Objective: Several mouse models of cardiac neural crest cell (NCC)-associated conotruncal heart defects exist, but the specific cellular and molecular defects within cardiac NCC morphogenesis remain largely unknown. Our objective was to investigate the underlying mechanisms resulting in outflow tract defects and why insufficient cardiac NCC reach the heart of the Splotch (Sp2H) mouse mutant embryo. Methods: For this study we used in vitro cell culture techniques, in vivo mouse–chick chimeras, BrdU cell proliferation labeling, TUNEL labeling to visualize apoptosis and the molecular markers AP-2, Wnt-1 and Wnt-3a to characterize NCC morphogenesis in vivo. Results: Expression of the NCC marker AP-2 revealed an extensive reduction in migratory NCC, however the rates of cell proliferation and apoptosis were unaffected, and do not account for the Sp2H NCC-associated heart defects. Further expression analysis revealed that Wnt-1, but not Wnt-3a, is expressed at decreased levels within Sp2H and that the cardiac NCC fail to undergo normal NC stem cell proliferative expansion prior to migration while still in the neural folds. However, when placed into a wild-type matrix or a tissue culture environment, the Sp2H cardiac NCC could migrate normally. Additionally, this reduced population of Sp2H NC stem cells do migrate properly within the Sp2H environment, as observed by neurofilament expression and cardiac innervation. Conclusions:. Taken together, all these data indicate that the Sp2H defect is intrinsic to the NC stem cells themselves and that there is a decrease in the number of pre-migratory cardiac NCC that form. It appears that this decrease in NCC number is the primary defect that ultimately leads to a lack of a cardiac NCC-derived Sp2H outflow tract septum.
Time for primary review 36 days.
This article is referred to in the Editorial by M.J.B. van den Hoff and A.F.M. Moorman (pages 212–216) in this issue.
Conotruncal heart defects are an important category of congenital disease with a prevalence of four per 10 000 births  and are well-known as part of the DiGeorge and velo-cardio-facial human syndromes [2–4]. Unlike most other cardiac malformations, conotruncal defects appear to have a specific embryonic pathogenesis, and often involves the cardiac neural crest cell lineage [5,6]. Failure of the cardiac neural crest cells (NCC) to populate the conotruncal ridges leads to a lack of fusion and descent of the conotruncal ridges towards the ventricles, resulting in a single undivided outflow tract. Since the conotruncal ridges also participate in formation of the interventricular septum, outflow tract defects are always accompanied by a defective interventricular septum [7,8], resulting in a single outflow tract vessel (called persistent truncus arteriousus — PTA) that exits the heart and receives blood from both the left and right ventricles.
Presently, the majority of data on the embryonic basis of these cardiac NCC-related defects comes from studies of avian systems [6,9] which have limited genetic data, and so are not easily extrapolated to the human situation. However, there are several mammalian models of cardiac NCC-related PTA in which experimental embryological studies can be performed: including Splotch Sp2H[5,10] and Sp alleles ; Patch and PDGFα null receptor ( and Dickman et al., submitted); RXRα; RARα/RARβ and RARα/RXRα double null mutants [15,16]; Nf-1 and c-Jun. Additionally, as all these mouse models incorporate a genetically based etiology, molecular analysis should provide evidence as to the genetic causation of these defects in humans. Despite the different genetic insults of all these mouse mutants, the relatively uniform phenotype of the conotruncal heart defects suggests that they may share a common developmental mechanism. Prior to the colonization of the outflow tract of the embryonic heart, pre-migratory cardiac NC stem cells are initially formed and undergo a proliferative expansion within the dorsal–lateral walls of the forming neural tube (i.e. neural folds). Subsequently, the NCC cells undergo an epithelial-to-mesenchymal transformation, delaminate from the neural tube and initiate migration towards their sites of terminal differentiation. Abnormal levels of cell death and/or cell proliferation, changes in cell fate and altered migration have all been proposed to account for these cardiac NCC-associated abnormal heart phenotypes. However, in the majority of these mutant mouse and chick models, the specific cellular and molecular defects within cardiac NCC development that are responsible for these morphological abnormalities remain largely unknown.
There are six classical and irradiation-induced alleles of Splotch, all of which have mutations within the Pax3 transcription factor: Sp, Sp-delayed, Sp-retarded, Sp1H, Sp2H and Sp4H; which all have mutations within the Pax3 transcription factor . Conotruncal heart defects within the Sp2H mutant mouse result from a primary defect in a failure of the cardiac NCC to colonize the conotruncal ridges of the outflow tract. Using a variety of molecular markers, including Pax3, CRABP1, Hoxa3, and Prx1/2, we have previously demonstrated that the cardiac NCC fail to colonize the cardiac outflow tract in homozygous Sp2H mutant embryos destined to develop PTA . In order to determine why the cardiac NCC do not reach the outflow tract of the Sp2H mutant hearts, we sought to follow the initial development of the NCC, by taking advantage of molecular markers known to be expressed by this cell population during various stages of NCC expansion, emigration and migration. Sp2H mutant NCC survival was assessed by determining whether altered cell proliferation and/or cell death could account for the cardiac NCC-related heart defect. Additionally, mouse–chick chimeras and cell culture techniques were used to determine whether the Sp2H NCC destined for the heart can migrate when provided with a wild-type matrix or a tissue culture environment. Significantly, our results show that the Sp2H mutant cardiac NCC fail to undergo normal NCC expansion, that this reduced population of NCC can migrate normally both in vitro and in vivo given a normal environment and that both cell proliferation and cell death rates are unaltered. Taken together this data suggests that the genetic defect does not reside in the environment through which the cardiac NCC migrate, but within the in Sp2H cardiac NCC themselves.
