomatogastric system (Jorge-Rivera et al. 1998; Richards and Marder 2000; Weimann et al. 1997). Recently, hemolymph levels of CCAP were shown to be greatly elevated during active phases of crustacean (crab and crayfish) ecdysis, and it was proposed that CCAP regulates the changes in physiology and behavior that occur during the exuviation (molting) process (Phlippen et al. 2000).
Although its cardioactive properties enabled the initial purification of CCAP, its mode of action on crustacean hearts remains incompletely understood. In this study, we examine the anatomy and physiology of its action on the Callinectes cardiac system, with a view toward generating data and insights that will contribute to an eventual full understanding of the modulatory logic in this simple CPG–effector system.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Histology
CCAP-LIKE IMMUNOHISTOCHEMISTRY (CCAPLI).
Specimens were covered in ice (30 min) to achieve immobilization. Tissues were dissected, secured to Sylgard-lined petri dishes with minuten pins, and fixed for 1 h in freshly prepared 4% paraformaldehyde. Standard wholemount immunohistochemical protocols were followed (see Miller et al. 1991 for detailed buffer composition, incubation, and wash procedures). Ganglia were washed (five times, room temperature with agitation) in PTA (0.1 M phosphate buffer containing 2% Triton X-100 and 0.1% sodium azide). After preincubation with normal goat serum (0.8%), tissues were immersed (48 h, room temperature) in a 1:5,000 dilution of the primary CCAP antibody (rabbit polyclonal). After repeated PTA washes (five times, 
30 min each, room temperature), the ganglia were incubated in secondary antibodies (1:3,000 dilution) conjugated to a fluorescent marker [Alexa 488 goat anti-rabbit IgG (H+L) conjugate; Molecular Probes, Eugene, OR]. The Alexa 488 was viewed with the G-2A filter block of the Nikon Optiphot or using the preconfigured FITC channel of a Zeiss Pascal LSM5 laser-scanning confocal microscope. Standard images were captured using the ACT1 (Nikon) software package. Confocal images were reconstructed (AIM Software) from sequential images captured in the z-axis plane of the tissue. Images were transported as TIFF files to Adobe Photoshop (version 6) for adjusting overall contrast and brightness. Finally, they were imported to Corel Draw 9 for addition of labels, cropping, and arrangement of panels.
NERVE BACKFILLS.
The biotin–avidin protocol followed the methods of Fort et al. (2004). The thoracic ganglionic mass was pinned out near a small Vaseline well that was formed on the Sylgard surface. The segmental nerve of the first thoracic neuromere (SN1) was cut and drawn into the well. The crab saline inside the well was withdrawn and replaced with a saturated aqueous solution (1.6 mg/30 µl) of biocytin (Sigma Chemical, St. Louis, MO). The preparation was covered and incubated overnight at 14°C. The well was then removed and the ganglia were washed three to five times, repinned, and fixed in paraformaldehyde as described earlier. The fixed ganglia were transferred to microcentrifuge tubes, washed five times (30 min each) with PTA solution, and incubated overnight (room temperature, with shaking) in Rhodamine600 Avidin D (Vector Laboratories, Burlingame, CA) diluted 1:3,000 in PTA (24–48 h, room temperature). The ganglia were then washed five times with PTA and the quality of the backfill was assessed before further immunohistochemical processing. In the double-labeling experiments (Fig. 4), CCAPli was visualized using the Alexa 488 goat anti-rabbit secondary antibody (see above). A barrier filter (546-nm green interference) was used to eliminate "bleedthrough" of rhodamine when examining and photographing CCAPli fluorescence.
|
|
|
WORKING HEART (WH). The heart was removed intact; the sternal artery was cannulated with a modified syringe needle and mounted in a 20-ml organ bath. The heart was suspended using a fine monofilament nylon thread attached to the force plates of a Grass (Astro-Med, West Warwick, RI) FT03 isometric force transducer and placed under a resting load (about 0.5 g). Perfusion with saline was maintained at a constant rate (2 ml/min) and pressure. The crab saline composition was based on Pantin's saline for Cancer pagarus (in mM): 487 NaCl, 13.6 KCl, 13.4 CaCl2, 13.6 MgCl2, 1.4 sodium sulfate, 3 HEPES (N-[2-hydroxyethyl] piperazine-N'-[2-ethanesulfonic acid]), adjusted to pH 7.4 with sodium hydroxide. Perfusion rate and pressure were maintained when CCAP trials were performed.
