J, Cell Sci. Suppl. 13, 131-138 (1990) Printed in Great Britain © The Company o f Biologists Limited 1990
131
Transforming growth factor-/} receptors and binding proteoglycans FREDERICK T. BOYD*, SELA CHEIFETZ, JANET ANDRES, MARIKKI LAIHO a n d JOAN MASSAGUE Cell Biology and Genetics Program, M em orial Sloan-Kettering Cancer Center, SloanKettering Division o f the Graduate School o f M edical Sciences, Cornell University, New York, New York 10021, USA
Summary Transforming growth factors-beta (TGFs-/?) are representative of a superfamily whose members were first identified as regulators of morphogenesis and differentiation, and subsequently found to be structurally related. Other members of the family include the activins and inhibins, BMPs, MIS, the DPP-C gene product and Vg-1. When assayed by affinity-labelling techniques, TGFs-/? bind to three distinct cell surface proteins which are present on most cells. These proteins are all of relatively low abundance but bind TGFs-/? with affinities consistent with the biological potency of the factors. The Type I and Type II binding proteins are glycoproteins with estimated molecular weights of 53 and 7 3 x l 0 3Mr, respectively. They both bind TGF-/31 significantly better than TGF-/32. The Type I receptor has been identified as the receptor which mediates many of the responses of TGFs-/?, based on somatic cell genetic studies of epithelial cell mutants unresponsive to TGFs-/J. Betaglycan is the third binding protein present on many, but not all, cell types and is a large proteoglycan (~ 2 8 0 x l 0 3Mr) with 1 0 0 -1 2 0 x l 0 3 Mr core proteins. A soluble form of this molecule is present in conditioned media of many cell lines and may be derived from the cell surface-associated molecule by cleavage of a small membrane anchor. Betaglycan binds TGF-/31 and TGF-/J2 with similar affinity and this binding is to the core proteins, not the glycosaminoglycan side chains. This molecule may have a function in the localization and delivery or the clearance of activated TGFs-/i The molecular basis of TGF-/3 signalling is still largely unknown, but it is possible th at one or more of these cell surface molecules signals via a novel mechanism, as the TGFs-/3 are biologically quite distinct from other factors th at act via well-characterized signalling systems.
Introduction Transforming growth factors-beta (TGFs-¡3) are a family of polypeptide hormones which probably act over relatively confined physiological spaces in vivo. They are concentrated in platelets and released in an inactive form upon platelet degranulation (Pircher et al. 1986). Presumably they are activated in a local area by proteases or other specific activating conditions (Lyons et al. 1988). In addition, in situ hybridization and immunohistochemical studies have demonstrated that the factors are synthesized in many in vivo sites (Heine et al. 1987; Lehnert and Adhurst, 1988; Pelton et al. 1989). Our laboratory has attempted to identify and * Present address: Laboratory of Medicine and Pathology, University of Minnesota, Minneapolis, Minnesota 55455, USA. Key words: Transforming growth factor-beta (TGF-/3), receptors, binding proteoglycans.
132
F. T. Boyd et al.
characterize those cell surface proteins with which TGFs-/3 can interact and which presumably mediate the biological signalling of the factor into the cell. This is a somewhat complicated task for a number of reasons. As indicated above, TGF-/2 is a family of five closely related factors (TGF-/S1 to TGF-/35) within a superfamily of several other factors implicated in essential elements of development. These other factors include the decapentaplegic gene product of Drosophila, the Vg-1 gene product and mesoderm inducing factor in Xenopus, Mullerian Inhibiting Substance, the Bone Morphogenic Proteins, and the activins and inhibins (Massague, 1990). This suggests that there may be a family of related receptors with differential affinities for different members of the superfamily. Another complicating element is that the known effects of TGF-/? are varied and in some instances diametrically opposed in different cell systems. It is intriguing that TGF-/3 has a potent growth inhibitory effect on some epithelial cell lines, but it is also mitogenic in some culture systems, such as osteoblasts (Centrella et al. 1987) and some fibroblast lines (Leof et al. 1986). It is an inhibitor of differentiation in myogenic and adipogenic model systems in vitro (Massague et al. 1986; Ignotz and Massague, 1985), but stimulates differentiation of chondroblasts (Seyedin et al. 1985). And finally, experimentally, we have found that there is a relatively low level of specific cell-surface TGF-/3 binding and that this activity is distributed among three distinct cell surface binding proteins. Distribution of specific TGF-^-binding proteins In the course of investigating the activities of TGF-/3 over the last six years, our laboratory has characterized the TGF-/5 binding patterns of over one hundred different cell lines, primary cells and tissues (Fig. 1A) (Massague, 1990). It is striking that the general pattern is so similar in most cells analyzed. Briefly, the experimental approach is to bind iodinated TGF-/3 to cells, crosslink the ligand to cell surface proteins to which it is associated with a bifunctional crosslinking reagent such as disuccinimidyl suberate, solubilize the cell membranes with detergent and separate the labeled proteins on SDS-polyacrylamide gels with subsequent autoradiography. This method commonly identifies three proteins Fig. 1. Distribution of TGF-/3 receptors and binding proteoglycans in various cell types. A. Summary of TGF-/3 receptors in various cell types. All cell lines were screened for the presence of TGF-/3 receptors (I, II, III) using the affinity-labelling protocol outlined in the text. The presence of a receptor type in any cell line is signified with an open circle (O). The majority of cell lines screened express all three protein species, some lines express only the Type II and Type I receptors, a few lines express only the Type I receptor. No cell line which responds to TGF-/3 with established assays lacks the Type I receptor. Cell lines which lack any receptor type are signified with a closed circle (• ) under the column relating to th at receptor type. B. Receptor profiles from representative cell lines. Mouse BALB/c-3T3 cells, 3T3-L1 cells, ra t NRK cells, chick embryo fibroblasts (CEF) and mink lung epithelial cells (MvlLu) were screened by affinity labelling with 50 pM 125I-TGF-/3 in the presence of no (a), 50 (b), 100 (c), 200 (d), 700 (e), or 3000 (f) p M native TGF-/3. These experiments show the affinity of these proteins for TGF-/3 and the general nature of the affinity labeled species (Cheifetz et al. 1986).
