A

(KrtCmlpdr©

K

aEa

f

(

K

1

Respiratory Physiology & Neurobiology 154 (2006) 37–46

Ontogeny of central rhythm generation in chicks and rodents�

Chatonnet F. 1, Borday C., Wrobel L., Thoby-Brisson M.,Fortin G., McLean H., Champagnat J. ∗

UPR 2216, Neurobiologie Genetique et Integrative, Institut federatif de Neurobiologie Alfred Fessard,C.N.R.S. 1, Avenue de la terrasse, Gif sur Yvette, 91198 Cedex, France

Accepted 1 February 2006

bstract

Recent studies help in understanding how the basic organization of brainstem neuronal circuits along the anterior–posteriorAP) axis is set by the Hox-dependent segmentation of the neural tube in vertebrate embryos. Neonatal respiratory abnormalities inrox20−/−, Hoxa1−/− and kreisler mutant mice indicate the vital role of a para-facial (Krox20-dependent, rhombomere 4-derived)

espiratory group, that is distinct from the more caudal rhythm generator called Pre-Botzinger complex. Embryological studies inhe chick suggest homology and conservation of this Krox20-dependent induction of parafacial rhythms in birds and mammals.alcium imaging in embryo indicate that rhythm generators may derive from different cell lineages within rhombomeres. Inice, the Pre-Botzinger complex is found to be distinct from oscillators producing the earliest neuronal activity, a primordial

ow-frequency rhythm. In contrast, in chicks, maturation of the parafacial generator is tightly linked to the evolution of this

rimordial rhythm. It seems therefore that ontogeny of brainstem rhythm generation involves conserved processes specifyingistinct AP domains in the neural tube, followed by diverse, lineage-specific regulations allowing the emergence of organizedhythm generators at a given AP level.

2006 Elsevier B.V. All rights reserved.

eres; H

eywords: Rhyhthm; Brainstem; Development; Embryo; Rhombom

� This paper is part of a special issue entitled “Frontiers in Compar-tive Physiology II: Respiratory Rhythm, Pattern and Responses tonvironmental Change”, guest edited by W.K. Milsom, F.L. Powellnd G.S. Mitchell.∗ Corresponding author. Tel.: +33 1 69 82 34 06;

ax: +33 1 69 82 41 78.E-mail address: jean.champagnat@iaf.cnrs-gif.fr

J. Champagnat).1 Present address: Cellular and Molecular Biology Department,arolinska Institutet, 171 77 Stockholm, Sweden.

1

cmctcat

569-9048/$ – see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.resp.2006.02.004

ox; Krox20; Chick; Mouse

. Introduction

During development, the assembly of neural cir-uits that encode animal behaviour results from severalechanisms contributing to generate distinct neuronal

ell types in appropriate number and position. Among

hese mechanisms, regionalization of the neural tube isonserved among vertebrates: it controls cell prolifer-tion and exit from the cell cycle at precise location ofhe neural tube (Lumsden and Krumlauf, 1996; Tanabe

3 siology

apfa1r(ut1vattbg

(ls(tmHftnsfiig

wloar(bearbiprr

nbr(2t(ot1

grdeh

2e

effaHoeHla

reHtdsit1ae

8 F. Chatonnet et al. / Respiratory Phy

nd Jessell, 1996). Neurogenesis refers to a variety ofrocesses by which cell types differentiate, migrate andorm nuclei to eventually produce neuronal populationsnd their synaptic interconnections (Tanabe and Jessell,996). Once activity starts in primordial neurons, neu-onal circuitry is refined in a use-dependent mannerKatz and Shatz, 1996) and the size of neuronal pop-lations depends on the activity of neurotrophic fac-ors interacting with apoptotic processes (Oppenheim,991). Downstream mechanisms greatly vary amongertebrate species. By considering experiments in micend chick, the present review investigates to what extenthe conserved molecular signalling operating duringhe segmentation of the neural tube may be responsi-le for the anterior–posterior subdivision of the rhythmenerating networks in the brainstem.

In all vertebrates, regionalization of the brainstemrhombencephalon) along the antero-posterior axis,eads the partitioning of the neuroepithelium into aeries of cellular compartments, called rhombomeresr1–r8, Lumsden and Krumlauf, 1996). This processakes place between embryonic days E8 and E12 in

ice, between Hamburger and Hamilton (1951) stagesH9 and HH24 in chicks, and influences later dif-

erentiation and spatial distribution of neuronal pat-erns (Lumsden and Keynes, 1989). The branchiomotoruclei conform to the rhombomeric pattern with a two-egment periodicity. Trigeminal motoneurons originaterom r2 and r3 and send their axons to an exit pointn r2. In mammals, the facial branchial nucleus orig-nates in r4. The ventral and dorsal motor nuclei oflossopharyngeal–vagal nerves derive from r6–r8.

In rodents, respiratory groups have been locatedith the help of preparations isolated in vitro. A corre-

ation can be suggested with the rhombomere relatedrganization of the branchio-motor nuclei. Caudally,t the vagal/glossopharyngeal level, the inspiratoryhythm is generated in the pre-Botzinger complexPBC), and persists in coronal brainstem slices of new-orn rodents, isolated in vitro (Smith et al., 1991; Grayt al., 1999; Lieske et al., 2000). At the facial level,nother rhythm generator, called the para-facial respi-atory group (pFRG, Onimaru and Homma, 2003), haseen recently delineated by activity-dependent imag-

ng. The pFRG is characterized by a pre-inspiratoryattern of neuronal activity that stops during inspi-ation and resumes afterward by a post-inhibitoryebound. In the hindbrain isolated in vitro, perma-

twgc

& Neurobiology 154 (2006) 37–46

ent rhythm generation requires a balanced interactionetween the pFRG and the PBC, which can be dis-upted by potent respiratory depressants such as opioıdsDenavit-Saubie et al., 1978; Morin-Surun et al., 1984,001; Mellen et al., 2003; Manzke et al., 2003). At therigeminal level, ventral pontine controls of the rhythmBorday et al., 1997) include noradrenergic neuronsf the A5 group exerting a depressant effect uponhe more caudal respiratory group (Di Pasquale et al.,992).

