SEARCH

SEARCH BY CITATION

Keywords:

  • endocardial cushion;
  • valvulogenesis;
  • periostin;
  • OSF-2;
  • fasciclin;
  • epithelial-to-mesenchymal transition;
  • chordae tendenae

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. Acknowledgements
  6. LITERATURE CITED

The process that cardiac cushions undergo to form the mature septa and valves of the adult heart is poorly understood. Periostin is an extracellular molecule that is expressed during cushion mesenchyme formation and throughout valvulogenesis. Once thought to be an osteoblast-specific factor, studies have shown this molecule is antiosteogenic. We have produced an antibody to chicken periostin and examined periostin's localization in the developing avian heart. This antibody recognized proteins from chick heart lysates around 90 kD molecular weight as predicted from the chick periostin mRNA and other periostin orthologs. Periostin immunolocalization was first evident as fibrous strands in the cushion mesenchyme. At HH25, periostin was detected on the basal surface of the trabecular endothelium and also on the endocardial epithelium of the atrioventricular cushion. We hypothesize that periostin may function in the organization of extracellular matrix molecules, providing cues necessary for attachment and spreading during the epithelial-to-mesenchymal transitions of the endocardial epithelium. Enhanced secretion of periostin in the region of delamination may directly or indirectly promote change in the myocardium that precedes or mediates delamination of the leaflet. At later stages of development (HH34-38), periostin was seen predominantly in the fibrous regions of the heart, such as the left atrioventricular valve (LAV), annulus, cardiac skeleton, and adventitia. We propose that periostin is induced by sheer stress and may be an essential molecular component for structures of the heart that undergo mechanical stress or tension during the cardiac cycle. © 2005 Wiley-Liss, Inc.

It has been well established that the cardiac cushions are mesenchymal swellings that form within the wall of the developing heart and are the primordia for the septa and valves of the mature heart (de la Cruz et al., 1983; van den Hoff et al., 1999; Schroeder et al., 2003). However, the remodeling of cushion tissue to form the mature structures of the valve leaflets is poorly understood. Matrix macromolecules such as periostin, secreted during cushion mesenchyme formation, likely play a critical role in the process of expansion and differentiation of the leaflet primordia into the suspensory chordae tendinae and leaflet of the atrioventricular (AV) valve.

The mammalian periostin gene encodes a 90 kD secreted protein, which shares significant homology with the insect protein fasciclin I. Drosophila fasciclin has been shown to regulate pattern formation and differentiation through adhesive and motility guidance mechanisms in invertebrates (Zinn et al., 1988). In zebrafish, periostin is required for the adhesion of muscle fibers to the myoseptum (Kudo et al., 2004). Adhesion is mediated through the four fasciclin I-like repeats, each of which also contains two highly conserved sequences (H1 and H2) in the periostin molecule (Kawamoto et al., 1998; Norris et al., 2004). Although these repeats lack a typical RGD-integrin binding motif, they are still capable of interacting with a variety of integrins and mediating cell adhesion (Gillan et al., 2002; Bao et al., 2004; Shao et al., 2004).

In mammals, periostin was originally identified in a subtraction hybridization differential cDNA screen and referred to as OSF-2 (Takeshita et al., 1993; Horiuchi et al., 1999). Initially thought to be bone-specific and required for osteoblast differentiation, subsequent experiments revealed that OSF-2 expression diminished as osteoblasts became committed to the bone lineage. Additional studies went on to show the antiosteogenic function of this molecule (Hall and Miyake, 2000; Ji et al., 2000; Saito et al., 2002). OSF-2 was found to be strongly expressed in the periodontal ligament (PDL), the periosteium, endocardial cushions and valves of the developing heart and was renamed periostin (Horiuchi et al., 1999; Kruzynska-Frejtag et al., 2001; Wilde et al., 2003).