133
TGF-ß receptors and binding proteins Huma n N o r ma l G M 3 1 6 s k i n f ib r o b l a s t s G M 3 7 0 s k i n f ib r o b l a s t s G M 5 8 7 7 s k l n f ib r o b l a s t s G M 1 2 6 2 c o n j u n c t i v e c e lls P r i m a r y lu n g f ib r o b la s t P r i m a r y s k i n f ib r o b la s t s U i l - 3 8 E m b r y o lu n g f ib r o b la s t s I M R - 9 0 E m b r y o lu n g f ib r o b la s t s R e t in a , p r im a r y E1M C3T 3 O ste o b la sts M C 3T 3 O st e o b la st s H S 1 4 M u s c le
1 o o o o o o o o o o o o
Huma n Tr a ns f o i
ea
79 2 2 O ste o sa rc o m a R 1684 C h o n d ro sa rc o m a H T -1 0 8 0 F i b r o s a r c o m a M N N G -H O S O s t e o g e n ic s a r c o m a U R - 1 3 f ib r o b l a s t S U - 4 0 U le ri R e t i n o b l a s t o m a R B 7 R e t in o b la s t o m a V 7 9 R e t in o b la st o m a R B 2 7 R e t in o b la st o m a R B I 3 R e t in o b la st o m a R B 2 0 R e t in o b la st o m a R B 2 2 R e t in o b la st o m a R B 2 4 R e t in o b la st o m a
o o o • • • • • o o o o o o o o o o o o
FR S k in f ib r o b la s t s N R K K id n e y f i b r o b l a s t s H 9 H e a rt m y o b l a s t L 8 S k e le ta l m u s c le m y o b la s t s R 7 r S R o r ta s m o o t h m u s c l e m y o b l a s t B R L - 3 A N e w b o r n r a t l iu e r P r i m a r y c a lu a r ia o s t e o b l a s t s P r im a r y h e p a t o c y te s P r im a r y c a r d io c y t e s P r i m a r y p it u i t a r y F e ta l t issu e P la c e n t a
o o o o o o • o o o o o o o o o o o o o o • • • • • •
280^
b
i
c
d
u
O O O O O O o o o o o o
o o o o
• o o o o o o o o o
o o o • o o o o o o o o
p
\o
o o o o o o
Rat T r a n s f o r m e d o o o o • o o o
F e S U - F i s h e r r a t e m b r y o f ib r o b la s t t s K i M S U - N R K K i d n e y e p it h e lia l C 6 G liom a H 3 5 H e p a to m a PC12 P h e o ch ro m o cy to m a L 6 S k e le t a l m u s c l e m y o b l a s t L 6 E 9 S k e le t a l M u s c l e M y o b l o s t L6E9 m y o tu b e s
Mouse
N o r ma l
e
• • • •
o o o o o
o o o o o
S w i s s 3T3 e m b r y o f i b r o b l a s t s S w i s s 3 T 3 / N R -6 e m b r y o f ib r o b la s t s B a lb / c 3T3 e m b r y o f i b r o b l a s t s 3 T 3 - L I p re a d ip o c y t e 3 T 3 - L I a d ip o c y t e
o o o •
o o o o • o o o
o o o o o
a
i
b
c
m
d
f
116 S" 94-
i
3 -.140
"*130
-
85 * -
"
68-
CEF d
M I S
MvlLu e
f
a
- 280 ^
I
200-
11694_ 68 -
85 *65
c
Ruian C h ic k e m b r y o f ib r o b l a s t s
Monkey B S C - 1 K i d n e y e p it h e lia l C O S K i d n e y f i b r o b la s t - li k e
110 85»- * -6 5 ►
b
§
c
d
e
f
O O O O • • 9 9 * 0
o o o
o • o o o
o o o o o
o
o
o
o o
o o
o o
o
o
o
o
o
•
o
o
•
O •
Mink Bouine FBH E F e t a l h e a r t e n d o t h e l ia l
Dog M D C K k i d n e y e p it h e lia l
Hamster DOT! M y o s a r c o m a C C L - 3 9 L u n g F ib r o b l a s t BH K k id n e y C H 0 - K 1 O u a ry CHO p g s f l- 7 4 5 CHO Id ID
a
I » I I
o
b
3T3 M S U - t r e n s f o r m e d f ib r o b la s t s N 2R N e u ro b la st o m a E L S t r a i n E, R s c it e s c a r c i n o m a C 1 2 7 I M a m m a ry tum or M B 6 6 - M C R 1 0 T I/ 2 F i b r o b l a s t
NRK e
O O O 0 0
M o u s e Transformed
3 T 3 -LI f
200-
a
M K K e r a t in o c y t e R K R - 2 B F ib r o b la st B 6 S U t R H e m a t o p o ie t ic p r o g e n i t o r s B 6 S U t- C L 2 7 h e m a t o p o ie t ic p r o g e n it o r s 3 2 D - C 1 3 H e m a t o p o ie t ic p r o g e n i t o r s
M u l L u (C C L 6 4 ) L u n g e p it h e lia l
BALB/C-3T3 a
• • • o o o o o o o
Rat N o r ma l
•
B
O O O O O O O O O O
000000
o o o o o o o o o o o o o o o o o o o o
R 4 3 1 E p it h e lio id c a r c in o m a T 2 4 B l a d d e r c a r c in o m a M D R - M B 2 3 1 B r e a s t c a r c in o m a M C F - 7 / 2 B r e a s t c a r c in o m a M C F - 7 / 4 B r e a s t c a r c in o m a M C F - 7 / 5 B r e a s t c a r c in o m a M C F - 7 / 1 B r e a s t c a r c in o m a S C C 2 S S q u a m o u s c e ll c a r c in o m a R 875 M e la n o m a R 5 4 9 L u n g a d e n o c a r c in o m a B 1 6 F I M e la n o m a B 6F10 M e la n o m a H5 2 2 4 R h a b d o m y o s a rc o m a M G -6 3 O ste o sa rc o m a O S - 1P O s t e o s a r c o m a
H L -6 0 M o n o m y e lo c y t ic le u k . (p h a g e ) U 9 3 7 M o n o c y t ic le u k e m ia IM R 3 2 N e u r o b l a s t o m a S K -N -S H N e u ro b la st o m a H S4 Sch w a n o m a H S 0 9 2 9 LU ilm 's t u m o r R D - 1 2 R e t in a , a d e n o u ir u s G M 0 I 9 7 2 R P r o g e r ia T E 6 7 1 M e d u ll o b l a s t o m a R D L - D G lio b la s t o m a