2.1 Breeding of mice and histological analysis
Splotch (Sp2H) mutant mice are maintained by random breeding on a mixed background comprising approximately 50% CBA/Ca, 25% C3H/He and 25% 101 . A breeding colony resulting from a single pair of Sp2H mice was established at the Medical College of Georgia and after 14 generations of sibling matings, the mice derived from this pair, give rise to offspring that have approximately 91% Sp2H/Sp2H embryos with PTA and associated interventricular septum defects at 13.5 days post-coitum (dpc) (n=34/37 embryos histologically examined had PTA).
Females were paired with males overnight and were checked for vaginal copulation plugs the following morning. Embryos were designated 0.5 dpc on the morning of finding a plug. Pregnant dams were sacrificed by cervical dislocation from 8.5 to 13.5 dpc, uterine horns removed and transferred to PBS (4°C) and embryos subsequently dissected free of the decidua. Tissue was collected for genotyping, and embryos were either used immediately for in vitro and in vivo culture; or fixed overnight in 4% paraformaldehyde/PBS (4°C), dehydrated to 100% methanol through a graded PBT (PBS+0.1% Tween 20) series, and stored at −20°C until used for in situ hybridization and/or apoptosis analysis. The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).
2.2 PCR genotyping
All embryos harvested were genotyped using the polymerase chain reaction (PCR) to identify the 32 base pair deletion present in the Sp2Hmutant allele of Pax3. For this purpose, genomic DNA was extracted from a fragment of embryonic tissue, or from the yolk sac, and PCR was performed using primers and conditions as described previously .
2.3 In vitro neural crest cell cultures
To obtain embryos for cardiac NCC outgrowth cultures, heterozygotes Sp2H mice were mated and embryos collected at 8.5 dpc, staged by counting somite number (only embryos having four to six somites were used as this is before initiation of NCC migration) and cultured essentially as described . The portions of the neural folds containing the ‘cardiac NCC region’, the region from the otic sulcus to the third somite as defined by Kirby et al. , were removed from the embryo and placed in DMEM medium (Cellgro) plus 10% fetal calf serum (Sigma). The rest of each embryo was simultaneously harvested for PCR genotyping. The cardiac NCC region was then treated with 1 mg/ml of collagenase type L (Sigma) in PBS for 5–10 min as the surrounding mesenchyme was removed from the neural folds using dissection with 25 Gauge needles and mechanical disruption with a pulled glass pipette that was slightly larger than the neural fold. Isolated neural folds were washed several times in PBS followed by quenching in 10% fetal calf serum plus DMEM medium, then placed on un-coated clean glass coverslips and flooded with M199 medium (Cellgro) plus 10% fetal calf serum (Sigma). Explants were incubated at 37°C with 5% CO2 for up to 72 h. Several isolated neural folds (n=3) were processed for histology in order to confirm that the neural folds were explanted without any contaminating mesenchymal tissue (see Fig. 1a and b).
Digital images of the cultures were collected at 24, 48 and 72 h after isolation using a SPOT camera (Model #1.1.0, Diagnostic Instruments, Inc.). NIH Image software was used to measure and analyze the area of outgrowth as described previously . This was obtained by taking the area of the entire explant and subtracting from it the area of the dense central neural fold mass. In order to control for differences in migration area due to variations in the shape of the explants, the area of outgrowth (mm2) was divided by the perimeter (mm) of the explant. This normalized number (in mm) is referred to as the migration index and provides an estimate of the migration rate of the NCC. Statistical analysis were carried out using ANOVA.
Immunostaining was used to determine whether the cultured NCC expressed various differentiation markers. Coverslips with the attached NCC were stained with either 3A10 IgG anti-neurofilament monoclonal antibody (1:100 dilution; Developmental Studies Hybridoma Bank) or α-smooth muscle actin (1:5000 dilution; Sigma) using the ABC kit (Vector) following manufacturer's directions.
2.4 Mouse–chick chimeras
Wild-type, heterozygous and homozygous Sp2H mutant neural folds from the cardiac NCC region were isolated as described above, stained for 1 min on ice with the fluorescent marker DiI (Molecular Probes, Eugene, OR) at a dilution of 40 μg/ml in DMEM medium (Cellgro) and then transplanted into stage 9  normal chick host embryos that had undergone cardiac NCC ablation as previously described . The rest of each embryo was simultaneously harvested for PCR genotyping. After transplantation, the egg was sealed and incubated for a further 48 or 72 h before being fixed in 4% paraformaldehyde and photographed using the rhodamine filter on a BioRad MRC-1020 confocal microscope.