In the following article, we consider the role of Hoxenes responsible for the segmental organization of thehombencephalon and branchio-motor nuclei. We williscuss their influence on the emergence of rhythm gen-rators, at specific antero-posterior levels of the activeindbrain.

. The murine parafacial controller requiresxpression of Krox20 in r3 and Hoxa1 in r4

Strong regulatory constraints couple Hox genexpression to the progression of embryogenesis. There-ore, the chromosomal organization of Hox genes intoour clusters is highly conserved in vertebrates. In birdsnd mammals, the formation of territories patterned byox genes is accompanied by a sequential activationf these genes from 3′ to 5′ in the clusters. As a result,arly structures are given an anterior identity with 3′ox genes as key determinants, while progressively

ater structures start expressing more 5′ Hox genes andcquire a more posterior identity.

Thus, Hox genes play a key role in anteroposte-ior patterning of the rhombencephalon at pre- andarly-segmental stages of development. The most 3′ox genes, Hoxa1 and Hoxb1 are expressed up to

he rhombomeric r3/r4 boundary and they are directlyown regulated in r3 by Krox20, a zinc-finger tran-cription factor also known as Egr-2. This signallings required for the proper development of r3, r4 and ofhe boundary between r3 r4 (Lumsden and Krumlauf,996; Schneider-Maunoury et al., 1993; Giudicelli etl., 2001). Neurogenesis starts within rhombomeres atnd segmental stages, shortly before hindbrain segmen-

ation disappears. It is therefore important to considerhether this early developmental regulation of pro-enitors influences the function of mature brainstemircuits at later stages.

siology

K(CtgthrptnrdffuTtnndtha(ltin

dK1eaar2TooHfrD2l

einan

Hdmc2rwtftdntmstibTifioDiTnpagep

3cr

F. Chatonnet et al. / Respiratory Phy

Investigation of the respiratory behaviour ofrox20−/−, Hoxa1−/− and kreisler−/− mutants

Jacquin et al., 1996; Domınguez del Toro et al., 2001;hatonnet et al., 2002) has established links between

he transient rhombomere-related expression of theseenes and deletions, neoformations and reconfigura-ion of the respiratory network. These three mutationselped in unveiling a vital role of the rhombomeric3r4 domain for the specification of an adapted res-iratory rhythm at birth (Borday et al., 2003). Inac-ivation of Krox20 and elimination of r4 in Hoxa1ull mutants lead to neonatal deaths due to centralespiratory deficits and probable collapse of the car-iovascular system. The Hoxa1−/− neonates show pro-ound hypoplasia in the ventral pons including theacial branchiomotor nucleus (r4-derived) and of thenderlying adjacent area where the pFRG is located.hese mutants revealed the “anti-apneic” function of

he pFRG at birth. In contrast to Hoxa1 and Krox20ull mice, homozygous kreisler mice lacking r5, showeither life-threatening apneic respiratory patterns norramatic reduction of the ventral area of the pons con-aining the facial branchiomotor nucleus, but rather anypoplasia of the ventral reticular formation causing anbnormal rostral positioning of the ambiguus nucleusChatonnet et al., 2002). Therefore, the ventral reticu-ar neurons in r4 and r5 greatly differ with respect toheir respiratory fates, and lethality in Hoxa1 mutantsnvolves the elimination of parafacial structures origi-ating from r4 rather than from r5.

In vivo and in vitro analysis show that the cau-al medulla including the PBC is not affected by therox20−/− and Hoxa1−/− mutations (Jacquin et al.,996; Domınguez del Toro et al., 2001; Chatonnett al., 2003). Opioid peptides, such as enkephalins,cting through � and � receptors of the PBC inducepowerful depressant modulatory control of respi-

atory rhythm (Bianchi et al., 1995; Manzke et al.,003; Mellen et al., 2003; Onimaru and Homma, 2003).his system remains functional after the impairmentf para-facial structures. Subcutaneous administrationf the opioid antagonist naloxone to Krox-20−/− andoxa1−/− animals increased the respiratory frequency

or several hours and definitively reversed the occur-

ence of apneas and lethality (Jacquin et al., 1996;omınguez del Toro et al., 2001; Chatonnet et al.,003). These observations indicate that neuromodu-ation by endogenous opioids contributes to the gen-

aim

& Neurobiology 154 (2006) 37–46 39

ration of life-threatening apneas that develop aftermpairment of the parafacial system and that antago-izing the enkephalinergic transmission might providen effective pharmacological tool to be used when theeonatal anti-apneic function is disrupted.

Interestingly, progenitor cells mis-specified inoxa1−/− mutants were not necessarily eliminateduring later developmental stages and their develop-ental fate mutants could be identified by examining

ell migration patterns (Domınguez del Toro et al.,001). In wild-type E11.5 embryos, facial motoneu-ons migrate caudally to form the facial motor nucleus,hereas trigeminal motoneurons migrate dorsally. In

he mutated r4 region, a greatly reduced number ofacial motoneurons migrate caudally and an abnormalrigeminal-like lateral migration of r4 motoneurons isetected at E11.5 and completed at about E12.5. Thus,eurons with an anteriorized trigeminal-like pheno-ype appear to survive at the para-facial level of Hoxa1

utant hindbrains. Using thick horizontal brainstemlices, the effects of pharmacological stimulation ofhe r4-derived dorsal pons on the respiratory activ-ty recorded from motor cranial nerve were comparedetween control and Hoxa1 mutants (Domınguez deloro et al., 2001). In contrast to wild type preparations

n which such stimulations had no effect on the rhythmrequency, stimulation in the mutant induced a robustncrease in frequency, followed by a transient inhibitionf the rhythm. In the mutants, labelling the VRG withiI revealed an axonal pathway and running laterally

n the mutant pons that is not present in the wild-type.hese results strongly suggest the presence of super-umerary functional efferent connections of the dorsalons to the rhythm generator (Domınguez del Toro etl., 2001; Chatonnet et al., 2003). These observationsive insights into the crucial role of Krox20/Hoxa1 onarly phenotypic choices that affect parafacial neuronalrogeny, leaving the PBC functionally intact.