Periostin expression is significantly increased in response to BMP and TGF-β growth factor signaling in mesenchymal cells undergoing differentiation (Horiuchi et al., 1999; Ji et al., 2000). In the heart, BMP and TGF-β are secreted by the myocardium and stimulate the endocardial cushions during growth and development (Eisenberg and Markwald, 1995; Nakajima et al., 1997, 2000; Sugi et al., 2004). Recently, it has been shown that periostin is induced as a result of stress responses in the heart and periodontal ligament (Stanton et al., 2000; Wang et al., 2003; Wilde et al., 2003; Katsuragi et al., 2004; Li et al., 2004). The precise role of periostin in heart development is unknown. In situ hybridization experiments have localized periostin mRNA to the cardiac cushions of the developing mouse and chick heart (Kruzynska-Frejtag et al., 2001; Norris et al., 2004). This pattern of periostin gene expression suggests an intriguing role for periostin protein in valvulogenesis. Here we report the production and characterization of an antibody that identifies chick periostin isoforms and describe the localization of periostin protein in the developing chick heart.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. Acknowledgements
  6. LITERATURE CITED

Production of Antichicken Periostin Antibody

For use in the preparation of antibodies against chick periostin, the peptide KLREEIEGRGSFTFFAPSN (representing amino acids 122–140 of chick periostin) was synthesized and coupled to KLH at the Emory University Microchemical Facility. The peptide was synthesized with an additional C-residue at the N-terminus to facilitate its attachment to Sulfolink Coupling Gel (Pierce, Rockford, IL). The injection of rabbits, the purification of total IgG, and the affinity purification of IgG using the peptide attached to Sulfolink Coupling Gel were all performed as previously described (Zanin et al., 1999).

Western Blotting

Tissues to be analyzed were homogenized in at least five volumes of 2% SDS/25 mM Tris (pH 8.0) containing a cocktail of protease inhibitors (Zanin et al., 1999) and sonicated for 30 sec using a Fisher sonic dismembrator Model 300. The homogenates were then clarified by centrifugation at 14,000 g for 5 min and their protein concentration was determined using the Micro BCA Protein Assay (Pierce). Equal amounts of protein reduced with β-mercaptoethanol were separated by SDS-PAGE. The gel was transblotted and the transfer incubated overnight with primary antibody (0.15 μg/ml). Secondary antibodies and detection procedures were as previously described (Zanin et al., 1999), using enhanced chemiluminescence to detect HRP-conjugated secondary antibody. For lane 4 of Figure 1, the primary antibody was preincubated with 0.5 μg/ml of the immunogenic peptide for 1 hr at room temperature before use on Western blots.

thumbnail image

Figure 1. Western analysis of the polyclonal chick periostin antibody. Twenty μg of SDS protein extracts of (lane 1) 4-day embryonic heart tissue, (lane 2) 6-day embryonic heart tissue, and (lanes 3 and 4) 10-day embryonic heart tissue were resolved by SDS-PAGE under reducing conditions and Western-blotted using antichicken periostin as primary antibody. For lane 4, the primary antibody was preincubated with the immunogenic peptide. The migration of the 97 kD standard protein is indicated.

Download figure to PowerPoint

Immunohistochemistry

Chick embryos were staged (Hamburger and Hamilton, 1951) and then fixed in 4% paraformaldehyde, Dent's fixative (80% methanol/20% DMSO) or Amsterdam Fix (35% methanol, 35% acetone, 25% dH20, and 5% acetic acid), embedded in paraffin, and sectioned at 5 μm. Deparaffinized sections were rehydrated through a graded series of ethanols to PBS. Sections were subjected to antigen unmasking (H-3300; Vector Laboratories, Burlingame, CA) based on a high-temperature citric acid formula. Next, sections were blocked 1 hr at room temperature with blocking buffer (PBS; Sigma; containing 5% normal goat serum; NGS; Cappel, Malvern, PA), and incubated with affinity-purified rabbit antichick periostin (4 μg/ml) in blocking buffer overnight at 4°C. Following primary antibody incubations, specimens were washed five times in PBS and incubated at room temperature with fluorescein-conjugated antirabbit IgG secondary antibody (Jackson Laboratories) diluted 1:100 in PBS. All samples were then washed extensively in PBS and coverslipped using Dabco mounting media (Sigma). For some sections, nuclei were labeled with propidium iodide (Molecular Probes C-7590) in PBS for 5 min prior to the final washes in PBS.

Controls for immunohistochemistry included omission of the primary antibody and preabsorption of the immunopeptide with the primary antibody prior to addition on the section. Immunostained sections were viewed with a Leica TCS SP2 AOBS Confocal Microscope System (Leica Microsystems, Inc Exton, PA). Files were transferred to Adobe Photoshop for labeling and figure preparation.

