OUP user menu

Extraembryonic venous obstructions lead to cardiovascular malformations and can be embryolethal

Bianca Hogers, Marco C. DeRuiter, Adriana C. Gittenberger-de Groot, Robert E. Poelmann
DOI: http://dx.doi.org/10.1016/S0008-6363(98)00218-1 87-99 First published online: 1 January 1999


Objective: To expand our knowledge concerning the effect of placental blood flow on human heart development, we used an embryonic chicken model in which extraembryonic blood flow was manipulated. Methods: First, one of the three major vitelline veins was ligated, while blood flow was visualized with Indian ink. In this way, we could study the effect of different ligation positions on intracardiac flow patterns. Secondly, these vitelline veins were ligated permanently with a microclip until cardiac septation was completed, thereafter, the hearts were morphologically evaluated. In this way, we could study the impact of the ligation position on the severity and frequency of heart malformations. On combining the results, we were able to study the effect of different intracardiac flow patterns on heart development. Results: Although ligation of each vein resulted in different intracardiac flow patterns, long-term ligation resulted in similar cardiovascular malformations in survivors. These consisted mainly of ventricular septum defects (VSDs), semilunar valve anomalies, and pharyngeal arch artery malformations. There was no significant difference (p>0.05) between the ligation position and the incidence of cardiovascular malformations. However, the percentage mortality after clipping the left lateral vitelline vein was significantly higher (p<0.05) than after ligation of either the right lateral or posterior vitelline vein. Conclusions: Early extraembryonic venous obstruction leads to altered flow patterns, which probably result in shear stress changes. In postseptation stages, these result in a spectrum of cardiovascular malformations irrespective of the ligation position. A diminished incidence of VSDs in the oldest stage was attributed to delayed closure of the interventricular foramen.

  • Blood flow
  • Heart development
  • VSD
  • Chick
  • Embryo

Time for primary review 28 days.

1 Introduction

The development of the heart is a very complex process, as the heart has to be functional from the earliest stages, being a muscle-wrapped tube, to a four chambered heart with separate blood flow. Abnormal genetic information is probably responsible for most disturbed developmental mechanisms during the earliest stages of cardiovascular development, while the influence of environmental factors becomes more important after placental circulation has started. Well known environmental factors are radiation, viruses, maternal diabetes or teratogens like thalidomide and retinoic acid. We were interested in the role of blood flow, not as a transportation fluid but as a physical factor in normal and abnormal heart development.

From the literature, it is known that combined transvaginal and transabdominal Doppler ultrasonography demonstrated fetal, cardiac and arterial flow velocity waveforms as early as 8–12 weeks of gestation [1, 2]. Growth retardation is accompanied by changed arterial, atrioventricular (AV) and venous flow velocity waveforms [3]and increased uteroplacental vascular resistance [4]compared to normal age-matched control fetuses, indicating a relation between hemodynamics and growth retardation. However, to date, it is not possible to distinguish any causal relation between abnormal placental blood flow, growth retardation and cardiovascular malformations.

The essential phases in human cardiovascular development take place between three and eight weeks of gestation [5], which is too early for echocardiographic evaluation of the interaction between hemodynamics and cardiac development. To expand our knowledge of the effect of placental blood flow on human heart development, we used an embryonic chicken model in which extraembryonic blood flow could be both visualized and manipulated. In this way, we could follow the changes in flow pattern that were introduced by obstructing part of the extraembryonic circulation. Former studies have already demonstrated the presence of stable intracardiac blood flow patterns in the earliest developmental stages by blood flow visualization with dye indicators [6, 7]and suggested a role for flow patterns in normal heart development.

Although many mechanical hemodynamic obstruction models resulted in abnormal heart development [8–10], the change in blood flow was rarely studied. Rychter and Lemez [11]very accurately followed the change in blood flow through the pharyngeal arch arteries after transection or ligation of several vitelline veins simultaneously, 4 h to three days after the manipulation, thereby explaining left eye abnormalities observed earlier by Orts Llorca et al. [12], resulting from the same type of experiments. Changes in intracardiac flow patterns, or the resulting cardiac malformations, have not been studied. In an earlier study [13], we were able to relate intracardiac malformations to extraembryonic venous obstruction by visualization of intracardiac flow patterns directly after ligation. However, a very important question remained unanswered, namely, whether the severity or frequency of observed cardiac malformations was determined by the ligation itself or by the position of the obstruction. Therefore, we performed experiments in which the visualized intracardiac flow patterns were used to reveal direct changes in flow pattern after ligation of the different vitelline veins. This was followed by morphologic analysis of cardiovascular malformations after permanent ligation. In this way, we could study the impact of ligation site on the quality of heart development and the frequency of malformations. Combining the results, we were able to study the effect of different intracardiac flow patterns on heart development.

