Main

Much of our current understanding of vertebrate heart development has been gained from studies of amphibian and avian embryos and from dissection and reconstruction of human hearts. Chick embryos are particularly accessible for experimental manipulation and observation compared with normal mouse embryo hearts which are poorly accessible without disturbing the placental blood flow. In addition, there is a lack of noninvasive techniques for physiologic assessment. Several studies describing hemodynamic change in the developing chick embryonic heart, both in physiologic and pathologic states, have been carried out using Doppler techniques(1–5). Such a 20 MHz Doppler system without spectral analysis was used successfully to quantitate dorsal aortic blood flow, and to qualitatively assess the inflow velocities. However, most of these studies were invasive in nature such that a tiny probe was inserted into the embryo and the hemodynamic status may have been disturbed. Doppler with spectral analysis has not previously been applied to this problem. A noninvasive method of embryonic cardiac assessment of blood velocities using Doppler was described by our group in the chick(6), and later in the mouse(7). A comparable understanding of mammalian embryos would be invaluable for the understanding of both normal and abnormal human heart development, especially during the stages of heart organogenesis. There is a need to obtain data on the developing mammalian heart that emphasizes the mechanistic, rather than the anatomical, perspective. Nakazawa et al.(8) have shown that ventricular function matures in the rat embryo. Two preliminary studies of normal and abnormal heart organogenesis in the mouse showed that abnormal trisomy embryos from both female and male matings could be detected by abnormal cardiac Doppler findings(9, 10) in a manner similar to human fetal echocardiography. Improved assessment of the mouse embryo may also be useful in the future using new imaging technology(11). Wladimiroff et al.(12) illustrated changes in normal fetal cardiac flow velocities in the late first trimester of human pregnancy using transvaginal Doppler.

In the present study, noninvasive Doppler velocimetry was applied prospectively to C57B1/6J mouse embryos to evaluate the potential applications and to study hemodynamic parameters of normal mammalian cardiac developmental sequences. Trisomic embryos on the same background were also evaluated. Three questions were addressed: 1) What Doppler patterns are found in normal embryonic mouse heart development? 2) What quantitative changes occur in normal embryonic mouse cardiac function with increasing gestational age from the time of cardiac septation? and 3) What types of abnormalities are encountered during screening and evaluation of a model of abnormal cardiac anatomy in trisomy embryos?

The methodology and patterns reported here may be also of importance in the analyses of transgenic mammalian embryo models noninvasively and especially in litters where heart defects may be suspected. Observations in animal models may give clues as to the abnormal physiology which affects human fetuses with similar genetic defects.

METHODS

Animals. Male and female C57B1/6J mice (Jackson Laboratories, Bar Harbor, ME) were housed in a room with a 12-h alternating light and dark cycle and maintained on Purina mouse chow and water ad libitum. Timed pregnancies were obtained by mating a single mature male, overnight, with two females. The presence of a vaginal plug the following morning was considered evidence of mating and counted as d 0 of gestation. Male mice heterozygous for the Robertsonian translocation Rb(16.17)7Bnr and Rb(6.16)24Lub, both on a C57B1/6J background, were obtained from Jackson Laboratories. The doubly heterozygous F1 males were mated with conventional C57B1/6J females carrying all acrocentric chromosomes to produce trisomic, monosomic (dead at or before implantation), and singly metacentric fetuses (euploid). This mating scheme results in a high probability of nondisjunction resulting in aneuploid progeny for chromosome 16(13).

Data presented are based on examinations of 129 embryos (24 litters) at different times of gestation beginning at d 10. The echocardiographic examinations were carried out on the morning of the day of gestation. In most cases, these examinations were done on odd days of gestation, except for d 10(the earliest day during the gestational period that Doppler patterns could be obtained) and therefore analysis of data on even days is not presented. The numbers of embryos examined each day were 12 embryos (gestational d 10), 25 (d 11), 27 (d 13), 27 (d 15), 27 (d 17), and 11 (d 19).

