| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Neuroscience, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
Submitted 1 August 2006; accepted in final form 28 September 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
; Marder and Bucher 2001; Marder et al. 2005; Ramirez et al. 2004). Network modulation results from the local neuronal release of modulatory transmitters and from circulating hormones. Distinct modulatory inputs to a neuronal network can work independently, but there are likely times when their influences overlap. In such cases, one neuromodulator can alter the influence of the co-active one, a condition termed metamodulation (Katz and Edwards 1999). There are, however, few examples documenting the consequences of metamodulation for network activity. Thus far, changes resulting from metamodulation include the generation of distinct motor patterns and/or behaviors and limits on the maximal effect of each co-released modulator on the assayed motor pattern relative to when they act separately (Crisp and Mesce 2004; Edwards et al. 2002; McLean and Sillar 2004; Svensson et al. 2001; Wood et al. 2000). Mechanistically, a metamodulator can act both in series and in parallel with its target modulator (Edwards et al. 2002; McLean and Sillar 2004; Svennson et al. 2001; Wood et al. 2000).
We are addressing the consequences of metamodulation at the circuit level, using the stomatogastric nervous system (STNS) of the crab Cancer borealis (Marder and Bucher 2007; Nusbaum and Beenhakker 2002). The STNS contains two well-defined central-pattern-generating (CPG) circuits in the unpaired stomatogastric ganglion (STG), including the gastric mill (chewing) and pyloric (filtering of chewed food) circuits. These circuits are modulated by a set of projection neurons that innervate the STG from the paired commissural ganglia (CoGs) and unpaired esophageal ganglion (OG). There are also identified peptide hormones that influence these two circuits (Marder and Bucher 2007). Many of these hormones are released into the cardiac chamber from the pericardial organs (POs), from where they are pumped through the ophthalmic artery to influence the STG, which is located within this artery (Li et al. 2003; Skiebe 2001; Turrigiano and Selverston 1990). One of these neuromodulators, crustacean cardioactive peptide (CCAP), influences the STG only as a circulating hormone (Billimoria et al. 2005; Marder and Bucher 2007). CCAP modulates the pyloric rhythm in C. borealis by activating a voltage-dependent, depolarizing current in several pyloric circuit neurons (Swensen and Marder 2000, 2001; Weimann et al. 1997).
In this study, we assessed the CCAP influence on gastric mill rhythms in the isolated STNS (Beenhakker and Nusbaum 2004; Coleman and Nusbaum 1994). CCAP application to the STG did not activate a sustained gastric mill rhythm, but it directly excited several gastric mill neurons including the gastric mill CPG neurons lateral gastric (LG) and interneuron 1 (Int1). CCAP slowed the gastric mill rhythms elicited by modulatory commissural neuron 1 (MCN1) and by coactivation of MCN1 and commissural projection neuron 2 (CPN2), even when applied at low concentrations (10–10 M). It also reduced the threshold firing frequency at which MCN1 activated the gastric mill rhythm. Further, only some gastric mill targets of CCAP exhibited altered activity during these CCAP-modulated rhythms, and the activity of one nontarget was also altered. These results indicate that a peptide hormone can facilitate and alter neuronal modulation of motor circuit activity, thereby expanding the available output patterns of rhythmically active motor circuits.
Some of these data were published in abstract form (Kirby and Nusbaum 2003, 2004).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Jonah crabs (C. borealis) were obtained from commercial suppliers (Commercial Lobster and Seafood, Boston, MA; Marine Biological Laboratory, Woods Hole, MA). The crabs were housed in commercial tanks containing recirculating, aerated and filtered artificial seawater (10–12°C). Before dissection, the crabs were cold-anesthetized by packing them in ice for 
30 min. The foregut was then removed and maintained in chilled physiological saline while the STNS was dissected from it.
Solutions
Under most experimental conditions, the isolated STNS was maintained in physiological saline (5–10 ml) containing (in mM) 439 NaCl, 26 MgCl2, 13 CaCl2, 11 KCl, 10 Trizma base, and 5 maleic acid (pH 7.4–7.6). In some experiments, transmitter release was eliminated using saline that contained a reduced (0.1x normal) concentration of Ca2+ plus a compensatory addition of Mn2+ ("low-Ca2+ saline") (Blitz and Nusbaum 1997). Low-Ca2+ saline contained (in mM) 439 NaCl, 26 MgCl2, 1.3 CaCl2, 11.7 MnCl2, 11 KCl, 10 Trizma base, and 5 maleic acid (pH 7.4–7.6).
