During early cardiac morphogenesis at looped heart stages, some of the endothelial cells constituting the atrioventricular (AV) and outflow tract (OT) regions of the single heart tube change phenotype to that of mesenchymal cells and migrate into the subjacent thick acellular extracellular matrix (cardiac jelly; Davis, 1927) to form endocardial cushion tissue, primordia of valves, and septa of the adult heart. During this endothelial–mesenchymal transformation (EMT), endothelial cells show sequential cellular phenotypic changes, including endothelial hypertrophy, polarization of the Golgi apparatus, loss of cell–cell contact, formation of migratory appendages, and invasion into the cardiac jelly (Markwald et al., 1975, 1977). This embryonic phenomenon is regulated by myocardially emitted inductive signals, which include bone morphogenetic protein, transforming growth factor-β, and other unknown inductive molecules (Mjaatvedt and Markwald, 1989; Mjaatvedt et al., 1991; Potts et al., 1991; Nakajima et al., 1994, 1998; Yamagishi et al., 1999; Sinning and Mckay, 2004). It is also reported that pertussis toxin, a specific inhibitor for either GTP-binding proteins or small GTP-binding proteins (Ueda et al., 2001), inhibits EMT (Boyer et al., 1999). This finding suggests that one or more GTP-binding proteins are involved in signal transduction pathways during EMT.
GTP-binding proteins are signal-transducing proteins that couple a large number of membrane-bound receptors regulating a variety of intracellular effector systems. Rho is one of the small GTP-binding proteins and cycles between GDP-bound inactive and GTP-bound active forms as an intracellular molecular switch (Aelst and D'Souza-Schorey, 1997; Aelst and Symons, 2002). The processes of the cycling between GTP-Rho and GDP-Rho are regulated by guanine nucleotide exchange factors (GEFs), GTPase-activating proteins (GAPs), and guanine nucleotide-dissociated inhibitors (GDIs). GTP-Rho activates downstream effectors, Rho-associated coiled-coil kinases (ROCKs), p160ROCK/ROCK1/ROKβ, and Rho kinase/ROCK2/ROKα (Matsui et al., 1996; Nakagawa et al., 1996; Bishop and Hall, 2000). ROCKs are serine/threonine protein kinases implicated in the regulation of cytoskeletal reorganization, such as stress-fiber formation in nonmuscle cells (Amano et al., 1997). The Rho–ROCK pathway also regulates the phosphorylation of myosin light chain (MLC) to promote contraction of actomyosin bundles by the direct phosphorylation of MLC and by inactivation of myosin phosphatase (Amano et al., 1996; Fukata et al., 2001).
During vertebrate development, Rho–ROCK pathways are important for various processes, including cardiogenesis and neural development, through effects on the actin cytoskeletal reorganization that controls cell migration and differentiation (Wunnenberg-Stapleton et al., 1999; Henderson et al., 2000; Wei et al., 2001, 2002; Zhao and Rivkees, 2003, 2004). Recent experiments showed that Rho–ROCK pathways are involved in the early development of the heart and that ROCK-inhibitor blocks either the fusion of bilateral heart mesoderm or the formation of appropriate heart chambers (Wei et al., 2001; Zhao and Rivkees, 2003; Kaarbo et al., 2003). Zhao and Rivkees (2004) reported that ROCKs play a role in endothelial cell differentiation and migration during the formation of endocardial cushion tissue in mice. However, the spatiotemporal expression pattern of ROCK protein during cardiogenesis and how ROCKs control endocardial cell differentiation and migration are still unknown. Using immunohistochemical staining for ROCK1 and -2, and a three-dimensional collagen gel culture assay, we showed that transforming endothelial and migrating mesenchymal cells express ROCK1 and -2, and that a ROCK inhibitor, Y27632, inhibited the establishment of endothelial cell polarity and mesenchymal invasion during the EMT.