Protocol. Embryonic Doppler echocardiography was performed noninvasively from d 10 to 19 of gestation. The pregnant female was sedated with pentobarbital, administered intraperitoneally (40 μg/g of adult mouse weight). Echocardiography imaging and Doppler velocities were obtained from the embryonic hearts using the Interspec XL (Conshohocken, PA) instrument by placing the ultrasonic transducer (7.5 MHz) over the dam abdomen and coupling the ultrasonic energy with an acoustic gel. An ultrasonic standoff made from a gel-filled latex glove allowed placement of the embryonic heart at the focal depth of the transducer (Fig. 1A). Each of the beating embryonic hearts was systematically scanned to identify its position(Fig. 1B). Imaging at this frequency was adequate to locate the embryos and the beating heart and to orient the direction of the sagittal plane of the embryo. Because of the small size of the embryonic heart at these ages, this method of imaging at 7.5 MHz was not adequate to directly image the intracardiac anatomy. After identifying the heart, the sample volume of the pulsed Doppler was placed over the entire heart to obtain blood velocities. The sample volume length was adjusted to 2-4 mm to completely insonate the beating embryonic heart. The high pass filter was at its lowest setting of 50 Hz. It was critical to minimize the angle of insonation to obtain the maximal velocity of the blood flow because of the relation between the angle of insonation and the actual velocity (cosine Θ)(6). The same embryo was evaluated from several angles of insonation to obtain the maximal velocities. These maximal velocities(outlined below) were recorded independent of each other, but it was common to obtain a pattern of inflow and outflow waveforms together. Sequentially, inflow A, inflow E, and outflow wave forms were obtained(Fig. 2). After the last noninvasive evaluation and at the time of sacrifice of the dam, the abdomen was quickly opened to scan the embryos with the Doppler one last time and check the location of the embryos with respect to the Doppler patterns that were obtained. In this way, we gained semiquantitative information on the accuracy of embryo identification. We found, with experience, that the same embryo could be repeatedly examined, but the consistency of the orientation of the uterine horns at each examination could not be confirmed. Embryonic heart rates dropped with opening the maternal abdomen and so valid quantitative comparisons to noninvasive data were not possible. Applicability of this technique to embryos with smaller size, or strains with increased litter sizes (greater than 10/litter) is yet to be determined. Echocardiographic patterns were obtained on 20 trisomic embryos on 11, 13, and 14 d of gestation.

Figure 1
figure 1

(A) Illustration of the technique of examination of the sedated pregnant dam. The ultrasonic transducer is placed on the abdomen, and ultrasound is transmitted to the animal through a gel-filled standoff. (B) Schematic of multiple mouse embryos that were evaluated by Doppler (in this case there were eight embryos imaged and confirmed by autopsy).

Full size image
Figure 2
figure 2

The Doppler blood velocity wave from tracing from the inflow and outflow tracts of the embryonic mouse heart. A = peak A velocity, E = peak passive inflow velocity, ET = ejection time, FT = diastolic filling time measured from the onset of the passive E filling wave to the end of the A wave. Outflow = peak outflow velocity, Inflow = passive and active velocities of ventricular inflow, IRT = IRT measured from the end of ejection to the onset of ventricular filling, ICT = ICT measured from the end of the A wave to the onset of ejection.

Full size image

The maximal embryonic heart inflow and outflow velocity wave forms were recorded using a sweep speed of 100 mm/s on videotape using a Panasonic video cassette recorder AG6300 (Tokyo, Japan) and analyzed using an echocardiography analysis system (Digisonics, Inc., Houston, TX) for the peak velocities, the systolic ejection, and diastolic filling time intervals. Multiple angles were tried empirically to maximize the peak velocities of the inflow and outflow. The E and A velocities and embryonic heart rate were measured on the inflow velocity wave form. The peak ejection velocity was measured on the outflow wave form. The IRT, was measured as the time period from the end of ejection to the onset of filling, and the ICT, as the interval from the end of filling to the onset of ejection. All time intervals were measured in milli-seconds and expressed as a percentage of the beat to beat cardiac cycle interval(Fig. 2). All time intervals were measured to the nearest 10 ms. The data were plotted in relation to the day of gestation that the animal was examined.

