ARVCF depletion cooperates with Tbx1 deficiency in the development of 22q11.2DS-like phenotypes in Xenopus

Knockdown of ARVCF Results in Malformations of the Craniofacial Cartilage and Abnormalities in the Aortic Arches

The in situ hybridization expression pattern of Xenopus ARVCF and the fact that ARVCF is deleted in 22q11.2DS prompted us to investigate if this protein plays a role in the development of the structures affected in 22q11.2DS. To deplete ARVCF, a previously described morpholino directed against its 5′ UTR (ARVCF MO, Fang et al., 2004) was injected in four- or eight-cell-stage embryos. As a control, a morpholino directed against an irrelevant sequence was used (control MO). For phenotypic analysis, embryos were injected with ARVCF MO or control MO in the dorsal blastomeres in the animal region in order to target the cranial and cardiac NCCs (Moody, 1987). Injection of 40 ng of the ARVCF MO caused 47% reduction in endogenous ARVCF levels (see Supp. Fig. S1, which is available online) and, when injected at the one-cell stage, led to gastrulation defects (data not shown and Fang et al., 2004). However, most embryos injected with a similar dose of ARVCF MO at the four- or eight-cell stage underwent normal gastrulation and initially developed normally. The resulting tadpoles were processed at stage 48 for immunostaining with an antibody directed against the von Willebrand factor (vWF) to observe the aortic arches. In normal development, only three of the six aortic arches remain at stage 48, namely arches 3, 4, and 5 (Levine et al., 2003). The development of these aortic arches was severely affected in embryos injected with the ARVCF MO (Fig. 2A and Table 1), and aortic arch 5 was usually missing (Fig. 2A, d–f) or severely underdeveloped (data not shown). In addition, fewer bifurcations were observed on aortic arches 3 and 4. Importantly, these aortic arch defects could be rescued to a large extent by coinjecting 20 pg of ARVCF RNA lacking the morpholino-recognition sequence (Supp. Fig. S2 and Table 1).

Alcian Blue staining, which reveals the cartilage structures of the head, showed that all the facial structures, such as Meckel's cartilage, the ceratobranchials, and the ceratohyal cartilage, were present in embryos injected with the ARVCF MO (Fig. 2B,C). However, malformations of Meckel's and ceratohyal cartilage were observed in 42% of the embryos injected with ARVCF MO (Fig. 2B and Table 1) and heads were more narrow (Fig. 2C). These defects were substantially rescued by coinjection of ARVCF RNA (Fig. 2C and Table 1). Together, these results show that depletion of ARVCF in Xenopus embryos generates developmental malformations comparable to the defects observed in 22q11.2DS.

Knockdown of Tbx1 Results in Defects of the Facial Cartilage and Aortic Arches

Several studies have demonstrated that absence of Tbx1 in mice and zebrafish generates phenotypes that resemble malformations observed in humans with 22q11.2DS (Jerome and Papaioannou, 2001; Lindsay, 2001; Merscher et al., 2001; Piotrowski et al., 2003). We wanted to determine if morpholino-mediated depletion of Tbx1 in Xenopus embryos results in similar defects. Hence, we designed a morpholino against a region encompassing the start codon to effectively deplete the Tbx1 protein (Supp. Fig. S1). Embryos were injected with 50 ng of either Tbx1 MO or control MO. At stage 48, embryos were stained with the vWF antibody to observe the aortic arches, and with Alcian Blue to reveal the cartilage structures (Fig. 2). Depletion of Tbx1 generated defects in the aortic arches (Fig. 2A) similar to those seen after knockdown of ARVCF. Again, aortic arch 5 was underdeveloped and sometimes absent (Fig. 2A, g–i), and fewer bifurcations were observed on aortic arches 3 and 4 compared to the controls (Fig. 2A). Injection of 50 ng of Tbx1 MO also led to malformations of the cartilage structures of the head at stage 48 compared with embryos injected with control MO (Fig. 2B,C and Table 1). Meckel's cartilage and the ceratohyal cartilage were frequently malformed (Fig. 2B,C). Like the effects of ARVCF MO, a significant reduction in the width of the head was observed in tadpoles injected with Tbx1 MO (Fig. 2C). Both the defects in the aortic arches and the narrowing of the head could be rescued by the coinjection of a synthetic Tbx1 RNA that could not be recognized by the Tbx1 MO (Supp. Fig. S3).

