During early embryonic development of mammals, the cloaca acts as a collecting chamber into which the intestinal, urinary, and genital tracts converge (Martinez-Frias et al.,2000; McBride et al.,2003; Dravis et al.,2004; Valasek et al.,2005). Later in development, the cloaca is partitioned into separate urogenital and alimentary tracts (Kimmel et al.,2000; de Santa Barbara et al.,2002; Penington and Hutson,2003; Hynes and Fraher,2004). In contrast to mammals, in adult birds, amphibians, reptiles, monotreme mammals, and chondrichthyan fish, the cloaca persists throughout their lives as a single opening for urinary, digestive, and genital ducts (King and McLelland,1981). In mammals, the equivalent of cloacal muscle are the perineal sphincters guarding the lower urogenital and gastrointestinal tract. The origin of cloacal muscles has only recently been shown to lie in the hypaxial muscle precursors in the somites which are known to populate the limbs and the ventral body wall (Valasek et al.,2005).
Following hypaxial myogenic precursors epithelio-mesenchymal transition from the ventrolateral dermomyotome, they either migrate individually into the limb mesenchyme or form cell sheets, myotomal extensions, that eventually give rise to ventral body wall muscles (Galis,2001). The single individually migrating myogenic progenitor cells generate hypaxial muscles of the extremities, tongue and diaphragm (Chevallier et al.,1977; Christ et al.,1977; Christ and Ordahl,1995). The formation of the hypaxial compartment in the dermomyotome, the delamination, the migration, and the eventual differentiation of these myogenic precursors are tightly regulated by several genes. Pax3 and Lbx1 are required for the specification of these myogenic precursors in the dermomyotome, cMet and scatter factor (SF/HGF) are involved in the delamination process, Lbx-1 enables the navigation of the myogenic precursors along migratory routes, while MyoD and Myogenin play a role in the myogenic differentiation program (Bober et al.,1994; Tajbakhsh et al.,1997; Dietrich et al.,1999; Schafer and Braun,1999; Brohmann et al.,2000; Gross et al.,2000). The experimental evidence that cloacal muscles originate from the somites that provide myogenic precursors for the hindlimb has come from the work of Valasek and colleagues. They report that the cloacal muscles in the developing chicken embryo derive their origin from the same somites that also give rise to the leg muscles (Valasek et al.,2005). In fact, cells that participate in cloacal muscle formation initially migrate from the somites into the developing hindlimb bud (Christ et al.,1974,1977; Chevallier et al.,1977; Jacob et al.,1979; Rees et al.,2003; Bonafede et al.,2006) and then translocate from the ventral muscle mass in the hindlimb toward the cloacal tubercle (Valasek et al.,2005; Evans et al.,2006). These cloacal precursors express Pax7 and MyoD as they extend out of the proximal limb toward the cloacal tubercle.
In placental animals, the hindgut initially ends blindly in the cloaca where the cloacal membrane separates it from proctodeum. Mesenchymal condensations inside the cloaca later form the urorectal septum dividing the cloaca into a ventral urogenital sinus and a dorsal anorectal canal. Later during development, the urorectal septum thickens and forms the deep perinium where the anal and urogenital sphincter muscles are located. Only after the rupture of the cloacal membrane a connection of the endodermal and ectodermal part of the anal canal is established. An epithelial plug ensures a closed anal canal till late into the fetal period. Defects in the formation of the urorectal septum and failure in the rupture of the urorectal membrane can give rise to severe forms of anorectal malformations. Persistent cloaca is normal in birds, reptiles, and some fish, but in human infants appears as a complex anorectal and genitourinary malformation whereby the rectum, vagina, and urethra are fused. A common opening larger than 3cm in diameter at birth is associated with a poor prognosis as the sphincter function is compromised and is not amenable to an effective surgical correction (Keith et al.,2005). Although work over the years using the mouse and chicken model has revealed the involvement of Shh and Wnt signaling in anorectal malformations (Mo et al.,2001; Tai et al.,2009), the etiology and pathogenesis of anorectal malformations remains poorly understood and controversial.
