Blood flow patterns underlie developmental heart defects

Congenital heart defects result from genetic anomalies, teratogen exposure, and altered blood flow during embryonic development. We show here a novel “dose-response” type relationship between the level of blood flow alteration and manifestation of specific cardiac phenotypes. We speculate that abnormal blood flow may frequently underlie congenital heart defects.

Abstract

Although cardiac malformations at birth are typically associated with genetic anomalies, blood flow dynamics also play a crucial role in heart formation. However, the relationship between blood flow patterns in the early embryo and later cardiovascular malformation has not been determined. We used the chicken embryo model to quantify the extent to which anomalous blood flow patterns predict cardiac defects that resemble those in humans and found that restricting either the inflow to the heart or the outflow led to reproducible abnormalities with a dose-response type relationship between blood flow stimuli and the expression of cardiac phenotypes. Constricting the outflow tract by 10–35% led predominantly to ventricular septal defects, whereas constricting by 35–60% most often led to double outlet right ventricle. Ligation of the vitelline vein caused mostly pharyngeal arch artery malformations. We show that both cardiac inflow reduction and graded outflow constriction strongly influence the development of specific and persistent abnormal cardiac structure and function. Moreover, the hemodynamic-associated cardiac defects recapitulate those caused by genetic disorders. Thus our data demonstrate the importance of investigating embryonic blood flow conditions to understand the root causes of congenital heart disease as a prerequisite to future prevention and treatment.

NEW & NOTEWORTHY Congenital heart defects result from genetic anomalies, teratogen exposure, and altered blood flow during embryonic development. We show here a novel “dose-response” type relationship between the level of blood flow alteration and manifestation of specific cardiac phenotypes. We speculate that abnormal blood flow may frequently underlie congenital heart defects.

the genesis of heart defects during embryonic development is rooted in complex and unexplained dysregulation of normal heart formation. Congenital heart disease (CHD) is the most common and lethal structural disease present at birth in newborns (14). Although several genetic disorders have been associated with congenital heart defects, less than 20% of defects are clearly linked to specific human gene mutations (43). Furthermore, even known cardiac-defect-inducing genetic mutations do not affect individuals equally and are typically associated with a range of anatomical outcomes. Cardiogenesis is a finely orchestrated interplay between both genetic and environmental factors, including exposure to teratogens and abnormal blood flow. Nevertheless, scientists investigating the underlying causes of CHD often neglect the confounding effects of blood flow alterations in favor of signaling pathways derived from gene anomalies. Meanwhile, the multifactorial combination of environmental and genetic influences is attracting ever increasing attention as the explanation for complex congenital heart defects (14, 42), but individual effects are difficult to tease apart. Here, we focus on the effects of altered blood flow patterns alone.

Hemodynamic forces associated with blood flow play an important role in early embryonic cardiac morphogenesis (17, 24) as well as in structural features of the mature heart (7, 18, 25). In the embryo, the heart is the first functional organ and starts beating as soon as a primitive tubular heart is formed. Once beating, pump-derived blood pressure and shear (frictional) forces exerted on tissue walls trigger nascent mechanotransduction mechanisms that lead to physical, biochemical, and gene regulatory responses within the heart and blood vessels (11, 24). The embryonic cardiovascular system quickly responds to abnormal mechanical cues (due to altered blood flow) by passively altering contractile function (26, 54), which is followed by abnormal tissue composition remodeling that eventually leads to subtle or overt cardiac defects (8, 9, 15, 21, 23, 24, 40, 46, 53). Genetic anomalies and environmental exposures can alter blood flow in human embryos and fetuses by inducing structural cardiovascular malformation, contractile deficiencies, and inadequate vascularization of the placental and vitelline beds (Fig. 1A). For example, intrauterine growth restriction is associated with decreased placental and fetal blood flow (5), while maternal smoking during pregnancy increases fetal vascular resistance and umbilical artery blood velocity (37). Although it is clear that hemodynamic forces are crucial components of heart development, the relationship between altered hemodynamic conditions and cardiac malformation phenotypes remains unclear.

Fig. 1.

Abnormal hemodynamics. A: human embryonic circulation with factors contributing to hemodynamic perturbations. B: schematic of a normal chick embryo at day 3 [Hamburger and Hamilton stage 18 (HH18)], corresponding to ~4 wk in human development. Example optical images of chick embryos at HH18, after surgical interventions used to alter blood flow with vitelline vein ligation (VVL; C) and outflow tract banding (OTB; D). Scale bar = 1 mm.

Previous studies have found a spectrum of congenital heart defects after hemodynamic interventions (8, 9, 23, 34, 46, 54), suggesting that multifactorial CHD may be the result of genetic factors compounded by hemodynamic signals. In such studies, however, altered hemodynamics induced by surgical interventions have only been investigated as a group and the possible large range of hemodynamic change within the group has been neglected. Previous studies, therefore, have not investigated whether different levels of hemodynamic perturbation may lead to distinct cardiac phenotypes. Herein, using chicken embryo models of cardiac development, we show that specific, reproducible, congenital heart defect phenotypes can be induced by precisely controlled blood flow patterns during early cardiac development. We thus demonstrate that congenital cardiac malformations are finely regulated by the specific hemodynamic environment. Future diagnostic imaging of early embryonic/fetal circulation will be critical to help guide management decisions during pregnancy to possibly revert or prevent cardiac malformations.