2 Methods

Fertilized White Leghorn eggs (Gallus domesticus) were incubated at 37°C and 60–70% relative humidity. Embryos were staged according to the age-determination criteria of Hamburger and Hamilton [14]and handled in accordance with the Guide for the Care and Use of Laboratory Animals, published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1985).

2.1 Procedure for Indian ink injections during short term ligation

In a pilot study, we investigated whether removal of the embryo from the egg, leaving the yolk sac circulation intact could change the normal flow patterns. We checked heart rate, cardiac performance by video recording and made identical injections at several stages. As there was no difference between the in-ovo and ex-ovo situation, we preferred the ex-ovo set-up as visual contrast was much better [7].

Chick embryos were carefully removed from their yolks and transferred to an agar-covered plastic Petri dish, immersed in Locke solution (0.94% NaCl, 0.0045% KCl and 0.004% CaCl2) at 37°C. The yolk sac was flattened and adhered to the agar. A pair of forceps was connected to a micromanipulator and positioned around one of the vitelline veins at the indicated sites (Fig. 1). The extraembryonic vasculature was injected by means of a glass needle (tip diameter, 2 to 6 μm) with controlled administration of minute amounts of Indian ink (1:5, v/v, diluted in Locke solution). The embryo was turned to the right to reveal a ventral view of the heart. When performed carefully, no alterations of the intracardiac flow patterns were observed. The route of Indian ink through the heart was recorded with a video camera (Sony, DXC-151P) and a U-matic videocassette recorder (Sony, VO-5630). After registration of the normal intracardiac route, the forceps were closed and the newly acquired intracardiac route was recorded. Each embryo was injected only once. The injection volume was estimated at 0.002 μl, which is less than the volume measured to be effective in raising the stroke volume index in stage 18 embryos [15]. Only experiments with a complete ligation of the vitelline vein that demonstrated a clear intracardiac flow pattern both before and after ligation were included. A total of 67 embryos were used (23 embryos for ligation of the right lateral vitelline vein, 24 embryos for ligation of the left lateral vitelline vein and 20 embryos for ligation of the posterior vitelline vein).

Fig. 1

Schematic representation of the vitelline veins of a stage 17 chicken embryo. The defined yolk sac regions are indicated by the broken lines. Indian ink injections were made into a capillary or small venule within each of these yolk sac regions. The insert shows a magnification of the three ligation sites (asterisks) at the right lateral (rl), the left lateral (ll) and the posterior (p) vitelline vein, respectively. The capillary plexus situated under the caudal trunk of the embryo (the latter indicated by a broken line) acts as a short cut after ligation. By rerouting the blood immediately to the other side, the embryo is not deprived of blood and oxygen. Preferential flow remodels the plexus into a new vein. A, anterior yolk sac region; LL, left lateral yolk sac region; RL, right lateral yolk sac region; P, posterior yolk sac region; om, omphalomesenteric vein.

2.2 Ligation in ovo

Egg shells were cleaned with ethanol (70%) and windowed. The eggs were kept warm in isolating foil. All manipulations were performed using a dissecting microscope. Above the intended ligation site (Fig. 1), the vitelline membranes were removed and a small incision was made with a tungsten needle in the yolk sac membrane, adjacent to the vitelline vein. A microclip, devised from a nickel transmission electron microscopy carrier grid, was clamped around the vein (Fig. 2). The cessation of blood flow downstream from the microclip was confirmed. Eggs were sealed with Scotch tape and returned to the incubator. Shams (n=20), in which all procedures were similar except for the ligation with the microclip, and normal eggs (n=10) served as controls. The eggs were reincubated and the embryos were examined from embryonic day 5 (stage 27) until hatching (stage 45). In total, 125 embryos were ligated right laterally, 57 embryos left laterally and 51 embryos posteriorly. The embryos as well as the hearts were macroscopically evaluated, prior to fixation in a mixture of 2% glacial acetic acid in 100% ethanol at 4°C. After paraffin embedding, the embryos were serially sectioned at 5 μm and stained with hematoxylin–eosin for light microscopical study of the hearts and pharyngeal arch arteries. Embryos were evaluated as normal or abnormal, presenting with a subarterial versus subaortic ventricular septal defect (VSD) or other malformations. Moreover, abnormal embryos were further characterized by additional cardiac malformations, such as abnormal semilunar valves, enlarged right ventricle, abnormal atrioventricular (AV) valves and partially absent epicardium. We also encountered pharyngeal arch artery malformations, such as hypoplastic right brachiocephalic artery, persistent right ductus caroticus, interruption of the aortic arch, double aortic arch and hypoplastic right pulmonary artery. In addition, ten ligated embryos (stage 34) were prepared for scanning electron microscopy to acquire a 3D-illustration of the intracardiac malformations. Half strength Karnovsky's [16]perfusion fixed hearts were opened transversally with iridectomy scissors. They were rinsed in 0.1 M sodium-cacodylate buffer (pH 7.2) and postfixed for 2 h at 4°C in 1% OsO4 in the same buffer, followed by dehydration in graded ethanol solutions. The preparations were critical-point dried over CO2 by conventional methods, sputter-coated with gold for 3 min (Balzers MED 010) and studied in the scanning electron microscope (Philips SEM 525M).

