• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of janatLink to Publisher's site
J Anat. Jun 2004; 204(6): 487–499.
PMCID: PMC1571319

Alx4 and Msx2 play phenotypically similar and additive roles in skull vault differentiation


Alx4 and Msx2 encode homeodomain-containing transcription factors that show a clear functional overlap. In both mice and humans, loss of function of either gene is associated with ossification defects of the skull vault, although the major effect is on the frontal bones in mice and the parietal bones in humans. This study was undertaken to discover whether Alx4 and Msx2 show a genetic interaction in skull vault ossification, and to test the hypothesis that they interact with the pathway that includes the Fgfr genes, Twist1 and Runx2. We generated Alx4+/–/Msx2+/– double heterozygous mutant mice, interbred them to produce compound genotypes and analysed the genotype–phenotype relationships. Loss of an increasing number of alleles correlated with an incremental exacerbation of the skull vault defect; loss of Alx4 function had a marginally greater effect than loss of Msx2 and also affected skull thickness. In situ hybridization showed that Alx4 and Msx2 are expressed in the cranial skeletogenic mesenchyme and in the growing calvarial bones. Studies of the coronal suture region at embyonic day (E)16.5 revealed that Alx4 expression was decreased, but not abolished, in Msx2−/– mutants, and vice versa; expression of Fgfr2 and Fgfr1, but not Twist1, was reduced in both mutants at the same stage. Runx2 expression was unaffected in the coronal suture; in contrast, expression of the downstream ossification marker Spp1 was delayed. Double homozygous pups showed substantial reduction of alkaline phosphatase expression throughout the mineralized skull vault; they died at birth due to defects of the heart, lungs and diaphragm not previously associated with Alx4 or Msx2. Our observations suggest that Alx4 and Msx2 are partially functionally redundant, acting within a network of transcription factors and signalling events that regulate the rate of osteogenic proliferation and differentiation at a stage after the commitment of mesenchymal stem cells to osteogenesis.

Keywords: coronal suture, craniofacial development, mouse, ossification, parietal foramina


The skull vault forms by intramembranous ossification, in which osteoprogenitor cells (preosteoblasts) are continuously produced by cell proliferation and converted to bone matrix-secreting osteoblasts at the bone margins. Osteogenic differentiation requires activity of Runx2, which encodes a transcriptional activator (CBFA1) that directly regulates the expression of osteoblast-specific genes (Komori et al. 1997; Otto et al. 1997). Alx4 and Msx2 encode transcription factors that play regulatory roles in osteogenesis. Their importance for intramembranous ossification has been revealed by studies on mutant mice (Qu et al. 1997b, 1998; Satokata et al. 2000) and by the identification of loss-of-function mutations of these genes in defective ossification of the human skull vault, enlarged parietal foramina (PFM) (Wilkie et al. 2000; Wuyts et al. 2000a, b; Mavrogiannis et al. 2001; Garcia-Miñaur et al. 2003). Additionally, an activating mutation of MSX2 causing a Pro→His change at the seventh position of the homeodomain that enhances its DNA binding affinity has been identified in a family with synostosis of the sagittal and other sutures (Jabs et al. 1993; Warman et al. 1993; Ma et al. 1996).

These apparently opposite effects of gain- and loss-of-function mutations suggest that craniosynostosis and PFM may be caused by reciprocal forms of imbalance of the process by which proliferating osteogenic stem cells are converted to differentiating osteoblasts. Craniosynostosis is also associated with activating mutations of genes encoding the fibroblast growth factor receptors FGFR1, -2 and -3 and haploinsufficiency of the transcription factor TWIST1 (see Wilkie & Morriss-Kay, 2001, for details and references). Evidence from Drosophila (Shishido et al. 1993) and mouse (Rice et al. 2000) suggests that Twist1 may be upstream of the FGFR pathway. In mouse studies, proliferation of osteogenic precursor cells has been shown to be functionally associated with expression of Fgfr2, whereas osteoblast differentiation is associated with expression of Fgfr1 (Iseki et al. 1999). Paradoxically, PFM are also associated with some cases of craniosynostosis resulting from mutations in TWIST1 (Saethre–Chotzen syndrome: Wilkie et al. 1995; Gripp et al. 1999).

Alx4−/– and Msx2−/– mouse mutants both show defective skull vault ossification, involving retarded growth of the frontal bones towards each other in the midline; in Alx4−/– mutants, the parietal bones are also affected (Qu et al. 1997b; Satokata et al. 2000). Analysis of Msx2−/– mice has revealed that the defect is associated with reduced proliferation of osteoprogenitors at the ossification front (Satokata et al. 2000); this effect is exacerbated by haploinsufficiency of Twist1 (Ishii et al. 2003). Msx2−/– mice show features additional to those of the human haploinsufficiency syndrome, including abnormalities of tooth, hair follicle and mammary gland development; there are also effects on endochondral ossification, involving deficient PTH/PTHrP signalling (Satokata et al. 2000). Msx2 is involved in diverse developmental processes, including patterning of neural crest, limb, branchial arches and skin appendages, neural induction, eye development, and the regulation of cell proliferation, differentiation and survival during organogenesis (Bendall & Abate-Shen, 2000; Kwang et al. 2002). Alx4−/– mice are allelic with Strong's luxoid (lst) and related spontaneous mutants; they show patterning defects of the limbs (preaxial polydactyly) and abdominal wall defects (gastroschisis), which have no equivalent in PFM patients (Qu et al. 1997b, 1998). Alx4 is expressed in several mesenchymal populations and is involved in pattern formation and control of growth (Meijlink et al. 1999). Unlike Msx2, Alx4 does not appear to participate in endochondral ossification.