Electrophysiology
All experiments were conducted using the completely isolated STNS (Fig. 1). The preparation was pinned down in a saline-filled silicone elastomer (Sylgard)-lined petri dish (Sylgard 184, KR Anderson, Santa Clara, CA) and superfused continuously (7–12 ml/min) with physiological saline and/or low-Ca2+ saline (10–12°C). Intra- and extracellular recordings of STNS neurons were made using routine methods for this system (Beenhakker and Nusbaum 2004). Glass microelectrodes (15–30 M
) filled with 4 M K-acetate plus 20 mM KCl or 0.6 M K2SO4 plus 10 mM KCl were used for intracellular recording. Intracellular recordings were made with Axoclamp 2 amplifiers (Molecular Devices, Sunnyvale, CA). Intracellular current injections were performed in single-electrode discontinuous current-clamp (DCC) mode with sampling rates of 2–5 kHz. To facilitate intracellular recordings, the STNS ganglia were desheathed and visualized with light transmitted through a dark-field condenser (Nikon, Tokyo, Japan).
|
The influence of the neuropeptide CCAP on individual neurons within the STG was tested by direct application of the peptide onto the desheathed STG neuropil during superfusion of low-Ca2+ saline. For these experiments, a recording electrode was filled with a CCAP solution (10–4 or 10–5 M) in low-Ca2+ saline to be consistent with the bath condition. The tip of the electrode was then broken to create an electrode resistance of 1–1.5 M. The CCAP solution was forced from the pipette using a Picospritzer II pressure ejection device (General Valve, Fairfield, NJ). CCAP was applied at 4–6 psi for a duration of 0.5–1.0 s during continual superfusion of low-Ca2+ saline. In between CCAP puffer applications, the CCAP-containing pipette was maintained at a distance away from the STG to prevent the possibility of leak-mediated CCAP actions on STG neurons.
To determine the effects of CCAP on the gastric mill rhythm, a CCAP solution (10–6–10–11 M) was superfused across the STG. The CCAP solution flowed for 10–15 min before any experimental manipulations, ensuring that the CCAP solution had completely displaced the normal saline solution. This time period was verified by determining the time taken for the effects of CCAP on the pyloric circuit to reach steady state (Weimann et al. 1997).
Gastric mill rhythms were elicited using one of two methods. First, we used extracellular stimulation of one or both dorsal posterior esophageal nerves (dpons), in preparations with the CoGs still connected with the STG, to activate the mechanosensory ventral cardiac neurons (VCNs) (Beenhakker et al. 2004). The VCNs trigger a long-lasting activation of MCN1 and CPN2, producing a long-lasting MCN1/CPN2-elicited gastric mill rhythm (Fig. 2A) (Beenhakker and Nusbaum 2004; Beenhakker et al. 2004). Second, we selectively activated MCN1 and thereby evoked MCN1-elicited gastric mill rhythms by extracellular stimulation of one or both inferior esophageal nerves (ions), after their transection to separate them from the CoGs (Fig. 2C) (Bartos and Nusbaum 1997; Bartos et al. 1999). Each ion stimulation (1-ms stimulus duration) elicits a single action potential in MCN1 (Bartos and Nusbaum 1997). Activation of MCN1 action potentials was confirmed by maintaining an intracellular recording of the LG neuron, which receives an electrical EPSP from each MCN1 action potential (Coleman et al. 1995).
|
Individual STNS neurons were identified by their axonal pathways, activity patterns and interactions with other neurons (Beenhakker and Nusbaum 2004; Blitz et al. 1999; Weimann et al. 1991). Data were collected onto a chart recorder (Models No. MT 95000 and Everest: Astromed) and, in parallel, digitized (
5 kHz) and collected onto a PC computer using data-acquisition/analysis tools (Spike2, Cambridge Electronic Design, Cambridge, UK). Figures were made from Spike2 files incorporated into Adobe Photoshop (Adobe, San Jose, CA) and Powerpoint graphics programs (Microsoft, Seattle, WA).
Data analysis was facilitated with a custom-written program for Spike2 that determines the activity levels and burst relationships of individual neurons (freely available at http://www.uni-ulm.de/~wstein/spike2/The_Crab_Analyzer.s2s). Unless otherwise stated, each datum in a data set was derived by determining the average of 10 consecutive gastric mill- or pyloric rhythm-timed impulse bursts. Briefly, burst duration was defined as the duration (s) between the onset of the first and last action potential in an impulse burst (e.g., see LG burst in Fig. 2C). The intraburst firing rate was determined by dividing the number of action potentials minus one by the burst duration. The cycle period of gastric mill and pyloric rhythms was determined by calculating the duration between the onset of two successive LG neuron bursts and two successive pyloric dilator (PD) neuron bursts, respectively (e.g., see Fig. 2C). The burst relationship among gastric mill neurons (phase relationships) was determined for normalized gastric mill cycles in which the onset and offset of a cycle was determined by the onset of successive LG neuron bursts. Specifically, we determined the mean point in a normalized gastric mill cycle at which the burst onset and offset occurred for each gastric mill neuron. To identify these points, we divided the duration from each cycle onset to burst onset (and offset) by its associated cycle period.
In the case of STG neurons (Int1, VD, IC, MG) whose gastric mill rhythm-timed activity was subdivided into pyloric-timed bursts, we determined their gastric mill-related burst duration as the duration of their entire period of activity (i.e., the cumulative duration of all of their pyloric-timed bursts) during a gastric mill cycle. In contrast, for these neurons we analyzed their number of action potentials per burst as well as their intraburst firing frequency based on their individual pyloric-timed bursts.
We determined the MCN1 firing frequency threshold for eliciting the gastric mill rhythm by tonically, and simultaneously, stimulating both ions (60- to 300-s train duration) at a series of constant instantaneous stimulus frequencies in the range of threshold for rhythm activation (2–8 Hz). There was a minimum lag of 1 min between successive stimulus trains. It is noteworthy that the MCN1-elicited gastric mill rhythm consistently stops immediately when MCN1 stimulation is terminated (Bartos and Nusbaum 1997).