SEMI-INTACT WORKING HEART (S-IWH). The heart was pinned in a Sylgard-lined petri dish in an arrangement as similar as possible to that in the intact crab. A small incision was made in the ventral wall of the heart to expose part of the nerve ring containing the motor neuron axons. The ring was cut, usually across one of the posterolateral connectives (see Fig. 1), and the severed end proximal to the ganglion was drawn into an extracellular suction electrode. Motor patterns were recorded with a differential AC amplifier (Model 1700, A-M Systems, Carlsborg, WA) and digitized with a PowerLab (ADInstruments, Colorado Springs, CO) data-acquisition system (total sampling rate 100 kHz). The heart was then connected to a Grass FT03 isometric force transducer with a hook and nylon thread and placed under a resting load (about 0.5 g). The preparation was continually internally perfused with saline at a constant rate (2 ml/min) and pressure. Perfusion rate and pressure were maintained when CCAP trials were performed.
ISOLATED CARDIAC GANGLION (ICG).
The heart was pinned ventral side up in a Sylgard-lined petri dish. A cut was made in the ventral musculature exposing the cardiac ganglion. Dissection was achieved principally by teasing away the adhering muscles. Previous investigators (Tazaki and Cooke 1979) noted that the region within the confluence of the motor roots at each end of the ganglion contains the dendritic endings of the ganglionic neurons. A substantial noncontracting remnant was therefore retained at either end of the ganglion. Extracellular suction electrode recordings were obtained as in the S-IWH from at least one of the four cut ganglionic roots. Intracellular recordings were obtained from anterior and/or posterior motor neurons using microelectrodes filled with 2 M KCl (10–30 M
). The preparation was continuously superfused with saline (2 ml/min).
CCAP APPLICATION.
In all physiological preparations, CCAP was applied by switching the continuously perfusing saline to CCAP-containing saline, then, after several minutes, back to control saline for wash. The relatively slow rate and large dead volume of the perfusion system meant that the CCAP response developed only 30–60 s after the switch to CCAP-containing solution and washed out over a number of minutes (see, e.g., Figs. 7 and 8, A–C). The preparation was washed for 20 min before another CCAP application. Typically a series of increasing CCAP concentrations, each 10-fold higher than the previous, was tested, separated by washes; no cumulative CCAP applications were performed. The dose–response relations in Figs. 6, 9, and 11 were constructed by comparing the parameter values averaged over a 1-min period during the strongest response to each CCAP concentration to control values averaged over a 1-min period immediately before the application of that CCAP concentration. The values were fitted with the equation: % Change = a/{1 + (log [CCAP]/EC50)b}, where [CCAP] is the CCAP concentration and a, b, and EC50 are parameters of the fit.
|
|
|
|
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
First, we used immunohistochemical methods to identify the possible modes and sources of regulation of the Callinectes cardiac system by CCAP (Fig. 2). In agreement with previous observations in Carcinus maenas (Dircksen and Keller 1988), Cancer borealis (Christie et al. 1995), Cancer productus (Fu et al. 2005), and several other arthropods (Dircksen 1998), CCAP-like immunoreactivity (CCAPli) was observed in varicosities and terminals throughout the Callinectes pericardial organs (Fig. 2B). The varicosities, which were located on the surface of all bars and trunks (nomenclature of Alexandrowicz 1953), appeared to originate from fibers that passed through the core of the PO (Fig. 2B, arrows). In contrast to our previous findings with dopamine (Fort et al. 2004), no CCAPli fibers were present in the dorsal nerve that projects from the POs to innervate the heart and no CCAPli was detected in the cardiac ganglion (four preparations; not shown). These observations suggest that, in contrast to dopamine, regulation of cardiac activity by CCAP is achieved in a purely neurohormonal fashion.
The distribution of CCAPli neurons in the Callinectes nervous system followed the pattern previously observed in other arthropods (reviewed by Dircksen 1998). Symmetric pairs of immunoreactive neurons were located in each of the fused ganglia of the thoracic ganglion complex (TGC) (Fig. 2, C–F). In the anterior neuromeres of the TGC (referred to as the subesophageal or mouthpart ganglia), one neuron in each pair was substantially larger (60–70 µm; arrows in Fig. 2D) than the other (20–30 µm; arrowheads in Fig. 2D). The larger of the two CCAPli neurons (corresponding to the type-1 cell in the nomenclature of Dircksen 1998) was usually positioned slightly posterior to the smaller (type-2) neuron. No consistent differences in their staining intensities were observed. Similar pairs of CCAPli neurons were located in the more posterior (leg) neuromeres (Fig. 2E), but their relative position was more variable, with the smaller neuron frequently being located posterior to the larger. In the most posterior (abdominal) neuromeres, CCAPli neurons also occurred in pairs, but differences in their sizes were less evident (Fig. 2F, arrows). At least one strongly immunoreactive fiber crossed the midline in each neuromere (Fig. 2, C, arrows, E, and F, arrowheads). It was not possible to trace these commissural fibers to their destinations or to their origins.
