Molecular and Cellular Endocrinology, 82 (1991) 191-198 0 1991 Elsevier Scientific Publishers Ireland, Ltd. 0303-7207/91/$03.50

191

MOLCEL 02646

~ca~izati~~

of transforming growth factorj31, $32 and $3 gene expression in bovine mammary gland

Rainer Maier, Peter Schmid, David Cox, Graeme Bilbe and Gary IS. McMaster Ciba-Geigy Ltd., Pharma Diuision, Biotechnology Department, CH-3002 Basel, Switzerland

(Received 4 June 1991: accepted 10 August 1991)

Key words: Transforming growth factor-pl,

$2 and $3; Hybridization, in situ; Mammary gland; Lactation

Summary We have studied the expression of transforming growth factor (TGF)$l, $32, and $3 in the non-lactating and lactating bovine mammary gland by in situ hybridization. All three isoforms were expressed in the lobuloalveolar framework of the non-lactating and lactating gland although marked differences were apparent in their spatial distribution. TGF-j31 was expressed predominantly by the epithelial ceils of the lobules although expression was also observed in the intralobular stroma cells lining the epithelium. In contrast, TGF+2 expression was only observed in the epithelial cells. TGF-@3 transcripts were expressed at the highest levels and were observed in almost all cells of the lobule. No TGF-/3 signals were found in the interlobular regions of the mammary gland. The possible regulatory functions of these molecules in development of the mammary gland and on differentiation processes in the neonate are discussed.

Introduction Transforming growth factor beta (TGF-P) is the collective name for a family of multifunctional growth regulatory molecules for which three mammalian isoforms have currently been identified. namely TGF+l, 42 and $33. Biologically active TGF-Ps are 25 kDa polypeptide dimers, which are derived from the carboxy-terminal end of a latent precursor molecule. These mature peptide forms share high degrees of sequence identity (approximately 70%; Derynck et al., 1988) and elicit a broad variety of cellular responses

Address for correspondence: Rainer Maier or Gary K. McMaster, Ciba-Geigy Ltd., Pharma Division, Biotechnology Department, CH-4002 Base& Switzerland.

(see reviews by Lyons and Moses, 1990; Roberts and Spot-n, 1990). In vitro, TGF-Ps are known to induce the synthesis of extracellular matrix proteins, the inhibition of epithelial cell proliferation, chemota~tic effects on macrophages, as well as mitogenic effects on osteoblasts and connective tissue fibroblasts. In vivo, local application of TGF-Pl has been shown to accelerate wound healing and induce bone formation. However, relatively little is known about the in vivo bioactivities of TGF-/32 or TGF-/33. Recent evidence now suggests that all three moiecules have similar effects although quantitative differences have been reported in some systems (Bock and Marsh, 1991). Earlier studies in our laboratory indicated that a TGF-P2-related molecule is present in bovine milk (Cox and Biirk, 1991); recent work has now

192

revealed the presence of both TGF-P1 and TGF/32, the latter being the predominant isoform (Jin et al., 1991). Although certain mammary epithelial cell lines are known to secrete a TGF-P-like activity (Knabbe et al., 1987; Ethier et al., 1990), to date there have been no reports on TGF-/3 expression in the mammary gland. We have performed in situ hybridization studies to determine the patterns of TGF-P expression in non-lactating and lactating bovine mammary gland tissues. Materials

and methods

Tissue preparations Mammary gland tissue was removed from calves and adult lactating cows (Bos taurus) immediately after slaughter at a local abattoir. Samples were fixed overnight at 4” C in a freshly prepared solution of 4% paraformaldehyde (in phosphate-buffered saline (PBS)) followed by another overnight incubation in 0.5 M sucrose in PBS. Tissue samples were stored in liquid nitrogen. Prior to sectioning, the tissue blocks were embedded in OCT compound (Miles). 10 pm cryostat sections were placed on 3-aminopropyltriethoxy-silane treated slides (Rentrop et al., 1986) and stored at -70” C. Prior to hybridization with RNA probes, the sections were dried for 5 min at 50 “C, postfixed with 4% paraformaldehyde in PBS for 5 min and finally washed in PBS and H,O. The sections were depurinated for 20 min with 0.2 N HCl at room temperature, incubated for 30 min at 70 o C in 2 X SSC (1 X SSC = 0.015 M trisodium citrate and 0.15 M NaCl, pH 7.0), dehydrated with increasing ethanol solutions and air dried. Preparation of RNA probes Sense and antisense RNA probes with [a-“S]UTP (1200 Ci mmol-‘,

