“Soft” tissue patterning: Muscles and tendons of the limb take their form


  • Peleg Hasson

    Corresponding author
    1. Department of Anatomy and Cell Biology, The Rappaport Family Institute for Research in the Medical Sciences, Technion-Israel Institute of Technology, Bat Galim, Haifa, Israel
    • Department of Anatomy and Cell Biology, The Rappaport Family Institute for Research in the Medical Sciences, Technion-Israel Institute of Technology, POB 9649, Bat Galim, Haifa 31096, Israel
    Search for more papers by this author


The musculoskeletal system grants our bodies a vast range of movements. Because it is mainly composed of easily identifiable components, it serves as an ideal model to study patterning of the specific tissues that make up the organ. Surprisingly, although critical for the function of the musculoskeletal system, understanding of the embryonic processes that regulate muscle and tendon patterning is very limited. The recent identification of specific markers and the reagents stemming from them has revealed some of the molecular events regulating patterning of these soft tissues. This review will focus on some of the current work, with an emphasis on the roles of the muscle connective tissue, and discuss several key points that addressing them will advance our understanding of these patterning events. Developmental Dynamics 240:1100–1107, 2011. © 2011 Wiley-Liss, Inc.


Acquisition of shape and pattern during embryogenesis depends on orchestrated signaling and crosstalk between tissues. Our understanding of how these processes occur in vivo is very limited largely due to the highly dynamic and interlinked nature of these processes. The vertebrate musculoskeletal system serves as an ideal model for studying such processes as it is composed of easily identifiable tissues: the skeleton, the muscles and the tendons, and ligaments. These tissues must interact with each other in three-dimensions with high fidelity to form a functional musculoskeletal system. The limb has proven to be a powerful model for studying such events and much effort has been aimed at identifying the cues that pattern the limb skeleton along its proximodistal, anteroposterior, and dorsoventral axes (Tickle,2000,2003; Tabin and McMahon,2008; Towers and Tickle,2009a,b). Although much is known about the molecular pathways that determine patterning of the skeleton, the patterning and morphogenesis of its associated “soft tissues,” namely the muscles, tendons and ligaments and the interwoven connective tissues, have received much less attention. Recent work suggests that soft tissue and skeletal patterning can be uncoupled and that the two processes are to some extent autonomous (Hasson et al.,2010; Li et al.,2010). That, in addition to the large number of tendons and muscles and the complexity of their arrangement, suggests that lessons learned from patterning of the skeleton cannot be simply ascribed to those regulating that of the muscles and tendons. Thus, additional studies are needed to dissect the mechanisms regulating muscle and tendon patterning which will also assist in better understanding of the etiology of congenital diseases affecting soft tissue morphogenesis such as Holt Oram Syndrome (HOS; heart-hand syndrome; OMIM 142900).

With the identification of additional tissue specific molecular markers and the development of novel imaging techniques, insights into the molding of these soft tissues are beginning to emerge (reviewed in Tozer and Duprez,2005; Schweitzer et al.,2010). Early steps of limb muscle development and patterning as well as muscle–tendon interactions have been reviewed elsewhere (e.g., Christ and Brand-Saberi,2002; Buckingham et al.,2003; Vasyutina and Birchmeier,2006; Schweitzer et al.,2010). Hence, this review will revisit some of the recent findings relating to limb soft musculoskeletal tissue patterning with an emphasis on the roles of the connective tissues in this process. Finally, some open questions that need addressing to improve our understanding of tissue and organ patterning will be discussed.


The cells that make up the limb skeleton, tendons, ligaments, and other connective tissues arise from resident cells of the lateral plate mesoderm (Chevallier et al.,1977; Wachtler et al.,1981). In contrast, limb muscles originate from Pax3-positive precursors that delaminate and migrate from the hypaxial dermomyotome and enter the nascent limb bud where they then proliferate and differentiate to form muscle masses (Chevallier et al.,1977; Christ et al.,1977a; Wachtler et al.,1981; Ordahl and Le Douarin,1992). These masses then segregate and split to form the individual muscles (Chevallier et al.,1977; Christ et al.,1977a,b). This difference in embryonic origin enabled the use of classical methods such as chick:quail grafts (Chevallier et al.,1977; Christ et al.,1977a,b) and lineage tracing (Kardon et al.,2002) to determine that muscle precursors migrating into the limb are naïve and are not predetermined to form a specific muscle. Instead, they respond to external local cues within the limb mesoderm that regulate their patterning (Chevallier and Kieny,1982; Grim and Wachtler,1991).