III o o o • o o o o o o o o
11 o o o o o o o o o o o o
b
c
d
e
f
134
F. T. Boyd et al.
labeled specifically (Fig. IB). The Type I protein is an affinity-labeled species of approximately 65 x 103 Mr. The Type II species is approximately 85 x 103 Mz, and the Type III species, termed betaglycan, is a broad band typically centered around 2 8 0 x l0 3Mr. All of these apparent molecular weights include an associated monomer of TGF-/? of 12.5 X103 Mr, so the presumed size of the binding proteins is correspondingly 12.5 x 103 Mr smaller than the apparent molecular weight on SDS gels. Each of these proteins have high affinities for TGFs-/?, with Kd values in the range of 5-500 pM . They bind TGF-/31, TGF-/32 and TGF-/33, but not more distantly related members of the TGF-/? superfamily. The Type I binding protein is ubiquitous, with every cell type that responds to TGFs-/? having the Type I protein. There are several examples of hematopoietic progenitor cell lines which respond to TGF-/31, TGF-j31.2 and TGF-/32 differentially with an order of potencies that parallels the order of affinities of the factors for the Type I protein, the only TGF-/? binding protein detectable on these cell lines (Ohta et al. 1987; Cheifetz et al. 1988). Human and bovine vascular endothelial cells possess both the Type I and Type II proteins and respond differentially to TGF-/31 or TGF-/32. In addition, L6E9 myoblasts also have only the Type I and Type II proteins, yet respond equivalently to TGF-/?1 and TGF-/32. The most common pattern of TGF-/? cell surface binding proteins is the presence of the Type I protein, Type II protein and betaglycan together. These data are suggestive, though certainly not definitive, that the Type I protein is sufficient to confer TGF-ft responsiveness to cells. However, the other TGF-/? binding proteins may modulate the availability or activities of the TGFs-/3 or mediate particular responses in particular cell types.
Somatic cell mutants defective in TGF-/J responsiveness More direct evidence supporting specific roles of these proteins in TGF-/? signalling has come from studies of somatic cell mutants in the MvlLu epithelial cell line which are non-responsive to TGFs-/?. The parental line is exquisitely sensitive to TGF-/? and is virtually completely growth inhibited by 5 p M TGF-/J1; it also responds to TGF-/? with increased expression of extracellular matrix components such as fibronectin and plasminogen activator inhibitor (PAI-1). In addition, the parental line has all three types of putative TGF-/? receptor proteins. Ethyl methane sulfonate (EMS)-mutagenized MvlLu cells were selected which were able to grow in the presence of 100 pM TGF-/31. The mutant clones were completely resistant to the growth inhibitory effects of TGF-/?1 and TGF-/?2 and have also lost all other responses to TGFs-/? assayed for. In addition, several of the clones were defective in expression of the Type I binding protein (Fig. 2). Clones of MvlLu cells obtained from a non-mutagenized population and analyzed for TGF-/3 binding proteins always have all three binding proteins present. On this basis, we suggest that the Type I protein is the receptor which mediates epithelial cell responsiveness to TGF-/? (Boyd and Massague, 1989). In addition to these receptor defective mutants, termed R mutants, mutants were isolated which have normal binding protein patterns, yet are deficient in all TGF-/3 responses. These are
TGF-/3 receptors and binding proteins
135
TGF-/3 response mutants in M v lL u cells TGF-/3 receptor profiles in CO w c c c £ CO 03 "3 -5 c "5 "5 Ei C D i_ E E CC GC CL ¿ ch Q
Betaglycan-
Type I I - *
%
Type I -►
m
•
I
■
&
I
Fig. 2. TGF-/Î receptor profiles of TGF-/3 resistant mutants of mink lung epithelial cells. Mutants of M vlLu cells which are not responsive to TGF-/Î with respect to growth inhibition were generated as described in the text. Representative clones of several mutant phenotypes were affinity labeled with 100 p M 125I-TGF-/3. Three distinct types of mutants have been isolated as characterized by TGF-/J binding profiles. When compared with the parental M vlLu cells, it is apparent that two of the mutant types have distinct deficits in TGF-/3 binding. The R-mutants are lacking the Type I receptor. In addition, the DR-mutants are lacking both the Type I and the Type II receptors. In contrast, the S-mutants have an apparently normal receptor profile.
called signalling, or S, mutants. When complementation analysis was performed, none of the mutant hybrids were complementary, suggesting that all the R and S mutants isolated are mutants in the same gene, presumably the TGF-/? receptor gene. It is also apparent from this analysis that all the mutants isolated were recessive mutations. All mutant-parental fusions were fully responsive to TGF-/? with normal receptor profiles. These studies have been extended and selection of M vlLu mutants has been performed with lower doses (25 p M) of TGF-/J2. In addition to the R and S mutant classes, a third class of mutants defective in the expression of both the Type I and Type II proteins has been isolated (DR mutants) (M. Laiho and others,
136
F. T. Boyd et al.
unpublished observations). Complementation analysis of these mutants is not complete, so the genetic basis of these mutations is still unknown. Several explanations for this phenotype are possible. DR mutants may be the result of mutations of two distinct loci encoding for the Type I and Type II proteins, although the frequency of isolation of these mutants argues strongly against this. Another possibility is that the two genes are linked and a single large deletion can knock out expression of both genes. A third possibility is that the two proteins are associated intracellularly and some mutations in one or the other of the proteins prevent the expression of both. These possibilities are the subject of active investigation in the laboratory. We are also attempting to complement these mutants by cDNA and genomic transfection to clone the genes responsible for these mutations. Betaglycan While mutant analysis has revealed the functional significance of the Type I and Type II proteins, the fact remains that in the majority of cells we have screened, the major component of cell surface TGF-/3 binding activity is associated with a large proteoglycan species with apparent molecular weight of 2 00-400x l 0 3Mr. This molecule is a complex mixed chondroitin/heparan sulfate proteoglycan (Segarini and Seyedin, 1988; Cheifetz et al. 1988) with multiple deglycosylated core proteins of 100-120x103Mr. Unlike other growth factors which associate with proteoglycans via relatively non-specific binding to the glycosaminoglycan chains, TGFs-/3 bind to betaglycan via the core proteins. The core proteins are expressed and bind TGFs-/3 in metabolic mutants which do not synthesize