In decapod crustaceans, neurosecretory cells within the CNS project to the pericardial organs through the segmental nerves (SNs) that originate from each neuromere of the TGC (Cooke and Sullivan 1982; Maynard 1961a). To identify potential sources of the CCAPli in the POs, we performed backfills of the first segmental nerve (SN1) of the most anterior TGC neuromere (see Fig. 1). Nerve tracing using biocytin (Fig. 3) revealed a more extensive complement of central neurons projecting into SN1 than was reported previously (Fort et al. 2004; Maynard 1961b). The majority of SN1 fibers originated from two clusters (designated B-cells and C-cells by Maynard 1961b) in the dorsolateral region of the ipsilateral anterior TGC (Fig. 3, A and C). Additional neurons that projected into the ipsilateral SN1 included four neurons near the midline that were aligned in the rostral–caudal direction (Fig. 3D). The most posterior neuron in this group was larger than the others (Fig. 3D, arrow). It was irregularly shaped and gave rise to a very stout axon. Finally, a single neuron in the anterior region of the first ipsilateral leg neuromere was found to project to SN1 (Fig. 3E, arrow). In addition to these neurons, a labeled fiber could be followed the entire length of the TGC (Fig. 3, B, arrow, and E), but it could not be traced to a specific neuron.
In the contralateral anterior TGC, the backfills of SN1 produced staining of two neurons (Fig. 3, A and F). One of these neurons (Fig. 3F, arrow) was substantially larger than the other. The larger neuron had a prominent axon hillock and fiber coursing in the medial–anterior direction. The fiber originating from the smaller neuron was very thin and followed a more irregular trajectory toward the midline.
To determine the source of the CCAP in the POs, we performed the CCAP immunohistochemistry and the biocytin backfills of SN1 in combination (Fig. 4). No CCAPli was observed in any of the ipsilateral backfilled neurons. However, one of the contralateral neurons that were backfilled in the anterior TGC exhibited double labeling (Fig. 4, A–C, arrows). This was the larger of the two backfilled neurons in Fig. 3F, now revealed to be the most anterior of the type-1 CCAPli cells seen in Fig. 2, C and D. The adjacent type-2 CCAPli cell (arrowheads in Fig. 4, B and C) did not correspond to the smaller contralateral backfilled neuron, although its size and position were similar. These observations agree with findings in other species where the CCAP in the POs was shown to originate from a single contralateral type-1 neurosecretory cell in each thoracic neuromere (Dircksen 1998).
Physiological effects of CCAP: working heart and semi-intact working heart
In our previous work in the crab cardiac system, we developed two "working heart" preparations (Fort et al. 2004; see METHODS). The basic working heart (WH) preparation was essentially the intact heart, with the cardiac ganglion embedded within it, removed from the crab and perfused but not further dissected. This preparation allowed us to record the heartbeat under minimally perturbed conditions. In the semi-intact working heart (S-IWH) preparation, the heart was lightly dissected to allow us to record not only the heart muscle contractions (e.g., top traces of each panel in Fig. 5), but also extracellularly from one of the connectives through which the axons of the cardiac ganglion motor neurons project to the muscle, the underlying motor neuron spike patterns (bottom traces of each panel in Fig. 5). When in the previous work we systematically compared the two preparations, we found that the S-IWH preparation retained, as far as could be determined, all of the principal properties of the intact WH (Fort et al. 2004). The effects of CCAP, too, were essentially the same in both preparations (see, e.g., the statistical comparison of the dose–response relations in Fig. 11). We will therefore discuss both preparations jointly, focusing however primarily on the S-IWH preparation that provided more complete information.
Representative recordings of the effects of CCAP in the S-IWH preparation are shown in Fig. 5 and summary dose–response relations pooled from all S-IWH preparations are shown in Fig. 6. Bath application of increasing concentrations of CCAP had both inotropic and chronotropic effects, generally in the positive direction: increases in the amplitude and frequency of the heart contractions. The increase in the frequency of the contractions (top traces in Fig. 5) presumably reflected the increase in the frequency of the motor neuron spike bursts that produced each contraction (bottom traces in Fig. 5). At no CCAP concentration, however, was there any significant change in the duration of the spike bursts or in the number of spikes per burst (Fig. 6, C and D; for statistical analysis see legend).
