DEVELOPMENTAL
BIOLOGY
141,421-425 (1990)
Regulation of Proliferation of Embryonic Heart Mesenchyme: Role of Transforming Growth Factor-PI and the Interstitial Matrix MICHAEL *Division
of Pediatric
CHOY,*,’
MARGARET
T. ARMSTRONG,?
AND PETER
B. ARMSTRONG?
Cardiology, Department of Pediatrics, University of Calzfornia Davis Medical Center, Sacranumto, and TDepartment of Zoology, University of Cal$n-nia at Davis, Davis, California 95616 Accepted
California 95817,
June 20, 1990
Proliferation of atrioventricular cushion mesenchyme of the embryonic avian heart maintained in three-dimensional aggregate culture is stimulated by interaction with the interstitial matrix. Chicken serum or transforming growth factor-pl, which stimulates proliferation, induces matrix deposition in regions of the aggregate showing high labeling indices with tritiated thymidine. Dispersed heart mesenchyme interstitial matrix introduced into serum-free culture is incorporated into the aggregate and stimulates cellular proliferation similar to serum or transforming growth factorpl. Proliferation is reversibly inhibited by the peptide Gly-Arg-Gly-Asp-Ser-Pro. It is suggested that transforming growth factor-(31 stimulates the production of interstitial matrix and that a sufficient stimulus for proliferation in this system is the presence of the matrix, which acts as the adhesive support for cellular anchorage. o XIO Academic PEW, IM. INTRODUCTION
Transforming growth factor-p (TGF-P)’ type I is a pleiotropic peptide mediator that affects cellular proliferation (Roberts et aZ., 1988) and stimulates production of the interstitial extracellular matrix (Ignotz and Massague 1986). TGF-/31 is present in a number of mesenchymal tissues during embryonic development (Heine et ak, 1987; Lehnert and Akhurst, 1988), where it may conceivably play a role in morphogenesis and tissue growth. One tissue with high levels of TGF-Bl is the mesenthyme of the atrioventricular valves of the embryonic heart of the mouse (Heine et al., 198’7; Lehnert and Akhurst, 1988) and chick (M. Choy et ah, submitted for publication). This tissue is established at stage 15 (Hamburger and Hamilton, 1951) of chick embryonic development when the mesenchyme precursor cells are differentiated from endocardium possibly under the influence of fibronectin and TGF-/31 (Mjaatvedt et CL,!., 1987; Potts and Runyan, 1989). Mesenchymal cells migrate into the adjacent extracellular matrix (ECM) separating myocardium and endocardium and then proliferate to establish the primordia for the endocardial cushions and subsequent atrioventricular valves. To investigate the possible involvement of TGF-Pl in later stages of heart morphogenesis, we developed a novel ’ To whom all correspondence should be addressed. ’ Abbreviations used: ANOVA, analysis of variance; BSA, bovine serum albumin; CS, chicken serum; DMEM, Dulbecco’s-modified Eagles medium; ECM, extracellular matrix; GRGDSP, Gly-Arg-GlyAsp-Ser-Pro peptide; GRGESP, Gly-Arg-Gly-Glu-Ser-Pro peptide; TGF-0, transforming growth factor-@ 421
culture assay in which heart mesenchyme from the atrioventricular valve region was separated from the myocardium and then was maintained in organ culture as three-dimensional aggregates. It was reasoned that this established an organization that better reflects the three-dimensional organization of heart mesenchyme in tivo than the conventional monolayer culture. Proliferation was determined by the incorporation of tritiated thymidine. Our observations support the suggestion that TGF-pl stimulates the production of an interstitial matrix and that the presence of the adhesive matrix alone is a sufficient stimulus for proliferation. MATERIALS
AND METHODS
Ceil and Aggregate Culture The hearts of lo-day-old White Leghorn chick embryos were isolated, and the atrioventricular regions, including the atrioventricular valves, were carefully dissected out in a manner to exclude the epicardial mesenchyme. The atrioventricular tissue was dissociated in 1.0 mg/ml trypsin, the cell suspension was freed of myocytes leaving less than 1% of myocytes and unidentified epithelioid cells (Armstrong and Armstrong, 1978), and the remaining mesenchymal cells were grown to eonfluence in Dulbecco’s-modified Eagles medium (DMEM) plus 10% chicken serum (CS). Five by 5-mm segments of the confluent monolayers were scraped gently from the culture plates using a small rubber scrapper and transferred to a gyratory shaker flask culture in DMEM plus 1.0 mg/ml of bovine serum albumin (BSA). The medium did not contain any 001%1606/‘90 $3.00 Copyright 0 1990 by Academic Press, Inc. All rigbta of reproduction in any form reserved.
