Regionalized Twist1 activity in the forelimb bud drives the morphogenesis of the proximal and preaxial skeleton

Developmental Biology 362 (2012) 132–140

Contents lists available at SciVerse ScienceDirect

Developmental Biology journal homepage: www.elsevier.com/developmentalbiology

Regionalized Twist1 activity in the forelimb bud drives the morphogenesis of the proximal and preaxial skeleton David A.F. Loebel a, b,⁎, Angelyn C.C. Hor a, Heidi Bildsoe a, b, Vanessa Jones a, You-Tzung Chen c, Richard R. Behringer c, Patrick P.L. Tam a, b a b c

Embryology Unit, Children's Medical Research Institute, Sydney, Australia Sydney Medical School, University of Sydney, Sydney, Australia Department of Molecular Genetics, MD Anderson Cancer Center, University of Texas, Houston, USA

a r t i c l e

i n f o

Article history: Received for publication 9 June 2011 Revised 22 November 2011 Accepted 29 November 2011 Available online 8 December 2011 Keywords: Twist1 Tissue patterning Forelimb Conditional mutant Polydactyly Mouse

a b s t r a c t Development of the mouse forelimb bud depends on normal Twist1 activity. Global loss of Twist1 function before limb bud formation stops limb development and loss of Twist1 throughout the mesenchyme after limb bud initiation leads to polydactyly, the ulnarization or loss of the radius and malformations and reductions of the shoulder girdle. Here we show that conditional deletion of Twist1 by Mesp1-Cre in the mesoderm that migrates into the anterior-proximal part of the forelimb bud results in the development of supernumerary digits and carpals, the acquisition of ulna-like characteristics by the radius and malformations of the humerus and scapula. The mirror-like duplications and posteriorization of pre-axial tissues are preceded by disruptions to anterior–posterior Shh, Bmp and Fgf signaling gradients and dysregulation of transcription factors that regulate anterior–posterior limb patterning. © 2011 Elsevier Inc. All rights reserved.

Introduction Limb morphogenesis in the mouse embryo begins with the formation of a bud, followed by outgrowth, the shaping of the autopod and zeugopod segments, and the specification and formation of digits. Proximal–distal growth and patterning are controlled by Fgf10 and Bmp signaling in the mesenchyme and Fgf4/8 signals from the apical ectodermal ridge (AER) (Bénazet and Zeller, 2009). Specification of the number and identity of digits is controlled by their position in the gradient of Shh signaling from the zone of polarizing activity (ZPA) and the duration of exposure to the signal. Shh signaling is in turn regulated by feedback loops involving Fgf signals from the AER and Bmp in the mesenchyme (Robert, 2007; Sun et al., 2000). The characteristics of skeletal elements are specified by transcription factors including members of the Hoxd, Alx and Msx families that are subject to regulation by the Fgf, Shh and Bmp signaling pathways (Bensoussan-Trigano et al., 2011; Kuijper et al., 2005; Zakany et al., 2007). Twist1, a basic helix–loop–helix transcription factor, plays a key role in integrating the actions of signaling pathways and transcription factors during limb development (O'Rourke and Tam, 2002; Zuniga et

⁎ Corresponding author at: Embryology Unit, Children's Medical Research Institute, Locked Bag 23, Wentworthville, NSW 2145, Australia. Fax: + 61 2 96872120. E-mail address: [email protected] (D.A.F. Loebel). 0012-1606/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2011.11.020

al., 2002). In Twist1−/− mouse embryo outgrowth of the forelimb buds is impaired and they are smaller than wild-type limbs by E10 (Chen and Behringer, 1995). Poor forelimb outgrowth in Twist1 −/− embryos is accompanied by reduced expression of Fgf genes in the AER, and weaker Bmp and Shh signaling in the mesenchyme (O'Rourke et al., 2002; Zuniga et al., 2002). Outgrowth of the hindlimb buds is affected to a lesser degree, and the disruptions to Fgf, Bmp and Shh signaling are less severe (O'Rourke et al., 2002). Further investigation into the morphogenetic outcome of these defects has been prevented by the death of Twist1 −/− embryos at approximately E11.5, not long after the limb pattern emerges. In contrast, partial reductions in Twist1 activity affect anterior– posterior patterning of the limb buds. In Twist1-heterozygous mice, this manifests as pre-axial polydactyly of the hindlimbs (O'Rourke et al., 2002). Homozygosity for a hypomorphic Twist1 point mutation affecting protein–protein interactions (Ska10, also known as Charlie Chaplin, CC) or a combination of Ska10 and null alleles results in loss of digits, and abnormal or missing long bones of the forelimb (Krawchuk et al., 2010). However, conditional loss of Twist1 function widely in the mesenchyme after the limb bud has begun to develop, driven by Prx1-Cre, leads to a disruption of A–P patterning that causes polydactyly and ulnarization of the radius, as well as abnormalities of the humerus and scapula, (Krawchuk et al., 2010; Zhang et al., 2010). Genetic and biochemical studies suggest that Twist1 influences anterior–posterior patterning of the limbs through interactions with ETV-family proteins (Zhang et al., 2010) that influence Shh signaling.