ROCK1 and -2 Are Expressed in the Developing Heart
To investigate the spatiotemporal expression of ROCKs during heart development, embryonic hearts from stage 10 to 23 were examined immunohistochemically with antibodies against ROCK1 and ROCK2. Before stage 13, anti-ROCK1 and ROCK2 immunoreactivity was observed in the myocardium but not in the endocardium (data not shown). At stage 14, hypertrophied AV endothelial cells, presumably transforming cells, showed anti-ROCK1 and -2 immunoreactivity in their cytoplasm (arrowheads in Fig. 1a,b,o,p). Myocardial cells expressed ROCKs extensively (M in Fig. 1a,b,o,p). At stage 15–16, endothelial cells in the OT and AV regions begin to transform into mesenchyme and invade into the cardiac jelly. These transforming/invading endothelial cells expressed ROCK1 and -2 (arrowheads in Fig. 1c–e,q,r). At around stage 18–23, endothelial cells within the OT and AV regions, but not the ventricular region, normally undergo EMT and migrate into the cardiac jelly. Transforming endothelial cells and migrating mesenchymal cells possessed anti-ROCK1 and -2 immunoreactivity in their cytoplasm (arrowheads in Fig. 1f–j,s–v). In contrast, ventricular endothelial cells, in which EMT does not occur, did not show apparent staining of ROCK1 and -2 (arrows in Fig. 1m,n,x,y). Myocardium in the ventricular region begins to elicit ventricular trabeculation, in which anti-ROCK1 immunoreactivity appeared to be down-regulated (M in Fig. 1f–j), whereas the inner myocardial layer subjacent to the endocardial cushion tissue expressed ROCK1 (arrows in Fig. 1f–j). ROCK2 was expressed uniformly in the myocardial layer (Fig. 1s–v). In addition to the heart, developing somites as well as migrating neural crest cells expressed ROCK1 and -2 (Fig. 1k,l,w). To reconfirm the expression of ROCK1 and -2 during chick heart development, we performed an immunoblot analysis of ROCK1 and -2 in embryonic hearts from stage 14 to 23 (Fig. 2a). Results showed that a single band with a molecular mass of 160 kDa (ROCK1) or 180 kDa (ROCK2) was detectable in the whole hearts examined. The relative amounts of protein/β actin were calculated and revealed that ROCK1 and -2 were expressed extensively in stage 16–18 hearts, in which there is marked EMT (Fig. 2b). Immunohistochemistry in developing chick embryos indicated that ROCKs were expressed in the OT/AV endocardial cushion tissue, migrating neural crest cells, and delaminating somites, in which there is extensive epithelial–mesenchymal transformation. These results suggest that ROCKs play an important role in the regulation of valvuloseptal EMT as well as epithelial–mesenchymal transformation during early organogenesis/embryogenesis.
ROCK Inhibition Blocks Mesenchymal Invasion in Cultured AV Explants
The immunolocalization of ROCKs in the developing heart suggested an important role for these kinases in the cellular changes that occur during EMT. We next examined the kind of cellular changes that occurred in stage 14-minus AV explants, when the ROCK activity was inhibited by Y27632, a ROCK-specific inhibitor. AV explants (endocardium + myocardium) were collected from stage 14-minus hearts and cultured on a three-dimensional collagen gel lattice with or without Y27632. At this stage, AV endothelial cells are not activated for EMT, as the endocardium alone will retain the epithelial phenotype without transforming factors, such as AV myocardium and embryonic cardiocyte-conditioned medium (CCM; Krug et al., 1987). After 48 hr, explants were observed under Hoffman modulation optics and mesenchymal cells that had invaded the gel lattice were counted by optical sectioning. In explants cultured with CM199 (control culture), endothelial outgrowth was seen on the gel surface and many mesenchymal cells were found in the gel lattice (approximately 60–70 mesenchymal cells/explant invaded the collagen gel lattice, Fig. 3a,b,g). In contrast, when explants were cultured in CM199 supplemented with Y27632 (5 μg/ml or 50 μg/ml), the number of mesenchymal cells in the gel lattice was significantly reduced (approximately 15 or 5 mesenchymal cells/explant at each concentration, P < 0.01, Fig. 3c,d,g). Endothelial outgrowth was not inhibited in cultures treated with 5 μg/ml of Y27632. However, endothelial outgrowth in explants treated with 50 μg/ml of Y27632 appeared to be affected slightly. Results indicated that Y27632 (ROCK inhibitor) blocked the invasion of mesenchymal cells that were seeded from AV endocardium co-cultured with associated myocardium.