422
DEVELOPMENTALBIOLOGY
growth factors, but it did allow the tissue to remain cohesive. After 4 days the sheets of cells had remodeled into smoothed-surface, spherical cell aggregates that ranged between 0.2 and 0.5 mm in diameter. Aggregates were then transferred into experimental culture media and incubated further in a shaker flask culture. The potential effecters were added to basal medium (DMEM plus 1.0 mg/ml BSA), with 2.0 &i/ml tritiated thymidine to allow a determination of the effects of the experimental conditions on cell proliferation. Fibronectin-free serum was depleted of fibronectin by passage over a gelatin-Sepharose column (Engvall and Ruoslahti, 1977). Isolation. of the Interstitial
Matrix
Monolayers of cardiac mesenchymal cells grown in DMEM plus 10% CS produce an ECM composed predominantly of fibronectin and containing lesser amounts of tenascin, laminin, and collagen type I (Armstrong and Armstrong, 1990). The matrix was prepared by hypotonic ammonium hydroxide extraction of the monolayers described by Fairbairn et al. (1985). Transfbrming
Growth Factor-@1 and Synthetic
Peptides
TGF-Pl was obtained from Calbiochem (Cat. No. 619350) and R + D (Cat. No. lOl-Bl). Gly-Arg-GlyAsp-Ser-Pro (GRGDSP) and Gly-Arg-Gly-Glu-SerPro (GRGESP) peptides were synthesized as described by Armstrong and Armstrong (1990). Immunohistochemistry
and Autoradiography
All the aggregates were fixed in Carnoy’s fixative, embedded in paraffin, sectioned, and mounted on gelatin-coated slides. The sections of aggregates were stained for fibronectin with goat anti-chicken fibronectin antibody obtained from Dr. K. Yamada (Yamada, 1978) by the ABC indirect immunoperoxidase procedure (Vector Laboratories) or with the rhodamine immunofluorescence procedure. Control sections were stained with preimmune serum or fibronectin-adsorbed antiserum substituted for the first antibody. The slides destined for autoradiography were coated with Kodak NTB-2 photographic emulsion and were processed as described by Kopriwa and Leblond (1962). Within the aggregate two regions were identifiedcortex and medulla. The cortex consisted of the outer layers of cells where extracellular fibronectin was present in some of the groups of aggregates. If fibronectin was absent, then the cortical layer was arbitrarily defined as extending from the surface of the aggregate to four to five cell layers deep. The medulla included the remaining interior cells. The labeling index of each region of the aggregate was calculated by dividing the
VOLUME 141.1990
number of cells labeled with tritiated total number of cells in the particular plying by 100.
thymidine by the region and multi-
Statistics Ten to 20 aggregates were placed in a shaker flask with the potential effector or effecters, and each potential effector was tried at least five times on different sets of aggregates. Since all results were repeatable, the statistical analysis was performed on one set of experiments. In each of the four major experiments, five aggregates were randomly selected from each group, and the labeling indices were determined. Each experiment was analyzed with several statistical methods described in Table 1. RESULTS
Stimulation
AND
of Cell Prolifwation
DISCUSSION
by TGF-pl
and Serum
Fibronectin, which was a major component of the ECM of the mesenchymal cell monolayer, was lost by the second day of aggregate culture. After 4 days of culture in basal medium, aggregates were transferred to experimental conditions (Experiment 1) in fresh culture flasks, and the effects on cell proliferation were assessed by determining the tritiated thymidine incorporation index. Cells transferred to fresh basal medium showed a labeling index that was relatively low, varying between 23% in the interior of the aggregate to 31% at the surface (Fig. lA, Table 1). A closer examination of the individual cells of the aggregates revealed that the cell shape tended to be stellate throughout the aggregate. Aggregates transferred to basal medium containing 0.1 or 1.0 nMTGF-Bl had different morphology and proliferative behavior (Experiments 1 and 4). After 2 days of incubation, an extracellular fibronectin matrix was present in the periphery or cortex of the aggregates in a zone that extended four to five cell layers deep (Fig. 1B). The inner portion of the aggregates, the medulla, was fibronectin poor. The boundary between the two regions was sharp. Interestingly, the cells in the cortex tended to be flattened, while the cells in the medulla remained stellate. The labeling index of the cortex was elevated (80%) compared to the medulla (22%) or to either region of the control aggregates maintained in basal medium (Fig. lB, Table 1). Thus, exposure to TGF-Bl induced the formation of a fibronectin-rich ECM and stimulated cell proliferation. The inclusion of 3 or 10% CS had an identical effect to that of TGF-01 on matrix deposition, cell morphology, and cell proliferation (Experiments 2 and 3). After 1 day of incubation, aggregates developed a cortical region of
CHOY,
ARMSTRONG,
423
TGF-01 and Cell Prolgeratim
AND ARMSTRONG
TABLE 1 REGIONAL
Medium
Cortex Experiment
BSA TGF-01 (1.0 nM)b
31 + 1.6” 80 f 1.3”,”
25 80 33 72
Medulla
INDICES
Medium
1
Experiment BSA cs (3%)” GRGDSP + CS GRGESP + CSb
LABELING
+ f + -t
Experiment 23 f 1.0 22 f 1.5
BSA cs (lO%)b ECMb
3 1.4 1.4”’ 2.9 1.6””
Cortex
19 f 24 f 15 f 18 f
1.1 1.3 3.3 0.9
BSA TGF-@l (0.1 nM)b TGF-@l (1.0 n&f) b GRGDSP + TGF-01 GRGESP + TGF-fll
2
9 + 0.7 81 + 2.1’!’ 66 f 1.6”,” Experiment 23 67 67 18 51
Medulla
5 * 0.2 7 k 0.9 5 f 0.7
4 f 3.3 f 1.7”~” f 2.5”” I? 1.0” f 3.0’”
24 29 26 13 15
f + f f +
4.2 1.2 1.6 0.8 0.9
Note. The labeling index of each region was calculated by dividing the number of cells undergoing DNA synthesis by the total number of cells in the particular region and multiplying by 100. Values are the means f SEM of five randomly chosen aggregates within the designated group. Abbreviations: ANOVA, analysis of variance; BSA, bovine serum albumin; CS, chicken serum; ECM, sonicated fibronectin-rich extracellular matrix; GRGDSP, Gly-Arg-Gly-Asp-Ser-Pro peptide; GRGESP, Gly-Arg-Gly-Glu-Ser-Pro peptide; TGF-01, transforming growth factor-81. a Significantly different between regions within a group, P < 0.01, Student’s paired t test. b Significantly different from BSA group, P < 0.01, repeated-measures ANOVA with one grouping factor (the medium) and one within factor (medulla/cortex), Scheffe F test. ’ Significantly different from corresponding region of BSA group, P i 0.01, one-way ANOVA, Scheffe F test.
