Expression of Fibronectin, the Integrin as, and a-Smooth Muscle Actin in Heart and Lung Development Jesse Roman and John A. McDonald Department of Medicine, Washington University School of Medicine, St. Louis, Missouri

The developmentally regulated expression of fibronectin (FN) in developing organs and FN's ability to stimulate cell migration and differentiation in vitro suggest a role in organogenesis. We examined the distribution of FN and the a5 subunit of its receptor, the integrin a5(31, in the lungs and hearts of murine embryos at 11, 13, 16, and 18 days of gestation. In the lung, FN staining was present in the mesenchyme and parabronchial cells at day 11, increased at day 13, and decreased after day 16. Increases in FN coincided with the period of branching morphogenesis, and FN was concentrated at areas of airway bifurcation, suggesting a role for FN in cleft formation. The a5 subunit appeared later at 13 days, co-distributing with FN only in well-developed primary bronchioles. At all stages, a-smooth muscle actin expression correlated temporally and spatially with that of the a5 subunit. In the heart, staining for FN, the a5 subunit, and a-smooth muscle actin were present at day 11 and increased at day 13. FN was present in the outflow tract and developing atria and ventricles, where it was concentrated in the outer layer or visceral pericardium. Interestingly, a5 was detected at the inner layer, the endothelium, lining the outflow tract and atrioventricular cushions where endothelial cells migrate into the cardiac jelly in the process of epithelial-mesenchymal transformation. This suggests a potential role for a5(31 and FN in ventricular septation and valve formation. In contrast to the lung, a5 expression in the heart preceded that of a-smooth muscle actin, particularly in the cushions and the trabecular zone of the ventricles. These observations suggest that the interaction of embryonic cells . with FN may precede cell cytodifferentiation. FN and FN-binding integrins, including a5(31, may provide positional information necessary for lung branching and ventricular septation during development.

Fibronectin (FN) is a 500-leD dimeric cell-adhesive glycoprotein expressed in a highly developmentally specific fashion in many embryonic tissues, including the heart, brain, kidney, liver, and lung (1). This regulated expression (2), together with its ability to affect cell adhesion (3), migration (4), and cytodifferentiation (5) in vitro, has implicated FN in organogenesis. Although its exact function is not known, several studies support a role for FN in heart and lung development. In the lung, FN expression is increased during the glandular stage of avian and mammalian lung development, coinciding with the period of maximal cell proliferation and epithelial branching morphogenesis (2, 6). At this stage, FN is concentrated at areas of airway bifurcation and its deposition seems

(Received in originaljorm June 25, 1991 and infinal form January 2, 1992) Address correspondence to: Jesse Roman, M.D., VA Medical Center, Department of Medicine, 111, Pulmonary Division, 1670 Clairmont Road, Decatur, OA 30033. Dr. McDonald's current address: Mayo Clinic Scottsdale, 13400 East Shea Boulevard, Scottsdale, AZ 85259. Abbreviations: a-smooth muscle actin, a-SM actin; fibronectin, FN; ArgGly-Asp, ROD; 25 mM Tris, 150 mM NaC!, TBS; transforming growth factor-S, TGF-~. Am. J. Respir. Cell Mol. BioI. Vol. 6. pp. 472-480, 1992

critical for lung branching, as inhibitors of ligand binding to FN receptors, such as synthetic peptides containing the sequence RGD (Arg-Gly-Asp), prevent in vitro murine lung branching morphogenesis (7). Furthermore, a 70-kD aminoterminal fragment of FN that inhibits FN matrix assembly also prevents lung branching morphogenesis (8). In the heart, FN has been implicated in the regulation of precardiac cell migration, which is important for the formation of the heart tube during the early phase of heart organogenesis. Indeed, antibodies to FN inhibit normal development of early chicken hearts (9-11). During later stages of heart development, FN is believed to stimulate epithelialmesenchymal transformation, a critical process for remodeling the heart tube into a multichambered organ (12). Clearly, knowledge of the distribution of FN in developing organs at different stages is essential for understanding its function. In order to fully understand the function of FN, we also need to examine the tissue distribution of receptors for FN, as their expression will presumably playa major role in determining cell responsiveness. Therefore, we have also examined the distribution of one component of a wellcharacterized FN receptor, the a5 subunit of the integrin a5(31. The integrin a5(31 is thought to mediate many of the cellular effects of FN in vitro (3, 13). We performed immunohistochemical studies to examine the distribution of FN and the a5 subunit in the lung and heart of mouse em-

Roman and McDonald: FN and FN Receptor in Heart and Lung Development

bryos at different stages of gestation, hoping to gain insight into their role in development. In view of its intimate relationship with a5~1 expression in developing lungs, we have also examined the distribution of a-smooth muscle (a-SM) actin (7). Our observations suggest that although the expression of FN in developing lungs and hearts coincides temporally, the expression of the a5 subunit and a-SM actin does not. These findings also support previous observations suggesting a critical role for FN in providing positional information necessary for cleft formation during lung branching morphogenesis and endothelial cell migration for epithelial-mesenchymal transformation during cardiac septation. This information may be transmitted via a5~1 receptors present in mesenchymal cells adjacent to the developing airways expressing a-SM actin (parabronchial cells) and migrating cardiac endothelial cells.