Dissociation of the inotropic and chronotropic effects
The increase in contraction amplitude was observed at lower CCAP concentrations than the increase in burst and contraction frequency. In the summary dose–response relations, the apparent threshold for the CCAP effect on contraction amplitude was between 10–10 and 10–9 M (Fig. 6A), but for the effect on burst and contraction frequency it was between 10–8 and 10–7 M, two orders of magnitude higher (Fig. 6B). The dose–response relation for the latter effect then rose steeply, however, reaching near-maximal values within approximately another two orders of magnitude (Fig. 6B). Tests of statistical significance carried out on the summary dose–response relations (see Fig. 6 legend) showed that the effect of CCAP on contraction amplitude was statistically significant (as indicated by the asterisks in Fig. 6) already at 10–9 M (and at all higher concentrations), but on the burst and contraction frequency only at 10–7 M (and higher concentrations).
In the summary dose–response relations, however, the CCAP effects were already normalized and averaged across multiple preparations whose raw values might have been quite different from each other. We therefore sought to confirm that the dissociation seen in the averaged dose–response relations could also be observed in the raw values measured at the same time in the same preparation. We performed experiments in which we directly compared the effects of a low (10–9 M) and a high (10–5 M) CCAP concentration in the same S-IWH preparation. Figure 5 is drawn from one of these experiments. At 10–9 M, CCAP had essentially no effect on the burst and contraction frequency, the burst duration, or the number of spikes per burst, yet it clearly increased the contraction amplitude (Fig. 5A; the control and CCAP bursts and contractions boxed in A1 and A2, respectively, are superimposed in A3 for direct comparison). At 10–5 M, CCAP increased the contraction amplitude considerably more and also increased the frequency of the bursts and contractions, but still had no effect on (or even decreased) the burst duration and the number of spikes per burst (Fig. 5B). In a comparison of 33 control heartbeats and 39 beats with 10–9 M CCAP in this preparation, there was a significant increase in contraction amplitude [control (mean ± SE) 1.44 ± 0.01 g, CCAP 2.00 ± 0.03 g, increase by 39%, P < 0.001, Mann–Whitney rank-sum test] but no significant change in burst and contraction frequency [control 25.37 ± 0.20 bursts and heartbeats per minute (BPM), CCAP 25.92 ± 0.15 BPM, P > 0.05], burst duration (control 104.33 ± 3.95 ms, CCAP 108.79 ± 3.42 ms, P > 0.05), or the number of spikes per burst (control 5.33 ± 0.13, CCAP 5.49 ± 0.10, P > 0.05). In a comparison of 36 control heartbeats and 68 beats with 10–5 M CCAP, there was an even larger increase in contraction amplitude (control 0.60 ± 0.010 g, CCAP 1.32 ± 0.010 g, increase by 120%, P < 0.001), now with a significant increase in the burst and contraction frequency (control 28.46 ± 0.16, CCAP 41.11 ± 0.25 BPM, increase by 44%, P < 0.001), but still no significant change in the burst duration (control 55.82 ± 3.40 ms, CCAP 55.85 ± 1.04 ms, P > 0.05) and a slight decrease in the number of spikes per burst (control 5.36 ± 0.08, CCAP 4.91 ± 0.03, decrease by 8%, P < 0.001). Such a decrease in the number of spikes per burst would have tended to decrease, rather than increase, the contraction amplitude (see, e.g., Fig. 8D).
The inotropic and chronotropic effects of CCAP could also be dissociated in the temporal dimension. Figure 7, A and B, shows the time course of the effects of CCAP on contraction amplitude (red) and frequency (blue) at 10–8 M (A) and 10–6 M (B) in a representative WH preparation. When the CCAP was applied, it increased the contraction amplitude similarly at both concentrations, either without (at 10–8 M) or with (at 10–6 M) a concomitant increase in the contraction frequency. The onset of both effects was rapid, within several heartbeats, and was almost certainly limited by the speed of access of the CCAP to the tissue. Subsequently, although the contraction amplitude remained elevated, the contraction frequency progressively declined to values even lower than the initial control frequency. The effects on both amplitude and frequency reversed when the CCAP was washed out. Figure 7C shows the averaged time courses from six such experiments, three in WH preparations and three in S-IWH preparations, with 10–6 M CCAP. The dissociation was robust across the six preparations: the contraction frequency (blue) declined relatively rapidly to values even lower than the initial control frequency, whereas the contraction amplitude (red) declined much more slowly; the difference was statistically significant (see Fig. 7C legend).