flattened, mitotically active cells (Table 1) that were associated with a fibronectin-rich ECM (Fig. 1C). The cells in the interior remained stellate and showed a low labeling index and a sparse matrix. The medullary tissue was viable with an identical rate of protein synthesis as the cortical tissue (Armstrong and Armstrong, 1990). Fibronectin-depleted CS had the same effect on matrix deposition and cell proliferation as complete CS (data not shown), suggesting that serum fibronectin was not the factor in the CS responsible for stimulating matrix deposition and proliferation. The matrix stimulated by serum and TGF-01 contained a variety of adhesive glycoproteins, including fibronectin, laminin, tenascin, and small amounts of collagen type I (Armstrong and Armstrong, 1990).
Role of the Extracellular
Matrix
in Cell Proliferation
The mitotically active cortex of TGF+land serumtreated aggregates is characterized by the presence of a prominent fibronectin-rich ECM and a flattened cell morphology. To study the role of the matrix in the regulation of cell shape and cell proliferation, aggregates were transferred to basal medium supplemented with a sonicated preparation of heart mesenchyme ECM (Experiment 2). The sonicated matrix incorporated selectively into the cortical zones of the aggregates (Armstrong and Armstrong, 1990) (Fig. 1D). The cortical cells associated with the matrix developed the flattened mor-
phology and the increased proliferative rate similar to aggregates exposed to TGF-01 or serum (Table 1). Further evidence for the involvement of the ECM in the regulation of cell morphology and proliferation is provided by the perturbation of matrix deposition by peptides containing the sequence Arg-Gly-Asp (Ruoslahti, 1988; Yamada and Kennedy 1984). Peptides with substitutions at the aspartate residue are, typically, markedly less inhibitory and are used as negative controls. Since Arg-Gly-Asp-containing peptides containing the usual L-isomer of arginine were inactivated within a few hours in this system (Armstrong and Armstrong, 1990), we used peptides containing the D-isomer of arginine because they proved to be stable and biologically active (Pierschbacher and Ruoslahti, 1987). GRGDSP (1.0 mg/ml) prevented the formation of the cortical fibronectin-rich ECM in aggregates exposed to 3% serum (Experiment 3) or 1.0 nM TGF-fil (Experiment 4). The cortical cells failed to develop the flattened morphology and revealed a low tritiated thymidine labeling index (Fig. lE, Table 1). GRGDSP did not appear to be toxic since its effects were reversible: aggregates that were transferred from peptide-containing medium to peptide-free medium established a cortical fibronectin matrix and showed a flattened cell morphology and a high proliferative index (data not shown). In addition, the GRGESP control (1.0 mg/ml) did not inhibit fibronectin production or cell proliferation in the cortical layers of the aggregates. Fibronectin is a matrix adhesion molecule that has
424
DEVELOPMENTALBIOLOGY
VOLUME141,1%0
FIG. 1. Autoradiograms of aggregates in different media containing tritiated thymidine and stained for fihronectin by a standard immunoperoxidase technique. Bar = 0.10 mm. (A) In basal medium, fibronectin ECM is absent. Labeled nuclei are present, but are relatively few both in the outer layer of cells called the cortex (c) and in the interior, the medulla (m). (B) In basal medium containingTGF+l, fibronectin appears as a gray fibrillar ECM in the cortical region. The percentage of mitotically active cells is much greater in the cortex than that of the medulla. (C) In CS, the cortex contains a fibrillar matrix of fibronectin and abundant labeled nuclei. (D) In basal medium supplemented with a sonicated preparation of the fibronectin-rich ECM isolated from confluent heart mesenchymal monolayer, staining for fibronectin demonstrates the cortical localization of the ECM. Similar to aggregates grown in TGF-j31 or CS, the cortical cells are mitotically active. (E) Cultured in medium containing both CS and the peptide GRGDSP, which binds to the fibronectin receptor sites of the mesenchymal cells, the intact fibronectin-rich matrix is absent, and cell proliferation in both regions is low.
