|Home | About | Journals | Submit | Contact Us | Français|
Clubfoot affects 1 in 1000 live births, although little is known about its genetic or developmental basis. We recently identified a missense mutation in the PITX1 bicoid homeodomain transcription factor in a family with a spectrum of lower extremity abnormalities, including clubfoot. Because the E130K mutation reduced PITX1 activity, we hypothesized that PITX1 haploinsufficiency could also cause clubfoot. Using copy number analysis, we identified a 241 kb chromosome 5q31 microdeletion involving PITX1 in a patient with isolated familial clubfoot. The PITX1 deletion segregated with autosomal dominant clubfoot over three generations. To study the role of PITX1 haploinsufficiency in clubfoot pathogenesis, we began to breed Pitx1 knockout mice. Although Pitx1+/− mice were previously reported to be normal, clubfoot was observed in 20 of 225 Pitx1+/− mice, resulting in an 8.9% penetrance. Clubfoot was unilateral in 16 of the 20 affected Pitx1+/− mice, with the right and left limbs equally affected, in contrast to right-sided predominant hindlimb abnormalities previously noted with complete loss of Pitx1. Peroneal artery hypoplasia occurred in the clubfoot limb and corresponded spatially with small lateral muscle compartments. Tibial and fibular bone volumes were also reduced. Skeletal muscle gene expression was significantly reduced in Pitx1−/− E12.5 hindlimb buds compared with the wild-type, suggesting that muscle hypoplasia was due to abnormal early muscle development and not disuse atrophy. Our morphological data suggest that PITX1 haploinsufficiency may cause a developmental field defect preferentially affecting the lateral lower leg, a theory that accounts for similar findings in human clubfoot.
Isolated clubfoot, also called talipes equinovarus, is one of the most common serious congenital birth defects with an estimated birth prevalence of 1 per 1000 live births (1). Clubfoot consists of malalignment of the bones and joints of the foot and ankle and is distinguished from positional foot anomalies because it is rigid and not passively correctable. Clubfoot is associated with chromosomal abnormalities, or known genetic syndromes such as distal arthrogryposis, and myelomeningocele in 20% of cases (2,3). However, ~80% of clubfeet occur as isolated birth defects (4).
Several theories of clubfoot pathogenesis, including vascular, neurogenic, myofibrosis and disuse from lack of in utero fetal movement, have been proposed based on specific morphological abnormalities. Vascular anomalies, consisting of anterior tibial artery hypoplasia in the lower limb, are present in ~80% of clubfoot patients in several small case series, although the genetic basis of this abnormality has not been demonstrated (5,6). Unilateral clubfoot is often associated with smaller calf circumference compared with the unaffected limb, resulting in quantitative muscle abnormalities (7,8). However, electrophysiological studies of muscle and nerve are typically normal in clubfoot, and histological evaluations of muscle biopsies from clubfoot limbs are most often normal or show nonspecific abnormalities (9–12). Although skeletal structures are only minimally affected in clubfoot, leg length discrepancy may be present, indicating possible effects on skeletal growth (13). These widespread anatomic abnormalities suggest that clubfoot is either etiologically heterogeneous or that a single primary underlying etiology may be responsible for these effects on multiple tissues.
Important insight into the pathogenesis of clubfoot is being revealed through a better understanding of its genetic basis. Recent data from our laboratory support a role for the PITX1-TBX4 developmental pathway in clubfoot etiology. We identified a missense mutation in the bicoid-related homeobox transcription factor PITX1 gene in a multigenerational family with predominantly isolated clubfoot segregating with reduced penetrance (14). The PITX1 E130K mutation is located in the highly conserved homeodomain, reduces its ability to transactivate a luciferase reporter and causes dominant-negative effects on transcription (14). Although PITX1 gene mutations are not common in clubfoot (14), chromosome 17q23 copy number variants (CNVs) involving the T-box transcription factor TBX4, a transcriptional target of PITX1 (15,16), occur in ~5% of all familial clubfoot (17), suggesting the importance of this pathway. However, the morphological mechanisms by which these genetic abnormalities cause clubfoot have not been elucidated.