It is known that ejection and inflow velocities for both ventricles in human fetal studies are similar comparing the left and right sides because both ventricles are pumping in parallel against similar afterloads(14). After cardiac septation, if an abnormality in the velocity pattern were detected shortly after ventricular septation, it would not be possible to localize its source to the side of the heart using current echocardiographic techniques. Because the heart size does not allow imaging of the separate left and right sides of the embryonic heart, the peak ejection velocities are obtained for both sides of the heart (pulmonary and aortic) after cardiac septation, and for the truncus arteriosus before septation. Similarly, the Doppler insonation of the embryonic inflow interrogates both the tricuspid and mitral valves after septation and the atrioventricular canal before septation. Furthermore, Doppler hemodynamic measurements are not absolute values, even for a particular strain.

The relationship between the Doppler measurements and gestational age in the normal mouse embryo was implemented as follows. First, mean values were obtained for each litter. The mean number of embryos per litter was 5.4. Measurements were obtained only at one gestational age. By computing mean values per litter, we obtained samples that contained independent observations. This allowed valid analysis using correlational and linear regression techniques. First, Pearson correlation coefficients were computed for each measure. Then, Spearman rank correlations were computed to assess whether the Pearson values were unduly affected by extreme values. Simple linear regressions were used to estimate the average changes per day using data only from normal embryos for parameters with evidence of a significant correlation. Mean values for each gestational age were computed for graphical presentation. t test was used to compare trisomy AVVR velocities and normal intervals for ejection and to study parametric variables for normal litter mates. Three observers made measurements on 13 embryos in duplicate, each evaluating the peak systolic velocity, the ejection time, and the cardiac cycle length. Two-way analyses of variance with random effects were used to analyze the data. Variance component estimation was used to compute estimates of interobserver variation, intraobserver variation, and the reliability coefficient from these sources of variation.

To provide a frame of reference for the presented study, a brief review of the relevant events of early heart development and the cardiovascular system of the mouse follows(15, 16). Events preceding the period of the present study: 8 d of gestation (1-7 somites). A horseshoe-like cardiogenic crescent begins to form, accompanied by the development at the 2-somite stage of the pericardial cavity that develops initially as lateral clefts in the mesoderm. These developmental processes proceed as an anterior to posterior wave that passes across the heart-forming region. In the 6-somite embryo the heart rudiment rapidly develops. The cardiac mesoderm cells form endocardial tubes that shortly will fuse anteriorly. In the mouse embryo as the endocardial tubes fuse to form a single tube, the fused region immediately begins to beat. Posterior heart forming regions in a gradient fashion are still differentiating and developing. Hence, the anterior (ventricular) region beats earlier, before the more posterior(atrial) regions which are still forming. (This is in contrast to the situation in the chick embryo, where both the ventricular and atrial regions develop and fuse to form a single tubular heart before beginning to beat.) The developing tube is continuous with the first aortic arch, which is developing along with the paired dorsal aortae. At 8.5 d of gestation (8-12 somites) the embryo begins to rotate. At 9 d of gestation (13-20 somites) the endocardial tube of the bulbus cordis is very narrow and begins contracting. The heart is now able to maintain some blood circulation. The ventricle and atrium are not yet septated. The placental circulation is beginning to form. The blood is circulating in the yolk sac via the paired dorsal aortae and a thick vitelline artery has formed. By 9.5 d of gestation (21-29 somites) the heart is a convoluted tube that is still not divided into left and right sides. First, second, and third aortic arches are developed. Events occurring during the period of this study. At 10 d of gestation the cardinal veins are present. At 10.5 d of gestation (35-39 somites) two complete extraembryonic circuits have been established: the yolk sac and the placental circulation. The heart is still a single, convoluted tube with two distinct constrictions, the sulcus atrioventricularis and the sulcus sinu-atrialis. The sixth aortic arch is developing and sends a branch to the lung primordium. On 11 d of gestation(40-44 somites) the heart remains undivided. Bulbar ridges are now discernible. At 11.5 d of gestation (over 45 somites) the first arch has degenerated; the second is reduced. The third, fourth, and sixth arches are both well developed. The sinus venosus receives the cardinal veins, right umbilical vein, and the primitive vena cava. Atrium is nearly completely partitioned off; the ventricle remains undivided. The atrioventricular cushion delimits the ventricle from the atrium. By 12 d of gestation the truncus arteriosus is being divided, and by d 13 of gestation the aortic and pulmonary trunks are completely separated. The interventricular septum is not completely closed. The valves of the heart are now present in a primitive form. By d 14 of gestation the definitive pattern of the prenatal circulatory system is established and the definitive shape of the heart as well. The septum dividing the ventricles is now closed. On d 15 of gestation the atrioventricular and semilunar valves are well developed. At the close of this stage the arteries and veins have the final fetal pattern that provides a larger umbilical circulation than yolk sac circulation.