All of the mouse–chick chimeras were subsequently processed and embedded in OCT mounting medium (Tissue-Tek) for cryostat sectioning as previously described . Serial transverse sections were cut at 25 μm and collected on slides.
2.5 Whole-mount in situ hybridization and RT-PCR analysis
Sp2H mutant embryos (8.5–10.5 dpc) were analyzed by whole-mount in situ hybridization as previously described , and representative embryos were subsequently processed for sectioning on a vibrating microtome .
Wnt-1 and Wnt-3a cDNA mouse probes were generated by PCR amplification, using primers based on GenBank sequences (M11943 and X56842). Wnt-1-specific primers were designed to amplify the region between base pairs 242–261 and 1165–1185 giving rise to a 943-bp fragment. Wnt-3a-specific primers were designed to amplify the region between base pairs 129–151 and 944–966 giving rise to a 837-bp fragment. Amplified fragments were cloned into pGEM-T vector (Promega) and sequenced to confirm identity.
Wnt-1, Wnt-3a and AP-2 plasmids were linearized with an appropriate restriction enzyme and used as a template to generate both digoxygenin-labeled sense and anti-sense RNA probes according to manufacturer's directions (Genius 3 kit, Boehringer Mannheim).
RNA isolation and RT-PCR analysis was carried out at 27, 30 and 33 cycles using the Wnt-1 and Wnt-3a primers, as described previously .
2.6 Whole-mount cell proliferation and apoptosis
Proliferating cells were detected in serial sections of embryos which had been labeled for 2 h by injection of dam with 0.3 mg BrdU/kg body weight and subsequent detection with anti-BrdU-alkaline phosphatase (Boehringer Mannheim, 1:30). The mean number (±S.E.M.) of cells labeled by BrdU was determined per 100 cells counted in both the mesenchyme adjacent to the neural tube and in the third and fourth branchial arches.
Apoptosis was visualized in whole embryos  using the TUNEL 5′-end labeling technique (Apoptag, Oncor). Washing, developing, and subsequent clearing/dehydration occurred as described for whole-mount in situ hybridization .
2.7 Whole-mount immunohistochemistry
The roof plate of the neural tube/brain of 11.5–13.5 dpc embryos was opened and the epidermal ectoderm was removed prior to fixation (40 min in 1% crude trypsin (Difco) at 4°C), followed by mechanical stripping in DMEM (Cellgro) containing 10% fetal calf serum, in order to allow antibody penetration. Embryos were subsequently fixed in 4% paraformaldehyde overnight. Embryos were then washed in PBT (PBS containing 0.5% Triton X-100) and endogenous peroxidase was blocked by incubation in 0.05% hydrogen peroxide. After washing embryos were incubated with primary antibody at 4°C for 3 days. Primary antibody: 3A10 anti-neurofilament mouse monoclonal antibody (Developmental Studies Hybridoma Bank) at 1:3 dilution (in PBS+10% newborn calf serum+1% Triton X-100+0.02% azide). After removal of primary antibody and washing in PBT+1% normal goat serum, secondary antibody was added and incubated at 4°C overnight. Secondary antibody: peroxidase conjugated goat anti-mouse IgG/IgM antibody (Jackson ImmunoResearch Laboratories, Inc.) used at 1:100 dilution (in PBS+10% newborn calf serum+1% Triton X-100). After washing, color reactions were performed in diaminobenzidine (5 mg/10 ml) and hydrogen peroxide. Embryos were cleared in glycerol for analysis and photographed on a SV11 Zeiss dissecting scope. Several representative embryos were embedded in wax and sectioned as previously described .
3.1 Sp2H cardiac NCC can migrate in vitro
As the NCC-derived outflow tract septum fails to develop in Sp2H mice, the observed phenotypes suggested a possible deficiency of cardiac NCC which may be related to problems in cellular migration. We therefore sought to determine if the environment was hindering the Sp2H cardiac NCC such that migration was affected, or whether the problem was intrinsic to the NCC themselves. To address this question, we isolated the portion of the neural fold (prior to NCC emigration) from the optic sulcus to the third somite which contains presumptive cardiac NCC from 8.5 dpc Sp2H mutants and placed the neural fold explant into culture (n=12 wild-type, 17 heterozygous and 13 homozygous mutant Sp2H). Upon attachment of the explant to the coverslip, the NCC were then able to migrate away from the neural tube explant, mimicking their in vivo pathway [21,29]. Relative migration distances was measured 24, 48 and 72 h after isolation.