Materials and Methods Embryonic Sections Balb/c female mice were mated with Balb/c males and checked every 8 h for a vaginal plug, the appearance of which represented day O. Female mice were killed at different stages of gestation, and uterine horns were exposed by laparotomy and excised into Hanks' buffered saline (Washington University Tissue Culture Support Center). The embryos were freed from the extraembryonic membranes and fixed in 4 % paraformaldehyde by immersion for paraffin embedding. Immunohistochemistry Immunohistochemical staining was performed on sagittal sections of whole mouse embryos obtained as above. Sections were stained with an immunoperoxidase-based method at room temperature (Vectastain ABC Kit; Vector Labs, Burlingame, CA). Embryos were obtained at 11, 13, 16, 17, and 18 days of gestation. A total of 85 embryonic sections were examined. Embryos were fixed in phosphate-buffered saline containing 4 % paraformaldehyde, dehydrated through a series of ethanol solutions, and embedded in paraffin. Sections (5 JLm) were deparaffinized, hydrated, and incubated with the indicated concentration of primary antibodies in diluent containing 25 mM Tris, 150 mM NaCI (TBS) and 0.05 % normal goat serum for 1 h, rinsed with TBS, and incubated with biotinylated secondary goat anti-rabbit and goat anti-mouse antibodies. Afterwards, the sections were rinsed with TBS, incubated with avidin-biotin-horseradish peroxidase complex for 1 h, and rinsed and developed with 0.5 % 3,3'-diaminobenzidine-0.01 % hydrogen peroxide solution in 0.1 M Tris (pH 7.2) for 7 min. Sections were counterstained with a solution consisting of 1% methyl-green plus 1% Alcian blue (Sigma Chemical Co., 51. Louis, MO) in 0.1 M acetate buffer (pH 6.5). Antibodies To detect FN distribution, tissue sections were stained with a polyclonal anti-FN antibody (Ab 1.5) developed in rabbits immunized with plasma FN (14). For the mouse a5 subunit, we used a monospecific synthetic peptide anti-of antibody (Ab 33) developed in rabbits immunized with a peptide containing the sequence of the last 12 amino acid residues of the a5 subunit cytoplasmic domain (13). The peptide sequence

473

of the human a5 cytoplasmic domain is identical to that in the mouse (15). A monoclonal antibody developed to a synthetic peptide of human a-SM actin was obtained from Sigma (cat. no. 29F4909). Rabbit normal IgG was used as control (cat. no. 1-5006; Sigma).

Results FN, as Subunit, and a-SM Actin Distribution in Lung Development In the developing mouse, the lung first appears as an epithelial evagination of the foregut during the embryonic stage of lung development (days 8 and 9) (16). This stage is followed by the glandular stage in which the lung undergoes repetitive branching in a process termed epithelial lung branching morphogenesis (9 to 16 days) during which the airways and the primitive bronchial tree are formed. At day 16, the embryonic lung is invaded by vascular structures, signaling the beginning of the canalicular stage. The alveolar stage follows and is characterized by the formation of alveoli with condensation of the mesenchyme and apposition of airspaces and vessels (after day 17). The alveolar stage extends into the postnatal period. Early glandular stage lungs (11 days) contain very few buds and are undergoing active monochotomous and dichotomous branching. At this stage, FN was detected at the epithelial-mesenchymal interface of developing airways (Figure lA, arrowheads). FN staining was also detected within the mesenchyme and the diaphragm but not in the airway epithelium. The a5 subunit was also present within the mesenchyme, but very little was detected at the epithelial-mesenchymal interface (Figure IB). a-SM actin staining was very faint around the airways and absent from the mesenchyme and epithelium (Figure IC). Mid-glandular stage lungs (13 days) have multiple primary and secondary bronchioles. Primary bronchioles are lined by columnar epithelium and are surrounded by spindleshaped mesenchymal parabronchial cells, whereas secondary bronchioles are lined with cuboidal epithelium, are round or not well defined, and are usually not surrounded by spindle-shaped cells. In 13-day-old lungs, FN was in the mesenchyme and concentrated at the epithelial-mesenchymal interface around spindle-shaped cells surrounding larger airways (parabronchial cells) (Figure ID). FN was particularly prominent in areas of airway bifurcation, but decreased or absent distally (not shown). FN was not present within the epithelium. The a5 subunit was also expressed by parabronchial cells of primary bronchioles (Figure IE) that were also expressing a-SM actin (Figure IE, arrow), suggesting that a5 expression increases concomitantly with cytodifferentiation. The epithelium of primary bronchioles seemed devoid of staining. However, staining for a-SM actin was detected in the epithelium of secondary bronchioles (Figure IF). Epithelial staining for a-SM actin was not detected after this stage. .. During the late glandular stage (16 days), vascular structures appear, initiating the canalicular stage of lung development which overlaps with branching morphogenesis. FN remained at the epithelial-mesenchymal interface and around parabronchial cells, as well as in vessels in 16-day-old lungs (Figure 2A). FN staining around parabronchial cells of pri-

