Myosin Light Chain 2a and 2vIdentifies the Embryonic Outflow

Tract Myocardium in the DevelopingRodent Heart

DIEGO FRANCO, MARY M.W. MARKMAN, GERRY T.M. WAGENAAR,JING YA, WOUTER H. LAMERS, AND ANTOON F.M. MOORMAN*

Department of Anatomy and Embryology, Academic Medical Center,University of Amsterdam, Amsterdam, The Netherlands

ABSTRACTThe embryonic heart consists of five segments comprising the fast-

conducting atrial and ventricular segments flanked by slow-conductingsegments, i.e. inflow tract, atrioventricular canal and outflow tract. Al-though the incorporation of the flanking segments into the definitive atrialand ventricular chambers with development is generally accepted now, thecontribution of the outflow tract myocardium to the definitive ventriclesremained controversial mainly due to the lack of appropriate markers. Forthat reason we performed a detailed study of the pattern of expression ofmyosin light chain (MLC) 2a and 2v by in situ hybridization and immunohis-tochemistry during rat and mouse heart development. Expression of MLC2amRNA displays a postero-anterior gradient in the tubular heart. In theembryonic heart it is down-regulated in the ventricular compartment andremains high in the outflow tract, atrioventricular canal, atria and inflowtract myocardium. MLC2v is strongly expressed in the ventricular myocar-dium and distinctly lower in the outflow tract and atrioventricular canal.The co-expression of MLC2a and MLC2v in the outflow tract and atrioven-tricular canal, together with the single expression in the atrial (MLC2a) andventricular (MLC2v) myocardium, permits the delineation of their bound-aries. With development, myocardial cells are observed in the lower endocar-dial ridges that share MLC2a and MLC2v expression with the myocardialcells of the outflow tract. In neonates, MLC2a continues to be expressedaround both right and left semilunar valves, the outlet septum and thenon-trabeculated right ventricular outlet. These findings demonstrate thecontribution of the outflow tract to the definitive ventricles and demonstratethat the outlet septum is derived from outflow tract myocardium. Anat Rec254:135–146, 1999. r 1999 Wiley-Liss, Inc.

Key words: gene expression; heart; morphogenesis; myosin; in situ hy-bridization; immunohistochemistry

During embryogenesis, the heart develops from a tubu-lar structure into a four-chambered organ. In the embry-onic heart tube, five morphological and functional seg-ments can be distinguished, i.e. inflow tract, atria,atrioventricular canal, ventricles and outflow tract (Moor-man and Lamers, 1994; Franco et al., 1998). With furtherdevelopment these embryonic segments become incorpo-rated into the mature atrial and ventricular chambers.Although the septation of the embryonic cardiac outflowtract into aortic and pulmonary trunks has been studied

Grant sponsor: NWO; Grant number: 902–16–219; Grant spon-sor: NHS; Grant number: 997206.

Jing Ya is presently at Department of Histology and Embryol-ogy, Shanxi Medical University, Taiyuan, Shanxi 030001 China.

*Correspondence to: Antoon F.M. Moorman, Department ofAnatomy and Embryology, Academic Medical Center, Meiberg-dreef, 15 1105 AZ Amsterdam, The Netherlands.

Received 4 August 1998; Accepted 10 September 1998

THE ANATOMICAL RECORD 254:135–146 (1999)

r 1999 WILEY-LISS, INC.

extensively in birds (Thompson and Fitzharris, 1979a,1979b; Laane, 1979; Thompson et al., 1987) and mammals(Icardo, 1990; Pexieder, 1995; Ya et al., 1998) includingman (Kramer, 1942; van Mierop et al., 1963; Orts Llorca etal., 1982; Bartelings and Gittenberger-de Groot, 1989), theprecise topographical contribution to the definitive adultheart remains unclear due to the lack of appropriatemarkers (Noden et al., 1995).

As a result of the coordinated formation of the aorticopul-monary septum and the proximal outflow tract septum,the ventricles obtain independent arterial connections.The first developmental process observed during septationof the outflow tract is the establishment of the aorticopul-monary septum by migrating neural crest cells (Kirby,1993). Several authors have suggested that fusion of theendocardial ridges of the outflow tract is followed bymuscularization (Goor et al., 1972; McBride et al., 1981;Asami, 1992; De Jong et al., 1992a; Ya et al., 1998). Thisnotion has been contested by others (Bartelings and Gitten-berger-de Groot, 1989; Sumida et al., 1989). In vivo tracingexperiments demonstrated that the arterial myocardialpole of the chicken heart contributes to the mature outletseptum as well as to the right ventricular free wall (De laCruz et al., 1977, 1989). More recently, evidence fromtransgene expression supports the notion that the outflowtract is a distinct transcriptional entity (Franco et al.,1997).

In this study we report the pattern of expression of twoendogenous genes that display a unique expression in theoutflow tract myocardium distinct from the ventricularmyocardium. Moreover, our findings demonstrate thatmyocardial cells derived from the embryonic outflow tractcolonize the proximal outflow tract ridges, giving rise tothe muscular outlet septum. Most of the embryonic outflowtract myocardium contributes to the non-trabeculatedright ventricular myocardial outlet and, to a lesser extent,to the left ventricular outlet. In line with previous trans-genic studies (Franco et al., 1997), the differential expres-sion of MLCs supports the notion that the embryonicoutflow tract constitutes a distinct transcriptional myocar-dial segment of the heart.

MATERIALS AND METHODSEmbryos

The generation of 3F-nlacZ-9 transgenic mice and theirtransgene expression pattern during cardiac and skeletalmuscle development have been previously described (Kellyet al., 1995; Franco et al., 1997). Hemizygous 3F-nlacZ-9and wild-type control embryos (CD1) ranging from embry-onic day (E) 9.5 to E16.5 [(strain CS7BL6/J 3 SJL) F1]were analyzed. Wistar rat embryos were obtained fromtimed-pregnant rats (E10–E20). The day of the vaginalplug was scored as 0.5 day of gestation. Neonatal (3, 8, 10and 23 days) rat hearts were dissected and processed asfollows. Tissues were fixed for 4 hr in 4% freshly preparedformaldehyde in phosphate-buffered saline (PBS) at roomtemperature for in situ hybridization and overnight inmethanol:acetone:water (2:2:1) at room temperature forimmunohistochemistry, dehydrated in a graded series ofethanol and embedded in paraplast. Serial sections werecut at 7 µm thickness and mounted onto RNAse-free3-aminopropyltriethoxysilane (AAS)-coated slides (in situhybridization) or onto poly-L-lysine-coated slides (immuno-histochemistry). Embryos for whole-mount in situ hybrid-

ization were fixed in 4% freshly prepared formaldehyde,rinsed in PBT (PBS with 0.1% Tween-20), dehydrated inmethanol:PBT series and stored at 220°C until used.

