Timed mutation and cell-fate mapping reveal reiterated roles of Tbx1 during embryogenesis, and a crucial function during segmentation of the pharyngeal system via regulation of endoderm expansion

Tbx1 encodes a transcription factor belonging to the T-box gene family. Its importance in development and in DiGeorge syndrome (DGS) was revealed through a search for DGS candidate genes using engineered segmental deletions in the mouse (Lindsay et al.,1999; Lindsay et al.,2001; Merscher et al.,2001) and single gene knockouts(Jerome and Papaioannou, 2001; Lindsay et al., 2001; Merscher et al., 2001). Mutational analyses in individuals have demonstrated that TBX1mutations can indeed cause the DGS phenotype(Yagi et al., 2003). In mice, Tbx1 is expressed in pharyngeal endoderm, mesodermal core of pharyngeal arches, head mesenchyme, the secondary heart field, the otocyst and sclerotome (Chapman et al.,1996; Hu et al.,2004; Vitelli et al.,2002; Xu et al.,2004). Tbx1 is haploinsufficient in the development of the aortic arch, thymus and parathyroids(Jerome and Papaioannou, 2001; Lindsay et al., 2001; Merscher et al., 2001; Taddei et al., 2001), and homozygous mutation causes cardiovascular abnormalities, craniofacial defects,thymic and parathyroid aplasia as well as severe ear developmental defects(Jerome and Papaioannou, 2001; Kelly et al., 2004; Moraes et al., 2005; Raft et al., 2004; Vitelli et al., 2002; Vitelli et al., 2003). Most birth defects associated with DGS and with Tbx1 mutations in mice are thought to derive from developmental abnormalities of the embryonic pharyngeal system, which is a transient, vertebrate-specific structure. The system is composed of `modules' (pharyngeal arches, pouches and clefts) that are added during embryogenesis in a cranial-to-caudal order. The process of `addition'of pharyngeal arches and pouches is usually referred to as pharyngeal segmentation. Loss of Tbx1 causes severe hypoplasia or aplasia of most pharyngeal arches (2nd-4th and 6th) and pouches (2nd-4th), and severe hypoplasia of the pharynx (Jerome and Papaioannou, 2001; Vitelli et al., 2002). The genes involved in segmentation are unknown, and whether or not Tbx1 is a direct player in this process is unclear. The severity of the early pharyngeal phenotype of Tbx1^-/-embryos makes it difficult, if not impossible, to establish the role of Tbx1 in development and organogenesis of pharyngeal derivatives, and to understand whether the cardiovascular phenotype is related to the pharyngeal phenotype or due to a direct role of Tbx1 in cardiovascular development. Tissue-specific deletion(Xu et al., 2004) and hypomorphic mutations (Hu et al.,2004; Xu et al.,2004) have partially addressed this issue. However, owing to early embryonic defects, the requirement of Tbx1 during the development of organs affected in DGS remains mostly unclear. Here, we have used a time-course deletion of Tbx1 during embryogenesis to map over time the role of the gene in the development of pharyngeal derivatives and cardiovascular system. We have associated this analysis with a timed map of the fate of Tbx1-expressing cells to correlate phenotypic findings with gene expression. Results show distinct time windows in which Tbx1 is required for different organs and structures, and that it has a direct role during pharyngeal segmentation. In addition, our data show that Tbx1 is required at multiple stages during the development of the same organ or system. We propose a unifying model for the mechanism of action of Tbx1 in various organs and systems based on regulation of cell proliferation.

All the experiments involving mice were carried out according to a protocol reviewed and approved by the Institutional Animal Care and Use Committee of Baylor College of Medicine, in compliance with the USA Public Health Service Policy on Humane Care and Use of Laboratory Animals.

