endocannabinoids mediate tachykinin-induced effects in the

Endocannabinoids Mediate Tachykinin-Induced Effects in the LampreyLocomotor Network

Carolina Thorn Perez, Russell H. Hill, Abdeljabbar El. Manira, and Sten GrillnerDepartment of Neuroscience, Nobel Institute for Neurophysiology, Karolinska Institutet, Stockholm, Sweden

Submitted 2 April 2009; accepted in final form 28 June 2009

Thorn Perez C, Hill RH, El Manira A, Grillner S. Endocannabi-noids mediate tachykinin-induced effects in the lamprey locomotornetwork. J Neurophysiol 102: 1358–1365, 2009. First published July1, 2009; doi:10.1152/jn.00294.2009. The spinal network underlyinglocomotion in lamprey is composed of excitatory and inhibitory inter-neurons mediating fast ionotropic action. In addition, several modulatorsystems are activated as locomotion is initiated, including the tachykininsystem and the metabotropic glutamate receptor 1 (mGluR1), the latteroperating partially via the endocannabinoid system. The effects ofmGluR1 agonists and tachykinins resemble each other. Like mGluR1agonists, the tachykinin substance P accelerates the burst rate andreduces the crossed inhibition in an activity-dependent fashion. Thepresent study therefore explores whether tachykinins also use theendocannabinoid system to modulate the locomotor frequency. Bymonitoring fictive locomotion, we were able to compare the facilita-tory effects exerted by applying substance P (1 �M, 20 min), on theburst frequency before and during application of the endocannabinoidCB1 receptor antagonist AM251 (2–5 �M). By using two differentlamprey species, we showed that the response to substance P on theburst frequency is significantly reduced during the application ofAM251. To examine whether endocannabinoids are involved in thesubstance P–mediated modulation of reciprocal inhibition, the com-missural axons were stimulated, while recording intracellularly frommotoneurons. We compare the effect of substance P on the ampli-tude of the contralateral compound glycinergic inhibitory postsyn-aptic potential (IPSP) in control and in the presence of AM251.The blockade of CB1 receptors reduced the substance P–mediateddecrease in the amplitude by 29%. The present findings suggest thatthe effects of substance P on the increase in the locomotor burstfrequency and depression of IPSPs are mediated partially via releaseof endocannabinoids acting through CB1 receptors.

I N T R O D U C T I O N

The motor pattern underlying locomotion is coordinated bycentral pattern generator networks (CPGs) in all classes ofvertebrates (see Grillner 2006). The locomotor CPGs areturned on from command regions in the brain stem via gluta-matergic reticulospinal neurons (Alford and Dubuc 1993; Bro-din et al. 1994; Dubuc et al. 2008; Ohta and Grillner 1989;Sirota et al. 2000). The intrinsic function of the CPG networkshas been elucidated in the lamprey and amphibian tadpoles(Buchanan and Grillner 1987; Cangiano and Grillner 2005;Cohen and Harris-Warrick 1984; Grillner et al. 1998; Robertset al. 2008; Sillar and Roberts 1993). Reticulospinal neuronsactivate groups of CPG interneurons in the spinal cord, whichgenerate the burst pattern, and the inhibitory glycinergic neu-rons ensure left–right alternation through reciprocal inhibition(Biro et al. 2008; Buchanan 1982; Cohen and Harris-Warrick

1984). The CPG network in lampreys consists of interneuronsthat act through ionotropic glutamate (N-methyl-D-aspartate[NMDA] and �-amino-3-hydroxy-5-methyl-4-isoxazolepropi-onic acid [AMPA]) and glycinergic receptors. In addition,numbers of metabotropic receptors contribute with a fine-tuning of the network properties by an action on neuronal andsynaptic properties. They include different metabotropic glu-tamate receptors (mGluRs): the 5-hydroxytryptamine (5-HT)system, �-aminobutyric acid type B (GABAB) receptors, andtachykinins (Harris-Warrick and Cohen 1985; Kettunen et al.2005; Krieger et al. 1994, 1998; Kyriakatos and El Manira2007; Photowala et al. 2006; Thorn Perez et al. 2007a; Zhangand Grillner 2000). mGluR1 receptors act partially via endo-cannabinoid signaling, through a depression of glycinergicsynaptic transmission (Kettunen et al. 2005). The net effects oftachykinins and mGluR1 receptors on the network level resem-ble each other, in that substance P is also known to reduce thecrossed inhibition in an activity-dependent fashion (Parker andGrillner 1999a). We therefore explored whether some of theeffects of tachykinins in the lamprey CPG are mediated byendocannabinoids.

Tachykinins and endocannabinoids exert powerful controlover the excitability in sensory and motor circuits in the spinalcord of the lamprey and other chordates (Farquhar-Smith et al.2000; Hokfelt et al. 2001; Kettunen et al. 2005; Kyriakatos andEl Manira 2007; Thorn Perez et al. 2007a; Waugh et al. 1995).Tachykinin receptors—named NK1, NK2, and NK3—are G-protein coupled and have been characterized both pharmaco-logically and genetically. Substance P is the preferred ligandfor the NK1 receptor, which is widely represented and distrib-uted in the central and peripheral nervous system of chordates(Fried et al. 1988; Maggi and Schwartz 1997), including thelamprey (Van Dongen et al. 1985, 1986). Activation of NK1,as with the mGluR1 receptor, is known to produce diacylglyc-erol with the subsequent activation of protein kinase C(PKC) (Ferguson et al. 2008; Parker et al. 1998; Wajimaet al. 2000). Eicosanoid ligands serve as endogenous ago-nists for CB1 and CB2, which are G-protein– coupled can-nabinoid receptors. CB1 receptors are among the mostabundant Gi/Go linked types in the brain. They have alsobeen found in the spinal cord and are known to mediateretrograde signaling in the CNS (El Manira et al. 2008;Yoshida et al. 2002). CB2 receptors are primarily found onimmune cells (Chevaleyre et al. 2006; Farquhar-Smith et al.2000; Kano et al. 2009; Salio et al. 2002).