D.A.F. Loebel et al. / Developmental Biology 362 (2012) 132–140

In addition, Twist1 interacts genetically with Gli3, which encodes a Shh signaling repressor (O'Rourke et al., 2002). These findings suggest that Twist1 function is critical for the acquisition of anterior limb characteristics and the formation of pre-axial limb structures. Whether Twist1 is required locally in the anterior tissues of the limb bud is not yet known. Precursors of limb bud tissues are derived from the lateral plate mesoderm, which forms the skeletal elements (Koussoulakos, 2004) and the hypaxial dermomyotome, which contributes to the muscles (Buckingham et al., 2003). We have discovered that Cre recombinase expressed from a knock-in allele at the Mesp1 locus (Saga et al., 1999) leads to the loss of Twist1 early in the mesoderm that eventually populates the mesenchyme in the anterior and proximal regions of the forelimb bud, providing a unique experimental system in which to investigate the contribution of Twist1 in a specific region to the anterior–posterior patterning of the limb. We show that loss of Twist1 function in this restricted domain within the forelimb bud mesenchyme results in anterior–posterior patterning defects reminiscent of those that result from deletion of Twist1 in the entire limb bud mesenchyme (Krawchuk et al., 2010; Zhang et al., 2010), but less severe defects in the humerus and scapula. The patterning defects are likely to be consequences of reduced Gli3 and Alx4 expression and disruptions to Fgf, Bmp and Shh signaling during a critical phase of limb patterning. Material and methods Mouse strains and genotyping Twist1 3loxPneo/3loxPneo and Twist1 del/+ mice were maintained and genotyped as previously described (Bildsoe et al., 2009; Chen et al., 2007). Mesp1-Cre mice (Saga et al., 1999) were maintained on a C57Bl/6 background and genotyped by PCR with the following primers: 5′-CTGACCGTACACCAAAATTTGCCTG-3′ (CreF), 5′-GATAATCGCGAACATCTTCAGGTTC-3′ (CreR). We first crossed Mesp1-Cre mice (Saga et al., 1999) to Twist1 del/+ mice to generate Twist1 del/+; Mesp1 cre/+ mice. To generate embryos with a mesoderm-specific Twist1 deficiency, Twist1 del/+; Mesp1 cre/+ mice were crossed with Twist1 3loxPneo/3loxPneo mice. Conditional knockout (CKO) embryos of Twist1 3loxPneo/del;Mesp1 Cre/+ genotype were compared with Twist1 3loxPneo/+, Mesp1 +/+ (WT) embryos and Twist1 3loxPneo/del;Mesp1 +/+ (heterozygous) embryos. To trace the distribution of the descendants of Mesp1-Cre expressing cells, we initially crossed the Rosa26R line (Soriano, 1999) to the Twist13loxPneo/3loxPneo line generating Twist1 3loxPneo/+; Gt(ROSA)26Sor tm1Sor/+ offspring which were to intercrossed generate Twist13loxPneo/3loxPneo; Gt(ROSA)26Sor tm1Sor/tm1Sor offspring. These were crossed with Twist1 del/+; Mesp1 cre/+ mice to generate βgalactosidase positive CKO and WT embryos for analysis. The Rosa26R allele was detected by PCR for lacZ with the following primers: LacZ-F: 5′ TTATCGATGAGCGTGGTGGTTATGC 3′ LacZ-R: 5′GCGCGTACATCGGGCAAATAATATC 3′ Bone and cartilage staining Embryos were collected at E16.5 and E17.5 in PB1 (Kinder et al., 2000) and rinsed in cold PBS (without calcium and magnesium). Bone and cartilage were stained with alizarin red and alcian blue (Hogan et al., 1994). Stained specimens were washed, stored and photographed in aqueous 20% ethanol: 20% glycerol. Images were captured using a SPOT camera and Leica microscope.

133

ß-galactosidase reporter staining For whole mount staining to detect ß-galactosidase activity, embryos were fixed for at least 2 h in glutaraldehyde solution (Igepal CA630 0.02%, Sodium Deoxycholate 0.01%, Glutaraldehyde, 0.2%, EGTA 5 mM, MgCl2 2 mM), washed in X-gal washing buffer briefly before being incubated in X-gal staining solution at 37 °C for 2–3 h for color development (Watson et al., 2008). The embryos were then washed twice in X-gal washing buffer and fixed in 4% PFA. For β-galactosidase staining of cryosections, limbs were dissected from E13.5 embryos in PB1 medium, rinsed in PBS and fixed in 4% PFA overnight at 4 °C. The limbs were then infiltrated with 15% sucrose/ PBS overnight, followed by 25% sucrose/PBS overnight at 4 °C. Tissues were embedded in equal volumes of 25% sucrose/PBS and OCT embedding medium (Tissue Tek). Sections were cut at 5–8 μm on a cryostat (M1900, Leica), collected on Superfrost Plus slides (Menzel– Glaser) and stored at −80 °C until needed. Sections were thawed at room temperature, rinsed in Lac Z buffer 3 times and then incubated in X-gal staining solution at 37 °C overnight for color development followed by washing in X-gal washing buffer and water, counterstained with nuclear fast red and mounted in Ultra-mount No.4 (Fronine). Immunofluorescence Embryos between E9.5–10.5 were dissected in PB1 and rinsed in cold PBS. The limb buds were fixed in 4% PFA overnight at 4 °C and cryoembedded as previously described (Bildsoe et al., 2009). Cryosections were stored at −20 °C until use. Prior to staining, slides were allowed to reach room temperature, washed three times for 3 min. in washing solution (PBS plus 0.1% BSA, 0.02% triton X-100) and blocked with 3% Donkey serum, 0.1% BSA 0.02% triton in PBS for 1 h at room temperature. Slides were incubated overnight at 4 °C with mouse anti-Twist1 (Abcam, 1/50 dilution) and rabbit anti-Ki67 (Abcam, 1/100 dilution) blocking solution and washed three times in washing solution for 5 min at room temperature. Secondary antibodies (AlexaFluor-488 Donkey anti-mouse and AlexaFluor-594 goat anti-rabbit, Invitrogen) diluted 1/500 in PBS plus 0.2% triton, 1% donkey serum were added in washing solution for 1–2 h at room temperature. Slides were then washed three times for 5 min in washing solution, stained with DAPI for 10 min, washed again for 5 min in PBS before mounting in 70% glycerol/PBS. Ki67 positive and DAPI stained nuclei were counted in identically sized and shaped regions within the Twist1-deleted zone of CKO limb buds and equivalent regions of wild-type limb buds. In situ hybridization Riboprobes for whole mount in situ hybridization to E10.5 or E11.5 mouse embryos were made from plasmid clones containing fragments of the following cDNAs: Alx1, Alx4, Emx2, Fgf4, Fgf8, Fgf10, Gli3, Grem1, Gsc, Hand2, Hoxd13, Msx1, Pbx1, Ptch1, Shh and Twist1. To generate a riboprobe to detect Cre transcript, a fragment of the open reading from was amplified using the following primers: Cre-F: 5′ CCGTACACCAAAATTTGCCTGCATT 3′ Cre-R: 5′ ACCATTGCCCCTGTTTCACTATCCA 3′ The amplified fragments were gel purified and re-amplified using primers identical to the first round of amplification, except that the reverse primer contained a T7 promoter sequence at the 5′ end. This product was gel purified and used for riboprobe generation. Digoxigenin-labeled riboprobes were synthesized using Ampliscribe (Epicentre Technologies). Automated whole mount in situ hybridization was carried out using an InsituPro machine (Intavis AG) as described previously (Bildsoe et al., 2009; Loebel et al.,