glycosaminoglycan side chains (Cheifetz and Massague, 1989). There is also a soluble form of betaglycan, found in the media of tissue culture cells which express betaglycan, which is capable of binding TGFs-/? (Andres et al. 1989). This form is slightly smaller than the membrane-associated form and is incapable of being incorporated into phospholipid vesicles. The soluble form of betaglycan also associates with the extracellular matrix. It is intriguing that TGF-/3, which is a well-characterized modulator of the extracellular matrix, binds specifically to a TGF-/3 binding proteoglycan which associates with the extracellular matrix and is the major species of TGF-/3 binding activity associated with cells. This may be a mechanism which can be modulated by TGFs-/3, by which TGFs-/? are sequestered in the intercellular space. A model of TGF-/3 binding protein function The definitive identification of the TGF-/3 receptor awaits cloning of the gene and reconstitution of a TGF-/3-responsive phenotype to TGF-/3-receptor mutants. However, the features of TGF-/3-binding outlined above suggests the following model. The identification of cell lines which possess only the Type I protein and the common defect in Type I binding activity in TGF-/3-response mutants suggests that the Type I protein is the signalling receptor for TGF-/3. The identification of
TGF-/3 receptors and binding proteins
137
TGF-/3-response mutants defective in both the Type I and Type II proteins suggests that the Type II protein may be part of a higher order complex with the Type I protein that associates and is a functional entity in some cell types. The dimeric structures of TGFs-¡5 are reminiscent of the structure of other dimeric growth factors which bind to dimeric receptors. Finally, betaglycan may be an extracellular storage site or mechanism of inactivation of the ligand. The physiology of the TGFs-/J suggests that specificity of action may come about as a result of acute regulation of availability of the active ligand. The fact that betaglycan is a proteoglycan and is present in membrane associated form as well as soluble and matrix associated forms suggests the possibility of acute regulation of the molecule, which might make it well suited to a role in clearance or storage of the ligands. References J . L., S t a n l e y , K., C h e i f e t z , S . a n d M a s s a g u e , J . (1989). Membrane-anchored and soluble forms of betaglycan, a polymorphic proteoglycan that binds transforming growth factor. J. Cell Biol. 109, 3137-3145. B o y d , F. T. a n d M a s s a g u jS , J . (1989). Transforming growth factor-beta-inhibition of epithelial cell proliferation linked to the expression of a 53 kD membrane receptor. J. biol. Chem. 264, 2272-78. C e n t r e l l a , M ., M c C a r t h y , T. L. a n d C a n a l i s , E. (1987). Transforming growth factor-beta is a bifunctional regulator of replication and collagen synthesis in osteoblast-enriched cell cultures from fetal ra t bone. J. biol. Chem. 262, 2 8 69-2874. C h e i f e t z , S. a n d M a s s a g u £ , J . (1989). The TGF-/3-receptor proteoglycan. Cell surface expression and ligand binding in the absence of glycosaminoglycan chains. J. biol. Chem. 264, 12 0 2 5 -1 2 0 2 8 . C h e i f e t z , S., A n d r e s , J . L. a n d M a s s a g u £ , J . (1988). The transforming growth factor-/j-receptor type III is a membrane proteoglycan. Domain structure of the receptor. J. biol. Chem. 263, 1 6 9 8 4 -1 6 9 9 1 . H e i n e , U. I., M u n o z , E . F . , F l a n d e r s , K. C., E l l i n g s w o r t h , L . R ., L a m , H. Y. P., T h o m p s o n , N. L ., R o b e r t s , A. B. a n d S p o r n , M . B. (1987). Role of transforming growth factor-/? in the development of the mouse embryo. J. Cell Biol. 105, 2 8 6 1-2867. I g n o t z , R. A . a n d M a s s a g u i s , J . (1985). Type-/? transforming growth factor controls the adipogenic differentiation of 3T3 fibroblasts. Proc. natn. Acad. Sci. U.S.A. 82, 8 5 30-8534. L e h n e r t , S . A . a n d A d h u r s t , R. J . (1988). Embryonic expression pattern of TGF-/?-type 1 RNA suggests both paracrine and autocrine mechanisms of action. Development 104, 2 6 3 -2 7 3 . L e o f , E. B., P r o p e r , J . A ., G o u s t i n , A . S . , S h i p l e y , G . D ., D i C o r l e t o , P . E. a n d M o s e s , H. L . (1986). Induction of c-sis mRNA and activity similar to platelet-derived growth factor-/} by transforming growth factor-/?: a proposed model for indirect mitogenesis involving autocrine activity. Proc. natn. Acad. Sci. U.S.A. 83, 2 4 53-2457. L y o n s , R. M ., K e s k i - O j a , J . a n d M o s e s , H. L . (1988). Proteolytic activation of latent transforming growth factor-/? from fibroblast-conditioned medium. J. Cell. Biol. 106, 1659-1665. M A S S A G u i, J . (1990). The transforming growth factor-/?-family. A. Rev. Cell Biol. 6, In press. M a s s a g u £ , J ., C h e i f e t z , S., E n d o , T. a n d N a d a l - G i n a r d , B. (1986). Type-/? transforming growth factor is an inhibitor of myogenic differentiation. Proc. natn. Acad. Sci. U.S.A. 83, 8206-8210. M i l l e r , D . A., L e e , A., M a t s u i , Y ., C h e n , E. Y ., M o s e s , H. L . a n d D e r y n c k , R. (1989). Complementary DNA cloning of the murine transforming growth factor-/?3 (TGF-/33) precursor and the comparative expression of the TGF-/?3 and TGF-/31 messenger RNA in murine embryos and adult tissues. Molec. Endocrinol. 3, 1926-1934. O h t a , M., G r e e n b e r g e r , J . S., A n k l e s a r i a , P., B a s s o l s , A . a n d M a s s a g u ^ , J . (1987). Two forms
A n d res,
138
F. T. Boyd et al.
of transforming growth factor-/? distinguished by multipotential haematopoietic progenitor cells. Nature 329, 5 3 9 -5 4 1 . P e l t o n , R. W., N o m u r a , S., M o s e s , H . L. a n d H o g a n , B. L. M . (1989). Expression of transforming growth factor-/82 R N A during murine embryogenesis. Development 106, 7 5 9 -768. P i r c h e r , R ., J u l i e n , P . a n d L a w r e n c e , D. A. (1986). Transforming growth factor-/? is stored in human blood platelets as a latent high molecular weight complex. Biochem. biophys. Res. Comm. 136, 3 0 -3 7 . S e g a r i n i , P. R. a n d S e y e d i n , S . M. (1988). The high molecular weight receptor to transforming growth factor-/? contains glycosaminoglycan chains. J. biol. Chem. 263, 8 3 66-8370. S e y e d i n , S . M., T h o m a s , T . C., T h o m p s o n , A. Y ., R o s e n , D. M. a n d P i e z , K. A. (1985). Purification and characterization of two cartilage-inducing factors from bovine demineralized bone. Proc. natn. Acad. Sci. U.S.A. 82, 2267—2271. S p o r n , M. B., R o b e r t s , A. B., W a k e f i e l d , L. M. a n d d e C r o m b r u g g e , B. (1987). Some recent advances in the chemistry and biology of transforming growth factor-/}. J. Cell Biol. 105, 1039-1045.