In Situ Hybridization on Tissue SectionsComplementary RNA probes of rat a and b myosin

heavy chain (MHC) mRNAs (Schiaffino et al., 1989; Bohe-ler et al., 1992), mouse regulatory atrial myosin light chain(MLC2a) mRNA (Kubalak et al., 1994), mouse regulatoryventricular myosin light chain (MLC2v) mRNA (O’Brien etal., 1993) and rat cardiac troponin I (cTnI) mRNA (Ausoniet al., 1991) were radiolabelled with 35S-UTP by in vitrotranscription according to standard protocols (Melton etal., 1984). To obtain representative samples of the degreeof expression for each gene transcript in the differentcardiac compartments, serial sections were hybridizedwith a range of probe concentrations varying between10,000 cpm/µl and 40,000 cpm/µl of hybridization solution.The hybridization conditions were as detailed elsewhere(Moorman et al., 1993, 1995). Briefly, the sections weredeparaffinized, rinsed in absolute ethanol and dried in anair stream. Pretreatment of the sections was as follows: 20min 0.2 M HCl, 5 min bi-distilled water, 20 min 2 3 SSC(70°C), 5 min bi-distilled water, 2–20 min digestion in 0.1%pepsin dissolved in 0.01 M HCl, 30 sec in 0.2% glycine/PBS, twice 30 sec in PBS, 20 min of postfixation in 4%freshly prepared formaldehyde, 5 min in bi-distilled water,5 min in 10 mM EDTA (pH 8), 5 min in 10 mM DTT andfinally drying in an air stream. The prehybridizationmixture contained: 50% formamide, 10% dextran sul-phate, 2 3 SSC, 2 3 Denhardt’s solution, 0.1% TritonX-100, 10 mM DTT and 200 ng/µl heat-denatured herringsperm DNA. The sections were hybridized overnight at52°C and washed as follows: a rinse in 1 3 SSC, 30 min at52°C in 50% formamide dissolved in 1 3 SSC, 10 min in1 3 SSC, 30 min in RNAse A (10 µg/ml), 10 min in 1 3 SSC,10 min in 0.1 3 SSC and dehydration in 50%, 70% and 90%ethanol containing 0.3 M ammonium acetate. The sectionswere dried and immersed in nuclear autoradiographicemulsion G5 (Ilford). The exposure time ranged from 4 to 7days with development time of 4 min. Images were takenusing a Photometrics camera coupled to a Zeiss Axiophotmicroscope and computerized files were recorded. Panelswere composed using Power Point software.

Whole-Mount In Situ HybridizationComplementary RNA probes against rat bMHC (Bohe-

ler et al., 1992), mouse MLC2a (Kubalak et al., 1994) andMLC2v (O’Brien et al., 1993) mRNAs were labeled withdigoxigenin-UTP by in vitro transcription according tostandard protocols (Hogan et al., 1994; Henrique et al.,1997). Hybridization conditions were as described by Hen-rique et al. (1997) with slight modifications (Zammit andBuckingham, 1998). Briefly, embryos were hydrated inmethanol:PBT (PBS with 0.1% Tween-20) graded series,washed gently in PBT, digested with 10 µg/ml proteinase Kat 37°C for 20–60 min, washed again gently in PBT andtransferred into a 1:1 PBT:hybridization mixture [HB;50% formamide, 1.3 3 SSC (pH 5), 5 mM EDTA (pH 8), 25µg/ml yeast tRNA, 0.2% Tween-20, 0.5% CHAPS, 0.01%heparin, 50 µg/ml heat-denaturated salmon sperm DNA]for 1 hr at 70°C. Embryos were hybridized overnight at68°C in hybridization mixture with a riboprobe concentra-tion of 100 ng/ml, followed by washes in 1:1 HB:TBST (1.5

136 FRANCO ET AL.

M NaCl, 25 mM KCl, 250 mM Tris-base salt, 0.1% Tween-20; pH 7.5) at 68°C and finally transferred to TBST atroom temperature. Prior to immunohistochemical detec-tion of the digoxigenin-labeled riboprobes with mouseanti-digoxigenin-alkaline phosphatase Fab fragments(Boehringer-Mannheim 1093–274; 1:2000; overnight), theembryos were treated with TBST containing 2% of block-ing reagent, 5 mM of levamisole (Boehringer-Mannheim1096–176) and 10% heat-inactivated rat serum for approxi-mately 1 hr. Embryos were gently washed after hybridiza-tion in MABT (100 mM maleic acid, 150 mM NaCl, 0.1%Tween-20, 5 mM levamisole, adding Tris-base salt to bufferat pH 7) for a few hours and finally rinsed in NTMT (100mM NaCl, 100 mM Tris-HCl pH 9.5, 50 mM MgCl2, 0.1%Tween-20, 5 mM levamisole). Alkaline phosphatase activ-ity was demonstrated using purpleAP reagent (Boehringer-Mannheim 1442–674).

