Hox10 and Hox11 Genes Are Required to Globally Pattern the Mammalian Skeleton

See allHide authors and affiliations

Science  18 Jul 2003:
Vol. 301, Issue 5631, pp. 363-367
DOI: 10.1126/science.1085672


Mice in which all members of the Hox10 or Hox11 paralogous group are disrupted provide evidence that these Hox genes are involved in global patterning of the axial and appendicular skeleton. In the absence of Hox10 function, no lumbar vertebrae are formed. Instead, ribs project from all posterior vertebrae, extending caudally from the last thoracic vertebrae to beyond the sacral region. In the absence of Hox11 function, sacral vertebrae are not formed and instead these vertebrae assume a lumbar identity. The redundancy among these paralogous family members is so great that this global aspect of Hox patterning is not apparent in mice that are mutant for five of the six paralogous alleles.

Hox genes have long been recognized as important transcriptional regulators of embryonic development. In mammals, this complex of 39 genes resides on four separate chromosomal linkage groups designated A, B, C, and D, which arose early in the evolution of vertebrates from successive duplications of a single ancestral complex. Homologous members within the separate linkage groups are divided into 13 sets of paralogous genes, each having two to four members. During development, paralogous sets of genes are activated sequentially, with Hox1 and Hox2 paralogous genes being expressed earlier and more anteriorly in the embryo and successive genes through paralogous group Hox13 appearing later and more posteriorly.

The spectrum of perturbations of the mammalian skeleton resulting from either gain- or loss-of-function mutations in individual Hox genes has been difficult to interpret in terms of a coherent model of how these genes participate in the patterning of the axial skeleton. Loss-of-function Hox mutations have yielded changes in vertebral morphology along the anteroposterior (AP) axis that have been interpreted as anterior homeotic transformations as well as posterior homeotic transformations. Typically, these morphological changes involve perturbations in one or a small number of vertebrae.

Among different vertebrate species, axial skeletal patterns have diverged considerably. A comparative survey of Hox gene expression in mice and chicks showed that Hox gene expression boundaries along the rostrocaudal axis shift in accordance with changes in the class of vertebrae produced at a given axial level (1). This observation suggests that Hox genes play a critical role in the global patterning of the vertebrate axial skeleton (2). Yet, over the past decade, loss-of-function studies of mice with single, and even subsets of, paralogous Hox gene mutations have shown, with variable expressivities and penetrance, only relatively minor changes in skeletal phenotypes, which is inconsistent with their proposed role as global regulators of axial skeletal patterning. What has complicated the analysis of Hox gene mutations is that these genes have retained considerable functional redundancy between paralogous groups (39). Therefore, we examined the effects of the loss of function of the entire group of Hox10 and Hox11 paralogous genes on skeletal patterning [see supporting online material for details regarding the generation of mice (9, 10)].

The axial formula in mice is 7 cervical, 13 thoracic, 6 lumbar, 4 sacral, and numerous (and slightly variable numbers of) caudal vertebrae. Mice with either Hox10 or Hox11 paralogous mutations show drastic alterations of the axial formula. Hox10 triple mutant skeletons completely lack lumbar vertebrae and exhibit rib processes that protrude from each vertebral segment beyond the 13th thoracic vertebra through the normal lumbar and sacral regions (compare Fig. 1, A and F). In addition to ectopic rib formation, these vertebral elements also display morphological characteristics that are normally associated with thoracic vertebrae (compare Fig. 1, B to E, with Fig. 1, G to J, and Fig. 2A with Fig. 2B). In the Hox10 triple mutant (Fig. 1A), the severely altered sacral vertebrae still form fusions at their lateral margins to produce a pseudosacrum. This fusion occurs at the appropriate position despite the severe perturbations in morphology of these vertebral elements. Despite the changes in axial morphology, the pelvis (which also displays patterning perturbations; compare Fig. 2, D and E) and the hindlimbs associate with the pseudosacral lateral fusion at the normal position along the vertebral axis (compare Fig. 2, A and B).

Fig. 1.

Axial skeletons of Hox10 and Hox11 triple mutants at embryonic day 18.5 (E18.5). Ventral views of the axial skeleton from the lower thoracic region through the early caudal region of a Hox10 triple mutant (A), a control (F), and a Hox11 triple mutant (K) are shown. Yellow asterisks indicate lumbar vertebrae; red asterisks indicate sacral vertebrae. A five-allele mutant from the Hox10 and Hox11 paralogous mutant group is shown in (P) and (Q), respectively (red arrows indicate sacral wing formation). Analogous vertebrae were dissected from the control and from each triple mutant to compare single vertebral identities. The 19th vertebral element, normally T12, is shown in (B), (G), and (L). The 23rd element, normally L3, is shown in (C), (H), and (M). The 28th element, normally S2, is shown in (D), (I), and (N). The 35th element, normally caudal vertebra 5 (C5), is shown in (E), (J), and (O). (Between two and seven E18.5 skeletons were collected for each of the triple mutant, five-allele, and control skeletons for each paralogous group.)

Fig. 2.