Genetic analyses in mice have revealed evidence on the molecular regulation that determines specific muscles. Thus, while Lbx1 is expressed in all migrating myoblasts, mice mutant for Lbx1 display specific limb muscle defects, presumably due to failure of progenitors to migrate (Mankoo et al.,1999; Schafer and Braun,1999; Brohmann et al.,2000; Gross et al.,2000; Uchiyama et al.,2000). Mox2 mutant mice also show muscle specific defects, yet in this case, this is caused by their inability to proliferate as a result of defective Pax3 expression (Mankoo et al.,1999). As the myoblast progenitors migrate into the nascent bud, they form dorsal and ventral muscle masses. This migration and organization is largely dependent on scatter factor/hepatocyte growth factor (SF/HGF) expressed in the limb mesenchyme, its receptor c-met expressed on the myogenic progenitors (Bladt et al.,1995; Brand-Saberi et al.,1996b; Heymann et al.,1996; Gross et al.,2000) and on EphA4/ephrinA5 and SDF1/CXCR4 (Swartz et al.,2001; Vasyutina and Birchmeier,2006). This step is largely complete by the 33-somite stage in mouse forelimb (around embryonic day [E] 10.5; Houzelstein et al.,1999) and Hamburger and Hamilton (HH) stage 24 in chick (Kardon,1998).

Once in the limb, these Pax3-positive premuscle masses enter the range of the limb's signaling centers secreting bone morphogenetic proteins (BMPs), Wnts, fibroblast growth factors (FGFs), and Sonic hedgehog (Shh). This further refines their positioning and initial patterning using additional modes such as regulation of their proliferation, differentiation and apoptosis (Amthor et al.,1998; Anakwe et al.,2002,2003; Church and Francis-West,2002). In this scenario BMPs, which are in part under the regulation of Shh, play a dual role. High BMP concentrations secreted from the ectodermal and subectodermal domains block myoblast progenitor proliferation and induce their apoptosis (Duprez et al.,1996; Amthor et al.,1998). In contrast, lower BMP levels are required for the proliferation of muscle progenitors and block muscle differentiation. As a result, the Pax3-positive muscle masses that make the myoblast progenitor pool are positioned underneath the ectoderm away from the distal tip of the limb; whereas muscle differentiation occurs in the central domains of the limb distant from the BMP source in the ectoderm (Amthor et al.,1998).


Classical embryological approaches have demonstrated that the limb mesenchyme and the presumed muscle connective tissue (MCT) participate in muscle patterning (Chevallier et al.,1977; Christ et al.,1977a,b; Wachtler et al.,1981; Chevallier and Kieny,1982). To identify the time window during which the limb mesenchyme regulates limb muscle patterning, Chevalier and Kieny have X-irradiated chick somitic mesoderm to generate muscle-less limbs. At specific stages of development, the muscle-less limbs were then recipients of quail myogenic masses and limb muscle patterning was later investigated. These experiments have lead to the suggestion that the MCT, the connective tissue embedded within and ensheathing the muscle, regulates patterning of the migrating myoblasts in a narrow time window, ranging between HH stages 22–27 (Hamburger and Hamilton,1951; Chevallier and Kieny,1982). In recent years, some molecular insights into the developmental functions of the MCT have begun to emerge. These experiments demonstrate that apart from participating in the prepatterning of the limb muscles, the MCT also contributes to the regulation of muscle development and differentiation (Kardon et al.,2003; Hasson et al.,2010; Mathew et al.,2011).