Research using avian and mammalian model systems are leading to a consensus on the evolutionary history of muscle formation, which enables us to use the avian model to depict the origin and formation of distinct muscles in mammals (Brand-Saberi and Christ,2000; Valasek et al.,2005). The external urogenital and anorectal sphincter muscles in mammals are functionally homologues to the cloacal sphincter muscles in birds. The underlying genetic and molecular mechanisms governing the origin and the development of the early common cloacal muscles are not understood. Moreover, the mechanism active during remigration of these myogenic precursors from the ventral muscle mass to the cloaca has not yet been described. In our present work, we attempted to decipher probable molecular players that are active during these migratory events. Chemokine receptor CXCR4 and its ligand SDF-1 have been implicated in numerous developmental processes and precursor migration events during embryonic development (Murdoch, 2000). Primordial germ cells express functionally active CXCR4 on their surfaces and respond to SDF-1 cues leading them toward the gonads (Stebler at al.,2004). CXCR4 and SDF-1 mutant mice show defective B-cell lymphopoiesis (Nagasawa et al.,1996; Zou et al.,1998) and abnormal vasculogenesis, developmental defects that occur due to defective cell migration. We and others have previously shown that the chemokine SDF-1 and its receptor CXCR4 are required for the migration of myogenic precursors in the limb buds of developing chicken embryos (Vasyutina et al.,2005; Yusuf et al.,2006). We hypothesized a role of chemokine SDF-1 and its receptor CXCR4 during cloacal myogenic precursor migration. Our hypothesis was based on the observation of SDF-1 expression in the cloacal region from stage HH27 onward (Rehimi et al.,2008) and CXCR4 expression in migratory cells (David et al.,2002; Li et al.,2004; Yusuf et al.,2005; Vasyutina et al.,2005; Haas and Gilmour,2006). To test our hypothesis, we targetted the SDF-1/CXCR4 signalling at the level of the proximal hindlimb and analysed effects on the migration of cloacal muscle precursors and eventual cloacal muscle formation. We show here that SDF-1 and CXCR4 play an important role in the remigration of myogenic precursor cells from the ventral part of the hindlimb into the cloaca. Our experimental manipulation targeting the SDF-1/CXCR4 signaling at the level of the cloacal muscle precursors in the proximal hindlimb affected their ability to migrate out of the hindlimb toward the cloaca and eventually leads to defects in cloacal muscle formation. We provide evidence based on manipulation experiments in chicken embryos that CXCR4/SDF-1 signaling is active during cloacal muscle formation and SDF-1 serves as a guiding cue to migratory cloacal muscle cells that express its receptor CXCR4 on their path from hindlimb toward the midline to form cloacal/perineal muscles.
SDF-1, CXCR4, Pax7, and MyoD Expression in the Cloaca
We have previously defined the expression pattern of SDF-1 in the developing chicken embryo. We observed that SDF-1 has a wide expression in different tissues during development. At stage HH26–29, SDF-1 transcripts are expressed in the cloacal cleft (Fig. 1A,B). Furthermore, at stage HH29, SDF-1 is strongly expressed in the proximal region of the hindlimb (Fig. 1B). At similar stages, CXCR4 transcripts are visible around the developing cloacal region (Fig. 1C), while a few CXCR4-positive cells were observed near the cloaca at stage HH29 (Fig. 1D). A band of Pax7 expression extending from the proximal side of hindlimb toward the cloaca is visible at stage HH26–HH29 (Fig. 1E,F). At similar stages, in MyoD expression domain is observable in the proximal hindlimb extending toward the midline cloacal region (Fig. 1G,H).
SDF-1-EGFP–Expressing Cells Affect the Migration of Cloacal Myogenic Precursors
We injected COS-1-SDF-1-EGFP–expressing cells at stage HH25–26 in the ventral proximal region of the hindlimb (Fig. 2A,G,M). We were initially interested in analyzing the effects of SDF-1 on its specific chemokine receptor CXCR4. An agglomeration of CXCR4-expressing cells around the SDF-1-EGFP–expressing cells (12 of 15 cases examined, Fig. 2B) was clearly observable. As a consequence, the number of CXCR4+ cells near the cloacal region was considerably decreased (Fig. 2B,C) as compared to the control side (12 of 15 cases examined, Fig. 2E,F). Similar effects were observed (agglomeration of cells around the SDF-1–transfected cells and decrease of the extension band from the hindlimb toward the cloaca) when Pax7 and MyoD probes were tested (9 of 10 cases examined, Fig. 2H,I,N,O). Implantation of control EGFP-expressing cells showed no effect on the CXCR4, Pax7, and MyoD population in the limb bud and near the cloaca (15 of 15 cases examined, Fig. 2K,L,Q,R). Therefore, implantation of SDF-1–expressing cells in the proximal hindlimb led to a decrease in myogenic cells moving out toward the developing cloaca, whereas no such effect was observed after implantation of control EGFP-expressing cells.