Trisomy 16 mouse embryos are considered a model for analyzing development related to human Down's syndrome due to homology between human chromosome 21 and mouse chromosome 16. Over 91% of the trisomy 16 mouse embryos have been shown to have heart defects anatomically, and most have atrioventricular canal defect(17). Trisomic mouse embryos usually die around d 16 of gestation.

RESULTS

The present report is an analysis of Doppler studies performed on 129 mouse embryos during different days of gestation. We were limited in our observations on gestational d 10 and before due to the sensitivity of the imaging and Doppler equipment. However, we observed consistent atrioventricular synchrony, such that each atrial contraction filling wave form was followed by a succeeding ventricular ejection wave form. We infer from these mechanical events that normal electrical atrioventricular coupling is established by d 10 of gestation. Over this time frame the valves are developing, but at gestational d 10 there is no sign of mature valvular tissue. The endocardial cushions are compressed by the outer myocardial layer, occluding the atrioventricular canal and preventing retrograde flow during ventricular systole (no AVVR). A typical Doppler pattern of AVVR in the fetal and postnatal heart has been described(18). During these Doppler evaluations of normal mouse embryos from gestational d 10 to 19, none demonstrated this pattern.

A repeatable qualitative pattern of the evolution of mouse embryonic blood velocities was observed (Fig. 3). Blood velocities into the ventricle during diastolic filling are displayed as deflections above the line of zero velocity in Figure 3. Deflections below the line of zero velocity are blood velocities of ventricular outflow ejection. Three different characteristic patterns of cardiac filling and ejection blood velocities were observed: early phase (gestational d 11)-monophasic inflow velocity of filling separated from ventricular ejection velocity by a period of isovolemic contraction with zero velocity; mid phase (gestational d 15)-biphasic inflow velocities of passive (E) and active (A) ventricular filling and an increased ventricular ejection velocity. ICT is decreased and IRT does not change; and late phase (gestational d 17)-a biphasic mature fetal pattern of ventricular filling without a measurable period of isovolemic contraction. A period of isovolemic relaxation remains. A higher peak velocity of ventricular ejection is present with increasing gestational age.

Figure 3
figure 3

Schematic of normal changes in the Doppler mouse embryonic blood velocities at increasing gestational age. There are three typical patterns illustrated: early (gestational d 11), mid (d 15), and late(d 17). The inflow wave form is illustrated consistently above the line of zero velocity and is monophasic early and biphasic late in gestation. The outflow ejection velocity increases with increasing age, the ICT becomes unmeasurable after gestational d 15, and the heart rate increases. The late pattern of biphasic filling and ventricular ejection is similar to human fetal Doppler in the second and third trimesters.

