a quantitative study of the ventricular myoarchitecture in europ j morph.chick... · its acute...

European Journal of Morphology 0924-3860/98/3602-0105$12.001998, Vol. 36, No. 2, pp. 105-119 0 Swets & Zeitlinger

A QUANTITATIVE STUDY OF THE VENTRICULAR MYOARCHITECTURE IN

THE STAGE 21-29 CHICK EMBRYO FOLLOWING DECREASED LOADING

David Sedmeral, Tomas Pexiederl, Norman Hu2 and Edward B. Clark2

lInstitute of Histology and Embryology, University of Lausanne, Lausanne, Switzerland2National Institutes of Health, Specialized Center of Research in Pediatric Cardiovascular Diseases, Strong

Children’s Research Center, Department of Pediatrics, University of Rochester School of Medicine &Dentistry, Rochester, New York 14642, U.S.A.

ABSTRACT

During the early developmental period, ventricular myoarchitecture undergoes a transition from a smooth-walledcardiac tube, to left and right ventricular chambers filled with a sponge-like network of trabecular struts. Wemeasured the quantitative changes of ventricular myocardium properties in normal stage 21-29 chick embryos andafter chronic verapamil suffusion, which is known to decrease work load and decelerate ventricular growth. Themorphologic parametres (compact layer thickness, ventricular wall composition, porosity of different layers andtrabecular orientation) were determined from scanning electron micrographs of transversely dissected perfusion-fixed hearts. A vascular bed of stage 21 chick embryos was suffused with 1 ng of verapamil at 1 pl per hour up tostages 24, 27 and 29 via a miniosmotic pump. From stage 24, the thickness of the compact myocardium in the leftventricle was greater than that of the right. The increase in thickness was minimal between stages 24 and 27, whilethe predominantly radially arranged trabeculae comprised up to 75% of total myocardial mass. The ratio of inter-trabecular spaces to trabeculae (local porosity) decreased from the ventricular center (70%) towards the compactmyocardium (0%). In verapamil-treated embryos, the hearts were smaller and showed delayed development. Thecompact myocardium was thinner than normal, and the proportion of trabeculae was higher than in controls. Thelocal porosity values were similar in control and experimental groups. Decreased load resulted in delayed growth andmorphogenesis, expressed as a persistence of trabeculae and a thinner compact myocardium. Embryonic heartpumping function is largely based on extensively developed trabeculation with regionally different properties.

KEYWORDS: Compact myocardium, image analysis, perfusion-fixation, trabecular orientation, chronic verapamilsuffusion.

ABBREVIATIONS: AV, atrioventricular junction; AVCa, anterior atrioventricular cushion; AVCp, posterior atrio-ventricular cushion; BT, basal trabecular layer; Co, compact layer of the ventricular wall; CLT, compact layerthickness; Ct, conotruncus; En, endocardium; Ep, epicardium; IS, interventricular septum; ITS, intertrabecularspace; LA, left atrium; LT, luminal trabecular layer; LV, left ventricle; RA, right atrium; RV, right ventricle; SEM,scanning electron microscopy; Tr, trabecula; V, ventricle

INTRODUCTION either macroscopic (Streeter, 1979) or microscopic(Usson et al., 1994) examination of the ventricular

The existence of a specific pattern of cardiomyocyte wall. During the embryonic period, the trabeculae areassembly in different regions of the heart is clear from radially arranged in the ventricles (Challice &

Address correspondence to: David Sedmera, Institute ofViragh, 1973; Ben-Shachar et al., 1985; Icardo &

Physiology, University of Lausanne, Rue du Bugnon 7, CH- Fernandez-Teran, 1987). Together with biomechani-

1005 Lausanne, Switzerland. Fax +41 21 692 5505; e-mail cal engineers (Taber et al., 1993), we believe that [email protected] architecture is of crucial importance for the heart’s

106 D. SEDMERA ET AL.

pumping function (de Jong et al., 1992; Broekhuisenet al., 199.5; Hu & Keller, 1995). The patterns oftrabecular architecture are modified during the courseof development (Rychterova, 1971; Sedmera et al.,1997) and are influenced by haemodynamics (Rychter& Rychterova, 1981; our unpublished data). Modifi-cations of the myocardial architecture were observedin the retinoic acid-treated mouse embryos (Pexiederet al., 1992), mice lacking the normal RXRa gene(Kastner et al., 1994; Sucov et al., 1994), as well assome other recently produced mutants (e.g. Kitsu-kawa et al., 1995). These studies prompted us to de-velop a simple method for systematic, both qualita-tive and quantitative, description of the embryonicmyoarchitecture.

Several investigations have been performed on themyocardial architecture of human fetal and adulthearts (Usson et al., 1994; Sanchez-Quintana et al.,1995), and various methods have been developed forstudying muscle fibre orientation (for review seeStreeter, 1979). Jouk et al. (1995) have recently pro-vided a new computer-based method enabling three-dimensional reconstruction and tracing of individualmuscle fibres. In the developing mouse heart, semi-automated technique based on serial sections was de-scribed by McLean et al. (1989), and demonstratedthe gradual development of three-layered system.However, only few qualitative descriptive studieshave been performed on the embryonic chick heart(Rychter & Rychterova, 1981; Steding et al., 1982;Ben-Shachar et al., 1985; Icardo & Fernandez-Teran,1987). Wenink (1992) contributed to the study of hu-man embryonic left and right trabecular patterns, andthe quantification of volume proportions of compactand trabeculated myocardium was done by Blausen etal. (1990). Unlike that of mature bird or mammalianhearts where the vast majority of muscle mass is con-tained in the compact myocardium, the embryonicmyocardium is a sponge-like structure with most my-ofibrils localised in the trabeculae (Markwald, 1969;Challice & Viragh, 1973; Tokuyasu, 1990).