Alx4 and Msx2 both encode homeodomain-containing transcription factors, but of different groups and with different properties. Alx4 is related to the Drosophila gene aristaless; the protein contains a paired-type homeodomain that binds to DNA either as a homodimer or as a heterodimer with other members of the same family such as CART1 (Qu et al. 1997a, 1999). It is a potent transcriptional activator that is expressed at sites of epithelial–mesenchymal interactions including hair follicles, and in osteoblast precursors (Hudson et al. 1998). ALX4 is coexpressed and interacts physically with lymphoid enhancer factor-1 (LEF-1) (Boras & Hamel, 2002). Msx2 shows a sequence similarity to the Drosophila gene muscle segment homeobox (msh) (Hill et al. 1989). It responds to BMP signalling (Kim et al. 1998), and is a transcriptional repressor (Bendall & Abate-Shen, 2000). A well-established target is the osteocalcin promoter, which is repressed when MSX2 inhibits binding of a DNA binding protein to a response element that is activated by both FGF signalling and CBFA1 (Newberry et al. 1997, 1999; Willis et al. 2002).

Although Alx4 and Msx2 have different spectra of biochemical activities, the similar phenotypic outcome of their loss-of-function mutations on skull vault ossification suggests that in this context they may act within the same molecular pathway. This study was undertaken to test the hypothesis that in skull vault ossification, Alx4 and Msx2 act through a common pathway that interacts with the Fgfr genes, Twist1 and Runx2. We generated double heterozygous Alx4+/–/Msx2+/– mutant mice and interbred them to produce all possible compound genotypes. We analysed and compared the effects of the loss of Alx4 and/or Msx2 alleles on the skull phenotype, the extracranial skeleton and soft tissues, and on the expression of Fgfr2, Fgfr1, Twist1, Runx2 and Spp1. The phenotypes show an incremental widening of the calvarial defect with increasing loss of functional Alx4 and Msx2 copies. Gene expression comparisons suggest that both genes exert an influence on transcription of Fgfr1, Fgfr2 and Spp1 but not Twist1 or Runx2. Expression of Alx4 was affected, but not extinguished, by loss of functional Msx2 and vice versa. These observations suggest that the two genes act within the same regulatory network in skull vault ossification, with a degree of functional redundancy. Analysis of double homozygotes revealed that Alx4 and Msx2 play an essential and mutually redundant role in development of the heart, lungs and diaphragm.

Materials and methods


Msx2+/– animals (Satokata et al. 2000) on an approximately 80% C57BL/6: 20% BALB/c background (a gift from Dr Robert Maxson) and Alx4+/– animals (Qu et al. 1997b) on an Sv129 background (a gift from Dr Ron Wisdom) were used in this study. Matings with a pure strain brought the C57BL/6 background representation to over 99% in the Msx2+/– line. Male Msx2+/– animals were mated with female Alx4+/– animals to generate double heterozygous animals (i.e. Alx4+/–/Msx2+/–). Offspring were genotyped to identify the double heterozygotes, which were then crossed to generate the nine possible genotype combinations: Alx4+/+/Msx2+/+, Alx4+/+/Msx2+/–, Alx4+/–/Msx2+/+, Alx4+/–/Msx2+/–, Alx4+/+/Msx2−/–, Alx4−/–/Msx2+/+, Alx4+/–/Msx2−/–, Alx4−/–/Msx2+/– and Alx4−/–/Msx2−/–. For skeletal staining and to assess the proportion of different genotypes, pups were collected at embryonic (E) day 17.5, and postnatal (P) days 1, P2, P3, P5 and P7. For in situ hybridization, wild-type and Msx2−/– C57BL/6 embryos, and wild-type and Alx4−/– Sv129 embryos, were generated from heterozygous matings and collected at E11.5, E13.5, E15.5 and E16.5. Genotyping to detect wild-type and null alleles of both Alx4 and Msx2 was performed as previously described (Qu et al. 1997b; Satokata et al. 2000). Either tail tips or pieces of skin were used, as appropriate.

In situ hybridization

Whole specimen in situ hybridization was carried out on embryos at E11.5, which were subsequently sectioned using a cryostat. In situ hybridization on transverse (horizontal) sections of heads from E12.5 to E16.5 was carried out on 10-µm frozen sections as described previously (Iseki et al. 1997). In situ hybridization of wild-type embryos was also carried out on sections of wild-type heads in the coronal and sagittal planes (data not shown; see Mavrogiannis, 2003). Serial sections from each head were mounted sequentially on separate slides for each of the five osteogenesis-related probes to be compared (Fgfr1, Fgfr2, Twist1, Runx2 and Spp1) so that no comparison involved more than a 50-µm difference of level. Alx4 expression was compared in wild-type and Msx2−/– heads; Msx2 expression was compared in wild-type and Alx4−/– heads. For each probe, wild-type and mutant slides were taken through the hybridization procedure together, with the same times for each stage, to enable quantitative comparisons to be made. The decision to stop colour development was taken on the basis of the appearance of wild-type sections; mutant sections were stopped at the same time.