several functional domains, including binding sites for structural elements of the ECM (collagen, glycosaminoglycans, and fibrin) and for receptors on the surfaces of cells (Ruoslahti, 1988). The multiplicity of binding sites
enables fibronectin to link cells to collagen and other components of the matrix. The colocalization of matrix and the region of active proliferation suggests that the mitogenic response in the three-dimensional system
CHOY, ARMSTRONG, AND ARMSTRONG
may consist of two steps: (1) the stimulation of matrix production by serum or TGF$l and (2) an increase in the rate of cell division consequent to the presence of the matrix. The ability of isolated ECM to stimulate division and the abolition of the mitogenic response to serum and TGF-@1 by GRGDSP peptides are consistent with this suggestion. It is suggested that the principal function of the fibronectin-rich matrix in this system is to provide anchorage for cell proliferation. The fibronectin-rich matrix in the three-dimensional aggregate is presumed to serve the same anchoring function as the adhesive glass or plastic surface in the two-dimensional monolayer culture system (Macpherson and Montagnier, 1964). Thus, in the three-dimensional system, mitogens such as TGF-/31 may operate by different cellular pathways than mitogens in monolayer culture, where mitogen activation of tyrosine kinase and phosphoinositide signaling pathways that directly activate a release from mitotic quiescence are the presumed cellular response. The demonstrated presence of TGF-/31 in the endocardial cushion mesenchyme in viva in the mouse (Heine et al., 1987; Lehnert and Akhurst, 1988) and chick (M. Choy et ak, submitted for publication) embryos suggests a growth stimulatory role in the embryo similar to that found in cultured aggregates of endocardial cushion mesenchyme. This work was supported by a grant-in-aid Heart Association and NIH Grant GM 35185.
from the American
REFERENCES ARMSTRONG, M. T., and ARMSTRONG, P. B. (1978). Cell motility in fibroblast aggregates. J. CeU Sci 33,37-52. ARMSTRONG, P. B., and ARMSTRONG, M. T. (1990). An instructive role for the interstitial matrix in tissue patterning: Tissue segregation and intercellular invasion. J. CeU Biol. 110,1439-1455. ENGVALL, E., and RUOSLAHTI, E. (1977). Binding of soluble form of fibroblast surface protein, fibronectin, to collagen. ht. J. Cancer 20, l-5. FAIRBAIRN, S., GILBERT, R., OJAKIAN, G., SCHWIMYER, R., and QUIGLEY, J. P. (1985). The extracellular matrix of normal chick embryo
TGF-@I and Cell Prolzferatim
425
fibroblasts: Its effect on transformed chick fibroblasts and its proteolytic degradation by the transformants. J. Cell BioL 101,17991798. HAMBURGER, V., and HAMILTON, H. L. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. 38,49-92. HEINE, U. I., MUNOZ, E. F., FLANDERS, K. C., ELLINGSWORTH, L. R., LAM, H. Y. P., THOMPSON, N. L., ROBERTS, A. B., and SPORN, M. B. (1987). Role of transforming growth factor-b in the development of the mouse embryo. J. CeU BioZ. 105,2861-2876. IGNOTZ, R. A., and MASSAGU~, J. (1986). Transforming growth factorj3 stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J. BioL Chem. 261,43374345. KOPRIWA, B. M., and LEBLOND, C. P. (1962). Improvements in the coating technique of radioautography. J. HistochemC&&em 10,269284. LEHNERT, S. A., and AKHURST, R. J. (1988). Embryonic expression pattern of TGF beta type-l RNA suggests both paracrine and autocrine mechanisms of action. Development 104,263-273. MACPHERSON, I., and MONTAGNIER, L. (1964). Agar suspension culture for the selective assay of cells transformed by polyoma virus. firology 23,291-294. MJAATVEDT, C. H., LEPERA, R. C., and MARKWALD, R. R. (1987). Myocardial specificity for initiating endothelial-mesenchymal cell transition in embryonic chick heart correlates with a particulate distribution of fibronectin. Den Biol. 119,59-67. PIERSCHBACHER,M. D., and RUOSLAHTI, E. (1987). Influence of stereochemistry of the sequence Arg-Gly-Asp-Xaa on binding specificity in cell adhesion. J. BioL Chem 262, 17,294-17,298. Pcrrrs, J. D., and RUNYAN, R. B. (1989). Epithelial-mesenchymal cell transformation in the embryonic heart can be mediated, in part, by transforming growth factor /3.Dev. BioZ. 134,392-401. ROBERTS, A. B., FLANDERS, K. C., KONDAIAH, P., THOMPSON, N. L., VAN OBBERGHEN-SCHILLING, E., WAKEFIELD, L., ROSSI, P., DE CROMBRUGGHE, B., HEINE, U., and SPORN, M. B. (1988). Transforming growth factor+: Biochemistry and roles in embryogenesis, tissue repair and remodeling, and carcinogenesis. Recent Prog. Hwm. Res. 44,157-197. RUOSLAHTI, E. (1988). Fibronectin and its receptors. Annu. Rev. Bie hem. 57,375-413. YAMADA, K. M. (1978). Immunological characterization of a major transformation-sensitive fibrohlast cell surface glycoprotein. J. Cell Biol. 78,520~541. YAMADA, K. M., and KENNEDY, D. W. (1984). Dualistic nature of adhesive protein function: Fibronectin and its biologically active peptide fragments can autoinhibit fibronectin function. J. Cell Biol 99,2936.