The following mouse mutant lines have been described previously:TgCAGG-CreER™ (Hayashi and McMahon,2002), Tbx1^flox/lox and Tbx1^mcm/+ (Xu et al.,2004), Tbx1^+/-(Lindsay et al., 2001), and R26R (Soriano, 1999). All lines were backcrossed into the C57Bl6 genetic background for at least two generations. Mice were genotyped using PCR as described in the original reports. To induce nuclear translocation of the inducible Cre, including Cre^ERTM and MerCreMer encoded by the CAGG-CreER™ transgene and the Tbx1^mcm allele, respectively, pregnant mice were treated with single intraperitoneal injection of tamoxifen (Sigma) with a dose of 75 mg/kg body weight at the desired time points. Tamoxifen was dissolved in absolute ethanol at the concentration of 100 mg/ml and then diluted 1:10 in autoclaved sesame oil (Sigma) for injection. The excision of the loxP-flanked exon 5 of the Tbx1 allele Tbx1^flox was evaluated as follows. Genomic DNA was extracted from whole E9.5 embryos exposed to TM for 3, 6, 12 or 24 hours and it was used as template in PCR. The PCR primers used to detect Tbx1^flox allele(Fig. 1A) are Tbx1loxP2-F(5′-cgacccttctctggcttatg-3′) and Tbx1loxP2-R(5′-aaagactcctgcccttttcc-3′). PCR products were separated in 1.5%agarose gel by electrophoresis and the intensity of the bands was measured using NIH image 1.63 software on digital images of the gel. The percentages of remaining Tbx1^flox allele in TgCAGGCre^ERTM; Tbx1^flox/+ mutants were calculated as the ratio of the intensity of Tbx1^flox band to that of Tbx1⁺ band, and the values were normalized with those of Tbx1^flox/+ embryos. A second PCR strategy was used to evaluate the disappearance of the Tbx1^flox allele and the appearance of the Tbx1^ΔE5 allele (see scheme in Fig. S1A in the supplementary material). To do this, we used a primer pair amplifying exon 5 and the flanking loxP sites, Tbx1E5-F(5′-ggccctgcctaactcagatt-3′) and Tbx1E5-R(5′-aaagactcctgcccttttcc-3′).

We have used RT-PCR to detect residual Tbx1 transcripts in TgCAGG-CreER™; Tbx1^flox/- and control(TgCAGG-CreER™; Tbx1^flox/+ or Tbx1^flox/+) embryos after exposure to TM for 24 hours (see scheme in Fig. S1B in the supplementary material). Embryos at different stages(E9.5-E12.5) were harvested 24 hours after TM injection and total RNA was extracted from whole embryos using the Trizol reagent (Invitrogen). The concentration of RNA samples was measured using a spectrophotometer and adjusted to 100 ng/μl. cDNA was synthesized from mRNA using Superscript first-strand synthesis system (Invitrogen) with random hexamers as the primers, and then subjected to PCR amplification (30 cycles) using Tbx1 and β-actin-specific primer pairs. The PCR primers used to examine Tbx1 mRNA level were Tbx1mRNA_F(5′-TTTGTGCCCGTAGATGACAA-3′) and Tbx1mRNA_R(5′-AATCGGGGCTGATATCTGTG-3′).

E18.5 embryos were examined and photographed after manual dissection under a stereomicroscope. After formalin fixation, the great arteries were visualized by India ink injected into the left ventricle. Earlier embryos were examined under the stereomicroscope, fixed and paraffin embedded for histological analysis. Pharyngeal arch arteries were visualized at E10.5 using intracardiac Ink injection. Embryos were then fixed and dehydrated in ethanol:water: acetic acid: chloroform (95:3:1:1) solution and cleared in methyl salicylate: benzyl benzoate (50:50) solution. β-Galactosidase activity was visualized by staining paraformaldehyde-fixed embryos with the X-gal substrate, according to standard procedures. Stained whole-mount embryos were photographed and then embedded in paraffin and cut into 10 μm histological sections. Sections were counterstained with Nuclear Fast Red. Cell proliferation was assessed by immunohistochemistry using an anti phosphorylated histone H3 antibody (Upstate Biotechnology) on sections of ethanol-fixed, paraffin-embedded E10.0 embryos. Whole-mount RNA in situ hybridization was performed as previously described(Albrecht et al., 1997) using a Pax1 probe kindly provided by Dr R. Balling.