were labeled NEN) to a

specific activity of > 10” dpm pg-’ using SP6, T3 or T7 RNA polymerase according to Schmid et al. (1991). The TGF-P riboprobe templates were 339 bp long fragments, corresponding to the cDNA sequences of the mature forms (including stop codon) of human TGF-/31, TGF-P2 and TGF-P3, subcloned into pGEM5 (Promega Biotec). The bovine TGF-Pl cDNA encoding the mature peptide has 92% homology to the human gene (Van Obberghen-Schilling et al., 1987). We have also cloned and sequenced the bovine TGFp2 and $33 cDNAs from mammary gland tissue (unpublished results) which have correspondingly high homologies (95%) to their human counterparts. However, comparison of the cDNA sequences encoding the mature peptide forms of human TGF-j31, $32 and -p3 reveals homologies of less than 75% between the three isoforms where the longest contiguous stretch of nucleotides is 14 bases long. The P-lactoglobulin riboprobe template was a 132 bp long fragment representing exon II of ovine P-lactoglobulin (which is 100% homologous to the bovine sequence) and was subcloned into Bluescript KS (Stratagene). In situ hybridization Prehybridization was carried out for a minimum of 3 h at 54 o C in hybridization buffer (50% formamide, 10% dextran sulfate, 0.3 M NaCl, 10 mM Tris, 10 mM sodium phosphate pH 6.8, 20 mM dithiothreitol (DTT), 0.2 X Denhardt’s, 0.1 mg ml-’ Escherichia cofi RNA, 0.2 PM [a-S] UTP). Hybridization was performed overnight at 54 o C in hybridization buffer containing the corresponding [a-“5S]UTP-labeled RNA probe at a concentration of 2 x lo5 dpm PI-’ in a humidified chamber. Coverslips were sealed with rubber cement to prevent changes in hybridization volume. Slides were washed 2 times for 1 h in

Fig. 1. Bright-field and corresponding dark-field illumination of TGF-Pl, $2 and $33 expression in cross-sections of the lobuloalveolar region of the lactating bovine mammary gland. (A, B) TGF-Pl; (C, D) TGF-P2; (E, F) TGF-/33. Expression of the TGF-P genes is restricted to the lobules (lo). No staining above background is observed in the surrounding interlobular stroma (is: A-F). TGF-61 mRNA signals are predominantly found in the alveolar epithelial cells (arrow) and in some subepithelial stromal cells bordering the basolateral side of the epithelium (A, B). TGF-62 mRNA is only observed in epithelial and myoepithelial cells of the lobule (C, D). TGF-P3 shows the highest level of RNA expression with signals detected in nearly all cells within the lobule (E, F).

193

Fig. 2. Bright-field C.4) Control

illuminations

(sense probe);

(B)

of P-lactoglohulin antisense

expression in the lobuloalveolar

region of the lactating bovine mammary

probe showing high levels of /3-lactoglohulin

expression

in cells (arrow)

gland.

bordering

the

alveolar lumen (1~).

hybridization buffer without dextran sulfate, E. coli RNA and UTP. Sections were then equilibrated at 37 ‘C for 15 min in equilibration buffer (0.5 M NaCl, 10 mM Tris-HCl pH 7.6, 5 mM EDTA) followed by RNase treatment (50 pug/ml RNase A) in the same buffer and then an additionai incubation without RNase A (37 “C. 15 min each). Slides were washed in 2 x SSC for 1 h and then in 0.1 x SSC for 1 h at room temperature. The sections were sequentially dehydrated in 65%, 85% and 95% (v/v> ethanol containing 300 mM ammonium acetate and then absolute ethano1 before being air dried. The sections were coated with a I:2 dilution of Ilford K5 photoemulsion, air dried and exposed for 2 weeks at 4” C in a light-safe box containing some silica gel. Development of the slides was performed using D19 developer (Kodak) and Agefix fixer (Agfa). Sec-

