Clonal analysis reveals a common origin betweennonsomite-derived neck muscles andheart myocardiumFabienne Lescroarta,1, Wissam Hamoua, Alexandre Francoub, Magali Théveniau-Ruissyb, Robert G. Kellyb,and Margaret Buckinghama,2

aUnité de Génétique Moléculaire du Développement, CNRS Unité de Recherche Associée 2578, Institut Pasteur, 75015 Paris, France; and bAix-MarseilleUniversité, Institut de Biologie du Développement de Marseille, CNRS Unité Mixte de Recherche 7288, 13288 Marseille, France

Contributed by Margaret Buckingham, December 23, 2014 (sent for review November 10, 2014; reviewed by Michael Levine and Drew Noden)

Neck muscles constitute a transition zone between somite-derivedskeletal muscles of the trunk and limbs, and muscles of the head,which derive from cranial mesoderm. The trapezius and sterno-cleidomastoid neck muscles are formed from progenitor cells thathave expressed markers of cranial pharyngeal mesoderm, whereasother muscles in the neck arise from Pax3-expressing cells in thesomites.Mef2c-AHF-Cre genetic tracing experiments and Tbx1mu-tant analysis show that nonsomitic neck muscles share a generegulatory network with cardiac progenitor cells in pharyngeal me-soderm of the second heart field (SHF) and branchial arch-derivedhead muscles. Retrospective clonal analysis shows that this groupof neck muscles includes laryngeal muscles and a component of thesplenius muscle, of mixed somitic and nonsomitic origin. We dem-onstrate that the trapezius muscle group is clonally related to myo-cardium at the venous pole of the heart, which derives from theposterior SHF. The left clonal sublineage includes myocardium ofthe pulmonary trunk at the arterial pole of the heart. Althoughmuscles derived from the first and second branchial arches alsoshare a clonal relationship with different SHF-derived parts of theheart, neck muscles are clonally distinct from these muscles anddefine a third clonal population of common skeletal and cardiacmuscle progenitor cells within cardiopharyngeal mesoderm. By link-ing neck muscle and heart development, our findings highlight theimportance of cardiopharyngeal mesoderm in the evolution of thevertebrate heart and neck and in the pathophysiology of humancongenital disease.

neckmuscles | myocardium | retrospective clonal analysis | mouse embryo | Tbx1

In mammals, all skeletal muscles are not equivalent. Entry intothe skeletal muscle program and subsequent differentiation

depend on transcription factors of the MyoD (myogenic differ-entiation 1) family. However, upstream regulators of the myo-genic program differ in different parts of the body. Skeletalmuscles of the trunk and limbs derive from the somites and thusare descendants of progenitors expressing Pax3, a paired boxtranscription factor that plays a major role in the control ofmyogenesis (1, 2). In Pax3−/−;Myf5−/− double mutants mostskeletal muscles are lost. However, muscles in the cranial part ofthe embryo are not affected in these mutant mice (3), showingthat myogenesis in head skeletal muscles is controlled by a dif-ferent upstream genetic network (2). Indeed, head muscles areformed from cranial mesoderm, derived from Pax3−;Mesp1+

(mesoderm posterior homolog transcription factor 1) cells (4, 5).Many of these myogenic progenitors express genes, includingIslet1 (Isl1 transcription factor), Nkx2-5, or Tbx1 (T-box tran-scription factor 1), that are also expressed in cardiac progenitorcells of the second heart field (SHF) within pharyngeal meso-derm populations that form myocardium at the poles of the heart(6, 7). Among these regulators, Tbx1 is required for developmentof both the arterial pole of the heart and head muscle specifi-cation (8–10).

Masticatory and facial expression muscles derive from themesodermal core that extends into the first and second branchial(or pharyngeal) arches, respectively (11). These arches also con-tain SHF cardiac progenitors (12). Cells with divergent cardiacand skeletal muscle fates in this cardiopharyngeal mesoderm aresimilarly labeled by genetic tracing (13). Furthermore, retrospec-tive clonal analysis in the mouse embryo has shown that masti-catory muscles of the head and right ventricular myocardiumoriginate from common progenitor cells, whereas facial expres-sion muscles share a common clonal origin with the arterial poleof the heart, such that myocardium at the base of the pulmonarytrunk or the aorta is clonally related to left or right facial ex-pression muscles, respectively (14). Cardiopharyngeal mesodermcells that give rise to both cardiac and skeletal muscle derivativesare already present in urochordates, such as the ascidian Cionaintestinalis, where the equivalent of an SHF can be distinguished(15). A number of genes, including homologs of vertebrate Islet1and Tbx1, are expressed in ascidian cardiopharyngeal mesodermand form a gene regulatory network that governs both heart andpharyngeal muscle formation (16, 17).Muscles of the neck play an essential role in coordination of

movement between the head and trunk. In this transition zone,muscles of both branchial and somitic origin are found. Myo-blasts in the more caudal branchial arches (3rd, 4th, and 6th) arethought to give rise to neck muscles such as the cucullaris muscle,

Significance

Head muscles, derived from the first and second pharyngealarches, share common progenitors with myocardial cells of theheart. This is in contrast to somite-derived skeletal muscles ofthe trunk and limbs. Neckmuscles, located in the transition zonebetween head and trunk, have both a somitic and nonsomiticorigin. We now demonstrate a clonal relationship betweennonsomitic neck muscles and myocardial cells located in theatria, inflow and outflow regions of the heart. This is distinctfrom that of the two head muscle lineages. Formation of theseneck muscles, like those in the head, depends on a gene regu-latory network shared with myocardial progenitors. We thusidentify a third clonal group within cardiopharyngeal meso-derm, with implications for human malformations.