Here we report PITX1 haploinsufficiency as a genetic mechanism responsible for human clubfoot through identification of a novel microdeletion in a three-generation family with isolated clubfoot. We also describe the first genetic mouse model of clubfoot that likewise results from Pitx1 haploinsufficiency. Our analysis of the Pitx1 clubfoot mouse reveals vascular, skeletal and muscle hypoplasia that significantly parallel findings in some humans with clubfoot, and supports a developmental field defect theory of clubfoot etiology.
To identify the genes responsible for isolated clubfoot, 40 probands with at least one affected first-degree relative were screened for genomic CNVs with the Affymetrix Genome-wide Human SNP Array 6.0. A 241 kb microdeletion involving 124 markers on chromosome 5q31 was identified in one proband (Fig. 1A). The microdeletion, chr5:134222383–134463022 (hg18 build), overlaps four RefSeq genes (TXNDC15, PCBD2, CATSPER3 and PITX1). Of the genes located within the microdeletion, only PITX1 does not overlap a copy number variation previously observed in healthy individuals from the UCSC Database of Genomic Variants. The microdeletion was verified by quantitative polymerase chain reaction (qPCR) and was not present in 700 controls (18,19) evaluated with the same platform (Affymetrix 6.0). The chromosome 5q31 microdeletion was also present in two additional family members with isolated clubfoot and was not identified in unaffected relatives (Fig. 1B) (14). All affected individuals had bilateral clubfoot (Fig. 1C and D). Short stature, with height more than 2 SD below the mean, was also present in all individuals with the PITX1 deletion. None of the other features that were seen occasionally with the PITX1 E130K mutation, including tibial hemimelia, preaxial polydactyly, patellar hypoplasia or developmental hip dysplasia (14), were present in this family.
Magnetic resonance imaging (MRI) of the lower limbs of a patient with PITX1 E130K mutation showed a reduction in the overall size of the affected clubfoot limb, with reduced muscle and bone volumes (BVs) (Fig. 2A and B). The limb was more severely affected below the knee. Although all muscle compartments were involved, the anterior compartment containing the tibialis anterior muscle was particularly small and partially replaced with fat (Fig. 2B). Magnetic resonance angiography (MRA) demonstrated diminution of the anterior tibial and peroneal arteries on the affected limb compared with the unaffected limb (Fig. 2C).
Because we had identified both PITX1 haploinsufficiency (above) and a PITX1 missense mutation in human patients with clubfoot (14), we were interested in determining whether Pitx1 haploinsufficient mice also had limb abnormalities. Homozygous Pitx1 knockout mice die immediately or shortly after birth, with significant hindlimb abnormalities including reduced femur and tibia lengths (20,21). Clubfoot or clubfoot-like features were not described in these mice and heterozygous knockout mice were described as normal (20,21). We obtained Pitx1 knockout mice that had been maintained on an SV129 background and inbred for at least 10 generations (kindly provided by M. Geoffrey Rosenfled, USCD). However, when we bred Pitx1 knockout mice in our laboratory, clubfoot-like abnormalities were noted in some of the Pitx1+/− mice (Fig. 3A–C). These abnormalities were always noted in the first week of life, with forefoot cavus in the frontal plane and hindfoot equinus in the sagittal plane, as shown by micro-computed tomography (microCT) imaging (Fig. 3D and E). These mice were ambulatory, although they walked on the dorsolateral aspect of the affected feet. This clubfoot-like malformation was observed in 20 of 225 Pitx1+/− mice, resulting in a penetrance of 8.9% (Table 1), and was never observed in Pitx1+/+ littermates. Unlike clubfoot in humans, which occurs at a higher incidence in males (male-to-female ratio of 2:1) (22), we observe a higher incidence in Pitx1+/− female mice (female-to-male ratio of 2.5:1, P-value < 0.04).
Asymmetric right-sided hindlimb and pelvic abnormalities are a hallmark of the complete loss of Pitx1 expression in mice (20,21), stickleback fish (23) and manatee (24). We therefore hypothesized that Pitx1 haploinsufficiency would likewise result in preferential involvement of the right limb. The clubfoot-like malformation was unilateral in 16 of the 20 affected Pitx1+/− mice; however, the right and left limbs were equally affected. This high unilateral expressivity differed from our three patients with PITX1 deletions who all had bilateral clubfoot but was consistent with clubfoot seen with PITX1 E130K mutations and the general clubfoot population in whom ~50% of all cases are unilateral (25). Investigation into environmental factors such as maternal age and litter size did not reveal any association with increased risk of clubfoot-like malformation (Table 1).