Fig. 3. Bright-field

and

c~)rrespondin~

dark-field

l(~bul(~alve(~lar region of the non-ia~tating the TGF-0

genes is restricted

(is. 1, A-F). bordering

TGF-PI

developing

mRNA

the basolateral

Results Differential counts of cell populations present in milk have revealed a predominance of macrophages and the absence of ductal secretory epithelial cells (Lee et al., 1980). We therefore performed our initial in situ hybridization experiments on cytospin preparations from freshly obtained bovine milk. Since no positive hybridization signals were obtained with the TGF-j3 probes on any of the cell types present in milk (Maier and Schmid, unpublished results) we focused our attention on the tissues of the mammary gland. In situ hybridization using antisense ribonucleotide probes corresponding to the human cDNA se-

illuminati~~n

of TGF-fit,

bovine rn~rnrna~f gland. (A,

$32

S) TGF-@I;

to the lobules (lo). No staining above background

signals are predominantly

side of the epithelium

alveoli (arrowed

tions were stained either with Giemsa or hematoxylin-eosin Y.

structures;

(A,

and

B). TGF-P2

mRNA

expression (e,

in cross-sections F) TGF-63.

is observed in the surrounding

found in the alveolar epithelial

C, D). Again, TGF-P3

$3

(C, Df TGF-/?Z:

is only observed in the multi-layered

nearly all cells within the lobule (E.

F).

interlobular

cells and in some subepithelial

shows the highest level of RNA

of the

Expression of

epithelial

stroma

stromal cells cells of the

expression with signals detected

in

19s

196

quences of the mature forms of TGF-Pl, $2 and $3 showed that all three isoforms were expressed in calf and adult bovine mammary gland. We have previously shown that by using these probes, no cross-hybridization was detected between the three TGF-P isoforms, either by in situ hybridization or Northern blot analyses (Schmid et al., 1991). This is easily explained since the three isoforms are less than 75% homologous to one another, where the longest contiguous stretch is 14 bases long. In the adult animal TGF-Pl RNA signals were predominantly detected in the alveolar epithelial cells (Fig. 1A and B) with some signafs also found in the subepithelial stromal cells bordering the basolateral epithelium. Stromal cells distant from the epithelia showed no hybridization signals with the TGF-/31 probe. Conversely TGF-/32 was only expressed by epithelial cells lining the lumen of the alveoli and the surrounding myoepithelial cells (Fig. 1C and D). TGF-/32 transcripts were not detected in any intralobular stromal cells. TGF-/33 expression showed the broadest and highest levels of distribution; strong signals were observed in nearly all cells within the lobules. Stromal cells distant from the epithelia also showed positive, but weaker, hybridization signals (Fig. 1E and F). In situ hybridizations using homologous ~-lactoglobulin sense and antisense RNA probes were performed on serial tissue sections in order to confirm that TGF-Pl and $2 transcripts were produced by milk protein secreting cells. Strong hybridization signals were seen in the cells bordering the alveolar lumen (Fig. 2). Comparison of the transcript distributions of the three TGFj3 isoforms revealed a clear pattern of TGF-/3 RNA expression in the mammal gland. Whereas TGFj32 was exclusively expressed in the epithelial cells of the gland, TGFpl and -/33 showed a gradient-like expression pattern; highest levels occurring in the alveolar epithelium and declining towards the subepithelial stromal areas. The only apparent difference between the expression of TGF-Pl and -j33 was the higher level of TGF-/33 transcripts in cells of the interalveolar stroma (Fig. 1 A/L3 and E/F). This pattern of expression was only discernible in lobules having a low percentage of luminal area and therefore a correspondingly high percentage

of interlobular stroma. In lobules with large numbers of alveoli these distinctions were not as clear since the close proximity of the stromal and luminal areas often resulted in a more uniform distribution of the TGF-P1 and $33 hybridization signals. In the premature, non-lactating mammal gland of the calf the signal distribution was almost identical (Fig. 3). The premature gland is characterized by only partially developed lobuloalveolar structures and a thick, multi-layered epithelium which is not fully differentiated. All three TGF-P genes were expressed in the epithelial cells (Fig. 3A-F). TGF-/32 transcripts were restricted to the epithelium (Fig. 3C and L)), whereas TGF-j31 and TGF-P3 mRNA were also detected in stromal cells (Fig. 3A/B and E/F) as found in the adult lactating mammary gland. However, due to the premature state of the calfderived tissue, the expression patterns were not as clear as the patterns found in the lactating gland. A gradient-like expression pattern of TGF~31 and TGF-P3 could not be seen but again TGF-P3 transcripts were found at a higher level and signals were more uniformly distributed in the intralobular stromal areas (Fig. 3A/B and E/F).