We have used drug-induced Cre-mediated recombination to resolve the roles(s) of Tbx1 during embryogenesis. We have previously described a Tbx1 conditional allele (Tbx1^flox) in which exon 5, encoding part of the T-box domain, is flanked by loxP sites(Xu et al., 2004). Cre-induced recombination of this allele generates the allele Tbx1^ΔE5, which is functionally null. To drive Cre recombinase expression, we used a transgenic mouse line expressing ubiquitously a tamoxifen-inducible Cre^ERTM protein(Hayashi and McMahon, 2002). This line is hereafter referred to as TgCAGG-CreER™. We tested TgCAGG-CreER™-induced recombination of the Tbx1^floxallele by crossing TgCAGG-CreER™ mice with Tbx1^flox/flox mice. We treated pregnant females with a single intraperitoneal injection of tamoxifen (TM) at 75 mg/kg body weight. We harvested embryos 3, 6, 12 and 24 hours after injection. Injections were timed so that all embryos were harvested at embryonic (E) day 9.5. We used a PCR strategy on whole embryo genomic DNA to evaluate the residual (undeleted) Tbx1^flox allele after TM injection, as illustrated in Fig. 1A,B. Results showed that after 3 hours, ∼28% of the floxed allele has been excised, and at 24 hours no floxed allele was detected in most embryos(Fig. 1B,C). We have also used an alternative PCR strategy to evaluate recombination, results were similar(see Fig. S1A in the supplementary material). To establish whether Tbx1 transcripts were also depleted after 24 hours of TM exposure in TgCAGG-CreER™;Tbx1^flox/- embryos, we have performed RT-PCR on whole embryo RNA at different stages (E9.5, E10.5, E11.5 and E12.5). Results showed undetectable or very low level of Tbx1 transcript at all stages tested (see Fig. S1B in the supplementary material). To establish whether the TgCAGG-CreER™-induced recombination is homogeneously distributed in the tissues of the pharyngeal apparatus, we examined TgCAGG-CreER™;R26R E9.5 and E10.5 embryos after 24 hours exposure to TM. Results showed homogeneous recombination (as detected by X-gal staining) in the pharyngeal apparatus (see Fig. S1C in the supplementary material). Overall, these results indicate that the induction of Tbx1^flox deletion is fast, progressive and leads to essentially ubiquitous deletion within 24 hours. TM can cause embryonic lethality (Fig. 1D). However,all the wild-type embryos exposed to a single dose of TM and examined at E10.5, E12.5 and E18.5 (n=35, 16, and 39, respectively), were grossly normal. Hence, under the conditions used, exposure to TM did not cause abnormalities that may confound phenotypic analysis.

We crossed TgCAGG-CreER™;Tbx1^+/- mice(Lindsay et al., 2001) with Tbx1^flox/flox mice in order to perform a time course inactivation of Tbx1. Pregnant females were injected with a single dose of TM at E7.5, E8.5, E9.5, E10.5 or E11.5, and we examined the fetuses at E18.5. Results (Fig. 2A,B)indicated that different organs and structures have distinct time requirements, but at no injection time point were the fetuses normal. Injection of TM at E7.5 recapitulated the germ line homozygous mutant phenotype, providing a functional confirmation that induced somatic recombination of the Tbx1 gene with this system is sufficient to generate a null phenotype. The two major components of the cardiovascular abnormalities in Tbx1 mutants(Vitelli et al., 2002), namely aortic arch patterning defects (in Tbx1^+/- and Tbx1^-/- mutants) and cardiac outflow tract (OFT) defects(mainly in Tbx1^-/- mutants) had distinct time-requirements(Fig. 2A, Fig. 3A-D). Tbx1^+/- animals exhibited aortic arch patterning abnormalities owing to defects of the 4th pharyngeal arch arteries (PAAs)(Jerome and Papaioannou, 2001; Lindsay and Baldini, 2001; Merscher et al., 2001; Vitelli et al., 2002). Injection of ink into the heart of E10.5 Tg-CAGG-CreER™;Tbx1^flox/+ embryos exposed to TM at E7.5 demonstrated that heterozygous deletion of Tbx1 at this time point causes 4th PAA hypoplasia (Fig. 3E). Exposure to TM at E8.5 of embryos with the same genotype resulted in normal or only mildly hypoplastic 4th PAAs(Fig. 3F, compare with 3G). Thus, the 4th PAA haploinsufficiency phenotype is due to an early requirement for Tbx1 that precedes the morphological appearance of the arteries,which occurs after E9.5.