Vascular abnormalities consisting of anterior tibial artery hypoplasia have been described in human patients with clubfoot (5,6,26,27). To characterize the vascular anatomy of the small vessels in the hindlimb of the Pitx1 haploinsufficient clubfoot-like mouse, we performed MRA on the affected and unaffected hindlimb of one Pitx1+/− mouse. In both humans and mice, the blood supply to the lower leg consists of three main arteries: anterior tibial, posterior tibial and peroneal. Although no differences between the posterior tibial or anterior tibial arteries were evident between the affected clubfoot and unaffected limbs, peroneal artery hypoplasia was apparent in the clubfoot limb (Fig. 4), suggesting that clubfoot in the Pitx1 genetic mouse model is associated with vascular deficiency.
Mild skeletal abnormalities of the long bones in the lower leg have been reported in some patients with human clubfoot (17,28). Yet previous reports based on non-quantitative observations indicated no morphological abnormalities in skeletal structures in Pitx1+/− hindlimbs (20,21). To examine bone structure in more detail, we used high-resolution microCT to measure the BVs of the entire tibia of both hindlimbs of adult wild-type Pitx1+/+ mice (n = 5), Pitx1+/− unaffected mice (n = 6) and Pitx1+/− unilateral clubfoot-affected mice (n = 7). Unilateral clubfoot mice were studied because this was the predominant phenotype, and because the unaffected limb could be used as an internal control. Tibial BVs were reduced in all Pitx1+/− mice compared with Pitx1+/+ wild-type littermates [Pitx1+/− unaffected mice (−26% ± 17%) and Pitx1+/− clubfoot-affected mice (−32% ± 20%) compared with wild-type Pitx1+/+ littermate controls] (Fig. 5A). This correlates with the overall growth impairment of Pitx1+/− mice (17.2 ± 0.4 g) compared with Pitx1+/+ mice (22 ± 1.4 g) that likely represents central effects of Pitx1 haploinsufficiency on pituitary function mediating general systemic growth (21,29). While left and right tibial volumes were uniformly equal and symmetrical in individual Pitx1+/+ wild-type and Pitx1+/− unaffected mice, the tibial volumes of unilateral Pitx1+/− affected clubfoot limbs were significantly and reproducibly reduced (−21% ± 8, P < 5 × 10−4) compared with the unaffected limb in every unilateral affected mouse (Fig. 5B). Clubfoot was always accompanied by a decrease in tibial BV, whereas no differences in BV between the right and left limbs were observed in the absence of clubfoot in unaffected Pitx1+/− or Pitx1+/+ wild-type mice. Fibular volumes were similarly reduced (data not shown). No obvious patterning defects were observed in any limbs from Pitx1+/−; all feet had the full complement of bones. Thus, clubfoot is associated with reduced tibial and fibular BVs in the Pitx1 genetic mouse model.
Histological evaluations of muscle biopsies revealed no significant morphological differences between the tibialis anterior muscle of Pitx1+/+ wild-type, Pitx1+/− unaffected and Pitx1+/− clubfoot limbs (data not shown). Skeletal muscle satellite cells were present in the affected and unaffected limbs at equal frequency when observed with PAX7 staining (data not shown). However, quantitative differences in muscle volumes were detected using MRI of the hindlimbs of Pitx1+/− mice with unilateral clubfoot (Fig. 6A). Total hindlimb volume was reduced (−21% ± 10, P < 0.002) and total muscle volume was reduced (−26% ± 11, P < 0.001) in the affected clubfoot limb compared with the unaffected limb (n = 7) (Fig. 6B). In contrast, total fat volumes were slightly increased in the affected limb. Specifically, the lateral muscle compartment containing the peroneus muscles appeared severely hypoplastic in all Pitx1+/− clubfoot limbs, and the tibialis anterior muscle was also small (Fig. 6A). A corresponding increase in fat tissue within the lateral compartment was noted in all clubfoot limbs.