METHODS AND METHODS

Experimental design.

The experiments outlined in this study were designed to determine the extent to which different blood flow conditions at early stages of cardiac development lead to distinct cardiac phenotype, heart defect incidence, and cardiac functional abnormalities. To conduct the studies, we used chicken embryo models of heart development, since it is relatively easy to surgically manipulate blood flow conditions in chicken embryos in ovo, while minimally disrupting other physiological processes. Chick embryos are not considered vertebrates under Institutional Animal Care and Use Committee and Oregon Health & Science University regulations; however, every effort was made to minimize the number of embryos needed. Results were compared between embryo groups that were selected based on hemodynamic (blood flow) alteration. Embryo groups included a normal group (no intervention), a surgical sham group (intervention control), and five surgical intervention groups, with distinct blood flow conditions. The hemodynamic interventions were performed at ~3 days of incubation at Hamburger and Hamilton stage 18 (HH18) (20), an early stage of embryonic development when the heart is a looped tubular structure, with atrium, ventricle, and outflow tract connected in series. At HH18 the heart has no valves nor chambers but is already pumping blood. Two hours after surgical intervention, the embryonic hearts were monitored with optical coherence tomography (OCT) in ovo to precisely determine the level of hemodynamic alteration in each embryo. Altered blood flow conditions lasted at most 24 h during tubular stages. Embryos that survived to about day 12 of incubation (HH38) had their hearts imaged with high-frequency ultrasound in vivo for functional analysis and then excised for imaging using microcomputed tomography for detailed structural analysis.

Hemodynamic intervention.

Fertilized white Leghorn chicken eggs were incubated blunt end up at 38°C and 80% humidity until HH18. Embryos that bled or had obvious structural defects at HH18 were discarded. We altered cardiac hemodynamics in the HH18 chicken embryo using two well-established interventions: 1) outflow tract banding (8, 9, 34, 46, 54); and 2) right vitelline vein ligation (VVL) (23, 44) (Fig. 1, B–D). Four types of embryo groups were included in this study: 1) normal group (NL), no interventions were performed; 2) control group (Con), a 10–0 nylon suture was passed under the heart outflow tract but not tightened; 3) outflow tract-banded groups (OTB), a 10-0 nylon suture was passed under the heart outflow tract and tied in a knot around the mid-section of the outflow tract (downstream of the cushion and near the flat portion of the outflow tract) to constrict the tract cross-sectional area; and 4) right VVL, a 10-0 nylon suture was tied in a knot around the right vitelline vein close to the embryo body to acutely stop venous return flow through that vein [reducing flow to the heart in about half for ~5 h (34, 48)] before blood flow was later restored by angiogenesis (see Fig. 1, B–D). Band tightness, which measures the degree to which the outflow tract is constricted by banding, ranged from 0 to 60%. Following interventions, windowed eggs were sealed with saran wrap and incubated until further evaluation. The band was removed from the outflow tract in the banded group ~24 h after placement when the embryo reached HH24 to increase survival to later stages, and then the embryo was allowed to develop to HH38 when the heart is nearly fully developed with four chambers and valves. The suture knot in the vein-ligated group remained on the vein until the final evaluation at HH38.

Band tightness measurement with OCT.

A custom-made OCT system was used to measure chick embryo band tightness as previously described (29, 30, 33, 44, 47). Briefly, the system has a spectral domain configuration with a superluminescent diode centered at 1,325 nm and a 1,024 pixel, 92-kHz maximal line-scan rate infrared InGaAs line-scan camera. It acquires 512 × 512 pixel, two-dimensional B-mode line-scan tomographic images at 140 frames/s with <10 μm resolution. Normal physiological temperature (38°C) was maintained during imaging with a thermocouple-controlled heating system. Two-hundred frames (~3 to 4 cardiac cycles) of the longitudinal outflow tract were acquired for each banded embryo before and after manipulation to measure the change in outflow tract diameter (D), calculate the degree of band tightness (ΔD/D), and measure blood flow velocity in the outflow tract, thus establishing the level of hemodynamic alteration.

Intervention velocity analysis.

A subset of embryos were imaged with Doppler OCT at HH18 and high-frequency ultrasound at HH24 to analyze changes in blood flow after hemodynamic surgical interventions. Constriction of banded embryos ranged between 21 and 52% tightness for this subset of embryos. Peak blood flow velocity at HH18 was measured with Doppler OCT, 2 h after each manipulation. The 2-h time period after we performed the sham surgery in the Con embryos, the banding in the banded embryos, or the ligation in the vein-ligated embryos allowed the embryo to warm back to physiological temperature in the incubator and physiologically respond to the intervention. A velocity sampling area was selected in the region of nonwrapped flow along the centerline and near the middle of the outflow tract to be most representative of changes due to band constriction in banded embryos, as well as consistent with the location within the outflow tract across all embryo groups [see our previous publication for a more detailed explanation (33)]. Peak blood flow velocity was then measured in the same embryos and outflow tract location at HH24 with our ultrasound system (Visualsonics, Vevo 2100). The positioning of the embryo and heart in the egg at HH18 allows us to employ the high spatial resolution of OCT imaging; however, by HH24 the embryo starts to sink in the egg and can no longer be imaged with OCT. Ultrasound provides enough tissue penetration and spatial resolution to observe cardiac wall features and blood flow at HH24. (See Echocardiography for more ultrasound details.) This data set included two groups of banded embryos: one that did not have the band removed at HH24, to assess the effect of banding after 24 h; and a second group that had the band removed 2 h before velocity measurement at HH24, to assess the effect of band removal on hemodynamics. The two groups were needed because flow measurements at HH24 are invasive and thus might affect the embryos, preventing longitudinal studies.