In the lamprey, a brief activation of tachykinin receptors byexogenous application of substance P to the spinal cord hasboth short- and long-term effects on the locomotor burstfrequency (Parker et al. 1998; Svensson et al. 2001; Thorn

Address for reprint requests and other correspondence: S. Grillner, Dept. ofNeuroscience, Karolinska Institute, 171 77 Stockholm, Sweden (E-mail:[email protected]).

J Neurophysiol 102: 1358–1365, 2009.First published July 1, 2009; doi:10.1152/jn.00294.2009.

1358 0022-3077/09 $8.00 Copyright © 2009 The American Physiological Society www.jn.org

by 10.220.33.3 on Septem

ber 27, 2016http://jn.physiology.org/

Dow

nloaded from

http://jn.physiology.org/

Perez et al. 2007a) and can modulate brain stem locomotorcontrol as well (Brocard et al. 2005). Further, tachykinins areendogenously released during locomotor activity in that block-ade of NK1 receptors reduces the burst frequency of fictivelocomotion (Thorn Perez et al. 2007a).

We show here that the effects of substance P on the networkactivity are appreciably reduced by a CB1 receptor antagonist.Furthermore, depression of the inhibitory postsynaptic poten-tials (IPSPs) in contralateral spinal neurons induced by sub-stance P is markedly reduced by a CB1 antagonist. It is suggestedthat a component of the effects of substance P is mediated byactivation of CB1 receptors. These results were previously re-ported in abstract form (Thorn Perez et al. 2007b, 2008).

M E T H O D S

Intact spinal cords from 65 adult lampreys of two species were used.Lampetra fluviatilis were collected in Ljusne, Sweden and Ichthyomyzonunicuspis were obtained from Aquatic Supply (Marquette, IA). Theywere kept in separate aerated aquaria at a temperature of 5°C. Allprotocols were approved by the animal research ethical committee(Stockholm). Lampreys were anesthetized with tricaine methanesulfonate(MS 222, 100 mg/l; Sigma–Aldrich, Stockholm) and the viscera, mus-culature, and ventral half of the notochord were removed.

The spinal cord and notochord of around 10–12 segments from theregion between the gills and the dorsal fin (Fig. 1A) were pinned to aSylgard-lined chamber and continuously perfused with physiologicalsolution at 8–10°C. The physiological solution for L. fluviatilis wascomposed of (in mM) 138 NaCl, 2.1 KCl, 1.8 CaCl2, 1.2 MgCl2, 4glucose, and 2 HEPES, and it was bubbled for 20 min with O2 and pHadjusted to 7.4 with NaOH. The physiological solution for I. unicuspiswas composed of (in mM) 91 NaCl, 2.1 KCl, 2.6 CaCl2, 1.8 MgCl2,4 glucose, and 23 NaHCO3, and it was bubbled with 5% CO2-95% O2

for 20 min, before the pH was adjusted to 7.65 with 1 M NaOH, andalso during the experiment.

To analyze the effects of different agonists and antagonists on thelocomotor frequency, the glycogen-containing tissue layer surround-ing the spinal cord (Rovainen 1970) was removed to increase visibil-ity, to facilitate drug penetration, and to allow extracellular recordingsusing glass suction electrodes placed gently on contralateral ventralroots (Fig. 1A). Fictive swimming was induced by adding N-methyl-D-aspartate (NMDA; Tocris, Bristol, UK) to the solution. A concen-trated stock was kept frozen and diluted to the final concentration(75–100 �M) during experiments. Burst frequency increased gradu-ally and after �3 h of NMDA perfusion there was no significantchange in frequency (see Thorn Perez et al. 2007a), thereafter allow-ing stable long-term recordings (�10 h). Agonists and antagonistswere added only after the burst frequency was stable. Substance P (1mM; Sigma–Aldrich), was dissolved in water with 0.05 M acetic acidto prevent oxidation and 1% bovine serum albumin to increase thesolution stability; frozen aliquots were stored at �20°C. They weredissolved in sufficient physiological solution to reach 1 �M concentrationof substance P (Parker et al. 1998) and applied for 20 min. Acetic acidalone at this final concentration had no effect on the locomotor frequency.1-(2,4-Dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-(1-piperidyl)pyra-zole-3-carboxamide (AM251, Tocris), an antagonist of cannabinoid re-ceptors, was prepared in 100 mM stock solutions, dissolved in DMSO,and stored at �20°C. The aliquots were dissolved in a physiologicalsolution to reach a final concentration of 2–5 �M (Kettunen et al. 2005).The final concentration of the solvent DMSO was 0.002–0.005%, whichis known not to affect the locomotor frequency or synaptic transmission(Tsvyetlynska et al. 2005).