134

D.A.F. Loebel et al. / Developmental Biology 362 (2012) 132–140

2004). Stained embryos were washed in 0.1% Tween20 in H2O, fixed in 4% PFA and photographed. Measurements were made using ImageJ and statistics calculated with Microsoft Excel. Cell death analysis Apoptotic cells were detected by whole mount TUNEL staining with the ApopTag Plus Peroxidase In situ Apoptosis kit (Millipore) following published methods (Martinez-Barbera et al., 2002). Briefly embryos were dissected in PB1, rinsed in cold PBS and fixed in 4% PFA overnight at 4 °C. For the detection step the embryos were incubated in blocking buffer containing 2% Blocking Reagent (Roche) and 20% heat inactivated fetal calf serum in PBS/0.1% Tween. After antibody labeling, the embryos were washed overnight at 4° in 2 mg/ml BSA in PBS, 0.1% Tween with at least 4 changes. After treatment embryos were fixed in 4% PFA, imaged using a Leica microscope and SPOT camera and sectioned. Results Mesp1-Cre labels cells in the anterior mesoderm Cre transcript expressed from the Mesp1 locus was detected in the nascent mesoderm adjacent to the primitive streak of E7.5 embryos (Fig. 1A), but at E8.5 and E9.5 Cre expression was restricted to a stripe of tissue at the cranial end of the presomitic mesoderm (Figs. 1B, C). This expression pattern recapitulates that of the endogenous Mesp1 transcript (Saga et al., 1996; Saga et al., 1999). In Mesp1 Cre; Rosa26R embryos at E7.5, β-galactosidase labeled cells were found throughout the mesoderm, consistent with the anterior migration of mesoderm cells that had previously expressed Cre [Figs. 1D, D(i), E]. By E7.75, stained cells were found predominantly in the anterior half of the embryo, indicating that the mesoderm present at E7.5 contributes predominantly to these tissues (Fig. 1F). At E8.5, the heart and the majority of the cranial mesoderm strongly expressed β-galactosidase [Figs. 1G, G(i)]. Widespread labeling of the lateral plate mesoderm of the splanchnopleure and somatopleure continued into the trunk region, but the paraxial mesoderm became progressively more sparsely labeled caudally [Figs. 1G, G(ii– iv)]. At the level of the hindgut, only the blood vessels, body wall and extraembryonic mesoderm expressed β-galactosidase [Fig. 1G (iv)]. At E9.0-E9.5, a similar rostral–caudal gradient of staining intensity was observed in the paraxial mesoderm, with the cranial, cervical and heart mesoderms being heavily labeled [Figs. 1I, I(i), J, J(i)], but progressively sparser labeling in caudal paraxial mesoderm [Figs. 1I (ii–iii), J(ii–iii)]. At E9.5 we observed β-galactosidase activity in the forelimb bud, contiguous with the stained lateral and ventral mesoderm populations [Fig. 1J(ii)]. Mesp1-Cre mediates ablation of Twist1 in cells that populate the anterior-proximal limb bud mesenchyme Closer examination of β-galactosidase staining in the forelimb buds at E9.5 revealed activity throughout the majority of the forelimb bud, with strongest staining toward the anterior and a patchy appearance in the posterior limb-bud mesenchyme (Figs. 2A, C, supplementary Fig. S1A). In the forelimb bud of E10.5 embryos, the distributions of β-galactosidase positive cells varied, but in most cases labeled cells were found predominantly in the mesenchyme of in the anterior two-thirds of the forelimb bud (Figs. 2E, G, supplementary Fig. S1 B). By E13.5, marked cells in the autopod were predominantly found in the tissues of digits 1, 2 and 3 and at a lower frequency in the postaxial digits (Figs. 2I, K). βgalactosidase positive cells contributed strongly to the radius, but patchily to the ulna, humerus and scapula (Fig. 2M).

Fig. 1. Mesp1-Cre is expressed in the anterior mesoderm. (A–C) Mesp1-Cre expression, revealed by in situ hybridization for Cre transcript in the nascent mesoderm near the primitive streak at E7.5 (A, embryo shown with anterior to the left) and in a restricted domain in the pre-somitic mesoderm at E8.5 (B) and E9.5 (C). (D–J) β-galactosidase from the Rosa26R reporter transgene detected widely in the mesoderm of embryos at E7.5 (D, E, shown with anterior to the left in D and D(i), and en face in E), and predominantly in anterior, cardiac and lateral mesoderm populations at E7.75 (F), E8.5 (G, H) E9.0 (I) and E9.5 (J). D (i), G (i–iv), I (i–iii), J (i–iii) show sections through embryos at the planes indicated in D, G, I and J respectively. Abbreviations: da, dorsal aorta; fig, foregut; h, heart; hg, hindgut; lb, forelimb bud; m, mesoderm; mg, midgut; ov, otic vesicle; pm, paraxial mesoderm; ps, primitive streak; so, somatopleure; sp, splanchnopleure; va, vitelline artery.

In conditional Twist1 mutant embryos (Mesp1 Cre/+; Rosa26R; Twist1 3loxPneo/del = CKO), β-galactosidase positive cells had a more restricted distribution in the forelimb buds. At E9.5, the distribution of labeled cells contracted proximally and anteriorly (Figs. 2B, D, supplementary Fig. S1 A). At E10.5 the distribution of labeled cells

D.A.F. Loebel et al. / Developmental Biology 362 (2012) 132–140

135

Fig. 2. Mesp1-Cre causes loss of Twist1 activity in a restricted region of the limb bud mesenchyme. (A–N) Staining for Rosa26R activity in E9.5 (A–D), E10.5 (E–H) and E13.5 forelimb buds (I–N). Rosa26R activity is detectable in a broader domain of anterior mesenchyme in the wild-type (A, C, E, G, I, K, M) than the Mesp1-Cre;Twist1 CKO embryos (B, D, F, H, J, L, N). (C, D) Sections through stained wild type and CKO limb buds at E9.5 and (G, H) at E10.5 reveal a more restricted distribution of labeled cells in CKO forelimbs. (K–N) Staining for Rosa26R activity on frozen sections of E13.5 limb buds shows that in wild-type embryos (K, M), labeled cells contribute strongly to digits 3, 4 and 5 and to the radius, and are found throughout the other digits and in the ulna, humerus and scapula. In CKO embryos labeled cells contribute mainly to digits 4 and 5, and to the radius, with weaker contributions to other digits and little contribution to the ulna, humerus or scapula (L, N). (O–P) Immunofluorescent staining for Twist1 and Ki67 on cryosections of wild-type (O) and CKO (P) forelimb buds at E9.5 revealing a region of Twist1-negative cells in the proximal mesenchyme with a greater concentration of these cells toward the anterior of the limb bud and a lower frequency of Ki67 in the CKO limb buds than in an equivalent region of the wild-type limb bud. The left panel shows Twist1 immunostaining, the middle panel shows merged DAPI and Ki67 and the right panel shows a merged image of all three. (Q–R) Immunofluorescent detection of Twist1 (left and merged with DAPI, right) at E10.5 in wild-type (Q) and CKO limb buds (R). Abbreviations: h, humerus; r, radius; s, scapula; u, ulna.

was variable but, like the E9.5 embryo, occupied a territory that was generally more restricted toward the anterior and proximal regions of CKO forelimb buds (Figs. 2F, H, supplementary Fig. S1 B). However,

some labeled cells were present in other parts of the limb bud mesenchyme. At E13.5, β-galactosidase positive cells congregated mainly in the two most pre-axial digits (Figs. 2J, L) in CKO embryos. β-