Whole Mount Immunostaining and Confocal Analysis

An incision was made in HH stage 29 embryos above the liver to expose the descending aorta and hepatic artery. The hearts were perfused with 1% lidocaine in PBS followed by 0.5 mg/ml FITC-poly-L-lysine (Sigma) in saline (5 min, RT), immersed in Dent's fixative, and rinsed in PBS/A (azide). Hearts were mounted onto a small silicone rubber support (170 silicone Eastover, Sylguard; Dow Chemical) for precise frontal viewing, immunostaining, and confocal analysis. For antibody staining (HH20, 29) antichick periostin (4 μg/ml) was used for overnight incubation (4°C). Hearts were extensively rinsed and secondary antibody was applied (1:100; overnight 4°C). Hearts were rinsed and then cleared for viewing with Murray's clear (50/50 benzyl benzoate/benzyl alcohol) and imaged on a Leica TCS SP2 AOBS confocal microscope using Leica Microsystems (Heidelberg, Germany) 1997–2002 version 2.0 0871 software. Hearts were imaged in precise frontal plane. Z-series data were collected using 25 μm steps.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. Acknowledgements
  6. LITERATURE CITED

We have produced a polyclonal antibody against a peptide that corresponds to a 20 amino acid sequence in a unique region of the chicken periostin molecule. A Blast search of this sequence showed identity with only periostin orthologs. The closest sequence similarity was found in another fasciclin family member, βigH3. The predicted molecular weight for βigH3 differs from periostin by approximately 20 kD. Western analysis of the chicken periostin antibody using heart extracts from day 4, 6, and 10 embryos showed a cluster of predominant bands around 90 kD (Fig. 1). This is consistent with the predicted molecular weight of periostin based on its cDNA sequence (Norris et al., 2004). Periostin is known in other organisms to form several different isoforms (Takeshita et al., 1993; Sasaki et al., 2001; Ito et al., 2002; Kudo et al., 2004; Litvin et al., 2004). The cluster of bands around 90 kD suggested that multiple periostin isoforms may also be present in the developing chick heart. Western analyses of unreduced gels have also shown reactive bands the appropriate size for multimers of periostin (data not shown).

Isoforms of periostin appear to differ in their C-terminus (Terasaka et al., 1989; Skonier et al., 1992; Huber and Sumper, 1994; LeBaron et al., 1995; Ulstrup et al., 1995; Horiuchi et al., 1999; Lal et al., 1999; Sasaki et al., 2001). Although the gene structure of the chick periostin has not been elucidated, preliminary data indicate that avian isoforms may also differ in the C terminus. The predicted amino acid sequence of the periostin cDNA sequence reported by Norris et al. (2004) is 41 amino acids longer at the C-terminal end than a previously reported chick periostin in the EST database. Others have shown by mutational analysis that the C-terminus of periostin is sufficient to suppress certain growth conditions in cancer cell lines (Yoshioka et al., 2002). Periostin isoform expression is also altered during osteogenesis, suggesting that isoform expression may play a role in growth and differentiation (Litvin et al., 2004). Periostin-like factor, which is an isoform of periostin, was isolated from mouse heart and differs from the isoform that was originally isolated from bone by exchanging small (27–28 amino acids) exon cassettes in the 3′ end. Others have proposed that due to the high conservation of these 3′ exon cassettes, it is likely that the different isoforms are functionally distinct (Litvin et al., 2004). There is also the possibility that the different forms of periostin are the result of posttranslational processing. In preliminary results, we have observed by Western blots a smaller band of approximately 50 kD (which was competed away with the immunizing peptide) that may represent a processed form of periostin. Also, when chick cardiomyocytes were infected with periostin antisense virus, the cluster of bands at 90 kD and a lower band were significantly suppressed (data not shown). In the heart, differential expression of periostin isoforms may more precisely modulate adhesion and migratory mechanisms that regulate morphogenetic interactions between cushion cells and the extracellular matrix (ECM).