Pelvic position and morphology in E18.5 Hox10 and Hox11 triple mutants. In (A to C), the axial skeleton associated with the pelvis is shown in lateral view from Hox10 triple mutants, control, and Hox11 triple mutants, respectively. Dissociated pelvic bones are shown in (D to F) for the same genotypes.

Mice with only five mutant alleles display a 14th rib and altered sacral processes, but the axial morphology is much less severely affected than in the Hox10 triple mutant (Fig. 1P). Combinations of any five of the six mutant alleles in the Hox10 paralogous group demonstrate similar mutant phenotypes to one another, indicating the approximately equal contribution of these alleles to axial patterning (11). Comparison of Hox10 five-allele mutants to Hox10 triple mutant animals demonstrates the extent of redundancy within this paralogous group.

Hox11 triple mutant skeletons show equally severe, but distinct, axial phenotypes. Rib formation terminates normally and the lumbar vertebrae appear normal; however, no sacral vertebrae are formed. Instead, these vertebrae assume a lumbar morphology (Fig. 1K). The lumbarlike vertebral elements continue far past the normal sacral region, and caudal vertebrae are not apparent until several elements more posterior than in controls (compare Fig. 1, L to O, with Fig. 1, G to J). Mice mutant for Hox11 paralogous genes also display severe perturbations of pelvic morphology (compare Fig. 2, F and E). However, despite the absence of sacral vertebrae in these mutants, the pelvis and hindlimbs again associate with the appropriate vertebral segments (Fig. 2C). The combined results demonstrate that the positioning of the pelvis and hindlimbs is not under the control of either Hox10 or Hox11 paralogous genes. Further, the AP positioning of these elements is not dependent on normal sacral development or on appropriate lumbosacral transitions.

Five-allele Hox11 mutants again demonstrate the redundancy within the paralogous group with respect to axial phenotype. No sacral wing fusion occurs in any of the five-allele skeletons, but sacral wings do appear on more posterior elements (Fig. 1Q), and there are fewer elements that are lumbarlike posterior to the normal sacral region. It is important to note that although Hox10 and Hox11 triple mutants both severely affect sacral formation, these paralogs clearly perform distinct functions on the same elements. Also, even though these sets of paralogous mutations result in the complete loss of lumbar or sacral vertebrae, the total number of vertebral elements is not altered.

The limbs of all vertebrates are composed of three basic elements: the stylopod (humerus/femur of the forelimb and hindlimb, respectively), the zeugopod (radius and ulna/tibia and fibula), and the autopod (numerous carpal, metacarpal, tarsal, metatarsal, and phalangeal elements). HoxA and HoxD complex paralogous group genes 9 to 13 are expressed and function in the developing forelimb; whereas HoxA and HoxD paralogous groups 10 to 13, as well as Hoxc10 and Hoxc11, are expressed and function in the developing hindlimb (1, 3, 7, 1217). It has previously been shown that Hoxa11/Hoxd11 and Hoxa13/Hoxd13 play major roles in the patterning of the forelimb zeugopod and of both the forelimb and hindlimb autopod, respectively (3, 7). We demonstrated that Hox10 and Hox11 paralogous genes are required for patterning of the hindlimb stylopod and zeugopod, respectively.

In Hox10 triple mutants, the humerus is only moderately decreased in length relative to controls, and the deltoid process is not formed (compare Fig. 3, A and B). In contrast, the formation of the femur is grossly affected in these mutants. The femur is greatly reduced in length and no patella is formed (compare Fig. 3, D and E). The Hox10 five-allele mutants show an intermediate stylopod phenotype between wild-type and triple mutants (11).

Fig. 3.

Limb skeletons of E18.5 Hox10 and Hox11 triple mutant mice. (A and D) show a Hox10 triple mutant forelimb and hindlimb, respectively. A control forelimb and hindlimb are shown in (B and E). (C and F) show a Hox11 triple mutant forelimb and hindlimb.

Hox11 triple mutants demonstrate dramatic mispatterning of the fore- and hindlimb zeugopods (Fig. 3, C and F). The forelimb mutant phenotype is similar to that reported for Hoxa11/Hoxd11 double mutants (7). However, in Hoxa11/Hoxd11 double mutants, the formation of the tibia and fibula is only mildly affected, whereas in the Hox11 triple mutants, the hindlimb zeugopod is grossly affected. These results are consistent with Hoxc11 being expressed only in the hindlimbs (12).