Skeletal staining and histology

Newborn mice were killed by intraperitoneal injection of 150 mg kg−1 sodium pentobarbitone (Euthatal, Rhone Merieux), skinned, fixed in 95% alcohol and stained with Alcian blue 8GX (BDH-Merck) and Alizarin Red S (Raymond Lamb). Specimens were photographed on a Wild Heerbrugg microscope equipped with a Leica MPS 60 camera. For histology, heart and lung tissue was fixed in Bouin's fluid, embedded in paraffin wax, sectioned and stained with Ehrlich's haematoxylin and aqueous eosin. Sections were photographed with a Leica DMRBE microscope equipped with a Leica MPS 60 camera.


Incidence of genotypes

Pups were collected and genotyped at E16.5, E17.5, P1 (as soon as possible after birth), P3, P5 and P7. Table 1 shows the number of pups of each genotype at each stage, compared with the expected incidence. Statistical analysis showed a marginal difference between the expected and actual incidence: for ν = 6 degrees of freedom following merging of the minor classes, χ2 = 13.61, P < 0.05. A clearer difference was observed for the animals that survived after birth: for ν = 6 degrees of freedom, χ2 = 14.57, P < 0.025. Of the four genotypes expected in a 1: 16 ratio, all three mutant groups were under-represented. Few double homozygotes survived to term and all of these died soon after birth, so were not found beyond P1. Gastroschisis has been reported in Alx4−/– newborn pups (Qu et al. 1997b); we found this abnormality only at prenatal stages, so although littering females were observed closely, some pups with Alx4−/– genotypes may have been missed at P1, and have been under-represented at later stages. The small number of Alx4+/+/Msx2−/– animals is likely to be coincidental, because several Alx4+/–/Msx2−/– mutants were obtained.

Table 1
Incidence of genotypes (n = 122) obtained from Alx4+/–×Msx2+/– matings

Skull phenotypes are independent of strain differences

On the Sv129 background, the Alx4 null mutation is recessive (Qu et al. 1997b), but the penetrance and severity of the limb phenotype is sensitive to strain-specific modifier genes (Forsthoefel, 1968; Qu et al. 1998). The phenotype associated with the Msx2 mutation has been reported not to be dependent on strain background (Satokata et al. 2000). The double mutants featured a 50% C57BL/6: 50% Sv129 background on average, but random assortment of chromosomes and a round of recombination have produced non-comparable genomes. In order to rule out the possibility that the observed skull phenotypes were influenced by this variability of genetic background, wild-type Sv129 females were crossed with C57BL/6 Msx2+/– males, and wild-type C57BL/6 females crossed with Sv129 Alx4+/– males. The resulting Msx2+/– and Alx4+/– offspring, both with an unfragmented 50% C57BL/6: 50% Sv129 genome, were collected at P1 and P3 and compared to Msx2+/– and Alx4+/– animals with patchy 50% C57BL/6: 50% Sv129 backgrounds. The skull phenotypes showed minor variation between individuals of the same genotype but no overall differences between the two groups, indicating that background differences had no significant effect.

Skull phenotypes

A detailed analysis of skeletons of each genotype was performed at P1, and for each available genotype at P3 and P7. A typical set of P1 skull vaults is illustrated in Fig. 1; some variation was observed, as described below. The figure has been ordered to show the increasing severity of the mineralization defect. For each equal number of null alleles, loss of Alx4 produced a wider calvarial defect than loss of Msx2. The defect caused by the loss of two alleles was less severe in double heterozygotes than in either single homozygote, where widening of different parts of the interfrontal (metopic) suture (anterior for Alx4−/–, posterior for Msx2−/–) was seen. In the ordered series, the incremental loss of alleles can be seen to be associated with a progressive decrease in skull size. The presence of small interfrontal bones was variable, and was observed in pups of all genotypes.

Fig. 1
Skulls of newborn (P1) pups, vertex view, arranged in order of increasing severity of the ossification defect. The skull base, which can be seen through the gap in the vault, is normal for all genotypes. Scale bar, 1 mm.

An effect on the sagittal (interparietal) suture was associated only with the loss of Alx4, but it was increased with the loss of one or both functional copies of Msx2. Alx4+/–/Msx2−/– skulls also showed reduction of the parietal bone in the rostrocaudal axis. Decreased mineralization at the sutural edges of the parietal bones was sometimes observed with the loss of one Alx4 copy (data not shown), but the effect was minor and variable except where two copies of Msx2 were also missing.

Reduction of the interparietal bone was observed with the loss of two Msx2 copies, and also with one Msx2 allele when both Alx4 copies were missing. Reduction of the interparietal bone was always associated with reduction or loss of the supraoccipital bone (an endochondral bone formed from the occipital sclerotome) (Fig. 2). Mandible length was normal in double heterozygotes (not shown) but reduction was associated with the loss of two copies of Msx2, as previously reported (Satokata et al. 2000), and with the Alx4−/–/Msx2+/– genotype; the effect was greatest in double homozygotes (Fig. 2). The squamosal bone was reduced in skulls lacking three functional alleles, and was represented only by its posterior process in double homozygotes. The coronal suture was widened (showing no overlap of frontal and parietal bones) in skulls lacking two functional alleles of Msx2, except for double homozygotes, in which alkaline phosphatase-stained sections revealed contact between the frontal and parietal bones (Fig. 3).

