The trisomy 16 (Ts16) mouse has for a long time been regarded as a useful animal model for the study of malformations in Down's syndrome (human trisomy 21). The characteristic cardiac malformations in this syndrome are found in the atrioventricular junction (Pexieder et al., 1981). In the Ts16 mouse, however, in addition to atrioventricular septal defects, a spectrum of outflow tract (OFT) and aortic arch anomalies has been described which is not typically associated with Down's syndrome. These anomalies include persistent truncus arteriosus (PTA), overriding aorta, double outlet right ventricle (DORV), interrupted aortic arch, and right aortic arch (Miyabara et al., 1982; Bacchus et al., 1987). These OFT malformations more closely resemble the cardiac lesions in patients with DiGeorge syndrome (Anderson et al., 1998). This syndrome is caused by submicroscopic deletions of the so-called DiGeorge critical region (DGCR) 22q11.21–q11.23 on human chromosome 22. The DGCR is syntenic to the proximal portion of the murine chromosome 16 (Nadeau, 1989).
The development of the OFT is intimately related to migration of neural crest cells into the endocardial ridges (Kirby et al., 1985; Conway et al., 1997; Waldo et al., 1998). Although disturbance of neural crest cells is suspected as the primary insult in DiGeorge syndrome, very little is known about the early developmental mechanisms leading to the OFT malformations. Unique neural crest-specific markers in the mouse are not available. However, previous studies in chicks and rats have shown that the expression of alpha-smooth muscle actin (SMA) in the endocardial ridges of the OFT is indicative of the presence of neural crest-derived cells (Beall and Rosenquist, 1990; Ya et al., 1998). The recent description of a transgenic mouse line in which a truncated Connexin43 promoter-construct is preferentially driving lacZ reporter expression in neural crest-derived cells (Lo et al., 1997; Waldo et al., 1999), enabled us to further establish the relationship between the presence of neural crest-derived cells and the expression of SMA in the mesenchymal tissues of the OFT. We sought to evaluate the role of the neural crest-derived cell population in the pathogenesis of the OFT malformations in the Ts16 heart at the earliest stages of aorticopulmonary septation, thereby establishing Ts16 as a model for the anomalies seen in DiGeorge syndrome.
To obtain normal disomic as well as embryos trisomic for chromosome 16, male mice carrying two Robertsonian translocations [Rb(6.16)24LubxRb(16.17)7Bnr]F1 (JAX Mice, The Jackson Laboratory, Bar Harbor, ME) were mated with Wild-type C57Bl/6 female mice (JAX Mice, The Jackson Laboratory, Bar Harbor, ME). Noon of the day the vaginal plug was observed was counted as embryonic day (ED) 0.5. After sacrificing the pregnant female, the uterine horns were isolated in sterile PBS and the embryos were separated; each was then transferred to individual culture dishes. The yolk sacs were collected in medium M199 (Gibco) supplemented with penicillin and streptomycine, containing 0.1 μg/ml colcemide, and then incubated for 1–2 hr at 37°C. Next, the yolk sacs were soaked in a hypotonic sodium citrate (1%) solution for 30 min, fixed in ethanol:acetic acid (1:1 v/v), and stored at 4°C. Chromosome spreads were produced by suspending the yolk sacs in a 60% acetic acid solution (in distilled water), followed by repeatedly pipetting drops of the suspension onto preheated clean (washed in 100% ethanol) microscope slides. After staining with Propidium Iodide (1:100 in distilled water), the chromosome spreads were inspected using a Biorad MRC 1000 scanning laser confocal microscope.