Echocardiography.

Blood flow velocity at HH24 and cardiac function of the mature heart at HH38 were evaluated with high-frequency ultrasound (Fig. 2). The windowed eggs were submerged in physiological temperature-controlled Hank’s balanced salt solution to make contact with the transducer probe. A 32- to 56-MHz linear array transducer probe was used with the smaller cardiac structures at HH24, and 18- to 38-MHz ultrasound linear array transducer probe was used to image the heart at HH38. Our high-frequency ultrasound Vevo 2100 system acquires images at 100 frames/s, with an axial resolution of ~30 μm. A complete two-dimensional B-mode (structure), Doppler velocity, and tissue Doppler imaging (TDI) echocardiographic examination was performed for each embryo to assess structural heart integrity and cardiac function at HH38. Peak blood flow velocity, velocity time integral, and heart rate were averaged over two to three cardiac cycles from the Doppler velocity signal from the left ventricular outflow tract. The flow angle was set during acquisition, so that Doppler velocities were automatically corrected for the direction of flow. Stroke volume, cardiac output, and left ventricular ejection fraction (LVEF) were calculated using the modified Simpson’s rule with four-chamber and two-chamber B-mode images (19, 27, 36). Additionally, TDI was used to record systolic (S′) and early diastolic (E′) peak radial and longitudinal myocardial velocities at the mitral lateral annulus from a four-chamber view. Please note that while longitudinal velocities are more widely used clinically (6, 22), the heart orientation required to measure longitudinal velocities was difficult to achieve with the embryo in the egg, and this reduced the functioning sample size. Color-Doppler, high-frequency ultrasound was used to identify ventricular septal defect (VSD) flow through the ventricular septum.

Fig. 2.

High-frequency ultrasound imaging at HH38. Example images from a normal (NL) embryo in diastole (A–C) and systole (D–F) at 4-chamber (A and D), cross-sectional mitral valve (B and E), and cross-sectional papillary muscle (C and F) views. Ventricle lengths in the long-axis view are noted in A, and ventricles perimeters are outlined in short-axis views in B and C. RV, right ventricle; LV, left ventricle. Scale bar = 1 mm.

Microcomputed tomography.

After ultrasound imaging at HH38, each heart was collected for microcomputed tomography (micro-CT) to identify structural defects and analyze structural geometry. The heart was arrested at end-diastolic phase by injecting chick ringer solution containing 60 mM KCl, 0.5 mM verapamil, and 0.5 mM EGTA (52), collected, and stained with 100% Lugol’s solution for 16 h to enhance tissue contrast. A Caliper Quantum FX Micro-CT system was used with a 10-mm field of view, 140-μA current, and 90-kV voltage to acquire high-resolution (~10 μm), three-dimensional heart scans. Scans were discretely sampled using the Amira software platform to generate two-dimensional image stacks and identify major cardiac defects. Double outlet right ventricle (DORV) was identified from micro-CT images as dextroposition of the aorta with alignment of both great arteries with the right ventricle, always combined with a VSD. VSDs were confirmed by Doppler ultrasound, a feature that enables color-coded identification of flow from ultrasound images. Tetralogy of Fallot (TOF), like DORV had dextroposition of the aorta but was separately characterized by unequal outflow sizes with a large aorta and small pulmonary artery, combined with a large VSD. Cardiac defects were mainly identified from micro-CT images and identification was aided by ultrasound.

Micro-CT images were also used to measure aortic arch diameter, compact and trabeculated myocardium thickness, septal thickness, and ventricle cavity diameters and length. These dimensions were measured from micro-CT images to ensure comparable measurement locations across embryos and to have adequate image resolution for fine measurements. Wall thicknesses and chamber diameters were measured at a midportion slice that contained both ventricles, just below the muscular rim of the tricuspid valve. Compact and trabeculated myocardium thicknesses were measured from the lateral segments of each ventricle wall. Relative wall thickness was calculated for each ventricle (posterior + septal wall thickness/end-diastolic cavity diameter) to categorize wall thickening and chamber dilation patterns of response. Diameters of the pharyngeal arch arteries were also measured from micro-CT images, where the base diameters of the aortic root, pulmonary root, right and left aortic arches, right and left brachiocephalic arteries, and right and left pulmonary arteries were compared with the sum of the aortic and pulmonary roots to allow for any differences in embryo size (9). A sphericity index was calculated for each ventricle (ventricle base-to-apex length/basal diameter) from a four-chambered image view (10). The sphericity index was used to detect elongation and shortened ventricles, where elongated and globular hearts were defined as having a sphericity index higher or lower than the average mean of the control group ±2 SD, respectively.