BIOLOGY
141,421-425 (1990)
Regulation of Proliferation of Embryonic Heart Mesenchyme: Role of Transforming Growth Factor-PI and the Interstitial Matrix MICHAEL *Division
of Pediatric
CHOY,*,’
MARGARET
T. ARMSTRONG,?
AND PETER
B. ARMSTRONG?
Cardiology, Department of Pediatrics, University of Calzfornia Davis Medical Center, Sacranumto, and TDepartment of Zoology, University of Cal$n-nia at Davis, Davis, California 95616 Accepted
California 95817,
June 20, 1990
Proliferation of atrioventricular cushion mesenchyme of the embryonic avian heart maintained in three-dimensional aggregate culture is stimulated by interaction with the interstitial matrix. Chicken serum or transforming growth factor-pl, which stimulates proliferation, induces matrix deposition in regions of the aggregate showing high labeling indices with tritiated thymidine. Dispersed heart mesenchyme interstitial matrix introduced into serum-free culture is incorporated into the aggregate and stimulates cellular proliferation similar to serum or transforming growth factorpl. Proliferation is reversibly inhibited by the peptide Gly-Arg-Gly-Asp-Ser-Pro. It is suggested that transforming growth factor-(31 stimulates the production of interstitial matrix and that a sufficient stimulus for proliferation in this system is the presence of the matrix, which acts as the adhesive support for cellular anchorage. o XIO Academic PEW, IM. INTRODUCTION
Transforming growth factor-p (TGF-P)’ type I is a pleiotropic peptide mediator that affects cellular proliferation (Roberts et aZ., 1988) and stimulates production of the interstitial extracellular matrix (Ignotz and Massague 1986). TGF-/31 is present in a number of mesenchymal tissues during embryonic development (Heine et ak, 1987; Lehnert and Akhurst, 1988), where it may conceivably play a role in morphogenesis and tissue growth. One tissue with high levels of TGF-Bl is the mesenthyme of the atrioventricular valves of the embryonic heart of the mouse (Heine et al., 198’7; Lehnert and Akhurst, 1988) and chick (M. Choy et ah, submitted for publication). This tissue is established at stage 15 (Hamburger and Hamilton, 1951) of chick embryonic development when the mesenchyme precursor cells are differentiated from endocardium possibly under the influence of fibronectin and TGF-/31 (Mjaatvedt et CL,!., 1987; Potts and Runyan, 1989). Mesenchymal cells migrate into the adjacent extracellular matrix (ECM) separating myocardium and endocardium and then proliferate to establish the primordia for the endocardial cushions and subsequent atrioventricular valves. To investigate the possible involvement of TGF-Pl in later stages of heart morphogenesis, we developed a novel ’ To whom all correspondence should be addressed. ’ Abbreviations used: ANOVA, analysis of variance; BSA, bovine serum albumin; CS, chicken serum; DMEM, Dulbecco’s-modified Eagles medium; ECM, extracellular matrix; GRGDSP, Gly-Arg-GlyAsp-Ser-Pro peptide; GRGESP, Gly-Arg-Gly-Glu-Ser-Pro peptide; TGF-0, transforming growth factor-@ 421
culture assay in which heart mesenchyme from the atrioventricular valve region was separated from the myocardium and then was maintained in organ culture as three-dimensional aggregates. It was reasoned that this established an organization that better reflects the three-dimensional organization of heart mesenchyme in tivo than the conventional monolayer culture. Proliferation was determined by the incorporation of tritiated thymidine. Our observations support the suggestion that TGF-pl stimulates the production of an interstitial matrix and that the presence of the adhesive matrix alone is a sufficient stimulus for proliferation. MATERIALS
AND METHODS
Ceil and Aggregate Culture The hearts of lo-day-old White Leghorn chick embryos were isolated, and the atrioventricular regions, including the atrioventricular valves, were carefully dissected out in a manner to exclude the epicardial mesenchyme. The atrioventricular tissue was dissociated in 1.0 mg/ml trypsin, the cell suspension was freed of myocytes leaving less than 1% of myocytes and unidentified epithelioid cells (Armstrong and Armstrong, 1978), and the remaining mesenchymal cells were grown to eonfluence in Dulbecco’s-modified Eagles medium (DMEM) plus 10% chicken serum (CS). Five by 5-mm segments of the confluent monolayers were scraped gently from the culture plates using a small rubber scrapper and transferred to a gyratory shaker flask culture in DMEM plus 1.0 mg/ml of bovine serum albumin (BSA). The medium did not contain any 001%1606/‘90 $3.00 Copyright 0 1990 by Academic Press, Inc. All rigbta of reproduction in any form reserved.