Fig. 2
Side views of P1 skulls to show reduction of the mandible and squamosal bone with loss of three or four alleles. The supraoccipital bone is reduced in Alx4−/–/Msx2+/– mutants (arrowhead) and absent in the other two (the right exoccipital ...
Fig. 3
Transverse sections through the frontal (f) and parietal (p) bones of (A) a wild-type and (B) a double homozygous mutant skull at P1, stained with alkaline phosphatase. Alkaline phosphatase levels are greatly reduced in the double mutant, which also shows ...

In P3 and P7 pups, ossification had progressed so that the phenotypes were progressively milder, confirming that the calvarial gaps were the result of an ossification delay rather than a permanent defect. Transverse sections of the skull at P3 showed that compared with wild-type pups, the calvarial bones were reduced in thickness in single homozygous mutants and in pups lacking two or more functional alleles of either or both genes (data not shown, but see gene expression section below for sections of single homozygous mutant skulls at E16.5). Alkaline phosphatase immunoreactivity was normal in double heterozygous skulls (data not shown) but reduced in double homozygotes (Fig. 3). Other genotypes were not tested.

Expression of Alx4 and Msx2

Expression of Alx4 in whole embryos at E11.5 was as described previously (Qu et al. 1997a, b; Beverdam & Meijlink, 2001; Rice et al. 2003), i.e. in a discrete anterior domain in the limb, in the frontonasal process and the first arch mesenchyme. In addition to these previously documented domains, transcripts were detected posterior to the eyes (Fig. 4A). Msx2 transcripts were detected around the periphery of the limb bud and at the distal tips of the nasal, maxillary and mandibular processes as described previously (Davidson et al. 1991; Bendall & Abate-Shen, 2000) but also in a similar domain to that of Alx4 posterior to the eye (Fig. 4B). Sections of these specimens at eye level and above (i.e. the levels in which the future skull vault will differentiate) revealed that expression of both genes was exclusively within the dermal mesenchyme (Fig. 4C–F). The Alx4 expression domain was more extensive anteriorly than that of Msx2, especially in sections cut at lower levels. Only Msx2 was expressed in dermal mesenchyme between the hindbrain and surface ectoderm.

Fig. 4
Expression of Alx4 and Msx2. (A,B) In situ hybridization carried out in whole, wild-type E11.5 embryos. (C–F) Transverse sections of the same E11.5 embryos (levels as indicated on A and B, through hindbrain and cerebral hemispheres (ch) in C,E, ...

In situ hybridization was also carried out on sections, from E12.5 to E16.5 (see Mavrogiannis, 2003, for detailed illustrations). At all stages, both Alx4 and Msx2 were expressed in the developing bones; Alx4 transcripts were more strongly detectable than Msx2 in the osteoid and preosteoblast regions of the sutural margins (Fig. 4G,H). Only Msx2 was expressed in differentiating cranial cartilages (Mavrogiannis, 2003), as previously observed for endochondral bones generally (Satokata et al. 2000).

Alx4 expression was also studied in Msx2−/– mutant embryos, and Msx2 in Alx4−/– embryos. Alx4 expression was slightly reduced in Msx2−/– mutant heads, and Msx2 expression was greatly reduced in the Alx4−/– heads (Fig. 4I,J). The latter effect was not only due to the reduced bone thickness, but was also observed in sections at a lower level, close to the eye, where the frontal bones are thickest (data not shown).

Osteogenesis-related gene expression in mutant embryos

To obtain more detailed information on possible differences between the function of Alx4 and Msx2 in cranial vault development, in situ hybridization was carried out on transverse (horizontal) sections of the head of littermate wild-type, Alx4−/– and Msx2−/– embryos at E13.5, E14.5, E15.5 and E16.5, using probes to detect expression of Runx2, Spp1, Fgfr1, Fgfr2 and Twist1 on adjacent sections for each genotype, in the frontal and parietal bones and in the coronal suture. The results were consistent at all stages; the sections illustrated are of E16.5 heads (Fig. 5). Because there were no differences in gene expression in C57BL/6 or Sv129 wild-type sections from littermates of either of the two mutants, the wild-type sections illustrated for each probe were selected from either group.

Fig. 5
In situ hybridization of frozen transverse sections through the frontal (f) and parietal (p) bones of E16.5 wild-type, Alx4−/–, and Msx2−/– heads at equivalent levels (see diagram, top) to show expression of the genes indicated. ...