ROCKs trigger the phosphorylation of myosin light chain, which in turn initiates the contraction of actomyosin bundles to generate the driving-force for cellular phenotypic change and locomotion (Amano et al., 1996; Fukata et al., 2001; Aelst and Symons, 2002). We next examined whether ML-9, a selective myosin light chain kinase inhibitor, inhibited mesenchymal cell invasion in cultured AV explants. As shown in Figure 3, stage 14-minus AV explants treated with ML-9 at a concentration of 1–2 μg/ml seeded few invasive mesenchymal cells in comparison with control cultures (Fig. 3e,f,h). To confirm the amounts of phosphorylated myosin light chain in AV explants that had been treated with Y27632 or ML-9, we performed an immunoblot analysis of myosin light chain and phospho-myosin light chain. As shown in Figure 3i, the amount of phospho-myosin light chain was reduced in AV explants treated with Y27632 or ML-9. The results indicated that inhibition of the downstream region of the ROCK pathway perturbed the mesenchymal cell invasion in a similar manner to that observed in cultured AV explants treated with Y27632. Taken together, these results suggest that the Rho–ROCK pathway followed by myosin light chain phosphorylation is involved in the mesenchymal cell invasion during EMT.
ROCK Inhibition Does Not Suppress the Expression of Transformation Markers
It has been reported that transforming AV endothelial cells express several differentiation markers, such as αSMA (Smooth muscle α-actin), Msx-1, JB3/fibrillin-2, and M38/type I procollagen (Sinning et al., 1988; Chan-Thomas et al., 1993; Wunsch et al., 1994; Nakajima et al., 1997). We previously reported that inhibition of αSMA expression perturbs EMT in cultured AV explants (Nakajima et al., 1999). Therefore, we first examined whether the expression of αSMA was affected by Y27632 in cultured AV explants. Stage 14-minus AV explants were cultured on a collagen gel lattice with or without Y27632. After 48 hr in culture, explants were fixed and stained with anti-αSMA antibody or subjected to immunoblot analysis. As reported previously (Nakajima et al., 1997), explants that were cultured in CM199 seeded many mesenchymal cells expressing αSMA (arrowheads in Fig. 4b). Endothelial cells maintaining the epithelial phenotype did not express αSMA (Fig. 4a). Transforming endothelial cells on the gel surface expressed αSMA (arrowhead in Fig. 4a). In contrast, AV explants treated with Y27632 seeded few mesenchymal cells. Some cells on the gel surface, which showed a hypertrophied polygonal shape with thin appendages, expressed αSMA (arrows Fig. 4c). Next, we performed an immunoblot analysis of αSMA expression after the explant experiment described above. After 48 hr in culture, the myocardium was removed with microforceps, and the resulting endothelial/mesenchymal cells were solubilized in sodium dodecyl sulfate (SDS) -sample buffer and subjected to immunoblotting (Fig. 4e,f). Results showed that the amount of αSMA protein was not reduced in endothelial/mesenchymal tissue even on treatment with Y27632. In contrast to this result, mouse embryonic AV endothelial cells, which were cultured on coverslips and treated with Y27632, showed a reduced expression of αSMA (Zhao and Rivkees, 2004). The difference between our results and those of Zhao and Rivkees may be attributed to the methodology (collagen gel culture vs. coverslip) and/or species (chick vs. mouse).