422
DEVELOPMENTALBIOLOGY
growth factors, but it did allow the tissue to remain cohesive. After 4 days the sheets of cells had remodeled into smoothed-surface, spherical cell aggregates that ranged between 0.2 and 0.5 mm in diameter. Aggregates were then transferred into experimental culture media and incubated further in a shaker flask culture. The potential effecters were added to basal medium (DMEM plus 1.0 mg/ml BSA), with 2.0 &i/ml tritiated thymidine to allow a determination of the effects of the experimental conditions on cell proliferation. Fibronectin-free serum was depleted of fibronectin by passage over a gelatin-Sepharose column (Engvall and Ruoslahti, 1977). Isolation. of the Interstitial
Matrix
Monolayers of cardiac mesenchymal cells grown in DMEM plus 10% CS produce an ECM composed predominantly of fibronectin and containing lesser amounts of tenascin, laminin, and collagen type I (Armstrong and Armstrong, 1990). The matrix was prepared by hypotonic ammonium hydroxide extraction of the monolayers described by Fairbairn et al. (1985). Transfbrming
Growth Factor-@1 and Synthetic
Peptides
TGF-Pl was obtained from Calbiochem (Cat. No. 619350) and R + D (Cat. No. lOl-Bl). Gly-Arg-GlyAsp-Ser-Pro (GRGDSP) and Gly-Arg-Gly-Glu-SerPro (GRGESP) peptides were synthesized as described by Armstrong and Armstrong (1990). Immunohistochemistry
and Autoradiography
All the aggregates were fixed in Carnoy’s fixative, embedded in paraffin, sectioned, and mounted on gelatin-coated slides. The sections of aggregates were stained for fibronectin with goat anti-chicken fibronectin antibody obtained from Dr. K. Yamada (Yamada, 1978) by the ABC indirect immunoperoxidase procedure (Vector Laboratories) or with the rhodamine immunofluorescence procedure. Control sections were stained with preimmune serum or fibronectin-adsorbed antiserum substituted for the first antibody. The slides destined for autoradiography were coated with Kodak NTB-2 photographic emulsion and were processed as described by Kopriwa and Leblond (1962). Within the aggregate two regions were identifiedcortex and medulla. The cortex consisted of the outer layers of cells where extracellular fibronectin was present in some of the groups of aggregates. If fibronectin was absent, then the cortical layer was arbitrarily defined as extending from the surface of the aggregate to four to five cell layers deep. The medulla included the remaining interior cells. The labeling index of each region of the aggregate was calculated by dividing the
VOLUME 141.1990
number of cells labeled with tritiated total number of cells in the particular plying by 100.
thymidine by the region and multi-
Statistics Ten to 20 aggregates were placed in a shaker flask with the potential effector or effecters, and each potential effector was tried at least five times on different sets of aggregates. Since all results were repeatable, the statistical analysis was performed on one set of experiments. In each of the four major experiments, five aggregates were randomly selected from each group, and the labeling indices were determined. Each experiment was analyzed with several statistical methods described in Table 1. RESULTS
Stimulation
AND
of Cell Prolifwation
DISCUSSION
by TGF-pl
and Serum
Fibronectin, which was a major component of the ECM of the mesenchymal cell monolayer, was lost by the second day of aggregate culture. After 4 days of culture in basal medium, aggregates were transferred to experimental conditions (Experiment 1) in fresh culture flasks, and the effects on cell proliferation were assessed by determining the tritiated thymidine incorporation index. Cells transferred to fresh basal medium showed a labeling index that was relatively low, varying between 23% in the interior of the aggregate to 31% at the surface (Fig. lA, Table 1). A closer examination of the individual cells of the aggregates revealed that the cell shape tended to be stellate throughout the aggregate. Aggregates transferred to basal medium containing 0.1 or 1.0 nMTGF-Bl had different morphology and proliferative behavior (Experiments 1 and 4). After 2 days of incubation, an extracellular fibronectin matrix was present in the periphery or cortex of the aggregates in a zone that extended four to five cell layers deep (Fig. 1B). The inner portion of the aggregates, the medulla, was fibronectin poor. The boundary between the two regions was sharp. Interestingly, the cells in the cortex tended to be flattened, while the cells in the medulla remained stellate. The labeling index of the cortex was elevated (80%) compared to the medulla (22%) or to either region of the control aggregates maintained in basal medium (Fig. lB, Table 1). Thus, exposure to TGF-Bl induced the formation of a fibronectin-rich ECM and stimulated cell proliferation. The inclusion of 3 or 10% CS had an identical effect to that of TGF-01 on matrix deposition, cell morphology, and cell proliferation (Experiments 2 and 3). After 1 day of incubation, aggregates developed a cortical region of
CHOY,
ARMSTRONG,
423
TGF-01 and Cell Prolgeratim
AND ARMSTRONG
TABLE 1 REGIONAL
Medium
Cortex Experiment
BSA TGF-01 (1.0 nM)b
31 + 1.6” 80 f 1.3”,”
25 80 33 72
Medulla
INDICES
Medium
1
Experiment BSA cs (3%)” GRGDSP + CS GRGESP + CSb
LABELING
+ f + -t
Experiment 23 f 1.0 22 f 1.5
BSA cs (lO%)b ECMb
3 1.4 1.4”’ 2.9 1.6””
Cortex
19 f 24 f 15 f 18 f
1.1 1.3 3.3 0.9
BSA TGF-@l (0.1 nM)b TGF-@l (1.0 n&f) b GRGDSP + TGF-01 GRGESP + TGF-fll
2
9 + 0.7 81 + 2.1’!’ 66 f 1.6”,” Experiment 23 67 67 18 51
Medulla
5 * 0.2 7 k 0.9 5 f 0.7
4 f 3.3 f 1.7”~” f 2.5”” I? 1.0” f 3.0’”
24 29 26 13 15
f + f f +
4.2 1.2 1.6 0.8 0.9
Note. The labeling index of each region was calculated by dividing the number of cells undergoing DNA synthesis by the total number of cells in the particular region and multiplying by 100. Values are the means f SEM of five randomly chosen aggregates within the designated group. Abbreviations: ANOVA, analysis of variance; BSA, bovine serum albumin; CS, chicken serum; ECM, sonicated fibronectin-rich extracellular matrix; GRGDSP, Gly-Arg-Gly-Asp-Ser-Pro peptide; GRGESP, Gly-Arg-Gly-Glu-Ser-Pro peptide; TGF-01, transforming growth factor-81. a Significantly different between regions within a group, P < 0.01, Student’s paired t test. b Significantly different from BSA group, P < 0.01, repeated-measures ANOVA with one grouping factor (the medium) and one within factor (medulla/cortex), Scheffe F test. ’ Significantly different from corresponding region of BSA group, P i 0.01, one-way ANOVA, Scheffe F test.