131
Transforming growth factor-/} receptors and binding proteoglycans FREDERICK T. BOYD*, SELA CHEIFETZ, JANET ANDRES, MARIKKI LAIHO a n d JOAN MASSAGUE Cell Biology and Genetics Program, M em orial Sloan-Kettering Cancer Center, SloanKettering Division o f the Graduate School o f M edical Sciences, Cornell University, New York, New York 10021, USA
Summary Transforming growth factors-beta (TGFs-/?) are representative of a superfamily whose members were first identified as regulators of morphogenesis and differentiation, and subsequently found to be structurally related. Other members of the family include the activins and inhibins, BMPs, MIS, the DPP-C gene product and Vg-1. When assayed by affinity-labelling techniques, TGFs-/? bind to three distinct cell surface proteins which are present on most cells. These proteins are all of relatively low abundance but bind TGFs-/? with affinities consistent with the biological potency of the factors. The Type I and Type II binding proteins are glycoproteins with estimated molecular weights of 53 and 7 3 x l 0 3Mr, respectively. They both bind TGF-/31 significantly better than TGF-/32. The Type I receptor has been identified as the receptor which mediates many of the responses of TGFs-/?, based on somatic cell genetic studies of epithelial cell mutants unresponsive to TGFs-/J. Betaglycan is the third binding protein present on many, but not all, cell types and is a large proteoglycan (~ 2 8 0 x l 0 3Mr) with 1 0 0 -1 2 0 x l 0 3 Mr core proteins. A soluble form of this molecule is present in conditioned media of many cell lines and may be derived from the cell surface-associated molecule by cleavage of a small membrane anchor. Betaglycan binds TGF-/31 and TGF-/J2 with similar affinity and this binding is to the core proteins, not the glycosaminoglycan side chains. This molecule may have a function in the localization and delivery or the clearance of activated TGFs-/i The molecular basis of TGF-/3 signalling is still largely unknown, but it is possible th at one or more of these cell surface molecules signals via a novel mechanism, as the TGFs-/3 are biologically quite distinct from other factors th at act via well-characterized signalling systems.
Introduction Transforming growth factors-beta (TGFs-¡3) are a family of polypeptide hormones which probably act over relatively confined physiological spaces in vivo. They are concentrated in platelets and released in an inactive form upon platelet degranulation (Pircher et al. 1986). Presumably they are activated in a local area by proteases or other specific activating conditions (Lyons et al. 1988). In addition, in situ hybridization and immunohistochemical studies have demonstrated that the factors are synthesized in many in vivo sites (Heine et al. 1987; Lehnert and Adhurst, 1988; Pelton et al. 1989). Our laboratory has attempted to identify and * Present address: Laboratory of Medicine and Pathology, University of Minnesota, Minneapolis, Minnesota 55455, USA. Key words: Transforming growth factor-beta (TGF-/3), receptors, binding proteoglycans.
132
F. T. Boyd et al.
characterize those cell surface proteins with which TGFs-/3 can interact and which presumably mediate the biological signalling of the factor into the cell. This is a somewhat complicated task for a number of reasons. As indicated above, TGF-/2 is a family of five closely related factors (TGF-/S1 to TGF-/35) within a superfamily of several other factors implicated in essential elements of development. These other factors include the decapentaplegic gene product of Drosophila, the Vg-1 gene product and mesoderm inducing factor in Xenopus, Mullerian Inhibiting Substance, the Bone Morphogenic Proteins, and the activins and inhibins (Massague, 1990). This suggests that there may be a family of related receptors with differential affinities for different members of the superfamily. Another complicating element is that the known effects of TGF-/? are varied and in some instances diametrically opposed in different cell systems. It is intriguing that TGF-/3 has a potent growth inhibitory effect on some epithelial cell lines, but it is also mitogenic in some culture systems, such as osteoblasts (Centrella et al. 1987) and some fibroblast lines (Leof et al. 1986). It is an inhibitor of differentiation in myogenic and adipogenic model systems in vitro (Massague et al. 1986; Ignotz and Massague, 1985), but stimulates differentiation of chondroblasts (Seyedin et al. 1985). And finally, experimentally, we have found that there is a relatively low level of specific cell-surface TGF-/3 binding and that this activity is distributed among three distinct cell surface binding proteins. Distribution of specific TGF-^-binding proteins In the course of investigating the activities of TGF-/3 over the last six years, our laboratory has characterized the TGF-/5 binding patterns of over one hundred different cell lines, primary cells and tissues (Fig. 1A) (Massague, 1990). It is striking that the general pattern is so similar in most cells analyzed. Briefly, the experimental approach is to bind iodinated TGF-/3 to cells, crosslink the ligand to cell surface proteins to which it is associated with a bifunctional crosslinking reagent such as disuccinimidyl suberate, solubilize the cell membranes with detergent and separate the labeled proteins on SDS-polyacrylamide gels with subsequent autoradiography. This method commonly identifies three proteins Fig. 1. Distribution of TGF-/3 receptors and binding proteoglycans in various cell types. A. Summary of TGF-/3 receptors in various cell types. All cell lines were screened for the presence of TGF-/3 receptors (I, II, III) using the affinity-labelling protocol outlined in the text. The presence of a receptor type in any cell line is signified with an open circle (O). The majority of cell lines screened express all three protein species, some lines express only the Type II and Type I receptors, a few lines express only the Type I receptor. No cell line which responds to TGF-/3 with established assays lacks the Type I receptor. Cell lines which lack any receptor type are signified with a closed circle (• ) under the column relating to th at receptor type. B. Receptor profiles from representative cell lines. Mouse BALB/c-3T3 cells, 3T3-L1 cells, ra t NRK cells, chick embryo fibroblasts (CEF) and mink lung epithelial cells (MvlLu) were screened by affinity labelling with 50 pM 125I-TGF-/3 in the presence of no (a), 50 (b), 100 (c), 200 (d), 700 (e), or 3000 (f) p M native TGF-/3. These experiments show the affinity of these proteins for TGF-/3 and the general nature of the affinity labeled species (Cheifetz et al. 1986).