Over the course of any single experiment, repeated activation of the gastric mill rhythm generally resulted in a progressively slower rhythm (see RESULTS). This occurred even when the sequentially activated gastric mill rhythms were each elicited under the same condition, such as during normal saline superfusion. This progressive slowing contrasted with the pyloric rhythm, which is often continuously active for hours in the isolated preparation at a relatively constant cycle period, presumably due to the presence of an endogenously oscillatory pacemaker neuron (Hooper and Marder 1987; Miller and Selverston 1982). There is no such pacemaker neuron within the gastric mill CPG (Bartos et al. 1999; Coleman et al. 1995).
Due to the progressive slowing of the gastric mill rhythm with repeated activation, it was difficult to generate matched control rhythms before and after CCAP application. It was, however, routinely possible to elicit two equivalent gastric mill rhythms (saline 1:saline 2 or S1:S2) with intervals that matched those of the pre-CCAP application and during-CCAP application rhythms (supplementary Tables 1–3;1 see RESULTS). To ensure that any changes that occurred in the gastric mill rhythm during CCAP superfusion were due to the presence of the peptide and not to the progressive slowing of the rhythm within each preparation, we compared two successive gastric mill rhythms during saline superfusion (S1, S2) with intervals comparable to our experimental interval (S, CCAP). We performed these controls in separate preparations from those in which CCAP was applied, so that the rhythms could be elicited at comparable times after the start of each experiment.
To most appropriately represent the results of our experiments, we used a two-stage process for determining which analyzed parameters were significantly different between the control and experimental conditions. In stage 1, we used a paired Student's t-test to obtain each within-group comparison (S1 vs. S2; S vs. CCAP). If, for any particular parameter (i.e., cycle period) a significant change was found in the S:CCAP data set, we progressed to stage 2.
The stage 2 test was used to determine whether a change in activity induced by the addition of CCAP was significantly different from the condition when saline was applied during consecutively elicited gastric mill rhythms. We determined the difference between S1 and S2 for each experiment (S2 minus S1) to measure the change in a particular gastric mill parameter induced by no intervention. For each experiment where saline was followed by an application of CCAP, we again determined the difference between values (CCAP minus saline). We then performed an unpaired Student's t-test to determine whether the change induced by CCAP across preparations was significantly different from the change, if any, induced in the experiments when saline was applied during both gastric mill rhythms. In some cases, where noted, we instead used the Mann-Whitney rank sum test in our stage 2 analysis. We used this alternative test when the variance between the data sets was unequal. In the text, any result for which a significant difference is noted during CCAP application represents the results of our stage 2 analysis.
To determine whether CCAP consistently increased the gastric mill cycle period during MCN1 stimulation, regardless of the value of the control cycle period, we used a sliding window analysis of the complete data set. Specifically, we compared the control (S1, S2) and experimental (S, CCAP) data within successive 4-s-duration windows. After each set of comparisons, these 4-s-duration windows were shifted 0.5 s to the right, and the analysis was performed again, until all possible windows were analyzed.
For each window, we used our stage 2 analysis (see preceding text) to compare the change induced in the gastric mill cycle period by the presence of CCAP with the change induced in the S1:S2 condition. The result (significant change or no change) was then assigned to the value at the center of each window (i.e., a significant change was assigned to 6 s for the 4-s-duration window that spanned cycle periods of 4–8 s). We note the distinction between significant and nonsignificant changes at the center of the 4-s-duration window encompassing the largest values of cycle period that showed a significant difference between S:CCAP and S1:S2.
We did not see any change in the results using time increments <0.5 s (
0.25 s; not shown). We also obtained the same results with 4- and 5-s-duration windows (not shown). We did not use a 3-s-duration window because it often resulted in too few data points to enable an accurate determination of significant changes.
Statistical analyses were performed with SigmaStat 3.0 and SigmaPlot 8.0 (SPSS, Chicago, IL). Data are expressed as means ± SE except where explicitly noted to be expressed as means ± SD.
Gastric mill model
We implemented a computational model modified as indicated in the following text from an existing conductance-based model of the gastric mill circuit (Beenhakker et al. 2005; Nadim et al. 1998). We retained all aspects of the model implemented by Beenhakker et al. (2005), including modeled versions of the LG, Int1, and MCN1 neurons having multiple compartments separated by an axial resistance, with each compartment possessing intrinsic and/or synaptic conductances, as documented originally by Nadim et al. (1998). The voltage-dependent trajectory of the MCN1 input to LG was based on the neuropeptide-activated current in pyloric neurons in C. borealis (Golowasch and Marder 1992; Swensen and Marder 2000). The only parameters that were altered from the model version presented in Beenhakker et al. (2005) were an increase in the maximal conductance value for the inhibitory synapse from Int1 to LG, from the previously used value of 1.4 to 2.1 nS, and elimination of the gastro-pyloric receptor (GPR) synapses onto MCN1 and Int1 in the STG.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
The gastric mill rhythm (cycle period: 5–20 s) controls the rhythmic protraction and retraction chewing movements of the teeth in the gastric mill stomach compartment (Heinzel et al. 1993). Thus this rhythm is composed of alternating impulse bursts of protractor and retractor motor neurons, plus a single retractor phase interneuron (Int1; Fig. 2) (Marder and Bucher 2007). Several gastric mill neurons also exhibit rhythmic impulse bursts that are time-locked to the pyloric rhythm, which is a faster rhythm (cycle period: 0.5–2 s) that controls the filtering of chewed food in the pylorus, immediately posterior to the gastric mill (Weimann et al. 1991). The neurons exhibiting this dual rhythmic pattern during the gastric mill rhythms studied in this paper include the inferior cardiac (IC), medial gastric (MG), and ventricular dilator (VD) neurons as well as Int1 (e.g., Fig. 2, A and C). The LG neuron also exhibits pyloric-timed activity during a distinct version of the gastric mill rhythm (Wood et al. 2004). All gastric mill neurons are present as single neurons in the C. borealis STG except the protractor gastric mill (GM) neuron, which is present as four apparently equivalent neurons.