Immunohistochemistry was essentially performed as described previously (Huang et al., 1998). After isolation, mouse embryos were fixed in Amsterdam's Fixative (methanol:acetone: acetic acid:water = 35:35:5:25 v/v), dehydrated in a graded series of ethanol, cleared in toluene, and embedded in Polyfin (Polysciences Inc., Warrington, PA). The specimens were serially sectioned (5 μm thick sections) and mounted on aminoalkylsilane coated slides (Sigma Diagnostics, St. Louis, MO). For the detection of monoclonal antibody binding on selected mouse tissue sections, the indirect conjugated method was applied. After deparaffination in xylene, and rehydration through a series of ethanols, the sections were pretreated for 30 min with hydrogen peroxide (3% v/v in phosphate buffered saline (PBS), pH = 7.4) to reduce endogenous peroxidase activity, and preincubated with TENG-T (10 mM Tris, 5 mM EDTA, 150 mM NaCl, 0.25% gelatin, 0.05% Tween-20, pH = 8.0) for an additional 30 min to minimize nonspecific binding of antibodies. Monoclonal antibodies were diluted in PBS/BSA/NRS (PBS with 1% bovine serum albumin [BSA], and 2.5% normal rabbit serum [NRS]). After incubation of the pretreated sections with the primary antibodies overnight, the antibody binding was detected with rabbit-anti-mouse-peroxidase (RAM-Po; Sigma, A-9044), also diluted (1:400) in PBS/BSA/NRS. All incubations were followed by three washes with PBS for 5 min. The immunocomplex formed was visualized using a Metal Enhanced Diaminobenzidine (DAB) Substrate Kit (Pierce, Rockford, IL; product number 34065). Sections were mounted in Accu-Mount 60 (Stephens Scientific, Riverdale, NJ). The monoclonal antibodies used included MF-20 (Developmental Studies Hybridoma Bank) recognizing atrial and ventricular myosin heavy chain, and anti-alpha-smooth muscle actin (Sigma, A2547). Digital images were collected using a Polaroid Digital Microscope Camera mounted on an Olympus BX40 microscope.
Colocalization of pCx43-lacZ and Alpha-Smooth Muscle Actin
Heterozygote transgenic mouse embryos from the pCx43/lacZ line, previously shown to express the lacZ reporter gene preferentially in neural crest-derived cells under control of a truncated Cx43 promoter construct (Lo et al., 1997; Waldo et al., 1999a) were, after being stained as whole-mounts in X-gal solution, serially sectioned. Sister sections were either stained with nuclear fast red or immunohistochemically processed to detect either alpha-smooth muscle actin with the monoclonal antibody A2547 (Sigma), or atrial myosin light chain, with a previously described polyclonal antibody (Kubalak et al., 1994).
The Outflow Tract in the Ts16 Mouse is Characterized by a Spectrum of Outflow Tract Malformations
Thirty-four normal embryos ranging from ED 8.5 to ED 14.5 and 26 Ts16 embryos ranging from ED 10 to ED 14.5 were investigated. Six of the Ts16 specimens (23%) showed a phenotype consistent with DORV, recognizable as early as ED 12, including one with Tetralogy of Fallot-like anatomy. Four Ts16 specimens (15%) had a phenotype consistent with PTA, recognizable as early as ED 12.5. Arch dominance was recognizable from ED 10.5 onward. Twelve of the Ts16 specimens (46%) had a dominant right aortic arch, whereas five Ts16 specimens (19%) were found to have a dominant left aortic arch. In the remaining embryos arch dominance could not be determined.
Alpha-Smooth Muscle Actin Expression Colocalizes With Neural Crest-Derived Cells
Transgenic pCx43/lacZ embryos express, under regulation of a truncated Cx43 promoter element, the reporter lacZ construct preferentially in neural crest derived cells (Lo et al., 1997; Waldo et al., 1999a). In addition, extensive studies on outflow tract development by Ya and colleagues strongly suggest that smooth muscle actin-positive cells in the endocardial ridges of the outflow tract also represent a population of neural crest derived cells (Ya et al., 1997, 1998a,b). In order to establish that both markers represent the same population of cells, transgenic pCx43/lacZ expression and SMA immunoreactivity was closely compared in transgenic and Wild-type embryos at different gestational stages. In Figure 1a,b it is demonstrated that at 10 ED both pCx43/lacZ (Fig. 1a; transgenic animal) as well as SMA (Fig. 1b; stage-matched Wild-type embryo) mark the same set of mesenchymal cells migrating into the endocardial ridges of the outflow tract. The serial sections shown in Figure 1c,d were derived from an ED 13.5 transgenic embryo. These panels show intense pCx43/lacZ expression in the population of SMA-expressing mesenchymal cells in the fused endocardial ridges in the OFT (note: the lacZ staining interferes slightly with the SMA immunostaining; wild type animals show more intense SMA staining). From these observations we conclude that SMA is a legitimate marker for tracing (the majority of) neural crest-derived cells into the endocardial cushion tissues of the developing murine heart.