ImmunohistochemistrySections were deparaffinized, hydrated in graded etha-

nol steps and briefly rinsed in PBS. After treatment withhydrogen peroxide (3% in PBS, 30 min) to reduce endog-enous peroxidase activity, an incubation in TENG-T (10mM Tris, 5 mM EDTA, 150 mM NaCl, 0.25% gelatin,0.05% Tween-20, pH 8.0; 30 min) was performed to avoidnon-specific binding. The sections were then incubatedovernight with specific primary monoclonal antibodiesagainst mouse MLC2a (Kubalak et al., 1994), rat MLC2v(kindly provided by W. Franz, Lubeck, Germany; Katus etal., 1982), human aMHC and bMHC (Wessels et al., 1991).Binding of the first antibody was detected using a rabbitanti-mouse immunoglobulin, followed by a goat anti-rabbit immunoglobulin and finally a rabbit peroxidaseanti-peroxidase (PAP) complex. Each incubation lasted 2hr and was followed by three washes in PBS (5 min each).All sera were diluted in PBS. The visualization of the PAPcomplex was performed by incubation with 0.5 mg/ml3,38-diaminobenzidine and 0.02% hydrogen peroxide in 30mM imidazole, 1 mM EDTA (pH 7.0) buffer. Alternatively,detection of the primary antibody was performed using agoat anti-mouse b-galactosidase-coupled secondary anti-body. Detection of b-galactosidase activity was performedas described by Sanes et al. (1986) at 37°C for 2–4 hr.Sections were dehydrated and mounted in Entellan(Merck).

RESULTSThe global expression pattern of MLC2a and MLC2v in

the atrial and ventricular compartments of the developingmouse heart have been previously reported in separatestudies (O’Brien et al., 1993; Kubalak et al., 1994). To getinsight into the important changes in the outlet of theventricular compartment during septation, we have stud-ied the detailed expression pattern of the regulatoryMLC2a and MLC2v isoforms (mRNA and protein) inrelation to the expression of aMHC, bMHC and cardiactroponin I. aMHC is mostly confined to the atrial myocar-dium (Lyons, 1994), bMHC is mostly confined to theventricular myocardium (Lyons, 1994) and cardiac tropo-nin I has been reported to be absent from the outflow tractbut present in the rest of the myocardial tube (Ausoni etal., 1991). We observed that the expression pattern ofMLC2a and MLC2v (mRNA and protein) was similar inmouse and rat throughout development. Although we have

observed temporal differences in the onset of expressionbetween these species, the expression in the embryonicand fetal stages is similar and therefore no further distinc-tion has been made.

Tubular Heart StageThe first developmental stage analyzed corresponds to

the early looping heart, i.e. approximately E8.5 in mice(E10.5 in rats). At this stage the heart can be considered asa tube with prospective atrial and ventricular regions.MLC2a mRNA expression declines gradually from thevenous pole (atrial) to the arterial (ventricular) pole (Fig.1b) in a similar fashion as aMHC (Fig. 1a). In contrast,homogeneous expression of MLC2a protein is observedalong the tube (Fig. 1e). MLC2v mRNA (Fig. 1c) andprotein (Fig. 1f) are already almost entirely confined to thepresumptive ventricle; low levels of expression are ob-served in the outflow tract and in the inflow part of thepresumptive atrium (Fig. 1c).

Embryonic Heart StageWith further development (E10.5 mouse; E12.5 rat), five

morphological segments become distinguishable in theheart, i.e. inflow tract, atria, atrioventricular canal, ven-tricles and outflow tract. The outflow tract is defined as theslowly conducting myocardial arterial pole of the heartoverlying the endocardial ridges. It has a sphincter-likefunction to substitute semilunar valves that have not yetdeveloped (De Jong et al., 1992b). Its distal (cranial) end isdelimited by the anterior myocardial margin whereas theproximal (caudal) boundary is delineated by initiation ofthe trabeculated myocardium. At this stage bMHC ishomogeneously expressed in the outflow tract and ventricu-lar myocardium with only residual expression in the atrialmyocardium (Fig. 2a). The expression pattern of bMHCdoes not change with further development and thereforecan be used as a reference marker of the distal outflowtract boundary. aMHC, at this time, is predominantlyconfined to the atrial/inflow tract myocardium, with lowlevels of expression in the ventricle and outflow tract(Lyons, 1994). Cardiac troponin I mRNA shows a caudo-cranial gradient along the heart (data not shown; see e.g.Franco et al., 1998).

At this stage the patterns of expression of MLC2a andMLC2v are virtually complementary. MLC2a (mRNA andprotein) remains highly expressed in the outflow tract, theatria and the inflow tract, whereas its expression becomesdown-regulated in the ventricles. A steep gradient in thelevel of MLC2a expression demarcates the morphologicalboundary of the outflow tract with the ventricles (Fig. 2b,e). On the other hand, the expression of MLC2v mRNAremains restricted to the outflow tract, the ventricles andthe atrioventricular canal. A distinct transition of expres-sion is observed at the morphological boundary of theoutflow tract with the ventricles (high) and at the atrioven-tricular canal boundary (Fig. 2c). Similar differences be-tween outflow tract and ventricular myocardium are ob-served at the protein level (Fig. 2f).

Gene Expression During Septationof the Outflow Tract

The outflow tract becomes septated into aortic andpulmonary trunks in a craniocaudal direction from E11.5

137FATE OF THE OUTFLOW TRACT MYOCARDIUM

(mouse; E13.5 in rat) onwards. At this stage, the morpho-logical boundaries of the outflow tract are still recogniz-able. Moreover, the boundaries of expression of MLC2aand MLC2v mRNAs remain well-defined and coincide withdistinct compartments.

With further development, the morphological bound-aries of the outflow tract become difficult to distinguish(after E12.5 in the mouse; after E14.5 in the rat). At thetime that the endocardial ridges are fused in the distal poleof the outflow tract, MLC2a mRNA expression remains

Fig. 1. a,b,c: In situ hybridization on sagittal serial sections of rat E10heart hybridized against aMHC (a), MLC2a (b) and MLC2v (c) mRNAs.MLC2a mRNA shows a posteroanterior gradient of expression, higher inthe venous pole (A) and lower in the arterial pole (O). Observe theexpression of MLC2v mRNA mostly confined to the presumptive ventricle(V) and with distinct lower expression in the arterial (O) and venous (A)poles. The arrow in panel c indicates expression in the atrial myocardium.

d,e,f: Immunohistochemical localization of bMHC (d), MLC2a (e) andMLC2v (f) protein on sagittal serial sections of a mouse E9.5 embryonicheart. bMHC protein is confined to the presumptive ventricular (V) andarterial pole myocardium (O). MLC2a is expressed along the entirecardiac tube without differences between each cardiac region. MLC2v isconfined to the presumptive ventricular myocardium. Scale bar 5 60 µm(a,b,c); 80 µm (d,e,f).