Author contributions: F.L. and M.B. designed research; F.L., W.H., A.F., and M.T.-R. per-formed research; F.L., R.G.K., and M.B. analyzed data; and F.L., R.G.K., and M.B. wrotethe paper.

Reviewers: M.L., University of California Berkeley; and D.N., Cornell University.

The authors declare no conflict of interest.1Present address: Institut de Recherche Interdisciplinaire en Biologie Humaine etMoléculaire,Université Libre de Bruxelles, B61070, Belgium.

2To whom correspondence should be addressed. Email: margaret.buckingham@pasteur.fr.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1424538112/-/DCSupplemental.

1446–1451 | PNAS | February 3, 2015 | vol. 112 | no. 5 www.pnas.org/cgi/doi/10.1073/pnas.1424538112

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 19

, 202

0

or its mammalian derivatives the trapezius and sternocleido-mastoidius, which, in keeping with a branchial origin, are in-nervated by a cranial nerve (18). Recent analysis of geneexpression and genetic tracing supports this view. The cucullarismuscle in birds and turtles derives from cells in adjacent lateralmesoderm that do not express Pax3 but express Tbx1 (19). De-spite evolutionary changes in cranial somites, this is also the casefor the progenitors of the trapezius and sternocleidomastoidmuscles in the mouse (19), where progenitors were also shown tohave expressed Islet1. Furthermore the trapezius muscle is missingin Tbx1 mutants (10, 19), consistent with an origin in cranial me-soderm rather than somites, in the mouse embryo.Here we address the question of a link between nonsomite-

derived neck muscles and the myocardium. Using genetic tracingin WT and Tbx1 mutant embryos and retrospective clonal anal-ysis, we characterize such muscles and show that neck muscleprogenitor cells share a common clonal origin with the venouspole of the heart. Our results identify a third common skeletaland cardiac progenitor population in cardiopharyngeal meso-derm, and we discuss the importance of this tissue in the evo-lution of the vertebrate heart and head/neck and the implicationsfor human pathology.

ResultsRegulatory Gene Expression. We first verified regulatory gene ex-pression in the progenitors of neck muscles, by using specific Credrivers with a conditional Rosa26R-nlacZ reporter. Pax3 is ex-pressed in myogenic progenitor cells in all somites (1), and as ex-pected, many neck muscles were labeled when the Rosa26R-nlacZ/+

reporter line was crossed with a Pax3Cre/+ line (Fig. 1 A and B),indicating that they originate from somites. However, the tra-pezius and sternocleidomastoid muscles are negative with onlyresidual labeling, which is due to connective tissue, as previouslyreported (19). With the Mesp1Cre/+ line, which is activated incardiac progenitors in the primitive streak (20), the heart and

some neck muscles, including the trapezius and sternocleido-mastoid muscles, are labeled (Fig. 1 C and D). The Mef2c(myocyte enhancer factor 2c) enhancer is activated in the ante-rior SHF (21), and when Mef2c-Cre transgenic mice are crossedwith the Rosa26R-nlacZ/+ reporter line, labeling is seen in headmuscles, notably those derived from the first branchial arch (14),as well as in myocardium of the pulmonary trunk, aorta, and rightventricle. The trapezius and sternocleidomastoid muscles are alsolabeled using this Cre line, as shown for the trapezius muscles onwhole-mount X-gal–stained embryos at embryonic day (E)14.5and E12.5 (Fig. 1 E and F). At E10.5 (Fig. 1G), labeling can beobserved in the posterior pharyngeal region, where trapeziusprogenitors are potentially present. In keeping with this, thesecells also express MyoD, indicating commitment to the skeletalmuscle lineage (1) (Fig. 1H). These results show that progenitorsof nonsomitic neck muscles express common markers with, andoriginate in close proximity to, cardiac progenitors.

No Clonal Relation with Head or Somite-Derived Muscles. To inves-tigate a potential clonal relationship between these neck muscles,muscles derived from the first and second branchial arches, andsomite-derived muscles of the neck and body, we performeda retrospective clonal analysis. This approach (Fig. S1A) dependson a rare intragenic recombination event in an nlaacZ sequencethat converts it to a functional nlacZ reporter (22). In this casethe nlaacZ sequence is targeted to an allele of the α-cardiac actin(αc-actin) gene (23), so that after recombination β-galactosidase(β-gal) positive cells will be generated in the myocardium and indeveloping skeletal muscles where the gene is expressed. Statis-tical analysis on the collection of embryos determines whether thefrequency of labeling corresponds to one or more recombinationevents (Fig. S1B). In the case of the former, β-gal positive cellsare clonally related and derive from a common progenitor.In a collection of 2,018 embryos with the αc-actin

nlaacZ allele,at E14.5, we found 30 embryos (1.5%) with labeling in the neck,limited to the trapezius and/or sternocleidomastoid muscles (Fig.2A). These muscles (Fig. 2B), referred to as the trapezius group,are often labeled together, and statistical analysis of the fre-quency of this labeling shows that they are a clonal unit (TableS1). When these embryos were sectioned, we observed that la-ryngeal muscles are also β-gal positive, suggesting a lineage re-lationship between these muscle groups (Fig. 3 B–E). Thesemuscles are not labeled in Rosa26R-nlacZ/+;Pax 3Cre/+ embryos(Fig. 3A), in keeping with a nonsomitic origin (11). No clonalrelationship was observed between the trapezius group of mus-cles and masticatory, or facial expression, head muscles (Fig. 2Aand Table 1), derived from the first and second branchial arches,respectively (14). Somitic neck muscles, marked by genetictracing with Pax3Cre, but not by the Mef2c-AHF-Cre transgene,and muscles in the trunk or limbs also showed no clonality withthe trapezius group (Fig. 2A and Table 1). This was also the casefor the tongue muscle, which is clonally related to somite-derivedneck muscles. The nonclonal nature of the labeling observed inthese muscles is also indicated by random left/right distributioncompared with labeling in the trapezius group (Fig. 2A). Wetherefore conclude that the trapezius and sternocleidomastoidmuscles derive from common progenitors that are distinct fromthose that give rise to first and second arch-derived head mus-cles, as well as to somite-derived muscles, and thus constitute anindependent muscle lineage.