Full size image

Table 1 summarizes the correlation analyses for normal litters and indicates evidence of significant linear associations with gestational age (d 10-15) for outflow velocity, A velocity and E velocity. Spearman rank correlations were similar with the exception of outflow velocity where the rank correlation value was 0.66 (p = 0.002).Table 2 presents the results for the simple linear regression for measures with evidence of a significant linear association and sufficient data. The slope values in Table 2 represent estimates of the average daily changes in the Doppler parameters. Predicted(mean) values as a function of day can be obtained by adding the intercept value to the product of gestational age times the estimated slope. The root mean square is an estimate of the SD around the mean values. Thus, there is evidence that peak velocity, A velocity, and E velocity increase as a function of gestational age. The mean heart rate values increased from 124 beats/min at d 11 of gestation up to 230 beats/min on d 19 of gestation(Fig. 4A). The mean maximal outflow velocity increased from 18 cm/s on gestational d 11 to 32 cm/s on d 19 (Fig. 4B). The mean peak inflow velocity (A wave) increased from 22 to 31 cm/s between gestational d 11 and 19 (Fig. 5A). The mean E wave velocity increased from 3 to 14 cm/s between d 11 and 19 of gestation(Fig. 5B). The diastolic filling time as a percentage of the cardiac cycle increased from 38 to 50% (Fig. 6). The systolic ejection time as a percentage of the cardiac cycle remained constant at approximately 33%. The IRT as a percentage of the cardiac cycle was constant averaging 20%. The ICT decreased from 15% of the cardiac cycle to unmeasurable.

Table 1 Normal mouse embryo Pearson correlation coefficients for relationships with gestational age
Full size table
Table 2 Regression analyses results for normal litters
Full size table
Figure 4
figure 4

(A) Normal mouse embryonic heart rate during gestation. (B) Normal mouse embryonic cardiac outflow peak velocity. The peak outflow tract velocity increases with gestational age. Mean ± SD of age group data are indicated.

Full size image
Figure 5
figure 5

Normal mouse embryonic inflow velocities. (A) The A peak inflow velocity values with increasing gestational age between d 10 and 19 of gestation. (B) The E wave peak velocity increased significantly between d 11 through 19 of gestation. Mean ± SD of age group data are indicated.

Full size image
Figure 6
figure 6

Mouse embryonic Doppler cardiac time intervals. The diastolic filling time as expressed as a percentage of total cardiac cycle time showed a significant increase between d 10 and 19 of gestation. There is no significant change in systolic ejection time or in IRT during the course of gestation. The ICT was no longer present after gestational d 15. SD intervals are indicated.

Full size image

To study the variability of the Doppler technique the variables analyzed were inspected by a comparison of the results of two normal litters on gestational d 15 (Table 3). There is a significant difference in the peak outflow velocity (p < 0.01), whereas there is no significant difference in the other blood velocity or time interval variables. The variability that was seen in echocardiographic patterns among a litter was most likely due to variability in size and development of the different embryos within a litter, but also could have been due to the different angles of insonation which affect the peak velocity obtained. Embryo sizes are known to vary by the location in the uterine horn, which could create differences. Total errors for Doppler measurements were similar to those in Table 3. For all three measurements in the intra- and interobserver variation study the random interaction terms were statistically significant. There was no significant variation among observers for peak velocity (p = 0.33) and cycle length (R-R interval)(p = 0.22). Variation among observers was statistically significant for ejection time (p = 0.002). The reliability coefficients for the three variables (peak velocity, R-R interval, and ejection time) were 0.93, 0.99, and 0.80, respectively. Interobserver SD was 1.82 cm/s, 0.0054 s, and 0.0122 s, respectively, whereas intraobserver SD was 1.17 cm/s, 0.0039 s, and 0.0076 s respectively.