Statistical analysis.

To compare significant changes among embryo groups, banded embryos were divided into four different band tightness ranges to capture distinct hemodynamic conditions induced by the intervention: 11–20, 21–35, 36–45, and 46–60% band tightness. The χ² statistic was used to compare proportions of embryos that survived to stage HH38 in each group to the NL group, assuming significance with corresponding two-tail P < 0.05. All surviving embryos in banded groups and the vein-ligated group were compared with control groups and analyzed as the means ± SD. Statistical significance between functional and structural parameters was determined with a two-sample Student’s t-test, assuming significance with two-tail P < 0.05. Because the defects that arise in late stages create separate and dissimilar hemodynamic conditions, only embryos without major defects (at least 3 embryos with no major structural defects per group) were compared with controls to evaluate changes in cardiac function and structure after hemodynamic manipulation in seemingly normal hearts.

RESULTS

Altered hemodynamics.

We altered cardiac hemodynamics in the early chicken embryo using two well-established interventions: OTB (8, 9, 34, 46, 54) and 2) right VVL (23, 44) (Fig. 1, B–D). Both interventions were performed at HH18, when the heart is a looped tubular structure. The heart at this stage is sensitive to hemodynamic changes that affect later septation and valve formation (16, 23, 24), and the interventions lead to heart defects seen in human babies. Outflow tract banding (referred to in this paper as “banding” for simplicity) was used to reduce the diameter of the early embryonic heart outflow tract and increase hemodynamic load on cardiac tissues (33, 44, 47, 54). A contrasting blood flow alteration, vitelline vein ligation (referred to as “vein ligation” for simplicity), was used to generate an acute reduction in hemodynamic load by decreasing venous return to the heart (34, 48).

While vein ligation resulted in a consistent flow decrease through the embryonic heart, the degree to which blood pressure and flow velocity increased following banding varied widely. Doppler optical coherence tomography and servo-null pressure characterization of the immediate blood flow response after banding revealed that the altered hemodynamic conditions are defined by the degree of diameter constriction in the outflow tract, also referred as band tightness (Fig. 3) (33, 47). Peak blood flow velocity increased with constriction up to ~35% band tightness, above which it plateaued and decreased. Peak ventricular pressure increased approximately exponentially with constriction. We thus considered five embryo groups according to the level of hemodynamic alteration at HH18: one vein-ligated embryo group (the VVL group) and four outflow tract-banded embryo groups (OTB groups) with different ranges of band tightness that correspond to separate peak outflow tract velocity and peak ventricular pressure conditions (Fig. 3, vertical lines separate groups). The first banded group (OTB 10–20%) exhibited almost normal blood pressures but had increased outflow tract velocity and thus wall shear stress (friction exerted by flowing blood on the endocardial tissue). The second banded group (OTB 21–35%) was characterized by maximum outflow tract velocity and increased wall shear stress, while blood pressure was only moderately elevated. The third group (OTB 36–45%) exhibited maximum outflow tract velocity and increased wall shear stress, as well as high blood pressure levels. In contrast, the fourth group (OTB 46–60%) had reduced outflow tract velocity and wall shear stress (with respect to maximum in the banding intervention) but highly elevated blood pressure levels. As shown later, the level of hemodynamic alteration strongly influences cardiac outcomes.

Fig. 3.

Altered hemodynamics after outflow tract banding. Hemodynamic response to outflow tract band tightness, produced from previously published data (33, 47). Vertical lines outline ranges of constriction used for analysis, and SD of controls are displayed as error bars.

The time period of perturbed flow in this study was limited to 24 h from HH18 to HH24. Two hours after intervention at HH18, peak blood flow velocity in the outflow tract was significantly increased in the banded group and significantly decreased in the vein-ligated group compared with control embryos. By HH24, peak velocities in the outflow tract remained significantly elevated in the banded embryos with the band in place with respect to controls but returned to control levels after band removal. Similarly, peak velocities in vein-ligated embryos were restored to control levels at HH24 (Fig. 4).

Fig. 4.

Maximum blood velocity after interventions. Interventions performed at HH18; band removed at HH24. Constriction of banded embryos ranged between 21 and 52% tightness. SD is displayed as error bars. *P < 0.05, statistically significant differences between experimental and control embryos (n = 8); CON, surgical sham control; VVL, vitelline vein ligated; OTB, outflow tract banded.

Embryo survival and cardiac defects.

We characterized the effects of altered early stage hemodynamics by comparing banded and vein-ligated groups with surgical sham control and normal embryos at HH38 once the heart was four chambered with valves. Rates of embryo survival to HH38 (~12 days of incubation) varied by group (Fig. 5). Roughly half of all embryos in normal (55%) and control (50%) groups survived to HH38, while the percentage of banded embryos that survived to HH38 decreased to ~38% in the band tightness ranges between 10 and 45%. Lowest survival rates corresponded to the highest peak velocity (and therefore wall shear stress) ranges of the banding alteration, OTB 21–35% group, with the proportion of embryos that survived to HH38 in this group significantly below normal survival ($X_1^2$ = 4.07, P = 0.044). Conversely, survival rate increased to within control variability in the tightest banded range (OTB 46–60%; $X_1^2$ = 0.087, P = 0.768) at the point where the peak velocity curve began to decrease while the pressure continued to increase with increased constriction. Vein-ligated embryos also had high survival rates that were comparable with controls ($X_1^2$ = 0.475, P = 0.491).