CXCR4 Inhibitors (T140/TN14003) Affect the Migration of Cloacal Myogenic Precursors
We implanted acrylic beads soaked in T140 and TN14003 at stage HH25–26 in the ventral proximal region of the hindlimb. The expression of CXCR4 was prominently reduced in the ventral proximal side of hindlimb where the beads were placed as well as in the extension band toward the cloaca (16 of 18 cases examined, Fig. 3A). Similar effects were observed with Pax7 (Fig. 3E) and MyoD (12 of 15 cases examined, Fig. 3I). The reduced expression of CXCR4 (Fig. 3B), Pax7 (Fig. 3F), and MyoD (Fig. 3J) could be likewise appreciated in histological sections of these animals. Control phosphate buffered saline (PBS) beads showed no effect on the CXCR4, Pax7, and MyoD population in the limb bud (10 of 10 cases examined, Fig. 3C,D,G,H,K,L).
SDF-1-EGFP Cells and CXCR4 Inhibitors Affect the Formation of Cloacal Muscle Mass
In our next experiments, we were interested in determining if application of SDF-1–expressing cells and CXCR4 inhibitors to the proximal hindlimb would have an impact on cloacal muscle formation. After implantation of SDF-1 cells or T140/TN14003 beads into the proximal hindlimb (Figs. 4A, 5A,B), we reincubated manipulated embryos for extended time periods and analyzed these embryos at stage HH33 to look for any possible alteration in cloacal muscle formation. We observed that the cloacal muscle is decreased due to decreased muscle precursor cell migration from the ventral side of the hindlimb toward the cloaca (7 of 8 cases examined, 8 of 10 cases examined after SDF-1 cells injection and T140/TN14003 bead implantation, respectively, Fig. 4B,C, and Fig. 5A–C). Control embryos that were injected with control EGFP-expressing cells or PBS beads showed no effect on cloacal musculature (10 of 10 cases examined, Fig. 4D–F and Fig. 5D–F). The beads do not cause a mechanical hindrance or shielding effect to perturb the migration of myogenic precursors as no decrease of muscle mass is observed in the embryos in which PBS beads were inserted.
T140/TN4003 Do Not Cause Cell Death
To ensure that the reduction in expression of CXCR4, Pax7, and MyoD was not a result of cell death following T140/TN14003 bead implantation, we used the Nile blue sulfate cell death detection method (van den Eijnde et al.,1997; Zuzarte-Luis and Hurle,2002) after T140/TN14003 and PBS bead implantation at intervals of 24 hr. No increased uptake of Nile blue sulfate was observed indicating absence of cell death (5 of 5 cases examined, Fig. 6A,B). We have also previously documented that T140/TN14003 at concentrations of 5–10 mg/ml does not induce cell death (Yusuf et al.,2006).
Cloacal Muscle Precursors Are CXCR4- and Pax7-Positive
We were interested to know whether CXCR4+ cloacal muscle precursor cells also express Pax7. For this reason, we have performed double immunohistostaining on paraffin sections of mouse embryos at stage embryonic day (E) 13.5. We observed that the cloacal muscle precursor cells coexpress CXCR4 (Fig. 7A) and Pax7 (Fig. 7B; 4 of 5 cases examined, Fig. 7C).