474

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992

A

c

B

A

A

A

E E

M

Figure 1. Distribution of fibronectin (FN), as, and a-smooth muscle (a-SM) actin in 11- and l3-day-old lungs . Eleven-day-old (panels A through C) and thirteen-day-old (panels D through F) embryos were stained for FN (panels A and D) , as (panels B and E), and a-SM actin (panels C and F) . Panel A: In ll-day-old lungs, FN staining was found within the mesenchyme (M) and at the epithelial-mesenchyme interface of developing airways (A) (arrowheads) . None was detected in the epithelium (E) . Panel B: The as subunit was detected within the mesenchyme, very little in the parabronchial cells surrounding the airways, and none in the epithelium. Panel C: Staining for a-SM actin was absent from the lung mesenchyme. However, the anti-a-SM actin antibody stained very faintly in parabronchial cells surrounding the developing airways. Note that the staining for a-SM actin at the epithelial-mesenchymal interface was remarkably lighter than in vascular structures such as the dorsal aorta (asterisk) . Panel D: In 13-day-old lungs, FN staining, as well as that for as and a-SM actin, was increased and localized to the mesenchyme as well as around parabronchial cells surrounding the airways. Panel E: as staining was detected in parabronchial cells surrounding the airway mesenchymal cells. Note that parabronchial cells ofless developed secondary bronchioles, lined by cuboidal epithelium, express little or no as (arrow). Panel F: a-SM actin staining was present in parabronchial cells around primary bronchioles. Secondary bronchioles did not show staining for a-SM actin at parabronchial cells. However, the epithelium of these airways showed staining for a-SM actin. Staining was also detected in mesenchymal cells.

mary bronchioles was co-distributed with staining for a5 (Figure 2B) and a-SM actin (Figure 2C), whereas the secondary bronchioles showed little or no staining for a5{11 and a-SM actin (Figures 2B and 2e, arrow). Alveolar stage lungs (18 days) contained widened airspaces and had greatly diminished FN staining compared with previous stages (Figure 2D). No FN staining was detected in basement membranes at this stage but some remained within the mesenchyme and within epithelial cells of larger airways. Similarly, staining

for a5 was also decreased and resembled that of FN (Figure 2E). In contrast, a-SM actin was detected in smooth muscle parabronchial cells surrounding the airways at 17 days of gestation (Figure 2F). FN, a5131 Receptor, and a-8M Actin Distribution in Developing Heart Heart development has been divided into three phases (17). The first or early phase of organogenesis is further divided

Roman and McDonald : FN and FN Receptor in Heart and Lung Development

475

Figure 2. Distribution of FN, a5, and , a-SM actin in 16- and 18-day-old lungs. Sixteen-day-old (panels A through C), eighteen-day-old (panels D and E), and seventeen-day-old (panel F) embryos were stained for FN (panels A and D), a5 (panels B and E), and a-SM actin (panels C and F) . Panel A : In 16-day-old lungs, FN staining was decreased compared with 13-day-old lungs. FN was present in the mesenchyme (M), airway basement membranes, and around parabronchial cells. Also, FN was in the walls of vessels (V). Panel B: The a5 subunit was expressed by parabronchial cells of developing airways and mesenchymal cells. Note that secondary bronchioles express little or no a5 (arrow) . Panel C: a -SM actin was located in the parabronchial cells of primary bronchioles and the walls of blood vessels, while very little staining was detected in secondary bronchioles and within the mesenchyme (note that this image is of lower magnification , 13 X versus 26x). Panel D : In 18-day-old lungs, FN staining was strikingly decreased compared with day 13, with some staining detected within the mesenchyme and around blood vessels as well as within airway epithelial cells. Panel E: The a5 subunit was detected within the mesenchyme of 18-day-old lungs. Panel F: Seventeen-day-old lungs show staining for a-SM actin around the airways and vessels and very little within the mesenchyme.

into three stages and starts sometime before day 7 with the formation of the cardiac crescent (pre-tube stage). This is followed by the appearance of a tubular heart at day 8 (tube stage). During the last stage of the early phase of heart organogenesis, the tubular heart has bent into an S-shape and the ventricular loop is formed (9 days; loop stage). The second phase, or phase of advanced organogenesis, commences with the appearance of widening of the atrial segment of the loop and with the septation of the ventricles (days 9 and 10). Septation of the heart tube into multiple chambers depends on the migration of epithelial cells

(cardiac endothelium) into the cardiac jelly and their transformation into mesenchyme (18, 19). This phase ends with the closure of the interventricular communication (day 15). The third and last phase of heart development is termed the phase of fetal growth. This phase begins when the two ventricles are completely separated from one another and ends with birth . Although this encompasses the longest phase of heart development, very little is known about it. In this study, we have focused on the second (days 11 and 13) and third phases (days 16 and 18) of heart development. At day 11, FN staining was most prominent in a thin cell

476

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992

,

...