. Development of para-facial rhythmontrolling systems in the segmentedhombencephalon of the chick

Krox20 and Hox expression, boundary formationnd neurogenesis do not take place at the same timen adjacent rhombomeres. Between E9.5 and E10.5 in

ice, and between HH10 and HH15-20 in chick, r3

4 siology

dbaS2cnsshhniWt

wmn1eieasecg1

3dn

fwtptatwedpas

eTf(csbdrcrbw(

3p

rbrpiisaeleprrtebi

etrbTu

0 F. Chatonnet et al. / Respiratory Phy

isplays a marked delay compared with even rhom-omeres in the timing of neuronal differentiationnd axonal outgrowth (Lumsden and Keynes, 1989;chneider-Maunoury et al., 1993; Giudicelli et al.,001). Therefore, heterochrony of neurogenic pro-esses at the parafacial level of the neuraxis allowseuronal differentiation in r4 and continuing expres-ion of segmental genes such as Krox20 in r3 at theame developmental stage. Given that rhombomericeterochrony is found in both chick and mouse, weypothesize that there are likely to be conserved sig-aling interactions by which the expression of Krox20n r3 may influence neural circuits developing in r4.

e used a chick embryological approach to investigatehese signaling interactions.

Intrinsic neuronal activity starts in the neural tubeith a primordial low frequency rhythm (LF, rangein−1) that can be recorded from branchiomotor

erves in isolated hindbrain preparation (Fortin et al.,994; see Fig. 1A, top trace). In chick, the maturation ofmbryonic central pattern generators of the hindbrains manifest in the course of embryonic day 6 (E6) whenach LF burst becomes suddenly followed by cycles ofdditional high frequency burst discharges (range s−1,ee Fig. 1A, bottom trace) that form an episode (Fortint al., 1994). Acute transversal sections of the rhomben-ephale show that in chick this episodic activity is firstenerated at the para-facial level (Fortin et al., 1995,999).

.1. Embryological approach to understandingevelopment of rhythm generators: isolation ofeural tube segments and AP reversal of r4

To investigate development of circuits responsibleor generation of episodic activities in chick embryos,e isolated segments of the neural tube and inverted

he anterior/posterior polarity of rhombomeres justrior to the beginning of neurogenesis. The neuralube, dissected free of adherent tissue was transectedt the level of rhombomere boundaries, making surehat the boundary and the immediately adjacent cellsere excised; the anterior–posterior orientation of the

xcised rhombomere was marked by leaving tags of

orsal ectoderm on one-half of the explant. For trans-lantation, stage-matched host eggs were windowednd a gap in the neural tube was created at the desiredite of insertion. The graft was then inserted, the ori-

iort

& Neurobiology 154 (2006) 37–46

ntation of the ectoderm being maintained or inverted.he egg was resealed and incubated for either 24 h,

or whole-mount in situ hybridization, or 5–6 daysuntil the embryo reached E7) for electrophysiologi-al recording. Preparations for electrophysiology con-isted of either whole hindbrains or isolated rhom-omere islands (particularly r3r4 segments), made byissecting out the tissues and transferring them to aecording chamber. Results suggest that a rhombomericode is required for the assembly of a specific neu-onal circuit within the hindbrain. Interestingly, assem-ling the para-facial rhythm-promoting episodic net-ork appears to require two-segment functional units

r3r4) in the chick embryo (Fortin et al., 1999).

.2. The inductive role of Krox20 on thearafacial rhythm generator

Starting with the r3r4 pair necessary for episodichythm generation, we have replaced homotopically r3y either r1 or r5 (Coutinho et al., 2004). An episodichythm resulted in r5r4 segment pairs, but not in r1r4airs that maintained an immature LF pattern. Thus, todentify a possible role of Krox20, which is expressedn r3 and r5 but not in r1 and which is known topecify aspects of r3 and r5 character (Giudicelli etl., 2001) we used a gain of function strategy. Welectroporated a Krox20 expression vector at the r1evel in donor embryos. Electroporated r1 were thenngrafted and isolated together with r4 to form r1r4airs, where Krox20 expression is restricted to the oddhombomere r1 (“r1[Krox20]”). An episodic rhythmesulted in r1[Krox20]r4 pairs indicating the instruc-ive role of Krox20 in the interaction with r4. Hence thexpression of Krox20 is sufficient to endow other rhom-omeres with the capacity to induce rhythmic activityn r4 (Coutinho et al., 2004; Borday et al., 2005).

An interesting aspect of this induction is the appar-nt anterior–posterior polarity requirement wherebyhe Krox20 expressing rhombomere can signal to evenhombomeres when these occupy a posterior (e.g. r3r4)ut not an anterior (e.g. r2r3 or r4r5) position in the pair.he chick LF activity is characterized by the high reg-larity of the rhythm, which was quantified by measur-

ng the mean coefficient of variation (CV, S.D./mean)f inter-LF burst intervals. Anterior–posterior (AP)eversal of r4 dramatically altered the development ofhe rhythm. Therefore, to identify which of the two