Dynamic modes
The results described to this point might give the impression that the heartbeat was regular: that the parameters with which each beat could be described, such as the contraction amplitude, burst and contraction frequency, burst duration, and the number of spikes per burst, had values that remained relatively constant from beat to beat under control conditions and, then, on application of CCAP, shifted to new values that again remained relatively constant from beat to beat. This was often the case. Frequently, however, the time series of these parameters had a more complex structure (Fig. 8). Sometimes the parameters merely varied from beat to beat in an irregular manner that did not exhibit any discernible pattern. At other times, however, the structure was again regular but at a higher level, exhibiting higher-order repeating patterns or cycles consisting of two, three, four, or more successive beats. As the recording continued, the system switched rather abruptly from one such dynamic "mode" of activity to another. This sometimes occurred spontaneously, but more often as CCAP was applied or as it was being washed off.
Figure 8, A–C shows examples from three different S-IWH preparations. The vertical dashed lines mark the times of the apparent switches between modes. In Fig. 8A, the system switched from an irregular mode to the simplest regular, constant mode. In Fig. 8B, a similar switch occurred, although in this case the initial "irregular" mode in fact exhibited some nonrandom alternation between beats with weaker spike bursts and smaller contractions and beats with stronger bursts and larger contractions.
In Fig. 8C, the system switched between three quite distinct modes. In Fig. 8E we have plotted sections of the time series from Fig. 8C in the space spanned by the three parameters measured: the contraction amplitude, the burst and contraction frequency, and the number of spikes per burst. In this space, a single point repeated in successive beats represents the simplest, constant mode, whereas regularly repeating cycles of two, three, four, or more points represent higher-order cycles of two, three, four, or more beats. Thus Fig. 8E shows how the system in Fig. 8C progressed, when CCAP was applied, successively from the simplest, constant mode (blue) to a higher-order cycle of five beats (green), and finally to a higher-order cycle of four beats (red).
Regularization of dynamic activity by neuromodulators was recently noted in other systems (e.g., Serrano and Miller 2006; Szücs et al. 2005). In the mode switches seen here, too, CCAP often made the activity more regular and constant from beat to beat (as in Fig. 8, A and B), although sometimes the opposite occurred (as in Fig. 8C). These phenomena were seen not only in S-IWH preparations, but in WH preparations as well. In Fig. 7, A and B, for example, the system appeared to switch, when either concentration of CCAP was applied, from an irregular mode to the simplest regular, constant mode and, then, as the CCAP was washed off, through a brief interval of higher-order cycles of two or four beats (arrows in Fig. 7A) back to a more irregular mode.
In the more irregular modes, the beat-to-beat variation of the parameter values constituted natural experiments in which the relationships between the parameters could be examined. For example, Fig. 8D plots, from all of the beats boxed in Fig. 8B, the contraction amplitude as a function of the number of spikes per burst. The beats varied only in these two parameters; the third parameter measured—the burst and contraction frequency—remained constant. Under these circumstances, the two varying parameters were highly correlated on a beat-to-beat basis. As the number of spikes per burst varied over a nearly fourfold range, apparently randomly from one beat to the next, so did the contraction amplitude. The linear regression fit (solid line in Fig. 8D) had a coefficient of determination R2 = 0.96.
Physiological effects of CCAP: isolated cardiac ganglion
In the isolated cardiac ganglion (ICG) preparation, the cardiac ganglion was dissected completely free of the heart musculature (Fort et al. 2004; see METHODS). The motor neuron spike bursts could then be recorded extracellularly as in the S-IWH preparation (e.g., bottom traces in Fig. 9, A1 and A2) and simultaneously with intracellular voltage recording from the motor neuron cell bodies (top traces in Fig. 9, A1 and A2).
Figure 9A is a representative recording of the effects of CCAP, and Fig. 9, B–D shows summary dose–response relations for the frequency of the motor neuron bursts, the burst duration, and the number of spikes per burst. CCAP had no effect on the burst frequency at any concentration 
10–6 M (Fig. 9B). It did, however, increase both the burst duration and the number of spikes per burst, at concentrations beginning around 10–7 M in a statistically significant manner (see Fig. 9 legend and asterisks in Fig. 9, C and D).