Ectopic Application of COS-1-SDF-1-EGFP Cells in the Chicken and Mice Hindlimb Attract Migrating Cloacal Myoblast
To determine if CXCR4+ muscle precursors respond to ectopic SDF-1 source, we applied COS-1-SDF-1-EGFP cells ectopically within the hindlimb of chicken and mouse embryos at stage HH26 and E12.5, respectively, in limb cultures. For this, COS-1 cells were transiently transfected with an SDF-1-EGFP expression construct and implanted into the hindlimb bud of chicken (HH26) and mouse (E12.5) embryos and thereafter the operated limbs were removed and cultured in collagen culture gel system. An accumulation of CXCR4+ muscle precursors ectopically around the SDF-1-EGFP–expressing cells was observable in the murine hindlimbs (16 of 18 cases examined, Fig. 8A,B). Furthermore, a similar agglomeration of Pax7+ cells around the SDF-1-EGFP–expressing cells occurred in cultured chicken hindlimb. These results show that CXCR4+/Pax7+ cells are able to respond and migrate to ectopic SDF-1 source (11 of 15 cases examined, Fig. 8C,D). Similar effects were absent in control limbs that received transplants of EGFP-transfected COS-1 cells (data not shown).
Cell migration is a central process in the normal physiology, development, and maintenance of multicellular organisms. During development, the chemokine, SDF-1 and its receptor, CXCR4 play an important role in different processes such as vascularization and embryogenesis. Mutations of CXCR4 or SDF-1 affect the migration of cerebellar granule cells and hippocampal and cortical neuronal progenitors (Lazarini et al.,2003; Pujol et al.,2005). CXCR4/SDF-1 signaling is also necessary for the migration of primordial germ cells in fish, avians, and mammals (Knaut et al.,2003; Stebler et al.,2004). We have previously analyzed the expression pattern of CXCR4 (Yusuf et al.,2005) and SDF-1 (Rehimi et al.,2008) in the developing chicken embryo and found that it is expressed in a multitude of developing embryonic structures undergoing active morphogenesis. In this present work, we report the expression of CXCR4 and SDF-1 during cloacal muscle myogenic precursor migration in chicken embryos and unravel their role during the migration of these cloacal muscle precursors from the proximal hindlimb toward the developing cloaca.
In an elaborate work, Valasek et al. (2005) could show that the cloacal musculature is of hypaxial origin and that cloacal muscle precursors temporarily reside within the ventral muscle mass of the hindlimbs in chicken and mouse embryos. Surgical experiments and detailed analyses of genetic mutants allowed us to conclude that the formation of a normal proximal hindlimb field and the delamination of the myogenic precursors from the dermomyotome are an essential prerequisite for the correct development of cloacal muscles (Valasek et al.,2005). Limbless (Prahlad et al.,1979) and met mutants (Dietrich et al.,1999), from which the cloacal muscles are completely absent, strengthen this hypothesis. Interestingly however, analysis of Lbx1 mutants showed a normal cloacal development (Valasek et al.,2005), probably because progenitors of the ventral premuscle masses develop normally in this mutant.
Cloacal muscle precursors originate predominantly from the same axial somitic levels that give rise to hindlimb muscles, namely somite 30 to 34 in chicken embryos. Only somite 34 provides muscle precursors exclusively for the cloaca, whereas somites 30–33 (together with somite 26–29) additionally provide limb muscle precursors. Stepwise analysis of the migratory pattern of cloacal muscle precursors shows that two separable migratory events are involved during cloacal muscle formation. In the chicken embryo, the myogenic precursors first delaminate from the hypaxial dermomyotome at appropriate axial levels and migrate into the hindlimb field as individual migrating precursors. This migration event requires an effective delamination of myogenic precursors from the hypaxial dermomyotome and is dependent on Met and SF/HGF signaling, while migration of these hindlimb muscle precursors deeper into the limb mesenchyme additionally uses the CXCR4/SDF-1 signaling (Vasyutina et al.,2005) and Lbx1 (Birchmeier and Brohmann, 2000). The second migration event occurs once the migration of muscle precursors from the somites toward the hindlimb mesenchyme has ceased around stage HH25–26. Cloacal precursors now extend similar to myotomal extensions for the ventral body wall muscles from the proximal ventral muscle mass of the hindlimb toward the future cloacal anlage in the midline (Valasek et al.,2005). During the initial migratory event, the cells that delaminate from the hypaxial dermomyotome express met, Pax3, Lbx-1, and only