The trabeculation in the embryonic ventricle servesseveral purposes distinct from those in adult trabecu-lation. It increases contractile myocardial mass in theabsence of coronary circulation (Tokuyasu, 1990),participates in the compartmentalization of blood inthe pre-septation heart (Hogers et al., 1995), and con-tributes to the formation of the muscular interven-tricular septum (Ben-Shachar et al., 1985; De La Cruzet al., 1997). The trabeculae are postulated to be an

important element of ventricular conduction (De Jonget al., 1992). Chuck et al. (1997) have recently de-scribed the development of ventricular activation pat-tern in the chick embryo. At tubular stages, in mosthearts it is baso-apical, corresponding with the peri-staltoid contraction pattern. In the majority of trabec-ulated hearts (stages 24-3 I), the left ventricular acti-vation is almost simultaneous, pointing out the role ofthe radially arranged trabeculae in excitation spread.The adult apico-basal activation pattern predominatesfrom stage 31 onwards. The compaction of the basalparts of the trabeculae contributes substantially to thethickening of the compact myocardium (Rychterova,1971). Remodeling of the remaining luminal partsgives rise to the adult trabeculation pattern and thepapillary muscles (Sedmera et al., 1997).

Our goal is to analyze the developmental changesof embryonic ventricular myocardial architecture. Weperformed quantitative analysis of the composition ofthe ventricular wall, local porosity and trabecular ori-entation in different areas in early chick embryonichearts using techniques of perfusion fixation and SEM(Moscoso & Pexieder, 1990) followed by planimetryand image analysis. The relevance of measured quan-titative parameters was tested under experimentallyreduced load conditions induced by chronic verapamilsuffusion (Clark et al., 1991). Verapamil is a calciumantagonist (class IV antiarythmic drug), decreasingthe heart contractility and pressure via vasodialata-tion. Its acute administration in the chick embryo ex-hibits dose-dependent lethality (Ostadal et al., 1987)and results in general growth retardation. It is not yetclear whether it is a direct systemic effect (Jaffee &Jaffee, 1989) or a result of hypoperfusion and there-fore decreased tissue metabolism. Chronic adminis-tration in dose used by Clark et al. (199 1) and in thepresent study decreases significantly heart rate, systo-lic blood pressure, ventricular dP/dt and dorsal aorticblood flow at matched developmental stages for 24-72 hours after beginning of suffusion at stage 21,therefore indicating a direct and specific influence onthe developing cardiovascular system.

Our results indicate that normal chick heart devel-opment between stages 21-31 is characterized by agradual increase in proportion and thickness of thecompact myocardium; this is more pronounced in theleft ventricle than the right. The hearts of verapamil-treated embryos are generally smaller and have lesscompact myocardium and more normally orientedtrabeculae than their normal counterparts. This shows

CHICK VENTRICULAR MYOARCHITECTURE 107

the correlation between cardiac growth and structualadaptations to developmentally increasing or experi-mentally changed load conditions.

MATERIAL AND METHODS

White leghorn eggs were incubated in a forced-draftincubator at 37.5”C and 70% relative humidity. Theembryos were sampled at HH stages (Hamburger andHamilton, 1951) 21, 24, 27 and 29. The hearts wereperfusion-fixed with 2% glutaraldehyde- 1% formal-dehyde in isotonic (280 mOsmol/l) 0.1 M cacodylate

Fig. 1. The schematic representation of part of the ventricular wallwith an illustration of the measurements taken. The wallcross-section is subdivided into the compact myocardium(dark gray), and the basal (BT) and luminal (LT) parts ofthe trabeculae (light gray). Their surface areas, togetherwith the intertrabecular spaces (ITS), were used for theevaluation of the ventricular wall’s composition. The di-ameter (feret) of one intertrabecular space measured at 0”and its length and orientation are also shown. CLT, com-pact layer thickness.

buffer and postfixed in 1% osmium tetroxide (Pex-ieder, 1981). They were photographed in 3 standardprojections (frontal, right and left profiles) using aM400 Wild Photomakroskop. This fixation removesall blood, arresting the hearts in end diastole. Thehearts were then cut transversely with speciallythinned microdissection scissors into slices of 0.25-0.50 mm thickness. These were then ethanol-dehy-drated and critical point dried (Balzers CPD 030) us-ing Freon. After mounting on stubs with colloidalsilver, the specimens were sputter coated (Edwards S150) with 300 nm of gold and examined using a JEOLJSM 630 OF scanning electron microscope.

For image analysis on a Quantimet 970 we usedline drawings of intertrabecular spaces traced fromSEM micrographs because variations in the greyscales of the prints prevented reliable automatic de-tection of tissue/space boundaries. The structure of across section of the chick embryonic ventricle is rath-er complex. From outside\ to inside, it consists of epi-cardium, compact myocardial mantle, a meshwork oftrabeculae covered with endocardium and, in someareas, trabecula-free lumen. To determine the relativevolumes of the three main wall components: the com-pact myocardium, trabeculae and intertrabecular spac-es, we measured the area contribution of each of thesecomponents, which together compose 100% of thewall cross section (Fig. 1). The histograms of ven-tricular wall composition are based on averages fromat least three hearts; they were corrected for slightlydifferent levels of slicing using linear regression. Thisapproach preserved the trends seen in individualhearts, but, rendered some of the statistical analysisinvalid. The porosity of individual areas (layers) ofthe trabeculated myocardium was determined as a per-centage of the total layer area occupied by the inter-trabecular spaces. By this definition, the porosity ofthe compact layer is 0%. ,

To obtain an image for analysis, we oriented theslices so they were viewed looking towards the apex,with the dorsal side at the top of the image (Fig. 2). Incases where they were viewed from the apical side(e.g. no trabeculae visible on the non-apical side) lat-eral invertion of the image preserved the standard ori-entation. The first of two techniques used for the eval-uation of trabecular orientation used a polar histogram(adapted from Usson et al., 1994) of the lengths of theintertrabecular spaces oriented in each direction. Thenecessary data were obtained by measuring the lengthand orientation of individual intertrabecular spaces


Fig. 2. Transversely dissected heart of HH stage 24 embryo. (a)Frontal view showing levels of slicing and the direction ofobservation (arrows). (b)-(e) Transversal slices from baseto apex. Note the differences between the patterns in theleft and right ventricles as well as bet ween the apex andmidportion slices. The atrioventricular cushions begin theseptkion of the atrioventricular canal. Scale bars, 100 pm. (f) Higher magnification of the lateral part of the right ventricle at HHstage 24. Details of the myocardial architecture can be seen and the three main wall constituents (compact layer, trabeculae andintertrabecular spaces) are easily distinguished. Arrows indicate the boundary between the basal and luminal layers in the trabec-ulated myocardium. Scale bar, 10 pm.