133
TGF-ß receptors and binding proteins Huma n N o r ma l G M 3 1 6 s k i n f ib r o b l a s t s G M 3 7 0 s k i n f ib r o b l a s t s G M 5 8 7 7 s k l n f ib r o b l a s t s G M 1 2 6 2 c o n j u n c t i v e c e lls P r i m a r y lu n g f ib r o b la s t P r i m a r y s k i n f ib r o b la s t s U i l - 3 8 E m b r y o lu n g f ib r o b la s t s I M R - 9 0 E m b r y o lu n g f ib r o b la s t s R e t in a , p r im a r y E1M C3T 3 O ste o b la sts M C 3T 3 O st e o b la st s H S 1 4 M u s c le
1 o o o o o o o o o o o o
Huma n Tr a ns f o i
ea
79 2 2 O ste o sa rc o m a R 1684 C h o n d ro sa rc o m a H T -1 0 8 0 F i b r o s a r c o m a M N N G -H O S O s t e o g e n ic s a r c o m a U R - 1 3 f ib r o b l a s t S U - 4 0 U le ri R e t i n o b l a s t o m a R B 7 R e t in o b la s t o m a V 7 9 R e t in o b la st o m a R B 2 7 R e t in o b la st o m a R B I 3 R e t in o b la st o m a R B 2 0 R e t in o b la st o m a R B 2 2 R e t in o b la st o m a R B 2 4 R e t in o b la st o m a
o o o • • • • • o o o o o o o o o o o o
FR S k in f ib r o b la s t s N R K K id n e y f i b r o b l a s t s H 9 H e a rt m y o b l a s t L 8 S k e le ta l m u s c le m y o b la s t s R 7 r S R o r ta s m o o t h m u s c l e m y o b l a s t B R L - 3 A N e w b o r n r a t l iu e r P r i m a r y c a lu a r ia o s t e o b l a s t s P r im a r y h e p a t o c y te s P r im a r y c a r d io c y t e s P r i m a r y p it u i t a r y F e ta l t issu e P la c e n t a
o o o o o o • o o o o o o o o o o o o o o • • • • • •
280^
b
i
c
d
u
O O O O O O o o o o o o
o o o o
• o o o o o o o o o
o o o • o o o o o o o o
p
\o
o o o o o o
Rat T r a n s f o r m e d o o o o • o o o
F e S U - F i s h e r r a t e m b r y o f ib r o b la s t t s K i M S U - N R K K i d n e y e p it h e lia l C 6 G liom a H 3 5 H e p a to m a PC12 P h e o ch ro m o cy to m a L 6 S k e le t a l m u s c l e m y o b l a s t L 6 E 9 S k e le t a l M u s c l e M y o b l o s t L6E9 m y o tu b e s
Mouse
N o r ma l
e
• • • •
o o o o o
o o o o o
S w i s s 3T3 e m b r y o f i b r o b l a s t s S w i s s 3 T 3 / N R -6 e m b r y o f ib r o b la s t s B a lb / c 3T3 e m b r y o f i b r o b l a s t s 3 T 3 - L I p re a d ip o c y t e 3 T 3 - L I a d ip o c y t e
o o o •
o o o o • o o o
o o o o o
a
i
b
c
m
d
f
116 S" 94-
i
3 -.140
"*130
-
85 * -
"
68-
CEF d
M I S
MvlLu e
f
a
- 280 ^
I
200-
11694_ 68 -
85 *65
c
Ruian C h ic k e m b r y o f ib r o b l a s t s
Monkey B S C - 1 K i d n e y e p it h e lia l C O S K i d n e y f i b r o b la s t - li k e
110 85»- * -6 5 ►
b
§
c
d
e
f
O O O O • • 9 9 * 0
o o o
o • o o o
o o o o o
o
o
o
o o
o o
o o
o
o
o
o
o
•
o
o
•
O •
Mink Bouine FBH E F e t a l h e a r t e n d o t h e l ia l
Dog M D C K k i d n e y e p it h e lia l
Hamster DOT! M y o s a r c o m a C C L - 3 9 L u n g F ib r o b l a s t BH K k id n e y C H 0 - K 1 O u a ry CHO p g s f l- 7 4 5 CHO Id ID
a
I » I I
o
b
3T3 M S U - t r e n s f o r m e d f ib r o b la s t s N 2R N e u ro b la st o m a E L S t r a i n E, R s c it e s c a r c i n o m a C 1 2 7 I M a m m a ry tum or M B 6 6 - M C R 1 0 T I/ 2 F i b r o b l a s t
NRK e
O O O 0 0
M o u s e Transformed
3 T 3 -LI f
200-
a
M K K e r a t in o c y t e R K R - 2 B F ib r o b la st B 6 S U t R H e m a t o p o ie t ic p r o g e n i t o r s B 6 S U t- C L 2 7 h e m a t o p o ie t ic p r o g e n it o r s 3 2 D - C 1 3 H e m a t o p o ie t ic p r o g e n i t o r s
M u l L u (C C L 6 4 ) L u n g e p it h e lia l
BALB/C-3T3 a
• • • o o o o o o o
Rat N o r ma l
•
B
O O O O O O O O O O
000000
o o o o o o o o o o o o o o o o o o o o
R 4 3 1 E p it h e lio id c a r c in o m a T 2 4 B l a d d e r c a r c in o m a M D R - M B 2 3 1 B r e a s t c a r c in o m a M C F - 7 / 2 B r e a s t c a r c in o m a M C F - 7 / 4 B r e a s t c a r c in o m a M C F - 7 / 5 B r e a s t c a r c in o m a M C F - 7 / 1 B r e a s t c a r c in o m a S C C 2 S S q u a m o u s c e ll c a r c in o m a R 875 M e la n o m a R 5 4 9 L u n g a d e n o c a r c in o m a B 1 6 F I M e la n o m a B 6F10 M e la n o m a H5 2 2 4 R h a b d o m y o s a rc o m a M G -6 3 O ste o sa rc o m a O S - 1P O s t e o s a r c o m a
H L -6 0 M o n o m y e lo c y t ic le u k . (p h a g e ) U 9 3 7 M o n o c y t ic le u k e m ia IM R 3 2 N e u r o b l a s t o m a S K -N -S H N e u ro b la st o m a H S4 Sch w a n o m a H S 0 9 2 9 LU ilm 's t u m o r R D - 1 2 R e t in a , a d e n o u ir u s G M 0 I 9 7 2 R P r o g e r ia T E 6 7 1 M e d u ll o b l a s t o m a R D L - D G lio b la s t o m a