Gross Anatomical Evaluation of the Larval Head Region by Micro-CT Scanning

We also investigated whether the aortic arch abnormalities observed in tadpoles depleted of ARVCF or Tbx1 could be secondary to gross general structural malformations in the head region. We injected embryos at the four-cell stage unilaterally with morpholino and a fluorescent tracer in one dorsal blastomere. Stage-48 tadpoles were fixed, stained with phosphotungstic acid, and scanned by micro-CT (Fig. 3). Several structures, including the brain, the cranial muscles, and the gill apparatus, could be observed in three dimensions. This examination confirmed that tadpoles injected with ARVCF or Tbx1 MO had a smaller gill apparatus, the basis of which is the ceratobranchial cartilage. It also showed that all structures that are visible in control tadpoles, including the muscles and the branchial basket, were present and correctly patterned. These results make it unlikely that the aortic arches are non-specifically affected in ARVCF or Tbx1 MO-injected embryos.

ARVCF and Tbx1 Depletion Delays Cranial Neural Crest Cell Migration

As described above, injection of either ARVCF or Tbx1 MO led to malformations in the cartilage structures of the head (Fig. 2). Because these structures are derived mainly from the cranial neural crest cells (CNCC), we wanted to find out if the specification and/or migration of these cells is disturbed by depletion of ARVCF or Tbx1. Embryos were injected unilaterally at the four-cell stage with ARVCF, Tbx1, or control MO, together with a tracer RNA. Neural crest specification and migration was then examined by in situ hybridization for the marker genes Twist (Fig. 4) and Slug (data not shown) at stage 20, when CNCC have initiated their migration, and at successive stages 27 and 32, when they have reached their final destination in the head. Knockdown of ARVCF or Tbx1 did not result in loss of CNCC, and their specification was apparently normal at stage 20 (Fig. 4). All three streams of migrating CNCC that can be revealed by the Slug or Twist probes were discernable, namely the mandibular, hyoid, and branchial streams. However, the migration of the CNCC in the different streams was delayed in embryos injected with ARVCF MO (57% affected, n = 36) or Tbx1 MO (80% affected, n = 21) compared to the non-injected side (Fig. 4), but it was not delayed in embryos injected with control MO. At stage 32, no differences between the injected and non-injected side could be observed for any of the injection set-ups (Fig. 4), indicating that the defect observed is indeed delayed migration and not a premature stop in migration. Together, our results show that depletion of neither ARVCF nor Tbx1 influences NCC specification. However, depletion of ARVCF or Tbx1 delays the migration of the CNCCs in the pharyngeal arches.

Cooperative Effect of ARVCF and Tbx1 Depletion on the Development of the Cartilage and the Aortic Arches