(Fig. 1), and then summing their lengths in 10” inter-vals. The expression of values as percentages enableda meaningful comparison of the different regions tobe made. This method is technically simple and theresults are essentially the same as direct measure-ments on individual skeletonized trabeculae (Sedmera& Pexieder, 1995). To obtain data enabling the con-struction of ellipses representing orientation in polarcoordinates, we adapted the mean linear intercepttechnique (Cowin, 1986) to use feature measurementsmade on a Quantimet 970. In each direction (unit step11.25”), the diameters (ferets, Fig. 1) of all inter-trabecular spaces were measured and their averagecalculated. These values were then plotted against theangle of measurement in polar coordinates, resultingin an ellipse-like curve representing the size and shapeof the “mean” intertrabecular space. This approachgives the same results as the original method (Cowin,1986), in which the dissector is rotated and the totallength of intercepts with intertrabecular spaces in eachdirection is divided by the number of intercepts. Forfurther morphometrical analysis, the intertrabecularspaces were classified according to their orientation,area, and roundness. All these techniques were usedseparately for different ventricular regions.

The thickness of the compact layer was measuredon SEM photographs using a digimatic caliper. Forconsistency and comparison, the midportion slice,containing both right and left ventricle, was selectedand standard points (at the intercepts of lines drawn atO”, +45” and -45” from the centre of the interven-tricular septum with the lateral wall) were used. Ac-cording to measurements of maximal transverse di-ameter performed on macrophotographs taken beforedissection (in isoosmotic buffer) and on SEM micro-graphs, between stages 2 l-29, the linear dimensionsshrink to approximately 52% of the wet values. Toapproximate the in viva diastolic dimensions, the val-ues of compact layer thickness can be multiplied byfactor 1.92. For all quantitative measurements a min-imum of three hearts per stage were analysed. Statis-tical analysis was done using the Wilcoxon one-waynon-parametrical test (SAS) for evaluating the differ-ences between stages and groups and a paired t-testwas used for comparison between the left and rightventricle at any one stage. Results ~~0.05 were con-sidered significant.

The experiments with chronic verapamil treatmentwere performed as described by Clark et al. (199 1).After opening of the membranes, one end of a piece

of PE-60 polyethylene tubing was positioned on thesurface of the extraembryonic vascular bed of HHstage 21 (incubation day 3 l/2) chick, and the otherend attached to the flow modulator of an Alzet mini-osmotic pump (Alza, Palo Alto, CA). The pump sys-tem was filled with 0.9% saline solution and im-mersed in normal saline. The pump had a constantflow rate of 1 l.tl/h with reservoir capacity 200 fl. Theextraembryonic vascular bed was suffused with vera-pamil at 1 ng per hour, and the controls with 0.9%saline alone. The eggs were then reincubated for 24,48 and 72 hours up to (approximately) HH stages 24,27 and 29, respectively. These stages were selectedbecause of approximately a two fold increase in em-bryo mass between them. Sampling, dissection, andanalysis was performed in the same way as describedearlier. As the values from embryos treated with 0.9%saline did not differ from those of normal embryos,measurements performed on normal embryos wereused for comparison.

RESULTS

Qualitative descriptionAt HH stage 21, the common ventricle had a consid-erable trabecula-free lumen and no distinguishableinterventricular septum. However, nascent trabecularpatterns were apparent. Thick, dorsoventrally alignedridges were found in the apex, and later in the central(luminal) part of the left ventricle. In the right lateralwall, the trabeculae were thinner, shorter, and orient-ed perpendicular to the outer compact myocardial lay-er.

Between HH stages 21 and 24 (Fig. 2), the trabec-ulation increased and the right ventricle was definedby the interventricular septum. From stage 24 thetrabecular patterns of the prospective left and rightventricles were dramatically different. In the left ven-tricle, the trabeculae were dorsoventrally orientedwith numerous interconnecting segments at the apex.They formed sickle-like folds (trabecular sheets) withno connections in the midportion of the ventricle andthe principal bundles were perpendicular to the outercompact myocardium. In the right ventricle, therewere more interconnecting segments and, as a result,the individual trabeculae were shorter, without a dis-tinct pattern; they were also thinner than those of theleft ventricle. This arrangement was reflected by morecircular and smaller intertrabecular spaces and con-


Fig. 3 . Stage 24 heart treated by chronic verapamil infusion. (a)The heart is distinctly smaller when compared with Fig. 2aand the single ventricle shows no external signs of septa-tion. (b) - (e) No internal interventricular partitioning isvisible; this delay is visible on the AV level where thecushions are much less developed than in the control atFigure 2c. The arrow indicates the entry into the conotrun-cus. Scale bars, 100 w.

firmed by mean linear intercept diagrams. In polarhistograms the patterns observed were narrow, verti-cal clusters in the left ventricle and wide almost sem-icircular ones in the right ventricle.

At the first sampling interval, the verapamil-treat-ed embryos were growth retarded (as appreciated ac-cording to HH stages) and the hearts resembled thoseof HH stage 21 (Fig. 3). The right ventricle could notbe distinguished and as such the common ventriclewas considered the left ventricle for measurement pur-poses. Its compact layer was significantly thinner thannormal for left ventricle on HH stage 24.