III o o o • o o o o o o o o
11 o o o o o o o o o o o o
b
c
d
e
f
134
F. T. Boyd et al.
labeled specifically (Fig. IB). The Type I protein is an affinity-labeled species of approximately 65 x 103 Mr. The Type II species is approximately 85 x 103 Mz, and the Type III species, termed betaglycan, is a broad band typically centered around 2 8 0 x l0 3Mr. All of these apparent molecular weights include an associated monomer of TGF-/? of 12.5 X103 Mr, so the presumed size of the binding proteins is correspondingly 12.5 x 103 Mr smaller than the apparent molecular weight on SDS gels. Each of these proteins have high affinities for TGFs-/?, with Kd values in the range of 5-500 pM . They bind TGF-/31, TGF-/32 and TGF-/33, but not more distantly related members of the TGF-/? superfamily. The Type I binding protein is ubiquitous, with every cell type that responds to TGFs-/? having the Type I protein. There are several examples of hematopoietic progenitor cell lines which respond to TGF-/31, TGF-j31.2 and TGF-/32 differentially with an order of potencies that parallels the order of affinities of the factors for the Type I protein, the only TGF-/? binding protein detectable on these cell lines (Ohta et al. 1987; Cheifetz et al. 1988). Human and bovine vascular endothelial cells possess both the Type I and Type II proteins and respond differentially to TGF-/31 or TGF-/32. In addition, L6E9 myoblasts also have only the Type I and Type II proteins, yet respond equivalently to TGF-/?1 and TGF-/32. The most common pattern of TGF-/? cell surface binding proteins is the presence of the Type I protein, Type II protein and betaglycan together. These data are suggestive, though certainly not definitive, that the Type I protein is sufficient to confer TGF-ft responsiveness to cells. However, the other TGF-/? binding proteins may modulate the availability or activities of the TGFs-/3 or mediate particular responses in particular cell types.
Somatic cell mutants defective in TGF-/J responsiveness More direct evidence supporting specific roles of these proteins in TGF-/? signalling has come from studies of somatic cell mutants in the MvlLu epithelial cell line which are non-responsive to TGFs-/?. The parental line is exquisitely sensitive to TGF-/? and is virtually completely growth inhibited by 5 p M TGF-/J1; it also responds to TGF-/? with increased expression of extracellular matrix components such as fibronectin and plasminogen activator inhibitor (PAI-1). In addition, the parental line has all three types of putative TGF-/? receptor proteins. Ethyl methane sulfonate (EMS)-mutagenized MvlLu cells were selected which were able to grow in the presence of 100 pM TGF-/31. The mutant clones were completely resistant to the growth inhibitory effects of TGF-/?1 and TGF-/?2 and have also lost all other responses to TGFs-/? assayed for. In addition, several of the clones were defective in expression of the Type I binding protein (Fig. 2). Clones of MvlLu cells obtained from a non-mutagenized population and analyzed for TGF-/3 binding proteins always have all three binding proteins present. On this basis, we suggest that the Type I protein is the receptor which mediates epithelial cell responsiveness to TGF-/? (Boyd and Massague, 1989). In addition to these receptor defective mutants, termed R mutants, mutants were isolated which have normal binding protein patterns, yet are deficient in all TGF-/3 responses. These are
TGF-/3 receptors and binding proteins
135
TGF-/3 response mutants in M v lL u cells TGF-/3 receptor profiles in CO w c c c £ CO 03 "3 -5 c "5 "5 Ei C D i_ E E CC GC CL ¿ ch Q
Betaglycan-
Type I I - *
%
Type I -►
m
•
I
■
&
I
Fig. 2. TGF-/Î receptor profiles of TGF-/3 resistant mutants of mink lung epithelial cells. Mutants of M vlLu cells which are not responsive to TGF-/Î with respect to growth inhibition were generated as described in the text. Representative clones of several mutant phenotypes were affinity labeled with 100 p M 125I-TGF-/3. Three distinct types of mutants have been isolated as characterized by TGF-/J binding profiles. When compared with the parental M vlLu cells, it is apparent that two of the mutant types have distinct deficits in TGF-/3 binding. The R-mutants are lacking the Type I receptor. In addition, the DR-mutants are lacking both the Type I and the Type II receptors. In contrast, the S-mutants have an apparently normal receptor profile.
called signalling, or S, mutants. When complementation analysis was performed, none of the mutant hybrids were complementary, suggesting that all the R and S mutants isolated are mutants in the same gene, presumably the TGF-/? receptor gene. It is also apparent from this analysis that all the mutants isolated were recessive mutations. All mutant-parental fusions were fully responsive to TGF-/? with normal receptor profiles. These studies have been extended and selection of M vlLu mutants has been performed with lower doses (25 p M) of TGF-/J2. In addition to the R and S mutant classes, a third class of mutants defective in the expression of both the Type I and Type II proteins has been isolated (DR mutants) (M. Laiho and others,
136
F. T. Boyd et al.
unpublished observations). Complementation analysis of these mutants is not complete, so the genetic basis of these mutations is still unknown. Several explanations for this phenotype are possible. DR mutants may be the result of mutations of two distinct loci encoding for the Type I and Type II proteins, although the frequency of isolation of these mutants argues strongly against this. Another possibility is that the two genes are linked and a single large deletion can knock out expression of both genes. A third possibility is that the two proteins are associated intracellularly and some mutations in one or the other of the proteins prevent the expression of both. These possibilities are the subject of active investigation in the laboratory. We are also attempting to complement these mutants by cDNA and genomic transfection to clone the genes responsible for these mutations. Betaglycan While mutant analysis has revealed the functional significance of the Type I and Type II proteins, the fact remains that in the majority of cells we have screened, the major component of cell surface TGF-/3 binding activity is associated with a large proteoglycan species with apparent molecular weight of 2 00-400x l 0 3Mr. This molecule is a complex mixed chondroitin/heparan sulfate proteoglycan (Segarini and Seyedin, 1988; Cheifetz et al. 1988) with multiple deglycosylated core proteins of 100-120x103Mr. Unlike other growth factors which associate with proteoglycans via relatively non-specific binding to the glycosaminoglycan chains, TGFs-/3 bind to betaglycan via the core proteins. The core proteins are expressed and bind TGFs-/3 in metabolic mutants which do not synthesize