Podosomal proteins as causes of human syndromes: A role in craniofacial development?

Abstract

Podosomes and invadopodia are actin-rich protrusions of the plasma membrane important for matrix degradation and cell migration. Most of the information in this field has been obtained in cancer cells, where the presence of invadopodia has been related to increased invasiveness and metastatic potential. The importance of the related podosome structure in other pathological or physiological processes that require cell invasion is relatively unexplored. Recent evidence indicates that essential components of podosomes are responsible for several human syndromes, some of which are characterized by serious developmental defects involving the craniofacial area, skeleton and heart, and very poor prognosis. Here we will review them and discuss the possible role of podosomes as a player in correct embryo development. genesis 49:209–221, 2011. © 2011 Wiley-Liss, Inc.

INTRODUCTION

Alterations in craniofacial development are the most common cause of birth defects, estimated to be three fourths of all congenital malformations in the United States (Chai and Maxson,2006). Although environmental factors can underlie some craniofacial anomalies, most cases appear in the context of hereditary syndromes caused by mutations in genes crucial for development. Head morphogenesis is an intricate process that involves all three germ layers and the neural crest cells, an evolutionary acquisition of vertebrates that gives rise to the viscerocranium, odontoblasts, and part of the neurocranium. Correct head development depends on precise spatiotemporal interactions among all these structures and their derivatives, and despite the efforts of recent years, the detailed picture of all the mechanisms and pathways involved is still far from complete. A cellular process present in several phases of craniofacial morphogenesis is the epithelial-mesenchymal transition (EMT), where selected populations of cells detach from the substrate and their neighboring cells, lose polarity and become migratory. EMT has also been found to drive pathological situations in adults such as organ fibrosis and cancer progression and metastasis, and seems to be regulated by the same molecular pathways present during embryonic development.

Podosomes and invadopodia are actin-rich protrusions of the cell membrane important for extracellular matrix degradation, adhesion to substrate, and migration. They were first described in 1980 in Src-transformed 3T3 cells (David-Pfeuty and Singer,1980), and since then they have gained increasing interest because their cellular functions make them attractive candidates to be part of numerous biological processes where matrix remodeling and cell motility are required. Despite the lack of specific markers that has impaired the study of podosomes in vivo, we know that podosomes are present in migratory cell types including macrophages, dendritic cells, endothelial cells and vascular smooth muscle cells (Saltel et al.,2010). Some cancer cells have podosome-related protrusions known as invadopodia, which have also been studied in vitro. Their presence is increased in invasive tumor cells, and they are believed to drive metastasis. Recently, mutations in podosomal proteins have been identified as the cause of human inherited diseases. Remarkably, some of these mutations markedly affect the correct development of the craniofacial structures in a similar fashion. The purpose of the present review is to describe these conditions and discuss the possible involvement of podosomes in craniofacial morphogenesis.

PODOSOMES AND INVADOPODIA: A BRIEF OVERVIEW

Podosomes and their related structures invadopodia are dynamic, actin-rich protrusions of the plasma membrane, located on the ventral surface of cultured cells, and which form close contact with the substrate. They were first described in the 1980s in v-Src-transformed fibroblasts (Chen,1989; David-Pfeuty and Singer,1980; Marchisio et al.,1984; Tarone et al.,1985). While some investigators focused on the invadopodia of cancer cells, podosomes in normal cells remained relatively unexplored until about a decade ago, when the discovery that Wiskott-Aldrich syndrome protein (WASp), mutation of which is responsible for the Wiskott-Aldrich syndrome, is an essential component of podosomes (Linder et al.,1999) was made.

The most accepted hypothesis is that the terms “podosomes” and “invadopodia” define a similar cellular structure in different cell types. The term podosome isused for nontransformed cells (and occasionally Src-transformed fibroblasts), whereas the term invadopodia is used to define the structures found in cancer cells. Currently there is a lot of effort put into defining the exact molecular and dynamic similarities and differences between podosomes and invadopodia. Briefly, podosomes appear to be more dynamic than invadopodia, with a turnover time of 2-12 min versus several hours for invadopodia. Podosomes often appear as numerous small (1–2 μm × 0.2–0.4 μm) dots organized in rings—rosettes—or distributed at the leading edge of cells. In contrast, invadopodia are less numerous and irregularly shaped, and in culture appear more aggressive in their matrix-degrading tasks (Caldieri et al.,2009; Linder,2007). The recent creation of the Invadosome Consortium (www.invadosomes.org) will help to unify current concepts and encourage research aimed at elucidating the distinct mechanisms underlying podosome and invadopodia biology.