Activity was recorded and amplified by a differential AC amplifier(Model MA 102; Biotrend, Cologne, Germany) and band-pass filteredbetween 100 and 500 Hz. The output from the amplifier was coupled

to an A/D converter (Digidata 1320, sampling rate 1 kHz; MolecularDevices, Sunnyvale, CA) and the data were acquired with pClamp 8.2software (Molecular Devices). One-minute recordings of ventral rootactivity were sampled every 5–20 min. Signals were analyzed withpClamp 9.0 and Spike2 4.16 software (Cambridge Electronic Design,Cambridge, UK). Burst frequency and rhythmic quality were analyzedby an event autocorrelogram (normalized between 0 and 1) of ventralroot recordings. Before calculating this function, the amplitude ofbaseline noise of the raw recordings was estimated; a multiple of thisvalue was determined empirically for each case and selected as athreshold above which events were set. The event autocorrelationgives a representative measure of the burst activity (Fig. 1B). In allcases, the raw recordings were manually measured and rhythmicity atthe frequency was predicted by the autocorrelogram was confirmed.To compare frequency of the controls with those in AM251, theaverage frequencies from three to five recordings taken prior tosubstance P application were normalized so that the percentage ofchange could be estimated in the controls. For experiments withAM251, the frequency decay was calculated with a linear regressionand the values were extrapolated to subtract the effect of the antag-onist; the percentage of change was estimated from the normalizedvalues. The rhythmic quality was described using the coefficient ofrhythmicity (Cr) (Cangiano and Grillner 2003), defined as Cr � (A �B)/(A � B), where A and B are, respectively, the height of the secondpeak and the first trough of the autocorrelogram (Fig. 1C). Cr valuescloser to 1 indicate greater rhythmicity. To compare rhythmic quality ofthe controls with those in AM251, the averages of three to five Cr valuesfrom recordings before and after substance P application were normal-ized.

Summary statistics and the values shown in the figures are reportedas �SE and n represents the number of experiments. The significancewas assessed with Student’s unpaired t-test and the significance levelwas defined as p � 0.05.

For the intracellular recordings, the notochord and the meningeswere removed to facilitate penetration into the tissue. The isolatedspinal cord was pinned ventral side up to prevent excitation of thesensory afferents. The spinal cord was activated by electrical stimu-lation (100–300 �A), with a glass pipette electrode, by delivering abrief train of 10–11 pulses of 3 ms duration, 20 Hz every 10 s, to theventral surface of the preparation near the midline. The intracellularelectrode was placed contralateral to this and four to five segmentscaudal to record responses, mainly from inhibitory crossed caudally(CC) projecting interneurons (Fig. 5A). Intracellular recordings weremade with sharp electrodes pulled to a final resistance of about 50 M�when filled with 3 M potassium acetate and 0.1 M KCl. Intracellularsignals were recorded using an Axoclamp 2A amplifier (MolecularDevices) in bridge mode, low-pass filtered at 1 kHz, and further ampli-fied. The signals were digitized at a sampling rate of 6.7 kHz per channelby a Digidata 1320A (Molecular Devices) and stored on a PC usingpClamp software (Molecular Devices).

Ionotropic glutamate receptors were blocked by adding the NMDAreceptor antagonist D-2-amino-5-phosphonopentanoic acid (AP5, 50�M; Tocris) and the AMPA receptor antagonist 6-cyano-7-nitroqui-noxaline-2,3-dione (CNQX, 40 �M; Tocris). Crossing inhibitorypremotor interneurons were monitored by recording their axonalaction potentials extracellularly on the side contralateral and caudal tothat of the stimulation. Motoneurons were identified by an extracel-lular spike in the ventral root recording (described earlier) in the samesegment. The glycinergic and monosynaptic nature of the IPSP (inaddition to the glutamatergic block) was further corroborated with theblockade of glycine transmission by the antagonist strychnine (5 �M;Sigma–Aldrich) and the constant delay between the stimulation arti-fact and the response. The peak amplitude of the IPSPs was measuredand 100 sweeps were averaged in control and in the presence of theendocannabinoid antagonist AM251. The effects of the cannabinoidreceptor agonist WIN55212-2 {(R)-(�)[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenyl-

1359ENDOCANNABINOIDS MEDIATE TACHYKININ EFFECTS

J Neurophysiol • VOL 102 • SEPTEMBER 2009 • www.jn.org



Dow

nloaded from


methan-1-mesylate, 5 �M; Tocris Bioscience} were blocked byAM251, thus confirming the antagonist effect.

R E S U L T S

Blockade of CB1 cannabinoid receptors by AM251 attenuatesthe acute response of the locomotor network to substance P

Fictive swimming was induced in isolated spinal cord prep-arations by perfusion of NMDA (75–100 �M). After approx-imately 3 h, a stable pattern of activity was reached and thequality of the rhythmic activity, expressed by the coefficient ofrhythmicity (Cr; see METHODS), indicated a high degree ofregularity (Cr �0.6).

Activation of tachykinin receptors by exogenous applicationof substance P enhances the burst rate and modulates the motoractivity in L. fluviatilis (Thorn Perez et al. 2007a). Theseresults were also obtained with a 20-min application of sub-stance P (1 �M) in the isolated spinal cord of I. unicuspis. Thisresponse was used as a control to test for a possible role ofendocannabinoid receptors on the effects exerted by substance P.As illustrated in Fig. 2, A and B, an application of substance Presults in a significant (p � 0.001) increase in frequency of thelocomotor rhythm from 2.5 � 0.11 to 3.5 � 0.12 Hz. A secondapplication of substance P in the same spinal cord in the presenceof the CB1 receptor antagonist AM251 did not result in anincreased frequency (Fig. 2C, 2.1 � 0.15 Hz). During the block-ade with AM251, the average frequency as a rule decreasesprogressively, compared with the initial value (Kettunen et al.2005). Figure 2D shows the time course of drug application inwhich, in addition to attenuation of the response to substance P,antagonizing CB1 also causes a steady decrease in frequency.

To test whether the decrease in the effect of substance P wasindeed due to the blockade of cannabinoid receptors, and not tothe previous application of substance P, two consecutive ap-plications of substance P in the absence of cannabinoid antag-onist were used as controls. In these experiments, the firstapplication of substance P increased the burst frequency by anaverage of 33 � 8.9% (p � 0.01, n � 10; Fig. 3A) for L.fluviatilis and 16 � 4.3% (p � 0.01, n � 3, Fig. 3E) for I.unicuspis. The second application of substance P, after 3 h ofwashout, increased the burst frequency by an average of 20 �9.5% (p � 0.01, n � 10, Fig. 3A) in L. fluviatilis and 14 �8.6% (p � 0.01, n � 3, Fig. 3E) in I. unicuspis. Thusattenuation in a second application of substance P cannotaccount for the reduced effect in the CB1 antagonist.