In contrast to the limited migratory behavior of Sp2H cardiac NCC observed using molecular markers in vivo , NCC from both control and Sp2H mutant neural folds emerged and migrated away from explanted neural tubes of mutant embryos in an essentially normal manner in vitro within hours of placement in culture. After 24, 48 and 72 h, there appeared to be no overt difference in relative migration (Table 1). Mesenchymal cells from wild-type (Fig. 1c–e), heterozygous (not shown) and homozygous (Fig. 1f–h) mutant explants could be seen at similar distances, suggesting that Sp2H NCC can migrate away from the neural tube explant in a normal fashion relative to wild type littermates and that there is not an intrinsic NCC migratory problem will give rise to the later cardiac NCC-deficient defects. Thus, when provided with a wild-type matrix or a tissue culture environment, the Sp2H mutant cardiac NCC can migrate in vitro.
|Genotype||No. cultures||Migration indexb|
|24 h||48 h||72 h|
|Genotype||No. cultures||Migration indexb|
|24 h||48 h||72 h|
Numbers indicated are mean±S.D.
No statistically significant differences detected by one way analysis of variance (ANOVA) between all three genotypes at each time-point. Compared to control group P>0.05
Moreover, at the end of the 3 days of culture, Sp2H NCC underwent relatively normal differentiation as these mutant cardiac NCC expressed both neurofilament and α-smooth muscle actin markers, in a similar pattern and intensity to both the wild-type and heterozygous cardiac NCC (data not shown).
3.2 Sp2H cardiac NCC can migrate normally in vivo
Because our in vitro migration assays suggested that Sp2H NCC have the ability to migrate in a normal fashion on glass cover-slips, we sought to determine their migratory ability in vivo along the cardiac NCC migration pathway, in order to determine where, if at all, do events become abnormal during their migration to the outflow tract.
To visualize the migration of the cardiac NCC in vivo, we constructed mouse–chick chimeras, which have previously been shown to be a reliable model to assess the developmental potentials of both normal and mutant mammalian cells [30,31]. As NCC are known to follow specific pathways to their sites of differentiation , mouse–chick chimeras were constructed where the chick cardiac NCC neural fold region was replaced by either wild-type, heterozygous or homozygous Sp2H mutant neural fold grafts from the mouse (at the four to six somite stage and before initiation of NCC migration) and marked with a fluorescent dye (Fig. 2). Using confocal microscopy, it was determined that cardiac NCC migrated from the neural folds into the pharyngeal region, from all three genotype grafts (n=four wild-type, seven heterozygous and five homozygous mutant Sp2H).
After 48 h, mouse–chick chimeras with either wild-type, heterozygous or homozygous Sp2H mutant grafts contained labeled cells in a pattern that was consistent with the normal pattern of chick cardiac NCC migration  into the circumpharyngeal region and pharyngeal arches (Fig. 2a–d). Two streams of cardiac NCC are migrating from the mouse graft into pharyngeal arches 3 and 4 of the stage 18 chick embryo. Note that very little or no migration was seen towards pharyngeal arch 2, as cardiac NCC do not normally colonize arches 1 and 2 . Cryostat sectioning revealed that the DiI-labeled mouse NCC cells were migrating via the normal dorsolateral pathway within the chick host embryo (data not shown). Furthermore, after 72 h of incubation, wild-type, heterozygous and homozygous Sp2H mutant grafts gave rise to DiI-labeled cells that were present in arches 3, 4 and 6 (Fig. 2e–f). Due the large size of the chick embryos, and that fact the NCC are dividing as they migrate (thereby diluting the DiI label) — the DiI marker could no longer be accurately followed after 72 h. However, a few DiI-labeled NCC (from wild-type, heterozygous and homozygous Sp2H mutant grafts) were detected within the outflow tract in stage 33 hearts (7.5 day old chick embryos). Additionally, because of variations in graft size and DiI-labeling, it was not possible to quantitatively determine whether there were any absolute differences in the number of migrating cells between the three genotypes. However, there did not appear to be any obvious differences in either the pattern or distance migrated among the three genotypes. Thus, given a relatively normal environment, the Sp2H mutant cardiac NCC appear capable of relatively normal migration in vivo.
3.3 Molecular analysis of Sp2H NCC in vivo
Because our in vitro and in vivo migration assays suggested that Sp2H NCC have the ability to migrate in a normal fashion, we sought to molecularly characterize their migration in vivo, in order to specifically determine where, if at all, cardiac NCC morphogenesis is abnormal.
In order to visualize the migration of the Sp2H cardiac NCC, we had previously shown that these cardiac NCC express a range of different mRNA transcripts during NCC migration and differentiation . To monitor the status of cardiac NCC emigrating from the neural tube and migrating through the extracellular matrix in vivo, we examined embryos for the expression of the AP-2 transcription factor, which is expressed at 8.5 dpc in the dorsal neural plate coincident with NC stem cell formation. By 9.5 dpc AP-2 expression is absent from the neural tube but is still expressed by migratory NCC after emigration from the neural tube and during subsequent migration through the extracellular matrix to the pharyngeal arches . Given that AP-2 null mice show a NCC-deficient phenotypes similar to the Sp2H mice, we questioned whether AP-2 may be mis-expressed in homozygous Sp2H mutants [34,35].