glycosaminoglycan side chains (Cheifetz and Massague, 1989). There is also a soluble form of betaglycan, found in the media of tissue culture cells which express betaglycan, which is capable of binding TGFs-/? (Andres et al. 1989). This form is slightly smaller than the membrane-associated form and is incapable of being incorporated into phospholipid vesicles. The soluble form of betaglycan also associates with the extracellular matrix. It is intriguing that TGF-/3, which is a well-characterized modulator of the extracellular matrix, binds specifically to a TGF-/3 binding proteoglycan which associates with the extracellular matrix and is the major species of TGF-/3 binding activity associated with cells. This may be a mechanism which can be modulated by TGFs-/3, by which TGFs-/? are sequestered in the intercellular space. A model of TGF-/3 binding protein function The definitive identification of the TGF-/3 receptor awaits cloning of the gene and reconstitution of a TGF-/3-responsive phenotype to TGF-/3-receptor mutants. However, the features of TGF-/3-binding outlined above suggests the following model. The identification of cell lines which possess only the Type I protein and the common defect in Type I binding activity in TGF-/3-response mutants suggests that the Type I protein is the signalling receptor for TGF-/3. The identification of
TGF-/3 receptors and binding proteins
137
TGF-/3-response mutants defective in both the Type I and Type II proteins suggests that the Type II protein may be part of a higher order complex with the Type I protein that associates and is a functional entity in some cell types. The dimeric structures of TGFs-¡5 are reminiscent of the structure of other dimeric growth factors which bind to dimeric receptors. Finally, betaglycan may be an extracellular storage site or mechanism of inactivation of the ligand. The physiology of the TGFs-/J suggests that specificity of action may come about as a result of acute regulation of availability of the active ligand. The fact that betaglycan is a proteoglycan and is present in membrane associated form as well as soluble and matrix associated forms suggests the possibility of acute regulation of the molecule, which might make it well suited to a role in clearance or storage of the ligands. References J . L., S t a n l e y , K., C h e i f e t z , S . a n d M a s s a g u e , J . (1989). Membrane-anchored and soluble forms of betaglycan, a polymorphic proteoglycan that binds transforming growth factor. J. Cell Biol. 109, 3137-3145. B o y d , F. T. a n d M a s s a g u jS , J . (1989). Transforming growth factor-beta-inhibition of epithelial cell proliferation linked to the expression of a 53 kD membrane receptor. J. biol. Chem. 264, 2272-78. C e n t r e l l a , M ., M c C a r t h y , T. L. a n d C a n a l i s , E. (1987). Transforming growth factor-beta is a bifunctional regulator of replication and collagen synthesis in osteoblast-enriched cell cultures from fetal ra t bone. J. biol. Chem. 262, 2 8 69-2874. C h e i f e t z , S. a n d M a s s a g u £ , J . (1989). The TGF-/3-receptor proteoglycan. Cell surface expression and ligand binding in the absence of glycosaminoglycan chains. J. biol. Chem. 264, 12 0 2 5 -1 2 0 2 8 . C h e i f e t z , S., A n d r e s , J . L. a n d M a s s a g u £ , J . (1988). The transforming growth factor-/j-receptor type III is a membrane proteoglycan. Domain structure of the receptor. J. biol. Chem. 263, 1 6 9 8 4 -1 6 9 9 1 . H e i n e , U. I., M u n o z , E . F . , F l a n d e r s , K. C., E l l i n g s w o r t h , L . R ., L a m , H. Y. P., T h o m p s o n , N. L ., R o b e r t s , A. B. a n d S p o r n , M . B. (1987). Role of transforming growth factor-/? in the development of the mouse embryo. J. Cell Biol. 105, 2 8 6 1-2867. I g n o t z , R. A . a n d M a s s a g u i s , J . (1985). Type-/? transforming growth factor controls the adipogenic differentiation of 3T3 fibroblasts. Proc. natn. Acad. Sci. U.S.A. 82, 8 5 30-8534. L e h n e r t , S . A . a n d A d h u r s t , R. J . (1988). Embryonic expression pattern of TGF-/?-type 1 RNA suggests both paracrine and autocrine mechanisms of action. Development 104, 2 6 3 -2 7 3 . L e o f , E. B., P r o p e r , J . A ., G o u s t i n , A . S . , S h i p l e y , G . D ., D i C o r l e t o , P . E. a n d M o s e s , H. L . (1986). Induction of c-sis mRNA and activity similar to platelet-derived growth factor-/} by transforming growth factor-/?: a proposed model for indirect mitogenesis involving autocrine activity. Proc. natn. Acad. Sci. U.S.A. 83, 2 4 53-2457. L y o n s , R. M ., K e s k i - O j a , J . a n d M o s e s , H. L . (1988). Proteolytic activation of latent transforming growth factor-/? from fibroblast-conditioned medium. J. Cell. Biol. 106, 1659-1665. M A S S A G u i, J . (1990). The transforming growth factor-/?-family. A. Rev. Cell Biol. 6, In press. M a s s a g u £ , J ., C h e i f e t z , S., E n d o , T. a n d N a d a l - G i n a r d , B. (1986). Type-/? transforming growth factor is an inhibitor of myogenic differentiation. Proc. natn. Acad. Sci. U.S.A. 83, 8206-8210. M i l l e r , D . A., L e e , A., M a t s u i , Y ., C h e n , E. Y ., M o s e s , H. L . a n d D e r y n c k , R. (1989). Complementary DNA cloning of the murine transforming growth factor-/?3 (TGF-/33) precursor and the comparative expression of the TGF-/?3 and TGF-/31 messenger RNA in murine embryos and adult tissues. Molec. Endocrinol. 3, 1926-1934. O h t a , M., G r e e n b e r g e r , J . S., A n k l e s a r i a , P., B a s s o l s , A . a n d M a s s a g u ^ , J . (1987). Two forms
A n d res,
138
F. T. Boyd et al.
of transforming growth factor-/? distinguished by multipotential haematopoietic progenitor cells. Nature 329, 5 3 9 -5 4 1 . P e l t o n , R. W., N o m u r a , S., M o s e s , H . L. a n d H o g a n , B. L. M . (1989). Expression of transforming growth factor-/82 R N A during murine embryogenesis. Development 106, 7 5 9 -768. P i r c h e r , R ., J u l i e n , P . a n d L a w r e n c e , D. A. (1986). Transforming growth factor-/? is stored in human blood platelets as a latent high molecular weight complex. Biochem. biophys. Res. Comm. 136, 3 0 -3 7 . S e g a r i n i , P. R. a n d S e y e d i n , S . M. (1988). The high molecular weight receptor to transforming growth factor-/? contains glycosaminoglycan chains. J. biol. Chem. 263, 8 3 66-8370. S e y e d i n , S . M., T h o m a s , T . C., T h o m p s o n , A. Y ., R o s e n , D. M. a n d P i e z , K. A. (1985). Purification and characterization of two cartilage-inducing factors from bovine demineralized bone. Proc. natn. Acad. Sci. U.S.A. 82, 2267—2271. S p o r n , M. B., R o b e r t s , A. B., W a k e f i e l d , L. M. a n d d e C r o m b r u g g e , B. (1987). Some recent advances in the chemistry and biology of transforming growth factor-/}. J. Cell Biol. 105, 1039-1045.