There are several distinct versions of the gastric mill rhythm in C. borealis, including one driven by selective activation of the projection neuron MCN1 and one triggered by the mechanosensory VCNs, which activate the rhythm via their excitation of MCN1 and CPN2 (Fig. 2) (Beenhakker and Nusbaum 2004; Beenhakker et al. 2004; Coleman and Nusbaum 1994). These two rhythms share several features, including the presence of alternating protractor and retractor phases (Fig. 2, A and C). Both rhythms also include rhythmic alternating bursting of the CPG neurons LG and Int1 and, in both cases, the retractor neurons VD and DG are coactive with Int1. Distinctions between these rhythms include the relative timing of the impulse bursts in the IC and MG neurons as well as the level of participation of the GM and anterior median (AM) neurons. For example, as shown in Fig. 2, IC neuron activity is distinct during the MCN1- and VCN-elicited gastric mill rhythms while GM neuron bursting occurs only during the VCN-elicited rhythm.
The projection neuron MCN1 excites all of the gastric mill neurons (Fig. 2B) (Bartos and Nusbaum 1997; Coleman and Nusbaum 1994; Coleman et al. 1995; Stein et al. 2007). However, its influence on the GM and AM neurons is modest, particularly when they are inactive prior to MCN1 stimulation. Consequently, GM and AM are generally silent or only weakly active during the MCN1-elicited gastric mill rhythm. During this version of the gastric mill rhythm, the core CPG includes the gastric mill neurons LG and Int1 plus the STG terminals of MCN1 (Bartos et al. 1999; Coleman et al. 1995).
During the VCN-triggered gastric mill rhythm, MCN1 is active during both phases but it fires tonically during protraction and is pyloric-timed during retraction (Beenhakker and Nusbaum 2004). The projection neuron CPN2 is also tonically active during the protractor phase but exhibits reduced or no activity during retraction (Beenhakker and Nusbaum 2004; Norris et al. 1994). CPN2 excites the LG and GM neurons but inhibits the IC and MG neurons, thereby shifting the activity of the latter two neurons during the VCN-triggered rhythm (Beenhakker and Nusbaum 2004; Norris et al. 1994) (Fig. 2, A and B). CPN2 also inhibits the retractor neuron DG (Norris et al. 1994) (Fig. 2B).
CCAP directly excites a subset of gastric mill neurons
Swensen and Marder (2000, 2001) showed previously that three of the seven pyloric circuit neurons, including the anterior burster (AB), lateral pyloric (LP) and IC neurons, are direct targets of CCAP in C. borealis. We determined whether CCAP had any direct actions on the gastric mill circuit neurons. To this end, we pressure ejected CCAP (10–4 or 10–5 M) onto the desheathed STG neuropil under conditions where neurotransmitter release was suppressed (low-Ca2+ saline, see METHODS) while recording intracellularly from each gastric mill neuron.
In low-Ca2+ saline, CCAP excited a subset of the gastric mill neurons. For example, as shown in Fig. 3, brief (1 s) pressure application of CCAP in low-Ca2+ saline caused a depolarization and action potential burst in the LG neuron (n = 19/19) and Int1 (n = 9/11). For both of these neurons, their response to CCAP outlasted its application by many seconds (Fig. 3, A and B). There was also an excitatory response in the MG (n = 11/11), IC (n = 13/13), and AM (n = 7/7) neurons, although the MG neuron response tended to be weaker than that of the other neurons. In contrast, the DG (n = 13/13; Fig. 3C) and GM (n = 20/20) neurons never responded to CCAP application in low-Ca2+ saline.