Both podosomes and invadopodia have a characteristic structure composed of a core of actin and actin-binding proteins surrounded by a ring rich in integrins and integrin-associated proteins. Proteins in podosomes/invadopodia seem to be specific to either the ring or the core (Caldieri et al.,2009; Gimona et al.,2008; Linder,2007; Linder and Kopp,2005; Saltel et al.,2010). Proteins necessary for podosome/invadopodia formation include F-actin and multiple actin interacting proteins: nucleators, polymerization activators, crosslinkers, and binders. Also, a number of kinases, such as Src and FAK, and small GTPases are essential for their formation. Different integrins and adhesion molecules such as CD44 can form part of the ring structure in different cell types. Adaptor proteins that provide scaffolding for other components of the podosome machinery are also necessary. In this regard, Tks5 is the only component of podosome/invadopodia that seems to be exclusive to these structures. Tks5 was identified as a novel Src-substrate and contains a PX domain and 5 SH3 domains (Lock et al.,1998). Tks5 is required for podosome and invadopodia formation, and its expression is elevated in Src-transformed cells, invasive human cancer cells and tumor tissue, suggesting a link with invasiveness (Seals et al.,2005). Podosomes/invadopodia are also rich in proteases, since one of their best known functions is to degrade the extracellular matrix. Other functions include adhesion to substrate and migration, and microenvironment sensing (Gimona et al.,2008; Saltel et al.,2010; Spinardi and Marchisio,2006).

The biogenesis of podosomes and invadopodia seems to be triggered by adhesion to specific substrates, stimulation by certain molecules, or transformation. Podosomes have not only been found in cells derived from the bone marrow: macrophages (Lehto et al.,1982), lymphocytes (Carman et al.,2007), dendritic cells (Binks et al.,1998) and osteoclasts (Marchisio et al.,1984), but also in endothelial cells (Moreau et al.,2003), myoblasts (Thompson et al.,2008) and vascular smooth muscle cells (Gimona et al.,2003). Invadopodia have been found in a number of cancer cells (Seals et al.,2005), and tyrosine kinase transformed fibroblasts (David-Pfeuty and Singer,1980; Mueller et al.,1992), where they were first described. In transformed cells, the presence of invadopodia correlates with metastatic potential.

MUTATION OF PODOSOMAL PROTEINS AS A CAUSE OF HUMAN GENETIC DISEASES

The study of podosomes is a young field and most of the information we have about their biogenesis has been obtained in cell culture models. However, there are a few hints that suggest these structures may also be important regulators of correct physiology. In 1991, it was demonstrated in mice that deficiency of c-src causes osteopetrosis due to the inability of osteoclasts to do proper bone resorption (Soriano et al.,1991). When analyzed in vitro, c-src null osteoclasts were not able to form podosomes, which are required for adherence to bone, ruffled border formation, and subsequent bone remodeling (Lowe et al.,1993). The first direct link between podosomes and human disease was established in 1999, when WASp, a hematopoietic, actin-binding protein mutation of which is responsible for the Wiskott-Aldrich syndrome, was found to regulate podosome assembly in human macrophages (Linder et al.,1999). Since then, the study of podosome biology has expanded and many other proteins have been described to localize and/or be essential for their formation (Caldieri et al.,2009; Linder and Kopp,2005). Mutations in some of these proteins are the cause of genetic human diseases (Table 1) of variable seriousness. Interestingly, most of them present maldevelopment of the same body structures (Table 2) and, most remarkably, affect the correct development of craniofacial structures in a similar manner (Table 3). Below we describe the traits of these diseases, their genetic causes, and the possible links with podosome/invadopodia biology.

Table 1. Human Hereditary Syndromes Associated With Mutations in Podosomal Proteins

Syndrome	MIM #	Gene	Protein	Localization/Function in podosomes/invadopodia	References
Wiskott-Aldrich	300392	WAS	WASp	Actin-binding protein. After activation by Cdc42, promotes actin polymerization by the Arp2/3 complex. Essential for leukocyte podosomal formation.	(Calle et al.,2008; Gimona et al.,2008)
Focal segmental glomerulosclerosis 1	603278	ACTN4	αActinin4	Actin-binding and crosslinking protein. Incorporates to nascent podosomes in A7r5 cells after phorbol ester stimulation. Found also in invadopodia of pancreatic ductal adenocarcinoma cells	(Aga et al.,2008; Fultz et al.,2000; Gimona et al.,2003; Mueller et al.,1992; Welsch et al.,2009)
Frank-Ter Haar	249420	SH3PXD2B	TKS4	Adaptor protein essential for mature podosome formation.	(Buschman et al.,2009)
X-linked myxomatous valvular dystrophy	314400	FLNA	Filamin A	Cytoskeletal protein that crosslinks actin into networks or stress fibers. Found in the podosome belt of osteoclasts and invadopodia of oral carcinoma cells.	(Marzia et al.,2006; Takkunen et al.,2010)
Periventricular nodular heterotopia	300049
Melnick-Needles	309350
Otopalatodigital type 1	311300
Otopalatodigital type 2	304120
Frontometaphyseal dysplasia	305620
Faciogenital dysplasia/ Aarskog-Scott	305400	FGD1	FGD1	Guanine nucleotide exchange factor specific for Cdc42. Required for invadopodia formation and function by activation of Cdc42.	(Ayala et al.,2009)