Alpha-Smooth Muscle Actin Expression in the Outflow Tract and Development of the Aorticopulmonary Septum of the Normal Murine Heart
At ED 8.5-9, SMA is abundantly expressed in the myocardial component of the heart (Fig. 2a). No SMA-positive cells are found either in the endocardial OFT ridges or in the atrioventricular cushions. The roof of the aortic sac between the foregut and the distal OFT is also negative for SMA at this time (Fig. 2a). Midline in the roof of the aortic sac, i.e. the area in which the aorticopulmonary septum (APS) will eventually develop, a small group of SMA-negative condensed mesenchymal cells is observed.
At ED 9.5–10 the first SMA-positive cells appear in the anlage of the APS as well as in the distal portion of the endocardial OFT ridges (Fig. 2b,e). In the region of the future APS, underneath the SMA-positive cells, an area of condensed, SMA-negative mesenchyme, is again seen. This midline group of mesenchymal cells is even more prominent at ED 10–10.5 (Fig. 2c,f). The heart at this stage shows a pronounced D-loop, the junction between the future right ventricle and proximal portion of the OFT being curved sharply along the rightward side (Fig. 2b,c). The aortic arch arteries and their derivatives (viz. the aorta, pulmonary trunk and the ductus arteriosus) are positive for SMA from ED 10 onward (Fig. 2c,d). It is important to note that the SMA staining around the dorsal aortae is a result of the presence of non-neural crest derived smooth muscle cells (see also Waldo et al., 1999).
At ED 10–11, more SMA-expressing mesenchymal cells in the OFT are seen and the cohorts of SMA-positive cells now penetrate deeper (i.e. more proximal) into the endocardial ridges (Fig. 2c,d,f,g). The APS is now a more prominent structure and protrudes ventrally into the aortic sac. The leading edge of the APS is heavily populated with SMA-expressing cells (Fig. 2c,d,f,g); its core, however, still consists of a condensed, SMA-negative, mesenchyme.
At ED 11.5–12 the prominent SMA-positive APS begins to fuse with the distal OFT ridges, resulting in the separation of an equally-sized pulmonary trunk and ascending aorta (Fig. 3a–c). Within the aortic sac, the APS is positioned in a pronounced angle, running from the superior rightward/ventral aspect to the inferior leftward/dorsal aspect (Fig. 3c,d). This alignment is responsible for the separation of the presumptive semilunar valves at different levels, with the developing pulmonary valve more distal in the OFT than the aortic valve. The position in which the ventral edge of the APS approaches the cushion tissue in the OFT is dictated by the axis of the OFT at its junction with the aortic sac. Within the dorsal aspect of the APS, at the junction with the leftward ridge, the SMA-positive cells are aligned with the axis of the OFT (Fig. 3d). Also notable is the rightward interface between myocardium and arterial walls; the outer layers of the developing vessel walls are primarily SMA-negative except for their subendothelial aspect. The rightward edge of the myocardium also does not extend as far distally as the leftward edge.
By ED 13 the presumptive semilunar valves approach their final positions relative to the ventricles (Fig. 4d). The aortic valve is caudal, posterior, and rightward of the pulmonary valve, but is not yet in continuity with the mitral valve. A prominent feature at this stage is the heavy band of SMA-positive cells in the developing conal septum. This mesenchymal outlet septum, formed by fusion of the proximal (or conal) OFT ridges, gradually becomes myocardialized as protrusions from the flanking myocardium align with the SMA-positive mesenchymal cells and cell-cell interactions are observed (Fig. 4a–c).