During an attempt to identify genes expressed in muscle forming domains but not in the muscle itself, Tabin and colleagues identified Tcf4, a member of the Tcf/Lef family of transcription factors which is expressed in the MCT and acts downstream of Wnt signaling (Kardon et al.,2003). Early Tcf4 expression generates a template for the forming limb muscles (Kardon et al.,2003) and Tcf4 mutant mice display reduced limb mobility and muscle mispatterning (Mathew et al.,2011). As development proceeds, Tcf4 remains highly expressed in mesodermal fibroblasts associated with the developing muscles that give rise to the anatomical/histological MCT that can be observed at the later stages of development. However, Tcf4 expression is also turned on at low levels in developing fetal myogenic cells where it participates in the regulation of myofiber differentiation (Mathew et al.,2011). Deletion of Tcf4 or the ablation of MCT cells in the developing mouse limb, demonstrated that the MCT by means of Tcf4-dependent and -independent mechanisms, extrinsically regulate myofiber type maturation and participate in the regulation of the switch from fetal to adult muscles (Mathew et al.,2011). In keeping with these results, ectopic activation of Wnt signaling in the chick limb mesoderm demonstrated that Wnt signaling, although not specifically by means of Tcf4, promotes myofiber differentiation. Conversely, blocking Wnt signaling activity prevents myofiber differentiation (Anakwe et al.,2003; Kardon et al.,2003). These results are in line with lineage tracing experiments which demonstrate that myofiber type is not predetermined in the migrating myoblasts, confirming a role for the limb mesenchymal environment in this process (Kardon et al.,2002). That said, Tcf4 expressed in the MCT contributes to, but is not sufficient for, underlying muscle differentiation, as it is also expressed in domains in which muscles do not form and which express scleraxis (scx), a marker of tendon progenitors (Kardon et al.,2003). Whether this is due to a lack of Wnt signaling in these domains or due to a Scx/Tcf4-mediated inhibition of myogenesis is currently not clear. Altogether, Tcf4 is the earliest marker identified in the developing MCT.

What regulates Tcf4 expression to form a template for the myoblast progenitors? Recent observations reveal that BMP-2 and -4 can repress Tcf4 expression, providing a mechanistic insight into the early role of BMPs in positioning the dorsal and ventral muscle masses (Amthor et al.,1998; Bonafede et al.,2006). In contrast to these BMP-mediated ectodermal inhibitory signals, Tcf4 is also a target of the Wnt pathway and is up-regulated following Wnt activation (Kardon et al.,2003). One putative regulator, Wnt3a, is also expressed in the ectoderm and promotes the expression of markers of the connective tissue, including the MCT (e.g., Collagen type I, Decorin and Tenascin; ten Berge et al.,2008). Thus, although not sufficient to explain the details of its expression, opposing signals from the ectoderm refine Tcf4 expression in the underlying mesoderm leading to its expression in the dorsal and ventral muscle masses.


The MCT has been defined as the connective tissue, rich in a mixture of proteoglycans and collagen fibers that ensheathes the muscle fibers, fasciculi, and muscle (Gray et al.,2005). However, this description relates to its histological/anatomical characteristics which can be detected only from late stages of development (around E15.5 onward in the mouse limb). This definition does not include the early MCT, which is rich in fibroblasts yet with relatively little extracellular matrix, and which has shown to regulate muscle prepatterning mediated by Tcf4 (Kardon et al.,2003; Bonafede et al.,2006). How does the MCT direct muscle patterning? One way of approaching these functions is by dissecting the mechanisms underlying human syndromes with muscle patterning abnormalities. One such syndrome is HOS characterized by various musculoskeletal limb, heart, and other abnormalities (Newbury-Ecob et al.,1996; Spranger et al.,1997). It is caused by mutations in the human T-box transcription factor TBX5 (HOS; heart-hand syndrome; OMIM 142900; Basson et al.,1997; Li et al.,1997). In vertebrate model systems, Tbx5 is expressed in the forelimb-forming regions of the lateral plate mesoderm before and during limb bud initiation (e.g., Ahn et al.,2002; Rallis et al.,2003). Once limb buds have initiated, it is expressed broadly in the limb mesenchyme. Yet as limbs develop, its expression becomes more restricted, and by E12.5 Tbx5 expression can be observed in the MCT, few skeletal elements, tendons, and other connective tissues (Hasson et al.,2007). Because at these postinitiation stages Tbx5 does not have an apparent function in the skeleton nor in the muscles (Hasson et al.,2007,2010), it serves as a good target to address MCT early roles in limb muscle patterning.