Clonal Relation with Myocardium. The Mef2c-AHF-Cre genetictracing of the trapezius group suggests a potential common ori-gin with a subset of heart progenitors. These are different fromthe common cardiac and skeletal myogenic progenitor cellsgiving rise to head muscles that, as we show, are clonally distinctfrom neck muscles. Six of 30 embryos (20%) with labeling inthe trapezius and/or sternocleidomastoid muscles also showed

Pax3

-Cre

; R26

RM

esp1

-Cre

; R26

R

A B

C D

E14.5

sterno

s-trap

s-trap

E14.5 E14.5

sterno

E14.5

a-trap

E14.5

E F1stBA1stBA

trap

heart

E12.5

E10.5

G1stBA

heart

trapprog

Mef

2c-A

HF-

enha

ncer

-Cre

; R26

R

s-trap

1stBA

E10.5

Htrap prog

heart

MyoD

Fig. 1. Genetic tracing of the trapezius and sternocleidomastoid neckmuscles. (A and B) Transverse sections of an E14.5 Pax3Cre/+;Rosa26R-nlacZ/+

(Pax3-Cre;Rosa26R) embryo, stained with X-gal and eosin. The level of thesections is indicated at the bottom right. The sternocleidomastoid (sterno)and spino-trapezius (s-trap) muscles are mostly negative for β-galactosidase(β-gal) (arrowheads). (C and D) Transverse sections of an E14.5 Mesp1Cre/+;Rosa26R-nlacZ/+ (Mesp1-Cre;Rosa26R) embryo, stained with X-gal and eosin.The level of the sections is indicated at the bottom right. Arrowheads in-dicate muscles that are positive for β-gal, including the trapezius (a-trap ands-trap) and sternocleidomastoid (sterno) muscles (outlined with a discontin-uous line). (E–G) Whole-mount X-gal staining of Mef2c-AHF-enhancer-Cre;R26R embryos at E14.5 (E), E12.5 (F), and E10.5 (G), showing β-gal positivecells (arrowheads) in the first branchial arch-derived head muscles (1stBA),acromio-trapezius (a-trap), and spino-trapezius (s-trap) muscles (E, F), ortheir progenitors (trap prog) (G) and in the heart. (H) Whole-mount in situhybridization with a MyoD riboprobe on an E10.5 embryo.

Lescroart et al. PNAS | February 3, 2015 | vol. 112 | no. 5 | 1447

DEV

ELOPM

ENTA

LBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 19

, 202

0

labeling in the arterial pole of the heart, specifically in pulmo-nary trunk myocardial cells. Unexpectedly, labeling was alsoobserved in myocardial cells at the venous pole of the heart in 16of 30 embryos (53.3%), 5 of which also had β-gal positive cells atthe arterial pole (Fig. 2A). Venous pole labeling was seen inmyocardial cells of the atria and in the superior caval vein (orvena cava) and pulmonary veins, which in the mouse havea myocardial sleeve labeled by cardiac actin (24, 25). Statisticalanalysis shows that this colabeling is very unlikely to be due toindependent recombination events (example for the left atrium,P = 6 × 10−7) (Table 1). We thus conclude that the trapeziusgroup of muscles is clonally related to the myocardium of boththe arterial pole, and of the atria and veins at the venous pole ofthe heart. In addition, we found a left/right regionalization, suchthat labeling in the trapezius muscle group on the left side of theneck correlates with labeling in the left side of the heart (pul-monary trunk, left atrium, left superior caval vein, and pulmo-nary vein) (Fig. 2 C, D, D′, and D″), and labeling in the trapeziusgroup on the right side correlates with labeling in the right ve-nous pole (right atrium and right superior caval vein) (Fig. 2 C,E, and E″). This regionalization is statistically significant (Fig. 2Cand Table 1). This is in keeping with our previous results showingclonal relationships with venous pole myocardium of the rightatrium and right superior caval vein or of the left atrium, leftsuperior caval vein, and pulmonary vein and of the pulmonarytrunk at the arterial pole of the heart (25).

Tbx1 Dependence. In addition to genetic labeling with the Mef2c-AHF-Cre (Fig. 1 C–G), the trapezius and sternocleidomastoidmuscles are also genetically labeled using a Tbx1-Cre allele (26).

Furthermore, these muscles are absent in Tbx1 null embryos atfetal stages (10, 19). We confirm this result at E14.5 in a crosswith the Mlc3f-nlacZ-2E transgene, expressed in differentiatedskeletal muscle cells, (Fig. 4 A and B), showing the loss of thetrapezius muscles as well as first and second arch-derived head

sterno

spino-trap

acromio-trap

* p=3x10-3

* p=2x10-5

* p=1x10-3

* p=2x10-4

* p=4x10-3

* p=1x10-2

* p=2x10-4

* p=7x10-3

left trapeziusright trapezius

A

C

B

LA

LAPV

s-trap

a-trap

14K1804-ventral

14K1804-dorsal

14K2215-ventral

14K2215-dorsal

RA

RSCV

14K2215-right

D’

D”

E E’

E”

Left trapezius muscles Right trapezius musclesratio (observ

RA

LA

RSCV

LSCV

PV

AP

0 5 10 15 20 25

s-trap

14K1804-left

a-trap

Dratio (observed/expected)