By HH stage 27, the pattern of trabeculation hadnot changed significantly. A fine trabeculation net-

e

work almost entirely filled both ventricular cavities.The interventricular septum as well as the compactmyocardial layer were about the same thickness as inthe previous stage. The hearts of verapamil-treatedembryos were visibly smaller. In the left ventricle,there was a decreased contribution of the compactlayer to the ventricular cross section area accompa-nied by an increase of relative trabecular mass; thecompact layer thickness was markedly reduced. Sim-ilar thinning was observed in the right ventricle.

At HH stage 29 the trabeculae and the left ven-tricular compact layer were distinctly thickened (Fig.4). A trabecula-free lumen reappeared in the basalportions of both ventricles which were previously

CHICK VENTRICULAR MYOARCHITECTURE

Fig. 4. (a) Midportion slice from HH stage 29 heart. Both ventriclesare visible and individual wall components are labeled. Theasterisks in the right ventricle indicate points of measurementof the compact layer thickness. The arrows indicate the bound-ary between the basal and luminal layers in the left ventriculartrabeculated myocardium. Scale bar, 100 /_tm. (b) The meanlinear intercept diagram quantifies the differences in shape,size and orientation of the intertrabecular spaces in differentareas. Dorsoventral trabecular sheets enclosing oblong intertrabecular spaces are represented by a rather oblong ellipse orientedalmost vertically. The variable orientation of smaller and more rounded spaces in the right ventricle gives rise to a more circularcurve. The basal layers of both ventricles are represented by much smaller, almost circular curves. (c) and (d) The orientationhistograms for these areas reveal the similarity (almost semicircular pattern with no prevailing direction) between the basal layersof both ventricles in contrast to differences in arrangement in the luminal ones. The direction of the main vertical cluster in the leftventricle matches the orientation of the long axis of the ellipse in (b). The most prominent cluster at about 145” in the rightventricle results from its sickle-like cross section being more than 180” wide so this direction is encountered twice. The diagramcan not distinguish between the two in this modality.


Fig. 5. Midportion slice from stage 29 heart treated with verapamil. In comparison with normal heart at this stage (Fig. 4a) the matchedslice is smaller, and the compact myocardium, trabeculae and interventricular septum are thinner. Despite this, the left and righttrabecular patterns are normal. Scale bar, 100 pm.

filled by trabeculations. There were regional differ-ences in the orientation of the trabeculae. In the apicalpart of the left ventricle, they maintained almost iso-tropic arrangement as seen in Figure 2. In the midpor-tion, the trabecular sheets preserved the dorsoventralalignment observed in earlier stages. They assume’dan alignment parallel to the lateral wall, or the inter-ventricular septum when they were close to thesestructures. In the right ventricle the trabeculae werearranged in a fan-like pattern, radiating from the in-terventricular septum perpendicular to the outer com-pact layer. Unlike the fenestrated trabecular sheets inthe left ventricle, they were more rod-like entities.

In the non-compact (trabeculated) part of the ven-tricular wall two different layers can be distinguishedfrom stage 24: basal parts of the trabecular sheetsadjacent to the compact myocardium, and central por-tions located luminally. In the basal area the trabecu-lae were fine, short, and oriented perpendicular to theouter compact layer. This layer was quite thin (onaverage 50 ~_un), the intertrabecular spaces were small,rounded (Figs. 2, 4), and there was no difference inpattern between the left and right ventricle. Thesefeatures were best demonstrated by the mean linear

COMPACT LAYER THICKNESS8 0

l LV normal * pco.05

7 0

6 0

A RV normal0 LV verapamilA RV verapamil

S.D.

I

Ofi21 2 4 2 7 2 9

HH STAGE

Fig. 6. Increase in compact layer thickness (mean&SD) during de-velopment. Note the much smaller slope for the right ven-tricle, as well as virtually no difference between stages 24and 27. The values for verapamil-treated hearts are de-creased.

CHICK VENTRICULAR MYOARCHITECTURE

Fig. 7. The histograms of ventricular wall composition at different developmental stages. The decrease in the proportion of trabeculae inthe left ventricle is accompanied by an increasing proportion of compact myocardium. The percentage of compact layer shows anapicobasal gradient. The proportion of trabeculae is relatively stable in the right ventricle and regional differences are lesspronounced. Trabeculae form most of the ventricular myocardial mass during these stages.


intercept diagram (Fig. 4). The differences betweenthe different layers and ventricles were verified bythe classification of the intertrabecular spaces accord-ing to their area, length, roundness, and orientation(Sedmera & Pexieder, 1995; and unpublished data).

The verapamil-treated hearts were markedly small-er than normal by stage 29 (Fig. 5). The compactlayers of both ventricles were thinner, but not signifi-cantly (Fig. 6). The proportion of trabeculae relativeto the other components was augmented in both ven-tricles (Fig. 8). The pattern of trabecular orientationwas normal.

Quantitative measurementsAt all stages examined, the compact layer in the leftventricle was significantly thicker than in the rightventricle (Fig. 6). The left to right ratio increased from2: 1 at HH 24 to 2.5: 1 at HH 29. The increase in thick-ness between the individual stages was not significantQ~0.05) between HH stages 24 and 27 in the left, and24 and 27, and 27 and 29 in the right ventricle.

The composition of the ventricular wall showedspatial and temporal gradients (Fig. 7). At HH stage2 1, the single non-septated ventricle was quite homo-geneous with the compact myocardium forming about20% of the wall cross-section and the trabeculae about57%. An apicobasal gradient in the proportion of thecompact myocardium soon became established in theleft ventricle accompanied by a decreased contribu-tion from the intertrabecular spaces. This gradientbecame more pronounced during subsequent devel-opment; at HH 29 the compact myocardium formedalmost 80% of the cross section of the basal region ofthe left ventricle. The wall composition was more ho-mogenous in the right ventricle: the compact myocar-dium comprising 15 to 25% and trabeculae 40 to 60%.This did not change much during the period of study.The highest proportion of intertrabecular spaces wasrepeatedly seen in the middle slice. From the datacollected it was possible to determine the ratio be-tween compact and trabeculated myocardium. Withthe exception of the left ventricle’s basal part of HHstage 29, the trabeculated component was the greater.