"

/'

/~~.p

/:

;,

,' 1

'; -'

',' "

1

.'

..... .; ':

.~

Figure 3. Distribution of FN, as, and a-SM actin in the l l-day-old heart . Embryos (11 days) were fixed, paraffinembedded, and stained for FN (panels A, D, and G), as (panels B and E), and a-SM actin (panels C, F, and H). Panel A: At 11 ' days of gestation, FN was detected in both atrial (a) and ventricular (v) structures, most prominently at the outermost layer. FN was also present in the pericardium (p). Panel B: as was found in all zones of the ventricles, including the outflow tract (0) , the cushions' lining (c), and the pericardium. Note the staining for as lining the developing cardiac valve (asterisk). Panel C: a-SM actin was present in the trabecular zone of the ventricle, in the atrium (A), and in the outer border of the outflow tract. No staining was detected in the pericardium. Panel D: Higher magnification (26x versus 6.SX) of panel A showing FN staining in the pericardium and compact, spongy, and trabecular zones of the ventricles. Panel E: Higher magnification of B (26x), showing as staining in all zones of the ventricles. Panel F: Higher magnification of panel C (26x), showing a-SM actin staining in the trabecular zone of the ventricles (26x). FN and as (not shown) were found mostly in the outer layer of the ventricle or compact myocardium, whereas a-SM actin was not. Panels G and H are higher magnification views (26x) of panels A and C, respectively, showing FN (panel G) and a-SM actin (panel H) staining in the atrium.

G

.

H

Figure 4 (above). Distribution of FN, as, and a-SM actin in 13-day-old hearts. Panel A: FN staining was increased at 13 days of gestation and found in all zones of the ventricles. Panel B: Staining for as was also increased at this stage and was found present in all zones of the ventricles. Panel C: a-SM actin was also present in all zones of the ventricles.

layer at the external border of both atria and ventricles and within the pericardium (Figure 3A). In the ventricles, it seemed to demarcate the visceral pericardium (Figure 3D) , although staining was also detected within the compact and spongy zones and developing ventricular trabeculae. FN was also detected in large vessels such as the dorsal aorta. Staining for as was detected in the pericardium, the compact and trabecular zones of the ventricles, and the atrium (Figure 3B). Interestingly; less staining for as was found in the myocardium at the atrioventricular junction and outflow tract (Figure 3B). We do not know if this is due to differences in the cellular expression of as in these regions or only apparent and due to differences in cell density. This finding plus the observation that cells located at the atrioventricular junction and outflow tract express a transforming growth factor-S (TGF-m-like molecule, whereas myocardial cells at other regions of the ventricles do not (20) , may suggest that myocardial cells at these sites are phenotypically different from cells within the ventricles . Staining for as was prominent in endothelial cells lining. the cushions at the atrioventricular junction and the outflow tract compared with the endothelium of the ventricles and atrium (Figure 3B). Staining for as was also detected within vascular structures such as the

Figure 5 (right). Distribution of FN and as in 16-day-old hearts. Sixteen-day-old hearts were stained for FN (panel A) and the as subunit (panel B) . Panel A: At this stage, the ventricular (V) FN staining pattern had reversed, being most prominent in the trabecular zone (inset) . However, note that FN staining continued to be detected in the outer zones as well. In the atrium (A), FN staining was detected in a fibrillar pattern surrounding clumps of cells. Panel B: The as subunit in 16-day-old lungs was present in all ventricular zones (see inset) . as staining in the atrial structures resembled that of FN. Panel C: Staining of 17-day-old heart with anti-a-SM actin antibody (magnification: l3x) . Note that staining is present in all areas of the ventricle.

478

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992

TABLE 1

Distribution of FN, FN receptor, and a-8M actin FN

as

Not detected in epithelium; present at EM interface and mesenchyme

Same distribution as for FN

Not detected in lung

13

Increased staining present at EM interface around PSMC and mesenchyme,

Present in PSMC at EM interface; only in primary airways; present in mesenchyme

Present in epithelium of secondary airways

16 (vascular stage) Vascular structures appear

Present at EM interface, around Present in vessels and around PSMC and mesenchyme, PSMC at the EM interface of and in vessels primary airways

Present in vessels and around PSMC of primary airways

18 (alveolar stage) Condensation of mesenchyme and alveolar formation

Overall diminished staining; present in epithelium and mesenchyme; not detected at EM interface