Table 3 Variability in Doppler measurements in gestational d 15 embryo
Full size table

Of the 20 trisomic embryos, 11 embryos showed abnormal echocardiographic wave forms (55%) (Table 4). Figure 7A is an echocardiogram of a trisomy 16 mouse embryo on d 10 of gestation showing outflow tract regurgitation. Note the early diastolic velocity superimposed on the A wave. The IRT that was seen in the normal heart pattern above is not present in the wave form pattern obtained on this trisomy 16 mouse heart. Six trisomic embryos (30%) have shown similar outflow tract regurgitation. Figure 7B is an echocardiogram of a heart demonstrating AVVR in the trisomy 16 embryonic mouse heart on d 11 of gestation. Note the holosystolic, high velocity jet which obscures the normal ejection wave form. The ICT is absent. Five trisomic embryos showed similar AVVR (25%). The peak velocity and duration of the wave form were significantly larger than that of normal embryos (Table 5). The remaining nine trisomic embryos showed normal blood flow patterns for their respective gestational ages. Although cardiac defects may be present in these mice, it did not alter blood flow patterns significantly to be detectable by current Doppler echocardiography. As in human trisomy 21 patients, there was variability resulting in a spectrum of defects ranging from relatively normal individuals to those severely affected. For d 11, three litters each containing one trisomic embryo were examined, and the results are shown inTable 5. For d 10, 13, 14, and 15 of gestation, respectively, two trisomic embryos, i.e. one embryo in each of two separate litters for each gestational age, were obtained. Trisomy was later confirmed by the demonstration of two metacentric Robertsonian chromosomes and a total of 41 chromosome arms in metaphase chromosome spreads of each trisomic embryo compared with 40 chromosomes in normal litter mates.

Table 4 Doppler findings in trisomic embryos (n = 20)
Full size table
Figure 7
figure 7

Echocardiographic patterns obtained on trisomic embryos on d 10 and 11 of gestation demonstrating outflow (A) and atrioventricular (B) regurgitation. (A) Doppler echocardiogram of a trisomy 16 mouse embryo on d 10 of gestation showing outflow tract regurgitation. Note that the IRT that was seen in the normal heart pattern above is now absent in the wave form obtained on the trisomy 16 mouse heart and a diastolic velocity is superimposed on the early filling. Six trisomic embryos showed semilunar valve regurgitation. (B) Echocardiogram of a heart demonstrating a holosystolic, high velocity wave form of AVVR (AVVR) in the trisomy 16 embryonic mouse heart on d 11 of gestation. Note that the ICT is obscured by the early onset of the jet. Five trisomic embryos showed AVVR (Table 2).

Full size image
Table 5 Normal ejection vs AVVR
Full size table

DISCUSSION

In the embryonic mouse, the heart septates and the atrioventricular valves appear from cushion precursors during 11-12 d of gestation(15, 16, 19). Over this time interval we obtained Doppler data from the embryonic mouse noninvasively. The normal C57B1/6J strain of mice used in this study is commonly used for experimental work and is not known to manifest cardiac or vascular abnormalities. In our study, competent atrioventricular function was a consistent observation throughout the study, even in the very early stage in which all cardiac valves had not completely developed.

The velocities detected in these small hearts are the instantaneous changes in the energy differences of the blood cells during acceleration and deceleration. These maximal velocities originate at the smallest area of the inflow or outflow. Using this technique we could not differentiate the exact location of the velocity. With experience it was straighforward to differentiate filling and ejection wave forms based on the characteristics of the signals. These patterns during gestation (Fig. 8,A andC) resembled very closely the wave forms of the inflow and outflow of normal human fetal Doppler echocardiography findings (Fig. 8,B andD). Also, similar changes were observed in early human Doppler experience(14, 19). We emphasize that these Doppler hemodynamic measurements are not absolute values even for a particular strain. However, in the era of transgenic embryos, this Doppler technique should be suitable for comparing wild and homozygote embryos. Another possible limitation of the technique is the need for sedation of the dam for echocardiographic evaluation of the embryos. We have not yet compared our current use of barbiturates with other medications systematically, but this should be done. The fact that our results are very similar to that of unsedated human fetuses and that similar results were found between litters suggests no systematic error from sedation. We have recognized the importance of maintaining the dam temperature, and we used a warming pad and radiant lamps for this purpose. Nonetheless, the absence of signs of embryonic heart failure such as valve regurgitation in a large number of normal embryos suggests that medication was not causing cardiovascular dysfunction. Serial studies require a consistent orientation of the uterine horns to compare serially results from the same embryo. Trisomic embryos with abnormal Doppler patterns gave us the first opportunity to track the same embryo unequivocally over time. It was possible to identify the abnormal embryo in a consistent location and to return to that embryo in this preliminary experience.