In situ analysis of the recently identified chicken periostin gene detected RNA expression in the endocardial cushions of the heart and the trabecular endothelium by stage 21 (Norris et al., 2004). In this study, we report the detailed localization of periostin protein in the developing heart ranging in stages from HH20 to HH38 (Hamburger and Hamilton, 1951). At stage 20, periostin expression was localized in the atrioventricular (AV) canal as strands that appeared to be organized into fibers near the adjacent ventricle (arrow in Fig. 2A). A light staining was also visible in the ventricle. This was consistent with the in situ results that demonstrate periostin mRNA expression in the AV cushion and in the trabecular endothelium (Norris et al., 2004). Since periostin is secreted as part of the extracellular milieu, it may not be surprising that it did not associate with the trabecular endothelium at this stage. However, it is also possible that myocardial cells express periostin mRNA below the level of detection. At HH22, the fibrous strands of periostin staining were observed between the cushion mesenchyme cells and are also associated with the basal surface of the endocardial epithelium (arrows in Fig. 2B and C). This epithelium has been shown to undergo an epithelial-to-mesenchymal transition (EMT) (Markwald et al., 1995) implicating periostin in this developmental process. It might be that cells from the epithelium, which have also been shown to express periostin mRNA (Norris et al., 2004), are secreting periostin as they migrate through the cushion. In doing so, periostin may provide a framework for attachment and spreading of these migrating cells while maintaining the shape and cellular organization of the cushion.

thumbnail image

Figure 2. Periostin distribution in the HH20–29 chick heart. At HH20–29, immunolocalization of periostin in chick heart is organized into strands of strong staining in the atrioventricular canal. Staining is also evident in the adjacent myocardial cells of the atrioventricular cushion. By HH29, ventricular staining is localized to the trabecular endothelium and strong expression is also found in the trucus of the developing outflow tract. A: Whole mount immunostaining with the antichick periostin antibody in an HH20 chick heart. Arrows show strong staining of periostin that appears as fibrous strands in the AV. Light staining in the ventricle is also evident. B: HH22 chick heart section reveals strong periostin staining in the AV that appears as fibrous strands in the developing cushions (arrows) and in the migrating myocardial cells (asterisk). Light staining is also observed in the myocardium at this stage in Dent's fixative. C: Higher magnification of the AV cushion (box from A, HH22) shows a strand of periostin reactivity as well as surrounding cells closer to the endothelium of the cushion. A similar gradient of periostin expression was observed by in situ hybridization (Norris et al., 2004). Adjacent myocardial cells are also strongly positive (asterisk). D: HH25 sagittal section through the AV of a developing chick heart. Periostin staining is found in strands in the superior (s) and inferior (i) cushions of the AV (arrows). The ventricular staining of periostin becomes evident around a limited number of trabeculae (asterisk). This is consistent with the in situ localization at this stage. E: High magnification of the box in D shows the strands of periostin localized in the cushions (arrows) and periostin that appears to be secreted into the ECM (asterisk). F: The control sister section of D that shows the competing peptide inhibits all staining of this chicken periostin antibody. This is indicative of the competition at other stages (data not shown). G: Low magnification through a 5 μm section of an HH29 chick heart. Strong periostin staining is evident in the truncus (Tr) and LAV and in cells of the ventricle. H: The high magnification of the Tr (box in G) showing the strong staining in the developing walls of the aortic arch arteries. I: The high magnification of the LAV showing more fibrous strands concentrated in the developing valve. J: A different low magnification of a transverse section of an HH29 chick heart. Strong staining is evident in the aortic arch arteries, the trabecular endothelium, and less staining in the myocardium. This section is the sister section to the in situ of chick periostin [Fig. 4B in Norris et al. (2004)]. The localization of periostin protein shows consistency with the in situ localization. K: High magnification of the RV of an HH29 chick heart shows localization to the endothelium of the trabeculae (arrow). L: Maximum projection of eight optical sections of a whole mount-immunostained HH29 heart. Arrow denotes green strands of periostin staining in the conus of the cardiac outlet and asterisk shows the high level of staining in the truncus. The purple color is background fluorescence and provides contrast that reveals structures of the heart for reference. Scale bars = (A) 200 μm (applies to panels G and L); (B) 80 μm (applies to panels D and F); (C) 40 μm (applies to E, H, I, and K; (J) 400 μm.

Download figure to PowerPoint

By stage 25, periostin was detected as fibrillar strands in the superior and inferior cushions of the AV (arrows in Fig. 2D and E) and throughout the extracellular matrix (asterisk in Fig. 2E). Strong patches of periostin staining were observed between the endothelium and myocardium of an isolated number of trabeculae (asterisk in Fig. 2D). This was consistent with the observation that cells within the trabeculated myocardium express periostin mRNA at HH25 (Norris et al., 2004). Controls for periostin immunostaining were negative (Fig. 2F). The chick periostin peptide, used for immunization, successfully competed out all of the observable staining in Figure 2D. Distribution of a low level of periostin throughout the ventricular myocardium is consistent with its mRNA localization (Norris et al., 2004).