Fig. 5.

Embryo percentage that survived to HH38 depended on intervention group.

Fully formed hearts were imaged with high-frequency ultrasound and micro-CT in late development at HH38 to determine the effects of varied early hemodynamics on cardiac function and structure (Fig. 6). Major defects included VSD that were further divided according to where the ventricular septum was disrupted, including conoventricular VSD (section of the ventricular septum just below the semilunar valves), perimembranous VSD (upper section of the ventricular septum), and muscular VSD (lower, muscular section of the ventricular septum). Other defects included DORV, TOF, and pharyngeal arch artery (fourth) malformation. DORV and TOF defect combinations were sometimes accompanied with pharyngeal arch artery (PAA) malformation, with persistence of the left 4th aortic arch, stenosis interruption of the left brachiocephalic artery, and tubular hypoplasia of right pulmonary artery. PAA malformation that developed without DORV or TOF included stenosis interruption of the right and left brachiocephalic arteries and pulmonary root and tubular hypoplasia of the right pulmonary artery.

Fig. 6.

Cardiac defects at HH38. Example microcomputed tomography (micro-CT) images of a normal (A, D, G, and J), banded (B, E, H, and K), and vein-ligated embryo (C, F, I, and L), depicting 3-dimensional reconstructions (A–C) and 3 cross-sectional planes (D–L). Plane 1 intersects the semilunar valves (D–F), plane 2 intersects the atrioventricular valves (G–I), and plane 3 intersects the ventricle midpoint (J–L). DORV in the banded embryo displayed with aortic valve rotated outward (E), along with both outflows from the RV and a perimembranous VSD (B and H). The vein-ligated embryo displayed stenosis of the right brachiocephalic artery (C). Examples ultrasound color-Doppler images (M–O) with detection of VSD flow after both interventions (N and O). Scale bars = 1 mm. NL, normal; VVL, vitelline vein ligated; OTB, outflow tract banded; DORV, double outlet right ventricle; VSD, ventricular septal defect.

Importantly, defect incidence and type of major structural cardiac malformation identified at HH38 depended on the distinct hemodynamic conditions in each group (Fig. 7). The resulting cardiac malformation for each banded embryo with its corresponding band constriction is shown in Fig. 8 to further describe the defect dependence on the level of hemodynamic alteration. The number of surviving embryos in each group was between 10 and 16, where the percentage of surviving banded embryos with major cardiac defects remained higher than 50% in all ranges above 20% band tightness, compared with only 10% of hearts with defects in the loosely constricted range (10–20% tightness), 33% in the ligated embryos, 8% in normal embryos, and no control hearts with defects. Defect incidence also varied with band tightness. DORV mostly occurred in hearts with greater than 36% band constriction while perimembranous VSD only occurred in the ranges below 36% band constriction. Contrastingly, the vein-ligated group exhibited a lower overall cardiac defect incidence compared with the banded groups with more than 20% band tightness. Furthermore, unlike the outflow banded groups, the vein-ligated group exhibited PAA malformations without DORV or TOF. Even though TOF and muscular VSD only occurred in a small subset of embryos, the distribution of defects uniquely depended on each embryo group. Thus the various early hemodynamic conditions translated into distinct cardiac phenotypes.

Fig. 7.

Cardiac defects depend on intervention group. A: overall defect incidence among surviving embryos. B: separate defect type incidence among surviving embryos. CON, surgical sham control; VVL, vitelline vein ligated; OTB, outflow tract banded; VSD, ventricular septal defect; CV VSD, conoventricular VSD; PM VSD, perimembranous VSD; M VSD, muscular VSD; PAA, pharyngeal arch arteries malformation; DORV, double outlet right ventricle; TOF, tetralogy of Fallot.

Fig. 8.

Cardiac defects depend on the level of outflow tract band constriction. A: overall defect incidence among surviving banded embryos across individual embryo band tightness. B: separate defect type incidence among surviving banded embryos across individual embryo band tightness. VSD, ventricular septal defect; CV VSD, conoventricular VSD; PM VSD, perimembranous VSD; M VSD, muscular VSD; DORV, double outlet right ventricle; TOF, Tetralogy of Fallot.

Additionally, TDI was used to identify impaired pumping function. TDI recorded systolic (S′) and early diastolic (E′) peak radial and longitudinal myocardial wall velocities at the mitral lateral annulus (39). As stroke volume is dependent on myocardial contractile function (51), our TDI analysis showed that longitudinal E′ and S′ was positively correlated with cardiac output across all embryos, most notably in DORV and TOF hearts (Fig. 9, A and B).

Fig. 9.