Figure 8
figure 8

A comparison of normal human and normal mouse Doppler-derived blood velocity wave forms at comparable stages of heart development. (A) Normal mouse embryonic echocardiographic pattern on d 11 of gestation. (B) Normal human embryonic echocardiographic pattern at 7 wk of gestation. (C) Normal mouse embryonic echocardiographic pattern on d 15 of gestation. (D) Normal human embryonic echocardiographic pattern at 20 wk of gestation.

Full size image

One may consider why the septating heart does not have an abrupt change in velocities of filling or ejection after the completion of septation. We know that, over the course of cardiac development, the inlet and outlet of the heart divide into nearly equal portions. This would not have any impact on the peak velocities at the inlet or outlet, however, because both the flowand the cross-sectional area of flow are halved.

The atrial contraction or “A” wave during diastolic filling of the embryonic ventricles is the only filling observed in the early phase embryos often with absence of the passive filling wave or “E” wave. This is consistent with a decreased compliance of the myocardium. The appearance of the E wave later suggests increased compliance. Human data during development shows a decreasing A velocity and an increasing E velocity in a similar manner(14, 19). As a percentage of the total cardiac cycle, the ventricular filling time shows an increase during development. The increasing E velocity later in gestation seen in this study and in human fetal Doppler from 14 to 40 wk gestation, with constant IRT, is the result of increasing blood flow and may not necessarily imply changes in ventricular compliance(14). Atrial contraction generates a velocity higher than the outflow velocity between d 11 and 16 of gestation, suggesting that the work of atrial systole is very important in early embryonic heart development. For technical reasons, venous Doppler was not obtained in this study. The atrial portion of embryonic ventricular filling changes from monophasic to biphasic with gestational age. What role a decrease in atrial work plays in the observed changes in A/E ratio is impossible to determine with only Doppler data. The filling pattern early in gestation in the chick is different from mammals with a single E wave followed by the development of the atrial contraction wave(2).

Earlier it was found that the peak outflow velocity in chick embryos was increasing in a linear pattern until mid-gestation and then decreased(6). This mouse study shows that the blood velocity across the outflow tract increases in a linear pattern throughout the period of gestation, which is consistent with studies in human embryos and fetuses(12, 19). The systolic ejection time as a proportion of total cardiac cycle time was constant throughout the gestation period in the mouse. Growth of the valve areas keeps the peak velocities below a level at which turbulent flow would result in increased work of ejection. The velocities measured by other workers in the chick embryo using different technology are significantly lower than those shown in this study(1–4). This can be explained by the differences in the Doppler techniques. The zero-crossing technology used with the 20 MHz crystal system used previously measured mean velocity and does correlate with cardiac flow. The instantaneous velocities obtained with state-of-the-art Doppler technology used here measure the instantaneous velocity spectrum at each instant and therefore higher velocities. This technique therefore assesses not the flow but rather the energetics of the embryonic circulation. Characterizing the nature of the velocities in the embryonic circulation will require improvements in the spatial resolution of this technique. Another possible explanation of the differences in the two methods of Doppler is the location of measurement. Chick embryo aortic arch data may have a lower velocity due to the dynamic narrowing of the outflow tract where the velocities in this study were obtained. However, later in human gestation the aortic arch velocity is higher than that at the aortic valve, and our data in the mouse show progressive increases in peak ejection velocity.