Table 2. Main Phenotypes Found in Syndromes Associated With Mutations in Podosomal Proteins

	Syndromes
	WAS	FSGS1	FTHS	PVNH	XMVD	OPD1	FMD	OPD2	MNS	FGDY
Growth Retardation	−	−	+	−	−	+	−	+	+	+
Immunodeficiency	+	−	−	−	−	−	−	−	−	−
Lethality	+	−	+	−	−	−	+	+	+	−
Skeletal dysplasia	−	−	+	−	−	+	+	+	+	+
Genital malformations	−	−	−	−	−	−	+	−	+	+
Renal/Urological malformations	+	+	−	−	−	−	+	+	+	+
Intestinal malformations	−	−	−	+	−	−	−	+	+	−
Respiratory malformations	−	−	−	+	−	−	+	+	+	−
Cardiovascular malformations	−	−	+	+	+	−	+	+	+	−
Mental retardation	−	−	−	−	−	−	−	+	−	−
Neurological defects	−	−	−	+	−	−	−	−	−	−
Craniofacial defects	−	−	+	−	−	+	+	+	+	+

Table 3. Craniofacial Features of Syndromes Associated With Mutations in Podosomal Proteins

	Syndromes
	FTHS	OPD1	FMD	OPD2	MNS	FGDY
Large anterior fontanel	+	−	+	−	+	−
Widow's peak	−	−	−	−	−	+
Exophthalmos	−	−	−	−	+	−
Megalocornea	+	−	−	−	+	−
Glaucoma	+	−	−	−	−	−
High/prominent forehead	+	+	+	+	+	−
Hypertelorism	+	+	+	+	+	+
Cleft palate	−	+	−	+	+	−
Mandibular hypoplasia	+	−	−	+	+	+
Broad nasal bridge	+	+	−	+	−	+
Anteverted nostrils	+	−	−	−	−	+
Dentition anomalies	+	+	+	+	+	+
Downward palpebral fissures	−	+	−	+	+	+
Deafness	−	+	+	+	−	−
Low-set ears	+	−	+	+	+	−
Broad mouth	+	−	−	−	−	−
Small mouth	−	+	−	+	+	−

Wiskott-Aldrich syndrome

Wiskott-Aldrich syndrome (WAS) is an immunodeficiency characterized by eczema, thrombocytopenia and recurrent infections. It was first described by Wiskott in1937 (Wiskott,1937), and later by Aldrich (Aldrich et al.,1954). WAS is an X-linked recessive syndrome with poor prognosis, with the average age at death being 8 years, although bone marrow transplantation has improved the life expectancy (Sullivan,1999; Sullivan et al.,1994). Causes of death include infections and bleeding, and some of the patients develop autoimmune disorders and malignancies.