138 FRANCO ET AL.

Fig. 2. a,b,c: In situ hybridization on sagittal serial sections of mouseE10 heart hybridized against bMHC (a), MLC2a (b) and MLC2v (c)mRNAs. Expression of bMHC mRNA is mostly confined to the outflowtract (OFT), ventricle (V) and atrioventricular canal (AVC). Only low levelsof expression are observed in the atrial myocardium (A). MLC2a mRNAexpression is highest in the atrial myocardium (A) and lower in the AVCand OFT. Within the ventricle, expression is distinctly lower than in theOFT or AVC. The arrows indicate the distinct transition of the expressionpattern. Note that the expression of MLC2v mRNA is mostly confined tothe ventricle, albeit lower levels of expression are observed in the OFT

and AVC myocardium. d,e,f: Immunohistochemical localization of aMHC(a), MLC2a (b) and MLC2v (c) protein on transverse serial sections of a ratE12.5 embryonic heart. Expression of aMHC protein is confined to theatrial myocardium. MLC2a protein is observed in the atrial myocardium aswell as in the OFT myocardium (arrows). MLC2v protein is confined to theOFT and right ventricular myocardium (RV), with a steep differentialexpression (arrows). The dashed line on panel d corresponds to the planeof sectioning on panels a, b and c. Scale bar 5 100 µm (a,b,c);120 µm(d,e,f).

139FATE OF THE OUTFLOW TRACT MYOCARDIUM

restricted to the non-trabeculated myocardium of the rightventricular outlet, the most distal myocardial portionsurrounding the right and left arterial valves, the atriaand the inflow tract with very low signal levels in theventricles and the former atrioventricular canal (Fig. 3a,d). A similar expression pattern is seen at the protein level(data not shown). On the other hand, MLC2v mRNA is stillexpressed at a high level in the ventricles and at a muchlower level in the former outflow tract and atrioventricularcanal (Fig. 3b, e). These boundaries coincide with thenegative expression domains of the 3F-nlacZ-9 transgene(Fig. 3c, f). Together these data demonstrate that theembryonic outflow tract myocardium has a program ofgene expression distinct from the ventricular myocardium.

To get insight into the process leading to the formation ofthe muscular component of the outlet septum, we investi-gated the MLC2a expression status of outflow tract myocar-

dial cells during septation. The myocardium of the OFTshows a homogeneously high expression of MLC2a and lowexpression of MLC2v transcripts (Fig. 4). Such differencescan be also observed for MLC2a protein and less pro-nounced for MLC2v protein (Fig. 5). At this time point,MLC2a mRNA is no longer expressed in the left ventricu-lar myocardium but can still be detected in the rightventricular myocardium (Fig. 4a). Interestingly, myocar-dial cells are observed between both arterial trunks aroundthe semilunar valve swellings (Fig. 5). These cells areaMHC negative (as shown in Fig. 5A at the protein level)and express low levels of MLC2v and cardiac troponin ImRNAs and high levels of bMHC and MLC2a mRNAs(Fig. 6a–d). More caudally, no myocardial cells within thecushion material are observed (Fig. 6e, f).

As the endocardial cushions continue to fuse in a cranio-caudal fashion along the outflow tract (E14.5 mouse; E16.5

Fig. 3. a,b,c: In situ hybridization on transverse serial sections of a3F-nlacZ-9 transgenic mouse E12.5 heart hybridized against MLC2a (a),MLC2v (b) and nlacZ (c) mRNAs. Expression of MLC2a is confined to theatrial myocardium (A) and the outflow tract (arrow), whereas low levels ofexpression are detected in the ventricular myocardium (LV). Expression ofMLC2v is mostly confined to the ventricular (LV) and outflow tractmyocardium. Lower expression is observed in the outflow tract comparedto the left ventricle. Note that the border line of expression of MLC2a andMLC2v mRNAs coincides with that defined by nlacZ transgene expres-sion (c). d,e,f: In situ hybridization on sagittal serial sections of a

3F-nlacZ-9 transgenic mouse E12.5 heart hybridized against MLC2a (a),MLC2v (b), nlacZ (c) mRNAs. Note the absence of nlacZ transgeneexpression in the outflow tract myocardium, which is similar in extensionto that observed for MLC2a positive expression. P, pulmonary trunk; LA,left atrium; RA, right atrium; AVC, atrioventricular canal; LV, left ventricle;A, atria. The dashed line on panel a correspond to the plain of sectioningon panels d, e and f. The dashed line on panel d corresponds to the planeof sectioning on panels a, b and c. Scale bar 5 250 µm (a,b,c); 300 µm(d,e,f).

140 FRANCO ET AL.

rat), myocardial cells below the semilunar valves migratetowards the fusion line and eventually become connected.During development, these myocardial cells seem to mi-grate into the endocardial ridge. These cells have a pheno-type that corresponds to the pattern of gene expressionobserved in the distal outflow tract myocardium (see e.g.Figs. 3 and 5). Conversely, the mesenchymal component ofthe outflow tract endocardial ridges gradually diminisheswith development (data not shown; Ya et al., 1998).

In the late fetal period (E16.5 mouse; E18.5 rat), MLC2a(mRNA and protein) has become restricted to the atria andinflow tract myocardium and remains detectable aroundthe base of both right and left semilunar valves (Fig. 7a).Conversely, MLC2v mRNA shows a patchy distributionaround the base of the semilunar valves, whereas expres-sion is homogeneous along the non-trabeculated region ofthe right ventricular outlet. Furthermore, low levels ofMLC2v expression are observed in discrete myocardialareas of the atria and inflow tract, i.e. lower rim of theatria (former atrioventricular canal) and the dorsal atrialmyocardium (Fig. 7b).

