Maruo et al.

JR, ed. Endocrinology of pregnancy. Chicago: Year Book, 1981:1-34.

19. Maruo T, Matsuo H, Hayashi M, Nishino R, Mochizuki M. Induction of differentiated trophoblast function by epidermal growth factor: relation of immunohistochem­ically detected cellular epidermal growth factor receptor levels. J Clin Endocrinol Metab 1987;64:744-9.

20. Ladines-Llave CA, Maruo T, Manalo AS, Mochizuki M. Cytologic localization of epidermal growth factor and its

July 1992 Am J Obstet Gynecol

receptor in developing human placenta varies over the course of pregnancy. AM J OBSTET GVNECOL 1991; 165: 1377-82.

21. Kao LC, Caltabiano S, Wu S, Strauss JF III, Kliman Hl The human villous cytotrophoblast: interactions with ex­tracellular matrix proteins, endocrine function and cy­toplasmic differentiation in the absence of syncytium for­mation. Dev Bioi 1988; 130:693-702.

Transforming growth factor-J3 opposes the stimulatory effects of interleukin-l and tumor necrosis factor on amnion cell prostaglandin E2 production: Implication for pre term labor

Kristina Bry, MD, and Mikko Hallman, MD

Irvine, California, and Helsinki, Finland

OBJECTIVE: In preterm labor increased concentrations of interleukin-' and tumor necrosis factor are present in amniotic fluid. These cytokines may promote labor by stimulating the production of

prostaglandins by intrauterine tissues. In many biologic processes, transforming growth factor-13 modifies the actions of cytokines. We studied the effect of transforming growth factor-J3 on the cytokine-induced prostaglandin E2 production by amnion cells. STUDY DESIGN: Human amnion cells in monolayer culture were treated with interleukin-1, tumor necrosis factor, or vehicle in the presence or absence of transforming growth factor-I3. The prostaglandin E2 production was measured. RESULTS: Transforming growth factor-13 decreased the interleukin-1- or tumor necrosis factor-induced prostaglandin E2 production by 70% to 80% and the basal prostaglandin E2 synthesis by 27%. The synergistic stimulation of prostaglandin E2 production by the combination of interleukin-1 with tumor necrosis factor was inhibited by 80% in cells treated with transforming growth faclor-l3. Transforming growth factor-j31, -132, and -131,2 were equipotent. CONCLUSION: Transforming growth factor-j3 suppresses the cytokine-induced prostaglandin E2 production by amnion cells and may be an important factor in maintaining pregnancy in the face of labor-promoting cytokines. (AM J OBSTET GVNECOL 1992;167:222-6.)

Key words: Preterm labor, prostaglandins, interleukin-l, tumor necrosis factor, transforming growth factor-/3, fetal membranes

Preterm labor often occurs in the setting of intra­uterine infection. I The concentrations of cytokines such as interleukin-l (IL-l) and tumor necrosis factor (TNF) increase in amniotic fluid in pre term labor.2~4 Prosta-

From the Department of Pediatrics, University of California, [rome, and the Department of Pediatrics, University of Helsinki. Supported by the Foundation for Maternal and Infant Care and the Sigrid Juselius Foundation. Received for publication October 14, 1991; revised January 21, 1992; accepted January 23,1992. Reprint requests: Kristina Bry, MD, DepaTtment of Pediatrics, Uni­versity of California, Imine, Med. SUTge /.. Rm 109 F, [mine, CA 92717. 611/36682

222

gland ins, major inducers of uterine contractions, are produced by intrauterine tissues in response to these cytokines.3

5·7 The amnion, which produces almost ex­clusively prostaglandin £2 (PG£2),8 is an important in­trauterine source of prostaglandin. The combination of [L-l with TNF has a synergistic stimulator effect on the PGE2 production by human amnion cells in culture.~ IL-I and TNF have been implicated in the induction of preterm labor associated with infections."' 3, 10

Transforming growth fac10r-/3 (TGF-/3) represents a family of multifunctional factors that influence cell growth, differentiation, immunity, and extracellular matrix formation. II Several forms of TGF -/3 and of

Volume 167 Number 1

0.6 -c:( Z 0.5 C

Cl 0.4 c -... Cl Q.

0.3 -N W (!) 0.2 c..