To determine any effects of prolonged disuse, we also compared MRI volume measurements from a single young Pitx1+/− clubfoot mouse (22 days) to an older Pitx1+/− clubfoot mouse (540 days). The affected limb showed reduced total volume (−47 versus −12%) and muscle volume (−53 versus −24%) compared with the unaffected limb in young and old mice, respectively, although the difference was greater in the younger mice. The fat volume of affected limb was increased compared with the unaffected limb, more so in the younger mice (172 versus 128%, young versus old, respectively).
The morphological abnormalities associated with clubfoot in the Pitx1+/− haploinsufficient mouse model suggest that the clubfoot phenotype results from an alteration in early embryonic limb development affecting all tissues in the limb. To identify genes regulated by Pitx1 during early hindlimb development, we compared gene expression differences between hindlimb buds from E12.5 Pitx1+/+ and Pitx1−/− mice on the Illumina MouseRef-8 expression Bead Chip. We identified 31 genes that were differentially expressed ≥ 2-fold between Pitx1+/+ and Pitx1−/− mouse hindlimbs: 27 genes were down-regulated and 4 genes were up-regulated (Table 2). Genes involved in muscle development GeneGO processes were highly enriched in our data set, specifically in down-regulated genes (P < 1.79 × 10−28). These data are consistent with the reduced muscle volumes that were observed in the clubfoot-like limbs of both young and old mice.
By demonstrating that PITX1 haploinsufficiency causes clubfoot in both humans and mice, we now have strong evidence linking clubfoot etiology to abnormal early limb development. We previously identified a missense mutation in PITX1 in a single large family with clubfoot, although the mechanism by which this mutation caused disease was unclear. Dominant-negative transcriptional effects of the PITX1 E130K mutation were demonstrated in vivo (14), and the additional phenotypes (i.e. preaxial polydactyly) present in some individuals also suggested the possibility of a gain-of-function mutation. Our current results demonstrate that haploinsufficiency of this gene is sufficient for the development of clubfoot both in humans and in mice. The association of clubfoot with duplications or deletions involving the T-box transcription factor TBX4 (17), a known transcriptional target of PITX1 (15,16), in patients with recurrent chromosome 17q23 CNVs also suggests that hindlimb development is extremely sensitive to the dosage of transcription factors in this pathway.
Our data also provide compelling evidence to support the reevaluation of haploinsufficient mouse models of human disease that were previously reported to be normal. Earlier studies did not report hindlimb abnormalities in Pitx1+/− mice (20,21) probably because of low penetrance of the clubfoot phenotype. Furthermore, highly sensitive quantitative measurements that are now obtainable with microCT or MRI may detect differences that cannot be detected with standard non-quantitative methods. Interestingly, based on our quantitative microCT measurements, there did not appear to be a graded response with a ‘partial’ clubfoot phenotype. The tibial volumes of the Pitx1+/− unaffected mice, although statistically significantly smaller than the Pitx1+/+ mice, never varied significantly from right to left in the absence of clubfoot. This non-linear relationship between genotypic and phenotypic values is characteristic of dosage-sensitive genes and the stochastic processes that influence their action (30). Because the Pitx1 knockout mice had been inbred for at least 10 generations and were therefore near-to-completely genetically identical, this suggests that other stochastic events or environmental factors are responsible for generating the clubfoot phenotype in some mice. Because of the low penetrant clubfoot phenotype, the Pitx1 haploinsufficient mouse may be an excellent model to investigate the effects of environmental risk factors for clubfoot, such as maternal diabetes or smoking (31,32) on clubfoot susceptibility.
The Pitx1+/− mouse is remarkable for being the first genetic model of clubfoot. Despite the high incidence of clubfoot in humans (1), very few mouse models of clubfoot have been described. Most of the previously described rodent models of clubfoot have relied upon the induction of clubfoot-like deformities with amniotic sac puncture (33) or maternal exposure to retinoic acid (34–36), neither of which are likely to play a major role in human clubfoot etiology. Although a spontaneous recessive mouse model of ‘clubfoot’ was described more than 50 years ago (37), forefoot clubbing was also present making it a better model of distal arthrogryposis. More recently, the peroneal muscle atrophy (pma) mouse was proposed as a possible model of human clubfoot (38–40). The pma mouse resembles the Pitx1+/− mouse in that the peroneus and tibialis anterior muscles are hypoplastic. However, the pma mouse lacks the common peroneal nerve (41), whereas in the Pitx1+/− mouse this nerve is present (unpublished observation), suggesting a possible different pathophysiological mechanism that may become apparent when the causative gene for the pma mouse is discovered.