Neonatal PeriodAfter birth (3 neonatal days), the expression of MLC2v

mRNA is restricted to the ventricular myocardium whereit acquires a homogenous distribution, resembling theexpression of MLC2v protein (data not shown). However,MLC2a (mRNA and protein) is still observed in myocardialcells around the semilunar valves. At the level of the aorticvalve, expression is restricted to a thin rim of myocardium.In the right ventricular outlet, MLC2a mRNA expressionis also observed in the non-trabeculated right ventricularmyocardium (distal region) and the myocardium separat-ing both semilunar valves, i.e. the muscular outlet septum(Fig. 8). In 10-day-old neonates, MLC2a mRNA has be-come undetectable in the myocardium surrounding thesemilunar valves and in the right ventricle (data notshown).

DISCUSSION

In the present study we demonstrate that the embryonicoutflow tract acquires a distinct program of gene expres-sion, indicating that it has a different transcriptionalregulation from that of the ventricular myocardium. Thisprogram becomes apparent in the embryonic segmentedheart and is subsequently maintained throughout the fetalstage. Myocardial cells having the same MLC2a expres-sion status as the embryonic outflow tract are observed aspart of the muscular outlet septum, the non-trabeculatedright ventricular outlet region and at the distal myocardialrim surrounding both right and left semilunar valves inthe neonatal heart. These findings support the notion thatthe endocardial cushions become muscularized by ingres-sion of myocardial cells derived from the embryonic out-flow tract, as it has been postulated in humans, rats andchickens (De Jong et al., 1992a; Lamers et al., 1995; Ya etal., 1998). Differences between the outflow tract and theventricular myocardium are observed at both the levels ofMLC2a and MLC2v mRNA and protein, suggesting thatthe main compartment-specific regulatory mechanism istranscriptional.

An understanding of the development of the arterialpole of the heart has been hindered by the use of ambigu-ous terminology and inaccurate identification of particularstructures during morphogenesis (Icardo, 1990; Pexieder,1995), mainly due to the absence of specific markers. Wehave adopted the outflow tract definition as defined byPexieder (Pexieder, 1995), that is, the myocardial portionof the heart overlying the endocardial ridges at the arterialpole of the cardiac tube because it matches best withfunctional differences (De Jong et al., 1992b) and patternsof gene expression (this study). The distal (cranial) end ofthe outflow tract is delimited by the boundary of themyocardium and the proximal (caudal) end by the bound-ary of endocardial ridges and by the trabeculated morphol-ogy of the ventricular myocardium. Our data demonstrate

Fig. 4. Whole-mount in situ hybridization in mouse E12.5 embryonicheart against MLC2a (a) and MLC2v (b) mRNAs. Note the expression ofMLC2a in the atria (LA, RA) and in the outflow tract myocardium(asterisk). At this stage, a markedly higher expression is observed in theright versus the left ventricle. Expression of MLC2v is confined to the

ventricles (RV, LV) and outflow tract (asterisk). Note the lower expressionof MLC2v in the outflow tract myocardium in comparison with theventricular myocardium. RA, right atrium; LA, left atrium Scale bar 5200 µm.

141FATE OF THE OUTFLOW TRACT MYOCARDIUM

that expression of MLC2a remains confined to the non-trabeculated (smooth) myocardium during outflow tractseptation. Although we are aware of the caveats of usinggene expression patterns as cell fate markers, we presentherein evidences that the boundaries based on morphologi-cal criteria match those based on the molecular phenotype;the embryonic outflow tract maintains its boundariesduring development, i.e. the transition between a smoothmyocardial layer and the trabeculated myocardium. Theseresults are in line with our previous findings in transgenicmice carrying regulatory sequences of the MLC31/3f locus(Franco et al., 1997). This finding contrasts with Pexieder(1995), who suggested that the trabeculations of the rightventricle extend distally into the former outflow tractduring fetal development. In contrast to the expression ofthe 3F-nlacZ-9 transgenic mice (Franco et al., 1997), the2-week-old neonatal (and adult) heart displays no longerdifferential expression of MLCs in the ventricular outlet.

This difference suggests that the transcriptional potentialof the embryonic outflow tract is ‘‘frozen’’ in the transgenic3F-nlacZ-9 mice.

The septation of the arterial pole of the heart is acomplex process involving different cell lineages: endocar-dium-derived mesenchyme, neural crest-derived cells andmyocardial cells (see e.g. Noden et al., 1995). Duringseptation, the endocardial ridges fuse in a craniocaudaldirection (Kramer, 1942; van Mierop et al., 1963; Goor etal., 1972; Asami, 1992). At the distal end of the outflowtract, the anlage of the semilunar valves is formed (Kra-mer, 1942; Maron and Hutchins, 1974; Hurle, 1979; Hurleet al., 1980; Ya et al., 1998), and at the proximal end,spindle-like myocardial cells are observed within the myo-cardial-mesenchymal edge (Ya et al., 1998; Lamers et al.,1995). Our results show that these myocardial cells havethe same gene expression as outflow tract myocardial cells.The contiguity of these cells with the surrounding outflow

Fig. 5. Immunohistochemical localization of aMHC (a), bMHC (b)MLC2a (c) and MLC2v (d) protein on transversal serial sections of a ratE16.5 embryonic heart. Expression of aMHC protein is confined to theatrial myocardium. MLC2a protein is observed in the atrial myocardium aswell as in the former outflow tract myocardium (arrow). bMHC proteinshows a similar pattern of expression as the MLC2v protein, that is mostly

confined to the outflow tract and ventricular myocardium. Slight differ-ences are observed at this stage at the protein level for the expression ofMLC2v between the right ventricle and the former outflow tract myocar-dium (see arrow in panel d). RA, right atrium; LA, left atrium; Ao, Aorta; P,pulmonary trunk. Scale bar 5 40 µm.

142 FRANCO ET AL.

tract myocardium and their molecular phenotype suggestthat myocardial cells of the outflow tract myocardiummigrate into the endocardial cushions during the processof fusion, resulting in the formation of the muscular outlet

septum. This assumption is based on the fact that myocar-dial cells in the outlet septum maintain expression ofMLC2a (mRNA and protein), a marker of the outflow tractmyocardium.