Eight additional spinal cord preparations of L. fluviatilis andthree of I. unicuspis were used to investigate the effect ofAM251 on the second consecutive application of substance P.To compare the controls with those in AM251, averages ofthree to five recordings before and after application of sub-stance P, alone and in AM251, were normalized to the steady-state control (Fig. 3, B and F). Due to the reduction in burstfrequency by the blockade of CB1 receptors (Figs. 2D and 3,B and F), the decay was calculated using a linear regressionand the values were extrapolated to subtract the projectedeffect of the antagonist. The values were then normalized andthe percentage of change estimated. There was no significantdifference between the samples during 30 min prior to thesecond application of substance P in the presence of AM251(6.8 � 7.0% in L. fluviatilis and 1.7 � 1.2% in I. unicuspis).Figure 3, C and G shows the percentage of change in frequency

after the second application of substance P in controls and inAM251 (p � 0.05, 14.9 � 1.9 and 7.5 � 2.1%, respectively,in L. fluviatilis; p � 0.001, 10 � 1.5 and 1.75 � 1.2%,respectively, in I. unicuspis) The increase in frequency wassignificantly lower in AM251 than that for the controls. Takentogether, these observations suggest that the increase in thelocomotor frequency obtained by substance P may include amodulation via CB1 receptors.

B

V

R-L

1 2 3 4

V

R-R

Events

Events

Time (s)

C

-2.4 -1.6 -0.8 0.0 0.8 1.6 2.4

AA

B0

TT = 0.69A = 0.92B = 0.035Cr = 0.93

Time (s)

Cr = (A - B)/(A +B)

A

ventral

VR-R

VR-L

FIG. 1. Preparation, experimental configuration, and analysis protocols.A: extracellular suction electrodes were placed on the left and right ventralroots (VR-L, VR-R) of the isolated spinal cord at segments lying between thegills and the dorsal fin, as shown by the dashed lines on the whole animalillustration. B: a representative sample of the ventral root records taken duringstable fictive locomotion induced by 75 �M N-methyl-D-aspartate (NMDA).Below each raw recording an event transformation is shown, which gives ameasurement of the bursting frequency. C: the event autocorrelation displaysoscillations of the period (T), established as 1/frequency. The amplitude of theoscillations is related to the regularity, which is described with values thatrange between 0 and 1, and was defined as the coefficient of rhythmicity [Cr �(A � B)/(A � B)].

1360 C. THORN PEREZ, R. H. HILL, A. EL MANIRA, AND S. GRILLNER




Dow

nloaded from


Decrease in coefficient of rhythmicity induced by substanceP is not affected by AM251

To investigate whether the blockade of endocannabinoidreceptors had also affected the quality of the burst rhythm, wefollowed the change in the coefficient of rhythmicity (Cr, asdefined in METHODS), induced by substance P alone and in thepresence of AM251.

A 20-min application of substance P during stable swim-ming with regular rhythmicity (Cr �0.6) resulted in a sig-nificant decrease in Cr by 24 � 8.3% (p � 0.005, n � 10)for L. fluviatilis and 16 � 3.1% (p � 0.05, n � 3) for I.unicuspis. This effect lasted for about 1 h and recovered toa similar control level after washout of substance P. Asecond application of substance P to the same spinal cords,3 h after washout of the first application, induced a similardecrease in rhythmicity in both species (16 � 6.5.%, p �0.05, n � 10, for L. fluviatilis; 20 � 6.9%, p � 0.01, n � 3,for I. unicuspis). In the same way, we compared the Cr inthe first and the second applications of substance P, with andwithout AM251. The Cr was not significantly differentduring the application of substance P with and withoutAM251 (Fig. 3, D and H), although the frequency wasdifferent (see Fig. 3, C and G). This might be explained bythe different mechanisms underlying responses of the burstfrequency and the coefficient of rhythmicity.

Blockade of CB1 cannabinoid receptors by AM251 obviatessubstance P–mediated depression of the IPSPs fromcommissural interneurons

The above-cited results show that the substance P–mediatedincrease in the locomotor frequency is significantly reduced bythe CB1 receptor antagonist AM251. At the synaptic level,

reduction of glycinergic inhibition with strychnine (Cohen andHarris-Warrick 1984; Grillner and Wallen 1980) or by surgicalseparation of the left and right halves of the spinal cord(Cangiano and Grillner 2003) results in an increase in thelocomotor frequency. We therefore examined whether glycin-ergic inhibitory synapses from commissural interneurons un-derlying left–right alternation during locomotor activity couldbe modulated by substance P via CB1 receptors.

For this, intracellular recordings were made from motoneu-rons, whereas crossing axons and inhibitory interneurons on thecontralateral side of the spinal cord were stimulated extracellularlywith trains of 11 � 1 pulses at 20 Hz (Alford and Grillner 1991;Parker and Grillner 1999a) (Fig. 4A). Because excitatory inter-neurons could also have been stimulated, the inhibitory compo-nent was isolated by blocking NMDA and AMPA receptors withAP5 (50 �M) and CNQX (40 �M), respectively.

The antagonistic action of AM251 was confirmed by thelack of effect of CB1 agonist WIN55212-2 (Fig. 4B), whichreduces the IPSP amplitude in motoneurons (Kettunen et al.2005). The glycinergic nature of the IPSP was corroborated byits blockade with the antagonist strychnine (Fig. 4D). In thepresence of glutamate blockers, the amplitude of the IPSPsinitially decreased during the first 5 min to subsequentlyremain stable (Fig. 4C). We therefore applied substance P onlyafter �10 min of stable recording and took as control the dataimmediately before the application.