136

D.A.F. Loebel et al. / Developmental Biology 362 (2012) 132–140

galactosidase positive cells colonized the radius of the CKO embryo as in the wild type embryo, but fewer were found in the ulna, humerus and scapula (Figs. 2M–N). Examination of Twist1 expression by whole mount in situ hybridization confirmed the region-specific reduction in Twist1 expression at E9.5 and E10.5 (Supplementary Fig. S2). Anti-Twist1 immunostaining at E9.5 showed Twist1-deficient cells were present along most of the length of the limb bud, but more concentrated proximally and in the anterior two-thirds of the limb buds (Figs. 2O, P), reflecting the pattern of Rosa26R reporter activity in CKO limb buds. At E10.5, Twist-deficient cells were mostly localized to the proximal half of the forelimb bud, and mostly in the anterior sector (Figs. 2Q, R). To test whether Twist1-deficient cells in CKO limb buds might be less proliferative we compared Ki67 immunostaining of the Twist1deficient regions to equivalent regions in wild-type limb buds (Figs. 2O–P). At E9.5, the proportion of Ki67-positive Twist1-deficient cells in CKO limb buds (mean 2.1% ± s.e.m. 0.20, n = 5) was significantly lower than in wild-type limb buds (3.2% ± 0.43, n = 4; p = 0.02 by a two-tailed t-test). At E10.5, there was a similar trend of reduced proliferation, though not statistically significant (wild-type, 3.10% ± 0.3, n = 4; CKO 1.94% ± 0.41, n = 4; p = 0.091). At E10.5, the cell density was significantly greater in Twist1-deficient mesenchyme than in corresponding regions of wild-type forelimb buds (wildtype, 3965.12 nuclei/mm 2, n = 4; CKO 4680.23 nuclei/mm 2, n = 4, p = 0.008). Analysis of apoptosis by whole mount TUNEL staining of CKO embryos revealed no increased cell death in the limb bud mesoderm at E9.5, or in the prospective limb field at E8.75, prior to limb bud initiation (Fig. S3. A–D). An increase in cell death was observed at E9.5 in the β-galactosidase positive paraxial mesoderm anterior to the forelimb bud, where Twist1 expression was lost (Fig. S3C–H). Regionalized loss of Twist1 activity affects patterning of pre-axial structures By E13.5, CKO forelimb buds were broader and contained more digit primordia than their wild-type counterparts (Figs. 2I, J). When examined at E17.5, the forelimbs of CKO embryos had 6–8 digits (Supplementary Fig. S4), ranging from an apparent duplication of digit 1 (Supplementary Fig. S4 A) to extensive polydactyly with bifurcated digits (Supplementary Fig. S4 E). Skeletal preparations revealed that the major skeletal elements of the limbs were present, but malformed in CKO embryos (Figs. 3A, B). Anterior–posterior asymmetry was less pronounced in the autopod of CKO embryos than in wild-type embryos. The most pre-axial digit was often longer in CKO embryos than in wild-type embryos (Figs. 3C, D). While the phalanges were not yet ossified, the metacarpals of all the central digits of E17.5 CKO forelimbs were ossified as were the three central metacarpals (II, III and IV) of the wild type limb (Figs. 3E, F). The arrangement of carpal bones in the wrist of the CKO forelimb differed from the wild type limb. Supernumerary bones were present in the distal row of carpals (Figs. 3C–F) and the proximal carpals of the CKO limb were flattened and the pisiform was reduced (Figs. 3E, F). In some cases the appearance of the anterior carpals of the CKO limb was consistent with a mirror duplication of posterior elements (Figs. 3E, F), but not in others (e.g. Fig. 3D). In 11/ 12 limbs examined, the radius was abnormal, with some morphological characteristics of the ulna including an olecranon process, resulting in a more symmetrical elbow joint (Figs. 3A, B, G, H). The humerus was short, curved and lacking the deltoid tuberosity in CKO embryos (10/ 10 limbs examined). Its proximal end was often fused with the scapula to form an immobile shoulder joint (6/9 limbs; Figs. 3A, B, I, J). The scapula of CKO embryos was hypo plastic with a small or absent spine (7/7 limbs; Figs. 3I, J). The data show that loss of Twist1 in the anterior mesenchyme results in significant changes in the anterior–posterior patterning of the distal pre-axial structures and affects the morphogenesis of proximal skeletal elements of the limb.

Fig. 3. Polydactyly and posteriorization of the limb skeleton in CKO embryos. (A–D) Alcian blue staining of E16.5 wild-type (A, C) and CKO embryos (B, D) shows polydactyly, a longer digit 1, ulnarization of the radius, foreshortened and bent humerus and a smaller scapula in the CKO forelimb. (E–J) Alcian blue and alizarin red staining of forelimb skeleton at E17.5. The wild-type autopod (E) displays anterior–posterior asymmetry of carpal and metacarpal elements, and appearance of ossification centers in the central three digits. The CKO autopod (F) contains two supernumerary digits and the middle five digits all display ossification. Additional carpals are formed and the morphology of preaxial and postaxial counterparts (marked pair-wise by asterisks of the same color) are reminiscent of mirror duplications. The wild-type ulna (G) is distinguished from the radius by its olecranon process at the elbow end, but the radius of the CKO embryo (H) develops an olecranon-like extension at the elbow end (black arrowhead) that is reminiscent of those of the radius. The wild-type humerus and scapula (I) are larger than their CKO counterparts (J). The CKO humerus is curved and lacks the anterior deltoid tuberosity (see G: wild type, yellow arrow head). The scapula has a malformed and reduced spine (see G: wild type, red arrow head) and the bones are fused at the shoulder joint (white arrowhead). Digits are numbered I–V. Abbreviations: ca, capitate; hm, hamate; p, pisiform; r, radius; s, scapula; sc-cn, scaphoid-centrale; tq, triquetral; tm, trapezium; tz, trapezoid; u, ulna.