At stage 29, periostin was more readily detected and was concentrated around developing structures of the heart such as the truncus region of the outflow tract, the endothelium of the trabeculae, and the left atrioventricular valve (LAV) (Fig. 2G and H). Figure 2H and J show periostin associated with the developing walls of the aortic arch arteries. Fibrous strands of periostin were associated in the LAV and appeared more concentrated at this stage of cushion formation (Fig. 2I). Remarkably strong expression of periostin was localized to the endothelium of the trabeculae (Fig. 2J and K). In the developing outflow tract, periostin staining was significantly stronger in the truncus than the conus. This is best illustrated in Figure 2L as a maximum projection of eight optical sections of a whole mount-stained HH29 heart. The arrow shows the fibrous strands of periostin expression in the conal cushions and the asterisk identifies the significantly higher levels of periostin associated with the walls of the developing outlet near the truncal cushions.

The protein localization using this periostin antibody is very consistent with its mRNA expression [Fig. 2J compared with Fig. 4B of Norris et al. (2004)]. Importantly, it differs with the in situ pattern of βigH3 [Fig. 2J compared with Fig. 4E of Norris et al. (2004)]. Since the greatest potential for cross-reactivity of this antibody would be for its highly conserved family member βigH3, the divergent patterns of expression clearly verify the high specificity of this antibody used for the localization of periostin in tissue sections. At stage 29, the expression of βigH3 does not appear to overlap with periostin in the developing avian heart (Fig. 2J) (Norris et al., 2004). To date, there has not been a thorough study of βigH3 expression in the developing chick heart. In the mouse, βigH3 is expressed in the cardiac valves at embryonic day 12.5–14.5 and is downregulated by embryonic day 15.5–18.5 (Ferguson et al., 2003). βigH3 was also seen in the myocardium of late fetal development (Schorderet et al., 2000). It is unclear if these highly related, yet distinct, molecules will overlap in expression and provide redundancy during chick heart development.

As development proceeds, the endocardial cushions change in size and shape (van Mierop et al., 1962; de la Cruz et al., 1983) as well as their molecular and cellular composition (Markwald et al., 1975, 1977; Kinsella and Fitzharris, 1980; Arguello et al., 1988; Icardo, 1989a, 1989b). Strong periostin staining was evident in the LAV and the developing aortic wall (Fig. 3A and B). The LAV shows strong staining of periostin arranged in fibrous strands predominantly localized to the ventricular portion of the leaflet (Fig. 3A and C). Detectable staining was present in the right atrioventricular valve (RAV), albeit at a much lower level (Fig. 3A). At this stage, the RAV cells expressing periostin are mesenchymal and have not yet committed to any particular lineage. In the chick heart, the mesenchymal cells on the right side remodel to form a muscularized leaflet, whereas the left side forms fibrous connections that are subject to sheer stress. At HH34, the superior and inferior cushions are fused and a portion of these cells progressively undergo a morphological change into the structural components of the right and left septal leaflets. Although the molecular mechanisms that lead to these changes are poorly understood, the localization of periostin suggests that this protein may play a role in the remodeling of the cushions into a mature valvular apparatus.