Of the phenotypic abnormalities assessed in this study, clefting of the secondary palate is the only one that occurred after Tbx1 ablation at E11.5 (Fig. 2A and Fig. 4F,G). Thus, clefting in these mutants is not due to early abnormalities of 1st pharyngeal arch patterning, which are present in Tbx1^-/- embryos(Kelly et al., 2004), but most likely to a role of Tbx1 in secondary palatogenesis, which occurs from approximately E12.5.

We used the Tbx1^mcm allele(Xu et al., 2004), which carries a cDNA encoding a tamoxifen-inducible Cre recombinase(Verrou et al., 1999) knocked into the Tbx1 allele, to obtain a timed fate map of Tbx1-expressing cells. Tbx1^mcm allele is an inactivating allele for Tbx1. However, we have shown that Tbx1 heterozygosity does not alter Tbx1 expression pattern(Vitelli et al., 2002). Exposure of Tbx1^mcm/+;R26R(Soriano, 1999) embryos to a single dose of TM should induce recombination of the R26R allele and permanent expression of β-galactosidase in cells expressing Tbx1 within 24-36 hours of injection [the half-life of TM is 11.9 hours(Robinson et al., 1991)] and their progeny. We thereby induced Cre recombination at different developmental stages and then harvested and stained embryos with X-gal at E10.5. Induction at E7.5 labeled, among other tissues, the endothelium of the 4th PAAs and the pharyngeal ectoderm, while induction at E8.5 or E9.5 did not (although very few, labeled ectodermal cells could be detected with induction at E8.5, Fig. 3H-J), suggesting that Tbx1 may have a role in endothelial precursors and/or pharyngeal epithelia in the formation of the 4th PAAs. The crucial time requirement for Tbx1 in the development of the OFT (as determined by scoring the presence of persistent truncus arteriosus) is between E8.5 and E9.5(Fig. 2A,B). Cell fate induction at these time points revealed that contribution of labeled cells to the OFT was very strong after induction at E8.5 (mainly myocardial and endothelial cells, as well as mesenchymal cells possibly derived from transformation of endothelial cells), but was very weak and restricted to the most distal segment of the OFT with induction at E9.5(Fig. 3L,M). Some blue cells were also detected in the atria with induction at E8.5 (data not shown). Cell-fate mapping with induction at E7.5 produced similar results to those obtained with induction at E8.5, but we noted more labeled cells invading the right ventricle (Fig. 3K). These results are consistent with our previous data indicating a crucial role of Tbx1 in cardiomyocyte precursors fated to the OFT(Xu et al., 2004), and they identify an OFT-specific subpopulation of cardiomyocyte precursors that migrates into the OFT between E8.5 and E9.5. The crucial time for Tbx1 function in thymus formation also spans the E8.5-E9.5 time interval (Fig. 2A,B; Fig. 4A-C). The labeling of the OFT was more robust and extensive than previously reported with the same Tbx1^mcm line (Xu et al., 2004). We found that the induction with a single, high dose of TM (used here) produces a more robust labeling than repeated, low dose injections (Xu et al.,2004).