Our data show that although ARVCF and Tbx1 may be active in different cell types, they both regulate the formation of craniofacial and cardiovascular structures. To look for potential cooperative activity, we depleted ARVCF and Tbx1 simultaneously. As ARVCF and Tbx1 morpholinos work in a dose-dependent way, we combined silent doses of both MOs and examined the cartilage and aortic arch phenotypes. Injection of 20 ng of either Tbx1 MO or ARVCF MO did not lead to any noticeable phenotypic change in Xenopus embryos. In contrast, combined injection of 20 ng of each morpholino at the one-cell stage generated phenotypes similar to those observed with single injections of the effective dose (50 ng), namely defects in the cartilage structures of the head and in the aortic arches (Fig. 5 and Table 1). After depletion of both ARVCF and Tbx1, aortic arch 5 was usually missing, and, when present, it was severely underdeveloped, as were arches 3 and 4 (Fig. 5A, f). In addition, Meckel's and ceratohyal cartilage were malformed in embryos injected with both ARVCF MO and Tbx1 MO, but not in the different controls (Fig. 5A). Heads of embryos injected simultaneously with both morpholinos were significantly smaller than in controls (Fig. 5B). Also, migration of CNCC was markedly delayed in the embryos injected unilaterally with 20 ng ARVCF MO in combination with 20 ng Tbx1 MO (76%, n = 25) (Fig. 4). No delays in migration were observed when the subphenotypic MO doses were supplemented with 20 ng of control MO in order to compensate for potential non-specific effects of higher morpholino concentrations (n = 25 for each set-up). All together, these data indicate that ARVCF and Tbx1 act cooperatively in the generation of structures derived from CNCC.

Whole-Mount In Situ Hybridization

To make sense and antisense probes directed against xARVCF, an 807-bp fragment (bp 1611–2418) from pCS2+ARVCF (Paulson et al., 2000) was cloned in pBluescript. The resultant plasmid was linearized with BamHI and SmaI. The plasmids containing the Twist and the Slug probe were linearized with EcoRI and NdeI, respectively. Digoxigenin was incorporated in the probes according to the manufacturer's instructions (Roche, Indianapolis, IN). In situ hybridization was performed as described (Ciesiolka et al., 2004).

Morpholino and RNA Injections

To deplete ARVCF, the ARVCF MOII was used as described (Fang et al., 2004). The morpholino directed against Tbx1 was designed by Gene Tools (Philomath, OR) and has the following sequence: 5′CCGAGATCATCCCAGGAATAGACAG3′. The control MO is a standard morpholino provided by Gene Tools. The morpholinos and RNAs were microinjected in X. laevis embryos as described (Ciesiolka et al., 2004). For tracing experiments, RNA encoding β-galactosidase or a lysamine-labeled control MO was coinjected with the morpholinos targeting ARVCF or Tbx1. To assess the efficiency of the Tbx1 morpholino, a HA-tag was added to the C-terminus of xTbx1 in the plasmid βUT2-xTbx1a (kindly provided by Peter Scambler) using the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The tagged cDNA was subcloned in pCS2+, linearized, and used for RNA synthesis using the mMESSAGE mMACHINE® SP6 Kit (Ambion, Austin, TX). For rescue experiments, RNAs were injected that lack the sequence targeted by the morpholinos.

Immunostaining, Alcian Blue Staining, and Head Measurements

Embryos were fixed in Dents fixative (20% DMSO, 80% methanol) overnight at −20°C. Embryos were rehydrated stepwise, washed in PBS, pre-incubated in TBS with 2% BSA, and incubated overnight at 4°C with primary antibody (von Willebrandt Factor, 1:100, Dako, Heverlee, Belgium). Next, embryos were washed in PBS containing 0.1% Triton X100, blocked with 2% BSA in PBS, and incubated overnight with secondary antibody (anti-rabbit Alexa 594, 1:500, Molecular Probes, Eugene, OR; Invitrogen, Carlsbad, CA). Finally, they were washed with PBS/0.1% Tween 20. Images were taken with a Leica MZFLIII fluorescence stereomicroscope. Alcian blue staining was performed as described (Pasqualetti et al., 2000). The width of the head was plotted by measuring the widest point of the ceratobranchial and Meckel's cartilage structures as indicated in Figure 2Ba and adding the two values.

Micro CT-Scanning

Embryos were injected at the four-cell stage with ARVCF, Tbx1, or Control MO in one dorsal blastomere and left to develop until stage 48, when they were fixed in 4% paraformaldehyde in PBS and further stained with 2.5% phosphotungstic acid. Next, the tadpoles were dehydrated stepwise until ethanol 70%. Finally, they were scanned by micro-CT. The images were imported into Amira 5.3.3 order to generate a 3D reconstruction of each tadpole.