The genetic cause of WAS was identified in 1994 (Derry et al.,1994) by creation of a clone contig of the region X11.23-11.22, critical for WAS. Analysis of the different candidate cDNAs led to the identification of one with expression limited to the lymphocytic and megakaryocytic cell lineages and altered in WAS patients. The gene was named WASP (Wiskott-Aldrich syndrome protein), and it encodes an actin-binding protein of 502 aa. WASp belongs to the WASp family of proteins which also contains N-WASp, ubiquitously expressed, and WAVE, of which 3 isoforms exist (Takenawa and Miki,2001). WASp and its closest homolog N-WASp have a WASp homology domain at the N-terminal end that allows binding to the WASp interacting protein (WIP), followed by a basic sequence that binds phosphatidylinositol 4,5-biphosphate and a CRIB domain for binding to the activated form of Cdc42. The rest of the protein is similar to the other members of the family and contains a central proline-rich region, followed by an Arp2/3 binding-activating module (see Fig. 1) (Calle et al.,2004a). WASp is an essential component of podosomes in cells of hematopoietic origin. Macrophages and dendritic cells from WAS patients completely fail to form podosomes (Burns et al.,2001; Linder et al.,1999). Deficiency of WASp in mice also results in impaired hematological functions linked to altered podosome function. Calle et al. described how WASp null osteoclasts are unable to form podosomes, and therefore actin rings at the sealing zone, resulting in defects in bone resorption (Calle et al.,2004b). Reexpression of WASp in murine WASp null myeloid cells restores podosome formation (Blundell et al.,2008; Calle et al.,2004b). In macrophages from WAS patients, podosome formation can be rescued by injection of the C-terminal acidic part of WASP (Linder et al.,2000). A model has been proposed for the role of WASp in podosome formation and dynamics, in which WASp would be recruited to focal adhesions in response to chemotactic factors, where it would promote actin polymerization at the podosome core and provide a scaffold between integrins at the ring and the forming actin filaments (Monypenny et al.,2010). To promote podosome initiation, WASp is required to bind to Cdc42 (Dovas et al.,2009), and for disassembly of podosomes, WASp needs to be cleaved by calpain (Calle et al.,2006). In nonhematopoietic cells, WASp is substituted by N-WASp during the podosome core formation (Calle et al.,2008).

Figure 1. Protein structure and location of mutations causative of the syndromes discussed in this review. Arrows and lines indicate the domains where mutations have been found. In the case of Filamin A syndromes where specific mutations have been found are also indicated. PH indicates pleckstrin homology; WH1, WAS homology 1; GBD, GTPase binding domain; PRD, proline rich domain; VD, verprolin homology domain; CD, cofilin homology domain; ABD, actin-binding domain; SP, spectrin repeat; CAM-LD, calmodulin-like domain; PX, Phox-Homology domain; SH3, Src homology³ domain; CP, calponin homology domain; FL, filamin repeat; DH, Dbl homology domain; FYVE, Zinc-finger FYVE domain.

Focal Segmental Glomerulosclerosis 1

Focal segmental glomerulosclerosis 1 (FSGS1) is an inherited renal disease caused by mutations in the ACTN4 gene, encoding the actin bundling and crosslinking protein α-Actinin 4 (Kaplan et al.,2000). α-Actinin 4 is one of four isoforms belonging to the actin-binding superfamily of spectrins. α-Actinin 2 and 3 are only expressed in muscle, while α-Actinin 1 and 4 are widely expressed, and are usually associated with stress fibers and focal contacts (Edlund et al.,2001; Sjoblom et al.,2008). α-Actinin 4 is the only isoform expressed in human kidney (Kaplan et al.,2000). α-Actinins act as antiparallel dimers, and their structure consists of an actin-binding domain at the N-terminus, a central rod formed by four spectrin domains and a calmodulin-like domain at the C-terminus (see Fig. 1) (Davison and Critchley,1988; Trave et al.,1995).

FSGS1 segregates in an autosomal dominant manner, and is characterized by proteinuria starting in adolescence or later, and progressive glomerular sclerosis that can end in kidney failure. Analysis of several affected families with disease mapping to a locus in chromosome 19q13 allowed the identification of mutations in ACTN4 as the cause of FSGS1 (Kaplan et al.,2000). All mutations found affect highly conserved residues located between the actin-binding domain and the first spectrin domain of α-Actinin 4, and give rise to mutant forms with increased binding affinity to actin (see Fig. 1) (Kaplan et al.,2000). Later studies in cell models and mutant mice suggest that mutations found in FSGS1 patients would result in both loss-of-function and gain-of-function mutant α-Actinin 4 forms. Loss-of-function mutations create mutant forms of α-Actinin 4 that form cellular aggregates, are less dynamic and have a higher degradation rate than the wild type protein (Yao et al.,2004). In contrast, gain-of-function mutations create α-Actinin 4 forms with increased affinity for F-actin (Michaud et al.,2003). Localization of α-Actinin to podosomes/invadopodia has been described in several reports, but only one specifically states α-Actinin 4 as the object of analysis. α-Actinin was first described to localize to invadopodia of v-Src transformed chicken embryonic fibroblasts (Mueller et al.,1992). Later it was shown to incorporate into nascent podosomes in the vascular smooth muscle cell line A7r5 after stimulation with phorbol 12,13-dibutyrate (Fultz et al.,2000; Gimona et al.,2003), and in podosome- or invadopodia-like structures (PILS) of porcine trabecular meshwork cells (Aga et al.,2008). α-Actinin 4 has also been found in invadopodia of pancreatic ductal adenocarcinoma cells (Welsch et al.,2009). Despite these reports, the importance of α-Actinin in podosome/invadopodia biology is unknown.