At this stage of development, the frontal bones of wild-type embryos are already thicker and more trabeculated than the parietal bones; both bones were thinner in sections of Alx4−/– and Msx2−/– heads compared with wild-type, the thinning being more pronounced in Alx4−/– sections. Expression of Fgfr2 and Fgfr1 was normal in the differentiated bone domains of Msx2−/– sections; there may have been a slight decrease in the sutural region but if so this was minimal. Fgfr2 expression was decreased throughout the whole frontal/parietal bone domains in Alx4−/– sections; an apparent decrease in Fgfr1 expression in Alx4−/– sections may simply be due to the obvious thinning of the bones. Runx2 expression was identical to wild-type in both mutant genotypes, but Spp1 expression was decreased, showing a wider sutural gap than in wild-type sections. Twist1 expression was identical in head sections of wild-type and both mutant genotypes (data not shown). These observations at E16.5 correspond to those of Ishii et al. (2003) at earlier stages with respect to Twist1 (E10.5 to E12.5) but not Runx2 expression, which was found to be decreased in the frontal bone primordium at E14.5. Sections at higher levels than those illustrated show a lack of expression of Runx2 in the frontal but not the parietal bone domain, corresponding with the reduced vertical extension of the frontal bone observed in coronal sections by Ishii et al. (2003), and consistent with the ossification defects observed at P1.

Extracranial features of Alx4−/–/Msx2−/– pups

The thoracic region of double mutant pups was greatly reduced. Skeletal preparations showed downward-angled ribs, absence of the manubrium, and reduced clavicles (Fig. 6A–D). In the pelvis the pubic bone was reduced in size in Alx4−/– pups and absent in double mutants (Fig. 6E). Similarly, the limb bones were shortened and the degree of luxation of both fore- and hind-limbs was greater in double homozygotes than in Alx4−/– or Alx4−/–/Msx2+/– pups (data not shown). The heart was reduced in size with atria embedded within the surrounding lungs, which were composed of an increased number of small lobes (Fig. 6F,G). Histological sections of the heart (Fig. 6H,I) revealed multiple abnormalities, including incomplete atrioventricular valves and reduced lumens of all chambers. Analysis of the whole series of sections revealed that although the pulmonary and aortic trunks arose from the correct ventricles, the outflow tract was shifted to the left: with regard to the section illustrated in Fig. 6(I), the rudimentary left ventricular lumen communicates with the aortic trunk above it on a nearly adjacent section, whereas the right ventricular lumen continues as a long channel over the muscular interventricular septum and behind the aorta before reaching an identifiable pulmonary trunk. The anatomical arrangement suggests that blood entering the upper parts of the ventricles was pumped out directly into the outflow tract; consistent with this interpretation, no blood cells were observed in the muscular part of the ventricles. There was no diaphragm, and the stomach and duodenum were located between lobes of the lung (Fig. 6J). Only one lung lobe showed expanded, air-filled alveoli (data not shown), although the larger airways of all lobes were open. No cartilaginous rings were present in the lower part of the trachea, which was collapsed. The upper trachea was also unstained on the skeletal preparations (compare Fig. 6B,D), suggesting that cartilaginous rings were absent along the whole length of the trachea.

Fig. 6
Extracranial features of the Alx4−/–/Msx2−/– phenotype at P1. (A–D) Alcian blue/Alizarin red-stained skeleton, cervical and thoracic region of wild-type pup (A,B) and double mutant (C,D) (right forelimb removed). ...


Genotype–phenotype observations

This study has shown that the skull defects associated with loss-of-function mutations of Alx4 and Msx2 are exacerbated when the two mouse mutant strains are combined. Loss of an increasing number of wild-type alleles has an incremental effect on the severity of the skull vault ossification defect; furthermore, it affects bones other than the frontals and parietals and overall skull growth. For the same number of null alleles, loss of Alx4 has a more severe effect than the loss of Msx2. Overall, the genetic data indicate gross additivity of the two genes as far as the calvarial phenotype is concerned; however, a single Msx2 null allele appears to act as a modifier in Alx4 mutants.

Unlike human PFM, in which the skull ossification defect resulting from haploinsufficiency of either gene cannot be distinguished on the basis of phenotype alone, the skull features of the two single homozygous mutant mice are anatomically distinct. Msx2−/– skulls show a defect that is more severe in the posterior region of the interfrontal suture, whereas that of Alx4−/– skulls is most severe in the anterior frontal region and also includes the parietal bones, as well as having a greater effect on skull thickness. There are also distinct roles in endochondral ossification, Msx2 being required for the somite-derived supraoccipital bone, and Alx4 for the pubic bone. These detailed aspects of the phenotypes suggest that the two transcription factors play similar but subtly different roles and can partially, but not fully, functionally substitute for each other.

Normal expression patterns

Similarities and differences were also revealed by expression patterns of the two genes. Both genes are expressed in the dermal mesenchyme of E11.5 heads, but only Alx4 is expressed posterior to the eye, in the dermal mesenchyme in which parietal ossification is initiated (Iseki et al. 1997; Kim et al. 1998). This correlates with the effects of loss of Alx4 function on parietal ossification, although both genes were expressed at more superior levels of the future parietal bone domain. Msx2, but not Alx4, was expressed in the dermal mesenchyme overlying the hindbrain, correlating with its later requirement for ossification of the endochondral supraoccipital bone.

Liu et al. (1995) suggested that the ossification abnormality caused by the Pro→His activating mutation of MSX2 might be due to an effect on neural crest cells during migration, specification, or later stages. Neural crest cell lineage mapping in the mouse has indeed shown the frontal bone and the mesenchyme within the sagittal suture to be neural crest-derived, but the domain of neither Alx4 nor Msx2 expression in the cranial dermal mesenchyme at E11.5 respects the neural crest–mesoderm boundary at that stage, as detected by the Wnt1–Cre/R26R neural crest lineage marker (Jiang et al. 2002). The functions of these two genes in the skull appear to be related to the process of osteogenic proliferation and/or differentiation or apoptosis, in both the neural crest-derived frontal bone and the mesodermal parietal bone. Different sites of the murine and human defects similarly suggest an effect on ossification that is unrelated to tissue origin.