F. Chatonnet et al. / Respiratory Physiology & Neurobiology 154 (2006) 37–46 41

Fig. 1. Low frequency pattern and maturation of episodic activity is a cyclic-AMP dependant process in the chick embryo. (A) Integrated facialneurograms recorded at different developmental stages shows initially single burst LF discharges (upper trace) and episodes of activity 2.5 hlater (lower trace). Time calibration bar: 1 min. (B) Normal (black) and pharmacologically altered (red) maturation (abscissae: time in min) ofthe frequency of bursts (“Burst”: filled symbols, top traces) and episodes (“Epi”: empty symbols, bottom traces) during the time of the recordingsession (min, abscissae). In all cases, recordings start at E6 when episodes contain a single bursts (at about min−1). Each symbol correspondsto one cycle (one episode and one interval between episodes), in which the number of bursts or the number of episodes (n = 1) is divided bythe total duration of the cycle. Fast upward deflections of “Burst” frequency indicate incorporation of an additional burst discharge into theepisodes (see also Borday et al., 2003). The rate of pattern maturation was measured by the slope of burst frequency increase (linear fits in B1).(B) Pharmacological analysis of the maturation of episodic activity (see also Fortin et al., 1999). Tests (in red traces) are indicated by horizontalbar above tracings. (B1) A 45 min application of SQ22536 100 �M, a blocker of adenylate cyclase, reduces LF (Epi) and burst frequency. Afterreversal of inhibition, a 60 min delay is observed (indicated by vertical black and red lines corresponding to the incorporation of a fourth burstin episodes), despite an otherwise unchanged maturation rate (note parallelism of fits at the end of recording). (B2) An increased intracellularc es LF at

robtrraTb

tinft

oncentration of cAMP caused by forskoline (5 �M, 45 min) increaso that obtained normally at later developmental stages.

hombomeres in the r3r4 pair exerts directional controln the other we inverted the AP polarity of each rhom-omere independently. AP inverted (flipped, f) orienta-ion of r3, in the pair f(r3)r4, did not change the episodichythm. However, AP reversal of r4 in the pair r3f(r4)

esulted in both absence of episodic rhythm generationnd an unexpected loss of regularity in the LF rhythm.his experiment suggests that r3, not r4, appears toe the source of the inter-rhombomeric influence and

rpKi

nd the frequency of bursts so that the resulting activity corresponds

hat the developing rhythmogenic network originatesn r4, not r3 (Coutinho et al., 2004). Therefore, the sig-aling initiated by Krox20 expression, leading to theormation of rhombomere boundaries may influencehe wiring of para-facial rhythm generators. The hete-

ochronic development of the r3r4 segment provides arecise window of time, of about one day, during whichrox20-initiated signaling persists in r3 while neurons

n r4 start to differentiate.

4 siology

4a

rictcehTibpagnt

4tp

aihgmtmam

fddc122a2SALb

rm

rdiettmca(poaoaFEitrTcrorFod

4r

naifa1fii

2 F. Chatonnet et al. / Respiratory Phy

. Evolution of the low frequency primordialctivity in chick and mouse

The early mechanisms of regionalization and neu-ogenesis involve large domains of the neural tubencluding a variety of neuronal types and developingircuits. It is therefore likely that the Krox20 induc-ion targets different cell populations within r4. As aonsequence, to what extent murine respiratory gen-rators and the chick episodic generators derive fromomologous or distinct progenitors remains unknown.he mature respiratory function of rhythm generators

s difficult to assess in birds, in which recordings of therainstem in vitro have not, as yet, identified central res-iratory rhythm generators (see, for example Sturdy etl., 2003). Some comparison between chick and mouseenerators is nevertheless possible by considering howewly active generators interact with the ongoing LFhat appears first in the neural tube.

.1. Generation of episodes in the chick embryo isightly linked to the maturation of low frequencyrimordial activity

The general scheme of early rhythm developmentppears to be the same in chick and mouse, sincen both species, before the appearance of sustainedigh frequency patterns (HF), in vitro preparationsenerate LF, observed in all preparations at E13 inice and E6 in chicks. Therefore, understanding how

his transition from LF to HF occurs represents aajor challenge of research. Interestingly, the mech-

nisms involved seem to be different in chick andice.In fact, spontaneous activity occurring at low

requency and invading large territories has beenescribed in various species and a wide variety ofeveloping structures in the CNS such as the spinalord (O’Donovan and Landmesser, 1987; O’Donovan,999; Chub and O’Donovan, 1998; Branchereau et al.,000; Hanson and Landmesser, 2003; Ren and Greer,003; Yvert et al., 2004), the cortex (Garaschuk etl., 2000; Peinado, 2000), the hippocampus (Ben-Ari,001), the retina (Galli and Maffei, 1988; Katz and

hatz, 1996) and the brainstem (Fortin et al., 1994;badie et al., 2000; Gust et al., 2003). Therefore, theF activity observed in the embryonic chick and mouserainstem probably corresponds to an immature, non-

dnmp

& Neurobiology 154 (2006) 37–46

espiratory primordial activity that is required for theaturation of neuronal circuits.In keeping with this hypothesis, the Hox-related

egulations do not greatly affect LF. In fact, primor-ial rhythms persists when Krox20 expression is eithernduced or inactivated in chick embryos (Coutinhot al., 2004). Nevertheless, pharmacological observa-ions show that the generation of episodes in chick isightly linked to the maturation of LF. In these experi-

ents maturation of motor output over a 2.5 h windowould be followed in the same experiments (Borday etl., 2003; Fig. 1B). Over this period, isolated burstsFig. 1A, top trace) were replaced by episodes com-rising four bursts (Fig. 1A, bottom trace). Formationf episodes (“Burst” in Fig. 1B1; incorporation of andditional burst in episodes leads to the upward shiftf the “Burst” curve) is consistently associated withdecrease of LF (Coutinho et al., 2004; see “Epi” inig. 1B1 and compare top and bottom traces in Fig. 1A).pisodes result from the synaptic activity of LF burst-

ng neurons initiating a sequence of GABAergic inhibi-ion, activation of the hyperpolarization-activated cur-ent IH and rebound excitation (Fortin et al., 1999).herefore, episodes are affected by pharmacologi-al manipulation of synaptic inhibitions via GABA-Aeceptors, of time intervals between bursts via IH, andf time intervals between episodes via beta-adrenergiceceptors and adenylate-cyclase (see Fortin et al., 1999;ig. 1B). These experiments suggest that maturationf parafacial episodes in the chick is an LF activity-ependent process.