Fig. 2

Stage 19 chick embryo in ovo with microclip. Six hours after ligation of the right lateral vitelline vein (RL), a newly developed vein is visible (arrow), which is connected to the left side via the posterior vitelline vein (P). The original vessel segment (arrowhead) is in regression. DAO, dorsal aorta; E, eye; H, heart.

2.3 Statistical analysis

Comparison of mortality percentages was performed using the Chi-square test, comparing the mortality percentages in the three groups induced by the classification of the veins. All further analyses pertain to survivors. The risk of an ‘abnormal outcome’ was estimated using a logistic regression approach. The basic question was whether or not the null-hypothesis of ‘equal risks’ in the three subgroups could be rejected. The veins were classified as three subgroups and were entered into the regression analysis as a factor, while the gestational stage was used as a (continuous) covariate (using the exact, i.e. non-recoded values). First, a test on the interaction of ‘type-of-vein’ by ‘stage’ was performed to test whether or not the supposed differences between the veins were stage-dependent. If not, the model without the interaction term (stage) was recomputed and tested for the supposed difference between the veins for the main effect. The subclassification by ‘type-of-vein’ was assumed to be associated with the risk of detecting the abnormal outcome if either the interaction test or the main effect was significant at the 5% level (two-sided).

To analyze the risk of specific abnormalities (as a refined classification of the dichotomy ‘normal–abnormal'), the same approach was used, conditional on the presence of an abnormality, thus answering the question: ‘if an abnormality is present, is the risk of a specific abnormality associated with the type-of-vein?'

3 Results

3.1 Intracardiac flow patterns

Blood flow was visualized continuously during ligation. A shift in pattern was visible within two–four heartbeats, except for those cases in which the yolk sac region (Indian ink marked current) drained via the ligated vein. In these cases, the Indian ink upstream of the microclip was directly diverted via the capillary plexus underneath the embryo to one of the other main vitelline veins and then returned to the heart, which took about five–eight heartbeats. Because we did not observe any acceleration or change in heartbeat, we assumed that clipping did not alter the steady state.

Typical intracardiac blood flow patterns, as seen after injection of Indian ink into different yolk sac areas before as well as after ligation, are demonstrated by photographs from the VCR-recordings (Fig. 3). All blood flow data have been combined schematically in Fig. 4. Each colored line represents the mean intracardiac blood flow pattern for a certain yolk sac region. All colored lines within one heart represent the mean total extraembryonic blood flow for the concerning situations, like normal and ligation of the three vitelline veins. We compared normal flow patterns with those after ligation, by studying e.g. the green current in the normal (Fig. 4a) and in the ligated situation (Fig. 4b–d). Moreover, the combined general pattern (all colored lines), the distribution of streamlines in a certain cardiac segment and the direction of flow were compared to the normal situation.

In normal embryos (Fig. 4a), blood from the anterior yolk sac region (green) followed the inner curvature of the heart. Blood from the lateral yolk sac regions (blue, pink) followed a central course, while blood from the posterior yolk sac region could follow two courses. One course ran centrally in the primitive ventricle and continued along the outer curvature of the conotruncus (yellow). The other course was shown to be a double one, splitting just before the still undivided AV canal. This resulted in a central and in an outer course through the ventricle, recombining in the conotruncus to a single streamline in the outer curvature (red). The transverse section through the conotruncus showed two preferential currents in this plane, namely a ventral and a dorsal one.

Ligation of the right lateral vitelline vein (Fig. 4b) caused a small change of pattern in the AV canal. Blood from posterior (yellow, red) coursed slightly more centrally, leading to a small reduction of streamlines through the left half of the AV canal. Moreover, remarkable changes were observed in the overall pattern within the conotruncus. All streamlines were located along the lateral-most half of the conotruncus. In particular, blood from anterior (green) had shifted considerably. In the transverse section, it was shown that all streamlines coursed almost ventrally after ligation. Blood from anterior (green) and posterior (red) shifted from dorsally to centrally, or even ventrally.