In 9.5 dpc wild-type embryos, streams of AP-2 positive NCC could be seen emigrating from the neural tube and migrating towards pharyngeal arches 1, 2, 3 and 4 and into the face (Fig. 3a, b). AP-2 appears to mark individual NCC emigrating and migrating from the neural tube, and two streams of AP-2 positive cardiac NCC were detected within pharyngeal arches 3 and 4, as well as the circumpharyngeal region (n=8). This staining pattern was consistent with that previously reported for migratory cardiac NCC . In Sp2H homozygous mutant embryos (n=11), fewer AP-2 positive cardiac NCC could be seen emigrating from the neural tube and migrating towards pharyngeal arches 3 and 4. However, those AP-2 positive cardiac NCC that were present, migrate approximately the same distances and along the normal migration pathways as the wild-type littermates. Also note, that the intensity of AP-2 expression was not altered within pharyngeal arches 1 and 2 and within the face, suggesting that the cranial NCC are un-affected within the Sp2H mutant.
In 10.5 dpc wild-type embryos (n=10), similar streams of individually marked AP-2 positive cardiac NCC can seen emigrating from neural tube in two streams and migrating into pharyngeal arches 3, 4 and 6, as well as the circumpharyngeal region (Fig. 3c, d). Both 9.5 and 10.5 dpc heterozygous Sp2H embryos gave identical AP-2 expression patterns to that of their wild-type littermates (n=23). However, in Sp2H homozygous mutant embryos (n=9), significantly fewer AP-2 positive cardiac NCC could be seen emigrating from the neural tube (Fig. 3d), and migrating towards the pharyngeal arches and outflow tract for the heart. Similarly, the distance migrated and pathway taken by the reduced population of Sp2H cardiac NCC was comparable to the wild-type littermates. Vibratome sectioning did not reveal any AP-2 expression within the neural tube (Fig. 3e, f), indicating that unlike the Wnt-1 and Wnt-3a double knockouts , the Sp2H cardiac NCC are not trapped within the neural tube. Since AP-2 expression was observed at the appropriate time and place in Sp2H mutants, these data suggest that mutant cardiac NCC are capable of migrating along the appropriate pathways at the appropriate times in vivo, but do so in significantly reduced numbers.
3.4 Cardiac NCC deficiencies are not due to elevated apoptosis or decreased cell proliferation
As Sp2H cardiac NCC migration can occur normally both in vitro and in vivo, we sought to determine if abnormal cell proliferation, cell death or both could account for the reduced numbers of migrating NCC (as observed by the AP-2 expression patterns) that are ultimately responsible for the Sp2H NCC-deficient cardiac defects. Recently, Pax3 has been shown to be essential for the survival of uncommitted somitic progenitors and inhibition of apoptosis within both the Sp and Sp2H mutant embryos .
We sought to determine if mutant NCC selectively failed to undergo sufficient cell proliferation. Upon comparing serial sections through both the mesenchyme adjacent to the neural tube and the pharyngeal arches of 9.5 dpc wild-type control littermates (23.6±5.0; n=3 embryos), heterozygous (25.3±5.5; n=3 embryos) and homozygous Sp2H mutants (24.8±4.0; n=3 embryos), we observed no overt difference in the mean (±S.E.M.) number of BrdU-labeled cells per 100 cells counted. As there is already an observable reduction of cardiac NCC, these results suggest that there is no overt decrease in cell proliferation that could account for the NCC-deficient phenotype.
Because cell proliferation was unaffected, we sought to determine if elevated apoptosis could account for the observed deficiencies in cardiac NCC using the TUNEL 5′-end-labeling technique. As seen in Fig. 4, apoptosis normally occurs at a low level in arches 1 and 2, in the circumpharyngeal region and within the somites at 9.5 dpc (Fig. 4a, a′). No significant differences in apoptosis were observed, either within the neural tube or along the cardiac NCC migration pathway among the three genotypes (wild-type control littermates n=3, heterozygous n=5, and homozygous Sp2H mutant embryos n=5). In fact, there was a decrease in apoptosis within the circumpharyngeal region of the homozygous Sp2H mutant embryos (Fig. 4b, b′), suggesting that the apoptosis that is normally present is due to NCC cell death. This is consistent with the findings of normal patterns of apoptosis within cranial NCC in chick embryos .
One day later in development at 10.5 dpc, when there is already a dramatic reduction of cardiac NCC, no significant differences were observed along the cardiac NCC migration pathway, between all three genotypes (Fig. 4c, d; wild-type control n=5, heterozygous n=7, and homozygous Sp2H mutants n=9). Taken together, these results suggest that the reduced numbers Sp2H mutant cardiac NCC that migrate along the appropriate pathways at the appropriate times, do survive and proliferate.