flattened, mitotically active cells (Table 1) that were associated with a fibronectin-rich ECM (Fig. 1C). The cells in the interior remained stellate and showed a low labeling index and a sparse matrix. The medullary tissue was viable with an identical rate of protein synthesis as the cortical tissue (Armstrong and Armstrong, 1990). Fibronectin-depleted CS had the same effect on matrix deposition and cell proliferation as complete CS (data not shown), suggesting that serum fibronectin was not the factor in the CS responsible for stimulating matrix deposition and proliferation. The matrix stimulated by serum and TGF-01 contained a variety of adhesive glycoproteins, including fibronectin, laminin, tenascin, and small amounts of collagen type I (Armstrong and Armstrong, 1990).
Role of the Extracellular
Matrix
in Cell Proliferation
The mitotically active cortex of TGF+land serumtreated aggregates is characterized by the presence of a prominent fibronectin-rich ECM and a flattened cell morphology. To study the role of the matrix in the regulation of cell shape and cell proliferation, aggregates were transferred to basal medium supplemented with a sonicated preparation of heart mesenchyme ECM (Experiment 2). The sonicated matrix incorporated selectively into the cortical zones of the aggregates (Armstrong and Armstrong, 1990) (Fig. 1D). The cortical cells associated with the matrix developed the flattened mor-
phology and the increased proliferative rate similar to aggregates exposed to TGF-01 or serum (Table 1). Further evidence for the involvement of the ECM in the regulation of cell morphology and proliferation is provided by the perturbation of matrix deposition by peptides containing the sequence Arg-Gly-Asp (Ruoslahti, 1988; Yamada and Kennedy 1984). Peptides with substitutions at the aspartate residue are, typically, markedly less inhibitory and are used as negative controls. Since Arg-Gly-Asp-containing peptides containing the usual L-isomer of arginine were inactivated within a few hours in this system (Armstrong and Armstrong, 1990), we used peptides containing the D-isomer of arginine because they proved to be stable and biologically active (Pierschbacher and Ruoslahti, 1987). GRGDSP (1.0 mg/ml) prevented the formation of the cortical fibronectin-rich ECM in aggregates exposed to 3% serum (Experiment 3) or 1.0 nM TGF-fil (Experiment 4). The cortical cells failed to develop the flattened morphology and revealed a low tritiated thymidine labeling index (Fig. lE, Table 1). GRGDSP did not appear to be toxic since its effects were reversible: aggregates that were transferred from peptide-containing medium to peptide-free medium established a cortical fibronectin matrix and showed a flattened cell morphology and a high proliferative index (data not shown). In addition, the GRGESP control (1.0 mg/ml) did not inhibit fibronectin production or cell proliferation in the cortical layers of the aggregates. Fibronectin is a matrix adhesion molecule that has
424
DEVELOPMENTALBIOLOGY
VOLUME141,1%0
FIG. 1. Autoradiograms of aggregates in different media containing tritiated thymidine and stained for fihronectin by a standard immunoperoxidase technique. Bar = 0.10 mm. (A) In basal medium, fibronectin ECM is absent. Labeled nuclei are present, but are relatively few both in the outer layer of cells called the cortex (c) and in the interior, the medulla (m). (B) In basal medium containingTGF+l, fibronectin appears as a gray fibrillar ECM in the cortical region. The percentage of mitotically active cells is much greater in the cortex than that of the medulla. (C) In CS, the cortex contains a fibrillar matrix of fibronectin and abundant labeled nuclei. (D) In basal medium supplemented with a sonicated preparation of the fibronectin-rich ECM isolated from confluent heart mesenchymal monolayer, staining for fibronectin demonstrates the cortical localization of the ECM. Similar to aggregates grown in TGF-j31 or CS, the cortical cells are mitotically active. (E) Cultured in medium containing both CS and the peptide GRGDSP, which binds to the fibronectin receptor sites of the mesenchymal cells, the intact fibronectin-rich matrix is absent, and cell proliferation in both regions is low.