.2. Emergence of the pre-Botzinger respiratoryhythm generator in the mouse embryo

In rodents, the PBC, a region immunoreactive for theeurokinin receptor NK1 (Gray et al., 1999; Wang etl., 2001; Stornetta et al., 2003; Pagliardini et al., 2003),s thought to contain all of the necessary elementsor respiratory rhythm generation in rodents (Smith etl., 1991; Koshiya and Guyenet, 1996; Ramirez et al.,998; Lieske et al., 2000). This group is readily identi-able and amenable to analysis in vitro in preparations

n which the pFRG is eliminated by a transection cau-

al to the facial motor nucleus. The respiratory neuraletwork contains rhythmic neurons connected by gluta-atergic synapses, some of them exhibiting pacemaker

roperties (Johnson et al., 1994; Koshiya and Smith,

siology

1Rnt2r

aoietalvriLncmetaagi1tisteaMet(smtrtmK

ed

HameaPNottttEf

5

bcotaceisitanptiermuetFre

F. Chatonnet et al. / Respiratory Phy

999; Thoby-Brisson et al., 2000; Thoby-Brisson andamirez, 2001; Pena et al., 2004). The destruction ofeurons expressing the NK1R leading to ataxic respira-ion (Gray et al., 2001) and sleep apneas (McKay et al.,005) supports the view that the PBC plays a primaryole in respiratory rhythmogenesis.

Using immunohistochemical, electrophysiologicalnd calcium imaging techniques on transverse slicesf the brainstem, recent evidence showed that the PBCs anatomically and functionally defined in the mousembryo at E15 (Thoby-Brisson et al., 2005). The spa-ial organization of neuronal circuits supporting LF waslso examined at the same stage. LF bursts arose uni-aterally from a dorso-medial area flanking the fourthentricle. The LF events were initiated from either theight or the left sides of the slice and often alternatedn time from one side to the other. Generation of thisF activity in the embryonic mouse brainstem doesot seem to rely on single anatomical structures or spe-ific circuits since it was observed in both facial andedullary slices. In contrast, at later stages, the bilat-

ral PBC rhythm generation site was spatially restrictedo a region ventral to the nucleus ambiguous whosectivity is preserved in a transverse slice (Smith etl., 1991). Rhythm generation depends on glutamater-ic transmission through AMPA/kainate receptors, asn older preparations (Gray et al., 1999; Smith et al.,991; Koshiya and Smith, 1999). This glutamatergicransmission is concomitant to establishment of NK1Rmmunoreactivity in this region. At this stage, sub-tance P and the opioid agonist DAMGO exert respec-ive excitatory and depressing neuromodulatory influ-nces on the rhythm (Gray et al., 1999; Janczewski etl., 2002; Mellen et al., 2003; Manzke et al., 2003).oreover the onset of respiratory activity in the mouse

mbryo during E15 corresponds to the developmen-al stage at which fetal movements become detectableSuzue, 1996; Abadie et al., 2000). In keeping with thehorter gestational period of mice compared to rats,ouse developmental time E15 probably corresponds

o rat E17 (Viemari et al., 2003), when formation of theespiratory rhythm generator is completed and beginso drive respiratory-like activity and fetal breathing

ovements (Di Pasquale et al., 1992; Greer et al., 1992;

obayashi et al., 2001; Pagliardini et al., 2003).

Therefore, calcium imaging of the embryonic PBCstablishes that the HF generator in mice does noterive from the pre-existing LF generating network.

rago

& Neurobiology 154 (2006) 37–46 43

owever, none of the characteristics tested so far havellowed identification of a candidate population ofigrating neural cells destined to form this HF gen-

rator prior to E15, at a stage when the LF generator islready active. In a previous study in the rat embryo,agliardini et al. (2003) similarly reported the lack ofK1 positive cells in the PBC location before the onsetf HF activity. There seems to be no neuron exhibitinghe PBC-like phenotype that would be located betweenhe germinative zone and the final ventrolateral loca-ion of the PBC. This strongly supports the view thathe entire set of PBC characteristics differentiates at15 in ventro-lateral post-migratory neurons, distant

rom the dorso-medial sites where LF is generated.

. Conclusion

It seems that a conserved signaling pathway initiatedy Krox20 expression induces development of parafa-ial rhythm generators, thereby establishing homologyf the parafacial territory in chick as in mouse. Inhis respect, it is important to note that we have beenble to induce formation of an episodic activity inhick by mis-expressing mouse mKrox-20 (Coutinhot al., 2004). Thus, major molecular pathways defin-ng the parafacial territory are compatible in the twopecies. Homology between the generators themselvess unclear. Rhombomeres are polyclonal developmen-al units that include several progenitor types (Clarkend Lumsden, 1993; Glover, 2001) and there areo cell-specific markers available to distinguish sub-opulations of developing parafacial rhythm genera-ors. Using neuronal activity as a marker, electrophys-ological and imaging studies show that rhythm gen-rators derive from subpopulations of progenitors inhombomeres which otherwise include various otherotor, reticular and sensory neuronal lineages. It is

nlikely that the lineage forming rhythm generatorsxhibit specific competency with respect to the induc-ive signals initiated by the expression of Krox20.or example, all growing axons of r4-derived neu-ons avoid territories expressing Krox20 or avoid thectopic patches of Krox20 expression after electropo-

ation (Giudicelli et al., 2001). Thus, indiction targetsre rhombomere- rather than cell type-specific. Theene encoding EphA4, a member of the Eph familyf the tyrosine kinase receptor might be interesting to

4 siology

igieiEtnEta

toaaistaaecaIt(api2ic2athvpiitp2imtr

cgipt

A

E0M

R

A

B

B

B

B

B

B

B

C

C

4 F. Chatonnet et al. / Respiratory Phy

nvestigate (Coutinho et al., 2004). It is a known tar-et of the Krox20 transcription factor. It is expressedn odd rhombomeres while their ephrin ligands arexpressed in alternating even rhombomeres and partic-pate in inter-rhombomeric cell sorting. Expression ofph receptors and their ephrin ligands have been shown

o affect neuronal wiring at several levels of the centralervous system, particularly in the spinal cord, wherephA4-positive neurons are excitatory components of

he locomotor central pattern generator (Kullander etl., 2003).