Altered transcription factor expression across the anterior–posterior axis of the forelimb buds To analyze the origins of the anterior–posterior patterning defect in the limb buds, we examined the expression of a set of transcription factor genes that are differentially expressed across the limb bud at E10.5. CKO embryos were compared to both wild type and Twist1heterozygous embryos to account for any differences in gene expression that could be due to a general reduction in Twist1 expression, since the CKO embryos are heterozygous for a Twist1-deleted allele in all tissues. Alx1 is expressed in the anterior-proximal limb bud mesenchyme of wild-type embryos but in CKO limb buds it was

D.A.F. Loebel et al. / Developmental Biology 362 (2012) 132–140

confined to a smaller domain at the anterior margin (Figs. 4A–C; the proportion of anterior limb perimeter covered by Alx1 expression domain is shown as mean ± SEM: wild-type, 0.226 ± 0.009, n = 6; het, 0.242 ± 0.008 n = 2; CKO, 0.163 ± 0.011 n = 4; two-tailed t-test comparing wild-type and CKO, p = 0.0025). Expression of Emx2 and Pbx1, which act upstream of Alx1 in scapula development, (Capellini et al., 2010) was similar in wild-type, heterozygous and CKO embryos (Fig. S5 A–F). In contrast, the expression of Gsc, which is also required for normal development of the shoulder joint (Belo et al., 1998), was slightly reduced in the proximal mesenchyme of CKO and heterozygous forelimb buds (Fig. S5 G–I). Alx4 is normally expressed in the anterior limb bud mesenchyme and loss of function mutations in Alx4 result in polydactyly (Qu et al., 1998). Alx4 expression was reduced in 6/6 CKO limb buds but not in heterozygous buds (2/2, Figs. 4D– F). In heterozygous forelimb buds, the expression domain of the posterior transcription factor Hoxd13 expanded toward the distal limb region (2/2, Figs. 4G–H), suggesting that the overall reduction in Twist1 expression resulted in de-repression of Hoxd13 in some cells. In CKO limb buds the ectopic expression extended into the anterior rim of limb bud (6/6, Fig. 4I). In contrast, expression of Hand2, which interacts with Twist1 (Firulli et al., 2005), was unchanged the posterior mesenchyme in 2/2 heterozygous and 6/6 CKO limb buds (6/6, Fig. S5. J–L). These data indicate that Twist1 is required for maintaining the expression of genes in the anterior mesenchyme, as well repressing some genes that are normally expressed in the posterior mesenchyme.

Fig. 4. Expression of transcription factor genes associated anterior posterior forelimb patterning at E10.5. Alx1 (A–C) expression covers a larger domain of the anterior proximal mesenchyme in the forelimb bud of the wild-type (A) and heterozygous embryos (B) than CKO embryo (C). The length of the Alx1 expression domain in the anterior limb bud (black bar) was measured and the proportion of the distance from the anterior margin to the distal tip of the limb bud calculated (see text). (D–F) Alx4 is expressed in the anterior one-third mesenchyme of the forelimb bud in wild-type (D) and Twist1 heterozygous embryos (E), but is reduced and confined to the anterior margin in CKO embryos (F). (G–I) Hoxd13 expression, which marks the posterior mesenchyme in the wild type limb bud (G), expands anteriorly in the distal mesenchyme of the heterozygous (H) and further in CKO forelimb buds (I). Arrows denote the region of expanded expression.

137

Abnormal signaling in CKO limb buds The Shh, Bmp and Fgf signaling pathways are essential for tissue outgrowth and patterning during limb development. At E10.5 Gli3, which encodes a transcription factor that represses the downstream response to Shh signaling, is robustly expressed in an anterior–posterior gradient in mesenchyme of the wild type forelimb bud (Fig. 5A). Whereas no change in Gli3 expression was observed in Twist1 heterozygous buds (2/2), its expression was reduced in the anterior mesenchyme of 6/6 CKO buds (Figs. 5B, C). Ptch1, a repressor and transcriptional target of Shh signaling, was expressed in the posterior mesenchyme of the limb bud of wild type and heterozygous embryos

Fig. 5. Altered Shh, Bmp and Fgf signaling in CKO embryos. (A–C) Gli3 is expressed most strongly in the anterior limb bud mesenchyme, becoming weaker posteriorly (A). Expression is not changed in heterozygous embryos (B), but is reduced in the anterior region (asterisk) of the CKO limb bud. (D–F) Ptch1 expression is similar in wild-type (D) and heterozygous (E) embryos, but is found ectopically in the anterior mesenchyme in 2/8 CKO limb buds (F). (G–I) Grem1 is expressed in the distal limb mesenchyme, with a broader domain on the posterior side in wild-type limb bud (G). An anterior expansion was observed in heterozygous limb bud (H) which was further augmented in CKO limb bud (I). (J–L) Msx1 is strongly expressed in the anterior mesenchyme (J) in wild-type and heterozygous limb buds (K). Expression was reduced (asterisk) in CKO limb bud (L). (M–O) Fgf10 is expressed in the posterior and distal mesenchyme in wild-type (M) and heterozygous (N) limb buds, but additional expression is found in the anterior-proximal mesenchyme in CKO limb bud (O). White asterisks indicate regions of reduced expression in CKO embryos. Black arrows indicate regions of ectopic or expanded expression.

138

D.A.F. Loebel et al. / Developmental Biology 362 (2012) 132–140

(Figs. 5D, E). In 2/8 CKO limb buds examined, Ptch1 was expressed ectopically in the anterior mesenchyme (Fig. 5F). Consistent with this, ectopic Shh expression was detected only in 1/8 CKO forelimb buds at E11.5 (Fig. S5 M–O). Grem1, encoding a Bmp signaling antagonist, was expressed in the distal mesenchyme of the wild type forelimb bud (Fig. 5G). In Twist1 heterozygous embryos, a slight anterior spread of Grem1 expression was observed (Fig. 5H, 2/2), which was more pronounced in CKO limb buds (Fig. 5I, 6/6). Consistent with this, expression of a Bmp responsive gene, Msx1, was reduced in the anterior mesenchyme of CKO forelimb buds (Figs. 5J–L, 5/6). Fgf10 expression, which is strongest in the posterior limb bud mesenchyme, expanded anteriorly in 6/6 CKO embryos but not in 2/2 heterozygous embryos (Figs. 5M–O). Fgf4 and Fgf8 expression in the AER is required for normal outgrowth of the limb buds. In CKO embryos, both genes were expressed, indicating that the conditional ablation of Twist1 has no early effect on the formation of the AER (Fig. S5 P–W). Fgf8 expression, which is activated by Fgf10 signaling (Min et al., 1998) expanded toward the anterior margin of the forelimb bud (Fig. S5 P–R; the proportion of unstained apical surface anterior to the Fgf8 expression domain is shown as mean± SEM: wild type, 0.132 ± 0.009, n = 6; het, 0.137 ± 0.011 n = 2; CKO, 0.0894 ± 0.009, n = 6; two-tailed t-test comparing wild-type and CKO, p = 0.007). Fgf8 expression remained within a well-defined band in the distal limb bud indicating that there was no appreciable dorso-ventral expansion of the AER in CKO embryos (Fig. S5 S, T). Fgf4 expression was reduced in the CKO buds compared to the wild type (Fig. S5 U–W, 6/6). These results are consistent with previous findings of a reciprocal relationship in the expression of Fgf4 and Fgf8 (Delgado et al., 2008; Lewandoski et al., 2000). Together, these findings show that loss of Twist1 activity in the anterior proximal mesenchyme disrupts the signals that regulate anterior–posterior patterning, resulting in polydactyly and the posteriorization of the pre-axial structures. Discussion In this study we have used Mesp1-Cre to produce a conditional deletion of Twist1 in the mesoderm during gastrulation. These cells contribute to anterior mesodermal populations in the embryo, including a restricted region of the forelimb bud. Mesp1-Cre is expressed broadly in the nascent mesoderm only during gastrulation,