The results from the genetic analysis of Hox10 and Hox11 paralogous genes suggest that Hox genes are indeed involved in global patterning of the mammalian axial skeleton. Further, one can begin to postulate mechanisms of how changes in Hox gene expression could account for variation of the axial formula in different vertebrate taxa. For instance, one would predict that shifts of the boundaries of Hox10 paralogous gene expression, rostrally or caudally, would alter the number of thoracic vertebrae present in an animal. Similarly, shifts in the expression of the Hox11 paralogous genes would predict an alteration in the position and number of sacral vertebrae. Many primitive tetrapods have ribs projecting from all vertebrae, extending from the head to the tail. This has led to the suggestion that the ground state for vertebrae includes rib projections (18). Our data from the mouse supports this hypothesis and provides a mechanism whereby Hox genes have been used during evolution to suppress and modify rib formation in the lumbosacral region. It is curious and perhaps not insignificant that the normal patterning of thoracic, lumbar, and sacral vertebrae, as well as the changes in the axial pattern resulting from mutations in the Hox10 and Hox11 paralogous genes, can be explained by a cascade of negative regulatory events among these genes that is analogous to the model first proposed by E. B. Lewis to explain the patterning of the Drosophila thoracic and abdominal segments by the Bithorax complex (19). That is, Hox10 paralogous genes suppress the formation of thoracic ribs in the lumbar through sacral region. Hox11 genes, in turn, partially suppress Hox10 activity in the sacral region, thereby activating the formation of modified ribs that, under the control of Hox11 genes, fuse and form the sacrum (Fig. 4A). If the ground state for rib formation extends from the head to the tail, then a similar rib-suppressive mechanism, mediated by more anteriorly expressed Hox genes, may be used to suppress rib formation in the cervical vertebrae. It remains to be shown whether the genetic suppressive mechanisms described above are direct or indirect at the molecular level. In more recent evolutionary history, snakes are a dramatic example of vertebrates acquiring, in a sense, a more primitive vertebral body plan through potential changes of Hox gene expression pattern (20).

Fig. 4.

Schematic representation of Hox patterning. (A) diagrams the axial phenotypes resulting from loss of Hox function. The axial vertebrae are shown as green blocks (for simplicity, only three caudal vertebrae are shown). The function of the Hox10 paralogous genes is to suppress thoracic development posterior to the 13th thoracic vertebra. In this model, Hox11 paralogous genes positively regulate the formation of sacral vertebrae by partially suppressing Hox10 function in the sacral region. In (B), functional domains of the AbdB Hox genes in forelimb patterning are diagrammed. Hox9 and Hox10 paralogs function together to pattern the forelimb stylopod. Hox10 paralogs also display some phenotype in the zeugopod (lighter orange shading). Hox11 paralogous genes function mainly in patterning the developing zeugopod, with a lesser contribution to autopod patterning (lighter yellow shading). Hox13 paralogs function predominantly in the autopod. In the hindlimb (C), Hox9 paralogs do not function. Hox10 paralogs function predominantly to pattern the stylopod. Hox11 paralogous genes function mainly in patterning the developing zeugopod, with some contribution to autopod patterning (lighter yellow shading). Hox13 paralogs function predominantly in the autopod. [Recent work provides evidence that Hoxd12 can substitute for Hox13 function in the autopod; patterning therefore is represented with light green shading for Hox12 function in the autopod (20). In (B) and (C), S denotes stylopod, Z zeugopod, and A autopod.]

The results from this study also extend our understanding of the roles of Hox genes in patterning the principal elements of the limbs (Fig. 4, B and C). In the hindlimb, Hox10 paralogous genes are required to pattern the stylopod, and Hox11 paralogous genes are required to pattern the zeugopod. In mice that are triple mutant for Hox10 or Hox11 paralogous genes, the femur, or the tibia and fibula, respectively, are grossly mispatterned. Fromental-Ramain et al. have previously shown that in the absence of Hoxa13 and Hoxd13 function (the only Hox13 paralogs that are expressed in the developing limb bud), the autopods of the forelimb and hindlimb are grossly malformed (3). In the forelimb, disruption of the Hox10 paralogous group affects the formation of the stylopod (that is, the humerus) but to a substantially lesser degree than in the hindlimb. Previous work has shown that mice mutant for both Hoxa9 and Hoxd9 exhibit humerus defects very similar to those in the Hox10 triple mutants (6). The Hox9 paralogous mutations reported, however, had no effect on the patterning of the hindlimbs. This suggests that Hox9 and Hox10 paralogous genes may function together in the patterning of the humerus. Taken together with the previous results (3, 6, 21), the above results complete the assignment of the principal Hox genes involved in the patterning of the major limb elements: the stylopod, zeugopod, and autopod.

This study has highlighted the extent of functional redundancy retained among Hox paralogous groups as well as the importance of the AbdB-group Hox genes in patterning the axial and appendicular skeleton. The Hox1 through Hox8 paralogous groups are related to individual Drosophila HomC homologs. Before vertebrate radiation, the most 5′ HomC member, AbdB, underwent additional tandem duplications, resulting in the Hox9 through Hox13 paralogous groups. These AbdB-related genes comprise 16 of the 39 mammalian Hox genes. The degree to which expansion of the vertebrate AbdB group of genes has contributed to the evolution of the vertebrate body plan is remarkable. The Hox9 through Hox13 genes appear to be largely responsible for Hox patterning of the limbs as well as the axial skeleton posterior to the thoracic vertebrae. We have shown that the Hox10 and Hox11 paralogous genes are global regulators of the lumbosacral region of the axial skeleton and are integral in patterning principal limb elements. Both in the formation of the axial skeleton and in the limbs, all members of a paralogous family that are expressed in a given structure must be disrupted before the full nature of the mutant phenotype is realized. By removing the redundancy in this system, we are beginning to understand the fundamental role these genes play in patterning the vertebrate skeleton.

Supporting Online Material

SOM Text


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article