AP

nb of -galpositive fibers

1331

2383

2010

2810

Sternoa-traps-trapMastFac

FLAPVA

SCVPV

1553

1996

2940

2469

2185

2128

2353

2436

3057

2573

1471

2036

2634

1342

1677

2215

2188

2703

2724

1445

1804

1631

1942

1317

1374

3198

R L L L L R R L R L R R L L+RL+RR L R L R L R L L R R L L R L R L L R R L R R L+RL+RL+R

L L L R R L R L L R R R L R L R L L R R L R R L+RL+R RR

R

R LL L L L L L

L L L+R R R L L+R L+R LL R L L+R R L L R L+R L L R L L+R

L L L+R R L R L+R L L+RL L L L L L L

10 10 10 14 12 19 20 26 27 30 30 31 40 + + + + + + + + + + + + + + + + +

Tongue

Fig. 2. Trapezius neck muscles share common progenitors with myocardium. (A) All αc-actinnlaacZ/+ embryos at E14.5 with labeling of >10 β-galactosidase

positive fibers (β-gal+) in the trapezius or sternocleidomastoid muscles of the neck are shown, with labeling indicated by a box. L, left side; R, right side; nb,number of fibers with β-gal+ nuclei in the trapezius muscles; +, >50 positive fibers; Sterno, sternocleidomastoid muscle; a-trap, acromio-trapezius muscle;s-trap; spino-trapezius muscle; Mast, masticatory muscles; Fac, facial expression muscles; Tongue, tongue muscles; FL, forelimb muscles; AP, arterial pole; V,ventricles; A, atria; SCV, superior caval veins; PV, pulmonary vein. (B) Schematic representation of an E14.5 embryo showing superficial labeling of the twotrapezius (trap) and sternocleidomastoid (sterno) muscles. (C) Significant clonal relationships estimated on the basis of the ratio between the observed andexpected numbers of double labeling events in trapezius muscles on the left or right side and myocardium in the collection of αc-actin

nlaacZ/+ embryos. Theblack vertical line indicates a ratio of observed labeling compared with labeling expected from more than one recombination event of 1. *P values showinga statistically significant relationship are indicated. (D–D″) Example of an αc-actin

nlaacZ/+ embryo (reference number 1804) with β-gal+ cells (arrowheads) in theleft a-trap and s-trap muscles (D) and in the arterial pole (AP), left atrium (LA), and pulmonary vein (PV) (D′ ventral, D″ dorsal views of the heart of the sameembryo.). (E–E″) Example of an αc-actin

nlaacZ/+ embryo (reference number 2215) with β-gal+ cells (arrowheads) in the right a-trap and s-trap muscles (E) and inthe right atrium (RA) and right superior caval vein (RSCV) (E′ ventral, E″ dorsal views of the heart of the same embryo).

c-actinnlaacZ/+ embryos

A A’

14K2929

a-trap sterno

14K2260sternoa-trap

D

E

sterno

a-trap

14K2703

a-trapsterno

14K1317

B

C

Pax3-Cre; R26R

trap muscle group trap muscle group+ splenius muscle

Fig. 3. Genetic tracing of laryngeal muscles and representative examples ofαc-actin

nlaacZ/+ embryos. (A and A′) Transverse sections of an E14.5 Pax3Cre/+;Rosa26R-nlacZ/+ (Pax3-Cre;R26R) embryo, stained with X-gal and eosin. Thelevel of the section is indicated at the bottom left. β-galactosidase (β-gal)negative muscle masses in the region of the larynx are surrounded by adotted line. (B–E ) Transverse sections of αc-actin

nlaacZ/+ embryos (numbersat the bottom left refer to the specific embryo) in which β-gal positivemuscle masses are indicated (blue arrowheads). a-trap, acromio-trapezius;sterno; sternocleidomastoid; black arrowheads indicate β-gal positive cellsclose to the larynx. Of 6 embryos with nonsomitic neck muscle labeling thatwere sectioned, all had labeling both in the trapezius group and in laryn-geal muscles.

1448 | www.pnas.org/cgi/doi/10.1073/pnas.1424538112 Lescroart et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 19

, 202

0

muscles, also marked byMef2c-AHF-Cre genetic tracing (Fig. 4 Eand F). In addition, we observe that other neck muscles such asthose of the larynx (Fig. 4 C and D), that are clonally relatedto the trapezius group, are also absent in Tbx1−/− embryos. Wefurther show, using Mef2c-AHF-Cre genetic tracing and expres-sion of the myogenic determination gene, MyoD, that trapeziusmuscle progenitor cells fail to emerge from the posterior pharyn-geal region from E10.5, in the absence of Tbx1 (Fig. 4 G–R).

The Splenius Muscle, of Somitic and Nonsomitic Origin. The spleniusmuscle, lying in an epaxial location caudal to the ear, provides anexample of a muscle of mixed somitic and nonsomitic origin. Of

the 22 embryos in the E14.5 collection that had labeling in thismuscle, 9 also had labeling in somitic neck muscles (40.9%), and5 showed labeling in the trapezius muscle group (22.7%) (Fig. 5A–E). In both cases this is statistically significant, indicatingclonality (Table 1 and Table S1). Genetic tracing with a Pax3Cre

allele (Fig. 5F) confirms that some progenitor cells for thismuscle had expressed Pax3 and are therefore somite-derived.We conclude that this muscle, which is located in a transitionzone, is of mixed origin, formed by progenitors derived from thesomites that are clonally related to other neck muscles and alsoby cardiopharyngeal progenitor cells that contribute to the tra-pezius group of skeletal muscles and to myocardium.