after they have entered the limb mesenchyme, CXCR4 expression is detectable. Data from mouse and chicken support the notion that CXCR4-positive cells form a subpopulation of a larger Lbx1-expressing pool of the muscle precursors invading the limb (Vasyutina et al.,2005). Eventually, as the initial migration event ceases, genes necessary for the initiation of the myogenic differentiation such as MyoD, Myogenin, and Mrf4 are turned on. It has been shown that, as differentiation progresses in the ventral and dorsal premuscle masses in the limb, there is a gradual decrease of the Pax3+/Lbx1+–expressing precursor pool coupled with a gradual rise of MyoD+ precursors. Analysis at the cellular resolution has shown that apart from the exclusively Lbx1+- and MyoD+-expressing precursors, a mixed population of the MyoD+/Pax3+/Lbx1+ pool is frequently encountered; however, none or very few CXCR4+/MyoD+ precursors are seen (Vasyutina et al.,2005). Therefore, CXCR4 expression in limb muscle precursors is lost as the differentiation begins. The scenario is different during the second migratory event when cloacal muscle precursors move from the proximal ventral hindlimb toward the cloacal tubercle. These precursors have already initiated the myogenic program and express MyoD. This is similar to the hypaxial myotomal extensions that partially express MyoD and form the ventral body wall musculature.
We report here for the first time that CXCR4 and its ligand SDF-1 are expressed at appropriate stages during the second migratory event and that they play a role during this migration. After experimentally perturbing the CXCR4/SDF-1 signaling pathway in the proximal hindlimb, the migration of cloacal muscle precursors was affected. The second migration event during cloacal muscle formation is reminiscent of the formation of myotomal/dermomyotomal extensions giving rise to ventral body wall muscles and may possibly share common molecular effectors. We have observed the expression of CXCR4 in the hypaxial myotomes/dermomyotome that extend ventrally to form the ventral body wall muscles (data not shown).
Whether CXCR4+ cells represent a subpopulation of the cloacal muscle precursors participating in the second migratory event needs to be further studied. There are indications for this as never during our experiments do we observe the entire cloacal muscle formation can be impaired, pointing toward the existence of a CXCR4− subpopulation that is not affected by our experimental manipulations targeting the CXCR4+ cell pool. However, this issue needs to be further addressed at the cellular level to show if any such subpopulations exist. As mentioned, similar CXCR4+ and CXCR4− myogenic presursor cells have been described in the context of limb muscle precursors (Vasyutina et al.,2005).
The role of chemokine receptor CXCR4 and its ligand SDF-1 during myogenic cell migration for the formation of cloacal musculature add up to the list of several roles of this signaling pair during development. The formation of the cloaca is of utmost importance for the development and patterning of the lower gastrointestinal and urogenital tract. Defect in the cloacal formation could lead to varying degress of anorectal malformations often accompanied with functional disability for the newborn infants needing immediate surgical intervention. How far defective cloacal muscle formation resulting from perturbed cell migration contribute to cloacal defects and eventual anorectal malformation has not been addressed as molecular mechanisms of cloacal muscle formation have just started to emerge. Understanding developmental cloacal myogenic migration and cloacal muscle formation, would help in understanding the pathogenesis of anorectal anomalies where defective myogenic cell migration is the primary underlying cause. Several tissue interactions have been described in the literature that influence myogenic patterning like epithelio-mesenchymal crosstalk (Anakwe et al.,2003), myogenic precursor-limb mesenchyme precursor interactions (Bonafede et al.,2006) and blood vessel formation and muscle splitting (Tozer et al.,2007). Therefore, further studies investigating the role of signaling molecules such as Shh and Wnts signaling aimed at analyzing the interaction of cloaca endodermal lining with the surrounding mesenchymal tissue and muscle tissue would greatly help in understanding the pathogenesis of anorectal malformations.
Fertilized chicken eggs obtained from a local breeder were incubated at 37°C and 80% humidity. Following operative procedures, the eggs were re-incubated at 37°C and 80% humidity to develop further to the required stages. Staging was done according to Hamburger and Hamilton (1992).