WT

reflecting the expression pattern of the wild-type Mesp1 transcript (Saga et al., 1996). Descendants of this population populate the heart, cranial mesoderm, the paraxial mesoderm in the cervical region rostral to the forelimb bud and the lateral plate mesoderm that contributes to the body wall and forelimb. These results are in agreement with previous fate mapping data based on cell transplantation, showing that the mesoderm of late-streak stage embryos contributes primarily to heart, cranial, lateral plate and anterior somite mesoderm in the early somite-stage embryo (Parameswaran and Tam, 1995). Since Cre expression is not detectable in the forelimb bud, βgalactosidase-labeled cells in the limb bud probably share ancestry with the Rosa26R expressing cells adjacent to the limb field. A recent cell labeling study (Wyngaarden et al., 2010) showed that cells entering the limb bud move generally in a caudal direction. This is consistent with some of the labeled cells in the forelimb bud being derived from the β-galactosidase-positive mesoderm anterior to the forelimb bud. It is also likely that the β-galactosidase-positive cell population medial to the forelimb bud contributes to the bud mesenchyme. Wild-type forelimbs also contain unlabeled cells, predominantly toward the posterior part of the limb bud. Although we observed variations in the distribution of reporter labeled cells in Twist1 CKO embryos at E10.5, in general they were more concentrated in the mesenchyme in the anterior and proximal parts of the limb bud. By E13.5, the labeled cells (which are likely to be lacking Twist1) in CKO embryos clearly contributed most strongly to pre-axial skeletal structures, notably the radius and the two most anterior digits. Since the radius and preaxial autopod elements are the most severely affected, having undergone duplications or loss of anterior identity, Twist1 may be acting autonomously within the progenitor cells to regulate anterior–posterior identity. This may be due to reduced expression of transcription factors (such as Aristaless-related genes) in Twist1-null cells. However, Twist1 may also be acting in a non-cell-autonomous fashion to affect patterning and identity in the limb. Consistent with this, the effects on skeletal patterning spread beyond the zones where Twist1 expression has been lost and our results show that the signaling environment in the anterior half of the limb bud has been disrupted. Therefore, localized loss of Twist1 can disrupt patterning non-cell-autonomously by perturbing signaling gradients across the limb (Fig. 6). In both Mesp1-Cre and Prx1-Cre CKO embryos (Krawchuk et al., 2010; Zhang et al., 2010), supernumerary digits were formed and

CKO Alx1

Alx1

4

Twist1

1,

7

2

Shh 6

3, 4

1

Gli3

Gli3 Shh BMP signaling

Alx4

Ptch h1

3

BMP signaling

Alx4 4, 6

5 5

Fgf10 Grem1

Fgf10

Hoxd13 Grem1

8

Hoxd13

BMP signaling

BMP signaling

9 10

Ptch1

11

Shh

Shh

Ptch1

Fig. 6. A proposed network of molecular interactions regulating anterior–posterior patterning of the limb bud. In the wild-type (WT) limb bud, Twist1 is required for normal expression of Gli3 and Alx4 in the anterior limb mesenchyme, consistent with previous limb bud expression studies. Gli binding sites are present upstream of Alx1, suggesting that Gli3 might regulate Alx1 expression. Gli3 and Alx4 repress Fgf10, Grem1 and Hoxd13, restricting their expression to the posterior or distal mesenchyme. Expanded Grem1 expression in CKO limb buds causes a reduction in Bmp signaling in the anterior limb, in addition to alterations to Shh signaling downstream of Gli3. Ligand independent ectopic activity of the Shh pathway is augmented by ectopic Shh ligand expression in a minority of cases (indicated by the dotted arrow). In the posterior mesenchyme, which is populated by Twist1 heterozygous cells, gene expression patterns do not change. Lighter shading in CKO shows the Twist1-deficient region. References: 1, (O'Rourke et al., 2002); 2, (Loebel et al., 2002); 3, (Te Welscher et al., 2002); 4, (Vokes et al., 2008); 5, (Kuijper et al., 2005); 6, (Zakany et al., 2007); 7 (Buscher et al., 1997); 8, (Khokha et al., 2003); 9, (Zakany et al., 2004); 10, (Bastida et al., 2009); 11, (Goodrich et al., 1996).

D.A.F. Loebel et al. / Developmental Biology 362 (2012) 132–140

139

Table 1 Impact of conditional loss and reduced Twist1 gene activity on the morphogenesis of skeletal elements and molecular phenotype of the forelimb bud of mouse embryos: this study compared with Zhang et al., 2010 (1) and Krawchuk et al., 2010(2). Phenotype Limb skeleton Supernumerary digits Mirror-like digit duplication Bifid digits Atypical “digit 1” Extra carpals Ulnarization of radius Fewer digits Loss of radius Humerus Fused shoulder joint scapula Molecular markers Tissue patterning Alx4 Hand2 Hoxd11 Hoxd13 Alx1 Emx2 Pbx1 Gsc Shh signaling Gli3 Ptch1 Shh FGF signaling Fgf8 Fgf4 Fgf10 Fgfr1 Mkp3 Spry1 Spry4 BMP signaling Grem1 Msx1 a

Mesp1-Cre;Twist13loxPneo/del

Prx1-Cre;Twist13loxPneo/−(1) Prx1-Cre;Twist13loxPneo/3loxPneo(2)

√ √ √ √ √ √

√ √ √ √ √ √

Short and deformed √ Reduced size, small or malformed spine

Twist1Ska10/Ska10(1,2)a

Twist1Ska10/−(2)a

√ √

√ √ No tuberosity

√

√ Short Rudimentary

Reduced size

Rudimentary

anterior Unchanged

Rudimentary

anterior Unchanged anterior

anterior anterior Unchanged Unchanged

anterior anterior (low freq.) anterior (low freq.) anterior anterior

anterior anterior Unchanged anterior anterior

anterior

anterior

anterior

anterior

Unchanged posterior anterior anterior

anterior

Upregulated, down-regulated and expanded expression. Ska10 and CC (Charlie Chaplin) are synonymous.