Functional comparisons from in vivo ultrasound data and micro-CT evaluations. A: tissue Doppler Imaging (TDI) longitudinal E′ (A) and S′ velocities (B) for all nondefected embryos. C: left ventricle (LV) stroke volume with respect to LV sphericity for all nondefected embryos.

Cardiac defects following interventions produced diverse functional and structural outcomes. This is likely due to the cyclical relationship between altered hemodynamic forces and subsequent cardiac remodeling. Defects that result from the initial hemodynamic changes impact blood flow patterns, which in turn influence further anatomic changes.

Hearts with no major defects.

Functional and structural parameters measured from hearts with only minor structural defects were compared to determine the more subtle effects of altered hemodynamic forces induced by outflow tract banding and vitelline vein ligation alone and when unaccompanied by secondary flow modification. Despite having no major defects at HH38, functional anomalies persisted in the embryos that were hemodynamically altered. These results are summarized in Table 1. Hearts that were banded in early development and did not exhibit major cardiac defects showed significant changes in function (peak left ventricular outflow tract flow velocity and ventricular wall radial velocities), as well as structure (relative ventricular wall thickness, compact myocardium thickness, and septal thickness). These anomalies also uniquely occurred in response to the various ranges of band tightness. Similarly, vein-ligated embryos displayed a separate abnormal response with altered LVEF, ventricular septal thickness, and aortic and pulmonary root diameters.

Table 1.

Functional and structural parameter evaluation of hearts without defects

Parameter	NL	CON	OTB 11–20%	OTB 21–35%	OTB 36–45%	OTB 46–60%	VVL
Blood flow parameters, n	11	11	9	6	5	4	8
Heart rate, beats/min	156 ± 17	151 ± 23	154 ± 40	160 ± 32	147 ± 31	168 ± 24	182 ± 49
Stroke volume, mm3	9.7 ± 2.1	9.6 ± 2.5	10.2 ± 5.8	8.1 ± 1.9	9.7 ± 4.2	10.9 ± 3.0	9.2 ± 4.5
Cardiac output, ml/min	1.5 ± 0.4	1.4 ± 0.4	1.6 ± 1.0	1.3 ± 0.4	1.4 ± 0.6	1.8 ± 0.6	1.9 ± 1.2
LVEF, %	53.4 ± 9.9	64.4 ± 16.1	56.6 ± 15.1	52.6 ± 2.1	56.5 ± 17.0	62.2 ± 6.4	70.9 ± 12.4*
Peak velocity, mm/s	423 ± 120	414 ± 166	425 ± 191	411 ± 170	369 ± 167	669 ± 209*	378 ± 142
Velocity time integral, mm	34.8 ± 7.7	38.6 ± 17.2	34.7 ± 16.0	33.7 ± 14.5	31.7 ± 14.7	53.5 ± 22.8	35.9 ± 12.0
TDI radial annular velocities, n	11	9	7	4	4	3	7
E′, mm/s	27.7 ± 14.6	22.6 ± 9.0	32.8 ± 22.2	25.3 ± 17.8	21.4 ± 14.3	17.6 ± 3.2*	21.7 ± 8.1
S′, mm/s	16.5 ± 6.9	13.2 ± 5.1	20.1 ± 11.5	17.4 ± 9.7	13.5 ± 3.3	12.5 ± 2.4	18.3 ± 7.4
TDI longitudinal annular velocities, n	3	2	2	1	0	3	2
E′, mm/s	25.7 ± 6.0	32.7 ± 3.1	35.1 ± 2.1	14.9		24.8 ± 3.3	35.8 ± 12.6
S′, mm/s	10.2 ± 2.0	17.2 ± 2.5	16.0 ± 0.7	8.3		13.1 ± 2.3	26.5 ± 2.2
LV structure, n	10	14	9	7	6	5	10
Relative wall thickness	0.58 ± 0.08	0.61 ± 0.13	0.58 ± 0.11	0.52 ± 0.13	0.65 ± 0.10	0.47 ± 0.09*	0.57 ± 0.14
Compact myocardium thickness, µm	257 ± 34	241 ± 38	231 ± 44	229 ± 51	231 ± 48	159 ± 20*	226 ± 57
Trabecular myocardium thickness, µm	160 ± 51	149 ± 53	158 ± 47	152 ± 97	164 ± 27	158 ± 63	146 ± 62
Sphericity	2.03 ± 0.14	2.06 ± 0.14	2.05 ± 0.21	2.08 ± 0.27	2.01 ± 0.24	1.99 ± 0.32	1.97 ± 0.19
RV structure, n	10	14	9	7	6	5	10
Relative wall thickness	0.73 ± 0.08	0.77 ± 0.12	0.82 ± 0.12	0.73 ± 0.19	0.94 ± 0.43	0.64 ± 0.07*	0.81 ± 0.10
Compact myocardium thickness, µm	168 ± 44	182 ± 57	177 ± 45	178 ± 34	136 ± 27*	118 ± 30*	161 ± 55
Trabecular myocardium thickness, µm	164 ± 40	195 ± 54	183 ± 30	218 ± 60	168 ± 60	159 ± 10	201 ± 74
Sphericity	1.55 ± 0.09	1.60 ± 0.27	1.59 ± 0.22	1.50 ± 0.20	1.54 ± 0.19	1.61 ± 0.46	1.66 ± 0.18
Septal thickness, µm	653 ± 86	635 ± 104	616 ± 68	536 ± 47*	623 ± 150	547 ± 90	534 ± 133*
Relative great vessel diameters, n	10	14	9	7	6	5	10
Aortic root	0.50 ± 0.03	0.52 ± 0.04	0.50 ± 0.05	0.46 ± 0.06	0.48 ± 0.03	0.44 ± 0.08	0.45 ± 0.07*
Pulmonary root	0.50 ± 0.03	0.48 ± 0.04	0.50 ± 0.05	0.54 ± 0.06	0.52 ± 0.03	0.56 ± 0.08	0.55 ± 0.07*
Right aortic arch	0.40 ± 0.09	0.47 ± 0.10	0.45 ± 0.06	0.47 ± 0.14	0.44 ± 0.06	0.52 ± 0.08	0.42 ± 0.11
Right brachiocephalic artery	0.40 ± 0.06	0.39 ± 0.07	0.45 ± 0.07	0.37 ± 0.11	0.43 ± 0.06	0.42 ± 0.09	0.43 ± 0.10
Left brachiocephalic artery	0.35 ± 0.05	0.35 ± 0.04	0.39 ± 0.06	0.35 ± 0.08	0.37 ± 0.10	0.36 ± 0.08	0.38 ± 0.07
Right pulmonary artery	0.43 ± 0.08	0.41 ± 0.10	0.46 ± 0.11	0.41 ± 0.10	0.49 ± 0.11	0.48 ± 0.07	0.48 ± 0.12
Left pulmonary artery	0.46 ± 0.09	0.50 ± 0.05	0.59 ± 0.07*	0.60 ± 0.11*	0.56 ± 0.13	0.53 ± 0.05	0.53 ± 0.10