Some, but not all, features of Pitx1 haploinsufficient clubfoot mice are clinically similar to clubfoot in humans. The Pitx1+/− mice are smaller than their wild-type littermates, which is consistent with the short stature we observed in individuals with PITX1 deletions. These growth defects are likely due to pituitary haploinsufficiency of Pitx1 (21,29), as the mice were derived from a non-conditional knockout strategy that replicates the human abnormality. Although short stature has not typically been associated with human clubfoot, it has not been systematically studied. In contrast, the Pitx1+/− mice do not replicate the 2:1 expected male predominance of clubfoot seen in the general population (22). Although we previously reported a higher penetrance of clubfoot in our male patients with the PITX1 E130K mutation, we found a female bias with the Pitx1 deletion in mice, although larger numbers need to be studied to determine the significance of these findings. The mechanism responsible for the gender bias in clubfoot remains unknown.
We anticipated finding a right-sided bias in the clubfoot phenotype based on the numerous studies of evolutionarily distant vertebrates demonstrating that complete loss of Pitx1 resulted in more significant effects on the right (20,21,23–25), but we observed that right and left sides were equally involved. Because the right-sided bias has been ascribed to compensation by the related gene, Pitx2, that is asymmetrically expressed and also plays a role in limb development (42), our current findings suggest that the influence of Pitx2 is too small to be observed when Pitx1 is in the haploinsufficient state. Instead, the equal involvements of either the right or the left limb in the Pitx1+/− clubfoot mice are more suggestive of a stochastic process.
The Pitx1+/− mouse replicates many of the morphological abnormalities that have previously been described in some patients with clubfoot for which no unifying pathophysiological mechanism has been proposed. In the Pitx1+/− affected clubfoot-like limb, many tissues (e.g. muscle, vasculature and skeleton) appear to be involved. This is consistent with the expression of Pitx1 during the early limb-bud stage of development (15,20,21,43) and the known roles of Pitx genes in proliferation, cell survival and differentiation of multiple cell types (44–48). We hypothesize that Pitx1 haploinsufficiency disrupts the proliferation and differentiation of a mesenchymal cell population from which these tissues derive, leading to defects in multiple tissues. However, we cannot exclude the possibility that some morphological effects in the clubfoot-like affected mice are secondary to major effects on a single tissue (i.e. vascular abnormalities may result in the secondary loss of muscle or bone). Future experiments with conditional knockouts of Pitx1 in specific developmental lineages may allow us to investigate the mechanism by which defects in Pitx1 causes clubfoot.
Although it is possible that some morphological effects in the clubfoot-like mouse are secondary to disuse atrophy, our gene expression studies support a major primary effect of Pitx1 on muscle development. In the absence of Pitx1, expression of genes characteristic of differentiated muscle (e.g. myosin and actin) was reduced along with several genes that are essential for muscle differentiation (e.g. myogenin). These results are consistent with previous studies that have shown Pitx genes to be critical for limb myogenesis (49). Large differences in gene expression of vascular and bone developmental markers were not noted early in development, although the skeletal and vascular anomalies observed in mature Pitx1+/− clubfoot mice suggest that Pitx1 also regulates development of these tissues. These cells likely represent a smaller portion of the E12.5 hindlimb; therefore, gene expression studies of the entire limb bud may not be sensitive enough to detect expression changes in vascular and bone developmental markers. Interestingly, we only identified four genes that were up-regulated >2-fold in Pitx1−/− compared with the Pitx1+/+ hindlimb. One of these genes is the zinc finger protein (Zfp42), a marker of murine (50,51) and human (52) embryonic stem cells that are known to play a role in stem-cell differentiation (53). Taken together, the developmental anomalies observed in the clubfoot leg are consistent with Pitx1 acting as an important regulator of cellular proliferation and differentiation (44–48), perhaps by activation of Tbx4 expression (15,16).