3.5 Cardiac NCC expansion is abnormal within the Sp2H mutant neural tube
NCC arise from the dorsal neural tube where they initially reside and undergo expansion (8.0–8.5 dpc). Because NCC migration, proliferation and apoptosis levels appear unaffected within the Sp2H homozygous mutants which have a phenotype consistent with a cardiac NCC-deficiency, we sought to determine if insufficient numbers of NC stem cells were present within the Sp2H neural tube itself, accounting for the reduced numbers of emigrating NCC. To do this, we examined the expression of both the molecular markers Wnt-1 and Wnt-3a, as these cystein-rich secreted signaling molecules have been previously shown to be expressed by NCC around the time of neural fold closure  and may be essential for the initial expansion of NCC in the neural tube [36,41]. Double homozygous null Wnt-1/Wnt-3a mutant embryos have a significant reduction within NCC-derived structures that do not seem to result from abnormalities within NCC emigration, migration, survival or differentiation but rather a lack of NCC expansion within the neural tube .
In 8.5 dpc wild-type embryos, Wnt-1 expression could be seen along the dorsal edge of the neural folds and closed neural tube, and also in a broad band extending dorsal to ventral just anterior to the midbrain/hindbrain junction (Fig. 5a). When 8.5 dpc wild-type (n=6; Fig. 5a) embryos were compared to homozygous Sp2H mutant littermate embryos (n=9; Fig. 5a) for the expression of Wnt-1, there already was a significant decrease in Wnt-1 expression within the homozygous Sp2H mutant embryos (arrow in Fig. 5a). Similarly, when 9.5, 10.0 and 10.5 dpc homozygous Sp2H mutant embryos (n=24) were examined, Wnt-1 expression was always reduced when compared to either wild-type (n=16) or heterozygous (n=21) littermates (Fig. 5b, c). Vibratome sectioning revealed that the Wnt-1 expression was restricted to a more dorsal region of the Sp2H mutant neural tube (Fig. 5e), when compared to wild-type littermates (Fig. 5d). Similarly, using semi-quantitative RT-PCR, a 50% reduction in Wnt-1 mRNA levels was observed within both 9.5 and 10.5 dpc homozygous Sp2H mutant embryos when compared to control littermate samples (data not shown).
To confirm the presence of NCC in the Sp2H mutant neural tube, we also examined embryos for the expression of another known marker of NCC expansion, Wnt-3a. However, at all ages examined (8.5–10.5 dpc) the expression of Wnt-3a appeared unchanged within the homozygous Sp2H mutant embryos (n=8) when compared to either wild-type (n=4) or heterozygous (n=7) littermate embryos (Fig. 5f–i). Semi-quantitative RT-PCR analysis also indicated that Wnt-3a mRNA levels were similar in both 9.5 and 10.5 dpc control and homozygous Sp2H mutant embryos (data not shown). Thus, as there is an alteration within the normal relative levels of these NCC expansion markers, it appears that Sp2H mutant cardiac NCC fail to undergo normal NCC expansion and that this is the initial event that will eventually give rise to the cardiac NCC-related heart defects.
3.6 Effect of reduced NCC migration on Sp2H neuronal derivatives
Given the drastic reduction of migrating NCC within the Sp2H mutant, we also examined the possible affects this may have on neuronal derivatives and on neuronal cardiac innervation of the cardiac outflow tract. Pax3 has recently been shown to regulate the differentiation of peripheral neurons  and the Sp mutant is known to have abnormalities within the cranial ganglia and nerves .
Despite the extensive reduction in NCC within the Sp2H mutant, neurofilament staining revealed that the IXth (glossopharyngeal), Xth (vagus) and XIth (accessory) nerves were present in all 11.5, 12.5 and 13.5 dpc homozygous mutant embryos examined (n=14). There was some disorganization, and most of the nerves were reduced in size, but all were present within the expected pattern in the Sp2H mutants (Fig. 6a–d). Histological sectioning revealed the homozygous Sp2H mutant outflow tract was innervated at 13.5 dpc (n=4; data not shown). However, the dorsal root ganglia were all absent, as reported for the Sp1H mutant . This data suggests that the reduced numbers of NCC that do migrate within the Sp2H mutant, are relatively normal and can form some of their normal neuronal derivatives, albeit in a somewhat limited and disorganized manner.
The Pax3 transcription factor contains two DNA-binding domains, a paired domain and a homeodomain, both of which are required for normal DNA-binding . Several Splotch alleles exist, and although all have different mutations within Pax3, only the Sp and Sp1H/Sp2H[8,10] homozygous mutants have NCC-associated heart defects. Sp1H and Sp2H alleles have the same mutation as they were both derived from a single irradiated male that was mated to different females. Sp-delayed homozygous mutants do not develop PTA but die during late embryogenesis or neonatally , presumably due to the presence of spina bifida. The Sp4H and Sp-retarded homozygous mutants die before outflow tract development . Whereas, the Sp mice have a mutation in the splice acceptor site of Pax3 at the intron 3/exon 4 boundary adjacent to the paired domain , the Sp2H mice have a 32-bp deletion that gives rise to a truncated protein lacking a functional homeodomain and the carboxyl region . Both Sp and Sp2H homozygous embryos have grossly similar abnormal heart phenotypes (PTA with associated VSD), but there are also significant differences. Sp mutants have an abnormally thinned myocardium  which is not present within the Sp2H homozygous mutant embryos . Additionally, there is an over-expression of p57Kip2 mRNA within the Sp myocardium that is not present at within Sp2H (Wang and Conway, unpublished results). However, apart from the similarity of Splotch phenotypes to the chick cardiac NCC ablation model , the effects of neither mutation on the cardiac NCC morphogenesis has been vigorously investigated.