Ligation of the left lateral vitelline vein forced blood from both posterior and left lateral yolk sac regions to take a detour (Fig. 1) and resulted in alterations both in the AV canal and in the inflow portion of the ventricle (Fig. 4c). Compared to normal, there were more streamlines through the left part of the AV canal, which is the future mitral orifice, and there was a change in direction of the streamlines. Blood from right lateral (pink) and from posterior (red) crossed within the AV canal from the inner curvature to the outer curvature and vice versa (from a central position to the inner curvature). This resulted in an additional streamline coursing along the outer curvature of the ventricle, directly downstream from the left part of the AV canal. In the conotruncus, the general pattern was similar to normal, although individual streamlines have changed position. The transverse section of the conotruncus showed that most streamlines were running ventrally after left lateral ligation, which was caused by a displacement of anterior (green) and posterior (red) blood from dorsal to ventral. Blood from right lateral (pink) was located dorsally, which made the overall pattern in the transverse plane roughly comparable to normal.

Ligation of the posterior vitelline vein resulted in changes within the AV canal and the ventricle (Fig. 4d). Most of the visualized bloodstreams were running centrally through the AV canal. There was a decrease in the number of streamlines through the right part of the AV canal, which is the future tricuspid orifice. All streamlines derived from extraembryonic sources were concentrated centrally in the ventricle. In the conotruncus, a more or less even distribution of visualized streamlines was observed, but with less flow through the inner curvature. The transverse section showed that all streamlines were equally distributed through the conotruncus, with the exception of the dorsomedial part.

3.2 Permanent ligation in ovo

3.2.1 Mortality

Mortality percentages, as determined for the three groups (ligated vitelline vein), were significantly different (p<0.05). It turned out that the remarkably high mortality of 49% (28/57) after clipping the left lateral vitelline vein caused this difference. Ligation of the right lateral vitelline vein resulted in a mortality of 22% (30/136) and ligation of the posterior vitelline vein resulted in a mortality of 20% (10/51).

Fig. 3

Photographs from video recordings of Indian ink injections before (a,c,e) and after (b,d,f) ligation. (a–d) Ventral view of stage 17 chicken hearts and (e–f) right lateral view of the conotruncus and ventricle. (a) Indian ink injected into the anterior yolk sac region runs through the inner curvature of the ventricle (V) and the conotruncus (CT). (b) Ligation of the left lateral vitelline vein results in a deviation of the streamline to the outer wall of the proximal ventricular outflow tract (arrow). (c) Indian ink injected into the posterior yolk sac region runs centrally through the ventricle and through the outer curvature of the conotruncus. (d) Ligation of the left lateral vitelline vein results in a remarkable outward shift in the atrium (arrow), while a shift from the outer to the inner curvature is observed in the conotruncus (two picture frames ahead). (e) Indian ink injected into the posterior yolk sac region runs dorsally (d) through the conotruncus (arrow). The transition from conotruncal cushion to ventricular lumen is indicated by the small arrow. (f) Ligation of the right lateral vitelline vein causes a shift to ventral (v), resulting in a central position in the conotruncus. E, eye; Hd, head; DAO, dorsal aorta.

Fig. 4

Combined schematic representation of intracardiac flow patterns (ICFP) after ligation of the different vitelline veins. Each colored line represents the average flow pattern of one particular yolk sac region. (a) Normal ICFPs (color compilation of Fig. 3a from Hogers et al. [13]). (b) ICFPs observed after ligation of the right lateral vitelline vein (color compilation of Fig. 3b from Hogers et al. [13]). There is a reduction of streamlines through the left half of the AV canal (red, yellow) and, in the conotruncus, all streamlines were now located along the lateral-most half and also more ventrally. (c) ICFPs observed after ligation of the left lateral vitelline vein. By crossing from the inner curvature to the outer curvature, the pink streamline is added to the left half of the AV canal. Both red streamlines are sharply edged in the left atrium, ventricular inflow and outflow. (d) ICFPs observed after ligation of the posterior vitelline vein. Most streamlines were running centrally through the AV canal, with hardly any flow along the outer curvature of the ventricle and the inner curvature of the conotruncus. Green=anterior yolk sac; pink=right lateral yolk sac; blue=left lateral yolk sac; yellow=posterior yolk sac, single streamline; red=posterior yolk sac, double streamline. A, atrium; AV, atrioventricular canal; CT, conotruncus; D, dorsal; V, ventral; VE, ventricle.

Fig. 5

Frequency distribution of the number of surviving embryos, evaluated after ligation of the three vitelline veins.

3.2.2 Survivors

Fig. 5 shows the frequency distribution of evaluated (survived) embryos after ligation of the different vitelline veins. Permanent ligation of the different vitelline veins with a microclip resulted in anomalies of the heart and pharyngeal arch arteries in 79% of the survivors (139/177). Although solitary VSDs and solitary pharyngeal arch artery malformations were observed, most embryos displayed a VSD in combination with at least one pharyngeal arch artery malformation (Table 1). During septation, there was a decrease in the number of VSDs (with or without pharyngeal arch artery malformations).