DiscussionIn Fig. 6, we integrate our results on the trapezius group ofnonsomitic neck muscles with a lineage tree for the secondmyocardial lineage contribution to the heart (27). We identifythree distinct sublineages that contribute to first or secondbranchial arch-derived head muscles or to neck muscles thatderive from the caudal branchial arches. These sublineages alsocontribute to myocardium on the anterior/posterior axis, withright ventricular, arterial pole, and venous pole contributionsthat follow the progressive posterior positioning of the heart inthe pharyngeal region as the branchial arches form. Commonprogenitors for both myocardial cells and skeletal muscles areprobably a small proportion of the overall cardiac progenitorpopulation marked by Mesp1 expression because they representonly 10% (out of 161 embryos studied) of the clones in Mesp1-inducible clonal analysis (28) and were not reported in the 38embryos studied by clonal analysis using the MADM systemdriven by Mesp1-Cre (29). The common progenitors that con-tribute to each of these three groups of skeletal and cardiacmuscle derivatives probably segregate early. Inducible genetictracing with clonal resolution of the kind performed to determinethe timing of first and second myocardial lineage segregation (28)will be required to investigate when these sublineages arise. It isstriking that neural crest cells that migrate through the pharyn-geal region also do so in three distinct evolutionarily conservedstreams, through the first, the second, and the more caudal arches(30), thus paralleling the patterning of cardiopharyngeal meso-derm revealed by our clonal analysis. The development of cranialneural crest and mesoderm are closely connected, crest being

Table 1. Statistical analysis of the probability of independentrecombination events

Muscle/myocardium Trapezius m Splenius m Somitic neck m

Skeletal musclesTrapezius m — 3 × 10−5* 0.51Splenius m 3 × 10−5* — 8 × 10−11*Somitic neck m 0.51 8 × 10−11* —

Masticatory m 0.24 1 0.27Facial expression m 0.27 0.18 0.30Tongue m 0.09 0.15 2 × 10−2*Forelimb m 0.13 0.04* 6 × 10−8*

MyocardiumAP 4 × 10−5* 1 0.51RV 0.18 1 0.19LV 0.64 0.22 0.18RA 3 × 10−3* 1 0.27LA 6 × 10−7* 1 1RSCV 9 × 10−6* 1 1LSCV 2 × 10−7* 0.13 0.38PV 1 × 10−9* 0.07 1

We have used the nonparametric Fisher’s exact test to assess whether dou-ble labeling results from two independent events in the 30 αc-actin

nlaacZ/+

embryos with labeling in the trapezius group of muscles at E14.5. Each cate-gory of neck muscles was tested for independence with other skeletal musclesor myocardium. *P < 5 × 10−2, indicating that the two regions are clonallyrelated. m, muscle; AP, arterial pole; RV, right ventricle; LV, left ventricle; RA,right atrium; LA, left atrium; RSCV, right superior caval vein; LSCV, left superiorcaval vein; PV, pulmonary vein.

Mlc

3f-2

E

A B

E14.5 E14.5

E14.5 E14.5

Tbx1-/-WT

C D

Tbx1-/-WT

E10

.5E

11.5

E12

.5

Mef2c-Cre; R26R Mef2c-Cre; R26RMyoD ISH MyoD ISH

G H I J

K L M N

O P Q RE F

E14.5 E14.5Mef

2c-C

re; R

26R

**

Fig. 4. The trapezius muscle group and la-ryngeal muscles are affected in Tbx1−/− em-bryos. (A–D) The comparison between Mlc3f-2E transgene expression, that marks all skel-etal muscle, in WT (A and C) or Tbx1−/−

mutants (B and D) stained with X-gal showsthe loss of the trapezius and sternocleido-mastoid muscles (blue arrowheads in A) andthe laryngeal muscles, both intrinsic and ex-trinsic (black arrowheads in C), in the mutant.The tongue muscles (black arrow in C and D)are still present in the absence of Tbx1. (E andF) Whole-mount X-gal staining ofMef2c-AHF-enhancer-Cre (Mef2c-Cre);Rosa26R-nlacZ/+ (R26R)embryos on a WT (E ) or Tbx1−/− mutant (F )genetic background at E14.5, showing thatthe trapezius muscle (blue arrowheads in E )as well as the branchial arch-derived headmuscles (white arrowhead in E ) are severelyaffected in the mutant. (G–R) Whole-mountin situ hybridization with a MyoD riboprobe,that marks skeletal muscle, on WT (G, K, and O) and Tbx1−/− mutant (I, M, and Q) embryos at E10.5 (G and I), E11.5 (K and M) and E12.5 (O and Q). X-galstaining of Mef2c-Cre;R26R embryos in WT (H, L, and P) or in Tbx1−/− mutant (J, N, and R) embryos at E10.5 (H and J), E11.5 (L and N) or E12.5 (P and R). Theforming trapezius (blue arrowheads) and head muscles (white arrowheads) fail to develop in the absence of Tbx1. The asterisk in F and R indicates asymmetricresidual first arch-derived muscles that form stochastically in the absence of Tbx1.