Construction of the SDF-1 Transgene Plasmid
For obtaining the chicken SDF-1 gene, reverse transcriptase-polymerase chain reaction (RT-PCR) was performed with a pair of cSDF-1 gene-specific primers (sense primer: 5′ GCCTGCACCGTCGCCAGAATG 3′; and antisense primer: 5′ AGGCC AACTCCAAACCCATCTTCA 3′). The RT-PCR product (425 bp) was cloned into pDrive vector and the cSDF-1 insert was confirmed by sequencing. The full coding sequence of the cSDF-1 cDNA can be re-harvested from the constructed pDrive- cSDF-1 plasmid digested with both NheI and PstI. For construction of the cSDF-1 overexpression plasmid, the NheI–PstI fragment containing the full-length chicken SDF-1 coding sequence was re-cloned into a prelinearized pIRES2-EGFP vector (BD Biosciences Clontech) treated with the same restriction enzymes.
Cell Culture, Transfection, and Implantation In Ovo
COS-1 cells were a gift from Prof. Dr. Carmen Birchmeier (Max-Delbrück-Center for Molecular Medicine, Berlin). COS-1 cells were grown and transfected with SDF-1-EGFP constructs as described previously (Vasyutina et al.,2005). In brief, COS-1 cells were grown in DMEM (Dulbecco's modified Eagle's medium; Invitrogen, USA) supplemented with 10% FCS (fetal calf serum; Invitrogen) in a 37°C humidified atmosphere of 5% CO2 in air. Cells were passaged by treatment with 0.25% trypsin (w/v)/1 mM EDTA (ehtylenediaminetetraacetic acid; Invitrogen), harvested, centrifuged, and then resuspended in DMEM (DMEM; Invitrogen, USA) supplemented with 10% FCS (Invitrogen) onto fresh flasks.
Cells were seeded 1 day before transfection at a density of 3 × 105 cells per T25 flask. Cells were approximately 70% confluent on the day of gene transfection. SDF-1-EGFP transfection was carried out with Lipofectamine (Invitrogen) according to the manufacturer's protocol. Transfection efficiency was estimated by enhanced green fluorescent protein (EGFP) expression. One or two days after gene transfection, cells were subjected to drug selective medium, which was additionally supplemented with 400 μg/ml G418 (Invitrogen). Individual colonies appeared after 2–3 weeks and were picked by micropipette controlled by a three-dimensional controller. Resuspended colonies were transferred into new T25 flasks. For controls, 10 μg pcDNA-EGFP was transfected into COS cells.
Eggs were windowed, and after removal of the extraembryonic membranes with a tungsten needle, concentrated SDF-1-EGFP–transfected COS-1 cell suspensions were injected proximally into the developing hindlimb bud of stage HH25–26 chicken embryos using a capillary glass needle. EGFP transfected COS-1 cells were injected as controls.
CXCR4 Inhibitors (T140/TN14003) and Beads Implantation In Ovo
Highly specific peptidic inhibitors of CXCR4 (T140/TN14003), obtained as a gift from the group of Hirokazu Tamamura (Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Chiyoda-ku, Tokyo, 101-0062, Japan), were dissolved in PBS to a concentration 15 mg/ml (Yusuf et al.,2006). For application of inhibitor, the acrylic beads approximately 80–120 μm in diameter (Bio-Rad Laboratories) were rinsed in PBS and incubated with CXCR4 inhibitor (T140/TN14003) solution overnight at 4 °C. Control beads were incubated in PBS. For bead implantation, extraembryonic membranes (stage HH25–26) were removed with a sharpened tungsten needle, and a bead was inserted into the ventral proximal side of the hindlimb mesenchyme using a blunt glass needle. Embryos were reincubated to reach stage HH29–HH30. PBS-soaked beads were used as controls. All embryos were fixed in 4% paraformaldehyde (PFA; VWR International) and further submitted for in situ hybridization.
In Situ Hybridization
Selected embryos were tested for CXCR4, Pax7, and MyoD using highly specific RNA probes. The CXCR4 probe used was a 693-bp chicken probe. The Pax7 probe used was a 582-bp chicken probe. The MyoD probe was a 1,518-bp specific chicken probe. Riboprobes were labeled with a digoxigenin RNA labeling kit from Boehringer, Mannheim, Germany. Whole-mount in situ hybridization was performed as described previously (Nieto et al.,1996).