Same distribution as for FN

Present in PSMC of larger airways and in mesenchyme

Present mostly in pericardium and epicardium of ventricle and atrium; also present in compact, spongy, and trabecular zones of ventricles

Present in pericardium and compact and trabecular zones of ventricles and atrium; staining was diminished in cardiac jelly but prominent in endothelium of AV and outflow tract cushions

Present in trabecular zones of ventricles and outer border of conotruncus; not detected in outer zone of ventricle and pericardium

Days

A. In lung development 11 (glandular stage) Branching morphogenesis

B. In cardiac development 11 (phase of advanced organogenesis) Septation of heart; EM transformation

13

Subunit

a-SM Actin

Present in all ventricular zones, atria, and pericardium

16 (phase of fetal growth)

Intensity of stain decreased; most prominent in trabecular zones of ventricles; present in atrium

Present in all ventricular zones, atrium, and pericardium

Present in all ventricular zones, atrium, and pericardium

18

Overall staining markedly diminished; distributed in myocardium

Same as for FN

Present in all ventricular zones, atrium, and pericardium

Definition of abbreviations: FN = fibronectin; a-SM AV = atrioventricular.

=

a-smooth muscle; EM = epithelial-mesenchymal; PSMC = para bronchial smooth muscle cells;

conotruncus, aortic sac, and posterior vena cava. a-SM actin was detected in the trabecular zone of the ventricles (Figure 3C). Contrary to as, no staining for a-SM actin was present in the outer zones of the developing ventricles or the atrioventricular cushions (Figure 3C). Also, a-SM actin was detected in the outer border of the conotruncus and none in the pericardium. This is in striking contrast to lung development in which a5 and a-SM actin did not appear until later stages of development and a-SM actin staining corresponded closely to that of a5{jl (Figure 1). Thus, the expression of a5 precedes that of a-SM actin in the compact, peripheral ventricular myocardium and trabecular zone, as well as the atrioventricular junction and outflow tract (Figures 3B, 3C, 3E, and 3F). Similar to 13-day-old lungs, there was a remarkable increase in staining for FN in 13-day-old hearts. FN was distributed in all ventricular zones (compact, spongy, trabecular, and smooth) (Figure 4A). FN was also detected within the atria and pericardium. Staining for a5 (Figure 4B) and a-SM actin (Figure 4C) was also increased and localized to all ventricular zones and the atria. By day 16, FN staining had decreased and was present

within the spongy and, most prominently, at the trabecular zones of the ventricles and not at the outer zone of the heart (Figure 5A). In the atrium, FN seemed to be expressed both in myocardial and endothelial cells (Figure 5A). Staining for a5 was present in all zones of the ventricle and atria (Figure 5B). Similar to a5, staining for a-SM actin was very intense in both atria and all zones of the ventricles and remained intense after this stage of gestation (Figure 5C). At the fetal phase of organogenesis, FN staining was diminished but present in both ventricles and atria (not shown). Staining for a5 was similar to that of FN with respect to its distribution (not shown). The intensity of staining for a5 also seemed decreased at day 18 compared with day 16.

Discussion FN in Lung Development (Table tA) The increase in FN staining in developing murine lungs at the mid-glandular stage agrees with the observations of Chen and colleagues in avian lung (2) and those of Snyder and associates in rabbit lung (6). This increase in FN expression during epithelial lung branching morphogenesis may be crit-

Roman and McDonald: FN and FN Receptor in Heart and Lung Development

ical, as RGD peptides inhibit FN matrix assembly and prevent lung branching morphogenesis (7, 21). At 11 days of gestation, FN was found at the epithelialmesenchymal interface and around smooth muscle parabronchial cells of developing airways (this study). In these early glandular stage lungs, there was scant staining for a5 and a-SM actin around developing airways. By 13 days, FN staining was increased, particularly at areas of cleft formation, and a5 and a-SM actin appeared. After this day, staining for FN and a5 decreased whereas staining for a-SM actin continued to increase. The expression of FN before that of a5 and a-SM actin suggests that FN may promote mesenchymal cell migration or smooth muscle differentiation in developing proximal airways. The positional information transmitted from FN to embryonic lung cells is likely to be transmitted via a5{31 receptors (7), resulting in synthesis and secretion of matrix components or growth factors necessary for cleft formation at sites of future airway bifurcation. This hypothesis is supported by the fact that FN transforms arterial smooth muscle cells in vitro from a proliferative to a secretory and growth factor-responsive phenotype (22). This effect is mediated via a5{31 receptors. The alterations in extracellular matrix composition and growth factor production resulting from the transformation of smooth muscle cells may alter the balance between epithelial cell proliferation and the restrictive forces of the surrounding matrix, resulting in cleft formation with subsequent airway bifurcation. Finally, it is intriguing that the expression of a-SM actin was intimately related to a5 expression. Although the consequences of this are unknown, the regulation of expression of this and other cytoskeletal proteins (vimentin, desmin, and smooth muscle myosin) in developing lungs suggests a role for cytoskeletal phenotypic changes in lung branching morphogenesis (23). FN in Heart Development (Table lB) Similar to developing lungs, FN was already present in hearts at 11 days of gestation, increased at 13 days, and decreased thereafter. The increase in FN staining coincided with the phase of advanced heart organogenesis when septation of the ventricles and outflow tract and closure of the interventricular foramina occur (17). Expression of a5 and a-SM actin occurred at the same stage of gestation as that of FN although they were not always colocalized. Their expression also increased in day 13 and was concentrated within the developing ventricles and outflow tract. Several pieces of evidence support a role for FN in heart development. After gastrulation, certain embryonic cells move to two lateral mesodermal zones that contain heartforming capacity (24, 25). Precardiac cells within these lateral zones migrate anteriorly, cease their movement at the midline, and form two endocardial tubes that will fuse and forma beating tubular heart by the second day of gestation. FN appears during this stage at the mesoderm-endoderm interface of the chick heart, coinciding spatially and temporally with the directional migration of precardiac cells, suggesting a functional role in this event (9). Indeed, anti-FN antibodies and the disruption of FN-containing migratory pathwaysprevent normal chicken heart development (10, 11).