In 9 of 12 cases, superfusion of substance P (1 �M) causeddepolarization in motoneurons, which was compensated withcurrent injection to maintain the cell at a stable membranepotential. In all cases, substance P caused a reduction in theamplitude of monosynaptic IPSPs elicited by the extracellularstimulation (Fig. 5A) and the first glycinergic IPSP was signif-icantly reduced by 53 � 8.7% (p � 0.05, n � 4).

SPSP

AM251

A B C

D

1

1.5

2

2.5

3

3.5

4

0 100 200 300 400 500

Time (min)

Fre

quency

(H

z)

AM251 + Substance P 2.1 Hz

Substance P3.5 Hz

Control (NMDA)2.5 Hz

VR

- L

VR

- R

VR

- L

VR

- R

VR

- L

VR

- R

1 s

A

B

C

1 s 1 s

FIG. 2. The effect of substance P on ven-tral root burst frequency in the presence ofan endocannabinoid antagonist. A: represen-tative example of a stable ventral root burst-ing induced by 100 �M of NMDA in Lam-petra fluviatilis. B: 20-min application ofsubstance P (1 �M) during fictive locomo-tion caused an acceleration of the frequency(p � 0.01). C: 2nd 20-min application ofsubstance P (1 �M) during fictive locomo-tion, following 2 h of blockade of endocan-nabinoid receptors by the antagonist AM251.Blockade of cannabinoid receptors by AM251significantly reduced the increase in frequencyinduced by substance P (p � 0.001). D: timecourse of the change in frequency with theapplication of substance P alone or in AM251(A, B, and C refer to the ventral root tracesshown above).





Dow

nloaded from


This depression was used as the control condition to becompared with the effect of substance P after the endocannabi-noid receptors had been blocked by AM251 (2–5 �M) (Fig.5B). In the preparations pretreated with AM251, the depressionof the first IPSP caused by the application of substance P wassignificantly reduced by 24.3 � 5.6% (p � 0.05, n � 6, Fig. 5,C and D).

These data suggest that the response to substance P is mediatedpartially by an endogenous release of endocannabinoids actingthough presynaptic mechanisms to depress inhibitory synaptic trans-mission within the locomotor network. The fifth and the last IPSPs ofthe train were also analyzed and, in this case, no significant changewas detected (Fig. 5D). Thus the effects of both substance P and itsblock by AM251 are primarily observed in the initial IPSP.

C

0

4

8

12

16

20

% o

f cha

nge

in fr

eque

ncy

with

2nd

app

licat

i on

AM251+SPSP

G

0

2

4

6

8

10

12

14

% o

f cha

nge

in fr

eque

ncy

with

2nd

ap p

l ica t

ion

SP AM251+SP

H

D

0

5

10

15

20

25

% o

f dec

reas

e in

Cr

with

2nd

app

licat

ion

SP

n = 3n = 3

AM251+SP

0

10

20

30

40

% o

f dec

reas

e in

Cr

with

2n d

app

licat

ion

SP

n=10n = 8

AM251+SP

F

BLampetra fluviatilis

Icthyomyzon unicuspis

A

E

n = 8

220

180

140

100

60

Bur

st F

requ

ency

(%

of C

ont r

ol)

Time (min)55045035025015050

SP SP

AM251

100

120

140

160

50 150 250 350 450

Time (min)

Bur

st F

requ

ency

(%

of C

ontr

ol)

SP SP

n = 10

80

90

100

110

120

Bur

st F

requ

ency

(%

of C

o ntr

ol)

50 150 250 350 450 550Time (min)

SP

AM251

SP

n = 3100

110

120

130

Bur

st F

requ

ency

(% o

f con

trol)

50 150 250 350 450

Time (min)

SP SP

n = 3

FIG. 3. The effect of endocannabinoids on the response to substance P. A and E: plots of averaged burst frequency over time from 2 different species showingsimilar responses to 2 consecutive 20-min applications of substance P (1 �M). B and F: averaged burst frequency showing a gradual decrease in the burstfrequency in the presence of the cannabinoid antagonist, AM251 (2–5 �M); and during which the response of a 2nd application of substance P (20 min, 1 �M)was greatly attenuated. C and G: graphic illustration of the percentage of change in frequency with the 2nd application of substance P in control solution andin AM251 (*p � 0.05, ***p � 0.001). D and H: graphic illustration of the percentage of decrease between Cr of the 2nd application of substance P in controlsolution and in AM251. No significant difference was detected.

FIG. 4. Preparation and intracellular con-figuration. A: an extracellular stimulationelectrode was placed close to the midline anda recording electrode was positioned on thecontralateral side to follow the spike of thecrossing inhibitory interneurons. A sharpelectrode was used to record inhibitorypostsynaptic potentials (IPSPs) of motoneu-rons that were identified by a spike in theventral root (VR) electrode. B: representativeexample of the action of the cannabinoidantagonist AM251 (4 �M), by the lack ofeffect of the agonist WIN55212-2 (4 �M) inreducing the amplitude. C: plot of the de-crease in IPSP amplitude recorded in mo-toneurons over time. The black line repre-sents the averaged decrease (n � 3). Iono-tropic glutamate receptors were blocked byAP5 (50 �M) and CNQX (40 �M). D: theIPSP was blocked by the glycinergic antag-onist strychnine (5 �M).





Dow

nloaded from


D I S C U S S I O N

The lamprey spinal cord allowed us to explore whether thereis an interaction between tachykinins and endocannabinoids inan active network and at an inhibitory synapse. The resultsprovide knowledge on possible inherent mechanisms playing arole in the coordination and modulation of rhythmic locomotornetwork activity. We show here that the effects of the tachy-kinin agonist substance P on the locomotor frequency andinhibitory synaptic transmission in the lamprey spinal cord arereduced if AM251 is applied, suggesting that the substance

P–induced effect may be elicited at least partially by a releaseof endocannabinoids.