Fig. 6. a,b,d,e: In situ hybridization on serial sections of rat E14 hearthybridized against bMHC (a), cardiac troponin I (b), MLC2a (d) andMLC2v (e) mRNAs. Note that spindle-like myocardial cells are present inthe endocardial cushions (d), displaying a similar expression pattern tothe surrounding outflow tract myocardial cells. c,f: In situ hybridization on

serial sections of rat E15 heart hybridized against MLC2a (c) and bMHC(f) mRNAs. Observe that spindle-like myocardial cells are observed in theendocardial cushions material, showing a similar expression pattern tothe surrounding outflow tract myocardial cells. Ao, Aorta; P, pulmonarytrunk. Scale bar 5 150 µm.

Fig. 7. In situ hybridization on serial sections of mouse E16 hearthybridized against MLC2a (a) and MLC2v (b) mRNAs. Observe thecomplementary patterns of expression of MLC2a and MLC2v between theformer outflow tract myocardium, surrounding the pulmonary valve

(arrows), and the right ventricular myocardium (RV). Note also theexpression of MLC2v transcript in the ventral portion of the atrial chamber.RA, right atrium, LA, left atrium. Scale bar 5 170 µm.

143FATE OF THE OUTFLOW TRACT MYOCARDIUM

The migration of the outflow tract myocytes into theendocardial ridges follows a craniocaudal gradient duringseptation. Similarly, migrating neural crest cells enter theoutflow tract in a craniocaudal fashion prior to the fusionof the endocardial ridges and therefore prior to the muscu-larization of the outflow tract septum (Kirby et al., 1983;Kirby, 1993; Conway et al., 1997). It is tempting tospeculate that the neural crest-derived cells are involvedin the myocardial migration into the cushions mesen-chyme (Ya et al., 1998; Van den Hoff et al., 1998; Poelmannand Gittenberger-de Groot, 1998). However the mecha-nisms leading to myocardial migration are as yet uncer-tain.

Regionalised gene expression in the developing hearthas been widely documented, i.e. differential expression inthe atrial and the ventricular myocardium (de Groot et al.,1987, 1989; De Jong et al., 1987, 1990; Wessels et al., 1991;Moorman and Lamers, 1994; Ya et al., 1997) and differen-tial expression in the ventricular trabeculations and theventricular compact myocardium (see for a review, Francoet al., 1996, 1998). Transient differences of gene expressionbetween right and left ventricles have been observed for asubset of endogenous genes (Lyons, 1994) and more widelydemonstrated using transgenic mice (Kuisk et al., 1996;Biben et al., 1996; Ross et al., 1996; Kelly et al., 1995).Recent data on the cardiac pattern of expression of tran-scription factors of the basic helix-loop-helix family, i.e.eHAND and dHAND, suggests that the right/left ventricu-lar compartments correspond to independent genetic units(Srivastava et al., 1997). However, the transient differ-ences between left and right ventricular myocardial expres-sion might be a reflection of a different degree of matura-tion at a particular developmental stage.

Compartmentalization of cardiac gene expression inrodents (e.g. mouse and rat) has been first observed aftercardiac looping, as most genes are evenly expressed (orexpressed in gradients) along the tubular heart (see e.g.Franco et al., 1998) in contrast to the avian system in

which regionalized atrial gene expression becomes appar-ent prior to the tubular heart stage (Yutzey et al., 1994;Yutzey and Bader, 1995). It remains unknown how thesecardiac compartments acquire their own distinct transcrip-tional potential. We favor a transcriptional model of initialgradients of signals along the anteroposterior axis of thecardiac tube (‘‘primary myocardium’’ characterized by co-expression of MHCs and MLCs). The subdivision into fivedifferent compartments (i.e. inflow tract, atria, atrioven-tricular canal, ventricles, outflow tract; for a review seeFranco et al., 1998) would be directed by selective repres-sion of the atrial program in the future ventricles andrepression of the ventricular program in the future atriabut in the flanking segments the original program hasbeen largely maintained. The maintenance of co-expres-sion of MLC2a and MLC2v in the outflow tract supportsthis model. It reflects the primitive character of the outflowtract, reminiscent of the primary heart tube that allows itto function as a sphincter to substitute for the valves. Theregionalization of gene expression in the outflow tract andatrioventricular canal distinct from the ventricles sup-ports the modular arrangement of transcriptional poten-tial in the myocardium.

ACKNOWLEDGMENTSWe gratefully thank Dr. Robert Kelly, Dr. Peter Zammit

and Dr. Margaret Buckingham (Department of MolecularBiology, Pasteur Institute) for providing us with 3F-nlacZ-9 transgenic mice, and for helpful assistance withthe whole-mount in situ hybridization technique. We areindebted to Dr. Steve Kubalak for the cDNA probe againstMLC2a mRNA, to Dr. Kenneth Boheler for the cDNAprobes against aMHC and bMHC mRNAs, to Dr. KennethChien for the cDNA probe against MLC2v mRNA and theMLC2a antibody, to Dr. Robert Kelly for the cDNA probeagainst lacZ mRNA, to Dr. Stefano Schiaffino for the cDNAprobe against cardiac troponin I mRNA, and to Dr. Wol-gang Franz for supplying the MLC2v antibody. We alsothank Dr. Maurice J.B. van den Hoff for his valuablecomments, Corrie de Gier-de Vries for her technical help,and Cars. E. Gravemeijer and Cees J. Hersbach for theirvaluable help with imaging.

LITERATURE CITEDAsami I. 1992. Partitioning of the arterial end of the embryonic heart.

In: Van Praagh R, Takao A, editors. Etiology and morphogenesis ofcongenital heart disease, vol 4. Mount Kisco, NY: Futura. p 51–61.

Ausoni S, de Nardi C, Moretti P, Gorza L, Schiaffino S. 1991.Developmental expression of rat cardiac troponin I mRNA. Develop-ment 112:1041–1051.

Bartelings MM, Gittenberger-de Groot AC. 1989. The outflow tract ofthe heart: embryologic and morphologic correlations. Int J Cardiol22:289–300.

Biben C, Hadchouel J, Tajbakhsh S, Buckingham M. 1996. Develop-mental and tissue-specific regulation of the murine cardiac actingene in vivo depends on distinct skeletal and cardiac muscle specificenhancer elements in addition to the proximal promoter. Dev Biol173:200–212.