Tbx5 deletion leads to the mispatterning and ectopic splitting of all limb muscles and tendons as early as E12.5, when tendon progenitors align between the forming muscles (Hasson et al.,2010). Close inspection shows that early expression patterns, but not levels, of a battery of MCT molecular markers, such as Tcf4, Osr1, and -2, SF/HGF among others, is subtly altered in the mutants. Nevertheless, later embryonic histological analysis showed that the MCT is disorganized, demonstrating that MCT organization, but not primary differentiation is affected following Tbx5 deletion. Expression levels of membrane-bound β-Catenin and N-Cadherin within the Tcf4 expression domain are down-regulated, but not eliminated, in the Tbx5 mutants. β-Catenin plays essential roles in both cadherin-mediated cell adhesion and in Wnt signaling. Although a reduction in the levels of β-Catenin was observed in the Tbx5 mutant limbs, expression levels of Wnt targets such as Axin2 and Tcf4 were not affected in these mutants. These observations suggest that Tbx5 primarily affects β-Catenin–mediated cell adhesion rather than its Wnt signaling activities and reinforce the notion that Tbx5 regulates muscle patterning independently of Tcf4 expression or activity. However, these results do not rule out the possibility that the observed muscle mispatterning defects are caused by the alterations in Tcf4 expression. Limb mesenchymal deletion of β-Catenin resulted in MCT disorganization defects and muscle mispatterning among other abnormalities (Hasson et al.,2010). Because β-Catenin is not required for embryonic myogenesis (Hutcheson et al.,2009), these results raise the possibility that β-Catenin is an important target of Tbx5 in this process. Together, these results demonstrate that the organization of the early MCT is key for executing its muscle patterning activities (Hasson et al.,2010).


Once muscle masses have formed, muscle splitting begins to take place. Light and electron microscopy studies have revealed that, in the incipient cleavage sites, the extracellular space increases, cells within the myotube cluster gradually dissociate, muscle cells are phagocytosed, and stellate mesenchymal cells take their place. With time, connective tissue similar to that surrounding the whole muscle builds up in the newly cleaved site (Schroeter and Tosney,1991a,b). This process proceeds in an orderly and stereotyped manner, initially from the dorsal and ventral muscle masses, generally progressing in a proximal to distal direction (Schroeter and Tosney,1991a,b; Kardon,1998). By the end of this process, the tetrapod limb has more than 40 muscles arranged in a precise and reproducible pattern. This complex arrangement of muscles, coordinated along the three limb axes, allows a high degree of strength and manipulation in limb movement. Surprisingly, although critical for generating the complex musculoskeletal morphology, our understanding of the splitting process and its regulation is very limited.

Several tissues and signaling pathways have been suggested to play a role in regulating muscle splitting. While the nervous system has been proposed to participate in the process, this has been ruled out as the absence of limb nerves following neural tube ablation does not affect muscle splitting pattern (Edom-Vovard et al.,2002). Recently, however, the involvement of two signaling pathways has been demonstrated. Retinoic acid signaling has been shown to regulate apoptosis in muscle bellies thus generating cleavage in a muscle mass (Rodriguez-Guzman et al.,2007). Meanwhile, although Duprez and colleagues have not identified any cell death in cleavage sites, they have highlighted a link between the splitting site and vascular development. Using overexpression assays in chick, they established that platelet derived growth factor-B (PDGFB) secreted from the endothelial cells participates in muscle splitting. PDGFB is secreted as vessels develop throughout the limb, leading to the up-regulation of extracellular matrix and connective tissue markers, including Tcf4. In turn, Tcf4 participates in the recruitment of endothelial cells (Levy et al.,2002). As a consequence, a positive feedback loop is generated in which extracellular matrix proteins and Tcf4 are up-regulated along the length of the vessel at the expense of muscle formation. The end product of this sequence is the promotion of muscle mass separation which enables splitting of the muscle (Tozer et al.,2007).