Values are means ± SD. NL, normal; CON, control; OTB, outflow tract banded groups; VVL, vitelline vein ligated; LV, left ventricular; RV, right ventricular; LVEF, left ventricular ejection fraction.

^*P < 0.05, significant difference from control groups.

A sphericity index was used to define ventricle shape, which highlights alterations in muscle fiber architecture (2, 50) and pumping efficiency (10). Hearts with no major structural defects across all groups displayed a loose correlation between stroke volume and left ventricle sphericity index (Fig. 9C), where decreased sphericity (more globular heart) was associated with decreased stroke volume. While examples of elongated and globular hearts developed in most experimental groups, differences in the overall sphericity indexes compared with control indexes were not significant.

DISCUSSION

Hemodynamic intervention models were used to mimic perturbed blood flow conditions in human embryos that can be induced by genetic anomalies, teratogens, or irregular placental or vitelline circulations. In this study, the short initial period of altered hemodynamics (≤1 day) during tubular heart stages triggered a detrimental growth and remodeling cascade that eventually led to major cardiac malformations. These results further support the notion that abnormal hemodynamics may be the main regulator of malformation in many CHD cases that develop following transient exposure to teratogens or other environmental conditions that are known to alter blood flow, considering that similar heart defects were induced after only a short-term developmental period of perturbed blood flow. Previous studies have reported cardiac malformations and abnormal remodeling in a number of embryonic heart models following blood flow perturbations (34); however, the degree of hemodynamic change was not assessed. This study is the first to show that the level of hemodynamic alteration predicts distinct abnormalities in the maturing heart.

Even hearts that did not show overt structural anomalies exhibited more subtle structural (myocardium thickness, septal thickness, and great vessel diameter) and functional (peak velocity, E′, and LVEF) challenges, which depended on early hemodynamic modifications with varied and distinct dependence on intervention group. Thus, even in seemingly normal hearts, exposure to altered blood flow in early development may increase the risk for cardiac dysfunction and cardiovascular disease later in life. These subtle challenges were especially prevalent in the most severely constricted banded group (OTB 46–60%). This group had a decreased diastolic radial wall velocity (E′) and increased aortic outflow velocity. The latter was combined with decreased aortic arch diameter and therefore increased aortic wall shear stress, while stroke volumes remained similar to controls. Since ventricular wall thickness is decreased in this same group, the reduction in E′ is perhaps associated with less compliant walls that inhibit chamber filling. The OTB 46–60% group had a decrease in posterior ventricle wall thickness while ventricle cavity diameter remained similar to controls. Lateral compact myocardium thickness also decreased in both OTB 36–45% (right ventricle) and 46–60% (left and right ventricles) groups. The compact myocardium adapts to physiologically increased load by thickening during development, initially by increased cell proliferation and then followed by compaction of trabeculae and invasion of vasculature from the epicardium (a cell layer that covers the external myocardium after cardiac looping) (31, 41, 45). Multilayered compact myocardium is important for force generation, and noncompaction of the myocardium (thin compact myocardium syndrome) presents severe consequences for the developing heart (49). While compact myocardial thickness decreased with tight bands in this study, trabecular thickness remained similar to controls, suggesting that the differences in compaction were not due to disturbed trabeculae compaction. Previous works have shown that continuous banding between HH21 and HH34 results in compact myocardium thickening and advanced trabeculation (46); however, our study suggests that the increased load in tightly banded embryos between HH18 and HH24 interferes with cell proliferation and epicardial vasculature invasion processes that contribute to normal myocardium compaction. The decrease in compact myocardium thickness was not accompanied by any changes in septum thickness. Nevertheless, the septum thickness was significantly decreased in the OTB 21–35% and vein-ligated groups. Vein-ligated embryos without major structural defects had an elevated LVEF compared with controls, which has been previously linked to high mortality and sudden cardiac death in humans (4). Since no hypertrophy (increase in cell volume) was observed in vein-ligated hearts, this compensation is likely due to input/output impedance mismatch (3). Even in hearts without major structural defects, deviations from normal physiology and morphology strongly depend on the level of early hemodynamic alteration.