Acute effect of substance P on the burst frequency isdependent on endogenous activation of CB1 receptors

Both substance P and endocannabinoid agonists increase thefrequency of locomotor rhythm in the lamprey spinal cord andboth are released endogenously during fictive locomotion (Ket-tunen et al. 2005; Kyriakatos and El Manira 2007; Thorn Perez

Stimulus artifact

100 ms

2 m

VControl

Substance P

SP SP + AM 251

A

B

C

D

0

10

20

30

40

50

60

70

SP

AM251 + SP

% o

f dec

rea s

e of

the

1 st

IPS

Ps

80

20

40

60

80

% o

f dec

reas

e of

the

I PS

Ps

1st IPSP 5th IPSP Last IPSP

100 ms

2 m

V

Stimulus artifactControl (AM251)

AM251 + Substance P

FIG. 5. The endocannabinoid antagonistAM251 (5 �M) reduces the substance Peffect on the glycinergic inhibitory synapsesof contralaterally projecting interneurons tomotoneurons. A: representative example ofaveraged data (100 sweeps) from a motoneu-ron showing a significant reduction of theamplitude of the IPSP elicited by a train of11 pulses at 20-Hz stimulation of inhibitorycontralateral interneurons, before and afterapplication of substance P (1 �m, 20 min).The trains were followed by a test pulse after250 ms, to measure the extent of the recov-ery from depression. B: representative exam-ple of averaged data (100 sweeps) from amotoneuron preincubated with the cannabi-noid antagonist, AM251 (4 �m), and duringwhich application of substance P had a re-duced effect on the IPSP elicited by a 20-Hzstimulation of inhibitory contralateral mo-toneurons. C: summary graph of the percent-age of decrease with the first IPSP betweenthe application of substance P in controlsolution (n � 4) and with AM251 (n � 6)(*p � 0.05). D: graphic illustration summa-rizing the percentage of decrease with the1st, 5th, and last IPSP of the stimulatingtrain, between the application of substance Pin control solution (n � 4) and with AM251(n � 6) (*p � 0.05).

PKC

IP3

PKA

SP

CB1

GlyRNMDA

cAMP

AC

Burst regularity Burst frequency

Endocannabinoids

Glycine vesicles

2+Ca

PIP2PLC

NK1

DAG

FIG. 6. Schematic illustration of the in-tracellular signaling pathways proposed tomodulate the locomotor burst frequency andburst regularity following NK1 activationthrough substance P. NK1 is a G-proteinreceptor and by its activation through sub-stance P, 2 main pathways are initiated. Theone depends on activation of a phospho-lipase C (PLC) signaling pathway that hy-drolyzes phosphatidylinositol-4,5-bisphos-phate (PIP2) into diacylglycerol (DAG) andinositol triphosphate (IP3). This pathway isresponsible for the increase in the locomotorfrequency that is mediated partially by po-tentiation of NMDA receptors via proteinkinase C (PKC) and partially through endo-cannabinoids. The latter can either be syn-thesized from DAG on activation of DAG-lipase or released by Ca2� from internalstores in response to IP3. The released endo-cannabinoids act as retrograde messengers todepress inhibitory synaptic transmission viapresynaptic receptors (CB1), thereby reduc-ing glycinergic inhibition. The 2nd pathwayactivated by NK1 activation regulates theburst regularity though protein kinase A(PKA) via an adenylyl cyclase (AC) signal-ing cascade.





Dow

nloaded from


et al. 2007a). The question at hand is whether the two types ofagonist act by a common signaling pathway. Our resultsindicate that, indeed, the effect of substance P on the locomotornetwork is dependent on activation of a cannabinoid receptorthat is blocked by AM251, a CB1 antagonist. However, themechanisms involved in endocannabinoid release by mGluR1and the action of substance P on the endocannabinoid systemmay differ since antagonizing NK1 receptors depresses theburst frequency to a steady-state level (Thorn Perez et al.2007a), whereas antagonizing CB1 receptors induces a contin-uous decrease in the burst frequency as long as the antagonistis present (Kettunen et al. 2005). This may indicate that NK1receptors are involved in setting the level of network activity,whereas CB1 receptor activation may be necessary to maintainlocomotion over extended periods.

Endocannabinoid-independent effects of substance P onlocomotor network regularity

Application of substance P reproducibly and significantlyreduced burst regularity (Cr) during fictive locomotion. Itcould be viewed as another aspect of the acute response tosubstance P. This effect, however, was not influenced byblocking cannabinoid receptors, whereas in the same experi-ments and during the same timeframe, the frequency wasaffected. The result is important in that it shows other compo-nents of the effect of substance P that do not involve canna-binoid receptors. Previous studies showed that NK1 receptorsactivate both adenylyl cyclase (AC) and phospholiphase C(PLC) second-messenger systems (Maggi and Schwartz 1997). Inan overview of the role of protein kinases in the tachykinin-mediated spike modulation, it was suggested that protein kinase Cis responsible for the potentiation of NMDA receptor activity seenwith exogenous application of substance P (Parker et al. 1998). Itis also suggested that by a different protein-synthesis–indepen-dent pathway, protein kinase A modulates the burst regularity(Parker and Grillner 1999b) (Fig. 6). These two mechanismsmight underlie the different responses of the burst frequencyand the coefficient of rhythmicity.

The responses to substance P and the attenuation by AM251were essentially the same in both L. fluviatilis and I. unicuspis.Likewise, the substance P–induced irregularity in locomotoractivity and the lack of effect by AM251 were similar in bothspecies.