Both shams and control embryos were macroscopically and microscopically normal, except for one sham embryo, which had a small subaortic VSD.

3.2.3 Spectrum of intracardiac malformations

VSDs were classified as either subarterial or subaortic. The most distal border of the subarterial VSD is at the semilunar valve level and is therefore doubly committed (Fig. 6c). These VSDs are characterized by the absence of a muscular outflow tract septum combined with a ventral displacement of the condensed mesenchyme of the aorticopulmonary septum. The most distal border of the subaortic VSDs is below the semilunar valve level (Fig. 6b). These VSDs are characterized by a muscular outflow tract septum, sometimes still present with a fine central band of mesenchyme. Note that the subarterial VSDs were earlier classified by us as a severe form with high VSD within a spectrum of subaortic VSDs [13]. Only nomenclature has changed, with morphologic criteria being identical.

View this table:
Table 1

Cardiovascular malformations after long-term ligation of the different vitelline veins in post-septation stages

Post-septation stagesEarlyLate
Ligated veinRLPRLP
Survivors (n)561916501125
Normal (%)02112402736
solitary PAA(%)14161914032
Miscellaneous (%)2150184616
  • Number of embryos that survived for each ligated vein.

    The malformations are expressed as predicted mean percentages during early post-septation stages (stages 27–35) and during late post-septation stages (stage 36–45), as regression analysis showed no differences.

    L, lateral vitelline vein; P, posterior vitelline vein; R, right lateral vitelline vein; PAA, pharyngeal arch artery malformation; VSD, ventricular septal defect; Miscellaneous, other cardiovascular malformations, such as semilunar valve anomaly, enlarged right ventricle, abnormal atrioventricular valves or partially absent epicardium.

Fig. 6

Scanning electron micrographs of transversely opened hearts of stage 34 chick embryos. The view is from apex to cranial. (a) Normal heart. The muscular outflow tract septum (asterisk) is continuous with the interventricular septum (white star). (b) Heart of a ligated embryo (right lateral vitelline vein) with a small subaortic VSD (arrow). Note that the outflow tract septum (asterisk) is beneath orifice level. The insert shows a variation with a large subaortic VSD. (c) Heart of a ligated embryo (right lateral vitelline vein) with a subarterial VSD. Due to the absence of the muscular outflow tract septum, the arterial ostia are separated by mesenchyme only. The common valve leaflets of the aortic and pulmonary semilunar valves are indicated by the arrowhead. A, aorta; LV, left ventricular outflow tract; M, mitral valve; P, pulmonary artery; RV, right ventricular outflow tract; T, tricuspid valve. Bar=1 mm.

3.2.4 Pharyngeal arch artery malformations

The pharyngeal arch system of the chick develops into two brachiocephalic arteries, a right-sided aortic arch and two ductus arteriosi (Fig. 7a). Permanent ligation resulted in a restricted pattern of affected pharyngeal arch arteries (Table 2). As the embryos were sectioned transversely at the level of the arterial pole, the diameters of the great arteries could be compared visually to that of normal controls and to each other (Fig. 8). In the early post-septation stages, many pharyngeal arch artery malformations were observed in combination with a VSD, in contrast to the late post-septation stages where more solitary pharyngeal arch artery malformations were observed (Table 1). Anomalies of derivatives of the third pharyngeal arch artery consisted of a hypoplastic right brachiocephalic artery (Fig. 7b, Fig. 8b–c) and a persistent right ductus caroticus (Fig. 7c). Anomalies of derivatives of the fourth pharyngeal arch artery consisted of an interruption of the aortic arch (either by the absence of the dorsal aorta between the fourth and sixth pharyngeal arch artery (Fig. 7d), or by obliteration of the entire fourth pharyngeal arch artery (Fig. 7e, 8b) and a double aortic arch (by persistence of the left fourth pharyngeal arch artery (Fig. 7f). Anomalies of the sixth pharyngeal arch artery derivatives only consisted of reduced vessel size of the right pulmonary artery and the ductus arteriosus (Fig. 7g, Fig. 8b).