Frank-Ter Haar syndrome

Frank-Ter Haar syndrome (FTHS) is a rare, autosomal recessive syndrome affecting the skeletal and craniofacial areas and the heart. FTHS was first described by Frank et al. in1973 (Frank et al.,1973), and ter Haar et al. (ter Haar et al.,1982) as two separate entities. The craniofacial and skeletal abnormalities seen in FTHS patients resemble those of the Melnick-Needles syndrome, described below, and led to the hypothesis that FTHS was a form of it (ter Haar et al.,1982). Identification of further cases subsequently allowed the recognition of FTHS as a separate entity (Hamel et al.,1995; Maas et al.,2004).

FTHS patients have a wide array of anomalies, the most distinctive being megalocornea, congenital glaucoma, and prominent coccyx. The craniofacial malformations are characterized by brachycephaly, large anterior fontanels, wormian bones, micrognathia, hypertelorism, prominent forehead, anteverted nostrils, dentition abnormalities, high palate, and the aforementioned megalocornea and glaucoma. In the skeletal compartment, FTHS patients present thoracolumbar kyphosis, prominent coccyx, pectus excavatum, bowing of long bones, short hands, flexion deformity of fingers, and club feet. The cardiac defects include mitral valve anomalies, double right outlet, and ventricular septal defects (Iqbal et al.,2010; Maas et al.,2004; Megarbane et al.,1997; Wallerstein et al.,1997).

Recently, mutations in the gene SH3PXD2B were identified as the genetic culprit of FTHS. Iqbal et al. performed homozygosity mapping on patients from 12 FTHS families, identifying a locus on chromosome 5q35.1 for which patients from nine families shared homozygosity. Detailed analysis of the sequences led to the observation that a homozygous deletion present in one of the families mapped to the smallest region of overlapping homozygosity, and contained a single gene: SH3PXD2B (Iqbal et al.,2010). This gene encodes the adaptor protein Tks4, named by us as a consequence of its high homology to Tks5 (see Fig. 1) (Buschman et al.,2009; Courtneidge et al.,2005). We cloned Tks4 and studied its cellular localization and function. In Src-transformed 3T3 fibroblasts, Tks4 mainly localizes to rosettes of podosomes/invadopodia. Knock-down of Tks4 in these cells and murine embryonic fibroblasts abrogates the formation of mature and functional podosomes, since they fail to degrade the extracellular matrix. Introduction of Tks5 only rescues podosome/invadopodia formation but not function, while reintroduction of Tks4 rescues both phenotypes, indicating the essential and independent role of Tks4 in podosome/invadopodia biology (Buschman et al.,2009).

Analysis of Tks4 expression in dermal fibroblasts of FTHS patients demonstrated the absence of this protein where mutations in SH3PXD2B gene were identified, in all patients where samples were available. Interestingly, we also found reduced Tks4 expression in a FTHS patient with no identified mutations in the SH3PXD2B coding sequence, suggesting that other proteins regulating the stability of Tks4 may be involved in the pathogenesis of FTHS. Confirmation of Tks4 as the cause of FTHS was obtained by analysis of the sh3pxd2b null mice, which display most of the craniofacial, ocular, skeletal, and cardiac phenotypes observed in the patients (Iqbal et al.,2010). Of note, we observed a marked decrease of white adipose tissue in the sh3pxd2b null mice that is not reflected in the FTHS patients, although there is one case in literature reported to have “little or no subcutaneous fat” (Wallerstein et al.,1997). In support of a role of Tks4 in adipogenesis, Hishida et al. found Tks4 (called fad49 in this study) is necessary for the early phase of adipogenic differentiation of 3T3-L1 cells (Hishida et al.,2008).

Periventricular Nodular Heterotopia (PVNH), X-Linked Myxomatous Valvular Dystrophy (XMVD), and the Otopalatodigital Spectrum Syndromes: Otopalatodigital Syndrome-1 (OPD1), Frontometaphyseal Dysplasia (FMD), Otopalatodigital Syndrome-2 (OPD2), and Melnick-Needles Syndrome (MNS)