The values of the local porosity of different re-gions reflected the changes in ventricular wall com-position. To make a developmental comparison, thevalues measured in the midportion slice were select-ed; the measurements of the compact layer thicknesswere also performed in this region. The small samplesize and relatively high variability prevented any ex-

tensive statistical analysis of this quantitative data. Inthe midportion of the left ventricle the porosity of theluminal layer increased from 30%, during stages ex-amined, to almost 60% by HH stage 29. The porosityof the basal layer was around 20% with a small de-cline from HH 24 to HH 29. Unlike that of the leftventricle, the porosity of the right ventricle luminallayer did not change during the examined period; itremained at around 38%. The porosity of the basallayer increased from 18 to 36%, converging with val-ues for the luminal layer. As with the values of com-pact layer thickness, there were no obvious differenc-es between HH stages 24 and 27.

tI. _.. l._“_ _...... . . . . _ . . . . _ .. ..I I.^.

/

mlcompact trabec 0 spaces 9 normal

-40

-20

5% _. ._.... ._..,.,. ....,,~_X- 7%

80 80

,, .~” ^_.,.__” . . “” ,_” .v+._____I1-x,,.lxx-““<^“--““-“--“-

-----j

*

compact trabecm spaces +?+ normal

Fig. 8. Ventricular wall composition of stage 29 hearts treated withverapamil. Note the increased proportion of trabeculae(normal values are shown by markers).


Verapamil-treated hearts showed a decreased pro-portion of compact myocardium and an increased pro-portion of trabeculae; heart development was retard-ed by about one day in comparison to non-treatedembryos (Fig. 8). However, the local porosity waswithin normal limits. A statistically significant reduc-tion of the compact layer’s thickness was observed instages 24 and 27 in both ventricles (Fig. 6).

DISCUSSION

Development of trabecular architectureThe first signs of trabeculation, of the initially smoothinner ventricular wall, can be seen by HH stage 15 or16 for the chick (Steding et al., 1982; Icardo & Fern-andez-Teran, 1987) and between 8 and 9 dpc for themouse (Challice & Viragh, 1973). The trabeculae in-crease in number rapidly, filling most of the ventricu-lar cavity within 24 hours (HH stage 21). Transfor-mation to the definitive pattern, where trabeculationbecomes limited to the wall as a result of ventricularexpansion and a free lumen is re-established, occursat approximately HH stage 34 (Sedmera et al., 1997),when the cardiac septation is completed. A qualita-tive description of the trabecular arrangement of thechick embryonic heart has been made in earlier stud-ies (Rychter & Rychterova, 198 1; Ben-Shachar et al.,1985; Icardo & Fernandez-Teran, 1987). Our data re-fine the description of the non-compact (porous) partof the ventricular wall (called “spongy” by Icardo &Fernandez-Teran, 1987, or “trabeculated” by Rychter-ova, 197 1) by dividing it into two sublayers: basaland luminal, as suggested by Rychterova (1982),which are characterised by precise morphological cri-teria. The finer basal parts of the trabeculae weresometimes overlooked in previous investigations dueto the use of suboptimal fixation. The presence of thebasal trabecular layer is temporary; it regresses dur-ing the development of coronary circulation (Rychter& Ostadal, 1971; Vrancken Peeters et al., 1997). Itsfunction is probably mainly nutritive as suggested byMinot (1901). It is consistent in its high degree ofisotropy and homogeneity; i.e. uniform size and theabsence of a preferred orientation of the intertrabecu-lar spaces. These characteristics are clearly demon-strated by the mean linear intercept diagram (Fig. 4).Using such a diagram, all the parameters of direction-al strength obtained from measurements on confocalimages using Denslow’s algorithms (Denslow et al.,1993) can be determined which makes the technique

used in this paper available for people without a strongmathematical background.

Rychterova (1971) made a quantitative study ofthe further development of the right ventriculartrabeculae between incubation days 6 and 14 and de-scribed the process of the gradual disappearance ofthe basal layer. There was an absolute reduction inthe number of trabeculae as established from frontalsections from 24-26 on day 6 (HH stage 29) to 19 onday 14 (HH stage 40) and a gradual thickening of theremaining trabeculae as well as the compact layer. Asan underlying mechanism it was suggested that therewere variations in the orientation of mitoses in thedifferent regions of the trabeculae: random in the ba-sal portion adjacent to the compact layer, and parallelto trabecular axis in the luminal part. Our measure-ments of the local porosity ratio agree with these ob-servations. The decrease of porosity in the basal layerpreceeds the factual compaction in the left ventricle.In the right ventricle, where the process of compac-tion is delayed, we observed a diminution of the out-side-to-inside porosity gradient. Later it was impossi-ble to separate these two regions as the smallestintertrabecular spaces disappeared during the processof compaction and the remaining trabeculae thick-ened (Sedmera et al., 1997).

Methodological aspectsAs making conclusions about three dimensional myo-cardial properties relying solely on a single dissectionplane would be tricky, we make reference to dissec-tions performed in frontal (Steding et al., 1982; Ben-Shachar et al., 1985) and sagittal (Pexieder, 1978)planes as well as some additional dissections. Fromthis data it can be concluded that the three-dimen-sional orientation is radial, i.e. perpendicular to thecompact layer at any point. Differences between theleft and right trabecular patterns are in part caused bydifferent shapes of the embryonic ventricles, howev-er, the left ventricular luminal trabeculae are coarserand less branched, and the intertrabecular spaces larg-er and more oblong. This aspect is visible in bothdiagrams used for the evaluation of orientation. Polarhistograms are useful for samples with several differ-ent directions, while the mean linear intercept tech-nique is suitable for samples with one preferentialorientation; and can provide additional strength infor-mation. However, neither technique can distinguishbetween polarized features oriented in multiple direc-tions and unpolarized (circular) shapes giving both


cases an air of random distribution. Visual evaluationof a colour feature classification by orientation is anuseful complement.