The activity of the motor neurons is driven by activity of the other class of neurons in the cardiac ganglion: the four local premotor interneurons. Firing of the interneurons elicits excitatory postsynaptic potentials (EPSPs) in the motor neurons (Mirolli et al. 1987; Tazaki and Cooke 1979, 1983a,b). We did not record directly from the interneurons, but we inferred their spike bursts from the bursts of EPSPs that we could record intracellularly in the motor neuron somata (Fig. 10). The interneuron bursts (e.g., as indicated by the gray bars in Fig. 10B1) were longer, beginning earlier and ending later, than the corresponding motor neuron bursts (black bars in Fig. 10B1). At 10–6 M, CCAP significantly increased the duration of the interneuron bursts (dashed lines in Fig. 10A1; for statistics see Fig. 10A2 and legend). Because the frequency of the bursts was not significantly changed by CCAP (Fig. 9B), the increased burst duration resulted in a significantly increased duty cycle, defined as the fraction of the period occupied by the burst (Fig. 10B1; for statistics see Fig. 10B2 and legend). CCAP appeared to hyperpolarize the membrane potential of the motor neurons, but only by a modest 2–3 mV (Fig. 10A1), and had no noticeable effect on the motor neuron spike generation threshold. Together, these observations suggest an action of CCAP upstream of the motor neurons, at the level of the premotor interneurons.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Using immunohistochemistry, we found a characteristic pattern of CCAPli neurons, a larger type-1 cell and a smaller type-2 cell, in each neuromere of the thoracic ganglion of Callinectes. This pattern is consistent with observations in other arthropods examined to date (Dircksen 1998; Dircksen and Keller 1988; Pulver and Marder 2002). The type-1 cell in each neuromere typically gives rise to a fiber that crosses the midline and projects to a contralateral neurohemal structure where it terminates (Dircksen 1998). Projections from the type-2 cells remain within the CNS. Combining the immunohistochemistry with backfills of SN1, the segmental nerve of the first thoracic neuromere, we confirmed that the type-1 cell, but not the type-2 cell, of the first thoracic neuromere projects to the contralateral pericardial organ. It is thus likely that the entire plexus of CCAPli terminals in the POs derives from a small number of CCAP-expressing central neurons, one type-1 cell per thoracic neuromere.
Dopaminergic terminals in the POs similarly derive from a small number of central neurons, indeed just a single neuron—the L-cell—in each commissural ganglion (Cooke and Goldstone 1970; Fort et al. 2004; Wood and Derby 1996). After arborizing in the PO, however, each L-cell projects further, through the dorsal nerve, directly into the heart to innervate the cardiac ganglion (Fort et al. 2004). We found no such further projection of the CCAPli neurons and detected no CCAPli in the CG. We therefore propose that, whereas dopamine reaches the heart both as a blood-borne neurohormone liberated from the POs and a local neuromodulator released within the CG itself, CCAP acts on the heart in a purely neurohormonal fashion.
If CCAP acts solely as a circulating hormone, the entire cardiac system will be exposed to essentially the same concentration, i.e., that present in the blood (hemolymph) in the region of the heart. How high is this concentration likely to be? Stangier et al. (1988) quantified the CCAP contents of the Carcinus maenas POs and calculated that release of <1% of the total could achieve bulk hemolymph thresholds for cardioactivity (10–10 M) in that species. To date, the most comprehensive measurements of CCAP concentrations in crustacean hemolymph were made by enzyme immunoassay in a study examining the role of CCAP in ecdysis of the crab Carcinus maenas and the crayfish Orconectes limosus (Phlippen et al. 2000). In Carcinus, a peak concentration of 1.4 x 10–9 M was measured during active ecdysis, when the crab extricates itself from its former carapace. Several considerations, however, suggest that the heart will be exposed to substantially higher levels of CCAP at some point during the exuviation process: 1) hormones released from the POs are initially concentrated around the heart in the restricted volume of the pericardial sinus; 2) the hemolymph samples of Phlippen et al. (2000) were drawn from the hypobranchial sinus, a return vessel that is essentially the furthest point from the heart in the crab circulatory system (see McGaw and Reiber 2002); and 3) the half-life of CCAP in the hemolymph was shown to be very brief, about 6 min (Phlippen et al. 2000), making it difficult to sample the actual peak concentration. We suggest therefore that, although tonically the heart probably experiences only relatively low levels of circulating CCAP, transiently it may experience much higher levels.