The thymic primordium derives from the endoderm of the 3rd pharyngeal pouch(Gordon et al., 2004) and Tbx1 is required at this time interval for the formation of this pouch (Fig. 5K,L). In addition,timed fate mapping of Tbx1-expressing cells revealed a heavy contribution of labeled cells to the thymic primordia after induction at E8.5(Fig. 4D), suggesting a cell-autonomous role for Tbx1. Surprisingly though, Tbx1 is also required for thymic morphogenesis after E10.5, that is well after the formation of the 3rd pouch (Fig. 4B). Timed fate mapping induced at E9.5(Fig. 4E) and E10.5 (not shown)revealed a very limited contribution of labeled cells to the thymus,suggesting different mechanisms for aplasia and morphogenesis. The thymic dysmorphism associated with Tbx1 timed deletion at E9.5 and E10.5 is reminiscent of that associated with null mutation of Fgf10(Ohuchi et al., 2000; Revest et al., 2001), a target of Tbx1 (Xu et al., 2004).

Timed cell fate mapping induced at E7.5, E8.5 and E9.5 revealed a progressive `concentration' of labeled cells in the caudal region of the pharyngeal endoderm (Fig. 5A-C). This is consistent with the described cranial-to-caudal gradient of Tbx1 expression observed during the development of the pharyngeal system (Vitelli et al.,2002). These observations suggest a role of Tbx1 in the process of cranial-to-caudal addition of pharyngeal arches and pouches that occurs in the E8.0-E10.5 time interval. If Tbx1 expression is crucial for progressive addition of pharyngeal arches and pouches over time, then deletion of Tbx1 during this process should effectively arrest it when the gene deletion occurs. To test this, we exposed TgCAGG-CreER™;Tbx1^flox/- embryos to TM at E7.5 or E8.5 and examined them at E10.5 when the entire complement of pharyngeal arches and pouches should have formed. Embryos exposed at E7.5 exhibited severe hypoplasia of the 2nd pharyngeal arch(Fig. 5D) and lacked the 3rd-6th arches and intervening pharyngeal pouches(Fig. 5D′,G). This phenotype is identical to that caused by germline Tbx1^-/-mutation. RNA in situ hybridization showed that Pax1, a marker of pharyngeal pouch endoderm, is weakly expressed in the first pouch but does not identify the caudal pouches of these embryos(Fig. 5J, compare with 5L). Embryos exposed to TM at E8.5 had normally segmented 1st, 2nd, and 3rd arches and normal 1st and 2nd pouches (Fig. 5E,E′,H), but the 3rd pharyngeal pouch was not identified by Pax1 and consisted only of a small endodermal evagination that did not approach the surface ectoderm (Fig. 5H,K). Thus, deletion of Tbx1 at the time of 3rd pouch formation (between E9 and E9.5) resulted in developmental defects of the segments caudal to and including the 3rd pouch, but did not affect the development of segments cranial to the 3rd pouch. The addition of new arches and pouches probably requires multiple morphogenetic events, the most basic of which is the expansion of the endodermal cell population. Therefore, we tested whether loss of Tbx1 during segmentation is associated with reduced endodermal cell proliferation in TgCAGG-CreER™;Tbx1^flox/- embryos exposed to TM at E8.5. We have used immunohistochemistry with an anti-phosphorylated histone H3 antibody to detect proliferating cells in E10.0 embryos. Results showed that the number of proliferating pharyngeal endodermal cells is reduced by almost 50% in TgCAGG-CreER™;Tbx1^flox/- embryos when compared with TgCAGG-CreER™;Tbx1^flox/+ embryos(P=0.03) (Fig. 6A-C). By contrast, we could not detect changes in proliferation of pharyngeal mesenchyme in the same embryos (Fig. 6C). Thus, Tbx1, directly or indirectly, regulates the expansion by cell proliferation of the pharyngeal endoderm during segmentation(Fig. 6D).