Parafacial signalling in chick induces bursting pat-erns resembling episodic breathing, a common featuref many adult ectotherme vertebrates including mostirbreathing fish, amphibians and turtles. Adult birdsnd mammals exhibit regular and permanent breath-ng patterns even though they have very different lungtructure (Powell and Scheid, 1986) a different func-ional organization of the respiratory cycle (Fortin etl., 1995; Sturdy et al., 2003). Observations in chicknd mouse embryos (Abadie et al., 2000) show thatpisodic patterns are present in these animals at pre-ise stages of development, suggesting that permanentnd episodic breathing may co-exist in all vertebrates.t is interesting to note that under specific environmen-al conditions, adult mammals may breathe in episodesduring hibernation, for example) while amphibiansnd reptiles can exhibit continuous breathing patterns,articularly when respiratory drive is elevated as dur-ng hypoxic or hypercarbic exposure (Milsom et al.,004). In the adult frog, lung ventilation is initiatedn a parafacial-like location, just caudal to the sixthranial nerve root (McLean et al., 1995; Wilson et al.,002) and lung inflations are random events that comelone or sometimes in doublets. Lung ventilations tendo cluster into episodes upon exposure to hypoxic orypercapnic gases, a phenomenon observed both inivo and in vitro (Milsom, 1991). In anurans, rostralontine structures, such as the nucleus isthmi, may benvolved in neuromodulation and play an important rolen controlling ventilatory chemo-sensory responses andhe passage between permanent and episodic breathingatterns (Milsom et al., 2004; Gargaglioni and Branco,004). In the developing mouse, the permanent breath-

ng pattern appears to result from a selective developpe-

ental program expressed in the r6–r8 derived neuralube (Borday et al., in press), while certain traits of the3r4 para-facial rhythmic neurons and pontine episodic

C

& Neurobiology 154 (2006) 37–46

ontrols may be conserved. Imaging studies of rhythmenerators in murine embryos and heterospecific graft-ng on murine hindbrain segments (Borday et al., inress) are currently being performed in our laboratoryo further explore this possibility.

cknowledgements

This work was supported by CNRS, ACI BDPI#57,.C. grant “Brainstem Genetics” QLG2-CT-2001-1467, C.B. is supported by the FRM and L.W. by theRT.

eferences

badie, V., Champagnat, J., Fortin, F., 2000. Branchiomotor activi-ties in mouse embryo. Neuroreport 11, 141–145.

en-Ari, Y., 2001. Developing networks play a similar melody.Trends Neurosci. 24, 353–360.

ianchi, A.L., Denavit-Saubie, M., Champagnat, J., 1995. Centralcontrol of breathing in mammals: neuronal circuitry, membraneproperties and neurotransmitters. Physiol. Rev. 75, 1–45.

orday, C., Abadie, V., Chatonnet, F., Thoby-Brisson, M., Cham-pagnat, J., Fortin, G., 2003. Developmental molecular switchesregulating breathing patterns in CNS. Respir. Physiol. Neurobiol.135, 121–132.

orday, C., Chatonnet, F., Thoby-Brisson, M., Champagnat, J.,Fortin, G., 2005. Neural tube patterning by Krox20 and emer-gence of a respiratory control. Respir. Physiol. Neurobiol. 149,63–72.

orday, C., Coutinho, A., Gernon, I., Champagnat, J., Fortin, G.Pre-/post-otic rhombomeric interactions control emergence of afetal-like respiratory rhythm in the mouse embryo. J. Neurobiol.,in press.

orday, V., Kato, F., Champagnat, J., 1997. A ventral pontine path-way promotes rhythmic activity in the medulla of neonate mice.NeuroReport 8, 3679–3683.

ranchereau, P., Morin, D., Bonnot, A., Ballion, B., Chapron, J.,Viala, D., 2000. Development of lumbar rhythmic networks: fromembryonic to neonate locomotor-like patterns in the mouse. BrainRes. Bull. 53, 711–718.

hatonnet, F., Domınguez del Toro, E., Voiculescu, O., Charnay, P.,Champagnat, J., 2002. Different respiratory control systems areaffected in homozygous and heterozygous kreisler mutant mice.Eur. J. Neurosci. 15, 684–692.

hatonnet, F., Dominguez del Toro, E., Thoby-Brisson, M., Cham-pagnat, J., Fortin, G., Rijli, F.M., Thaeron-Antono, C., 2003.From hindbrain segmentation to breathing: developmental pat-

tern in rhombomeres 3 and 4. Mol. Neurobiol. 28, 277–294.

hub, N., O’Donovan, M.J., 1998. Blockade and recovery of spon-taneous rhythmic activity after application of neurotransmitterantagonists to spinal networks of the chick embryo. J. Neurosci.18, 294–306.

siology

C

C

D

D

D

F

F

F

G

G

G

G

G

G

G

G

G

H

H

J

J

J

K

K

K

K

K

L

L

L

M

M

M

M

M

F. Chatonnet et al. / Respiratory Phy

larke, J.D., Lumsden, A., 1993. Segmental repetition of neuronalphenotype sets in the chick embryo hindbrain. Development 118,151–162.

outinho, A.P., Borday, C., Gilthorpe, J., Jungbuth, S., Champag-nat, J., Lumsden, A., Fortin, G., 2004. Induction of a parafacialrhythm generator by rhombomere 3 in the chick embryo. J. Neu-rosci. 24, 9383–9390.

enavit-Saubie, M., Champagnat, J., Zieglgansbezrger, W., 1978.Effects of opiates and methionin-enkephalin on pontine and bul-bar respiratory neurones of the cat. Brain Res. 155, 55–67.

i Pasquale, E., Monteau, R., Hilaire, G., 1992. In vitro study ofcentral respiratory-like activity of the fetal rat. Exp. Brain Res.89, 459–464.

omınguez del Toro, E., Borday, V., Davenne, M., Neun, R.,Rijli, F.M., Champagnat, J., 2001. Generation of a novel func-tional neuronal circuit in Hoxa1 mutant mice. J. Neurosci. 21,5637–5642.

ortin, G., Champagnat, J., Lumsden, A., 1994. Onset and maturationof branchiomotor activities in the chick hindbrain. NeuroReport5, 1149–1152.