several functional domains, including binding sites for structural elements of the ECM (collagen, glycosaminoglycans, and fibrin) and for receptors on the surfaces of cells (Ruoslahti, 1988). The multiplicity of binding sites
enables fibronectin to link cells to collagen and other components of the matrix. The colocalization of matrix and the region of active proliferation suggests that the mitogenic response in the three-dimensional system
CHOY, ARMSTRONG, AND ARMSTRONG
may consist of two steps: (1) the stimulation of matrix production by serum or TGF$l and (2) an increase in the rate of cell division consequent to the presence of the matrix. The ability of isolated ECM to stimulate division and the abolition of the mitogenic response to serum and TGF-@1 by GRGDSP peptides are consistent with this suggestion. It is suggested that the principal function of the fibronectin-rich matrix in this system is to provide anchorage for cell proliferation. The fibronectin-rich matrix in the three-dimensional aggregate is presumed to serve the same anchoring function as the adhesive glass or plastic surface in the two-dimensional monolayer culture system (Macpherson and Montagnier, 1964). Thus, in the three-dimensional system, mitogens such as TGF-/31 may operate by different cellular pathways than mitogens in monolayer culture, where mitogen activation of tyrosine kinase and phosphoinositide signaling pathways that directly activate a release from mitotic quiescence are the presumed cellular response. The demonstrated presence of TGF-/31 in the endocardial cushion mesenchyme in viva in the mouse (Heine et al., 1987; Lehnert and Akhurst, 1988) and chick (M. Choy et ak, submitted for publication) embryos suggests a growth stimulatory role in the embryo similar to that found in cultured aggregates of endocardial cushion mesenchyme. This work was supported by a grant-in-aid Heart Association and NIH Grant GM 35185.
from the American
REFERENCES ARMSTRONG, M. T., and ARMSTRONG, P. B. (1978). Cell motility in fibroblast aggregates. J. CeU Sci 33,37-52. ARMSTRONG, P. B., and ARMSTRONG, M. T. (1990). An instructive role for the interstitial matrix in tissue patterning: Tissue segregation and intercellular invasion. J. CeU Biol. 110,1439-1455. ENGVALL, E., and RUOSLAHTI, E. (1977). Binding of soluble form of fibroblast surface protein, fibronectin, to collagen. ht. J. Cancer 20, l-5. FAIRBAIRN, S., GILBERT, R., OJAKIAN, G., SCHWIMYER, R., and QUIGLEY, J. P. (1985). The extracellular matrix of normal chick embryo
TGF-@I and Cell Prolzferatim
425
fibroblasts: Its effect on transformed chick fibroblasts and its proteolytic degradation by the transformants. J. Cell BioL 101,17991798. HAMBURGER, V., and HAMILTON, H. L. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. 38,49-92. HEINE, U. I., MUNOZ, E. F., FLANDERS, K. C., ELLINGSWORTH, L. R., LAM, H. Y. P., THOMPSON, N. L., ROBERTS, A. B., and SPORN, M. B. (1987). Role of transforming growth factor-b in the development of the mouse embryo. J. CeU BioZ. 105,2861-2876. IGNOTZ, R. A., and MASSAGU~, J. (1986). Transforming growth factorj3 stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J. BioL Chem. 261,43374345. KOPRIWA, B. M., and LEBLOND, C. P. (1962). Improvements in the coating technique of radioautography. J. HistochemC&&em 10,269284. LEHNERT, S. A., and AKHURST, R. J. (1988). Embryonic expression pattern of TGF beta type-l RNA suggests both paracrine and autocrine mechanisms of action. Development 104,263-273. MACPHERSON, I., and MONTAGNIER, L. (1964). Agar suspension culture for the selective assay of cells transformed by polyoma virus. firology 23,291-294. MJAATVEDT, C. H., LEPERA, R. C., and MARKWALD, R. R. (1987). Myocardial specificity for initiating endothelial-mesenchymal cell transition in embryonic chick heart correlates with a particulate distribution of fibronectin. Den Biol. 119,59-67. PIERSCHBACHER,M. D., and RUOSLAHTI, E. (1987). Influence of stereochemistry of the sequence Arg-Gly-Asp-Xaa on binding specificity in cell adhesion. J. BioL Chem 262, 17,294-17,298. Pcrrrs, J. D., and RUNYAN, R. B. (1989). Epithelial-mesenchymal cell transformation in the embryonic heart can be mediated, in part, by transforming growth factor /3.Dev. BioZ. 134,392-401. ROBERTS, A. B., FLANDERS, K. C., KONDAIAH, P., THOMPSON, N. L., VAN OBBERGHEN-SCHILLING, E., WAKEFIELD, L., ROSSI, P., DE CROMBRUGGHE, B., HEINE, U., and SPORN, M. B. (1988). Transforming growth factor+: Biochemistry and roles in embryogenesis, tissue repair and remodeling, and carcinogenesis. Recent Prog. Hwm. Res. 44,157-197. RUOSLAHTI, E. (1988). Fibronectin and its receptors. Annu. Rev. Bie hem. 57,375-413. YAMADA, K. M. (1978). Immunological characterization of a major transformation-sensitive fibrohlast cell surface glycoprotein. J. Cell Biol. 78,520~541. YAMADA, K. M., and KENNEDY, D. W. (1984). Dualistic nature of adhesive protein function: Fibronectin and its biologically active peptide fragments can autoinhibit fibronectin function. J. Cell Biol 99,2936.