Multiple physiological actions of CCAP
In Fig. 11 we have superimposed the CCAP dose–response relations from Figs. 6 and 9, and corresponding dose–response relations in the WH preparation, for the four principal CCAP effects that we quantified, on the contraction amplitude (A), motor neuron burst and contraction frequency (B), motor neuron burst duration (C), and the number of spikes per motor neuron burst (D), in the WH (red), S-IWH (green), and ICG (blue) preparations. Comparison of these dose–response relations strongly suggests that the effects do not reflect a single unified action of CCAP, but rather distinct actions exerted at separate loci in the cardiac system. Our view of how CCAP acts is summarized in the schema in Fig. 12: as the CCAP concentration increases (increasing density of background shading from left to right), different parts of the system are progressively modulated, with progressive changes in the contraction waveform (bottom).
|
Second, at somewhat higher concentrations (>10–8 M) in the WH and S-IWH preparations, CCAP increases the burst and contraction frequency (Fig. 11B). In our experiments, we observed no dissociation of bursts and contractions, such as might be caused, for example, by CCAP-induced myogenic rhythmicity (see Meyrand and Marder 1991). Because the bursts drive the contractions, we propose that CCAP accelerates the bursts by acting centrally on the CG (Fig. 12, right). An alternative possibility is that the increase in burst and so contraction frequency is a feedback consequence of the peripheral increase by CCAP of the contraction amplitude (see Sakurai and Wilkens 2003; and below). However, the effects of CCAP on amplitude and frequency can be dissociated. At the low CCAP concentrations and also with longer exposure to the higher concentrations (Fig. 7), the increase in amplitude occurs without the increase in frequency. Thus increased amplitude does not automatically produce increased frequency. We suggest, rather, that CCAP increases the frequency by acting centrally, on the CG. Within the CG, CCAP most plausibly acts on the premotor interneurons that initiate each burst (Hartline 1979; Tazaki and Cooke 1979).
A potential difficulty in locating this second action of CCAP in the CG is that the increase in burst frequency is not seen at all in the ICG preparation (Fig. 11B, blue). (This, indeed, strengthens the case for the peripheral feedback mechanism.) There is, however, a likely straightforward explanation. The burst frequency in the ICG preparation is significantly higher than that in the S-IWH (or, presumably, the WH) preparation. Indeed, the frequency in the ICG preparation under control conditions is about as high as that in the S-IWH preparation modulated by 10–6 M CCAP (see Fig. 11B, insets). This maximal frequency, around 30 BPM, may thus constitute a ceiling above which the cardiac ganglion cannot accelerate (see also Fort et al. 2004).
The increase in burst and contraction frequencies at CCAP concentrations >10–8 M is accompanied by a further large increase in contraction amplitude (Figs. 5B and 11A). In part this presumably reflects the more fully developed peripheral action of CCAP. Some part of the increase in contraction amplitude, however, is likely to be a consequence of the increased burst and contraction frequency (see Mahadevan et al. 2004; and below), arising, at least in part, from frequency-dependent facilitation of synaptic transmission at the cardiac neuromuscular junctions (Anderson and Cooke 1971; Mahadevan et al. 2004).
There is still a third primary action of CCAP. At concentrations >10–8 M, CCAP increases the burst duration and the number of spikes per burst (Fig. 11, C and D) in the ICG preparation (blue), but interestingly not in the S-IWH preparation (green). Because it is seen in the isolated ganglion, this is necessarily a central action of CCAP. The increase in burst duration, at least, is seen not only in the motor neurons but also upstream in the premotor interneurons (Fig. 10), where the locus of this action of CCAP may therefore lie. It is a separate primary action: it is not a consequence of the increase in burst frequency, which does not occur in the ICG preparation. Nor, conversely, is it a cause of the increase in burst frequency because it does not occur in the S-IWH preparation where the increase in burst frequency is robust.
Comparison with dopamine
In our previous work (Fort et al. 2004), we observed essentially the same constellation of effects with dopamine. With dopamine, however, the increase in contraction amplitude occurred only at higher concentrations (>10–8 M), those that also increased the contraction frequency. Consequently it was not clear whether the increase in amplitude reflected a separate primary action of dopamine or merely a secondary consequence, through the mechanism of frequency-dependent synaptic facilitation, of the increase in frequency. With CCAP, the increase in frequency similarly occurs at >10–8 M, but the increase in amplitude already occurs at >10–10 M (Fig. 6, A and B). Thus with CCAP, the increase in amplitude clearly can occur without the increase in frequency. Functionally, as the cardiac system becomes exposed to increasing concentrations of the modulator, CCAP will initially increase the contraction amplitude and only later increase the contraction frequency (Fig. 12), whereas dopamine will have rather the reverse effect. These differences in the sequential expression of the two effects will be enhanced by the different concentration profiles that are likely to be achieved by the different modes of delivery of the two modulators. Any activity of the L-cell that releases hormonal dopamine from the POs will also release dopamine in the CG to achieve (presumably) higher local concentrations there and so bring about a coordinated, simultaneous expression of all of the effects of dopamine. The effects of CCAP, in contrast, will probably be expressed differentially and in a graded fashion. Low activity of the CCAP-containing neurons will sufficiently elevate hormonal levels of CCAP to express just the peripheral action to increase contraction amplitude. Higher activity of the neurons, higher levels of CCAP, and the added expression of the central action on contraction frequency may occur only transiently during periods of intense demand for increased cardiac output in behaviors such as ecdysis.