Immunohistochemistry on Paraffin Sections of Mice and Chicken Embryos
For immunohistochemistry, histological sections of chicken and mouse embryos were fixed overnight in 4% PFA (VWR International) in PBS. Nonspecific binding sites were blocked with 1% bovin serum albumin (BSA, Sigma) in PBS for 30 min. Anti-CXCR4 antibody (ab2074, Abcam; 1:200 in BSA) and anti-Pax7 antibody (Hybridoma bank; 1:50 in PBS) were diluted in 1% BSA (Sigma) in PBS. Sections were incubated with the primary antibodies overnight in a humid chamber at 4°C. Primary antibodies were detected by incubating sections for 2 hr at room temperature with either fluorescein isothiocyanate- (FITC) or Cy3-labeled secondary antibodies (Cy3-conjugated goat anti-mouse IgG antibody; Jackson ImmunoResearch, 1:100 in PBS). Subsequently, sections were washed in PBS, embedded in Aquatex (Mounting agent, Merk), and analyzed with an epifluorescence microscope (Axioscope 20; Zeiss).
Immunohistochemistry on Hindlimb in Collagen Gel Cultures
Immunohistochemistry was performed on hindlimbs cultured in collagen gels following fixation in 4% PFA (8.18715, Merck) in PBS for 2 hr in well plates. Collagen gels with operated hindlimb were washed with PBS for overnight. Nonspecific binding sites were blocked with 5% lamb serum (LS, 16070096 Invitrogen), 20% dimethyl sulfoxide (DMSO; D8418, Sigma) in PBS for 1 hr. Anti-CXCR4 antibody (ab2074, Abcam; 1:200 in BSA) and anti-Pax7 antibody (Hybridoma Bank; 1:50 in BSA) were diluted in 5% lamb serum (LS, Sigma), 20% DMSO (D8418, Sigma) in PBS. The gels containing the hindlimbs were then incubated with the primary antibodies for 2 days in a humid chamber at 4°C and washed again with PBS overnight. Primary antibodies were detected by incubating gels overnight at room temperature in dark with either FITC- or Cy3-labeled secondary antibodies (Cy3-conjugated goat anti-mouse IgG antibody; Jackson ImmunoResearch, 1:100 in PBS). Before mounting the collagen gels on slides, we prepared spacers from plasticine to hold the cover slips. Coverslips with collagen gels were carefully mounted in 90% glycerol (G7757, Sigma) in PBS.
Cell Death Test
Cell death test was performed after the implantation of T140/TN14003-soaked beads, using Nile Blue sulfate (van den Eijnde et al.,1997). After removal of the extra embryonic membranes, the embryos were stained in toto for 30 min at 37°C in a Nile Blue sulfate solution (1:20,000 w/v). After incubation, the specimens were washed twice with PBS, put on ice, and immediately photographed under a microscope equipped with a digital camera connected to a computer (Leica DFG320).
Microscopy and Imaging
After in situ hybridization, samples were observed and photographed using the Leica MZFLIII microscope and Leica DC 300F digital camera. Hybridized sections were observed and photographed using an Axioscope 20 (Zeiss) and a Leica DFC320 digital camera.
We thank Prof. Dr. Carmen Birchmeier (Max-Delbrück-Center for Molecular Medicine, Berlin) for providing the COS-1 cells, Prof. Dr. Erez Raz (Center for Molecular Biology of Inflammation, Münster) for kindly providing the CXCR4 chicken probe and Hirokazu Tamamura (Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University) for sharing the CXCR4 inhibitor. We also thank Prof. Dr. Rujin Huang and Susanne Theis (Bonn) and Prof. Dr. B. Heimrich (Freiburg). All authors extend their thanks to Ellen Gimbel, Susanna Glaser, and Ulrike Pein for their excellent technical assistances. We also acknowledge the support of “Molecular mechanisms of migration, invasion and metastasis,” a Baden-Württemberg grant, the Myores Project (511978) funded by the EU's sixth framework Programme and GKR 1104.