Progression of Vertebrate Limb Development Through SHH-Mediated Counteraction of GLI3

See allHide authors and affiliations

Science  25 Oct 2002:
Vol. 298, Issue 5594, pp. 827-830
DOI: 10.1126/science.1075620


Distal limb development and specification of digit identities in tetrapods are under the control of a mesenchymal organizer called the polarizing region. Sonic Hedgehog (SHH) is the morphogenetic signal produced by the polarizing region in the posterior limb bud. Ectopic anterior SHH signaling induces digit duplications and has been suspected as a major cause underlying congenital malformations that result in digit polydactyly. Here, we report that the polydactyly of Gli3-deficient mice arises independently of SHH signaling. Disruption of one or bothGli3 alleles in mouse embryos lacking Shhprogressively restores limb distal development and digit formation. Our genetic analysis indicates that SHH signaling counteracts GLI3-mediated repression of key regulator genes, cell survival, and distal progression of limb bud development.

The Hedgehog(Hh) signaling pathway controls many key developmental processes during animal embryogenesis (1). InDrosophila embryos, all known functions of Hh signaling are mediated by the transcriptional effector Cubitus interruptus (Ci) (2). Several homologs of Hh and Cihave been identified in higher vertebrates. In particular, Sonic Hedgehog (SHH) and the Ci homolog GLI3 are required for vertebrate limb development (3–6). GLI3 acts first during the initiation of limb bud development and before the activation of SHH signaling in posterior restriction of the basic helix-loop-helix transcription factor dHAND. dHAND in turn prevents Gli3expression from spreading posteriorly (Fig. 1A, panel 1) (7). In addition, GLI3 restricts the SHH-independent early expression of5′HoxD genes and Gremlin to the posterior mesenchyme (8). Subsequently, dHAND functions in the activation of Shh expression (9). Limb bud morphogenesis is then controlled by reciprocal interactions of two signaling centers (Fig. 1A, panel 2): the polarizing region, an instructive organizer located in the posterior limb bud mesenchyme, and the apical ectodermal ridge (AER). SHH signaling by the polarizing region in combination with bone morphogenetic proteins (BMPs) and their antagonists instruct limb skeletal patterning, most likely by regulating the expression of 5′HoxD transcription factors (Fig. 1A, panel 2) (10, 11). The BMP antagonistGremlin mediates a signal relay from the polarizing region to the AER (12), and fibroblast growth factor (FGF) signaling by the AER maintains the polarizing region to enable progression of limb bud morphogenesis (SHH-FGF feedback loop) (10, 11).

Figure 1

(A) Interactions of key regulators of vertebrate limb development: ➁ Reciprocal antagonism of GLI3 and dHAND prepatterns the limb bud mesenchyme before activation of SHH signaling. ➀ dHAND is required to activateShh expression by polarizing region cells. SHH signaling inhibits the processing of GLI3 to GLI3-83, which acts as transcriptional repressor (GLI3R). SHH positively regulates5′HoxD (5′HOX) gene and Gremlin (GRE) expression in distal mesenchyme. The SHH-FGF feedback loop between the polarizing region and the AER is established through Gremlin-mediated BMP antagonism. Only the interactions relevant to the present study are shown. (B) Skeletal stains of forelimbs [embryonic day 16.5 (E16.5)] of Shh andGli3 single- and double-mutant mouse embryos (28). Arrowheads point to well-formed elbow joints. Genotype labels: Wt, wild-type; Gli3 +/–, one allele ofGli3 inactivated, arrow points to small ectopic cartilage condensation; Gli3 –/–,Gli3-deficient; Shh –/–,Shh-deficient; Shh –/–,Gli3 +/–, Shh –/–lacking in addition one functional Gli3 allele;Shh –/–, Gli3 –/–, forelimb of a double-homozygous embryo. (C) Skeletal stains of hindlimbs (E14.5) of Shh and Alx4 single- and double-mutant mouse embryos. Genotype labels:Alx4 –/–, Alx4 deficient, arrowhead points to a duplicated preaxial digit 2 (phenotype is 100% penetrant in hindlimbs);Shh –/–, arrowhead points to the one digit formed; Shh –/–,Alx4 –/–, hindlimb of a double homozygous embryo. (D) Detection of apoptotic cells (29) in forelimb buds of Shh–/– embryos (E10.75, 38 to 39 somites) lacking one or both Gli3alleles. Shown are representative sections at similar levels (anterior-dorsal is to the left and posterior-ventral is to the right). Anterior is to the top and posterior to the bottom in all panels. Arrowheads point to the AER. TUNEL, terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling.