ortin, G., Kato, F., Lumsden, A., Champagnat, J., 1995. Rhythmgeneration in the segmented hindbrain of chick embryos. J. Phys-iol. 486, 735–744.

ortin, G., Jungbluth, S., Lumsden, A., Champagnat, J., 1999. Seg-mental specification of GABAergic inhibition during develop-ment of hindbrain neural networks. Nat. Neurosci. 2, 873–877.

alli, L., Maffei, L., 1988. Spontaneous impulse activity of rat retinalganglion cells in prenatal life. Science 242, 90–101.

araschuk, O., Linn, J., Eilers, J., Konnerth, A., 2000. Large-scaleoscillatory calcium waves in the immature cortex. Nat. Neurosci.3, 452–459.

argaglioni, L.H., Branco, L.G.S., 2004. Nucleus isthmi and con-trol of breathing in amphibians. Respir. Physiol. Neurobiol. 143,177–186.

iudicelli, F., Taillebourg, E., Charnay, P., Gilardi-Hebenstreit,P., 2001. Krox-20 patterns the hindbrain through both cell-autonomous and non cell-autonomous mechanisms. Genes Dev.15, 567–580.

lover, J.C., 2001. Correlated patterns of neuron differentiation andHox gene expression in the hindbrain: a comparative analysis.Brain Res. Bull. 55, 683–693.

ray, P.A., Rekling, J.C., Bocchiaro, C.M., Feldman, J.L., 1999.Modulation of respiratory frequency by peptidergic input torhythmogenic neurons in the pre-Botzinger complex. Science286, 1566–1568.

ray, P.A., Janczewski, W.A., Mellen, N., McCrimmon, D.R., Feld-man, J.L., 2001. Normal breathing requires preBotzinger com-plex neurokinin-1 receptor-expressing neurons. Nat. Neurosci. 4,927–930.

reer, J.J., Smith, J.C., Feldman, J., 1992. Respiratory and locomotorpatterns generated in the fetal rat brainstem-spinal cord in vitro.J. Neurophysiol. 67, 996–999.

ust, J., Wright, J.J., Pratt, E.B., Bosma, M.M., 2003. Developmentof synchronized activity of cranial motor neurons in the seg-mented embryonic mouse hindbrain. J. Physiol. 550, 123–133.

amburger, V., Hamilton, H.L., 1951. A series of normal stages inthe development of the chick embryo. J. Morphol. 88, 49–92.

M

M

& Neurobiology 154 (2006) 37–46 45

anson, M.G., Landmesser, L.T., 2003. Characterization of thecircuits that generate spontaneous episodes of activity in theearly embryonic mouse spinal cord. J. Neurosci. 23, 587–600.

acquin, T.D., Borday, V., Schneider-Maunoury, S., Topilko, P.,Ghilini, G., Kato, F., Charnay, P., Champagnat, J., 1996. Reorgan-isation of pontine rhythmogenic neuronal networks in Krox-20knockout mice. Neuron 17, 747–758.

anczewski, W.A., Onimaru, H., Homma, I., Feldman, J.L., 2002.Opioid-resistant repiratory pathway from the preinspiratory neu-rons to abdominal muscles: in vivo and in vitro study in newbornrat. J. Physiol. (Lond.) 545, 1017–1026.

ohnson, S.M., Smith, J.C., Funk, G.D., Feldman, J.L., 1994. Pace-maker behavior of respiratory neurons in medullary slices fromneonatal rat. J. Neurophysiol. 72, 2598–2608.

atz, L.C., Shatz, C.J., 1996. Synaptic activity and the constructionof cortical circuits. Science 274, 1133–1138.

obayashi, K., Lemke, R.P., Greer, J.J., 2001. Ultrasound mea-surements of the fetal breathing movements in the rat. J. Appl.Physiol. 91, 316–320.

oshiya, N., Guyenet, P.G., 1996. Tonic sympathetic chemoreflexafter blockade of respiratory rhythmogenesis in the rat. J. Physiol.(Lond.) 491, 859–869.

oshiya, N., Smith, J.C., 1999. Neuronal pacemaker for breathingvisualized in vitro. Nature 400, 360–363.

ullander, K., Butt, S.J., Lebret, J.M., Lundfald, L., Restrepo, C.E.,Rydstrom, A., Klein, R., Kiehn, O., 2003. Role of EphA4 andEphrinB3 in local neuronal circuits that control walking. Science299, 1889–1892.

ieske, S.P., Thoby-Brisson, M., Telgkamp, P., Ramirez, J.M.,2000. Reconfiguration of the neural network controlling mul-tiple breathing patterns: eupnea, sighs and gasps. Nat. Neurosci.3, 600–607.

umsden, A., Keynes, R., 1989. Segmental patterns of neuronaldevelopment in the chick hindbrain. Nature 337, 424–428.

umsden, A., Krumlauf, R., 1996. Patterning the vertebrate neuraxis.Science 274, 1109–1115.

anzke, T., Guenther, U., Ponimaskin, E.G., Haller, M.,Dutschmann, M., Schwarzacher, S., Richter, D.W., 2003. 5-HT4(a) receptors avert opioid-induced breathing depressionwithout loss of analgesia. Science 301, 226–229.

cKay, L.C., Janczewski, W.A., Feldman, J.L., 2005. Sleep-disordered breathing after targeted ablation of preBotzinger com-plex neurons. Nat. Neurosci. 8, 1142–1144.

cLean, H.A., Perry, S.F., Remmers, J.E., 1995. Two regions in theisolated brainstem of the frog that modulate respiratory-relatedactivity. J. Comp. Physiol. A 177, 135–144.

ellen, N.M., Janczewski, W.A., Bocchiaro, C.M., Feldman, J.L.,2003. Opioid-induced quantal slowing reveals dual networks forrespiratory rhythm generation. Neuron 37, 821–826.

ilsom, W.K., 1991. Intermittent breathing in vertebrates. Annu.Rev. Physiol. 53, 87–105.