At ED 14, the conal septum is heavily, albeit not completely, myocardialized (Fig. 5a,b). The mesenchymal portion still expresses SMA in its central portion; there is, however, a portion of the mesenchyme which is SMA-negative (arrow in Fig. 5a). A myocardial cuff still separates the aortic valve from the LV, and the ostia of the coronary arteries are easily seen.
Abnormal Alpha-Smooth Muscle Actin Expression in the Outflow Tract and Perturbed Development of the Aorticopulmonary Septum in the Trisomy 16 Mouse
A spectrum of abnormalities is observed in our collection of Ts16 embryos. Although phenotypic classification in the older specimens is relatively easy, a phenotype can not always be assigned to the trisomic embryos at the earlier stages of development (before ED 12.5) in which abnormal SMA patterns are found. Hence, because of this, and the variety of malformations observed, the Ts16 specimens are being described on an individual basis and not, as is the case with the normal specimens, as developmental groups with generalized features.
In a severely affected Ts16 embryo at ED 10–10.5, SMA-positive cells are detected neither in the mesenchyme of the OFT nor in the presumptive region of the aortic arches (Fig. 6a). The SMA-positive myocardium of this particular specimen appears irregular and thin. The outflow tract of this specimen looks straight and elongated. Interestingly, a straight outflow tract is characteristically seen in chick neural crest ablated embryos (Waldo et al., 1999b).
At ED 11–11.5, Ts16 specimens present with a significantly underdeveloped APS associated with a dilated aortic sac (Fig. 6b,d). The SMA-expressing mesenchymal cells in the distal cushions of these specimens are not equally distributed, with a predominance of SMA-positive cells on the right-sided ridge (Fig. 6c,d). In several specimens, an irregular group of SMA-negative condensed mesenchymal cells is seen on the rightward and ventral aspect of the interface between the OFT myocardium and arch arteries (Fig. 6b).
At ED 12–12.5 the cardiac abnormalities of Ts16 specimens can be more easily discerned (Fig. 7). A specimen with DORV morphology (Fig. 7b) shows several abnormal features. In the OFT, the ventral surface has a dense pattern of SMA mesenchymal cells in a midline position (Fig. 7a). These cells do not extend as far into the proximal portion of the OFT and do not show normal axial alignment. Instead, they align laterally with the myocardium of the distal OFT (Fig. 7a,c). The distal cuff of the myocardium is irregular, especially along the left side. The APS is aligned in a vertical plane (Fig. 7b) as opposed to the usual angled plane (cf. Fig. 3). Secondly, it also fails to fuse with the distal OFT cushions except for at the most dorsal aspect (Fig. 7c). The OFT cushions are irregular; they are scalloped especially along the leftward ridge and there is underdevelopment of the right sided ridge. Other malformations include a dominant right aortic arch, a right-sided ductus arteriosus (Fig. 7c), and an abnormal curvature of the heart, reflected in the fact that the transition from RV to OFT is fairly straight and midline.
A specimen at ED 12.5 with features characteristic of PTA with no APS and no left-sided aortic arch structures is shown in Figure 8a–c. A small SMA-positive ridge (Fig. 8b) located on the dorsal aspect of the OFT most likely represents the dorsal column of condensed mesenchyme. The leftward position of the ventral column of SMA-positive cells is markedly abnormal. This ventral group of SMA-positive cells appears to be interacting with the adjacent OFT myocardium (Fig. 8b). There is a right-sided aorta (Fig. 8a) and the proximal aspect of the hypoplastic right sixth aortic arch can be seen giving rise to the pulmonary arteries (Fig. 8a,b). The pulmonary trunk is extremely hypoplastic and has no valvar tissue associated with it. The distal ridges of this specimen are quite irregularly scalloped and there is possibly a single truncal valve (Fig. 8c).