Ectopic anterior expression of SHH in limb buds induces digit duplications, and this ectopic expression is viewed as a major cause of limb polydactylous (extra digits) phenotypes (10,11). Spontaneous mutations affecting GLI3 cause a severe congenital malformation in mice [Extra-toes (Xt)] (3,13) and several related syndromes in humans (14). The associated polydactylous phenotypes have so far been attributed to ectopic Shh expression (13) and anteriorly expanded 5′Hoxd expression (8). In the absence of SHH signaling, full-length GLI3 is proteolytically cleaved to a smaller protein (GLI3-83) (15). Such truncated forms of GLI3 protein have repressor activity (16, 17). However, SHH signaling inhibits GLI3 processing, which may allow the full-length protein to function as a transcriptional activator (18, 19). Biochemical analysis showed that long-range SHH signaling and inhibition of GLI3 processing in limb buds (Fig. 1A, panel 2) result in formation of an anterior-to-posterior graded distribution (high to low) of the truncated GLI3-83 protein with repressor activity (15).

To identify the essential function(s) and to study the phenotypic consequences of these GLI3-SHH interactions, we analyzed the limb development of Shh and Gli3 double-mutant embryos (20) (Fig. 1B). The limbs of Gli3-deficient (Gli3 –/–) embryos are polydactylous (Fig. 1B) (13), whereas one fused forearm (zeugopod) bone and no digit arch (autopod) form in limbs of Shh-deficient (Shh –/–) embryos (Fig. 1B) (5,6). Disruption of one Gli3 allele inShh-deficient (Shh –/–,Gli3+/–) embryos improves distal limb development, as two zeugopodal bones and rudimentary digits form (Fig. 1B). The limbs of double-homozygous (Shh–/– ,Gli3 –/–) mouse embryos are gross-morphologically indistinguishable from the limbs ofGli3–/– embryos (Fig. 1B and fig. S1). This genetic analysis establishes that the polydactylous limb phenotype ofGli3-deficient embryos is not caused by ectopic SHH signaling as previously assumed (13). Another molecularly well-studied mouse mutation with a polydactylous phenotype is Strong's Luxoid (Lst), which is caused by disruption of the transcription factor Alx4 (Alx4 –/–, Fig. 1C) (21). In wild-type limb buds, Alx4 is expressed in the anterior mesenchyme, and a functional anterior SHH signaling center is established in the limb buds of Alx4 –/–embryos that is similar to the one in Gli3 –/–embryos. In contrast to Gli3, no rescue of distal limb development is observed in limbs lacking Shh andAlx4 (Fig. 1C) (20). These results establish that two different mechanisms can cause polydactyly. The polydactyly ofAlx4 –/– mice depends on SHH signaling, whereas the polydactyly of Gli3 –/– mice is SHH independent. The limb phenotype of Gli3–/– mouse embryos is similar to that of talpid chicken embryos (22). Biochemical evidence shows that GLI3-83 repressor levels are substantially reduced in talpid mutant embryos (15), which indicates that this polydactyly may also occur independent of SHH signaling (22).

The results shown in Fig. 1B reveal that SHH-independent restoration of distal limb development inShh–/– embryos is Gli3-dose dependent. Because massive cell death occurs in limb buds ofShh–/– embryos (Fig. 1D) (5), the progressive restoration of distal limb development (Fig. 1A) may be due to improved cell survival. Indeed, apoptosis is reduced in limb buds ofShh–/– embryos lacking one functionalGli3 allele but is still higher than apotosis in wild types (Fig. 1D). The low levels of apoptotic cells in limb buds of double homozygous embryos are, however, comparable to the levels in wild-type (Fig. 1D) and Gli3-deficient embryos (23).

To exclude the possibility that ligand-independent activation of intracellular SHH signal transduction restores limb development inShh and Gli3 double mutant embryos (Fig. 1B), we analyzed the expression of the SHH transcriptional targetsPatched (Ptc) and Gli1 (11,24). Neither Ptc (23) norGli1 were expressed in the limb buds ofShh –/– embryos lacking either one or bothGli3 alleles (Fig. 2A). TheGli3 dose-dependent improvement of limb development is also not linked to alterations of dHAND, because its expression remains posteriorly restricted in limb buds ofShh–/– , Gli3+/– embryos (Fig. 2B). The disruption of both Gli3 alleles results in SHH-independent anterior dHAND expression during early limb development (Fig. 2B) (7).

Figure 2

Expression of Gli1 (A) and dHAND (B) in limb buds of Shh andGli3 single- and double-mutant embryos at E10.25 (32 to 33 somites). All limb buds are oriented with anterior to the top and posterior to the bottom. Genotypes are as described in Fig. 1B.