The lower leg lateral muscle hypoplasia that we identified in the Pitx1+/− clubfoot-like mouse limb corresponds spatially with the peroneal artery hypoplasia, and suggests that Pitx1 haploinsufficiency leads to a developmental field defect preferentially involving the antero-lateral limb. Interestingly, the focality of these anatomic abnormalities may be explained by the higher expression of Pitx1 in the anterior limb bud at E11.5 (21). We did not find evidence of muscle patterning defects in Pitx1−/+ mice on MR imaging, although these have been described in mice with complete absence of Pitx1 (15), and our methods may not have been optimal to identify these differences. As a human correlate, previous studies found reduced muscle volume in the affected clubfoot limb (7), and here we extend these studies to demonstrate more significant loss of muscle in the anterior muscle compartment in the clubfoot of a patient with a PITX1 E130K mutation. Biomechanically, the development of a clubfoot deformity may therefore be explained by a muscle imbalance caused by hypoplasia of these muscle compartments with relative overpower of the remaining muscle groups. However, we do not yet know the frequency with which focal-muscle hypoplasia occurs in clubfoot, whether a phenotypic correlate exists with known mutations in the PITX1-TBX4 pathway, or whether these structural defects are associated with treatment refractory clubfoot.
Thus, our identification of PITX1 haploinsufficiency in humans and mice with clubfoot provides additional strong support for the PITX1-TBX4 developmental pathway in clubfoot pathogenesis. The morphological abnormalities identified in the Pitx1+/− clubfoot mouse suggest that Pitx1 haploinsufficiency leads to a developmental field defect preferentially involving the lateral lower limb, and adds genetic data to previous work implicating regional growth disturbances in human clubfoot etiology (7,28,54). Future studies will determine whether compensation for the focal-muscle hypoplasia can occur by stimulating the growth of satellite cells during the post-natal period. As the first genetic model of clubfoot, the Pitx1+/− mouse represents a significant advance for the study of clubfoot, and provides a unique tool to generate and test novel hypotheses of clubfoot pathogenesis.
We identified 40 isolated idiopathic clubfoot probands with at least one affected first-degree relative. Individuals were considered to have isolated clubfoot only in the absence of additional congenital anomalies or known underlying etiology (e.g. arthrogryposis, myelomeningocele and myopathy). The study protocol was approved by the Institutional Review Board and all subjects and/or their parents gave informed consent. Blood and saliva samples were collected from the affected individuals and family members. DNA was extracted following the manufacturer's protocol using either the Roche DNA Isolation Kit for Mammalian Blood (Roche) or the Oragene Purifier for saliva (DNA Genoteck). The human MRI/MRA was performed as a research study after obtaining informed consent using a previously described imaging protocol (26).
Forty familial isolated clubfoot probands were screened for genomic CNVs on the Genome-wide Human SNP Array 6.0 (Affymetrix). Genomic copy number calls were determined compared with a set of 270 normal HapMap reference controls with the Canary algorithm using the Genotyping Console software (Affymetrix). Segregation was confirmed by qPCR.
Pitx1+/− mice were provided by M. Geoffrey Rosenfeld, UCSD (21) and maintained on a SV129 background by heterozygous mating. Newborn mice were genotyped by PCR using DNA isolated from tails (Neo-F: AGGATCTCCTGTCATCTCACCTTGCTCCTG and Neo-R: AAGAACTCGTCAAGAAGGCGATAGAAGGCG, PITX1-F: CCGGCTACTCCTACAACAACTGG and PITX1-R: CTGTTGTACTGGCAAGCGTTGAG). Affected status was determined within 7 days of birth and defined as fixed internal rotation of the hindlimb.
MRI/MRA were performed in the Biomedical MR Laboratory using a 4.7 T Varian/Agilent DirectDrive small-animal MR scanner. Mice were anesthetized with isoflurane (2% for induction, 1% for maintenance). A purpose-built Helmholtz coil was used for excitation and signal reception. In muscle/fat volume studies, water- and lipid-selective images were acquired with a spin-echo sequence preceded by an optimized binomial frequency-selective excitation plus gradient spoiling to eliminate either the lipid or water signal. Image data were acquired from 36 contiguous, 400 μm thick slices with typical in-plane resolution of 70–90 μm.