Another problem is the possible distortion of theproportion of different layers in sections which arenot always perpendicular. This was sometimes viewedin the apical area. Supplementary information wasobtained from frontal dissections which showed theradial arrangement of the luminal trabeculae, as wellas the extent of the basal trabecular layer and thecircularity of its intertrabecular spaces. The shape ofthe embryonic chick heart is more cylindrical thanspherical, unlike in mammals, so use of the dissectionplane perpendicular to its long axis should lead tominimal error in comparison with other planes. Ourhistological series, not included in this study, confirmthe above, and the error in the estimation of compactlayer thickness is minor. Furthermore, for compara-tive measurements we have chosen the midportionarea where the section plane is perpendicular to theouter heart curvature.

The reported porosity values should be correctedfor section thickness (approximately 50 pm). This er-ror might be eliminated using semithin sections (ca. 1c,1111 thick) which can be considered to have zero thick-ness for this purpose. Such problem, which is still oneof the unresolved questions of stereology, does notoutweigh the advantages of using the method of “thickslices” (Jaygasinghe et al., 1994). It was shown byZhu et al. (1994), that the variations in the directionalproperties of porous biological materials such as can-cellous bone from one section plane to another areenormous, so the use of relatively thick “optical sec-tions” (or, more precisely, profiles) gives a more real-istic idea of existing patterns and properties. It issometimes difficult to define individual layers andintertrabecular spaces on histological sections, and assuch time-consuming three-dimensional reconstruc-tions (Blausen et al., 1990) and/or confocal microsco-py (Usson et al., 1994) are required. We have shownthat consistent results can be achieved with the tech-niques used in this paper. These methods were used indifferent stages and locations, giving data that wereconsistent and comparable allowing observed gradi-ents to be reproduced. It should be realized, however,that trabeculations are like fingerprints: there is acommon building plan, but each heart has its ownunique, precise pattern of trabeculation similar to thesituation described in human mitral valves by Solo-mon and Nayak (1994).

Functional and comparative aspectsWe agree with Challice and Viragh (1973), andWenink et al. (1996) that the extensive trabeculationwith a radial three-dimensional arrangement is im-portant for the pumping function of the embryonicheart. According to our results, the trabeculae formmost of the myocardial mass during early stages. Val-ues between 64 and 84% were reported for early hu-man embryos using three dimensional reconstructionsfrom histological series by Blausen et al. (1990). Thedifferences between left and right trabecular patternsin humans were studied quantitatively by Wenink(1992), and it was found that early trabeculae in theright ventricle are finer than those of the left ventri-cle; this is in agreement with our observations of em-bryonic chicks. The cells in the trabeculae are moredifferentiated, containing more myofibrils that arebetter aligned than the cells in the compact zone(Markwald, 1969; Challice & Viragh, 1973). Themyocytes in the conotruncus show a high degree ofdifferentiation and circumferential alignment whichis probably important for its sphincter-like function(Tokuyasu, 1990).

The recently produced mice mutants with a disrup-tion of genes for neuregulin or its receptors (reviewedby Marchionni, 1995) lack most or all ventriculartrabeculation and die in utero at gestation day 10.5.The probable cause of death was heart failure as anirregular heartbeat and atria1 dilatation were observed.These results give further evidence for the importanceof trabeculation in the proper pumping function of theembryonic heart; these mice provide a model in whichthe effects of the absence of trabeculation on contrac-tile efficiency can be studied.

During development the proportion of the compactlayer becomes more substantial, especially in the leftventricle. This suggests that from HH stage 29 on-wards, the left ventricle relies mostly on this layer asimplied by its increasing thickness. This seems to cor-respond with the results of a comparative study per-formed by Ostadal and Schiebler (197 1) on fish heartsof different species. Rather than finding a correlationwith the phylogenetic position, the proportion of com-pact myocardium increased parallel to body size (andblood pressure) suggesting that the increased propor-tion of the compact layer is a way to improve heartperformance. Greer Walker et al. (1985) showed thatproportion of the compact myocardium is dependenton the animal’s life style; more active fish have ahigher proportion of the compact myocardium. The


augmented proportion of the compact layer coincideswith the development of coronary circulation, whichstarts in the chick embryo after a certain thickening ofthe compact layer at HH stage 3 1. Details of this proc-ess and epicardial origin of the coronary endotheliumwere demonstrated by Poelmann et al. (1993) usingchick-quail chimeras. This thickening seems to becrucial for the proper pumping function of the heart atlater stages as shown by embryonic lethality at gesta-tion day 12.5 of neuropilin-overexpressing mouseembryos which have a very fine compact myocardiallayer and problems with ventricular septation (Kitsu-kawa et al., 1995). Embryonic lethality was reportedin VCAM-1 knockout mice which display a reductionof the compact layer of the ventricular myocardiumand interventricular septum (Kwee et al., 1995).

Importance for perturbation studiesMeasurements of the compact layer serve as markersfor experimental procedures that alter the load condi-tions of the embryonic heart. From the decreased pro-portion of compact myocardium relative to trabeculaein verapamil-treated hearts one can draw the conclu-sion that a developmental increase of compact layerthickness is a mechanism for increasing pumping ef-ficiency. So the smaller number of cardiomyocytes ofnormal size and composition found in verapamil-in-duced decreased afterload (Clark et al., 1991) is alsoabnormally arranged. Less remarkable differencesseen at HH stage 29 could be caused by higher mor-tality of embryos with more severely affected heartsor to insufficiency of dose due to the rapid embryonicgrowth. However, the treated hearts showed no spe-cific malformations. Compact layer thickness meas-urements could be useful in interspecies comparisons.The developmental trends of compact layer thicknessare similar to those seen in the mouse heart betweengestation days 12 and 17 (Pham, 1997), but the in-crease in thickness of the compact layer of the leftventricle is not as marked as in the chick. This resem-bles the adult state, where the left/right ventricularwall thickness ratio is 51 in birds, but about 3:l inhuman or other mammals (Komarek et al., 1985; ourunpublished observations).