|
; Coleman and Nusbaum 1994), that there was no change in the membrane potential at the STG terminals of the MCN1 and CPN2 projection neurons in response to CCAP application (MCN1: pre-CCAP, –45.7 ± 7.8 (SD) mV; CCAP, –46.7 ± 7.4 mV; n = 6, P = 0.20; CPN2: pre-CCAP, –56.0 ± 8.8 mV; CCAP, –54.0 ± 7.4 mV; n = 4, P = 0.18). Additionally, CCAP never elicited action potentials in either projection neuron (MCN1: n = 7; CPN2, n = 6). The LG, IC, and MG neurons are electrically coupled in C. borealis (M. P. Beenhakker, M. S. Kirby, M. P. Nusbaum, unpublished data). Therefore to obtain a more accurate assessment of whether each of them was a direct target of CCAP, we performed additional experiments in which we hyperpolarized two of them during any particular CCAP application. These current injections consistently suppressed activation of spiking in the hyperpolarized neurons in response to CCAP but never caused a hyperpolarization of >5 mV in the uninjected neuron (n = 12/12). When the LG and IC neurons were hyperpolarized, the MG neuron failed to respond to pressure applied CCAP (n = 9/9). Moreover, under these conditions MG remained unresponsive even when it was depolarized to spike threshold via intracellular current injection (n = 9). In contrast, when LG and MG were hyperpolarized, the IC neuron consistently responded as vigorously as when no neurons were hyperpolarized (n = 10/10). Similarly, the LG neuron exhibited an unchanged excitatory response when both the IC and MG neurons were hyperpolarized (n = 6/6). Thus four of the eight types of gastric mill neurons (LG, IC, Int1, AM) appeared to be direct targets of CCAP. The remaining neurons (DG, VD, MG, GM) as well as the STG terminals of the projection neurons MCN1 and CPN2 were not directly responsive to this peptide. Last, no gastric mill neurons ever responded to CCAP application in low-Ca2+ saline with an inhibitory response.
It remains possible that we obtained a false negative result regarding the neurons that were unresponsive to CCAP in low-Ca2+ saline because CCAP might also influence a Ca2+ current or Ca2+-sensitive current in these neurons. Such a current would have been minimal or absent due to the 10-fold reduced Ca2+ concentration in the low-Ca2+ saline. Although the presence of an additional, Ca2+-sensitive current remains a possibility, previous work indicated that CCAP influenced only a single ionic current in C. borealis pyloric neurons, and this current persisted in low-Ca2+ saline (Golowasch and Marder 1992; Swensen and Marder 2000).
CCAP elicits a transient, incomplete gastric mill rhythm
The reciprocally inhibitory neurons LG and Int1 are key gastric mill CPG neurons (Bartos et al. 1999; Coleman et al. 1995). Because both of these neurons were directly excited by CCAP application, we determined whether application of this peptide in normal saline would activate the gastric mill rhythm. CCAP superfusion (10–6 or 10–7 M) to the isolated STG often elicited a gastric mill rhythm (n = 30/38), but this rhythm never included the DG retractor neuron nor did it persist for the duration of the application. This rhythm did, however, include coordinated bursting of several gastric mill neurons, including LG and Int1 (Fig. 4, A and B). These CCAP elicited rhythms persisted for 7.4 ± 3.5 (SD) min (n = 19). Generally this rhythm gave way to intermittent spiking or inactivity in LG and the return to exclusively pyloric-timed activity in the other participating neurons before the end of each CCAP application (Fig. 4C). During these CCAP-elicited rhythms, not every gastric mill neuron participated. As mentioned in the preceding text, the DG neuron was never activated under these conditions (Fig. 4, B and C). The lack of DG participation in the CCAP-elicited rhythm was not a consequence of damage during dissection, because DG was effectively activated subsequently during the MCN1-elicited gastric mill rhythm (Fig. 4D).
|
As noted in METHODS, we found that repeated activation of the gastric mill rhythm with sufficiently long inter-stimulus intervals resulted in increased gastric mill cycle periods even without CCAP application. For example, in a sampling of 15 preparations where we superfused CCAP (10–7 M), the initial MCN1-elicited gastric mill cycle period during saline superfusion was 9.6 ± 0.6 s. After a 1-h washout of CCAP, the gastric mill rhythm had slowed (cycle period: 10.7 ± 0.9 s, P < 0.001, paired Student's t-test). Similarly, when we elicited consecutive gastric mill rhythms in normal saline with intervals (1.5 h) that mimicked the duration of the application and washout of CCAP, the gastric mill cycle period of the postinterval rhythm was prolonged relative to the preinterval rhythm (saline, pre-interval: 9.5 ± 1.3 s; saline, post-interval: 11.9 ± 1.9 s; n = 8; P < 0.05, paired Student's t-test).
In contrast to the gradual slowing of gastric mill rhythms during saline superfusion that occurred across intervals of 1 h or more, the MCN1-elicited rhythms in saline were equivalent when they had an interval comparable to the one occurring between a control rhythm and the rhythm elicited in CCAP. Comparison of these matched controls revealed that none of the studied parameters differed between consecutively elicited gastric mill rhythms in saline (supplementary Table 1, Table 2). For the VCN-triggered gastric mill rhythms, the only parameter that was distinct between the consecutively elicited rhythms in saline was the burst onset phase of the GM neuron (n = 10, P < 0.05; Fig. 5, supplementary Table 3). For example, there was no change in the VCN-triggered gastric mill cycle period (S1: 9.8 ± 0.6 s; S2: 9.8 ± 0.7 s; n = 11, P = 0.98, paired Student's t-test).
|
We analyzed the response of the gastric mill rhythm triggered by mechanosensory (VCN neurons)-mediated coactivation of MCN1 and CPN2 to CCAP application (Fig. 2A) (Beenhakker and Nusbaum 2004; Beenhakker et al. 2004). The VCN-triggered rhythm requires the CoGs to remain connected to the STG because this rhythm requires properly timed synaptic feedback from STG neurons onto MCN1 and CPN2 (Beenhakker and Nusbaum 2004). Consequently, in these experiments, we used a petroleum jelly wall to separate the STG from the anterior ganglia and selectively superfused CCAP onto the STG while the anterior compartment was continually superfused with normal saline.