Adding a temporal dimension to the functional analysis of developmental genes can provide important information. However, this can be technically challenging, especially for mammalian embryos. We have used a Cre-loxP system in which the Cre recombinase is TM inducible. Our data demonstrate that the system is efficient, relatively fast responding and capable of delivering the desired deletion of the loxP-flanked allele in a time-controlled fashion. The embryonic lethality caused by TM is a limitation of this approach. However, using single-injection delivery of TM,we observed a relatively low incidence of embryonic lethality associated with injections at E7.5 and E8.5, lower at E9.5 and no lethality associated with injections carried out at E10.5 and E11.5. In addition, we did not lose any pregnant female. The timing of embryo lethality-causing injection is consistent with previously reported data indicating interference of TM with placentation in rats (Sadek and Bell,1996). With our phenotypic methods, none of the surviving embryos presented with developmental abnormalities, except for Tbx1 mutants,and the latter only presented abnormalities previously described in these mutants. Hence, in our particular case, TM treatment did not introduce a phenotypic confounding variable.

The time course deletion of Tbx1 during embryogenesis dissected most of the phenotypic abnormalities that we have tested. Data indicated that Tbx1 is required throughout embryogenesis and that there are discrete and distinct time intervals crucial for the development of the different organs and structures affected in the model. The knowledge of these time windows provides clues as to the role of Tbx1 in the development of different pharyngeal derivatives. For example, thymic formation required Tbx1 at E8.5-E9.5, this did not coincide with formation of the thymic primordia (which occurs later) but with the formation of the 3rd pharyngeal pouch, from which the thymus derives(Gordon et al., 2004; Manley and Blackburn, 2003). Thus, the role of Tbx1 in the formation of the thymus is secondary to its role in the formation of the 3rd pouch. Surprisingly, however, Tbx1 continues to be important for thymic development also after the formation of the 3rd pouch, as demonstrated by organ dysmorphism caused by Tbx1 deletion after E10.5. This late phenotype is more likely to be the consequence of a role of Tbx1 in thymic organogenesis. We were also surprised by the timing of secondary palate closure requirement. It has been shown that the patterning of the first pharyngeal arch is abnormal in Tbx1^-/- mutants (Kelly et al., 2004), therefore we speculated that cleft palate may be a consequence of these early defects. By contrast, timed-deletion data support a role of Tbx1 in the secondary palatogenesis, which occurs between E12 and E14.5 (Kaufman and Bard,1999).

Tbx1 loss of function affects two segments of the cardiovascular system, the aortic arch and the OFT(Vitelli et al., 2002). We have previously shown that tissue-specific deletion of Tbx1 can separate these two groups of abnormalities(Xu et al., 2004) suggesting distinct pathogenetic mechanisms. Here, we show that Tbx1 is required in distinct time windows for the development of the aortic arch and OFT. Counterintuitively, Tbx1 is required early for aortic arch patterning(a relatively late process), and later for OFT growth and remodeling (a relatively early process). Our data are consistent with a role of Tbx1 that precedes or coincides with the formation of the 4th PAAs. Our data also exclude a direct role of Tbx1 in the remodeling or smooth muscle lining of the 4th PAAs, in contrast to hypotheses previously put forward (Kochilas et al.,2002; Lindsay and Baldini,2001). Indeed, deletion of Tbx1 after the 4th PAAs are formed, but before smooth muscle lining and remodeling occurs (TM injections at E8.5, E9.5 and E10.5), resulted in no phenotypic consequences for aortic arch patterning.

Timed-deletion revealed that the requirement for Tbx1 in OFT development is restricted to a relatively late and narrow time-window(E9.0-E9.5), which is not consistent with a crucial role in the early SHF cell populations located medially to the cardiac crescent(Kelly et al., 2001; Meilhac et al., 2004; Mjaatvedt et al., 2001; Waldo et al., 2001). Rather,our findings are consistent with a crucial role of Tbx1 in regions that provide precursors destined to the OFT at a later stage, such as the splanchnic mesoderm (Kelly et al.,2001; Mjaatvedt et al.,2001; Waldo et al.,2001). Timed cell-fate mapping showed that the crucial time window for the Tbx1 role in OFT development coincides with expression of Tbx1 in cells destined to populate the OFT myocardium. It is also possible that Tbx1-traced cells provide a cue for neural crest infiltration or outflow cushion development, once they have entered the heart tube. While several genes have been shown to mark cells of the SHF lineage[e.g. Nkx2.5 (Stanley et al.,2002; Waldo et al.,2001; Xu et al.,2004), Isl1 (Cai et al., 2003), and Mef2c(Dodou et al., 2004)], Tbx1 is the only gene known to mark a subpopulation of cardiomyocyte precursors destined predominantly to the OFT, demonstrating the presence of regionally specified cell populations within the SHF.