We next examined whether Y27632 affected the expression of other mesenchymal cell markers, JB3/fibrillin-2 and M38/type I procollagen. After the explant experiments, cultures were fixed and stained with JB3 or M38. As shown in Figure 5a,b, transforming endothelial/mesenchymal cells from the control AV explants expressed fibrillin-2 and type I procollagen. In explants treated with Y27632, mesenchymal cells did not seed; however, some cells on the gel lattice expressed fibrillin-2 and type I procollagen (Fig. 5c,d). A semiquantitative analysis (incidence of JB3/fibrillin-2 or M38/type I procollagen-positive cells) showed that the number of cells expressing JB3 or M38 was identical between control explants and explants treated with Y27632 (Fig. 5e). The results indicated that, although Y27632 affected the mesenchymal cell invasion into the gel lattice, it had negligible effects on the expression of the differentiation markers tested. These results suggested that the Rho–ROCK pathway appears to regulate the migration or invasion that occurs at the late onset of EMT.
ROCK Inhibition Blocks Mesenchymal Invasion in Cultured Stage 14–18 AV Explants
The above experiments suggested that the inhibition of ROCK by Y27632 perturbed the mesenchymal cell invasion rather than early transformation processes, such as cell–cell separation and the expression of early transformation markers. To confirm this schema, we next examined whether Y27632 blocked the mesenchymal cell invasion from stage 14–18 AV explants, in which endothelial cells are already committed to undergo EMT in a cell autonomous manner (Ramsdell and Markwald, 1997). Stage 14, 16, and 18 AV explants were cultured on a collagen gel lattice and treated with Y27632, and the resulting cultures were analyzed for mesenchymal invasion after 48 hr incubation. AV explants cultured in CM199 alone showed endothelial outgrowth and the seeding of numerous mesenchymal cells into the gel lattice (Fig. 6a,b,e). In explants treated with Y27632, the number of mesenchymal cells in the gel lattice was reduced significantly in all explants tested (Fig. 6c–e). The result indicated that Y27632 inhibited mesenchymal invasion in stage 14–18 AV explants, in which endothelial cells are already committed to undergo EMT, supporting the schema that Y27632 inhibited the process responsible for mesenchymal invasion rather than activation process that occurs at the onset of EMT.
ROCK Inhibition in Endothelial Cells Blocks Mesenchymal Invasion
The above experiments showed that ROCK inhibition affected mesenchymal invasion into the gel lattice. However, it is unclear whether Y27632 acted directly on the endothelial cells, as AV explants contain myocardium expressing ROCKs (Fig. 1). To address this issue, we performed an AV endothelial activation assay using embryonic CCM, which substituted for the AV myocardium to initiate EMT in culture (Krug et al., 1987). Stage 14-minus preactivated endothelial monolayers were prepared on a collagen gel lattice. The resulting cultures were supplemented with CCM, CCM + Y27632, or CCM + ML-9. After 36- to 48-hr incubation, cultures were assessed for EMT. As described previously (Krug et al., 1987), AV endothelial cells treated with CCM seeded mesenchymal cells into the gel lattice. In contrast, endothelial cells treated with CCM + Y27632 or CCM + ML-9 seeded few mesenchymal cells in comparison with control cultures (Fig. 7).