In conclusion, we have devised a simple, repro-ducible methodology to analyse embryonic ventricu-lar myocardial architecture. We use the above men-tioned methods to describe the chick embryonicventricle during organogenesis and in the experimen-tal model of decreased pressure load. It was shown

that left and right ventricle exhibit distinct morpho-genesis from the time of their appearance. The devel-opmental trends of parametres such as proportion ofthe compact myocardium reflect the adaptation of theheart to the gradually increasing demands of the grow-ing embryo. Myocardial growth and architecture areinfluenced by epigenetic factors. Thus, the in uteroenvironment is likely an important determinant of thefabric of the mature myocardium influencing longterm survival.

ACKNOWLEDGEMENTS

We would like to express our thanks to Mrs. Ariane Gerberfor her excellent technical assistance, Dr. Richard Hayden(Leica, England) for his advice about image analysis andDr. Penny Thomas (NHLI, London) for inspiring discus-sions and helpful criticism. Dr. Si Minh Pham is acknowl-edged for helpful discussions on mouse trabeculation de-velopment. Our further thanks belong to Mr. Didier Sevinfor his assistance in computing, to Ms. Tina Hernandez forhelping with the statistical analysis and Mr. Jay Legue forfinal language revision. This work was supported by grantsfrom the Swiss National Science Foundation 31-33889.92(T. Pexieder) and the National Heart, Lung and Blood Insti-tute (NHLBI) Specialized Center of Research in PediatricCardiovascular Diseases p50-HL51498 (E.B. Clark and N.Hu).

REFERENCES

Ben-Shachar G, Arcilla RA, Lucas RV, Manasek JF (1985):Ventricular trabeculations in the chick embryo heart andtheir contribution to ventricular and muscular septal de-velopment. Circ Res 57: 759-766.

Blausen BE, Johannes RS, Hutchins GM (1990): Compu-ter-based reconstructions of the cardiac ventricles of hu-man embryos. Am J Cardiovasc Path01 3: 37-43.

Broekhuisen ML, Bouman HG, Mast F, Mulder PG, Gitten-berger-De Groot AC (1995): Hemodynamic changes inHH stage 34 chick embryos after treatment with all-trans-retionoic acid. Pediat Res 38: 342-348.

Challice CE, Viragh S (1973): The architectural develop-ment of the early mammalian heart. Tissue Cell 6: 447-462.

Chuck ET, Freeman DM, Watanabe M, Rosenbaum DS(1997): Changing activation sequence in the embryonicchick heart: implications for the development of the His-Purkinje system. Circ Res 81: 470-476.

Clark EB, Hu N, Turner D, Litter JE, Hansen J (1991):Effect of chronic verapamil treatment on ventricularfunction and growth in chick embryos. Am J Physiol261: H166-H171.

Cowin SC (1986): Wolff’s law of the trabecular architec-


ture at remodeling equlibrium. J Biomech EnginlO8: 83-88.

De Jong F, Opthof T, Wilde AAM, Janse MJ, Charles R,Lamers WH, Moorman AFM (1992): Persisting zones ofslow impulse conduction in developing chicken hearts.Circ Res 71: 240-250.

De La Cruz MV, Castillo MM, Villavicencio L, ValenciaA, Moreno-Rodriguez RA ( 1997): Primitive interven-tricular septum, its primordium, and its contribution inthe definitive interventricular septum: in vivo labellingstudy in the chick embryo heart. Anat Ret 247: 5 12-520.

Denslow S, Zhang Z, Thompson RP, Lam CF (1993): Sta-tistically characterized features for directionality quanti-fication in patterns and textures. Pattern Recognition26: 1193-1205.

Greer Walker M, Santer RM, Benjamin M, Norman D(1985): Heart structure of some deep-sea fish (Teleostei,Macrouridae). J 2002 (London) 20.5: 75-89.

Hamburger V, Hamilton HL (195 1): A series of normalstages in the development of the chick embryo. J Mor-phol88: 49-92.

Hogers B, De Ruiter MC, Baasten AMJ, Gittenberger-DeGroot AC, Poelmann RE (1995): Intracardiac flow pat-terns related to the yolk sac circulation of the chick em-bryo. Circ Res 76: 871-877.

Hu N, Keller BB (1995): Relationship of simultaneous atri-al and ventricular pressures in stage 16-27 chick embry-os. Am J Physio1269: H1359-H1362.

Icardo JM, Fernandez-Teran A (1987): Morphologic studyof ventricular trabeculation in the embryonic chick heart.Acta Anat 130: 264-274.

Jaffee OC, Jaffee AL (1989): The effects of verapamil onthe developing cardiovascular system. Anat Ret 223:63A (Abstract).

Jayasinghe JAP, Jones SJ, Boyde A (1994): Three-dimen-sional photographic study of cancellous bone in humanfourth lumbar vertebrae. Anat Embryol 189: 259-274.

Jouk PS, Usson Y, Michalowicz G, Parazza F (1995): Map-ping of the myocardial cells by means of polarized lightand confocal scanning laser microscopy. Microscop ResTech 30: 480-490.

Kastner P, Grondona JM, Mark M, Gansmuller A, LemueurM, Decimo D, Vonesch J, Dolle P, Chambon P (1994):Genetic analysis of RXRa developmental function: Con-vergence of RXR and RAR signalling pathways in heartand eye morphogenesis. CeZZ 78: 987-1003.

Kitsukawa T, Shimono A, Kawakami A, Kordoh H, Fuji-sawa H (1995): Overexpression of a membrane protein,neuropilin, in chimeric mice causes anomalies in the car-diovascular system, nervous system and limbs. Develop-ment 121: 4309-4318.

Komarek V, Malinovsky L, Lemez L, eds. (1982): Anato-mia avium domesticarum et embryologia galli. Bratisla-va, Priroda.

Kwee L, Baldwin HS, Shen HM, Stewart CL, Buck C, BuckCA, Labow MA (1995): Defective development of theembryonic and extraembryonic circulatory systems invascular cell adhesion molecule (VCAM- 1) deficientmice. Development 121: 489-503.

Marchioni MA (1995): neu tack on neuregulin. Nature 378:334-335.