In this study, we set out to determine why the cardiac NCC do not reach the outflow tract of the Sp2H mutant hearts as previously reported  and whether the environment was hindering the Sp2H cardiac NCC such that migration was affected, or whether the problem was intrinsic to the NCC themselves. Both the cell adhesion molecules, N-cadherin and N-CAM which are important for the emigration of NCC , are aberrantly expressed in Sp mutant embryos  and N-CAM has also been shown to be prematurely polysialylated . There are also excessive gap junctional vesicles within the Sp mutant neural groove that may affect cellular adhesion . Additionally, there are alterations of laminin, fibronectin, type IV collagen and heparan sulfate proteoglycan expression within the Sp-delayed mutant extracellular matrix . And recently, the matrix in Sp2H has been shown to over-express versican within the somites through which the NCC migrate , and versican is known to have an inhibitory effect on avian NCC migration . However, our results show that the Sp2H mutant cardiac NCC can migrate normally both in vitro on glass slides and in vivo within the mouse–chick chimera system, given a wild-type environment. Unlike the Sp and Spd mutants , there is no delay in NCC emigration from the Sp2H neural folds in vitro. Although, Serbedzija and McMahon  demonstrated that caudal trunk Sp NCC are capable of relatively normal migration in vivo by transplanting transgenically marked Sp mouse neural tubes into normal chick hosts, the migratory ability of both cranial and cardiac Sp NCC within the mouse–chick chimera system remains unknown. Thus, the loss of a functional Pax3 protein appears to affect the normal development of the NCC themselves and not the environment through which they migrate.
This reduced population of Sp2H cardiac NCC can migrate along the normal migratory pathways at the appropriate developmental times, indicating that a functional Pax3 protein is also not required for NCC guidance or migration. The AP-2 molecular marker studies also indicate that the mutant Sp2H NCC present within the neural tube, can undergo normal epithelial-to-mesenchymal transformation and do emigrate from the neural tube. Additionally, cranial NCC do not appear to be affected by the Sp2H mutation, supporting the suggestion that NCC morphogenesis is differentially controlled at different levels along the body . This reduced population of Sp2H mutant NCC also express all the known cardiac NCC molecular markers . Interestingly, approximately 10% of the homozygous Sp2H mutant embryos have a completely divided heart (34/37 Sp2H homozygous mutant 13.5 dpc embryos had PTA) and survive to birth when they die because of the severe neural tube defects. Taken together, all these data indicate that once the Sp2H cardiac NCC have emigrated from the neural tube, they can migrate normally and can even form a functional septum if a threshold level of cardiac NCC colonize the outflow tract of the heart.
Similarly, the neuronal derivatives of the Sp2H NCC are present, but in a reduced manner. Thus, the NCC defect appears to only affect some NCC-derived structures such as the outflow tract of the heart, thymus, thyroid, parathyroids and dorsal root ganglia , but not the neuronal derivatives suggesting that the Sp2H mutation only affects a subpopulation of NCC. Recently, Poelmann and Gittenberger-de Groot [55,56] demonstrated that the ultimate fate of a subpopulation of mesenchymal cardiac NCC is to undergo apoptosis in the outflow tract of the heart, and our data is consistent with the idea that it is that subpopulation of NCC that is missing or diminished in the Sp2H mutant.
Significantly, the Sp2H mutant NC stem cells fail to undergo normal NCC expansion, as shown by the reduction in Wnt-1 marker expression within the 8.5 dpc Sp2H embryo. Interestingly, Pax3 transcripts are coexpressed in the 8.5 dpc mouse embryo within the dorsal part of the neuroepithelium , coincident with the site of Wnt-1/Wnt-3a expression and NCC induction and expansion. It is also interesting to note that the expression pattern of Pax3 is altered in the Wnt-1/Wnt-3a double homozygous mutant embryos , suggesting that there is either a direct or indirect control mechanism, but further studies are required to determine the genetic hierarchy regulating NCC development. Wnt-1/Wnt-3a double homozygous mutant embryos have a drastic reduction in NCC derivatives, but neither single mutant has a NCC-associated phenotype . As Wnt3a expression is unaffected by the Sp2H mutation, this suggests that either the down-regulation of the Wnt1 expression is merely a molecular marker of abnormal Sp2H NCC expansion, or that it is the relative levels of both Wnt-1 and Wnt-3a within the Sp2H mutant background that are important for normal cardiac NCC expansion. Alternatively, the Sp2H mutation within Pax3 may result in a gain-of-function aberrantly affecting the regulation of Wnt-1 and thereby normal NCC expansion. A recent study has shown that Pax3-dependent transactivation is mediated by the C-terminus and that Pax3 transcripts lacking a functional C-terminus/homedomain, which is absent within the Sp2H mouse, can result in gain-of-function . Using the Wnt-1 lacZ reporter construct to mark migratory NCC, significantly fewer marked migratory cardiac and trunk NCC (but not cranial NCC) were observed within the Sp mouse embryos, but the intense Wnt-1 transgene expression within the neural tube was unaffected . This may reflect differential effects of the different mutations within the Pax3 alleles, or the fact that lacZ expression may not precisely reflect the normal expression levels of Wnt-1 mRNA.