Lescroart et al. PNAS | February 3, 2015 | vol. 112 | no. 5 | 1449

DEV

ELOPM

ENTA

LBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 19

, 202

0

required for patterning of branchiomeric muscles as well as forsecond heart field progenitor cell addition to the arterial pole ofthe heart (30–32).We had previously analyzed clonal relationships between myo-

cardium at the venous pole of the heart, which derives from theposterior SHF (33), and were surprised to find clonality betweenthe left venous pole and pulmonary trunk myocardium at the ar-terial pole (25). This sublineage is distinct from the sublineage thatcontributes to pulmonary trunk myocardium and to left headmuscles derived from the second branchial arch, and, as we showhere, also gives rise to left nonsomitic neck muscles. Pulmonarytrunk myocardium thus has two different origins. Dye labeling ofcells in different regions of the SHF, followed by mouse embryoculture, showed that some cardiac progenitor cells move froma caudal to a more rostral location to contribute to outflow tractmyocardium (34). We therefore propose that some of the pro-genitors that are clonally related to neck muscles move rostrallybefore they enter the arterial pole of the heart.Tbx1 is required to regulate proliferation and delay differen-

tiation in the SHF and is essential for outflow tract development(9). However, Tbx1−/− embryos have recently been shown to haveinflow as well as outflow tract defects, with abnormal developmentof the dorsal mesenchymal protrusion and impaired addition ofcells to the venous pole of the heart, resulting in atrioventricularseptal defects (35). Tbx1 thus regulates the behavior of cardiacprogenitors that contribute to both poles of the heart tube as wellas being required for the development of nonsomitic neckmuscles,regulating all of the myogenic derivatives of the common lineagedefined here.The role of Tbx1 in the deployment of cardiopharyngeal me-

soderm is already presaged by its expression in ascidians (16).The sublineage that we characterize here probably originatedlater during radiation of the vertebrates, because the cucullaris

muscle first evolved in gnathostomes (bony fish) (36). The pres-ence of common lineages giving rise to specific compartments ofthe heart and specific head and neck muscle groups suggests thatmodulation of the developmental potential of cardiopharyngealmesoderm has played an important role in the coevolution of theneck and heart during vertebrate evolution. Indeed, the existenceof multiple common myogenic progenitor cell populations in themammalian pharyngeal region may reflect reiteration of a devel-opmental motif regulating the segregation of skeletal and cardiacmyogenic fate from common progenitor cells in cardiopharyngealmesoderm. The transition zone, constituted by the neck, betweenhead and trunk, shows no clear boundary between somitic andcraniopharyngeal muscle derivatives; and indeed our analysis ofthe splenius muscle, which is located relatively anteriorly, showsthat there is also no obligatory segregation between muscles, sothat a muscle like the splenius, with cervical innervation, can be ofmixed origin. In the future it will be important to define when thethree sublineages that we describe segregate and how the ap-pearance of different muscle and cardiac derivatives is regulatedby cardiopharyngeal patterning in the mouse. Our findings alsohave relevance for human pathology. TBX1 is the major geneinvolved in del22q11.2 or DiGeorge syndrome, characterized bya range of cardiovascular and craniofacial anomalies. Under-standing of the etiology of these defects will depend on identifyingthe mechanisms by which Tbx1 regulates cardiac and skeletalmyogenic fates within the three clonally related populations ofcommon cardiac and skeletal muscle progenitor cells in car-diopharyngeal mesoderm.

spleniusA

C

B

splenius

14K1560-left

splenius muscle splenius + Trap muscles

splenius + somitic neck muscles Pax3-Cre; R26R

D E F

14K2929-right 14K2421-left-gal

MF20

splenius

a-trap

splenius

s-trap splenius

som

L L R L L R L R R L R R R L R L R L L LL/RL R RL/R L/RL/RL L L LL/R

R R L10101015203030 + + 15 + + + 2020 + + + + + +

3037

3058

1793

1560

2026

2286

1384

2282

2483

2165

2760

2929

2260

2484

1340

1830

2717

2126

2421

3288

3004

spleniusTrap

other neckMastFacFL

nb of -gal+ fibres splenius

onlysplenius+ Trap

splenius + other neck muscles

Fig. 5. The splenius muscle has a dual clonal origin. (A) All embryos withlabeling of >10 β-galactosidase positive (β-gal+) fibers in the splenius muscleare shown, with labeling indicated by a box. L, left side; R, right side; nb,number of fibers with β-gal+ nuclei; +, >50 positive fibers; Trap, the trape-zius group of muscles; Other neck, somite-derived neck muscles; Mast,masticatory muscles; Fac, facial expression muscles; FL, forelimb muscles. (B)Schematic representation of an E14.5 embryo showing the labeled spleniusmuscle, situated behind the ear. (C–E) Examples of αc-actin

nlaacZ/+ embryoswith β-gal+ cells (arrowheads) in the left splenius muscle (C), in the rightsplenius and spino-trapezius (s-trap) muscles (D), and in the left splenius andsomitic neck muscles (som) (E). Figures at bottom right indicate the refer-ence numbers of individual embryos. (F) Transverse section of an E14.5Pax3Cre/+;Rosa26R-nlacZ/+ (Pax3-Cre; R26R) embryo stained for X-gal (blue).Skeletal muscles are marked by immunohistochemistry with a myosin anti-body (MF20) (brown). The splenius muscle has β-gal positive cells, whereasthe trapezius (a-trap) is negative. The level of the section is indicated at thebottom left.

Lescroart et al. 2010Lescroart et al. 2010

left neck nonsomitic muscles

pulmonary trunkleft atriumleft superior caval veinpulmonary vein

right neck nonsomitic muscles

right atriumright superior caval vein

left

right

EOMs

right ventricle

1st BA-derived muscles

aorta

right 2nd BA-derived muscles

pulmonary trunk

left 2nd BA-derived muscles

splenius muscles

splenius muscles

somitic-derived neck muscles

left

right

Fig. 6. Schema showing the lineage relationship between nonsomitic neckmuscles of the trapezius group and myocardium. The first two lineages (inblue) contribute to the nonsomitic head muscles, derived from the first (1st)and second (2nd) branchial arches (BA), and to arterial pole myocardium(14). A third lineage (in mauve) contributes to the trapezius muscle group aswell as to venous pole myocardium, with an additional contribution to thepulmonary trunk, and a contribution to the splenius muscle (orange). Anindependent lineage (green) contributes to the somite-derived neck muscles,including part of the splenius muscle.