Interaction between substance P and CB1 receptors atinhibitory synapses

Paired recordings from CC interneurons to motoneurons hadshown that substance P significantly depressed the monosyn-aptic IPSPs when activated at 20 Hz (Parker and Grillner1999a). This response was used as a control in the presentstudy to test for a possible role of endocannabinoids, althoughthe stimulation was done extracellularly. In all cases, the firstIPSP was significantly reduced following application of sub-stance P. However, since there is synaptic depression in thecontrol, the effect of substance P on the later pulses in the trainwould be less obvious. The difference between the first IPSPamplitude in control condition and with substance P (Fig. 5A)was significantly (p � 0.05) larger than the difference betweenthe control condition and substance P in the presence of

AM251 (Fig. 5B). From this we can infer that it is very likelythat NK1 receptors use a mechanism similar to that ofmGluR1, in which a reduced crossed inhibition is caused byendocannabinoid release.

Taken together, these observations suggest a possible scheme,shown in Fig. 6, for substance P activation of several pathwaysthat could explain its effects on burst regularity and frequency, thelatter involving endocannabinoid release.

G R A N T S

This work was supported by Swedish Research Council Grants VR-M 3026and VR-NT621-2007-6049, European Union Grant for Spinal Cord RepairHealth-F2.2007–201144/FP7, Karolinska Institute research funds, and Euro-pean Union Grant ICT-STREP 216100 Lampetra.

R E F E R E N C E S

Alford S, Dubuc R. Glutamate metabotropic receptor mediated depression ofsynaptic inputs to lamprey reticulospinal neurones. Brain Res 605: 175–179,1993.

Alford S, Grillner S. The involvement of GABAB receptors and coupledG-proteins in spinal GABAergic presynaptic inhibition. J Neurosci 11:3718–3726, 1991.

Biro Z, Hill RH, Grillner S. The activity of spinal commissural interneuronsduring fictive locomotion in the lamprey. J Neurophysiol 100: 716–722,2008.

Brocard F, Bardy C, Dubuc R. Modulatory effect of substance P to the brainstem locomotor command in lampreys. J Neurophysiol 93: 2127–2141,2005.

Brodin L, Shupliakov O, Pieribone VA, Hellgren J, Hill RH. The reticu-lospinal glutamate synapse in lamprey: plasticity and presynaptic variability.J Neurophysiol 72: 592–604, 1994.

Buchanan JT. Identification of interneurons with contralateral, caudal axonsin the lamprey spinal cord: synaptic interactions and morphology. J Neuro-physiol 47: 961–975, 1982.

Buchanan JT, Grillner S. Newly identified “glutamate interneurons” andtheir role in locomotion in the lamprey spinal cord. Science 236: 312–314,1987.

Cangiano L, Grillner S. Fast and slow locomotor burst generation in thehemispinal cord of the lamprey. J Neurophysiol 89: 2931–2942, 2003.

Cangiano L, Grillner S. Mechanisms of rhythm generation in a spinallocomotor network deprived of crossed connections: the lamprey hemicord.J Neurosci 25: 923–935, 2005.

Chevaleyre V, Takahashi KA, Castillo PE. Endocannabinoid-mediated syn-aptic plasticity in the CNS. Annu Rev Neurosci 29: 37–76, 2006.

Cohen AH, Harris-Warrick RM. Strychnine eliminates alternating motoroutput during fictive locomotion in the lamprey. Brain Res 293: 164–167,1984.

Dubuc R, Brocard F, Antri M, Fenelon K, Gariepy JF, Smetana R,Menard A, Le Ray D, Viana Di Prisco G, Pearlstein E, Sirota MG,Derjean D, St-Pierre M, Zielinski B, Auclair F, Veilleux D. Initiation oflocomotion in lampreys. Brain Res Rev 57: 172–182, 2008.

El Manira A, Kyriakatos A, Nanou E, Mahmood R. Endocannabinoidsignaling in the spinal locomotor circuitry. Brain Res Rev 57: 29–36, 2008.

Farquhar-Smith WP, Egertova M, Bradbury EJ, McMahon SB, Rice AS,Elphick MR. Cannabinoid CB(1) receptor expression in rat spinal cord. MolCell Neurosci 15: 510–521, 2000.

Ferguson AR, Bolding KA, Huie JR, Hook MA, Santillano DR, MirandaRC, Grau JW. Group I metabotropic glutamate receptors control metaplas-ticity of spinal cord learning through a protein kinase C-dependent mecha-nism. J Neurosci 28: 11939–11949, 2008.

Fried G, Franck J, Brodin E. Differential distribution of 5-hydroxytrypta-mine and substance P in synaptosomal vesicles of rat ventral spinal cord.Neurosci Lett 91: 315–320, 1988.

Grillner S. Neuronal networks in motion from ion channels to behaviour. AnR Acad Nac Med (Madr) 123: 297–298, 2006.

Grillner S, Ekeberg, El Manira A, Lansner A, Parker D, Tegner J, WallenP. Intrinsic function of a neuronal network: a vertebrate central patterngenerator. Brain Res Brain Res Rev 26: 184–197, 1998.

Grillner S, Wallen P. Does the central pattern generation for locomotion inlamprey depend on glycine inhibition? Acta Physiol Scand 110: 103–105,1980.





Dow

nloaded from


Harris-Warrick RM, Cohen AH. Serotonin modulates the central patterngenerator for locomotion in the isolated lamprey spinal cord. J Exp Biol 116:27–46, 1985.

Hokfelt T, Pernow B, Wahren J. Substance P: a pioneer amongst neuropep-tides. J Intern Med 249: 27–40, 2001.

Kano M, Ohno-Shosaku T, Hashimotodani Y, Uchigashima M, WatanabeM. Endocannabinoid-mediated control of synaptic transmission. PhysiolRev 89: 309–380, 2009.

Kettunen P, Kyriakatos A, Hallen K, El Manira A. Neuromodulation viaconditional release of endocannabinoids in the spinal locomotor network.Neuron 45: 95–104, 2005.