At ED 13–14 other PTA specimens are seen (Fig. 8d–f). Each of these has a dominant right aortic arch (Fig. 8d) and a right-sided sixth aortic arch associated with a hypoplastic pulmonary trunk arising from the posterior dorsal surface of the OFT above the level of a single “truncal” valve (Fig. 8d,e). The ED 13 specimen illustrates underdeveloped dorsal and ventral prongs of SMA-positive condensed mesenchyme above the level of the abnormally developing truncal valve (Fig. 8d). The prongs are oriented abnormally and likely contribute to an underdeveloped and malaligned APS. The ED 14 specimen illustrated in Figure 8e,f is characterized by the absence of an APS and by the presence of a quadricuspid truncal valve (Fig. 8f). On the leftward/ventral aspect there is a cluster of SMA-negative condensed mesenchyme within a larger group of SMA-positive mesenchymal cells (Fig. 8e). A ridge, possibly representing the remnant of the dorsal prong of SMA-positive mesenchyme, is seen near the valve level (Fig.8f).
Two specimens with DORV are depicted in Figure 9. The ED 14 specimen (Fig. 9a–c) shows side-by-side, symmetric great vessels without restriction of blood flow to either vessel from the RV. There is a right-sided aortic arch and ductus arteriosus. The APS is strongly positive for SMA, and oriented in a sagittal plane; it does not extend below the level of the semilunar valves. There is no conal septum. The specimen in Figure 9d–f shows a Ts16 heart at ED14.5. This specimen with DORV has a non-myocardialized outlet septum (Fig. 9d) which is malaligned with the ventricles in a vertical plane. The outlet septum has a small defect allowing restricted blood flow to the main pulmonary artery (Fig. 9e,f). This main pulmonary artery is hypoplastic and gives rise to the branch pulmonary arteries and a right-sided ductus arteriosus (Fig. 9e,f). This specimen, if the aorta were to ultimately be committed to some degree to the LV, would morphologically be consistent with TOF.
Neural Crest Cells in Murine Outflow Tract
The evidence for actual migration of neural crest cells into the cardiac OFT in mouse is only slowly accumulating, mainly because of the limited possibilities of performing experimental cell fate studies in murine embryos (Fukiishi and Morriss-Kay, 1992; Lo et al., 1997; Huang et al., 1998; Waldo et al., 1999a). However, the similarity between murine cardiac malformations consistent with neural crest perturbation (Conway et al., 1997) and those avian models known to be neural crest related makes it likely that neural crest-derived cells in the mouse serve the same purpose as in chick (Waldo et al., 1999a). In this study we have used the monoclonal antibody MF20 (Developmental Studies Hybridoma Bank) to delineate myocardial cells and the monoclonal antibody against SMA to visualize a specific subpopulation of mesenchymal cells in the OFT. There are a number of observations that indicate that the SMA-positive mesenchymal cells in the distal OFT are neural crest-derived. First, in the transgenic pCx43/lacZ mouse embryo, known to express the lacZ transporter gene almost exclusively in neural crest-derived cells (Lo et al., 1997: Waldo et al., 1999), lacZ positive cells co-localize with SMA-positive cells in the OFT. Secondly, in the normal specimens the SMA reactivity is mainly found in the condensed mesenchymal prongs in the endocardial ridges of the OFT (Thompson and Fitzharris, 1979); these prongs are known to be neural crest-derived in the chick (Kirby, 1993; Waldo et al., 1998, 1999). Furthermore, no SMA-positive mesenchymal cells are observed either in the mesenchymal atrioventricular cushions or in the mesenchymal cap of the primary interatrial septum. Finally, the SMA-positive mesenchymal cell population is contiguous with the SMA-positive neural crest-derived smooth muscle cell population in the walls of the aortic arches. In addition, the extensive studies on OFT development by Ya and co-workers (Ya et al., 1997, 1998a,b) in the developing rat and mouse also provide strong indications that the cells characterized by SMA expression are neural crest-derived. However, it can not be ruled out that some of the SMA-positive cells in the endocardial cushions of the outflow tract are of non-neural crest origin. In addition, it is also possible that not all neural crest cells are expressing SMA in this region.