The SHH-FGF feedback loop is established throughGremlin-mediated BMP antagonism (Fig. 1A, panel 2).Gremlin and Fgf4 expression are activated but not maintained in limb buds of Shh –/– embryos (Fig. 3) (12). Low levels ofGremlin but not Fgf4 transcripts are detected in limb buds of Shh–/– embryos lacking oneGli3 allele (Fig. 3). In double-homozygous embryos,Gremlin and Fgf4 expression in limb buds is comparable to that in Gli3–/– embryos. The differential restoration of Fgf4 indicates that its induction may depend on a threshold of BMP antagonism.

Figure 3

(Left) Differential rescue ofGremlin expression in the limb bud mesenchyme andFgf4 in the AER. Gremlin (mesenchyme) andFgf4 (AER) in limb buds (E 10.75, 37 to 38 somites) are simultaneously detected. Asterisks indicate anterior margin of the limb buds. Arrowheads point to Fgf4 expression in the AER. Genotypes are as described in Fig. 1B.

Hoxa11 and Hoxd11 are essential to pattern the zeugopod, whereas Hoxa13 and Hoxd13 are essential for autopod patterning (25). In Shh –/– limb buds, Hoxa11 remains expressed (23), whereas Hoxd11 is rapidly down-regulated (Fig. 4A). The disruption of one Gli3 allele in Shh–/– embryos partially restoresHoxd11 expression such that its anterior boundary in limb buds is located at a position similar to that in wild type (arrowheads,Fig. 4A). This restoration provides a likely molecular explanation for improved zeugopod development in Shh–/– ,Gli3+/– embryos (Fig. 1B). The expressions of both Hoxa13 and Hoxd13 are low in limb buds ofShh –/– embryos (Fig. 4, B and C) (4–6). The additional inactivation of oneGli3 allele restores Hoxa13 expression to intermediate levels (Fig. 4B), whereas Hoxd13 transcripts remain low (arrowhead, Fig. 4C). In limb buds of double homozygous embryos, all three 5′Hox genes are expressed at levels similar to those expressed in Gli3–/– embryos (Fig. 4, A to C). These results indicate that the progressive restoration of distal limb development (Fig. 1B) is likely due toGli3 dose-dependent restoration of the distal5′Hoxa and distal 5′Hoxd expression domains in limb buds of Shh –/– embryos.

Figure 4

(Right) Gli3 dose-dependent up-regulation of 5′HoxD and 5′HoxA expression in limb buds ofShh –/– embryos. All limb buds (E 10.75, 37 to 38 somites) are oriented with anterior to the top and posterior to the bottom. (A) Expression of Hoxd11. Arrowheads indicate the anterior expression boundary. (B) Expression ofHoxa13. (C) Expression of Hoxd13. Arrowheads in (B) and (C) indicate low expression of respective gene. Genotypes are as described in Fig. 1B.

Our study provides genetic evidence for an important sequential interaction of GLI3 with dHAND and SHH. The SHH-independent nature of the digit polydactyly in Gli3 –/– limb buds is most likely a direct consequence of the early anterior expansion of the expression of “posterior” genes such as 5′Hoxd genes (7, 8). The scapula and stylopod (not affected inShh–/– embryos) (5,6) may be patterned by the genetic interaction ofGLI3 with dHAND before the activation of SHH signaling (Fig. 1A, panel 1) (7). In limb buds ofShh –/– embryos, Gli3 expression expands posteriorly, concurrent with the down-regulation of “posterior” genes and the onset of apoptosis (5,6, 12), which cause the “shut-down” of distal limb development and antero-posterior patterning. TheGli3-dose-dependent restoration of cell survival and limb bud development in Shh –/– embryos indicates that, in wild-type limb buds, SHH counteracts GLI3 to enable progression of outgrowth and patterning. The polarizing region is propagated distally, and SHH signaling is up-regulated via the SHH-FGF feedback loop (Fig. 1A, panel 2) (11, 12). Such up-regulation of SHH signaling should increasingly inhibit GLI3 processing and should result in a graded GLI3-83 repressor distribution as shown biochemically (15). The resulting full-length GLI3 protein may also function in positive transcriptional regulation of SHH targets (18, 19). Although the SHH-GLI3 interactions enable distal limb development and formation of a digit arch, additional signals such as BMPs participate in specification of digit identities (26). Lastly, aspects of the dorso-ventral neural tube patterning (disrupted inShh –/– embryos) are also restored in mouse embryos mutant for both Shh and Gli3(27). This shows that regulation of GLI3 by SHH signaling (and vice versa) is of general functional importance during embryonic development.

Supporting Online Material

Fig. S1

  • * To whom correspondence should be addressed. E-mail: R.Zeller{at}


View Abstract

Navigate This Article