Fig. 7

Schematic representation of anomalies of pharyngeal arch artery derivatives. (a) Normal arteries, as derived from the pharyngeal arch arteries (PAA) (indicated at the left side in Roman numerals). Note that the chick has a right-sided aortic arch (AA), two brachiocephalic arteries and two ductus arteriosi. (b) Hypoplastic right brachiocephalic artery (HRBC). (c) Persistence of the part of the dorsal aorta between the third and fourth PAA results in persistent right ductus caroticus (PRDC). (d) Interruption of the aortic arch (IAA) due to disappearance of the dorsal aorta between the fourth and sixth PAA. (e) Absent aortic arch (AAA) due to the disappearance of both the complete fourth PAA and the dorsal aorta between the fourth and sixth PAA. (f) Double aortic arch due to persistence of the left fourth PAA. (g) Hypoplastic right sixth pharyngeal arch artery (HRSA). A, aorta; DA, dorsal aorta; LAA, left aortic arch; L+RBC, left and right brachiocephalic artery; L+RD, left and right ductus arteriosus; L+RDA, left and right dorsal aorta; L+REC, left and right external carotid artery; L+RIC, left and right internal carotic artery; L+RPA, left and right pulmonary artery; L+RSC, left and right subclavian artery; PT, pulmonary trunk.

3.2.5 Atrioventricular valve anomalies

Although the intracardiac malformations mainly consisted of conotruncal malformations, we observed some minor AV anomalies, which always occurred in combination with a VSD. Due to impaired wedging of the arterial trunk between the atrioventricular orifices, the tricuspid orifice was sometimes situated dorsally in relation to the aorta instead of to the right of the aorta. Although all hearts showed normally fused AV cushions, relatively immature AV valves were observed for the stages concerned. The number of embryos with AV malformations in each vitelline vein subgroup was too small for statistical analysis.

View this table:
Table 2

Malformations of the derivatives of the pharyngeal arch artery in relation to ligated vitelline veins

 Ligated veinRLPRLP
 Number of embryos with PAA36131016111
IIIHypoplastic right brachiocephalic artery1664503
 Persistent ductus caroticus1834202
IVInterruption aortic arch331302
 Double aortic arch611100
VIHypoplastic pulmonary artery854512
 Miscellaneous PAA1013807
  • For every pharyngeal arch artery (Roman numerals), the number of embryos with an anomaly of the concerning derivative is indicated.

    L and R, left and right lateral vitelline vein; P, posterior vitelline vein; PAA, pharyngeal arch artery malformations; Miscellaneous PAAs, very proximal branching of the subclavian arteries and thin-walled arteries.

Fig. 8

Transverse sections at the pharyngeal arch artery level of stage 34 embryos. (a) Normal arterial pole. Note that chick embryos have a right aortic arch (A) and two brachiocephalic arteries (magnification: 61×). (b) Absent aortic arch due to left lateral vitelline vein ligation. In addition, both the right brachiocephalic artery (RB) and the right pulmonary artery (RP) are smaller then the left ones (50×). (c) Hypoplastic right brachiocephalic artery due to left lateral vitelline vein ligation (40×). LB, left brachiocephalic artery; LP, left pulmonary artery.

3.2.6 Semilunar valve anomalies

Semilunar valve abnormalities were common in the affected hearts. The total number of embryos with semilunar valve anomalies during early post-septation stages is 21/56 for right lateral ligation, 7/19 for left lateral ligation and 8/16 for posterior ligation. The number of embryos with semilunar valve anomalies during late post-septation stages is 17/50 for right lateral ligation, 4/11 for left lateral ligation and 11/25 for posterior ligation (not shown). When the embryos were subdivided for the specific semilunar valve malformations, the numbers became too small for statistical analysis, allowing only trends on severity and variability to be indicated. We observed the more severe types of semilunar valve defects, such as bicuspid aortic and quadricuspid pulmonary valves as well as common valve leaflets, after ligation of the right lateral vitelline vein. Ligation of the left vitelline vein affected more specifically the pulmonary valve leaflets, and ligation of the posterior vitelline vein resulted more often in less severe types, such as additional leaflets and fused commissures of the facing valve leaflets of the aorta.

3.2.7 Statistical analysis

Logistic regression analysis demonstrated that possible differences between the ligated veins were not dependent on the stage of evaluation. Therefore, it was justified to adjust for the stage itself rather than analyze for different stages separately. The overall test (normal vs. abnormal) showed no significant difference when the different veins were compared (p=0.97). However, there was a significant decrease in the percentage of abnormal embryos with increasing age, which is illustrated in Fig. 9. Since there were also no significant differences between the three vitelline veins in the occurrence of each individually subclassified malformation (specific secondary tests), we abstained from any particular correction for multiple testing.

Fig. 9

Percentages of abnormal embryos after ligation of the different vitelline veins. The results of the logistic regression analysis are depicted, showing no differences between ligation sites. The percentage of abnormal embryos is expressed as an estimated probability for an embryo of a particular stage to be abnormal after ligation of either of the three vitelline veins. The decrease in the percentage of abnormal embryos with increasing stage is statistically significant. LL, left lateral vitelline vein; P, posterior vitelline vein; RL, right lateral vitelline vein.