Early in cardiogenesis, there is a decrease in the time of isovolemic contraction. Whether the delay in ventricular ejection after the atrial contraction wave is due to atrioventricular electrical delay or due to a slow rise of ventricular pressure was not clear initially. By combining these observations with those in the trisomic embryos with regurgitation, we conclude that the ventricular pressure is rising after the end of the A wave. Therefore, we may exclude the hypothesis that the ICT is due to delay in atrioventricular conduction. Wladimiroff et al.(12) have suggested that the transition of umbilical arterial and venous Doppler in the human fetus near the time of trophoblast invasion (10-14-wk gestation) in the placenta is linked to a decrease in cardiac work from this event. It is possible that the transition in cardiac Doppler described in this research experience (monophasic to biphasic ventricular filling) could be linked to the same event. Alternatively, there may be a change in the myocardial compliance as a primary cause of this change. Obstruction at the conotruncus could also be a factor(8).

A factor in the noninvasive early identification of the trisomic embryo in this study was the characterization of abnormal wave forms of atrioventricular or semilunar valve regurgitation. Because the blood velocity of AVVR is holosystolic and passing from the ventricle to the atrium instead of to the outflow as in normal ejection, one would expect to find this wave form to have several distinctive features. First, the velocity would occupy more of the cardiac cycle than that found in normal ejection, and these results were confirmed in some trisomic mice. The duration of the regurgitation jet is from the onset of ventricular pressure rise to the time when the atrial pressure again exceeds the ventricular pressure. This effectively eliminated the ICT from the wave form because the presence of regurgitation allows the jet to be detected even before ventricular ejection occurs. This was a further aid is separating normal ejection from regurgitation. Second, the peak velocity of AVVR would be higher than ventricular ejection because the end-diastolic pressure in the great arteries is much higher than that in the atrium. Therefore, the pressure gradient during systole between the ventricle and the outlet is less than between the ventricle and the atrium. The higher holosystolic velocity jet was detected only in a proportion of the trisomy 16 embryos. Heart rate differences in the trisomy 16 embryos were not large enough to explain the increased velocities. Further work is necessary to differentiate at these low systolic pressures the characteristics of AVVR which is a marker of a congenitally abnormal valve precusor and that which will lead to embryonic death from progressive congestive heart failure. Regardless of the cause, we conclude that the Doppler findings of a systolic jet with both increased duration and peak velocity is consistent with AVVR. Because even mild degrees of tricuspid valve regurgitation in the fetus often indicate abnormal physiology(20), improvements in ultrasonic technology to improve detection of this finding may improve detection of abnormal embryonic physiology. We recognize that not all trisomic embryos have abnormal physiology, but morphologic and genetic evaluation of those that do could add an important tool to the investigator of fetal cardiac failure. From open abdominal dam echocardiography correlations we were able to examine 90% of the viable embryos. Taken with a 90% chance of atrioventricular canal defect in trisomy 16 embryos, this gives a theoretical 81% chance of detecting the abnormal using this new technique.

This study indicates that changes in cardiac physiology as reflected by blood velocities can be assessed in the developing mammal during embryonic cardiac development as early as d 10 of gestation in the mouse. In comparison to human embryos in the first trimester period and beyond, the Doppler echocardiographic patterns are similar. This indicates that this methodology can be used to investigate physiologic changes occurring during mammalian heart development. The present Doppler evaluation of normal blood flow patterns in embryonic mouse hearts during a critical period of early heart development provides a control baseline for future studies where abnormalities of heart development may be expected. A high velocity holosystolic or diastolic velocity with altered time intervals was identified as atrioventricular or semilunar valvular regurgitation, respectively. An abnormal pattern of inflow or outflow valvular regurgitation was present in 55% of full trisomy 16 embryos; however, no normal embryos showed this pattern. Hydrops and cardiac failure in the trisomic embryos should be manifested as failure of valvular function, which is what was detected by Doppler techniques and what is found in anatomical examination of the embryos(17). This indicates that the mouse embryo is an appropriate animal model for further investigation to analyze normal and abnormal mammalian heart development and function.