thumbnail image

Figure 3. Localization of periostin in the HH34 chick heart. In later valvulogenesis (HH34–38) of the chick heart, periostin is predominant in the LAV at regions of attachment to the forming papillary muscle and areas undergoing delamination. It is also evident in the wall of the forming cardiac outlet and in the endothelial lining of the trabeculae (arrows). Regions of the RAV and the leaflets of the cardiac outlet show limited periostin reactivity (asterisk). A: Immunoreactivity with chick periostin shows strong staining in the LAV (box) and the wall of the developing aorta (Ao). In contrast, there is little reactivity in the right atrioventricular valve (asterisk). The arrow highlights the endothelial staining of the ventricle evident even at low magnification. B: High magnification of the developing aortic outlet with strong staining evident in the aortic wall but little or no staining in the aortic leaflets (asterisk). Nuclei are stained red with propidium iodide. C: Strong extracellular staining that appears fibrillar is evident in the LAV at HH34. This corresponds to regions of attachment to the forming papillary muscle and regions undergoing delamination. The arrow shows high magnification of periostin in the endothelial lining of the trabeculae. Propidium iodide (red) was used to reveal details of cardiac structure. D and E show a frontal section of an HH38 chick heart in the LAV. This section is from a whole mount in situ for periostin. This section was also immunostained with chick periostin. D: Immunoreactivity that predominantly localizes periostin protein to the delamination region of the leaflet. E: This is the identical section with peroxidase staining (purple) revealing the mRNA expression of periostin. F: Colocalization of the periostin mRNA (now an inverted image colored red) with protein (green). G: A transverse section of the LAV showing the mural (m) and septal (s) leaflets. Staining is found exclusively in the outer (ventricular) portion of the leaflet. H and I: Valvular apparatus of HH38 heart was dissected and whole mount-immunostained for periostin. Strong staining is evident in the chordae tendineae (CT) that attach the developing leaflet to the papillary muscle (PM). I: High magnification shows the fibrous nature of periostin in the region of attachment to the forming leaflet of the LAV. JL: Sections through an HH38 cardiac outlet showing strong staining in the tunica media of AA3 and AA4 as well as the adventitia (arrows) derived from the epicardium. K: High magnification of the pulmonary leaflets (box of J) reveals staining of periostin, albeit at a lower level than in other outflow tract (OFT) structures. L: Staining in the aortic branch of the HH38 cardiac outlet with strong staining in the aortic wall and adventitia. Scale bars = (A) 400 μm (applies to J and L); (B) 80 μm (applies to C and K). Magnification for D–G is approximately 100×.

Download figure to PowerPoint

Very little staining is evident in the aortic leaflet (asterisk in Fig. 3B), even though the developing wall of the aortic side of the outflow tract shows strong extracellular staining arranged into fibrous-looking strands (Fig. 3A and B). In the ventricle, periostin staining persisted near the trabecular endothelium (arrow in Fig. 3A and C). Since the epicardium highly expresses periostin (data not shown), the possibility exists that the strong staining in the ventricle at these later stages may be contributed by ventricular fibroblasts derived from the epicardium (EPDC).

The architectural pattern of fibrous tissues in the heart is more clearly apparent in the maturing leaflets as well as the annulus and adventitia by HH38. At this stage, periostin was found predominantly in these fibrous regions that are required to withstand the dynamic and repetitive changes that normally occur during the cardiac cycle. Periostin was expressed in the region of delamination of the preleaflet cushion from associated myocardium in the LAV (Fig. 3D–F). The enhanced secretion of periostin in this region may indirectly or directly induce the associated myocardium to transform into mesenchyme of a mixed phenotype. These mesenchymal precursors may undergo apoptosis to create and enlarge the extracellular spaces that precede delamination and/or transdifferentiate into cells that form the tendinous cords of the valvular suspensory apparatus. Since these mesenchymal cells have been shown to have the potential to form a diverse population of differentiated cells such as cardiac muscle, cartilage, bone, and bone marrow (Icardo, 1989b; Galvin et al., 2000), the expression of periostin at later stages in valvulogenesis may serve to maintain the fibrous or tendinous identity of these cells (Ji et al., 2000). We have shown that early in cardiac cushion formation, periostin is associated in fibrous strands long before the differentiated fibrous tissues could be identified within the forming valves.

The specificity of the periostin antibody was also evident in the sections of the LAV HH38 heart (Fig. 3D–F) that showed excellent colocalization of protein with periostin mRNA (Fig. 3E and F) within the same tissue section. A separate transverse section through the LAV (Fig. 3G) shows that periostin expression was even more discretely localized to the ventricular surface of the mural and septal leaflets (de La Cruz and Markwald, 1998). This is consistent with the earlier stage HH34 LAV shown in Figure 3C. Periostin may be an important molecular component of the subendocardial layer that participates in the development of the posterior leaflets of the LAV, particularly in the chordae tendinae. When the HH38 valvular apparatus is dissected and stained for periostin, confocal analysis reveals strong expression in the chordae tendinae and the more delicate fibrous connections to the papillary muscle and leaflet (Fig. 3H and I). Periostin may be an essential molecular component in leaflets that acquire a suspensory apparatus required to transmit significant stress and tension from the leaflet to the myocardium during the cardiac cycle.