Effects of loss of function mutations on gene expression patterns

Loss of Alx4 function decreased Fgfr2 expression throughout the parietal and frontal bones; the effects of loss of Msx2, if any, were located mainly in the proliferating/differentiating cells at the coronal sutural margins of the two bones. FGFR2 signalling is associated with proliferation of osteogenic stem cells, and FGFR1 with their differentiation to bone-forming osteoblasts (Iseki et al. 1999). This difference is consistent with the observation that Alx4−/– skulls were thinner than those of Msx2−/– pups, suggesting that the Alx4−/– defect may lead to a greater decrease in the number of cells available for bone formation. Hu et al. (2001) found that MSX2 inhibits differentiation and causes a slight increase in cell proliferation by maintaining cyclin D1 expression in a variety of cultured cell types and prevented exit of cells from the cell cycle. In Msx2−/– heads, Satokata et al. (2000) and Ishii et al. (2003) observed a decrease in cell proliferation but no enhanced apoptosis in the developing calvarial bones. The nature of the primary defect in the Alx4−/– skull vault remains to be elucidated.

The finding that Alx4 and Msx2 expression was reduced but not abolished in Msx2−/– and Alx4−/– mutants, respectively, suggests that each gene is involved rather indirectly in regulation of the other. The interpretation that Alx4 and Msx2 are at least partially interdependent in their calvarial osteogenic activity is in accordance with the observation that both genes show Foxc1-dependent activation by BMP2 (Rice et al. 2003).

No effect was detected in the expression of Runx2 in either Alx4−/– or Msx2−/– heads at E16.5, but expression of the bone marker Spp1 was excluded from a wider area of the differentiating sutural tissue in both mutants. These observations suggest that the involvement of both Alx4 and Msx2 in the transcriptional regulation of skull vault osteogenesis is downstream of Runx2.

The reduced frontal bone Runx2 expression domain in sections at apical levels in both Alx4−/– and Msx2−/– mutants is consistent with the delayed upward extension of the frontal bone primordium reported by Ishii et al. (2003). A recent cell labelling study has revealed that the frontal bone primordium, which is initiated as a group of Runx2-expressing cells just above the eye, extends upwards after E14 towards the interfrontal midline by the movement of its proliferating cells through mesenchymal tissue between the brain and surface ectoderm, and not by the recruitment of new mesenchymal cells (Y. Yoshida et al. personal communication). We suggest that the effect of loss of function of either Alx4 or Msx2 on Runx2 expression at the upper margin of the frontal bone is due to delayed vertical expansion of the frontal bone primordium, and not to a direct effect on Runx2 expression. This interpretation is supported by the observation that Runx2 expression was unaffected in the adjacent margins of the frontal and parietal bones in transverse sections of the coronal suture, which is established by E14 as a close juxtaposition of the two bones (Johnson et al. 2000). Further evidence that the effects are downstream of Runx2 and mainly related to intramembranous ossification is the relative lack of effect on osteogenesis elsewhere in the skeleton: even the double homozygotes show well-ossified bones, in contrast to the complete lack of bone in Runx2−/– mutants (Komori et al. 1997; Otto et al. 1997).

Investigation of a mouse model for the activating Fgfr1 mutation associated with Pfeiffer-type craniosynostosis indicated that Runx2 is downstream of Fgfr1 (Zhou et al. 2000). Increased FGFR signalling has been shown to result in up-regulation of Msx2 and Runx2 expression and premature sutural fusion (Ignelzi et al. 2003). Our observation of decreased Fgfr1 expression in Msx2−/– skulls suggests a reciprocal interaction between the two genes. Similarly, Msx2 can be both upstream and downstream of BMP signalling (compare Kwang et al. 2002, with Rice et al. 2003).

The results of this study indicate that Alx4 and Msx2 play additive and partially functionally redundant roles in skull vault osteogenesis, and show complete functional redundancy in heart, lung and diaphragm development. Differences in the skull phenotypes of single and combined mutants imply the existence of co-factors that are specific to each gene. Differences between the results of in vivo and in vitro studies, and between the skull vault and other skeletal structures, indicate context-specific features of the interactions of ALX4 and MSX2 with other factors. We suggest that in skull vault osteogenesis, Alx4 and Msx2 function within a network of transcription factors and signalling events that regulate the rate of osteogenic proliferation and differentiation, acting downstream of the events that determine the osteogenic potential of calvarial bone stem cells.


We thank Robert Maxson and Ron Wisdom for provision of mutant mice. Research expenses were provided by grants from Action Research (G.M.M.K.) and the Wellcome Trust (A.O.M.W.). I.A. was supported by an Anatomical Society studentship and L.M. by the MRC and the Alexander S. Onassis Foundation.