ilsom, W.K., Chatburn, J., Zimmer, M.B., 2004. Pontine influ-ences on respiratory control in ectothermic and heterothermicvertebrates. Respir. Physiol. Neurobiol. 143, 263–280.

orin-Surun, M.-P., Boudinot, E., Gacel, G., Champagnat, J.,Roques, B.P., Denavit-Saubie, M., 1984. Different effects of mu

4 siology

M

O

O

O

O

P

P

P

P

R

R

S

S

S

S

S

T

T

T

T

V

W

W

6 F. Chatonnet et al. / Respiratory Phy

and delta opiate agonists on respiration. Eur. J. Pharmacol. 98,235–240.

orin-Surun, M.-P., Boudinot, E., Dubois, C., Matthes, H.W., Kief-fer, B.L., Denavit-Saubie, M., Champagnat, J., Foutz, A.S., 2001.Respiratory function in adult mice lacking the mu-opioid recep-tor: role of delta-receptors. Eur. J. Neurosci. 13, 1703–1710.

’Donovan, M.J., Landmesser, L., 1987. The development ofhindlimb motor activity studied in the isolated spinal cord ofthe chick embryo. J. Neurosci. 7, 3256–3264.

’Donovan, M.J., 1999. The origin of spontaneous activity in devel-oping networks of the vertebrate nervous system. Curr. Opin.Neurobiol. 9, 94–104.

nimaru, H., Homma, I., 2003. A novel functional neuron group forrespiratory rhythm generation in the ventral medulla. J. Neurosci.23, 1478–1486.

ppenheim, R.W., 1991. Cell death during development of the ner-vous system. Annu. Rev. Neurosci. 14, 453–501.

agliardini, S., Ren, J., Greer, J.J., 2003. Ontogeny of the pre-Botzinger complex in perinatal rats. J. Neurosci. 23, 9575–9584.

einado, A., 2000. Traveling slow waves of neural activity: a novelform of network activity in developing neocortex. J. Neurosci.20, 1–6, RC54.

ena, F., Parkis, M.A., Tryba, A.K., Ramirez, J.M., 2004. Differ-ential contribution of pacemaker properties to the generation ofrespiratory rhythms during normoxia and hypoxia. Neuron 43,105–117.

owell, F.L., Scheid, P., 1986. Physiology of gas exchange in theavian respiratory system. In: King, A.S., McLellan, J. (Eds.),Form and Function in Birds, vol. 4. Academic Press, London, SanDiego, Berkeley, Boston, Sydney, Tokyo, Toronto, pp. 393–437.

amirez, J.M., Schwarzacher, S.W., Pierrefiche, O., Olivera, B.M.,Richter, D.W., 1998. Selective lesioning of the cat pre-Botzingercomplex in vivo eliminates breathing but not gasping. J. Physiol.(Lond.) 507, 895–907.

en, J., Greer, J.J., 2003. Ontogeny of rhythmic motor patterns gen-erated in the embryonic rat spinal cord. J. Neurophysiol. 89,

1187–1195.

chneider-Maunoury, S., Topilko, P., Seitandou, T., Levi, G., Cohen-Tannoudji, M., Pournin, S., Babinet, C., Charnay, P., 1993. Dis-ruption of Krox-20 results in alteration of rhombomeres 3 and 5in the developing hindbrain. Cell 17, 1199–1214.

Y

& Neurobiology 154 (2006) 37–46

mith, J.C., Ellenberger, H.H., Ballanyi, K., Richter, D.W., Feld-man, J.L., 1991. Pre-Botzinger complex: a region that maygenerate respiratory rhythm in mammals. Science 254, 726–729.

tornetta, R.L., Rosin, D.L., Wang, H., Sevigny, C.P., Weston,M.C., Guyenet, P.G., 2003. A group of glutamatergic interneu-rons expressing high levels of both neurokinin-1 receptors andsomatostatin identifies the region of the pre-Botzinger complex.J. Comp. Neurol. 455, 499–512.

turdy, C.B., Wild, J.M., Mooney, R., 2003. Respiratory and telen-cephalic modulation of vocal motor neurons in the zebra finch.J. Neurosci. 23, 1072–1086.

uzue, T., 1996. Movements of mouse fetuses in early stages ofneural development studied in vitro. Neurosci. Lett. 218, 131–134.

anabe, Y., Jessell, T.M., 1996. Diversity and pattern in the develop-ing spinal cord. Science 274, 1115–1123.

hoby-Brisson, M., Ramirez, J.M., 2001. Identification of two typesof inspiratory pacemaker neurons in the isolated respiratory neu-ral network of mice. J. Neurophysiol. 86, 104–112.

hoby-Brisson, M., Telgkamp, P., Ramirez, J.M., 2000. The role ofthe hyperpolarization-activated current in modulating rhythmicactivity in the isolated respiratory network of mice. J. Neurosci.20, 2994–3005.

hoby-Brisson, M., Trinh, J.B., Champagnat, J., Fortin, G., 2005.Emergence of the pre-Botzinger respiratory rhythm generator inthe mouse embryo. J. Neurosci. 25, 4307–4318.

iemari, J.C., Burnet, H., Bevengut, M., Hilaire, G., 2003. Perinatalmaturation of he mouse respiratory rhythm-generator: in vivoand in vitro studies. Eur. J. Neurosci. 17, 1233–1244.

ang, W.G., Stornetta, R.L., Rosin, D.L., Guyenet, P.G., 2001.Neurokinin-1 receptor-immunoreactive neurons of the ventralrespiratory group in the rat. J. Comp. Neurol. 434, 128–146.

ilson, R.J., Vasilakos, K., Harris, M.B., Straus, C., Remmers,J.E., 2002. Evidence that ventilatory rhythmogenesis in the froginvolves two distinct neuronal oscillators. J. Physiol. (Lond.) 540,

557–570.

vert, B., Branchereau, P., Meyrand, P., 2004. Multiple spontaneousrhythmic activity patterns generated by the embryonic mousespinal cord occur within a specific developmental time window.J. Neurophysiol. 91, 2101–2109.