Could the MCT be also implicated in regulating muscle splitting? Deletion of both Tbx5 in the forelimb and its paralogue Tbx4 in the hindlimb during early limb outgrowth, or deletion of their putative downstream target β-Catenin, strongly affects MCT organization, and limb muscles are ectopically split (Hasson et al.,2010). These results are suggestive of the involvement of the MCT in the splitting process. Importantly, in these Tbx5/4-dependent muscle splitting defects, vascular involvement or abnormal retinoic acid signaling were not identified in a way that could explain the observed phenotypes. This suggests that regulation of the forming cleavage domains is presumably carried out by several mechanisms.


Tendons and ligaments are highly related tissues, expressing many similar transcription factors (e.g., Scx) and extracellular matrix proteins (Pearse et al.,2009), yet they serve different functions: tendons attach muscles to bones and ligaments attach bones to bones. Thus, the key difference is in their anatomical connections, rather than their histology or gene expression profile. Most experimental data do not distinguish between the two tissues largely due to a lack of markers that separate each tissue. In this review, as has been done previously (Pearse et al.,2009), both are referred to as tendons. Due again to a lack of early and specific markers, our knowledge of the mechanisms regulating tendon and ligament patterning is in its infancy, and most of what is known relates to the initial formation of tendons (reviewed in Tozer and Duprez,2005). The identification of scx (Cserjesi et al.,1995; Schweitzer et al.,2001) as a marker of early tendon progenitors has allowed the study of pathways and genes involved in their development. Novel techniques such as genome-wide microarrays have further revealed new markers that can now be used for additional experiments (e.g., fibin2; Pearse et al.,2009).

Although originating from separate embryonic origins, early tendon progenitors and Pax3-positive myoblast progenitors are expressed in a mixed and partially overlapping sub-ectodermal domains along the dorsal and ventral regions of the limb mesenchyme (Schweitzer et al.,2001; Murchison et al.,2007; Fig. 1), suggesting they might respond to common inputs. Ectoderm removal inhibits the initiation, but not maintenance, of scx expression, whereas BMPs restrict scx expression (Schweitzer et al.,2001,2010), similar to the response observed by myoblast progenitors (Amthor et al.,1998). As yet, unlike the myotome which is involved in scx induction, in the limb, ectoderm is the only tissue demonstrated to be capable of initiating its expression (Schweitzer et al.,2001). While these observations are suggestive of cross-talk between the two progenitor populations, generation of muscle-less limbs has demonstrated that, at these early stages, the two populations develop independently of each other (Kardon,1998; Schweitzer et al.,2001; Edom-Vovard et al.,2002; Brent et al.,2005). It will be interesting to challenge this hypothesis by the generation of tendon-less limbs.

Figure 1.

Dynamic muscle and tendon patterning throughout limb development. A–H: Muscle (A,C,E,G) and tendon (B,D,F,H) limb pattern at embryonic day (E) 10.5 (A,B), E12.5 (C,D), E13.5 (E,F), and E15.5 (G,H). Note the early overlap between Pax3 (A, muscle) and scx (B, tendon) progenitors and how they become complementary and mutually exclusive as the limb develops. MyoD (C,E) or My32 (G) was used to monitor differentiated muscles, whereas scx (D,F) or Scx-GFP (H) to detect tendons.

The development of a tendon is a multistep process composed of induction, recruitment, and differentiation of scx-expressing tendon progenitors. TGFβ and FGF signaling, as well as Scx-transcriptional input, play key roles in this process (Edom-Vovard et al.,2002; Edom-Vovard and Duprez,2004; Murchison et al.,2007; Eloy-Trinquet et al.,2009; Pryce et al.,2009). TGFβ signaling is required for the recruitment and maintenance of tendon progenitors by E11.5–E12.5 in the mouse limb. In TGFβ2−/−; TGFβ3−/− double mutants, all limb tendon progenitors are lost (Pryce et al.,2009); in contrast, in scx mutants, early scx expression is not affected and phenotypes are only observed beginning at E13.5. In both cases however, the tendon progenitors do not die but assume a different fate, suggesting scx does not mark fully committed tendon progenitors (Murchison et al.,2007; Pryce et al.,2009). As development proceeds, tendon morphogenesis begins to differ along the proximal–distal axis of the limb. In the proximal limb, tendons are induced but do not segregate to form specific tendons in the lack of muscles. In contrast, in the distal autopod, where muscles are not present at these early stages, tendon formation is initiated, and segregation into individual tendons occurs. These distal tendons will however later require muscle contact, and will degenerate in a muscle-less limb environment (Shellswell and Wolpert,1977; Kieny and Chevallier,1979; Kardon,1998; Edom-Vovard et al.,2002). By E12.5, when divergent muscle masses are discernable, tendon progenitors become organized between the muscles of the proximal limb (Fig. 2), and by E13.5, these tendon anlagen condense and differentiate to specific tendons throughout the limb (Fig. 1). This process does not depend only on scx, because scx overexpression does not affect patterning or formation of either tendon or muscle (Schweitzer et al.,2001).