CCAP slowed the MCN1/CPN2-elicited gastric mill rhythm by prolonging the protractor (LG burst) phase (Fig. 5; supplementary Table 3). This was a consistent effect across preparations, regardless of the control cycle period which ranged from 7 to 14 s in different preparations (Fig. 5B). During saline superfusion, the mean gastric mill cycle period was 10.4 ± 0.5 s, whereas during superfusion of CCAP (10–7 M), the cycle period increased to 11.9 ± 0.6 s (n = 15; P < 0.001; Fig. 5C). In parallel, the LG burst duration was consistently increased by the presence of CCAP (10–7 M; Fig. 5C; supplementary Table 3). This increase in LG neuron burst duration was accompanied by increases in the number of LG neuron action potentials per burst and LG intraburst firing frequency (Fig. 5C; supplementary Table 3).
There were additional CCAP (10–7 M)-mediated changes in the activity of gastric mill neurons during the VCN-triggered rhythm. These included increased activity in the DG and AM neurons (supplementary Table 3). The DG neuron, despite not being a direct target of CCAP, exhibited an increase in its burst duration (P < 0.05, n = 13). The AM neuron, which was directly excited by CCAP, also showed an increased burst duration (P < 0.05, n = 7) and number of action potentials per burst (P < 0.01, n = 7). The activity of both DG and AM were unchanged in the sequentially triggered, saline control rhythms (supplementary Table 3). The activity (burst duration, number of spikes per burst, and intraburst firing frequency) of the other gastric mill neurons was unchanged by the presence of CCAP during the VCN-triggered gastric mill rhythm (supplementary Table 3).
Finally, CCAP (10–7 M) superfusion changed several aspects of the phase relationships among the gastric mill neurons during the VCN-triggered rhythm (supplementary Table 3). Phase relationships represent the fraction of a normalized cycle during which each neuron fires its burst of action potentials (see METHODS). Because the normalized gastric mill cycle is designated as extending from the onset of successive LG neuron bursts, the protractor neurons are generally active during approximately the first half of each cycle while the retractor neurons fire during the latter half of the cycle (i.e., Fig. 2). During the VCN-triggered rhythm, CCAP (10–7 M) superfusion altered the active phase of one protractor neuron (MG) and one retractor neuron (DG). Specifically, the onset phase of MG neuron activity was significantly delayed by CCAP superfusion (P < 0.05, n = 8) with no corresponding change in MG offset phase (P = 0.11, n = 8). Further, CCAP delayed both the onset (P < 0.05, n = 13) and offset phase (P < 0.05, n = 13) of DG activity.
CCAP reduces the MCN1 firing frequency threshold for gastric mill rhythm activation
Because CCAP effectively influenced the VCN-triggered gastric mill rhythm when applied selectively to the STG, we aimed to study these influences in more detail in a further reduced preparation. To this end, we removed the CoGs by transecting the ions and superior esophageal nerves (sons) and studied the CCAP actions on the MCN1-elicited gastric mill rhythm by selectively stimulating MCN1 extracellularly (see METHODS).
CCAP reduced the threshold MCN1 firing frequency for eliciting the gastric mill rhythm (Fig. 6, A and B). With the isolated STG superfused with normal saline, the mean threshold MCN1 firing frequency that elicited the gastric mill rhythm was 5 Hz (4.7 ± 0.4 Hz; n = 10). During CCAP superfusion (10–7 M), this threshold firing rate was reduced to 3 Hz (2.9 ± 0.2 Hz; n = 10; P < 0.01; Fig. 6B). This effect of CCAP on the threshold MCN1 firing frequency for activating the gastric mill rhythm was effective even at CCAP concentrations as low as 10–10 M [saline: 5.4 ± 0.3 Hz; CCAP (10–10 M): 4.3 ± 0.3 Hz; P < 0.05, n = 10; Fig. 6B]. At lower concentrations (10–11 M), CCAP had no effect on this parameter (P = 0.60, n = 7). In control experiments in which CCAP was not present during the second activation of the gastric mill rhythm, there was no change in the threshold firing frequency at which MCN1 elicited the gastric mill rhythm (saline 1: 6.6 ± 0.6 Hz; saline 2: 6.2 ± 0.4 Hz; P = 0.20, paired Student's t-test, n = 9; Fig. 6B).
|
Previous work showed that CCAP also influences the pyloric rhythm (Weimann et al. 1997), and the pyloric rhythm regulates the MCN1-elicited gastric mill rhythm (Bartos et al. 1999). Therefore the possibility existed that the CCAP influence on the gastric mill rhythm was a secondary consequence of the peptide action on the pyloric rhythm. We tested this possibility by superfusing the STG with CCAP (10–7 M) while suppressing the pyloric rhythm via intracellular hyperpolarizing current injection into the two PD neurons which, along with the AB neuron, are members of the electrically coupled pyloric pacemaker group (Fig. 7). As shown previously, the MCN1-elicited gastric mill rhythm was slower when the pyloric rhythm was suppressed (Bartos et al. 1999).