preaxial structures were posteriorized (Table 1). Proximal limb elements lost some anterior characteristics: the anterior tuberosity of the humerus and spine of the scapula were absent. Some Prx1-Cre CKO embryos lacked a radius altogether (Krawchuk et al., 2010), which was not seen in the Mesp1-Cre CKO embryos. Embryos in which the Ska10 missense mutant allele was the only functional Twist1 allele lacked a radius and had fewer digits, while showing signs of mirror-duplication (Krawchuk et al., 2010; Zhang et al., 2010). Although similarities were observed between Mesp1-Cre and Prx1-Cre CKO embryos in distal limb development, there are differences in the degree of abnormality in the proximal limb (Table 1). Mesp1-Cre CKO embryos displayed reductions and abnormalities in the humerus and scapula. In Prx1-Cre CKO embryos, more dramatic reductions in scapula development were reported (Krawchuk et al., 2010; Zhang et al., 2010) (Table 1). The loss of the deltoid tuberosity and scapula spine in CKO embryos is consistent with the loss of anterior characteristics in the limb. The more pronounced abnormalities and reductions in the humerus and scapula in Prx1-Cre CKO embryos suggest that there are additional roles for Twist1 in other parts of the limb buds mesenchyme in growth and patterning of these structures. We observed ectopic expression of Shh in a minority of CKO limb buds (Goodrich et al., 1996). This is perhaps surprising since Hoxd13 has previously been shown to activate Shh expression (Zakany et al., 2004) but it is possible that the level of anterior Hoxd13 is insufficient to activate detectable Shh transcription in most cases. Hindlimb polydactyly was significantly reduced in Twist1 +/Ska10; Shh +/− embryos (Krawchuk et al., 2010), confirming

that the Shh signaling pathway is involved in hindlimb polydactyly in Twist1 mutants, even though ectopic Shh expression was observed in only a small proportion of Twist1 +/Ska10 or Twist1 +/− hindlimb buds (Krawchuk et al., 2010; O'Rourke and Tam, 2002). Although it is possible that expression of Shh (and Ptch1) is below the level of detection in these assays, Shh ligand-independent polydactyly has been observed in Gli3 −/−; Shh −/− embryos (Litingtung et al., 2002; Te Welscher et al., 2002) and ectopic Hoxd13 and Fgf4 expression occurs in Alx4 −/−; Shh −/− embryos indicating an anterior–posterior patterning defect (Kuijper et al., 2005). Since both Alx4 and Gli3 are down-regulated in Twist1 CKO forelimb buds (Fig. 6) it is possible that ectopic Shh ligand expression is not a pre-requisite for limb anterior–posterior patterning defects. Rather, loss of Twist1 in the anterior-proximal forelimb bud mesenchyme might impact on Shh signaling at other points in the pathway, perhaps by reducing anterior Gli3 expression (Fig. 6). In a minority of cases, ectopic Shh expression, activated by anterior Hoxd13 expression and suppression of anterior Bmp signaling by Grem1 (Fig. 6) could reinforce the effects of Shh ligand-independent patterning defects, resulting in more pronounced mirror image-like polydactyly. Supplementary materials related to this article can be found online at doi:10.1016/j.ydbio.2011.11.020. Acknowledgments We thank Atsushi Yoshiki and Richard Harvey for providing the Mesp1-Cre mice, Phillipe Soriano for Rosa26R mice and the following

140

D.A.F. Loebel et al. / Developmental Biology 362 (2012) 132–140

scientists for providing cDNA clones for riboprobes: F. Meijlink (Alx1, Alx4), L. Selleri (Emx2, Pbx1), L. Niswander (Fgf4), G. Martin (Fgf8), D. Ornitz (Fgf10), C.C. Hui (Gli3), X. Sun (Grem1), W. Shawlot (Gsc, Twist1), E. Olson (Hand2), P. Chambon (Hoxd13), R. Hill (Msx1), M. Scott (Ptch1) and A. McMahon (Shh). We thank the staff of the CMRI Bioservices Unit for animal husbandry. This project is supported by a grant from the Australian Research Council (DP 1094008) and by Mr James Fairfax. DAFL is a CMRI Kimberly-Clark Research Fellow, HB is a National Health and Medical Research Council (NHMRC) postgraduate Scholar and PPLT is a NHMRC Senior Principal Research Fellow (Grant 1003100). References Bastida, M.F., Sheth, R., Ros, M.A., 2009. A BMP-Shh negative-feedback loop restricts Shh expression during limb development. Development 136, 3779–3789. Belo, J.A., Leyns, L., Yamada, G., De Robertis, E.M., 1998. The prechordal midline of the chondrocranium is defective in Goosecoid-1 mouse mutants. Mech. Dev. 72, 15–25. Bénazet, J.-D., Zeller, R., 2009. Vertebrate limb development: moving from classical morphogen gradients to an integrated 4-dimensional patterning system. Cold Spring Harbor Perspectives in Biology 1. Bensoussan-Trigano, V., Lallemand, Y., Saint Cloment, C., Robert, B., 2011. Msx1 and Msx2 in limb mesenchyme modulate digit number and identity. Dev. Dyn. 240, 1190–1202. Bildsoe, H., Loebel, D.A., Jones, V.J., Chen, Y.T., Behringer, R.R., Tam, P.P., 2009. Requirement for Twist1 in frontonasal and skull vault development in the mouse embryo. Dev. Biol. 331, 176–188. Buckingham, M., Bajard, L., Chang, T., Daubas, P., Hadchouel, J., Meilhac, S.N., Montarras, D., Rocancourt, D., Relaix, F.D.R., 2003. The formation of skeletal muscle: from somite to limb. Journal of Anatomy 202, 59–68. Buscher, D., Bosse, B., Heymer, J., Ruther, U., 1997. Evidence for genetic control of Sonic hedgehog by Gli3 in mouse limb development. Mech. Dev. 62, 175–182. Capellini, T.D., Vaccari, G., Ferretti, E., Fantini, S., He, M., Pellegrini, M., Quintana, L., Di, G.G., Sharpe, J., Selleri, L., Zappavigna, V., 2010. Scapula development is governed by genetic interactions of Pbx1 with its family members and with Emx2 via their cooperative control of Alx1. Development 137, 2559–2569. Chen, Y.T., Akinwunmi, P.O., Deng, J.M., Tam, O.H., Behringer, R.R., 2007. Generation of a Twist1 conditional null allele in the mouse. Genesis 45, 588–592. Chen, Z.F., Behringer, R.R., 1995. Twist is required in head mesenchyme for cranial neural tube morphogenesis. Genes Dev. 9, 686–699. Delgado, I., Dominguez-Frutos, E., Schimmang, T., Ros, M.A., 2008. The incomplete inactivation of Fgf8 in the limb ectoderm affects the morphogenesis of the anterior autopod through BMP-mediated cell death. Dev. Dyn. 237, 649–658. Firulli, B.A., Krawchuk, D., Centonze, V.E., Vargesson, N., Virshup, D.M., Conway, S.J., Cserjesi, P., Laufer, E., Firulli, A.B., 2005. Altered Twist1 and Hand2 dimerization is associated with Saethre–Chotzen syndrome and limb abnormalities. Nat. Genet. 37, 373–381. Goodrich, L.V., Johnson, R.L., Milenkovic, L., McMahon, J.A., Scott, M.P., 1996. Conservation of the hedgehog/patched signaling pathway from flies to mice: induction of a mouse patched gene by Hedgehog. Genes Dev. 10, 301–312. Hogan, B., Beddington, R., Costantini, F., lacy, E., 1994. Manipulating the Mouse Embryo: A Laboratory Manual, Second edition. Cold Spring Harbor Laboratory Press, Plainview, NY. Khokha, M.K., Hsu, D., Brunet, L.J., Dionne, M.S., Harland, R.M., 2003. Gremlin is the BMP antagonist required for maintenance of Shh and Fgf signals during limb patterning. Nat. Genet. 34, 303–307. Kinder, S.J., Tan, S.S., Tam, P.P., 2000. Cell grafting and fate mapping of the earlysomite-stage mouse embryo. Methods Mol. Biol. 135, 425–437. Koussoulakos, S., 2004. Vertebrate limb development: from Harrison's limb disk transplantations to targeted disruption of Hox genes. Anat. Embryol. 209, 93–105. Krawchuk, D., Weiner, S.J., Chen, Y.T., Lu, B.C., Costantini, F., Behringer, R.R., Laufer, E., 2010. Twist1 activity thresholds define multiple functions in limb development. Dev. Biol. 347, 133–146.