These six syndromes are all X-linked dominant conditions caused by mutations in the FLNA gene, which encodes Filamin A, a 280 kDa cytoskeletal protein that cross-links actin into tridimensional networks or stress-fibers in response to extracellular signals (Cunningham et al.,1992; Hartwig and Stossel,1975; Niederman et al.,1983; Tu et al.,2003). Filamin A contains an actin-binding domain located at the N-terminus, consisting of 2 calponin homology domains, and 24 filamin repeats interrupted by two hinge regions, the last one allowing homodimerization (see Fig. 1) (Robertson,2005). Filamin A can bind a vast array of proteins (Zhou et al.,2007), that modulate its multiple cellular functions. The most recognized is its ability to reorganize the actin cytoskeleton by binding to the β-chain of integrins (Zhou et al.,2010), which allows Filamin A to control cell polarization and migration in some cell types (Calderwood et al.,2001; Cunningham et al.,1992; Hart et al.,2006). Filamin A can interact with small GTPases and their effectors (Zhou et al.,2010), and modulate gene transcription (Berry et al.,2005; Sasaki et al.,2001). Filamin A has been found in the podosome belt of osteoclasts (Marzia et al.,2006) and in invadopodia of invasive and non-invasive oral squamous carcinoma cell lines (Takkunen et al.,2010). The exact role of Filamin A in invadopodia and podosome formation or function has not been explored yet.

The 6 hereditary diseases caused by mutations in the FLNA gene are very different in their clinical manifestations. PVNH is a neuronal migration disorder caused by mutations of the FLNA gene that cause loss-of-function of the protein (Fox et al.,1998). The most common manifestation of PVNH is seizures that usually start during adolescence. Other complications affect the vascular system, and include patent ductus arteriosus and coagulopathy. PVNH is lethal in males, who usually die early in gestation (Eksioglu et al.,1996), indicating the importance of Filamin A in embryogenesis. Analysis of mice deficient for Filamin A supports this finding. Male mice null for Filamin A die at E14.5 and present hemorrhage and edema. Analysis of the embryos revealed abnormal vascular patterning and severe cardiac defects involving atria, ventricles and the outflow tract. Surprisingly, migration of mutant neural crest cells in vivo and murine embryonic fibroblasts in vitro was not impaired, and the null brains appeared to develop normally (Feng et al.,2006). In a different loss-of-function mutant FLNA mouse model, unfused sternum and cleft palate were described in addition to the other phenotypes (Hart et al.,2006). In keeping with the effects of Filamin A on heart development, mutations in FLNA have recently been identified as the cause of XMVD (Kyndt et al.,2007). Myxomatous dystrophies are characterized by excessive valve tissue that leads to swelling of the valve leaflets, with or without valve prolapse and regurgitation, the most typical form being mitral valve prolapse (Levy and Savage,1987). XMVD has variable severity among carriers, and has complete penetrance in men and incomplete penetrance in women.

The otopalatodigital spectrum syndromes—OPD1, FMD, OPD2, and MNS—are caused by mutations in the FLNA gene that create gain-of-function changes without disturbing the length of the protein (Clark et al.,2009; Robertson,2007; Robertson et al.,2003). These four syndromes cause similar anomalies—craniofacial, skeletal, heart, genitourinary, and intestinal tracts dysplasias—but differ in their severity, being OPD1 the mildest and MNS the most serious of the spectrum (Tables 2 and 3). OPD1 is characterized by cleft palate, occipital, frontal, and supraorbital prominence, flat nasal root, hypertelorism, dental anomalies, and small mouth. Skeletal anomalies include short stature, digital, and sternum deformities, mild bowing of long bones and deafness due to malformations of the auditory ossicles (Dudding et al.,1967; Gall et al.,1972; Hidalgo-Bravo et al.,2005). Females present a milder form of the phenotypes than males, although occasionally can be as affected (Gorlin et al.,1973; Langer,1967). FMD patients have generalized skeletal dysplasia, characterized by bowing and flaring of bones, scoliosis, joint contractures, arachnodactyly, and deafness. They also present asthenia and urogenital defects (Fitzsimmons et al.,1982; Glass and Rosenbaum,1995; Gorlin and Cohen,1969; Kanemura et al.,1979; Morava et al.,2003). FMD craniofacial defects are characterized by increased bone density in the supraorbital area and jaw, low-set ears, bushy eyebrows, hypertelorism, dentition anomalies, and delayed fontanel closure (Fitzsimmons et al.,1982; Gorlin and Winter,1980; Morava et al.,2003). OPD2 affected males have disabling skeletal anomalies and malformations in the hindbrain, heart, intestines, and kidneys that frequently lead to death within the first year of life (Verloes et al.,2000). Craniofacial defects of OPD2 patients include large anterior fontanels, hypertelorism, prominent forehead, downslanting palpebral fissures, broad nasal bridge, low-set ears, cleft palate, small mouth, and micrognathia (Fitch et al.,1983; Fitch et al.,1976; Verloes et al.,2000). OPD2 affected females present a mild form of the craniofacial traits, but do not usually have the skeletal anomalies, and carriers can be asymptomatic (Ogata et al.,1990; Preis et al.,1994). MNS, the most serious of all four, causes malformations as severe as the ones seen in OPD2 patients, typically causing male death in utero or during the first months of life. (Donnenfeld et al.,1987; Melnick and Needles,1966; Santos et al.,2010). Some adult female patients have been reported to die from respiratory failure (Robertson,2007), but most have a normal lifespan. Craniofacial traits of the MNS comprise exophthalmos, high forehead, full cheeks, large anterior fontanel, micrognathia, and malalignment of teeth. Skeletal dysplasia is characterized by bowing of long bones with flared metaphyses, deformed clavicles, ribs, scapula, sclerosis in the base of the skull, contracted pelvis, long fingers and short stature. Analysis of affected male infants and fetuses added omphalocele, prune belly sequence, kidney hypoplasia, heart defects (tetralogy of Fallot, atrioventricular canal defect, and atrial septum malformation) and intestinal malrotation to the other traits. MNS heterozygotes present skeletal dysplasia. Curiously, some of the first FTHS patients described (ter Haar et al.,1982) were mistakenly thought to have an autosomal recessive form of MNS, due to the similar craniofacial and skeletal features present in both syndromes.