ROCKs Are Expressed in Regions Where Epithelial-Mesenchymal Transformation Occurs
During the formation of endocardial cushion tissue, endocardial cells in OT and AV regions transform into mesenchymal cells that migrate into adjacent cardiac jelly to form valvuloseptal endocardial cushion tissue (Markwald et al., 1975, 1977). At the onset of EMT, ROCK1 and -2 were first expressed in the transforming endothelial cells and their expression was maintained in the migrating mesenchymal cells. In addition to the heart, ROCKs were expressed either in migrating neural crest cells or in delaminating somites. Among intracellular signaling mediators, Rho-GTPases and their effectors have been found to be essential for many aspects of embryogenesis, such as neural crest cell delamination, epicardial–mesenchymal transformation during coronary vessel formation, and convergent extension movements during gastrulation (Liu and Jessell, 1998; Lu et al., 2001; Wei et al., 2001; Kim and Han, 2005). During mouse embryogenesis, both rho-B and rock1 are expressed in the endocardial cushion tissue as well as in the developing myocardium (Henderson et al., 2000; Wei et al., 2001). Our preliminary immunohistochemistry showed that Rho-B was expressed in regions where there is extensive epithelial–mesenchymal transition, such as in the endocardial cushion tissue, migrating neural crest, and delaminating somites during early chick embryogenesis (data not shown). In addition, premigratory and delaminating neural crest cells express rho-B and the zinc finger transcription factor slug during chick embryogenesis (Liu and Jessel, 1998). These findings suggest that, ROCKs, downstream effectors of Rho, play an important role in the regulation of valvuloseptal EMT as well as epithelial–mesenchymal transformation during early organogenesis.
ROCK Is Required for Mesenchymal Invasion During the Formation of Endocardial Cushion Tissue
Although the inhibition of ROCK by Y27632 did not suppress the expression of early transformation markers for cushion mesenchyme, mesenchymal invasion from AV endocardium was affected. In addition, Y27632 perturbed the seeding from stage 16–18 AV explants, in which endothelial cells had already committed to undergo EMT. Furthermore, ML-9, a myosin light chain kinase inhibitor, inhibited the mesenchymal cell invasion in a manner similar to that observed in AV explants treated with Y27632. ROCKs are thought to be involved in a wide variety of cell activities, including cytoskeletal reorganization, cell migration, cell adhesion, and non–muscle cell contraction (Kawada et al., 1999; Katoh et al., 2001). During EMT, some of the endothelial cells change their epithelial phenotype to that of mesenchyme. These changes include cellular hypertrophy, loss of cell–cell contact, polarization of Golgi, migratory appendage formation, and invasion into the cardiac jelly (Eisenberg and Markwald, 1995). At the onset of this embryonic phenomenon, transforming endothelial/mesenchymal cells express αSMA and the perturbation of αSMA-expression inhibits the mesenchymal invasion in cultured AV explants (Nakajima et al., 1997, 1999). The expression of αSMA at the onset of EMT is essential for the transformation processes, including cytoskeletal reorganization, migratory appendage formation, and cellular invasion (Nakajima et al., 1999). It is widely accepted that the major driving force of cell-migration is the extension of a leading edge protrusion or lamellipodium, the establishment of new adhesion sites at the front, cell body contraction, and detachment of adhesions at the rear edge. All these processes involve the assembly, disassembly, or reorganization of the actin cytoskeleton (Ridley, 2001; Raftopoulou and Hall, 2004). One important target of ROCK1 is phosphorylation of the myosin light chain involved in stimulating the assembly of actin-myosin filaments and, therefore, contractility to generate the driving force. Other experiments showed that ROCKs phosphorylate and activate LIM-kinase, which in turn phosphorylates and inactivates cofilin, leading to the stabilization of actin filaments within actomyosin filament bundles (Maekawa et al., 1999; Sumi et al., 2001). These findings together with our results strongly suggest that ROCKs are required to generate the driving force for the mesenchymal cell invasion/migration that occurs at the late onset of EMT.