The Development of the Aorticopulmonary Septum
This study on aorticopulmonary septation of the murine OFT confirms and further expands the results from earlier studies in the murine (Fananapazir and Kaufman, 1988), rat (Ya et al., 1998), and avian (Thompson and Fitzharris, 1979; Thompson et al., 1983; Kirby et al., 1983; Kirby and Bockman, 1984; Thompson et al., 1987) heart. Hence, the data show that there is a septation complex which forms primarily from the interaction between mesenchymal tissues and primitive myocardium of the OFT. The APS initially develops as a small structure in the “roof” of the aortic sac. As the APS descends into the cardiac OFT, a process very reminiscent of the development of the primary atrial septum, the separation of aorta and pulmonary trunk becomes established. The initial plane of fusion of the APS with cushion tissues of the OFT does not occur in a straight sagittal fashion. The ventral aspect is directed rightward and the dorsal aspect leftward. It is along this plane that the prongs of mesenchyme which stain positive for SMA then seem to migrate into the OFT cushions. These prongs are similar to those described in quail-chick chimera models (Waldo et al., 1998) and likely represent neural crest cells in the mouse mesenchyme. Furthermore, the appearance of SMA-positive cells in the distal OFT cushions and along the developing APS around ED 10–10.5 corresponds well with recent studies (Conway et al., 1997) describing the appearance of Pax3, a marker for neural crest cells, at ED 10.5.
The APS complex moves proximally into the OFT with the SMA-positive prongs leading the descent. The ventral prong remains rightward and the dorsal prong leftward, so that by ED 13 the APS is in a more horizontal plane which is mirrored in the conal septum. It is at this stage that a thick band of neural crest-derived cells aligns side-to-side with projections of myocardial cells within the conal ridges (Fig. 4). The attachments of elongated mesenchymal cells to myocardial cells protruding into the cushion tissues in rat and chick has been described previously (Thompson et al., 1984; Sumida et al., 1989).
We also observed in our normal specimens a subpopulation of SMA-negative, condensed mesenchymal cells midline in the roof of the aortic sac, in the region of the future APS. This group of cells can be distinguished as early as ED 8.5–9, i.e., prior to the appearance of the neural crest-derived cells in the OFT mesenchyme (Fig. 2a). At the earliest stages of APS development the condensed SMA-negative cells typically form the core of the APS, with the leading edge of the APS consisting of SMA-positive cells (Fig. 2b–g). By ED 12, at which point the fusion of the APS with the truncal ridges is well established, no group of clearly SMA-negative cells can be discriminated anymore within the core of the APS. At ED 11.5–12 the walls of the early ascending aorta and pulmonary trunk consist of an inner layer of SMA-positive cells, and an outer layer of condensed SMA-negative cells.
In the normal mouse embryos, SMA-negative condensed mesenchymal cells are also observed in the fusing conal ridges at ED 13 and 14. It is not clear whether this group of cells is related to the group seen at ED 8.5–10.5 in the distal OFT. The existence of a non-neural crest-derived condensed mesenchyme in the conal septum has been recognized before but its origin has never been established (Waldo et al., 1998). Recently, however, it has been shown that epicardially derived cells (EPDCs) accumulate in endocardial cushion tissues (Gittenberger-de Groot et al., 1998; Dettman et al., 1998). It is possible that these EPDC's contribute to this condensed mesenchyme.