Markwald RR ( 1969): Ultrastructural and histochemicalstudy of cardiac histogenesis in the fetal hamster. AnatRet 163: 226 (Abstract).

McLean M, Ross MA, Prothero J (1989): Three-dimension-al reconstruction of the myofiber pattern in the fetal andneonatal mouse heart. Anat Ret 224: 392-406.

Minot CS (1901): On a hitherto unrecognised form of bloodcirculation without capillaries in the organes of Verte-brata. Proc Boston Sot Nat Hist 29: 185-215.

Moscoso G, Pexieder T (1990): Variations in microscopicanatomy and ultrastructure of human embryonic heartssubjected to three different modes of fixation. Path ResPratt 186: 768-774.

Ostadal B, Schiebler TH (197 1): Ueber die terminal Strom-bahn in Fischherzen. ZAnat Entwickl Gesch 134: lOl-110.

Ostadal B, Janatova T, Krause EG, Pelouch V, Dusek J(1987): Different effects of propranolol and verapamilon isoproterenol-induced changes in the chick embryon-ic heart. Physiol Bohemosl36: 301-3 11.

Pexieder T (1978): Development of the outflow tract of theembryonic heart. Birth Defects: Original Article Series14: 29-68.

Pexieder T (198 1): Prenatal development of the endocar-dium: a review. Scan Electron Microsc 2: 223-253.

Pexieder T, Rousseil MP, Prados-Frutos JC (1992): Prena-tal pathogenesis of the transposition of great arteries. In:Vogel M, Buhlmeyer K, eds., Transposition ofthe GreatArteries 25 years after Rashkind Baloon Septostomy.Darmstadt, Steinkopf Verlag, pp. 1 l-27.

Pham SM (1997): Developpement de l’architecture myo-cardiaque de la souris. MD thesis, University ofLausanne.

Poelmann RE, Gittenberger-De Groot A, Mentink MMT,Bokenkamp R, Hogers B (1993): Development of thecardiac coronary vascular endothelium, studied with an-tiendothelial antibodies, in chicken-quail chimeras. CircRes 73: 559-568.

Rychter Z, Ostadal B (197 1): Fate of “sinusoidal” inter-trabecular spaces of the cardiac wall after developmentof the coronary vascular bed in chick embryo. FoliaMorphol (Praha) 19: 31-44.

Rychter Z, Rychterova V (198 1): Angio- and myoarchitec-ture of the heart wall under normal and experimentallychanged conditions. In: Pexieder T, ed., Perspectives incardiovascular research, Vol 5, Mechanisms of cardiacmorphogenesis and teratogenesis. New York, RavenPress, pp. 43 l-452.

Rychterova V (1971): Principle of growth in thickness ofthe heart ventricular wall in the chick embryo. FoliaMorph01 (Praha) 19: 262-272.

Rychterova V (1982): Analysis of model of the left hearthypoplasia. PhD thesis, Charles University, Prague.

Sanchez-Quintana D, Garcia-Martinez V, Climent V, HurleJM (1995): Morphological changes in the normal patternof ventricular myoarchitecture in the developing humanheart. Anat Ret 243: 483-495.


Sedmera D, Pexieder T (1995): SEM and image analysis inquantitative study of chick ventricular myoarchitecture.Biology of the Cell 84: 227 (Abstract).

Sedmera D, Pexieder T, Hu N, Clark EB (1997): Develop-mental changes in the myocardial architecture of thechick. Anat Ret 248: 421-432.

Solomon V, Nayak VM (1994): Definition and function ofcommisures, slits and scallops of the mitral valve: Anal-ysis in 100 hearts. Asia Pacific J Thorac CardiovascSurg 2: 10-16.

Steding G, Seidl W, Rahlf G (1982): Rasterelektronen-mikroskopische Untersuchungen zur Entwicklung desendocardialen Reliefs der Herzkammern. Verh Anat Ges76: 199-201.

Streeter DD Jr (1979): Gross morphology and fiber geome-try of the heart. In: Berne RM, Sperelakis N, Geiger S,eds., Handbook of physiology - Section 2, The cardio-vascular system. Am Physiol Sot (Bethesda), pp. 61-112.

Sucov HM, Dyson E, Gumeringer CL, Price J, Chien KR,Evans RM (1994): RXRa mutant mice establish a genet-ic basis for vitamin A signaling in heart morphogenesis.Gen Develop 8: 1007-1018.

Taber LA, Hu N, Pexieder T, Clark EB, Keller BB (1993):Residual strain of the stage 16-24 chick embryo. CircRes 72: 455-462.

Tokuyasu KT (1990): Co-development of embryonic myo-cardium and myocardial circulation In: Clark EB, TakaoA, eds., Developmental Cardiology: Morphogenesis andFunction. New York, Futura Publishing, pp. 205-218.

Usson Y, Parazza F, Jouk PS, Michalovicz G (1994): Meth-od for the study of the three-dimensional orientation ofthe nuclei of myocardial cells in fetal human heart bymeans of confocal scanning laser microscopy. JMicros-copy 174: 101-110.

Vrancken Peeters MPFM, Gittenberger-De Groot AC,Mentink MMT, Hungerford JE, Little CD, Poelmann RE(1997): The development of the coronary vessels andtheir differentiation into arteries and veins in the embry-onic quail heart. Dev Dyn 208: 338-348.

Wenink ACG (1992): Quantitative morphology of the em-bryonic heart: an approach to development of the atriov-entricular valves. Anat Ret 234: 129-135.

Wenink ACG, Knaapen MWM, Vrolijk BCM, Van Gronin-gen JP (1996): Development of myocardial fiber organi-zation in the rat heart. Anat Embryo1 193: 559-567.

Zhu M, Keller TS, Moeljaslo E, Spengler DM (1994): Mul-tiplanar variations in the structural characteristics of can-cellous bone. Bone 1.5: 251-259.

Received: May 1997Accepted after revision: December 24, 1997

a quantitative study of the ventricular myoarchitecture in europ j morph.chick... · its acute...

Documents