Figure 2.

A,B: Muscle:tendon intimate interactions at embryonic day (E) 12.5. Still images from 3D Optical Projection Tomography of Scx-GFP transgenic mouse limb stained for muscle (A4.1025; red) and tendons (green fluorescent protein [GFP], green) at E12.5. Individual muscles can be observed in very close proximity to tendon progenitors. Note the tendon development in proximity to muscle in the proximal limb but not in the distal autopod where muscles are not present at these early stages. Ventral (A) and lateral (B) views.

Recently, the involvement of additional transcription factors (Mohawk, Egr1, and Egr2) in tendon development and morphogenesis has been demonstrated (Ito et al.,2010; Lejard et al.,2010; Liu et al.,2010). Although these factors do not participate in tendon patterning, they may be required alongside scx to regulate tendon development. The atypical homeobox gene Mohawk (Mkx), is expressed in differentiating limb tendon progenitors from E12.5 as they align between the forming muscles, and is up-regulated once the progenitors undergo condensation and further differentiation (Liu et al.,2010). Mkx null mice display a reduction in type I collagen, the main collagen expressed in tendons (Ito et al.,2010; Liu et al.,2010) but no reduction in scx expression, suggesting that Mkx is required for tendon differentiation and morphogenesis but is not involved in their specification (Liu et al.,2010). Two additional transcription factors, members of the Early growth response-like family genes, Egr1 and Egr2, were similarly shown to be expressed in developing and differentiating tendons (Lejard et al., 2010). Forced expression of Egr1 and Egr2 is able to induce de novo scx expression while their deletion results in reduced expression of scx and Mkx. The endogenous temporal onset of Egr1 and Egr2 suggests that these genes are involved in scx maintenance and are required for expression of tendon-specific collagen. Egr1 and Egr2 expression can be induced by Fgf4, which is expressed at muscle extremities where the future tendons will attach (Edom-Vovard et al.,2002; Lejard et al.,2010). Of interest, the Drosophila Egr family member Stripe has previously been shown to play a role in fly tendon development (Frommer et al.,1996; Becker et al.,1997), suggesting that a functional analogy exists in the activity of Egr genes in tendon development in both flies and vertebrates (Lejard et al.,2010).


This review has tried to illustrate some of the novel findings in the field of muscle and tendon patterning, emphasizing the roles of the MCT and putting them into perspective with previous work. Overall, soft tissue patterning is a relatively unexplored territory in comparison to that of its associated skeleton. Because the two processes can be uncoupled (Hasson et al.,2010; Li et al.,2010) more research is needed to understand how these tissues are patterned and how their development is coordinated with that of the skeleton. The emergence of new tools and reagents has lead to significant progress in recent years. For example, advancement in three dimensional imaging (such as the Optical Projection Tomography or High Resolution Episcopic Microscopy; Sharpe et al.,2002; Weninger and Mohun,2002) has facilitated such studies and allowed investigation of dynamic tissue patterning in the space of an organ such as the limb (e.g., DeLaurier et al.,2006,2008; Boot et al.,2008; Hasson et al.,2010). Furthermore, identification of specific markers such as Tcf4 (MCT), scx (tendons), or the newly identified fibin2 (tendons), (Schweitzer et al.,2001; Kardon et al.,2003; Pearse et al.,2009) promotes the dissection of the specific pathways that regulate muscle and tendon patterning and the relations between them. With the generation of novel mouse lines, based around these markers (e.g., Mathew et al.,2011), novel insights should soon follow.