Abnormal Organization of Neural Crest-Derived Cells in the Outflow Tract of the Ts16 Mouse and the Pathogenesis of Conotruncal Malformations
When compared to normal embryos, a pronounced difference in number, distribution, and organization of the SMA-expressing cell population was observed in the “anlage” of the APS and in the endocardial OFT ridges of the Ts16 mouse (cf. Ya et al., 1998). This was evident as early as ED 10. The arrival of SMA-positive cells in the APS appeared to be delayed, although it can not be excluded that the lower number of cells could occur from lower division rates. The location of SMA-positive cells in the OFT ridges was abnormal.
The primary abnormality seen in the organization of the SMA-positive cells was in the alignment and extent of the prongs of SMA-positive cells penetrating into the endocardial OFT ridges. In specimens with PTA, the ventral column of SMA-positive cells was considerably larger than the hypoplastic dorsal column and was located on the leftward aspect of the OFT instead of penetrating rightward. In addition, the SMA-positive cells of the ventral prong were disorganized. The hypoplasia of the dorsal column effectively results in the existence of only one prong, which is likely the primary reason for failure of the APS to form. The immunohistochemical data suggest that the hypoplasia of the dorsal column is related to the absence of normal left sided aortic arches. Based on these observations we suggest that either the lack of proper signaling or the lack of appropriate numbers of neural crest cells from the left side is responsible both for the absence of aortic arch structures and for the underdevelopment of the dorsal column.
In our specimens with DORV, the location of the ventral SMA-positive column is midline. Although the dorsal column of SMA-positive cells in these hearts appears to be not as hypoplastic as in cases with PTA (Fig.8), it also does not penetrate normally into the conal cushions The ventral mesenchymal cells do not penetrate proximally into the OFT, but, instead, align laterally with the OFT myocardium at, or immediately below, the level of the developing semilunar valves, inducing ectopic myocardialization. This typical arrangement of the ventral column of SMA-positive cells appears to be responsible for the vertical alignment of the APS. The failure of these cells to penetrate deeply into the conal ridges prevents the formation of a normal conal septum and is likely also responsible for the absence of myocardialization in the conal ridges.
The partly formed mesenchymal outlet septum in the specimen with TOF-like anatomy is vertical, malaligned with the ventricular septum, and restricts bloodflow to the pulmonary artery (Fig. 9d–f). This differs from the other DORV specimens (e.g., Fig. 9a–c) in which there is no outlet septum and both great arteries are completely committed to the RV with unrestricted bloodflow. In this TOF-like case, the abnormality is likely produced if the ventral column descends in a position slightly leftward of midline, and perhaps with the prongs extending slightly deeper into cushions.
In addition, the SMA-negative condensed mesenchyme was not normal in the Ts16 specimens. It was often difficult to identify, but when present it was usually in an abnormal location, such as in the 11 ED specimen (Fig. 6b) where an SMA-negative group of cells was seen on the ventral rightward edge of the OFT. There were also groups of SMA-negative condensed mesenchyme seen in older ED 13–14 specimens in the more proximal OFT which were not in the normal position (Fig. 8e).
Abnormal Myocardialization in the Ts16 Mouse
Neural crest derived mesenchymal cells in the distal OFT and APS display distinct longitudinal arrays during septation in normal hearts but not in trisomic hearts. Our previous descriptive studies of matrix and cell realignment have implicated the onset of longitudinal mechanical tension as a common important feature of outflow septation in chicks, rats and humans (Thompson and Fitzharris, 1979; Thompson et al., 1984, 1985; Sumida et al., 1989), much as we now report in the normal mouse. Our observations in Ts16 suggest that the sparse or tardy immigration of neural crest derivatives lead to perturbation of the typical alignment normally seen, leading to abnormal mechanical tension resulting in malformations as described in this paper. Whereas in normal embryos myocardialization is typically found in the lower region of the conal cushions, in the trisomic situation abnormal myocardialization was observed in the more distal portion of the OFT. What triggers this ectopic myocardialization is not completely clear. It is possible that changes in the organization of the extracellular matrix, possibly induced by altered mechanical tension, play a role in the establishment of this particular phenomenon. It is also possible that the observed abnormalities are intrinsic to the neural crest cell population itself. In this respect it is important to note that it has recently been reported that fibroblasts from Ts16 mice interact with collagen and laminin gels less effectively than normal fibroblasts (Carver, 1998). With the recent development of an in vitro myocardialization assay, in which myocardialization of a collagen gel by cardiac explants can be studied under different conditions (Van den Hoff et al., 1999), some of these aspects can be specifically addressed.