FLNA mutations associated with the OPD syndromes have been located to specific domains of Filamin A (see Fig. 1) (Robertson,2007; Robertson et al.,2003): OPD1 mutations have only been found in the second calponin homology domain of the actin-binding region, MNS mutations only in filamin repeat 10, OPD2 in the second calponin homology domain and filamin repeats 14 and 15, and FMD, in the second calponin homology domain, and repeats 3, 9, 10, 14, 15, 22, and 23.

Faciogenital dysplasia/Aarskog-Scott syndrome

Faciogenital dysplasia (FGDY), or Aarskog-Scott syndrome, is a non-lethal, x-Linked recessive syndrome first described in the early 1970s (Aarskog,1970; Scott,1971), which affects mainly males, although female carriers present a mild form of the phenotypes. Manifestations of the FGDY syndrome comprise facial, skeletal and urogenital anomalies, the most characteristic being the disproportionate short stature with shortened distal extremities. Facial manifestations include hypertelorism, maxillary and mandibular hypoplasia, anteverted nostrils, broad upper lip, floppy ears, abnormally formed teeth and widow's peak. Skeletal defects consist of hypoplastic phalanges, retarded bone maturation, cervical spina bifida occulta, fused cervical vertebrae, odontoid hypoplasia, segmentation anomalies, and an additional pair of ribs. Urogenital malformations are less important clinically with respect to the skeletal defects and include scrotal anomalies, hypospadius, and kidney hypoplasia (Gorski et al.,2000; Porteous and Goudie,1991) (Tables 2 and 3). Severe mental retardation has not been associated with FGDY (Orrico et al.,2004; Porteous and Goudie,1991), but some patients may present infantile neuropsychiatric alterations such as attention deficit hyperactivity disorder (Orrico et al.,2005; Orrico et al.,2004).

Genetic linkage studies indicated that in most affected families FGDY was linked to the X chromosome and mapped to Xp11.21 (Porteous et al.,1992; Stevenson et al.,1994). FGD1, the gene responsible for FGDY, was isolated in 1994 and identified as a Rho/Rac guanine nucleotide exchange factor (Pasteris et al.,1994). Mutations in FGDY patients have been found throughout all the FGD1 coding sequence and all are predicted to function as null alleles (see Fig. 1) (Bedoyan et al.,2009; Orrico et al.,2004; Orrico et al.,2000).

FGD1 protein has a proline-rich N-terminal region, followed by a DBL-homology domain, a pleckstrin-homology (PH) domain, a FYVE-finger domain and another PH domain in the C-terminal region (Estrada et al.,2001). It has been shown to be a specific guanine-nucleotide exchange factor for Cdc42 (Zheng et al.,1996), and through its interaction can modulate the actin cytoskeleton (Nagata et al.,1998; Zheng et al.,1996) and regulate the export of cargo proteins from the Golgi complex (Egorov et al.,2009). Independently of this interaction, FGD1 can activate JNK and promote G₁ progression (Nagata et al.,1998). FGD1 can also interact with cortactin and mAbp1 (Hou et al.,2003), and it can be activated by Src (Miyamoto et al.,2003). Recently, FGD1 has been associated with invadopodia (Ayala et al.,2009). Ayala et al. demonstrated the localization of Fgd1 to nascent, but not mature, invadopodia in the melanoma cell line A375MM. Use of RNAi to knockdown FGD1 in this cell line as well as in MDA-MB-231 and PC3 cells correlated with decreased invadopodia activity and decreased Cdc42 activity, suggesting that FGD1 is necessary for invadopodia biogenesis and function.