Boheler KR, Chassagne C, Martin X, Wisnewsky C, Schwartz K. 1992.Cardiac expressions of a and bmyosin heavy chains and sarcomericaactins are regulated through transcriptional mechanisms. J BiolChem 267:12979–12985.

Conway SJ, Henderson DJ, Copp AJ. 1997. Pax3 is required forcardiac neural crest migration in the mouse: evidence from thesplotch (Sp2H) mutant. Development 124:505–514.

De Groot IJM, Sanders E, Visser SD, Lamers WH, De Jong F, Los JA,

Fig. 8. In situ hybridization on a 3 day neonatal rat heart hybridizedagainst MLC2a mRNA. Observe the remaining expression of MLC2amRNA in the surrounding area of the aortic valve and also in thenon-trabeculated myocardial area of the right ventricular infundibulum(arrows). Ao, Aorta; rv, right ventricle; P, pulmonary valve. Scale bar 550 µm.

144 FRANCO ET AL.

Moorman AFM. 1987. Isomyosin expression in developing chickenatria: a marker for the development of conductive tissue? AnatEmbryol 176:515–523.

De Groot IJM, Lamers WH, Moorman AFM. 1989. Isomyosin expres-sion pattern during rat heart morphogenesis: an immunohistochemi-cal study. Anat Rec 224:365–373.

De Jong F, Geerts WJC, Lamers WH, Los JA, Moorman AFM. 1987.Isomyosin expression patterns in tubular stages of chicken heartdevelopment: a 3D immunohistochemical analysis. Anat Embryol177:81–90.

De Jong F, Geerts WJC, Lamers WH, Los JA, Moorman AFM. 1990.Isomyosin expression pattern during formation of the tubularchicken heart: a 3D immunohistochemical analysis. Anat Rec 226:213–227.

De Jong F, Lamers WH, Moorman AFM. 1992a. The myocardializationof membranous septa in the developing chicken heart. Proceedingsof the Bilthoven meeting of the ‘‘Developmental Anatomy andPathology’’ working group of the European Society of Cardiology.

De Jong F, Opthof T, Wilde AAM, Janse MJ, Charles R, Lamers WH,Moorman AFM. 1992b. Persisting zones of slow impulse conductionin developing chicken hearts. Circ Res 71:240–250.

De la Cruz MV, Sanchez Gomez C, Arteaga MM, Arguello C. 1977.Experimental study of the development of the truncus and the conusin the chick embryo. J Anat 123:661–686.

De la Cruz MV, SanchezGomez C, Palomino M. 1989. The primitivecardiac regions in the straight tube heart (stage 9) and theiranatomical expression in the mature heart: an experimental studyin the chick embryo. J Anat 165:121–131.

Franco D, Ya J, Wagenaar GTM, Lamers WH, Moorman AFM. 1996.The trabecular component of the embryonic ventricle. In: Ostadal B,Nagano M, Takeda N, Dhalla NS, editors. The developing heart.New York: Lippincott Raven. p 51–60.

Franco D, Kelly R, Lamers WH, Buckingham M, Moorman AFM. 1997.Regionalised transcriptional domains of myosin light chain 3Ftransgenes in the embryonic mouse heart: morphogenetic implica-tions. Dev Biol. 188:17–33.

Franco D, Lamers WH, Moorman AFM. 1998. Patterns of geneexpression in the developing myocardium: towards a morphologi-cally integrated transcriptional model. Cardiovasc Res 38:25–53.

Goor DA, Dische R, Lillehei CW. 1972. The conotruncus. I. Its normalinversion and conus absorption. Circulation 46:375–384.

Henrique D, Adam J, Myat A, Chitnis A, Lewis J, IshHorowicz D. 1997.Expression of a delta homologue in prospective neurons in the chick.Nature 375:787–790.

Hogan B, Beddington R, Costantini F, Lacy E. 1994. In: Manipulatingthe mouse embryo. Cold Spring Harbor, NY: Cold Spring HarborLaboratory Press.

Hurle JM. 1979. Scanning and light microscope studies of the develop-ment of the chick embryo semilunar heart valves. Anat Rec 102:289–298.

Hurle JM, Colvee E, Blanco AM. 1980. Development of mouse semilu-nar valves. Anat Embryol 160:83–91.

Icardo JM. 1990. Development of the outflow tract: a study in heartswith situs solitus and situs inversus. In: Bockman DE, Kirby ML,editors. Embryonic origins of defective heart development. NewYork: New York Academy of Sciences. p 26–40.

Katus HA, Hurrell JG, Matsueda GR, Ehrlich P, Zurawski VR, KhawB-A, Haber E. 1982. Increased specificity in human cardiac-myosinradioimmunoassay utilizing two monoclonal antibodies in a doublesandwich assay. Mol Immunol 19:451–455.

Kelly R, Alonso S, Tajbakhsh S, Cossu G, Buckingham M. 1995.Myosin light chain 3F regulatory sequences confer regionalisedcardiac and skeletal muscle expression in transgenic mice. J CellBiol 129:383–396.

Kirby ML. 1993. Cellular and molecular contributions of the cardiacneural crest to cardiovascular development. Trends Cardiovasc Med3:18–23.

Kirby ML, Gale TF, Stewart DE. 1983. Neural crest cells contribute tonormal aorticopulmonary septation. Science 220:1059–1061.

Kramer TC. 1942. The partitioning of the truncus and conus and theformation of the membranous portion of the interventricular septumin the human heart. Am J Anat 71:343–370.

Kubalak SW, MillerHance WC, O’Briens TX, Dyson E, Chien KR.1994. Chamber specification of atrial myosin light chain2 expressionprecedes septation during murine cardiogenesis. J Biol Chem269:16961–16970.

Kuisk IR, Li H, Tran D, Capetanaki Y. 1996. A single MEF2 sitegoverns desmin transcription in both heart and skeletal muscleduring mouse embryogenesis. Dev Biol 174:1–13.

Laane HM. 1979. The septation of the arterial pole of the heart in thechick embryo. IV. Discussion. Acta Morphol Neerl-Scand 17:21–32.

Lamers WH, Viragh S, Wessels A, Moorman AFM, Anderson RH. 1995.Formation of the tricuspid valve in the human heart. Circulation91:111–121.