1450 | www.pnas.org/cgi/doi/10.1073/pnas.1424538112 Lescroart et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 19

, 202

0

Materials and MethodsMice. The αc-actin

nlaacZ1.1/+ (23), Pax3Cre/+ (37), Mesp1Cre/+ (38), conditionalRosa26R-nlacZ/+ (14), Gtrosa26tm1Sor (39), Tbx1+/− (40), and Mlc3f-nlacZ-2E (41)mouse lines and the Mef2c-AHF-enhancer-Cre transgenic line (21) have beendescribed previously. Animal care was in accordance with national andinstitutional guidelines.

Xgal Staining, Immunochemistry, in Situ Hybridization, and Histology. E14.5αc-actin

nlaacZ1.1/+ embryos were fixed in 4% paraformaldehyde, and whole-mount X-Gal staining was performed as previously described (23, 42). Sec-tions, obtained using a cryostat, were X-gal stained, and immunochemistrywas performed with MyoD (Dako) or MF20 (DSHB) antibodies with therabbit vectastain ABC kit (Vector Laboratories). Peroxidase activity was re-vealed with SIGMAFAST DAB tablets. In situ hybridization using an antisenseMyoD riboprobe was carried out as previously described (10).

Retrospective Clonal Analysis and Statistical Analysis. A total of 2,018 embryosat E14.5 were collected (14). Most of the E14.5 αc-actin

nlaacZ1.1/+ embryospresent multiple clusters of β-gal positive cells/fibers in skeletal muscles.Because small clusters are derived from a late recombination event and arenot relevant for investigating clonal relationships between skeletal musclesand heart myocardium, we scored in our analysis only embryos with labelingof more than 10 β-gal positive cells/fibers. A total of 83 embryos (4.11%)

were found with labeling in neck skeletal muscles. Because recombination israndom, there is a low statistical probability that such an event occurs in thesame location a second time, and therefore a cluster of labeled cells in theneck probably contains clonally related cells (23, 43).

To establish clonal relationships between two distinct regions, we esti-mated the expected frequency of double recombination events in two dif-ferent regions, which, according to the law of independent probabilities, isequal to the product of the frequency of labeling in each region (Fig. S1). Wethen assessed with the Fisher’s exact test whether this number differed fromthe observed frequency of colabeling in the two regions. The null hypothesisis that the labeling in both regions results from two independent events.When the P value is lower than 0.05, the null hypothesis may be confidentlyrejected, leading to the conclusion that the two regions are clonally related.

ACKNOWLEDGMENTS. We thank C. Bodin for technical help. The work inM.B.’s laboratory was supported by the Pasteur Institute and the CNRS, withgrants from the European Union Integrated Projects “Heart Repair” [LH SM-CT2005-018630 (also to R.G.K.)] and “CardioCell” (LT2009-223372). M.B. andR.G.K. also acknowledge the support of the Association Française contreles Myopathies (AFM). R.G.K. is an INSERM research scientist and acknowl-edges the support of the Fondation pour la Recherche Médicale (Equipe FRMDEQ20110421300). F.L. has benefitted from a doctoral fellowship from theIle de France region and was supported by the AFM.

1. Buckingham M, Relaix F (2007) The role of Pax genes in the development of tissuesand organs: Pax3 and Pax7 regulate muscle progenitor cell functions. Annu Rev CellDev Biol 23:645–673.

2. Buckingham M, Rigby PW (2014) Gene regulatory networks and transcriptionalmechanisms that control myogenesis. Dev Cell 28(3):225–238.

3. Tajbakhsh S, Rocancourt D, Cossu G, Buckingham M (2007) Redefining the genetichierarchies controlling skeletal myogenesis: Pax-3 and Myf-5 act upstream of MyoD.Cell 89(1):127–138.

4. Harel I, et al. (2009) Distinct origins and genetic programs of head muscle satellitecells. Dev Cell 16(6):822–832.

5. Sambasivan R, et al. (2009) Distinct regulatory cascades govern extraocular andpharyngeal arch muscle progenitor cell fates. Dev Cell 16(6):810–821.

6. Buckingham M, Meilhac S, Zaffran S (2005) Building the mammalian heart from twosources of myocardial cells. Nat Rev Genet 6(11):826–835.

7. Kelly RG (2012) The second heart field. Curr Top Dev Biol 100:33–65.8. Kong P, et al. (2014) Tbx1 is required autonomously for cell survival and fate in the

pharyngeal core mesoderm to form the muscles of mastication. Hum Mol Genet23(16):4215–4231.

9. Papangeli I, Scambler P (2013) The 22q11 deletion: DiGeorge and velocardiofacialsyndromes and the role of TBX1. Wiley Interdiscip Rev Dev Biol 2(3):393–403.

10. Kelly RG, Jerome-Majewska LA, Papaioannou VE (2004) The del22q11.2 candidategene Tbx1 regulates branchiomeric myogenesis. Hum Mol Genet 13(22):2829–2840.

11. Noden DM, Francis-West P (2006) The differentiation and morphogenesis of cranio-facial muscles. Dev Dyn 235(5):1194–1218.

12. Tirosh-Finkel L, Elhanany H, Rinon A, Tzahor E (2006) Mesoderm progenitor cells ofcommon origin contribute to the head musculature and the cardiac outflow tract.Development 133(10):1943–1953.

13. Nathan E, et al. (2008) The contribution of Islet1-expressing splanchnic mesodermcells to distinct branchiomeric muscles reveals significant heterogeneity in headmuscle development. Development 135(4):647–657.

14. Lescroart F, et al. (2010) Clonal analysis reveals common lineage relationships be-tween head muscles and second heart field derivatives in the mouse embryo. De-velopment 137(19):3269–3279.