Kuijper, S., Feitsma, H., Sheth, R., Korving, J., Reijnen, M., Meijlink, F., 2005. Function and regulation of Alx4 in limb development: complex genetic interactions with Gli3 and Shh. Dev. Biol. 285, 533–544. Lewandoski, M., Sun, X., Martin, G.R., 2000. Fgf8 signalling from the AER is essential for normal limb development. Nat. Genet. 26, 460–463. Litingtung, Y., Dahn, R.D., Li, Y., Fallon, J.F., Chiang, C., 2002. Shh and Gli3 are dispensable for limb skeleton formation but regulate digit number and identity. Nature 418, 979–983. Loebel, D.A., O'Rourke, M.P., Steiner, K.A., Banyer, J., Tam, P.P., 2002. Isolation of differentially expressed genes from wild-type and Twist mutant mouse limb buds. Genesis 33, 103–113. Loebel, D.A., Tsoi, B., Wong, N., O'Rourke, M.P., Tam, P.P., 2004. Restricted expression of ETn-related sequences during post-implantation mouse development. Gene Expr. Patterns 4, 467–471. Martinez-Barbera, J.P., Rodriguez, T.A., Greene, N.D.E., Weninger, W.J., Simeone, A., Copp, A.J., Beddington, R.S.P., Dunwoodie, S., 2002. Folic acid prevents exencephaly in Cited2 deficient mice. Hum. Mol. Genet. 11, 283–293. Min, H., Danilenko, D.M., Scully, S.A., Bolon, B., Ring, B.D., Tarpley, J.E., DeRose, M., Simonet, W.S., 1998. Fgf-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless. Genes Dev. 12, 3156–3161. O'Rourke, M.P., Soo, K., Behringer, R.R., Hui, C.C., Tam, P.P., 2002. Twist plays an essential role in FGF and SHH signal transduction during mouse limb development. Dev. Biol. 248, 143–156. O'Rourke, M.P., Tam, P.P., 2002. Twist functions in mouse development. Int. J. Dev. Biol. 46, 401–413. Parameswaran, M., Tam, P.P.L., 1995. Regionalisation of cell fate and morphogenetic movement of the mesoderm during mouse gastrulation. Dev. Genet. 17, 16–28. Qu, S., Tucker, S.C., Ehrlich, J.S., Levorse, J.M., Flaherty, L.A., Wisdom, R., Vogt, T.F., 1998. Mutations in mouse Aristaless-like4 cause Strong's luxoid polydactyly. Development 125, 2711–2721. Robert, B., 2007. Bone morphogenetic protein signaling in limb outgrowth and patterning. Dev. Growth Differ. 49, 455–468. Saga, Y., Hata, N., Kobayashi, S., Magnuson, T., Seldin, M.F., Taketo, M.M., 1996. MesP1: a novel basic helix–loop–helix protein expressed in the nascent mesodermal cells during mouse gastrulation. Development 122, 2769–2778. Saga, Y., Miyagawa-Tomita, S., Takagi, A., Kitajima, S., Miyazaki, J., Inoue, T., 1999. MesP1 is expressed in the heart precursor cells and required for the formation of a single heart tube. Development 126, 3437–3447. Soriano, P., 1999. Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat. Genet. 21, 70–71. Sun, X., Lewandoski, M., Meyers, E.N., Liu, Y.H., Maxson Jr., R.E., Martin, G.R., 2000. Conditional inactivation of Fgf4 reveals complexity of signalling during limb bud development. Nat. Genet. 25, 83–86. Te Welscher, P., Zuniga, A., Kuijper, S., Drenth, T., Goedemans, H.J., Meijlink, F., Zeller, R., 2002. Progression of vertebrate limb development through SHH-mediated counteraction of GLI3. Science 298, 827–830. Vokes, S.A., Ji, H., Wong, W.H., McMahon, A.P., 2008. A genome-scale analysis of the cisregulatory circuitry underlying sonic hedgehog-mediated patterning of the mammalian limb. Genes Dev. 22, 2651–2663. Watson, C.M., Trainor, P.A., Radziewic, T., Pelka, G.J., Zhou, S.X., Parameswaran, M., Quinlan, G.A., Gordon, M., Sturm, K., Tam, P.P., 2008. Application of lacZ transgenic mice to cell lineage studies. Methods Mol. Biol. 461, 149–164. Wyngaarden, L.A., Vogeli, K.M., Ciruna, B.G., Wells, M., Hadjantonakis, A.K., Hopyan, S., 2010. Oriented cell motility and division underlie early limb bud morphogenesis. Development 137, 2551–2558. Zakany, J., Kmita, M., Duboule, D., 2004. A dual role for Hox genes in limb anterior– posterior asymmetry. Science 304, 1669–1672. Zakany, J., Zacchetti, G., Duboule, D., 2007. Interactions between HOXD and Gli3 genes control the limb apical ectodermal ridge via Fgf10. Dev. Biol. 306, 883–893. Zhang, Z., Sui, P., Dong, A., Hassell, J., Cserjesi, P., Chen, Y.T., Behringer, R.R., Sun, X., 2010. Preaxial polydactyly: interactions among ETV, TWIST1 and HAND2 control anterior–posterior patterning of the limb. Development 137, 3417–3426. Zuniga, A., Quillet, R., Perrin-Schmitt, F., Zeller, R., 2002. Mouse Twist is required for fibroblast growth factor-mediated epithelial–mesenchymal signalling and cell survival during limb morphogenesis. Mech. Dev. 114, 51–59.