Discussion Considerable morphological changes take place during development of the mammary gland such as invasive branching of epithelial tissue, establishment of a continuous central lumen and formation of the secretory lobuloalveolar network (Dulbecco et al., 1982; Hogg et al., 1983). TGF-j31 has already been shown to inhibit ductal elongation in early mammary gland development but has little effect later on cellular proliferation of the lobuloalveolar structures (Daniel et al., 1989). The similar patterns of expression observed in the non-lactating and lactating mammary gland suggest that the key regulatory steps are posttranscriptional; the release of TGF-Ps’ precursors from their storage sites and their subsequent activation (Roberts and Sporn, 1990) being likely essential prerequisites for their autocrine/ paracrine effects in the mammary gland.

197

A number of recent reports show that the regulation of active TGF-/3 is a multifold event. It has been reported that at the transcriptional level TGF-/3 mRNAs are only transiently expressed, e.g. during embryogenesis (Wilcox and Derynck, 1988; Millan et al., 1991; Schmid et al., 1991) whereas constitutive expression of TGF-/? mRNA does occur, e.g. in monocytes although no protein is secreted. However, when the monocytes acquire the phenotype of macrophages these cells secrete latent TGF-/3 which still requires further activation in vitro, e.g. by acid or protease treatment, to produce the mature active peptide (Assoian et al., 1987; McCartney-Francis et al., 1990; Lazdins et al., 1991). In vivo activation of TGF-/3, as far as the autocrine/paracrine actions are concerned, may also be dependent upon cell-cell contact, as suggested by recently published experiments (Antonelli-Orlidge et al., 1989; Van Zoelen and Tertoolen, 1991). TGF-Ps have major influences on extracellular matrix (ECM) composition, notably by stimulating the expression of ECM components, such as fibronectin and collagen (Ignotz et al., 1986) and promoting the addition of chondroitin sulfate side chains to cell surface proteoglycans on mammary epithelium (Rapraeger, 1989). Since milk protein secretion by the mammary epithelium is controlled by the composition of the ECM (Li et al., 1987; Medina et al., 1987; review by Bissell and Hall, 1987) we propose that TGF-Ps also play important roles in regulating the functional changes which precede milk protein secretion. In the present study we observed specific TGF-Pl and TGF-P3 mRNA signals in the intralobular stromal cells of the mammary gland. Secretion and activation of the respective proteins by these cells may well influence the secretion of ECM components and/or induce final differentiation of epithelial cells in the calf as well as modulating milk expression in the adult. Since TGF-/32 mRNA signals were exclusively observed in epithelial cells and both TGF-/31 and TGF-P2 proteins have been isolated from milk (Cox and Biirk, 1991; Jin et al., 19911, we propose that the functions of this isoform may be dependent on its secretion into milk. The role of TGF-Ps in milk is also presently unclear. TGF-Pl has been shown to induce ter-

minal differentiation of intestinal epithelial cells in vitro (Kurokowa et al., 1987) and one possible function for milk TGF-fls may be to induce terminal differentiation of the gastrointestinal epithelia in the neonate. TGF-Pl is also known to induce IgA synthesis in gut associated lymph node cells (Chen and Li, 1990) and therefore another function for these polypeptides present in milk may be to induce IgA synthesis in the neonate thereby protecting against infection by microorganisms at the gut mucosal surface. The present study has shown for the first time TGF-P gene expression in non-lactating and lactating mammary gland, where almost identical patterns of expression were observed for the three genes. Previously we have observed the co-expression of all three TGF-/3s only in the root sheath epithelia of the hair follicle during mouse embryogenesis (Schmid et al., 1991). Hence the co-expression of all three genes in the same tissue and cell type, as occurs in the epithelium of mammary gland lobules, is a notably rare event.