The Development of the Outflow Tract in Ts16; a Model for DiGeorge Syndrome Rather Than for Down's Syndrome
The conotruncal abnormalities observed in the Ts16 mouse include DORV, PTA, interrupted aortic arch, right aortic arch, and overriding aorta (as seen in TOF). Except for DORV these defects are commonly seen in patients with DiGeorge syndrome (Marmon et al., 1984; Van Mierop and Kutsche, 1986; Wilson et al., 1993). An increasing body of evidence points to the involvement of the neural crest in the pathogenesis of DiGeorge syndrome (Farrell et al., 1999; Yamagishi et al., 1999). The DiGeorge Critical Region (DGCR) on human chromosome 22 contains candidate genes like UFD1L (Yamagishi et al., 1999) and HIRA (Farrell et al., 1999), which, when perturbed in animal models, have been shown to result in OFT abnormalities reminiscent of those seen in DiGeorge. The proximal portion of chromosome 16 in the mouse is syntenic to 22q11 and the DGCR in the human (Nadeau, 1989). Interestingly, trisomy 22, a rare condition in human, is also characterized by a high incidence of congenital cardiac malformation, including conotruncal abnormalities (Bacino et al., 1995). This suggests that excess of gene products of certain gene(s) on chromosome 22 as well as haploinsufficiency of the same gene(s) can lead to similar cardiac defects. A similar conclusion was drawn from our studies on the role of Cx43 in OFT development (Ewart et al., 1997; Huang et al., 1998; Lo and Wessels, 1998) in which gain or loss of Cx43 function lead to cardiac defects. Hence, it appears that the candidate genes now proposed to be involved in the pathogenesis of DiGeorge need also to be considered in the dysmorphogenesis of the OFT in the Ts16 mouse.
We agree with previous reports that the Ts16 mouse is not an ideal model for cardiac anomalies in Down's syndrome, despite the frequent occurrence of AV septal defects in both (Anderson et al., 1998; Webb et al., 1997). Given that DORV is not frequently seen in DiGeorge, we believe that it is possible that the relative high percentage of DORV in Ts16 occurs as a result from the combined remodeling and alignment defects seen in the distal outflow tract (or truncus) and AV junction.
The results described in this paper illustrate that the conotruncal anomalies observed in the Ts16 mouse are reminiscent of those seen in DiGeorge syndrome and that they are likely caused by perturbation of both the number, as well as the extent of migration of neural crest cells into the aortic sac and OFT of the developing heart. The proper formation of the APS and the conal septum is likely dependent upon the interaction and alignment of the neural crest cells with other, possibly non-neural crest-derived, mesenchymal cells, and on the interaction with surrounding myocardial cells in the OFT. We therefore conclude that the conotruncal anomalies of the Ts 16 mouse do not simply result from an arrest of normal development, but in fact are a direct result of abnormal neural crest cell interaction with the primitive OFT. Thus, we believe that, in addition to being helpful in the study of some aspects of Down's syndrome, the Ts16 mouse should also be considered as a model for conotruncal pathogenesis in DiGeorge syndrome.
This study was financially supported by NIH Training grant (to B.R.W.), by NIH (to A.L.P., R.R.M., R.P.T., and A.W.), and by AHA (to T.M.) and GIA (SC-Affiliate; to T.M.). The authors thank Kioina L. Myers and Tanya Rittmann for expert technical assistance. The monoclonal antibody MF20 developed by Dr. D.A. Fischman was obtained from the Developmental Studies Hybridoma Bank developed under auspices of the NICHD and maintained by the University of Iowa, Department of Biological Sciences, Iowa City, IA 52242.