|
With the pyloric rhythm suppressed, CCAP (10–7 M) still altered the MCN1-elicited gastric mill rhythm (Fig. 7). Under this condition, the CCAP-elicited changes included prolonging the cycle period of the rhythm (saline: 18.8 ± 1.2 s; CCAP: 22.6 ± 1.4 s; n = 10, P < 0.01) and the LG neuron burst duration (saline: 9.7 ± 0.8 s; CCAP: 11.4 ± 1.0 s; n = 10, P < 0.05), and increasing the number of LG spikes per burst (saline: 88.9 ± 10.1 spike; CCAP: 111.1 ± 12.5 spike; n = 10, P < 0.01). There were also increases in the DG neuron burst duration (saline: 9.3 ± 1.0 s; CCAP: 12.0 ± 1.0 s; n = 8, P < 0.01) and the number of DG spikes per burst (saline: 113.9 ± 13.0 spike; CCAP: 136.9 ± 12.9 spike; n = 8, P < 0.05).
CCAP modifies the MCN1-elicited gastric mill rhythm
We also examined the MCN1-elicited gastric mill rhythm response to CCAP application when the pyloric rhythm remained active. In these experiments, we delayed MCN1 stimulation until after any CCAP-elicited gastric mill rhythmicity had ended. At these times, CCAP influenced the MCN1-elicited gastric mill rhythm in a manner comparable to its effects on the VCN-triggered rhythm, including prolonging the gastric mill cycle period (Fig. 8). Across preparations, the gastric mill cycle period increased from 9.6 ± 0.6 s during saline superfusion to 10.5 ± 0.5 s in the presence of CCAP (10–7 M; n = 27, P < 0.05). However, there was no significant change in the duration of either the protractor (saline: 6.4 ± 0.5 s; CCAP: 7.1 ± 0.4 s; n = 27, P = 0.07) or retractor phase (saline: 3.2 ± 0.2 s; CCAP: 3.4 ± 0.2 s; n = 27, P = 0.15) of this rhythm.
|
).
Our current understanding of the mechanisms underlying the generation of the MCN1-elicited gastric mill rhythm include the periodic buildup and decay of a slowly developing excitation of LG by MCN1 (Fig. 9) (Beenhakker et al. 2005; Coleman et al. 1995). This slow excitation, which is mediated by the MCN1 peptide co-transmitter C. borealis tachykinin-related peptide Ia (CabTRP Ia) (Stein et al. 2007; Wood et al. 2000), appears to build continuously during each retractor phase (LG interburst) of the gastric mill rhythm until it overcomes the contemporaneous Int1 inhibition of LG and enables LG burst onset. During the LG burst, the excitatory response in LG from MCN1 decays due to LG inhibition of the STG terminals of MCN1, until the LG burst terminates and the next retractor phase begins (Fig. 9) (Bartos et al. 1999; Beenhakker et al. 2005; Coleman et al. 1995).
|
). Thus for a constant MCN1 firing frequency, increasing the Int1 inhibition of LG prolongs the time needed for the MCN1-activated current to enable LG to reach burst threshold (Fig. 9B). Prolonging the retractor phase in this manner prolongs the subsequent protractor phase due to the apparently unchanged decay rate of the MCN1 excitation of LG (Beenhakker et al. 2005).
In the pyloric circuit neurons of C. borealis, several neuropeptides including CCAP and CabTRP Ia selectively activate the same, voltage-dependent inward current (Swensen and Marder 2000, 2001). Because CCAP and CabTRP Ia converge to activate the same ionic current in pyloric neurons, we tested the hypothesis that there was a similar convergence in the gastric mill neurons and this convergence limited CCAP, at the concentrations used (10–7 M), to only influencing relatively fast gastric mill rhythms. As suggested in the preceding text, we anticipated that this limitation resulted from a given CCAP concentration activating a relatively constant amount of the aforementioned peptide-activated current whereas the amount of this current contributed by MCN1 would be larger when the retractor phase was prolonged, as occurs during slower gastric mill rhythms (i.e., Fig. 9). Thus the impact of the CCAP action would be proportionally larger during the faster rhythms.
There was indeed a cycle period-dependent influence of CCAP on the MCN1-elicited gastric mill rhythm (Fig. 10). As shown in Fig. 10Ai, CCAP (10–7 M) appeared to more consistently increase the cycle period during relatively fast gastric mill rhythms. We confirmed this qualitative observation by using a sliding window analysis to analyze 4-s-duration windows of the data from these 27 preparations relative to controls (see METHODS). We thereby determined that CCAP consistently increased the gastric mill cycle period when the control cycle period was 9.0 s (0.001 < P < 0.05 for each window between 4 and 8 and 7 and 11 s). Overall, when the control gastric mill cycle period was 9.0 s, CCAP (10–7 M) increased the cycle period from 6.9 ± 0.3 s in saline to 8.5 ± 0.4 s in CCAP (n = 12, P < 0.001; Fig. 10Aii). At the level of individual experiments, the presence of CCAP increased the gastric mill cycle period in 10/12 preparations with control cycle periods of 9 s, whereas the cycle period increase occurred in only 5/15 preparations with longer-duration control cycle periods (Fig. 10Ai).