Before the onset of mesenchymal cell migration, cells require the establishment of cellular polarity, i.e., leading and rear edges. In our experiments, the inhibition of ROCK affected the formation of cellular polarity in endothelial cells, which were expressing early transformation markers and showing a star-like phenotype with thin appendages. Zhao and Rivkees (2004) reported that endothelial cells cultured on coverslips and treated with Y27632 formed thin processes instead of a well-developed lamellipodium. The signaling pathway, which regulates such planar polarity, a cellular axis perpendicular to the apical–basal axis of the cell, is not well understood. Recently, it has been reported that the Wnt/Frizzeled pathway regulates planar cell polarity by means of Rho/ROCK as well as the Jun kinase pathway (called the PCP pathway; Habas et al., 2001). During the formation of endocardial cushion tissue, expression of Wnt6 is observed in the myocardium of the AV canal (Schubert et al., 2002) and Wnt9a in AV endothelial cells (Person et al., 2005). Moreover, mice deficient in endocardial β-catenin show a lack of endocardial cushion tissue (Liebner et al., 2004). Therefore, it is strongly suggested that, in addition to the Wnt–canonical pathway (Hurlstone et al., 2003), the Wnt/Rho/ROCK pathway (PCP pathway) may have an important role in the regulation of endocardial differentiation and invasion.
In conclusion, we showed that the Rho–ROCK pathway has an important role in the regulation of endocardial EMT and that ROCK appears to be involved in the endothelially derived mesenchymal cell invasion and migration that occur at the late onset of EMT. Further study is necessary to elucidate the relation between the Rho–ROCK pathway and other signaling pathways regulating EMT.
Fertilized eggs from the domestic fowl (Gallus gallus) were incubated for 2–4 days at 37.8°C and 90% humidity. Embryos were collected on ice-cooled phosphate-buffered saline (PBS) and staged according to the criteria of Hamburger and Hamilton (1951). The staged embryos were subjected to the experiments described below.
Gel Electrophoresis and Immunoblotting
Embryonic hearts or cultured cells were homogenized in lysis buffer (62.5 mM Tris, 0.1% glycerol, 2% SDS, and 5% 2-mercaptoethanol, pH 6.8). After heat-denaturation (95°C for 5 min), equal amounts of protein were subjected to SDS-polyacrylamide gel electrophoresis (7.5 and 10% polyacrylamide), and then transferred to Immunobilon-P membrane (Millipore). The membrane was subsequently treated with 5% nonfat dry milk and was incubated with the primary antibody against ROCK1-RB (Amano et al., 1997), ROCK2 (Santa Cruz Biotechnology, 1:1,000), αSMA (smooth muscle α-actin, Sigma, diluted 1:1,000), β-actin (Sigma, 1:1,000), total actin (Sigma, 1:1,000), myosin light chain-20kDa (Sigma, 1:200), phospho-myosin light chain (Cell Signaling, 1:1,000), or GAPDH (glyceldehyde-3-phosphate dehydrogenase, CHEMICON, 1:1,000) for 2 hr at room temperature. After extensive washing, the membrane was incubated with peroxidase-conjugated secondary antibody for 2 hr. Immunoreactive bands were visualized using ECL detection reagent (Amersham).
Three-Dimensional Collagen Gel Culture
Hydrated collagen gels (1 mg/ml, type I rat-tail collagen; Becton Dickinson) were prepared in four-well dishes (Nunc) as described by Bernanke and Markwald (1982). The AV region was resected from stage 14-minus chick embryonic heart, then cut longitudinally. The resulting AV explants were placed on a collagen gel lattice saturated with complete medium (CM199; medium 199 containing 1% chick serum, 5 μg/ml insulin, 5 μg/ml transferring, 5 ng/ml selenium, and streptomycin/penicillin; ITS, Becton Dickinson), and left for 6 hr to adhere to the gel surface. After that time, the cultures were subjected to various test conditions: they were supplemented with 300 μl of the following: (1) CM199 alone, (2) CM199 containing Y27632 (Yoshitomi Pharmaceutical Industries), and (3) CM199 containing ML-9 (BIOMOL). After a total of 48 hr incubation, cultures were assessed for endothelial cell proliferation and evidence of EMT under a microscope equipped with Hoffman modulation optics or a phase contrast microscope. Subsequently, the cultures were subjected to immunological detection of αSMA, fibrillin-2, and type I procollagen, as described below.