In the outflow tract of the HH38 heart, periostin staining was predominantly found within the tunica media of the aortic vessels but not in the pulmonary vessels (compare Fig. 3J and K with J and L). Periostin is also localized to the epicardially derived fibrous adventitia of the outflow tract vessels (arrows in Fig. 3J and L). Although detectable in the developing valvular structures of the outflow tract, it does not have a strong staining pattern like the LAV. The highest staining of periostin in the developing valves of the outflow tract was observed in an area undergoing separation (escavation) from the outflow tract myocardium to form the valvular sinuses (asterisk in Fig. 3K).

In summary, periostin was detected primarily in the mesenchymal cushions of the LAV, the tendenous connections of the differentiating AV leaflet, the distal outflow tract (truncus), and the trabecular endothelium of the developing chick heart. It was also present in the conal cushions, RAV, and myocardium, albeit at significantly lower levels. It is interesting that throughout chick cardiovascular development, periostin expression is observed predominantly in the LAV and not the RAV, which contains only the muscular valve apparatus. Therefore, avians provide an easily accessible, internally controlled environment to examine periostin expression in both types of leaflets.

We hypothesize that periostin may serve to organize the extensive array of extracellular matrix molecules of the endocardial cushions. In doing so, it could provide a framework for attachment and spreading of the migrating cells that are undergoing the EMT early on in valvulogenesis. This hypothesis is consistent with the cells of the tooth mesenchyme that also highly express periostin and undergo an EMT (Kruzynska-Frejtag et al., 2004). During cardiovascular development initially, periostin is expressed in response to signaling from the BMP and TGF-β family of growth factors (Horiuchi et al., 1999; Ji et al., 2000; Hall et al., 2001; Kruzynska-Frejtag et al., 2001, 2004; Lindner et al., 2004). However, as development progresses, the heart must withstand the increasing influence of hemodynamic sheer stress (Stekelenburg et al., 2004). Thus, we might predict an increase of periostin within the valvular apparatus and the walls of the outflow tract, structures that also undergo tension and mechanical stress during the cardiac cycle. In mouse hearts that were subjected to pressure overload, ECM expansion resulted in a greater than 40-fold increase in periostin mRNA (Wang et al., 2003). Periostin mRNA was also increased in the periodontal ligament during experimental tooth movement (Wilde et al., 2003). In addition to its induction by sheer stress, at late stages of cardiovascular development periostin may also serve to maintain the integrity of fibrous tissues of the heart that has been shown to have potential to become cardiac muscle, cartilage, bone, and even bone marrow (Icardo, 1989b; Galvin et al., 2000).

Based on our expression studies and the known function of periostin in other cell types, we hypothesize that the loss of periostin would result in a lack of organized migration and spreading of cells undergoing EMT in the cushion tissue. This would lead to an unorganized matrix and misshapened cushions. Later in valvulogenesis, periostin appears to be a major component of the ECM of the chordae tendinae that provides a fibrous attachment to both the papillary muscle and the valvular leaflet. Loss or misexpression of periostin at this stage may result in an extracellular matrix that may not adequately transmit and withstand the mechanical forces from the leaflet to the papillary muscle.

However, in some cases, a decrease of periostin may actually be advantageous. It has recently been shown that a decrease of periostin using an antisense strategy in Dahl salt-sensitive rats increases the survival rate and left ventricular function in this cardiac dilation model (Katsuragi et al., 2004). This is consistent with an increase in the colocalization of periostin with collagen, laminin, and fibronectin that is observed in the remodeled myocardium of the adult rat ventricle following myocardial infarction (Stanton et al., 2000). Therefore, the ability to inhibit periostin in the adult myocardium might become a new therapeutic approach for the treatment of heart failure (Katsuragi et al., 2004).

We report here the production and characterization of an antibody that recognizes several isoforms of chicken periostin. This antibody will serve as a valuable tool in future experiments that elucidate the precise role of periostin in valvulogenesis and maintenance of the cardiac skeleton of the mature heart.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. Acknowledgements
  6. LITERATURE CITED

The authors thank Joshua Spruill and Sally Fairey for their technical assistance. Supported by National Heart, Lung and Blood Institute/National Institutes of Health grants R01-HL66231 (to C.H.M.) and R01-HL33756 and Po1-HL52813 (to R.R.M. and C.H.M.).

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. Acknowledgements
  6. LITERATURE CITED