References

Antonelli-Orlidge, A., .Saunders, K.B., Smith, S.R. and D’Amore, P.A. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 4544-4548. Assoian, R.K., Fleurdelys, B.E., Stevenson, H.C., Miller, P.J., Madtes, D.K., Raines, E.W., Ross, R. and Sporn, M.B. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 6020-6024. Bisell, M.J. and Hall, H.G. (1987) The Mammary Gland (Neville, M.C. and Daniel, C.W., eds.), pp. 97-146, Plenum Press, New York. Bock, G.R. and Marsh, J. teds.) (1991) Ciba Foundation Symposium, Vol. 157, John Wiley, New York Chen, S.-S. and Li, Q. (1990) Cell. Immunol. 128, 353-361. Cox, D. and Biirk, R. (1991) Eur. J. Biochem. 197, 353-358. Daniel, C.W., Silberstein, G.B., Van Horn, K., Strickland, P. and Robinson, S. (1989) Dev. Biol. 135, 20-30. Derynck, R., Lindquist, P.B., Lee, A., Wen, D., Tanim, J., Graycar, J.L., Rhee, L., Mason, A.J., Miller, D.A., Coffey, R.H., Moses, H.L. and Chen, E.Y. (1988) EMBO J. 7, 3737-3743. Dulbecco, R., Henahan, M. and Armstrong, B. (1982) Proc. Natl. Acad. Sci. U.S.A. 79, 7346-7350. Ethier, S.P. and Van de Velde, R.M. (1990) J. Cell. Physiol. 142, 15-20. Hogg, N.A.S., Harrison, C.J. and Tickle, C. (1983) J. Embryol. Exp. Morphol. 73, 39-57. Ignotz, R.A. and Massague, J. (1986) J. Biol. Chem. 261, 4337-4345.

198 Jin, Y., Cox, D. and Cerletti, N. (1991) J. Protein Chem. (in press). Knabbe, C., Lippman, M.E., Wakefield, L.M., Flanders, K.C., Kasid, A., Derynck, R. and Dickson, R.B. (1987) Cell 48, 417-428. Kurokowa, M., Lynch, K. and Podolsky, D.K. (1987) Biochem. Biophys. Res. Commun. 142, 775-782. Lazdins, J.K., Klimkait, T., Woods-Cook, K., Walker, M., Aheri, E., Cox, D., Cerletti, N., Shipman, R., Bilbe, G. and McMaster, G.K. (1991) J. Immunol. 147, 1201-1207. Lee, C.-S., Wooding, F.B.P. and Kemp, P. (1980) J. Dairy Res. 47, 39-50. Li, M.L., Aggeler, J., Farson, D.A., Hatier, C., Hassell, J. and Bissell, M.J. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 136-140. Lyons, R.M. and Moses, H.L. (1990) Eur. J. Biochem. 187, 467-473. McCartney-Frances, N., Mizel, D., Wong, H., Wahl, L. and Wahl, S. (1990) Growth Factors 4, 27-35.

Medina, D., Li, M.L., Oborn, C.J and Bissell, M.J. (1987) Exp. Cell Res. 172, 192-203. Millan, F.A., Denhez, F., Kondaiah, P. and Akhurst, R.J. (1991) Development 111, 131-144. Rapraeger, A. (1989) J. Cell Biol. 109, 2509-2518. Rentrop, M., Knapp, B., Winter, H. and Schweizer, J. (1986) Histochem. J. 18, 271-276. Roberts, A.B. and Sporn, M.B. (1990) Handbook of Experimental Pathology, Vol. 95 (Sporn, M.B. and Roberts, A.B., eds.), pp. 419-472, Springer-Verlag, Heidelberg. Schmid, P., Cox, D., Bilbe, G., Maier, R. and McMaster, G.K. (1991) Development 111, 117-130. Van Obberghen-Schilling, E., Kondaiah, P., Ludwig, R.L., Sporn, M.B. and Baker, CC. (1987) Mol. Endocrinol. 1, 693-698. Van Zoelen, E.J. and Tertoolen, L.G. (1991) J. Biol. Chem. 266, 12075-12081. Wilcox, J.N. and Derynck, R. (1988) Mol. Cell. Biol. 8, 34153422.