Lyons GE. 1994. In situ analysis of the cardiac muscle gene pro-gramme during embryogenesis. Trends Cardiovasc Med 4:70–77.

Maron BJ, Hutchins GM. 1974. The development of the semilunarvalves in the human heart. Am J Pathol 74:331–344.

McBride RE, Moore GW, Hutchins GM. 1981. Development of theoutflow tract and closure of the interventricular septum in thenormal human heart. Am J Anat 160:309–331.

Melton DA, Krieg PA, Rebagliati MR, Maniatis T, Zinn K, Green MR.1984. Efficient in vitro synthesis of biologically active RNA and RNAhybridization probes from plasmids containing a bacteriophage SP6promoter. Nucleic Acids Res 12:7035–7056.

Moorman AFM, Lamers WH. 1994. Molecular anatomy of the develop-ing heart. Trends Cardiovasc Med 4:257–264.

Moorman AFM, de Boer PAJ, Vermeulen JLM, Lamers WH. 1993.Practical aspects of radioisotopic in situ hybridization on RNA.Histochem J 25:251–260.

Moorman AFM, Vermeulen JML, Koban MU, Schwartz K, LamersWH, Boheler KR. 1995. Patterns of expression of sarcoplasmicreticulum Ca21 ATPase and phospholamban mRNAs during ratheart development. Circ Res 76:616–625.

Noden DM, Poelmann PE, Gittenbergerde-Groot AC. 1995. Cell ori-gins and tissue boundaries during outflow tract development.Trends Cardiovasc Med 5:69–75.

O’Brien TX, Lee KJ, Chien KR. 1993. Positional specification ofventricular myosin light chain 2 expression in the primitive murineheart tube. Proc Natl Acad Sci USA 90:5157–5161.

Orts Llorca F, Puerta Fonolla J, Sobrado J. 1982. The formation,septation and fate of the truncus arteriosus in man. J Anat134:41–56.

Pexieder T. 1995. Conotruncus and its septation at the advent of themolecular biology era. In: Clark EB, Markwald RR, Takao A, editors.Developmental mechanisms of heart disease. Armonk, NY: Futura.p 227–247.

Poelmann RE, Gittenberger-de Groot AC. 1998. Apoptosis in theouftlow tract: a trigger for myocardialisation? Proceedings of Key-stone Symposia on Molecular Biology of the Cardiovascular System,p 38.

Ross RS, Navankasattusas S, Harvey RP, Chien KR. 1996. AnHF1a/HF1b/MEF2 combinatorial element confers cardiac ventricu-lar specificity and established an anteriorposterior gradient ofexpression. Development 122:1799–1809.

Sanes JR, Rubenstein JLR, Nicolas JF. 1986. Use of recombinantretrovirus to study postimplantation cell lineage in mouse embryo.EMBO J 5:3133–3142.

Schiaffino S, Samuel JL, Sassoon D. 1989. Nonsynchronous accumula-tion of askeletal actin and bmyosin heavy chain mRNAs duringearly stages of pressure overload induced cardiac hypertrophydemonstrated by in situ hybridization. Circ Res 64:937–948.

Srivastava D, Thomas T, Lin Q, Kirby ML, Brown D, Olson EN. 1997.Regulation of cardiac mesodermal and neural crest development bythe bHLH transcription factor, dHAND. Nat Genet 16:154–160.

Sumida H, Akimoto N, Nakamura H. 1989. Distribution of the neuralcrest cells in the heart of birds: a three dimensional analysis. AnatEmbryol 180:29–35.

Thompson RP, Fitzharris TP. 1979a. Morphogenesis of the truncusarteriosus of the chick embryo heart: tissue reorganization duringseptation. Am J Anat 156:251–264.

Thompson RP, Fitzharris TP. 1979b. Morphogenesis of the truncusarteriosus of the chick embryo heart: the formation and migration ofmesenchymal tissue. Am J Anat 154:545–556.

145FATE OF THE OUTFLOW TRACT MYOCARDIUM

Thompson RP, Abercrombie V, Wong M. 1987. Morphogenesis of thetruncus arteriosus of the chick embryo heart: movements of autora-diographic tattoos during septation. Anat Rec 218:434–440.

Van den Hoff MJB, Wessels A, Markwald RR, Lamers WH, MoormanAFM. 1998. Myocardialization as a tool in reconstruction of thefailing myocardium. Proceedings of the Weinstein Symposium,Nashville, Tennessee.

van Mierop LHS, Alley RD, Kausel HW, Stranahan A. 1963. Pathogen-esis of transposition complexes. I. Embryology of the ventricles andgreat arteries. Am J Cardiol 12:216–225.

Wessels A, Vermeulen JLM, Viragh S, Kalman F, Lamers WH,Moorman AFM. 1991. Spatial distribution of ‘‘tissue specific’’ anti-gens in the developing human heart and skeletal muscle. II. Animmunohistochemical analysis of myosin heavy chain isoform expres-sion patterns in the embryonic heart. Anat Rec 229:355–368.

Ya J, Markman MWM, Wagenaar GTM, Blommaart PJE, MoormanAFM, Lamers WH. 1997. Expression of the smooth-muscle proteins

alpha smooth-muscle actin and calponin, and of the intermediatefilament protein desmin are parameters of cardiomyocytes matura-tion in the prenatal rat heart. Anat Rec 249:495–505.

Ya J, van den Hoff MJB, de Boer PAJ, TesinkTaekema S, Franco D,Moorman AFM, Lamers WH. 1998. The normal development of theoutflow tract in the rat. Circ Res 82:464–472.

Yutzey KE, Bader D. 1995. Diversification of cardiomyocytes celllineages during early heart development. Circ Res 77:216–219.

Yutzey KE, Rhee JT, Bader D. 1994. Expression of the atrial-specificmyosin heavy chain AMHC1 and the establishment of anteroposte-rior polarity in the developing chicken heart. Development 120:871–883.

Zammit PS, Buckingham M. 1997. Non-isotopic whole mount in situhybridisation of mRNA in intact mouse embryos. In: Partridge T,Rickwood D, editors. Multimedia methods in molecular biology onCD-ROM, 1st ed. London: Chapman and Hall.

146 FRANCO ET AL.