|
Because we compared the results of S:CCAP to the results of a set of control experiments (S1, S2), we separated the control data sets into those with fast (9.0 s) and slow (>9.0 s) cycle periods to match them to the cycle-period-dependent actions of CCAP. All of the analyzed parameters were equivalent for the slow group, and nearly all parameters were equivalent for the fast group (supplementary Table 1, Table 2). For example, the gastric mill cycle period was unchanged in both control groups (fast group: S1, 6.5 ± 0.5 s; S2, 6.7 ± 0.4 s; n = 10, P = 0.20, paired Student's t-test; slow group: S1, 11.2 ± 0.6 s; S2, 11.3 ± 0.5 s; n = 11, P = 0.77, paired Student's t-test) as were the LG burst duration, intraburst firing frequency, and number of spikes per burst (Fig. 10; supplementary Tables 1, Tables 2). There were two parameters that were distinct for the S1 versus S2 controls in the fast group, including MG neuron burst duration (P < 0.01, n = 4) and the phase off of the MG neuron burst (P < 0.01, n = 4; supplementary Table 1).
The CCAP-mediated slowing of these relatively fast rhythms resulted largely from its causing a prolongation of the protractor (LG neuron active) phase (Figs. 8 and 10 and supplementary Table 1). For example, the LG neuron burst duration increased during CCAP (10–7 M) application to these fast rhythms (saline: 4.4 ± 0.3 s; CCAP: 5.6 ± 0.4 s; n = 12, P < 0.001; Fig. 10B). There was no change in LG burst duration during the slower rhythms (saline: 8.0 ± 0.5 s; CCAP: 8.3 ± 0.4 s; n = 15, P = 0.30). In contrast to the cycle-period-dependent actions of CCAP on the gastric mill cycle period and LG burst duration, CCAP caused an increase in the number of LG neuron action potentials per gastric mill cycle and in its intraburst firing frequency during both fast and slow rhythms (Fig. 10, C and D; supplementary Table 1, Table 2).
Based on the preceding information, we reasoned that if CCAP was providing a smaller fraction of the total excitation to LG when the rhythm was slow, then providing a higher CCAP concentration to slow rhythms would increase its influence on LG relative to the contribution from MCN1 and therefore effectively influence these rhythms in a manner comparable to the lower CCAP concentrations on faster rhythms. Consistent with this hypothesis, application of CCAP (10–6 M) to slow rhythms did cause an increased cycle period (saline: 10.8 ± 0.3 s; CCAP: 13.6 ± 1.1 s; n = 8, P < 0.05) and increased LG burst duration (saline: 7.6 ± 0.5 s; CCAP: 10.4 ± 1.2 s; n = 8, P < 0.05; Fig. 11).
|
CCAP also did not alter the activity of any of the retractor neurons during these fast rhythms, including Int1, DG, and VD (supplementary Table 1). The absence of increased activity in Int1 was surprising because it was directly excited by focally applied CCAP (10–5 M; Fig. 3). Moreover, the pyloric-timed firing frequency of Int1 was consistently increased by CCAP (10–7 M) superfusion prior to each MCN1 stimulation (saline: 6.4 ± 1.1 Hz; CCAP: 8.8 ± 1.3 Hz; n = 7, P < 0.05; Fig. 12). Unlike Int1, the retractor neuron AM was not activated during CCAP (10–7 M) application (n = 7) despite the fact that AM was directly excited by higher concentrations of focally applied CCAP (see preceding text). Aside from the LG neuron, none of the recorded gastric mill neurons exhibited altered activity when CCAP was applied during slow gastric mill rhythms (supplementary Table 2).
|
CCAP influence on the gastric mill rhythm is dose-dependent
We also determined the threshold concentration at which CCAP influenced the MCN1-elicited gastric mill rhythm. For these experiments, we used only preparations with relatively fast control cycle periods (9.0 s). As we found with the CCAP influence on the MCN1 firing frequency threshold necessary to generate the gastric mill rhythm, superfusion of CCAP concentrations as low as 10–10 M (n = 7) slowed the MCN1-elicited gastric mill rhythm (Fig. 13A). There was a comparable increase in the LG burst duration and intraburst firing frequency for this same range of CCAP concentrations (Fig. 13, B and C). As in the preceding text, none of the studied parameters were different in the matched saline controls (S1, S2; Fig. 13).
|
), and the mean (±SD) value for the same parameter from 2 successive gastric mill rhythms elicited during saline superfusion (S1, S2: 
), obtained from a separate set of preparations. —, equivalent points on the y and x axes. C: CCAP (10–7 M) application altered several parameters of the VCN-triggered gastric mill rhythm across preparations. Specifically, relative to the same parameter during saline superfusion (S), CCAP increased the gastric mill cycle period (Ci, n = 15), LG burst duration (Cii, n = 15), LG intraburst firing frequency (Ciii, n = 15), and number of LG action potentials per burst (Cvi, n = 15). There was no change in any of these parameters during consecutive rhythms triggered during saline superfusion (S1, S2: P > 0.05, n = 11). ***P < 0.001.