  • Bendall AJ, Abate-Shen C. Roles for Msx and Dlx homeoproteins in vertebrate development. Gene. 2000;247:17–31. [PubMed]
  • Beverdam A, Meijlink F. Expression patterns of group-I aristaless-related genes during craniofacial and limb development. Mech. Dev. 2001;107:163–167. [PubMed]
  • Boras K, Hamel PA. Alx4 binding to LEF-1 regulates N-CAM promoter activity. J. Biol. Chem. 2002;277:1120–1127. [PubMed]
  • Davidson DR, Crawley A, Hill RE, Tickle C. Position-dependent expression of two related homeobox genes in developing vertebrate limbs. Nature. 1991;352:429–431. [PubMed]
  • Forsthoefel PF. Responses to selection for plus and minus modifiers of some effects of Strong's luxoid gene on the mouse skeleton. Teratology. 1968;1:339–351. [PubMed]
  • Garcia-Miñaur S, Mavrogiannis LA, Rannan-Eliya SV, Hendry MA, Liston WA, Porteous ME, et al. Parietal foramina with cleidocranial dysplasia is caused by mutation in MSX2. Eur. J. Hum. Genet. 2003;11:892–895. [PubMed]
  • Gripp KW, Stolle CA, Celle L, McDonald-McGinn DM, Whitaker LA, Zackai EH. TWIST gene mutation in a patient with radial aplasia and craniosynostosis: further evidence for heterogeneity of Baller–Gerold syndrome. Am. J. Med. Genet. 1999;82:170–176. [PubMed]
  • Hill RE, Jones PF, Rees AR, Sime CM, Justice MJ, Copeland NG, et al. A new family of mouse homeo box-containing genes: molecular structure, chromosomal location, and developmental expression of Hox-7.1. Genes Dev. 1989;3:26–37. [PubMed]
  • Hu G, Lee H, Price SM, Shen MM, Abate-Shen C. Msx homeobox genes inhibit differentiation through upregulation of cyclin D1. Development. 2001;128:2373–2384. [PubMed]
  • Hudson R, Taniguchi-Sidle A, Boras K, Wiggan O, Hamel PA. Alx-4, a transcriptional activator whose expression is restricted to sites of epithelial–mesenchymal interactions. Dev. Dyn. 1998;213:159–169. [PubMed]
  • Ignelzi MA, Jr, Wang W, Young AT. Fibroblast growth factors lead to increased Msx2 expression and fusion in calvarial sutures. J. Bone Miner. Res. 2003;18:751–759. [PubMed]
  • Iseki S, Wilkie AOM, Heath JK, Ishimaru T, Eto K, Morriss-Kay GM. Fgfr2 and osteopontin domains in the developing skull vault are mutually exclusive and can be altered by locally applied FGF2. Development. 1997;124:3375–3384. [PubMed]
  • Iseki S, Wilkie AOM, Morriss-Kay GM. Fgfr1 and Fgfr2 have distinct differentiation- and proliferation-related roles in the developing mouse skull vault. Development. 1999;126:5611–5620. [PubMed]
  • Ishii M, Merrill AE, Chan YS, Gitelman I, Rice DP, Sucov HM, et al. Msx2 and Twist cooperatively control the development of the neural crest-derived skeletogenic mesenchyme of the murine skull vault. Development. 2003;130:6131–6142. [PubMed]
  • Jabs EW, Muller U, Li X, Ma L, Luo W, Haworth IS, et al. A mutation in the homeodomain of the human MSX2 gene in a family affected with autosomal dominant craniosynostosis. Cell. 1993;75:443–450. [PubMed]
  • Jiang X, Iseki S, Maxson RE, Sucov HM, Morriss-Kay GM. Tissue origins and interactions in the mammalian skull vault. Dev. Biol. 2002;241:106–116. [PubMed]
  • Johnson D, Iseki S, Wilkie AOM, Morriss-Kay GM. Expression patterns of Twist and Fgfr1, -2 and -3 in the developing mouse coronal suture suggest a key role for twist in suture initiation and biogenesis. Mech. Dev. 2000;91:341–345. [PubMed]
  • Kim HJ, Rice DP, Kettunen PJ, Thesleff I. FGF-, BMP- and Shh-mediated signalling pathways in the regulation of cranial suture morphogenesis and calvarial bone development. Development. 1998;125:1241–1251. [PubMed]
  • Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K, Deguchi K, et al. Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell. 1997;89:755–764. [PubMed]
  • Kwang SJ, Brugger SM, Lazik A, Merrill AE, Wu LY, Liu YH, et al. Msx2 is an immediate downstream effector of Pax3 in the development of the murine cardiac neural crest. Development. 2002;129:527–538. [PubMed]
  • Liu YH, Kundu R, Wu L, Luo W, Ignelzi MA, Jr, Snead ML, et al. Premature suture closure and ectopic cranial bone in mice expressing Msx2 transgenes in the developing skull. Proc. Natl Acad. Sci. USA. 1995;92:6137–6141. [PMC free article] [PubMed]
  • Ma L, Golden S, Wu L, Maxson R. The molecular basis of Boston-type craniosynostosis: the Pro148→His mutation in the N-terminal arm of the MSX2 homeodomain stabilizes DNA binding without altering nucleotide sequence preferences. Hum. Mol. Genet. 1996;5:1915–1920. [PubMed]