Neither elevated cell death or reduced rates of cell proliferation along the migration pathway were present within the Sp2H mutant, confirming that the molecular and morphological NCC-deficiencies observed must be due a defect prior to cardiac NCC emigration. Additionally, the environment through which the Sp2H mutant NCC migrate can support their survival. However, Pax3-expressing NCC appear to undergo apoptosis within the Sp mutant embryo . This suggests that mutations of either the paired or homeobox regions of Pax3 may give rise to different NCC morphogenetic abnormalities within the different Splotch alleles. The Sp2H and Patch mutants  both have phenotypically similar conotruncal heart defects but different genetic mutations, and both have widely different abnormalities within NCC morphogenesis. In contrast to the Sp2H mutants, Patch cardiac NCC are capable of normal NCC expansion, emigration and migration but undergo extensive apoptosis along the migratory pathway (Dickman and Conway, manuscript submitted). Additionally, Sox-4 deficient mice have a similar phenotypic conotruncal heart defect as both the Patch and the Sp2H mutants, but do not appear to have any cardiac NCC-associated abnormalities, rather a lack of endocardial cushion formation results in absent semilunar valves . Thus, although different genetic mouse mutant models may have phenotypically similar cardiac NCC-related heart defects, the underlying abnormalities can occur at several different stages and even independently of NCC morphogenesis. By comparing the effects of many different genetic models of cardiac NCC-related heart defects, mouse mutant models should be able to provide evidence as to the genetic causation and underlying mechanisms responsible for these defects in humans.
An alternative explanation of our data suggests that the environment immediately surrounding the neural tube may be inhibiting normal NCC morphogenesis, and that when the Sp2H NCC are released from their repressive environment, such as after transplantation into culture or into a chick embryo, they begin to migrate normally. This idea was thought to be relevant in the view of the fact that versican is over-expressed in the Sp2H somites and is known to have an inhibitory effect on NCC migration [53,54]. However, the NCC-related heart and neural tube defects within the Sp homozygous mutant embryo was recently rescued using a transgenic approach. Over-expression of Pax3 in the neural tube and NCC indicated a cell autonomous role for Pax3 within the NCC, as Pax3-deficient somites (over-expressing versican) do not seem to adversely affect NCC migration , indicating that the abnormal matrix does not play a role within the Sp NCC-related defects. Our results are consistent with this data and suggest that abnormal NCC-related heart defects within the Sp2H mutant allele are solely due abnormalities within the early neural fold/tube, and not due to abnormalities within the Sp2H matrix through which the cardiac NCC migrate. However, we cannot yet exclude the possibility that the environment within the Sp2H neural folds/tube or immediately surrounding it is aberrantly affecting NCC induction within the neural folds/tube itself, resulting in the observed abnormal NCC expansion. It is also possible that the reduced number of cardiac NCC may be due to contact inhibition. That is, if Sp2H cardiac NCC cannot move into the migratory pathway, there may be no room for additional proliferation. A number of candidate molecules have been thought to play a role in NCC induction/expansion; including bone morphogenetic proteins (member of the TGFβ gene family of growth factors), slug, dorsalin-1, Zic2 and growth factors such as bFGF (reviewed in Ref. ). Additionally, Wnt-1 has been shown to modulate cell–cell adhesion by stabilizing β-catenin binding to N-cadherin, thereby influencing both cell boundary formation and cell proliferation . Studies are currently underway to try to determine whether cell proliferation, cell–cell adhesion, NCC induction or contact inhibition is primarily responsible for the decreased number of NC stem cell precursors within the Sp2H mutant neural folds/tube.
We thank both Drs Andrew Copp and Margaret Kirby for invaluable technical advice and assistance at the start of this project, and Dr Pam Mitchell for the AP-2 cDNA probe. The 3A10 antibody, developed by Thomas Jessell and Jane Dodd, was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by the University of Iowa, Department of Biological Sciences, Iowa City, IA 52242. This work was supported by a Post-doctoral Research Fellowship (#9820132V) to ED from the Southeast Affiliate of the American Heart Association; and by a Basil O’Connor Starter Scholar Research Award (#FY97-0690) from the March of Dimes Birth Defects Foundation and NIH grants HL60104 and HL60714 to SJC.