The AV endothelial activation assay was performed as described by Krug et al. (1987). Stage 14-minus AV explants were cultured on a collagen gel lattice. After 4 hr incubation, myocardial layers were removed using microforceps. The resulting endothelial monolayers were cultured in embryonic cardiocyte-conditioned medium (CCM; Krug et al., 1987), CCM + Y27632, or CCM + ML-9. CCM contains myocardially derived inductive molecules to induce EMT. After a total of 36 or 48 hr incubation, cultures were assessed for EMT.
Indirect Immunofluorescence Microscopy
Embryos were equilibrated in a graded series of sucrose solutions (10–20%, wt/vol) in PBS at 4°C for 12 hr, embedded in OCT compound (Sakura), and frozen in liquid nitrogen. Frozen sections (8 μm) were cut on a cryostat, mounted on slides with 3-aminopropyltriethoxysilane, and air-dried. After rinsing in PBS for 15 min, sections were blocked with 1% bovine serum albumin (BSA) in PBS for 1 hr and incubated with anti-ROCK1 (mouse monoclonal, diluted 1:100 in blocking solution, Becton Dickinson) or ROCK2 (goat polyclonal, diluted 1:100 in blocking solution, Santa Cruz Biotechnology) antibody in a moist chamber overnight at 4°C, rinsed in PBS, incubated with fluorescein-5-isothiocyanate (FITC) -conjugated secondary antibody (goat antibody against mouse IgG [Cappel] or donkey antibody against goat IgG [CHEMICON]) for 1.5 hr at room temperature, rinsed with PBS, and coverslipped using mounting medium (5% triethylenediamine in 10% PBS/90% glycerol). Samples were observed under a conventional fluorescence microscope (BX50-FLA, Olympus), and images were recorded by CCD camera. The exposure time for FITC images was 1.8 sec.
Cultures were drained of medium, rinsed with PBS, and fixed in 4% paraformaldehyde in PBS for 1 hr at room temperature. After rinsing in PBS for 15 min, cultures were blocked with 1% BSA in PBS containing 0.1% Triton X-100 for 1 hr, then incubated with anti-αSMA antibody (mouse monoclonal diluted 1:400 in blocking solution, Sigma) in a moist chamber for 2 hr at room temperature. After rinsing with PBS, cultures were incubated in FITC goat anti-mouse IgG (diluted 1:100 in blocking solution) for 2 hr. For fibrillin-2 or type I procollagen staining, cultures were rinsed with PBS and fixed in 70% ethanol for 30 min followed by absolute ethanol for 1 hr at room temperature with constant agitation. Cultures were rehydrated in PBS, allowed to equilibrate overnight at 4°C, blocked for 1 hr at room temperature with PBS/BSA, and finally incubated with JB3/anti–fibrillin-2 (kindly provided by Dr. Isokawa) or M38/anti-type I procollagen (Developmental studies Hybridoma Bank, University of Iowa) for 2 hr at room temperature. After extensive washing in PBS, cultures were incubated in FITC goat anti-mouse IgG for 2 hr. After rinsing with PBS, cultures were incubated in DAPI (4,6-diamidino-2-phenylindole dihydrochloride, diluted 1:200), then transferred to slides and cover-slipped. Samples were observed under a conventional fluorescence microscope or laser confocal microscope (Zeiss). Semiquantitative analysis of fibrillin-2 or type I procollagen expression was performed using the following formula: percentage incidence of JB3- or M38-positive cells (%) = (number of JB3- or M38-positive cells/DAPI-positive cells) × 100.
The authors thank Dr. K. Kaibuchi for providing anti-ROCK1-RB, Dr. K. Isokawa for JB3, and Dr. M. Uehara for Y27632. M38 was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD, and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242. Y.N. was funded by the Ministry of Education, Science and Culture of Japan, The Japan Cardiovascular Research Foundation, The Takeda Science Foundation, and the Terumo Life Science Foundation.