As it was meaningless to continue to distinguish the two morphologically different types of VSD, we combined these data in Table 1. To avoid any suggestion of differences between observed numbers, the percentages of each malformation during the early and late post-septation stages were expressed as predicted percentages. These were calculated from the regression analysis data and are depicted in Table 1.

In conclusion, the position of the microclip is not important for the frequency or severity of the cardiovascular malformations observed after venous clipping.

4 Discussion

The venous clip model is a suitable animal model for studying both direct and long-term effects of venous hemodynamic disturbances on heart development using the avian yolk sac circulation as an analogon to the human placental circulation. Intracardiac blood flow patterns before and after ligation were determined in each embryo, so they served as their own control. Although we only visualized intracardiac blood flow patterns derived from the extraembryonic circulation, we considered this adequate, as more than 80% of all blood flow in a stage 17 chick embryo is of extraembryonic origin [17]. Ligation of three different vitelline veins resulted in a disturbance of normal intracardiac blood flow patterns, with each vein giving a different pattern. Ligation of the right lateral vitelline vein mainly affected conotruncal flow patterns, and ligation of the posterior vitelline vein resulted in very regular, but changed, parallel streamlines running centrally through the heart. Ligation of the left lateral vitelline vein showed a major change in distribution of streamlines within the AV canal and resulted in the most irregular general pattern. As the majority of the left lateral ligated embryos died at around stage 24 (not shown), we assume that the above-mentioned irregular pattern is related to a phenotype that is incompatible with life, resulting in a significantly (p<0.05) higher mortality rate than ligation of the other veins.

Due to ligation, the different areas draining on separate vitelline veins became connected to each other using an existing capillary plexus present underneath the embryonic trunk. With Indian ink, we visualized (not shown) that blood immediately followed a detour via this plexus and that, within minutes, this plexus was remodeled into a new vessel (Fig. 2). Differences in mortality percentages cannot be explained by differences in areas to drain. The right and left vitelline vein have comparable areas to drain, whereas the posterior vitelline vein has a somewhat smaller area. However, the impact of left vitelline vein ligation is larger, due to anatomical properties. Left lateral ligation resulted in a detour of blood from both the left lateral and the posterior vitelline vein via the caudal plexus underneath the embryonic trunk. The emergence of a new vein on the other hand was as fast as after ligation of other vitelline veins.

By clipping a vitelline vein, we might have unintentionally changed the hemodynamics (blood flow velocity, blood pressure, cardiac output). However, we think that it is of minor importance for the value of our model whether or not the observed final phenotype is solely the result of the changed intracardiac flow patterns, as demonstrated, or is caused by an additional, still unknown, secondary factor. To unravel the underlying mechanisms, we have started a new series of experiments in which hemodynamic parameters are determined directly after clipping a vitelline vein.

The decrease in malformations between early and late post-septation stages was striking. This was possibly caused by an initial minor delay in closure of the interventricular septum, which gives an overestimation of VSDs at the early post-septation stages. As small VSDs were only observed early (not shown), it is likely that these were closed at the late post-septation stages. Spontaneous closure of the interventricular septum can also be observed in human prenatal clinical care. Due to early diagnostic tools, like Doppler echocardiography and color-coded Doppler, it is possible to diagnose VSDs at a very early stage [18]. Some of these early diagnosed VSDs are no longer present or are considerably reduced in size postnatally [19, 20], indicating spontaneous closure in the human fetus. There is also the possibility that the embryo is able to restore its function and/or morphology and survives [21].

The observed intracardiac malformations after venous clipping are considered to result from a combination of two cooperating, but disturbed, processes of cardiogenesis. The earliest event is impaired heart looping, resulting in improperly looped hearts in the postseptation stages, characterized by a graded severity in dextroposed aorta. Due to the dextroposed aorta, there is malalignment of an otherwise normally septated outflow tract in relation to the superior AV cushion, resulting in a spectrum of subaortic VSD. The Indian ink data showed, irrespective of the ligated vein, a shift of blood flow predominantly to the ventral side of the outflow tract. This could result in centrifugal forces that were responsible for improper looping, just like a meandering river. The second affected developmental process is the myocardialization of the outflow tract septum. Orifice septation is only brought about by mesenchyme due to a lack of myocardialization, resulting in a subarterial VSD directly below semilunar valve level and a ventral displacement of the condensed mesenchyme of the aorticopulmonary septum. The combined outcome of the above-mentioned processes can lead to a cardiac phenotype that varies with respect of the position and size of the VSD as well as in the degree of dextroposition of the aorta and in the extent of myocardialization of the outflow tract septum. All combinations have been encountered, e.g. subarterial VSD with dextroposed aorta, or subaortic VSD with mesenchymal outflow tract septum. The fact that we could not statistically distinguish the two morphologically distinct types of VSDs suggested that we are actually dealing with one continuous spectrum.