479

During the advanced phase of heart organogenesis, the cardiac jelly swells in the areas of future atrioventricular canal and the outflow tract (26). These swellings, termed endocardial cushions, constitute the primordia of valves and membranous septa (26, 27). Stimulated by signals derived from the myocardial layer, endothelial cells migrate into the cushions and become mesenchymal cells via the process of epithelial-mesenchymal transformation (18, 19). The fact that FN is concentrated at the outer margin of the heart supports the role of FN as a stimulatory force for endothelial migration and transformation. However, FN alone does not stimulate endothelial cell migration into collagen gels (18). On the other hand, FN may function as an inhibitor of endothelial cell migration. The in vitro segregation of mesenchymal and myocardial cell aggregates is stimulated by a FNrich matrix elaborated by the cardiac mesenchyme as well as by TGF-{3, and this effect is inhibited by serumless media or RGD peptides (28). Although addition of FN to the serumless medium does not restore the ability of the cell aggregates to segregate, it suggests that FN may prohibit endothelial cell migration into the cardiac jelly. Therefore, we could predict that a decrease in FN expression at the cushions should occur to allow epithelial-mesenchymal transformation. Indeed, a decrease in FN staining occurs within the atrioventricular cushions at 10 days of gestation (29, 30). The expression of a5 receptors in endothelial cells lining the cushions (this work) supports the idea that the effects of FN are mediated via a5{31 receptors. Expression of endothelial a5 is most prominent at the atrioventricular junction and the outflow tract. This pattern of expression is also typical for TGF-{3 (31) and suggests that TGF-{3 may be responsible for integrin expression during heart development (32). Interestingly, a5 expression preceded that of a-SM actin in the cardiac cushions and the outer zones of the heart, suggesting that, at early stages of development, integrin expression may precede the expression of tissue-specific antigens. Of particular interest is the observation that the trabecular muscle expressed both a5 and a-SM actin at early stages of development. This suggests that the myocardium in the trabecular zone is more differentiated than the myocardium in the parietal or peripheral compact zone (33). In the atrium, the formation of mesenchyme is critical for the formation of the septum primum which will fuse with the atrioventricular endocardial cushions during the advanced stage of heart development (34). The immunohistochemical localization of FN and a5 suggests that mechanisms similar to those described above for septation of the ventricle and outflow tract may take place during atrial development. FN in Organogenesis Our observation that FN is increased after day 13 of murine gestation in both lung and heart (as well as brain; unpublished observations) suggests the existence of an embryonic clock responsible for turning on extracellular matrix production at a predetermined stage of embryogenesis. Increased FN deposition during specific stages of organ formation may provide positional information regulating the migration and differentiation of certain embryonic cells (1, 7, 35-39). The mechanisms responsible for enhancing FN expression and initiating this cascade of events are unknown. The polypep-

480

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992

tide growth factor TGF-t3 has been implicated in lung as well as heart development. TGF-t3 stimulates cellular production of extracellular matrix components including FN, collagens, elastin and proteoglycans, and their receptors (32). In the lung, TGF-t3 is distributed around developing airways in parabronchial mesenchymal cells colocalizing with FN, collagens type I and III, and proteoglycans (40). Its concentration at areas of airway bifurcation suggests a role in cleft formation. In the heart, in situ hybridization studies in developing mice revealed high levels of expression within the cardiac mesenchyme at 7 days post coitum (30). After this, TGF-t3 concentrates at the endothelium, particularly the area overlying the cushion tissue. The fact that antibodies to TGF-t3 inhibit mesenchymal transformation from atrioventricular endothelium in vitro further supports its role in cardiac septation (41). Acknowledgments: We would like to acknowledge Ms. Candice Little for her assistance in the generation of the data presented. This work was supported by an American Lung Association Research Grant and a Minority Medical Faculty Development Award from the Robert Wood Johnson Foundation to J. Roman and Grant 5-ROI-HL43418 from the National Institutes of Health to Douglas C. Dean, Ph.D. (Washington University School of Medicine) and J. A. McDonald.