  • Mavrogiannis LA, Antonopoulou I, Baxova A, Kutilek S, Kim CA, Sugayama SM, et al. Haploinsufficiency of the human homeobox gene ALX4 causes skull ossification defects. Nat. Genet. 2001;27:17–18. [PubMed]
  • Mavrogiannis L. Thesis. University of Oxford; 2003. The roles of homeobox genes ALX4 and MSX2 in skull development.
  • Meijlink F, Beverdam A, Brouwer A, Oosterveen TC, Berge DT. Vertebrate aristaless-related genes. Int. J. Dev. Biol. 1999;43:651–663. [PubMed]
  • Newberry EP, Boudreaux JM, Towler DA. Stimulus-selective inhibition of rat osteocalcin promoter induction and protein–DNA interactions by the homeodomain repressor Msx2. J. Biol. Chem. 1997;272:29607–29613. [PubMed]
  • Newberry EP, Latifi T, Towler DA. The RRM domain of MINT, a novel Msx2 binding protein, recognizes and regulates the rat osteocalcin promoter. Biochemistry. 1999;38:10678–10690. [PubMed]
  • Otto F, Thornell AP, Crompton T, Denzel A, Gilmour KC, Rosewell IR, et al. Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development. Cell. 1997;89:765–771. [PubMed]
  • Qu S, Li L, Wisdom R. Alx-4: cDNA cloning and characterization of a novel paired-type homeodomain protein. Gene. 1997a;203:217–223. [PubMed]
  • Qu S, Niswender KD, Ji Q, van der Meer R, Keeney D, Magnuson MA, et al. Polydactyly and ectopic ZPA formation in Alx-4 mutant mice. Development. 1997b;124:3999–4008. [PubMed]
  • Qu S, Tucker SC, Ehrlich JS, Levorse JM, Flaherty LA, Wisdom R, et al. Mutations in mouse Aristaless-like4 cause Strong's luxoid polydactyly. Development. 1998;125:2711–2721. [PubMed]
  • Qu S, Tucker SC, Zhao Q, deCrombrugghe B, Wisdom R. Physical and genetic interactions between Alx4 and Cart1. Development. 1999;126:359–369. [PubMed]
  • Rice DPC, Aberg T, Chan YS, Tang Z, Kettunen PJ, Pakarinen L, et al. Integration of FGF and TWIST in calvarial bone and suture development. Development. 2000;127:1845–1855. [PubMed]
  • Rice R, Rice DPC, Olsen BR, Thesleff I. Progression of calvarial bone development requires Foxc1 regulation of Msx2 and Alx4. Dev. Biol. 2003;262:75–87. [PubMed]
  • Satokata I, Ma L, Ohshima H, Bei M, Woo I, Nishizawa K, et al. Msx2 deficiency in mice causes pleiotropic defects in bone growth and ectodermal organ formation. Nat. Genet. 2000;24:391–395. [PubMed]
  • Shishido E, Higashijima S, Emori Y, Saigo K. Two FGF-receptor homologues of Drosophila: one is expressed in mesodermal primordium in early embryos. Development. 1993;117:751–761. [PubMed]
  • Warman ML, Mulliken JB, Hayward PG, Muller U. Newly recognized autosomal dominant disorder with craniosynostosis. Am. J. Med. Genet. 1993;46:444–449. [PubMed]
  • Wilkie AOM, Yang SP, Summers D, Poole MD, Reardon W, Winter RM. Saethre–Chotzen syndrome associated with balanced translocations involving 7p21: three further families. J. Med. Genet. 1995;32:174–180. [PMC free article] [PubMed]
  • Wilkie AOM, Tang Z, Elanko N, Walsh S, Twigg SR, Hurst JA, et al. Functional haploinsufficiency of the human homeobox gene MSX2 causes defects in skull ossification. Nat. Genet. 2000;24:387–390. [PubMed]
  • Wilkie AOM, Morriss-Kay GM. Genetics of craniofacial development and malformation. Nat. Rev. Genet. 2001;2:458–468. [PubMed]
  • Willis DM, Loewy AP, Charlton-Kachigian N, Shao JS, Ornitz DM, Towler DA. Regulation of osteocalcin gene expression by a novel Ku antigen transcription factor complex. J. Biol. Chem. 2002;277:37280–37291. [PubMed]
  • Wuyts W, Cleiren E, Homfray T, Rasore-Quartino A, Vanhoenacker F, Van Hul W. The ALX4 homeobox gene is mutated in patients with ossification defects of the skull (foramina parietalia permagna, OMIM 168500) J. Med. Genet. 2000a;37:916–920. [PMC free article] [PubMed]
  • Wuyts W, Reardon W, Preis S, Homfray T, Rasore-Quartino A, Christians H, et al. Identification of mutations in the MSX2 homeobox gene in families affected with foramina parietalia permagna. Hum. Mol. Genet. 2000b;9:1251–1255. [PubMed]
  • Zhou YX, Xu X, Chen L, Li C, Brodie SG, Deng CX. A Pro250Arg substitution in mouse Fgfr1 causes increased expression of Cbfa1 and premature fusion of calvarial sutures. Hum. Mol. Genet. 2000;9:2001–2008. [PubMed]

Articles from Journal of Anatomy are provided here courtesy of Anatomical Society of Great Britain and Ireland
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...