Dopamine and CCAP thus converge on a common set of effects, but they activate them in a different balance. This is a common theme in neuromodulation (Brezina and Weiss 1997; Nusbaum et al. 2001), well documented for example in the crustacean stomatogastric system (Marder and Thirumalai 2002; Skiebe 2001) and in the accessory radula closer neuromuscular system of Aplysia, where it was proposed that the different balance of effects of convergent, simultaneously acting modulators serves to maximize the range of responses of the system to diverse behavioral demands (Brezina et al. 1996, 2003a, 2005).
Modulation of the integrated CPG–effector system
A striking feature of our findings is that the cardiac ganglion behaves differently in isolation (in the ICG preparation) than when it remains embedded within the heart musculature (in the S-IWH and WH preparations). This is true with respect to the effects of CCAP (Fig. 11), the effects of dopamine (Fort et al. 2004), and the frequency and other parameters of the activity even of the unmodulated system (Fig. 11B, insets; Fort et al. 2004). We propose that this is because the intact system incorporates multiple feedforward and feedback mechanisms that integrate its activity (as schematized by the arrows in Fig. 12), which are disrupted when the ganglion is isolated. One feedforward mechanism—the facilitation at the cardiac neuromuscular junctions that increases contraction amplitude as burst frequency increases (Anderson and Cooke 1971; Mahadevan et al. 2004)—was in previous mentioned text. There are several additional mechanisms.
First, there are local mechanisms operating within the cardiac ganglion itself. In particular, the burst duration and the number of spikes per burst are often found to decrease as the burst frequency increases (Tazaki and Cooke 1990; but see Benson 1980; Hokkanen 2000). We propose that this mechanism may explain why, when CCAP acts to increase the burst duration and the number of spikes per burst, a net effect of CCAP on these two parameters is seen in the ICG preparation (Fig. 11, C and D, green) but not in the S-IWH preparation (blue). In the S-IWH preparation, unlike in the ICG preparation, CCAP also increases the burst frequency, which will tend to decrease the spike duration and the number of spikes per burst, oppose the increasing action of CCAP, and keep these two parameters approximately constant.
Second, two recent studies suggested that there are feedback mechanisms by which the state of the heart muscle can modify the parameters of the motor pattern produced by the cardiac ganglion. In the isopod Ligia, Sakurai and Wilkens (2003) showed that passive stretch or active contractions of the muscle alter the timing and frequency of the CG bursts. This appears to be mediated by beat-to-beat modification of the membrane voltage of the CG neurons, most likely through the mechanosensitive dendrites that the neurons extend into the heart muscle (Sakurai and Wilkens 2003; see Alexandrowicz 1932). In lobster hearts, Mahadevan et al. (2004) and Goy (2005) proposed that nitric oxide (NO) produced by the active heart muscle acts to reduce the CG burst frequency. The presence of high levels of NO synthase in the heart of Cancer productus and the sensitivity to NO of neurons in the Cancer CG (Scholz et al. 2002) suggest that the NO-mediated feedback may also operate in the cardiac systems of crabs. If such feedback mechanisms do indeed operate in Callinectes, they might explain two of our findings in this study. Disruption of negative feedback might explain why the cardiac ganglion bursts faster in isolation than in the intact heart (Fig. 11B, insets). Negative feedback could also be responsible for the secondary decrease in contraction frequency during a maintained CCAP-induced elevation of contraction amplitude (Fig. 7). The gradual time course of this decrease makes it more likely to be mediated by a diffusible messenger such as NO than by direct mechanosensitivity.
Dynamic modes and system complexity
The crustacean cardiac system is one of the simplest CPG–effector systems. Nevertheless, it is clear from the preceding discussion that it has a sufficient number of interacting elements to produce complex dynamical behaviors. That these are indeed produced was suggested previously. For example, studying just the unmodulated cardiac ganglion—not even the complete system—Benson (1980) provided experimental data on the relationship between the durat