References 1. Vartio, T., L. Laitinen, O. Narvanen, L. E. Thornell, and I. Virtanen. 1987. Differential expression of the cellular form of fibronectin in embryonic and adult human tissues. J. Cell Sci. 88:419-430. 2. Chen, W. T., J. M. Chen, and S. C. Mueller. 1986. Coupled expression and colocalization of 140K cell adhesion molecules, fibronectin, and laminin during morphogenesis and cytodifferentiation of chick lung cells. J. Cell BioI. 103:1073-1090. 3. Akiyama, S., S. Yamada, W. Chen, and K. Yamada. 1989. Analysis of fibronectin receptor function with monoclonal antibodies: roles in cell adhesion, migration, matrix assembly, and cytoskeletal organization. J. Cell BioI. 109:863-875. 4. Boucaut, J. C., and T. Darribere. 1983. Fibronectin in early amphibian embryos. Migrating mesodermal cells contact fibronectin established prior to gastrulation. Cell Tissue Res. 234:135-145. 5. Menke, A. S., and D. Boettiger. 1987. Occupation of the extracellular matrix receptor, integrin, is a control point for myogenic differentiation. Cell 51:51-57. 6. Snyder, J., J. O'Brien, and H. Rodgers. 1987. Localization and accumulation of fibronectin in rabbit fetal lung tissue. Differentiation 34:32-39. 7. Roman, J., and J. A. McDonald. 1991. Potential role of RGD-directed integrins in mammalian lung branching morphogenesis. Development 112:551-558. 8. Roman, J., E. C. Crouch, and J. A. McDonald. 1991. Reagents that inhibit fibronectin matrix assembly of cultured cells also inhibit lung branching morphogenesis in vitro: implications for lung development, injury and repair. Chest 99:20s. (Abstr.) 9. Linask, K. K., andJ. W. Lash. 1986. Precardiac cell migration: fibronectin localization at mesoderm-endoderm interface during directional movement. Dev. Biol. 114:87-101. 10. Linask, K. K., andJ. W. Lash. 1988. A role for fibronectin in the migration of avian precardiac cells. I. Dose-dependent effects of fibronectin antibody. Dev. BioI. 129:315-323. 11. Linask, K. K., andJ. W. Lash. 1988. A role for fibronectin in the migration of avian precardiac cells. II. Rotation of the heart-forming region during different stages and its effects. Dev. BioI. 129:324-329. 12. Mjaatvedt, C. H., and R. R. Markwald. 1989. Induction of an epithelialmesenchymal transition by an in vivo adheron-like complex. Dev. BioI. 136:118-128. 13. Roman, J., R. M. LaChance, T. J. Broekelmann et al. 1989. The fibronectin receptor is organized by extracellular matrix fibronectin: implications for oncogenic transformation and for cell recognition of fibronectin matrices. J. Cell BioI. 108:2529-2543. 14. McDonald, J. A., T. 1. Broekelmann, D. G. Kelley, and B. Villiger. 1981. Gelatin binding domain specific anti-human plasma fibronectin Fab' inhibits fibronectin mediated gelatin binding but not cell spreading. J. Bioi. Chern. 256:5583-5587. 15. Holers, V. M., T. G. Ruff, D. L. Parks, J. A. McDonald, L. L. Ballard, and E. J. Brown. 1989. Molecular cloning of a murine fibronectin recep-