Krieger P, Grillner S, El Manira A. Endogenous activation of metabotropicglutamate receptors contributes to burst frequency regulation in the lampreylocomotor network. Eur J Neurosci 10: 3333–3342, 1998.

Krieger P, Tegner J, El Manira A, Grillner S. Effects of metabotropicglutamate receptor activation on the cellular and network level in thelamprey spinal cord. Neuroreport 5: 1760–1762, 1994.

Kyriakatos A, El Manira A. Long-term plasticity of the spinal locomotorcircuitry mediated by endocannabinoid and nitric oxide signaling. J Neuro-sci 27: 12664–12674, 2007.

Maggi CA, Schwartz TW. The dual nature of the tachykinin NK1 receptor.Trends Pharmacol Sci 18: 351–355, 1997.

Ohta Y, Grillner S. Monosynaptic excitatory amino acid transmission from theposterior rhombencephalic reticular nucleus to spinal neurons involved in thecontrol of locomotion in lamprey. J Neurophysiol 62: 1079–1089, 1989.

Parker D, Grillner S. Activity-dependent metaplasticity of inhibitory andexcitatory synaptic transmission in the lamprey spinal cord locomotornetwork. J Neurosci 19: 1647–1656, 1999a.

Parker D, Grillner S. Long-lasting substance-P-mediated modulation ofNMDA-induced rhythmic activity in the lamprey locomotor network in-volves separate RNA- and protein-synthesis-dependent stages. Eur J Neu-rosci 11: 1515–1522, 1999b.

Parker D, Zhang W, Grillner S. Substance P modulates NMDA responsesand causes long-term protein synthesis-dependent modulation of the lam-prey locomotor network. J Neurosci 18: 4800–4813, 1998.

Photowala H, Blackmer T, Schwartz E, Hamm HE, Alford S. G proteinbetagamma-subunits activated by serotonin mediate presynaptic inhibitionby regulating vesicle fusion properties. Proc Natl Acad Sci USA 103:4281–4286, 2006.

Roberts A, Li WC, Soffe SR. Roles for inhibition: studies on networkscontrolling swimming in young frog tadpoles. J Comp Physiol A NeuroetholSens Neural Behav Physiol 194: 185–193, 2008.

Rovainen CM. Glucose production by lamprey meninges. Science 167:889–890, 1970.

Salio C, Cottone E, Conrath M, Franzoni MF. CB1 cannabinoid receptorsin amphibian spinal cord: relationships with some nociception markers.J Chem Neuroanat 24: 153–162, 2002.

Sillar KT, Roberts A. Control of frequency during swimming in Xenopusembryos: a study on interneuronal recruitment in a spinal rhythm generator.J Physiol 472: 557–572, 1993.

Sirota MG, Di Prisco GV, Dubuc R. Stimulation of the mesencephaliclocomotor region elicits controlled swimming in semi-intact lampreys. EurJ Neurosci 12: 4081–4092, 2000.

Svensson E, Grillner S, Parker D. Gating and braking of short- and long-termmodulatory effects by interactions between colocalized neuromodulators.J Neurosci 21: 5984–5992, 2001.

Thorn Perez C, Hill RH, El Manira A, Grillner S. Neuromodulatoryinteraction between cannabinoids and tachykinins in the lamprey spinalcord. Soc Neurosci Abstr 670, 2007b.

Thorn Perez C, Hill RH, El Manira A, Robertson B, Grillner S. Neuro-modulatory interaction between cannabinoids and tachykinins in the lam-prey spinal cord (part II). Soc Neurosci Abstr 675, 2008.

Thorn Perez C, Hill RH, Grillner S. Endogenous tachykinin release contrib-utes to the locomotor activity in lamprey. J Neurophysiol 97: 3331–3339,2007a.

Tsvyetlynska NA, Hill RH, Grillner S. Role of AMPA receptor desensitiza-tion and the side effects of a DMSO vehicle on reticulospinal EPSPs andlocomotor activity. J Neurophysiol 94: 3951–3960, 2005.

Van Dongen PA, Hokfelt T, Grillner S, Verhofstad AA, Steinbusch HW,Cuello AC, Terenius L. Immunohistochemical demonstration of someputative neurotransmitters in the lamprey spinal cord and spinal ganglia:5-hydroxytryptamine-, tachykinin-, and neuropeptide-Y-immunoreactiveneurons and fibers. J Comp Neurol 234: 501–522, 1985.

Van Dongen PA, Theodorsson-Norheim E, Brodin E, Hokfelt T, GrillnerS, Peters A, et al. Immunohistochemical and chromatographic studies ofpeptides with tachykinin-like immunoreactivity in the central nervous sys-tem of the lamprey. Peptides 7: 297–313, 1986.

Wajima Z, Hua XY, Yaksh TL. Inhibition of spinal protein kinase C blockssubstance P-mediated hyperalgesia. Brain Res 877: 314–321, 2000.

Waugh D, Bondareva V, Rusakov Y, Bjenning C, Nielsen PF, Conlon JM.Tachykinins with unusual structural features from a urodele, the amphiuma,an elasmobranch, the hammerhead shark, and an agnathan, the river lam-prey. Peptides 16: 615–621, 1995.

Yoshida T, Hashimoto K, Zimmer A, Maejima T, Araishi K, Kano M. Thecannabinoid CB1 receptor mediates retrograde signals for depolarization-induced suppression of inhibition in cerebellar Purkinje cells. J Neurosci 22:1690–1697, 2002.

Zhang W, Grillner S. The spinal 5-HT system contributes to the generationof fictive locomotion in lamprey. Brain Res 879: 188–192, 2000.





Dow

nloaded from


endocannabinoids mediate tachykinin-induced effects in the

Documents