We never observed persistent truncus arteriosus (PTA), an outflow tract septation defect, which accounts for 73–100% of the abnormalities in neural crest ablated embryos [22, 23]. Recently, Männer et al. [24]extensively studied the influence of different microsurgical procedures used for neural crest ablation on cardiac outcome. Compared to almost 100% PTA of the survivors obtained by micro cautery [22, 23], mechanical cutting produced PTA in only 28% of the survivors. The remaining 72% of the survivors demonstrated a spectrum of cardiovascular malformations that was identical to the spectrum observed after our venous clipping. Thus, neural crest is pertinent for proper aorticopulmonary septation, while outflow tract septation is also directed by mechanical processes.

Morphologic analysis of semilunar valve malformation in relation to the different ligation positions revealed that severe semilunar valve defects, like bicuspid aortic and quadricuspid pulmonary valve and common valve leaflets, were mainly observed after ligation of the right lateral vitelline vein. This was in accordance with the Indian ink data, as normal intracardiac flow patterns within the conotruncus were severely disturbed after ligation of the right lateral vitelline vein. The highest number of embryos with semilunar valve malformations were observed after posterior vitelline vein ligation (p=0.19). However, these semilunar valve malformations consisted of less severe types. The regular intracardiac patterns observed within the conotruncus are probably close to normal, although individual streamlines have changed position. This might explain why severe semilunar valve anomalies were rarely encountered.

It was experimentally impossible to study differences in distribution of Indian ink before and after ligation through both the left and right pharyngeal arch arteries. Therefore, we were not able to relate the observed pharyngeal arch artery malformations to changes in blood flow in the individual pharyngeal arch artery. As the left fourth pharyngeal arch artery is destined to disappear, persistence of that vessel segment during the early post-septation stages, resulting in a double aortic arch, was most likely the result of a delay in normal development, as this malformation was no longer encountered during the late post-septation stages.

Although ligation itself resulted in cardiovascular malformations, we could not detect any significant difference in malformations observed after ligation of the different veins.

Icardo [25, 26]showed that clipping of the left lateral vitelline vein resulted in an additional area of endothelial alignment in the right atrium, abnormal AV cushions at day 5 [25]and, eventually, a double outlet right ventricle [26]. Apparently, the endothelial cells are able to sense differences in laminar flow, probably via changed shear stress, and they can respond to these changes by changing their orientation. Growth factors, such as transforming growth factor-β (TGF-β) and endothelin-1 (ET-1) are known to be expressed differently under changing shear-stress conditions [27, 28]and are involved in the regulation of proliferation and differentiation of myocardium and smooth muscle cells [29–31], and in epithelial–mesenchymal transformation in the AV cushions [32–34]. Changing endocardial orientation by ligating vitelline veins might disturb the precarious balance of growth factors released by the endocardium, thereby affecting cardiac morphogenesis. TGF-β2 and ET-1 null mutant mice display a similar cardiac phenotype [35–37]as that observed after venous clipping, which is suggestive of endothelial signaling being a mechanism in cardiac morphogenesis.

We still do not encompass how different intracardiac flow patterns can result into similar cardiovascular malformations. Although we have studied intracardiac flow patterns directly after ligation of different vitelline veins, we did not investigate endothelial cell orientation two days later, as Icardo [25]did. We cannot exclude the possibility that the different intracardiac flow patterns demonstrated by us resulted in equally changed orientation of endothelial cells, followed by equally disturbed endothelial signaling. Another possibility is a general reaction of the heart to a small change in a particular affected region that results in a cascade of greater reactions affecting the whole heart.

In conclusion, early extraembryonic venous obstruction in chick embryos results in cardiovascular malformations. Neither the frequency nor severity of the major cardiovascular malformations observed after vitelline vein ligation were dependent on the ligation position. Shear stress-regulated expression of ET-1 or TGF-β is suggested to be involved in the disturbed cardiogenesis. To correlate human placental venous blood flow with congenital cardiac malformations in the future, extensive data collection of placental blood vessels and cardiac morphology is necessary.


This work was supported by grant 900-516-096 of the Netherlands Heart Foundation and the Netherlands Organization for Scientific Research (NWO). The authors are greatly indebted to Mmes AMJ Baasten, D Vermeij and MMT Mentink for their technical assistance, Mr SB Blankevoort and Mr J Lens for preparing the illustrations, Mr R Brand from the Department of Medical Statistics, Leiden University Medical Center for statistical analysis of the data and Mr JR Tooms for computer assistance.


View Abstract