The relative diameters of the pharyngeal arch arteries were measured as an index of altered blood flow. The only significant difference in banded group hearts without overt malformations was an increased left pulmonary artery diameter in both 10–20 and 21–35% tightness groups. Vein-ligated embryos displayed decreased aortic arch base and increased pulmonary root base diameter compared with controls. These results indicate that the distribution of flow to the arch arteries depends on the early hemodynamic intervention performed.

Most importantly, this study demonstrates a “dose response” type of relationship between mechanical stimuli and the expression of cardiac phenotypes, where in malformed hearts the frequency and severity of cardiac defects also depended on the level of flow alteration. The highest incidence of DORV occurred in the hemodynamic groups corresponding to the highest pressure, which perhaps is less commonly achieved in humans and reflects the relatively rare occurrence of DORV in 1–3% of babies with CHD (1). VSD, a common defect that affects ~50% of children with CHD (35), also frequently occurred in our study with the highest incidence in the OTB 21–35% tightness group. The incidence of TOF, a complex CHD that appears in ~1 in 3,000 live births (13), was also low in our studies. PAA defects not associated with DORV or TOF, in contrast, only developed in the vein-ligated group. Not only did the overall malformation incidence depend on intervention group, each defect phenotype uniquely occurred in response to separate hemodynamic alterations.

While this study did not investigate the disturbed mechanistic pathways that govern the various defect phenotypes, our results emphasize the need for future research to elucidate the initial mechanotransduction signaling changes induced by distinct levels of abnormal blood flow, which then lead to the development of specific cardiac defects. Most cardiac defects found in this study (TOF, DORV, and PAA) result from anomalous outflow tract septation and/or alignment. From the time we perturbed hemodynamic conditions (looped tubular heart stages) to the time the heart is fully formed with four chambers and valves, important developmental processes take place in the heart outflow tract. Among them, neural cardiac crest (NCC) and secondary heart field (SHF) cells migrate to the outflow tract, where they differentiate and proliferate; furthermore, endothelial-mesenchymal transition (EMT), in which cells delaminate from the endocardium and invade the adjacent tissue, occurs in the outflow tract cushions. These events populate the outflow tract and its cushions with cells and ensure proper development of outflow tract derived valves, a portion of the interventricular septum and the aortic and pulmonary trunks. Alterations in any of these processes lead to heart defects. For example, impaired activity of NCC cells in the outflow tract leads to PAA and DORV, while impaired migration of NCC cells into the outflow tract leads to PAA (28); ablation of the right side of the SHF cells results in pulmonary stenosis or atresia, which is one of the features of TOF (12); and altered EMT in the outflow tract cushions leads to failed alignment and DORV (38). Previous studies have reported that altered hemodynamic conditions affect EMT in the outflow tract portion of the heart (32) and our own (33a) results suggest that the degree of EMT perturbation depends on the degree of flow alteration in banded embryos, thus supporting the finding that defect incidence and phenotype uniquely occurred in response to distinct hemodynamic perturbations. Whether proper induction, migration, and differentiation of NCC and SHF cells are also compromised with altered hemodynamic conditions is unknown and requires further study. The fact that independent alterations in SHF, NCC, and EMT lead to the same cardiac defects found by hemodynamic alterations in this study suggest that the observed cardiac anomalies in this study resulted from complex mechanisms, which perhaps link all these developmental processes together.

In conclusion, our data emphasize that altered hemodynamics and cardiac defects are linked, with cardiac malformations dictated by early blood flow patterns. Future therapies for CHD should address the altered blood flow conditions produced alone or as part of genetic anomalies and environmental exposure, as the adverse outcomes of congenital heart defects may be largely propagated by the associated blood flow conditions. This finding suggests the need for new avenues for prevention, early diagnostics, and treatment of cardiac defects.

GRANTS

This work was funded by National Heart, Lung, and Blood Institute Grants R01-HL-094570 (to S. Rugonyi) and HL-129684 (to M. Midgett) and the American Heart Association Grant 16PRE31180006 (to M. Midgett).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

M.M. performed experiments; M.M. analyzed data; M.M., K.L.T., and S.R. interpreted results of experiments; M.M. prepared figures; M.M. drafted manuscript; M.M., K.L.T., and S.R. edited and revised manuscript; M.M., K.L.T., and S.R. approved final version of manuscript.