The embryonic pharyngeal system has a characteristic modular structure resulting from progressive addition of new segments in a cranial-to-caudal order. In Tbx1^-/- embryos, the pharyngeal cavity is very hypoplastic, pharyngeal pouches 2-4 are not recognizable, the first is abnormal, the second pharyngeal arch is severely hypoplastic and the 3rd-6th arches are not recognizable (Jerome and Papaioannou, 2001; Kelly et al., 2004; Vitelli et al.,2002). The zebrafish Tbx1 mutant also has similar pharyngeal defects (Piotrowski et al.,2003; Piotrowski and Nusslein-Volhard, 2000). These phenotypic observations led to hypothesize a role of Tbx1 in pharyngeal segmentation(Baldini, 2002). The timed deletion data reported here demonstrated that Tbx1 is required during segmentation because elimination of Tbx1 while the segmentation is in progress effectively stops it. Tbx1 timed deletion is followed by downregulation of the proliferative activity of endodermal cells, indicating that at least one role of Tbx1 in segmentation concerns regulation of endodermal cell proliferation. We propose a model(Fig. 6D) in which the cranial-to-caudal wave of Tbx1 expression would cause a wave of expansion of the pharyngeal endoderm that allows migrating neural crest-derived cells to populate (and thus shape) the pharyngeal arches. This model is also consistent with the hypothesis that the pharyngeal endoderm has a primary role in pharyngeal segmentation(Graham, 2001; Graham and Smith, 2001). However, because Tbx1 is also expressed in pharyngeal mesoderm and,transiently, in the pharyngeal ectoderm, we cannot exclude a role for the gene in these tissue during segmentation and (or) in triggering extracellular signals required for endodermal cell proliferation.

Tbx1 mutants and individuals with DGS have complex phenotypes. One of the goals of our time-based dissection of the phenotype is to reduce complexity so that individual abnormalities can be studied more effectively. In addition, the availability of an inducible deletion system allows one to test cellular phenotypes (e.g. proliferative activity) shortly after somatic gene deletion, hence reducing the chance of possible adaptive changes that may be effected in germline mutants. We speculate that Tbx1 may have a similar role in different tissues. Our finding of reduced proliferation in the pharyngeal endoderm is consistent with the finding that tissue-specific deletion of Tbx1 in the precursors of OFT cardiomyocytes is associated with reduced cell proliferation in the splanchnic mesoderm/secondary heart field (Xu et al.,2004). In addition, it has been shown that overexpression of Tbx1 in the OFT causes expansion of cellularity in that organ(Hu et al., 2004). Furthermore, the inner ear of Tbx1^-/- embryos presents a defect of otic epithelial cell expansion(Vitelli et al., 2003). Thus,expansion of specific but different cell populations may be a general mechanism of action of Tbx1. Future experiments should determine the molecular effectors that mediate the role of Tbx1 in cell proliferation.

We thank H. Sobotka, G. Ji, M. Reese and P. Terrell for valuable technical assistance; A. McMahon and P. Soriano for making available the TgCAGG-CreER™ and R26R mouse lines, respectively; and R. Behringer, E. Lindsay, J. Martin, R. Schwartz and F. Vitelli for critical reading of the manuscript. This work was funded by grants HL51524, HL64832, HL67155 and DE14521 from the National Institutes of Health (to A.B.).