tor and its expression during inflammation: expression of VLA-5 is increased in activated peritoneal macrophages in a manner discordant from MHC class II. J. Exp. Med. 169:1589-1605. 16. Ten Have-Opbroek, A. A. W. 1981. The development ofthe lung in mammals: an analysis of concepts and findings. Am. J. Anat. 162:201-219. 17. Pexieder, T., A. C. G. Wenink, and R. H. Anderson. 1989. A suggested nomenclature for the developing heart. Int. J. Cardiol. 25:255-264. 18. Mjaatvedt, C. H., R. C. Lepera, and R. R. Markwald. 1987. Myocardial specificity for initiating endothelial-mesenchymal cell transition in embryonic chick heart correlates with a particulate distribution of fibronectin. Dev. Biol. 119:59-67. 19. Markwald, R. R., E. L. Krug, C. H. Mjaatvedt, and A. R. Sinning. 1990. Endothelial formation of mesenchyme: induction by an extracellular glycoprotein complex. In Molecular Biology of the Cardiovascular System. Alan R. Liss, New York. 311-319. 20. Lyons, K. M., R. W. Pelton, and B. L. Hogan. 1990. Organogenesis and pattern formation in the mouse: RNA distribution patterns suggest a role for bone morphogenetic protein-2A (BMP-2A). Development 109:833844. 21. Darribere, T., K. Guida, H. Larjava, K. E. Johnson, K. Yamada, J. P. Thiery, and J. C. Boucaut. 1990. In vivo analysis ofintegrin b-I subunit function in fibronectin matrix assembly. J. Cell Biol. 110:1813-1823. 22. Hedin, U., B. A. Bottger, J. Luthman, S. Johansson, and J. Thyberg. 1989. A substrate of the cell-attachment sequence of fibronectin (ArgGly-Asp-Ser) is sufficient to promote transition of arterial smooth muscle cells from a contractile to a synthetic phenotype. Dev. Biol. 133:489-501. 23. Mitchell, J. J., S. E. Reynolds, K. O. Leslie, R. B. Low, andJ. WoodcockMitchell. 1990. Smooth muscle cell markers in developing rat lung. Am. J. Respir. Cell Mol. BioI. 3:515-523. 24. Dehaan, R. L. 1963. Migration patterns of the precardiac mesoderm in the early chick embryo. Exp. Cell Res. 29:544-560. 25. Dehaan, R. L. 1964. Cell interactions and oriented movement during development. J. Exp. 2001. 157:127-138. 26. Markwald, R. R., T. P. Fitzharris, and F. J. Manasek. 1987. Structural development of endocardial cushions. Am. J. Anat. 148:85-120. 27. Van Mierop, L. H. S., R. D. Alley, H. W. Kausel, and A. Stranahan. 1963. The anatomy and embryology of endocardial cushion defects. J. Thorac. Cardiovasc. Surg. 43:71-96. 28. Krug, E. L., R. B. Runyan, and R. R. Markwald. 1985. Protein extracts from early embryonic hearts initiate cardiac endothelial cytodifferentiation. Dev. BioI. 112:414-426. 29. Armstrong, P. B., and M. T. Armstrong. 1990. An instructive role for the interstitial matrix in tissue patterning: tissue segregation and intercellular invasion. J. Cell BioI. 110:1439-1455. 30. Icardo, J. M. 1985. Distribution of fibronectin during the morphogenesis of the truncus. Anat. Embryol. 171:193-200. 31. Arkhurst, R., S. A. Lehnert, A. Faissner, and E. Duffie. 1990. TGF beta in murine morphogenetic processes: the early embryo and cardiogenesis. Development 108:645-656. 32. Heino, J., R. A. Ignotz, M. Hemler, C. Crouse, and J. Massague. 1989. Regulation of cell adhesion receptors by transforming growth factor-beta. J. BioI. Chern. 264:380-388. 33. Tokuyasu, K. T. 1990. Co-development of embryonic myocardium and myocardial circulation. In Developmental Cardiology: Morphogenesis and Function. E. B. Clark and A. Takao, editors. Futura Publishing Co., Mount Kisco, NY. 205-218. 34. Arrechedera, H., M. Alvarez, M. Strauss, and C. Ayesta. 1987. Origin of mesenchymal tissue in the septum primum: a structural and ultrastructural study. J. Mol. Cell. Cardiol. 19:641-651. 35. French-Constant, C., and R. O. Hynes. 1988. Patterns of fibronectin gene expression and splicing during cell migration in chicken embryos. Development 104:369-382. 36. Boucaut, J. C., T. Darribere, S. D. Li, H. Boulekbache, K. M. Yamada, and J. P. Thiery. 1985. Evidence for the role of fibronectin in amphibian gastrulation. J. Embryol. Exp. Morphol. 89:211-227. 37. Bronner-Fraser, M. 1986. An antibody to a receptor for fibronectin and laminin perturbs cranial neural crest development in vivo. Dev. Biol. 117:528-536. 38. Bottger, B. A., U. Hedin, S. Johansson, and J. Thyberg. 1989. Integrintype fibronectin receptors of rat arterial smooth muscle cells: isolation, partial characterization and role in cytoskeletal organization and control of differentiated properties. Differentiation 41:158-167. 39. Roman, J., and J. A. McDonald. 1991. Fibronectins. In The Lung: Scientific Foundations. R. G. Crystal, J. B. West, P. Barnes, N. S. Cherniack, and E. R. Weibel, editors. Raven Press, New York. 399-411. 40. Heine, U. I., E. F. Munoz, K. C. Flanders, A. B. Roberts, and M. B. Sporn. 1990. Colocalization of TGF-beta 1 and collagen I and III, fibronectin and glycosaminoglycans during lung branching morphogenesis. Development 109:29-21. 41. Potts, J. D., and R. B. Runyan. 1989. Epithelial-mesenchymal cell transformation in the embryonic heart can be mediated, in part, by transforming growth factor beta. Dev. BioI. 134:392-401.