15. Stolfi A, et al. (2010) Early chordate origins of the vertebrate second heart field.Science 329(5991):565–568.

16. WangW, Razy-Krajka F, Siu E, Ketcham A, Christiaen L (2013) NK4 antagonizes Tbx1/10to promote cardiac versus pharyngeal muscle fate in the ascidian second heart field.PLoS Biol 11(12):e1001725.

17. Razy-Krajka F, et al. (2014) Collier/OLF/EBF-dependent transcriptional dynamics con-trol pharyngeal muscle specification from primed cardiopharyngeal progenitors. DevCell 29(3):263–276.

18. Edgeworth F (1935) The Cranial Muscles of Vertebrates (Cambridge Univ Press,Cambridge, UK).

19. Theis S, et al. (2010) The occipital lateral plate mesoderm is a novel source for ver-tebrate neck musculature. Development 137(17):2961–2971.

20. Saga Y, Kitajima S, Miyagawa-Tomita S (2000) Mesp1 expression is the earliest sign ofcardiovascular development. Trends Cardiovasc Med 10(8):345–352.

21. Verzi MP, McCulley DJ, De Val S, Dodou E, Black BL (2005) The right ventricle, outflowtract, and ventricular septum comprise a restricted expression domain within thesecondary/anterior heart field. Dev Biol 287(1):134–145.

22. Bonnerot C, Nicolas JF (1993) Clonal analysis in the intact mouse embryo by intragenichomologous recombination. C R Acad Sci III 316(10):1207–1217.

23. Meilhac SM, et al. (2003) A retrospective clonal analysis of the myocardium reveals

two phases of clonal growth in the developing mouse heart. Development 130(16):3877–3889.

24. Mueller-Hoecker J, et al. (2008) Of rodents and humans: A light microscopic and ul-trastructural study on cardiomyocytes in pulmonary veins. Int J Med Sci 5(3):152–158.

25. Lescroart F, Mohun T, Meilhac SM, Bennett M, Buckingham M (2012) Lineage tree for

the venous pole of the heart: Clonal analysis clarifies controversial genealogy basedon genetic tracing. Circ Res 111(10):1313–1322.

26. Huynh T, Chen L, Terrell P, Baldini A (2007) A fate map of Tbx1 expressing cells revealsheterogeneity in the second cardiac field. Genesis 45(7):470–475.

27. Meilhac SM, Lescroart F, Blanpain C, Buckingham ME (2014) Cardiac cell lineages that

form the heart. Cold Spring Harb Perspect Med 4(9):a013888.28. Lescroart F, et al. (2014) Early lineage restriction in temporally distinct populations of

Mesp1 progenitors during mammalian heart development. Nat Cell Biol 16(9):829–840.29. Devine WP, Wythe JD, George M, Koshiba-Takeuchi K, Bruneau BG (2014) Early

patterning and specification of cardiac progenitors in gastrulating mesoderm. eLife,

10.7554/eLife.03848.30. Keyte AL, Alonzo-Johnsen M, Hutson MR (2014) Evolutionary and developmental

origins of the cardiac neural crest: Building a divided outflow tract. Birth Defects ResC Embryo Today 102(3):309–323.

31. Rinon A, et al. (2007) Cranial neural crest cells regulate head muscle patterning and

differentiation during vertebrate embryogenesis. Development 134(17):3065–3075.32. Noden DM (1983) The role of the neural crest in patterning of avian cranial skeletal,

connective, and muscle tissues. Dev Biol 96(1):144–165.33. Galli D, et al. (2008) Atrial myocardium derives from the posterior region of the second

heart field, which acquires left-right identity as Pitx2c is expressed. Development 135(6):

1157–1167.34. Domínguez JN, Meilhac SM, Bland YS, Buckingham ME, Brown NA (2012) Asymmetric

fate of the posterior part of the second heart field results in unexpected left/rightcontributions to both poles of the heart. Circ Res 111(10):1323–1335.

35. Rana MS, et al. (2014) Tbx1 coordinates addition of posterior second heart field

progenitor cells to the arterial and venous poles of the heart. Circ Res 115(9):790–799.36. Kuratani S, Schilling T (2008) Head segmentation in vertebrates. Integr Comp Biol

48(5):604–610.37. Engleka KA, et al. (2005) Insertion of Cre into the Pax3 locus creates a new allele of

Splotch and identifies unexpected Pax3 derivatives. Dev Biol 280(2):396–406.38. Saga Y, et al. (1999) MesP1 is expressed in the heart precursor cells and required for

the formation of a single heart tube. Development 126(15):3437–3447.39. Soriano P (1999) Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat

Genet 21(1):70–71.40. Jerome LA, Papaioannou VE (2001) DiGeorge syndrome phenotype in mice mutant

for the T-box gene, Tbx1. Nat Genet 27(3):286–291.41. Kelly R, Alonso S, Tajbakhsh S, Cossu G, Buckingham M (1995) Myosin light chain 3F

regulatory sequences confer regionalized cardiac and skeletal muscle expression intransgenic mice. J Cell Biol 129(2):383–396.

42. Bajolle F, et al. (2006) Rotation of the myocardial wall of the outflow tract is impli-cated in the normal positioning of the great arteries. Circ Res 98(3):421–428.

43. Meilhac SM, Esner M, Kelly RG, Nicolas JF, BuckinghamME (2004) The clonal origin ofmyocardial cells in different regions of the embryonic mouse heart. Dev Cell 6(5):685–698.

Lescroart et al. PNAS | February 3, 2015 | vol. 112 | no. 5 | 1451

DEV

ELOPM

ENTA

LBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 19

, 202

0