SPECULATIONS ABOUT THE POSSIBLE ROLE OF PODOSOMES IN CRANIOFACIAL DEVELOPMENT

The study of podosomes/invadopodia is an emerging field. Growing evidence indicates that they are crucial players in integrating external and internal signals that allow focalized degradation of the extracellular matrix, and migration of cells (Linder,2007). These properties and the discovery that mutation of one of their essential components, WASp, is responsible for Wiskott-Aldrich Syndrome made them objects of increasing interest. Most of what we know about podosome biology has been discovered using cellular models, where podosomes can be easily identified by specific staining correlating with focal zones of matrix degradation. However, the lack of markers exclusive for podosomes has impaired the study of these structures in vivo. Tks5, an adaptor protein and Src substrate, is essential for podosome and invadopodia formation (Seals et al.,2005), and seems to be the missing marker that might allow the field to expand from the culture plate to living organisms (Quintavalle et al.,2010). Until then, we can only speculate about the presence and function of podosomes in physiology.

Matrix degradation and migration are key features in the process of epithelial-mesenchymal transition, an essential phenomenon during embryo development (Acloque et al.,2009; Thiery et al.,2009; Yang and Weinberg,2008). Likewise, it has been implicated to be the underlying process by which cancers become invasive and metastatic. The expression of invadopodia in transformed cells has been known since their discovery in Src-transformed fibroblasts, and since then findings indicate that their presence correlates with tumor aggressiveness (Seals et al.,2005). Therefore, it would not be out of place to think that podosomes may play a role during cell movement in embryogenesis. This review has focused on human genetic syndromes that originate when recognized podosomal proteins are mutated. It is remarkable that three of them - FTHS, MNS and FGD - share many distinctive traits causing some to postulate that FTHS was an autosomal recessive form of MNS (ter Haar et al.,1982) (Tables 2 and 3). These three syndromes have similar craniofacial facies - hypertelorism, mandibular hypoplasia, anteverted nostrils, dentition problems -, and MNS and FTHS also similarly affect the heart. This suggests that a similar developmental defect may be responsible for these syndromes. For instance, it could be possible that some of the stem cells involved in craniofacial morphogenesis, such as the neural crest cells or mesenchymal cells, could have podosomes. In support of this, there are some developmental syndromes -such as the branchiooculofacial syndrome, campomelic dysplasia or the Marfan syndrome- where genes important for neural crest cell migration or epithelial-mesenchymal transition are mutated that highly resemble the ones reviewed here (Acloque et al.,2009). Furthermore, we have found that mouse neural crest stem cells can elaborate podosomes in tissue culture (Danielle Murphy and SAC, unpublished observations). It could also be possible that embryonic mesenchymal stem cells, that originate most of the bone structures of face and skull, could have podosomes.

Six of the ten syndromes reviewed here present some degree of craniofacial dysmorphogenesis. The rest affect the heart (XMVD), the brain (PVNH), the hematopoietic lineage (WAS), and the kidney (FSGS1), with little or no defects in any other tissue. Assuming impairment of podosome biology is behind these diseases, specific tissue expression of certain podosomal proteins might explain this heterogeneity of phenotypes. For instance, WASp expression is limited to hematopoietic tissues (Derry et al.,1994), and α-Actinin4 is the only α-Actinin isoform found in human kidney (Kaplan et al.,2000). Restrictive expression of certain proteins would modulate podosome activity in a tissue/cell type specific way. It is important to note that although all the syndromes reviewed here are caused by mutation of proteins present in podosomes, it has not been demonstrated that an actual defect in podosome formation and/or function is the cause of any of them. We cannot rule out that mutation of these proteins may be causing these pathologies through podosome-independent mechanisms. However, given the properties of podosomes it is tempting to speculate that disruption of these structures could be the cause of some of the symptoms/pathologies reviewed here. In the same way, it is possible that other syndromes such as megalocornea mental retardation syndrome (MIM 249310) or serpentine fibula-polycystic syndrome (MIM 600330), with clinical manifestations similar with the ones described in this review, and whose molecular cause is unknown, could be caused by mutations in proteins important for podosome biology.

The study of podosomes/invadopodia is an emerging field. Many of the cellular events driving craniofacial morphogenesis are unknown. We are convinced that in the next years the paths of these two fields will converge.

Acknowledgements

The authors thank Begoña Diaz and Emanuela Ghia for critical reading of this manuscript.