Numerous basic questions still remain unanswered in this field and their exploration should promote the understanding of fundamental questions also relevant to other developmental settings. One example is a basic definition of the term MCT. It seems unlikely that the early Tcf4 expressing cells play a similar role later on when a histologically distinct tissue rich in extracellular matrix can be observed. Is there a specific time window during which Tcf4 directs muscle and tendon pattern? Chevallier and Kieny proposed that the MCT directs muscle patterning in a narrow time window ranging from HH 22 up to HH 26–27 in the chick (Chevallier and Kieny,1982). The Tbx5 (and Tbx4) mouse mutant analysis has identified a comparable time window. These data demonstrate that these T-box genes regulate muscle and tendon patterning, presumably by means of MCT organization, at around E11.5–E12.5, but not later (Hasson et al.,2010). Importantly, this analysis also showed that muscle and tendon patterning and morphogenesis can be dissociated from other steps of their development (Hasson et al.,2010). The recent analysis performed by Kardon and colleagues demonstrated that Tcf4 and the MCT further regulate myofiber differentiation (Mathew et al.,2011). It will be interesting to test whether the MCT organization defects resulting from the Tbx4/5 deletion similarly affect these properties. Altogether these studies demonstrate that the MCT controls muscle development and patterning at different levels with distinct mechanisms.

What is the nature of the signals produced by the MCT that direct soft tissue patterning? Are they secreted or membrane bound? The recent analysis by Kardon and colleagues (Mathew et al.,2011) demonstrated that at least to some extent, the MCT secretes a signal to the underlying muscle to regulate its differentiation. However, the MCT also participates in muscle patterning. That loss of MCT organization leads to muscle mispatterning defects (Hasson et al.,2010) cannot distinguish between the distinct types of signals. The primary phenotype in the Tbx5 mutants is a patterning defect. No proliferation or cell death abnormalities were observed; processes regulated by secreted factors such as Wnts, BMPs, FGFs, and Shh, therefore, evoking the involvement of a membrane bound signal. In such a scenario, the loss of MCT organization would mainly affect the interaction between the MCT and the myoblasts without affecting other muscle developmental processes. If this is indeed a membrane bound signal, how do these interactions occur, and what players partake in them? Could N-Cadherin expressed on myoblasts and on MCT cells underlie such interactions? N-Cadherin and β-Catenin expressed in the MCT have been shown to regulate muscle patterning (Brand-Saberi et al.,1996a; Hasson et al.,2010). In addition, cadherins and β-Catenin expressed in craniofacial connective tissue play similar roles in patterning adjacent head muscles (Rinon et al.,2007), which, similar to the limb, also arise from a distinct embryonic origin. Yet both genes are also required autonomously within the myoblasts for normal muscle development (e.g., Brand-Saberi et al.,1996a; Goichberg and Geiger,1998; Anakwe et al.,2002,2003; Geetha-Loganathan et al.,2005,2006; Hutcheson et al.,2009). N-Cadherin has been shown to participate in homophilic interactions between neighboring cells (e.g., Levenberg et al.,1998); thus, it serves as an excellent candidate to fulfill such a role.

Tcf4 and scx expression partially overlaps (e.g., Kardon et al.,2003). This should not come as a surprise when looking at the complexity and proximity of muscle and tendon progenitors (Fig. 2). Of interest, disruption of MCT organization not only affects muscle morphogenesis, but also interferes with scx and tendon patterning. This disruption is independent of the associated muscle abnormalities because the tendon patterning defects are observed before, or just as, the interdependence of the two tissues begins (Hasson et al.,2010; Schweitzer et al.,2010). These results, along with Tcf4 expression, promote the suggestion that the MCT is not solely a connective tissue required for muscle development, but also may play a role in tendon development. It would thus be interesting to test whether this tissue serves as a junction for cross talk between the distinct elements of the musculoskeletal system allowing integration of signals required for their coordinated development.


I thank Natalie Butterfield and Malcolm Logan for critical reading and insightful comments on the manuscript.