For MRA studies of mouse leg, a water-selective 3D gradient echo imaging sequence was employed with a 20 mm × 10 mm × 10 mm (64 × 64 × 64) field of view. A flip angle of 90° was used for excitation with TR = 200 ms, TE = 2.36 ms and two signal averages. A maximum intensity projection was used to display the angiography.
The arterial patterns were subjectively compared between limbs, using the contralateral normal limb as a control. The presence or absence of the anterior tibial, posterior tibial, peroneal and dorsalis pedis arteries was assessed for each limb. The arteries were considered normal when the anterior and posterior tibial and peroneal arteries were of equal caliber. An artery was considered abnormal if it was undetectable (i.e. absent) or small (i.e. hypoplastic) compared with the opposite limb.
Muscle, fat and total limb volume measurements were quantified using the semi-automated image software ImageJ 1.44 (55). Primary volume measurements including bone, adipose tissue and all muscle compartments were quantified across 18 consecutive slices. BVs were isolated and quantified using semi-automated brightness thresholds. Cutaneous and subcutaneous adipose tissue was quantified from corresponding fat-selective images. Muscle volumes were calculated by fat and bone subtraction from total volumes for each limb and P-values were generated using a one class T-test.
To evaluate skeletal structures, 3D microCT scanning was performed (VivaCT 40, Scanco Medical; 45 kVp, 177uA, 0.031 mm isometric voxel resolution, 100 ms integration time). Affected mice were scanned between 3 and 8 months of age. Mice were anesthetized with isoflurane (2–3% inhalation), and placed in the imaging-positioning tray in a prone position with the limb of interest fully extended. Full scans of each leg were obtained and analyzed using the Image Processing Language (IPL) software provided by the manufacturer. Contours were drawn manually to define the outer surface of the tibia, and the volume of interest was defined as the entire tibia (~15–16 mm length). Segmentation was done using consistent settings (sigma = 1.2; support = 2; threshold density = 475 mg HA/ccm). Tibial BV was then computed; this represents the size of the entire tibia.
Hindlimb buds were collected from Pitx1+/− mated mice at embryonic day E12.5 and genotypes were determined by PCR of DNA isolated from tails. Age was determined according to the appearance of the vaginal plug (day 0) and confirmed by morphological criteria. Hindlimb buds from two embryos (four hindlimb buds) were combined for each biological sample with three biological samples per genotype. Total RNA was extracted using Trizol (Invitrogen) and further purified using RNeasy (Qiagen). RNA quality was determined with the Agilent Bioanalyzer. Microarray hybridizations were performed by the Washington University Microarray Core Facility using the MouseRef-8 expression Bead Chip (Illumina). Expression values were quantile normalized and mean expression values for three biological samples were compared using a Student's t-test. Differentially expressed genes are defined as ≥ 2-fold change in expression and P-value < 0.05. All microarray data for the present study have been deposited with NCBI GEO with accession number GSE27363.
This work was supported by grants from Shriners Hospital for Children, the National Institute of Health (K12 HD001459), The Children's Discovery Institute, The March of Dimes Basil O'Connor Starter Scholar Research Award, St Louis Children's Hospital Foundation, The Orthopaedic Research and Education Foundation, Pediatric Orthopaedic Society of North America, and the Mallinckrodt Institute of Radiology. The microCT analysis was performed at the Washington University Center for Musculoskeletal Research, supported by the National Institute of Arthritis, Musculoskeletal and Skin Diseases (NIH P30AR057235).
We kindly thank M. Geoffrey Rosenfeld for providing the Pitx1+/− mice, and Seth Crosby and Mike Heinz at the Washington University Genome Center for processing the Affymetrix microarray. We also thank Alan Pestronk for histological analysis of mouse muscle, and Susan Mackinnon for quantitative analysis of peripheral nerve. The microarray data have been deposited into the Gene Expression Omnibus (GEO) database under accession number GSE27363.
Conflict of Interest statement. None declared.