0.1

0.0 0

--0-- no TGF·B1

~ TGF·B1 (10 ng/ml)

.05 .5 5 50

TNF- (X (ng/ml)

TGF·jj, PGE2 production, and labor 223

0.4

0.3

0.2

0.1

0.0 500 5000 0 .005 .05 0.5 5 50 500

IL-1B (ng/ml)

Fig. 1. Effect of TGF-jj on amnion cells stimulated with different concentrations of TNF-a or IL­Ijj. Human amnion cells were treated with indicated concentrations ofTNF-a or IL-Ijj in presence (e) or absence (0) of human TGF-jj\ (10 ng/ml). PGE2 production (picograms per nanogram DNA) was measured after 42-hour incubation. Results are means ± SE (n = 4) of one representative experiment.

their receptors have been identified. TGF-J31 and TGF-132 are homodimers, and TGF -131,2 is a heterodimer containing one polypeptide chain each from TGF-J31 and TGF-J32. Most tissues found at the maternal-fetal interface contain TGF-J3 and express TGF-J3 messenger ribonucleic acid.'2 In many biologic processes, TGF-J3 acts as a biologic switch, antagonizing or modifying the action of other growth factors or cytokines. 13 We there­fore studied the effect of TGF-(3 on the cytokine-in­duced PGE2 production by human amnion cells in cul­ture. In this report we show that TGF-J3 suppresses the IL-l- or TNF-stimulated PGE2 production by amnion cells.

Material and methods

Chemicals. Recombinant human TNF-a (specific ac­tivity 4.8 x 107 U I mg, endotoxin content <0.13 endo­toxin unit [EU]/mg) and TNF-J3 (specific activity 2.1 x 10" U I mg, endotoxin 3.1 E U I mg) were generously pro­vided by Genentech, South San Francisco. Recombi­nant human IL-IJ3 (specific activity 1.8 x 107 U I mg, endotoxin 2.5 EU Img) was a generous gift from the National Cancer Institute. Recombinant human IL-la (specific activity 3 x 108 U I mg, endotoxin 0.7 EU I mg) was a kind gift from Hoffmann-La Roche, Inc., Nutley, N.]. The PGE. kit was from New England Nuclear, Boston. Human TGF-J31 and porcine TGF-J31, TGF-132, and TGF-J31,2 were from R&D Systems, Minne­apolis.

Preparation and culture of human amnion cells. Human fetal membranes were obtained from normal pregnancies at cesarean sections conducted at term be-

Table I. Effect of TGF-J3 on PGE. production by amnion cells stimulated with IL-l or TNF

Control IL-Ia IL-I jj TNF-a TNF-jj

PGE2 production (fractions of control)

Without J With TGF-f31 TGF-f31

I 25.5 ± 9.6 27.3 ± 10.6 48.4 ± 7.9

9.6 ± 2.6

0.7 ± 0.1 4.1 ± 1.8 4.4 ± 2.0 9.7 ± 3.1 1.4 ± 0.3

Inhibition (%) in PGE2 production

caused by TGF-f31

27 ± 9 77 ± 4 75 ± 5 75 ± 8 71 ± 9

Amnion cells were treated with IL-Ia (6.8 ng/ml), IL-Ijj (5 ng/ml). TNF-a (50 ng/ml), TNF-jj (10 ng/ml), or vehicle (control) in the presence or absence of human TGF-jjI (10 ngl ml). After 42 hours of incubation, the media were removed and analyzed for PGE2• PGE2 production is shown as fractions of production in untreated (control) cells; in these cells PGE2

production was 0.031 ± 0.008 pg/ng DNA. The inhibition in percent caused by TGF-jj in each experiment was calculated. The means ± SE of six separate experiments are shown.

fore the onset of labor. Cells from the reflected amnion were isolated as described. 14 The cells were seeded in tissue culture plates at a density of 1.4 x 105 cells I cm2

and maintained at 37° C in an atmosphere of air and 5% carbon dioxide in Eagle's minimum essential me­dium with Earle's salts supplemented with glutamine, 10% fetal calf serum, antibiotics, and antimycotics. Af­ter 24 hours the wells were washed with minimum es­sential medium, and the medium was changed to the culture medium supplemented with cytokines, TGF-J3, or the appropriate vehicle. The cells were incubated for 42 hours before the culture media were removed

224 8ry and Hallman

0.4

0.3

0.2

0.1

0.0

0.20

Ci" Z C 0.15

CI I: C, 0.10 S: N W 0.05 CI 0.

0.00

0.03

0.02

0.01

0.000 .01

--0-- TNF (50 ng/ml) + TGF-B1 _ TNF (50 ng/ml) + TGF-B2

-tr- TNF (50 ng/ml) + TGF-B1,2

--0-- IL-1 (5 ng/ml) + TGF-B1 _ IL-1 (5 ng/ml) + TGF-B2

-tr- IL-1 (5 ng/ml) + TGF-B1,2

--0-- TGF-B1 _ TGF-B2

-tr- TGF-B1,2

0.1 10

TGF-B (ng/ml)

Fig. 2. Comparison of effects of TGF-131, TGF-132, and TGF-I3I,2 on PGE2 production by amnion cells. Human am­nion cells were treated with TNF-a (50 ng/ml) (upper panel), IL-ll3 (5 ng/ml) (middle panel), or vehicle (lower panel), in pres­ence of indicated concentrations of porcine TGF-131 (0), TGF-132 (e), or TGF-131,2 (~). The PGE2 production (pico­grams per nanogram DNA) was measured after 42-hour in­cubation. Results are means ± SE (n = 4) of one represen­tative experiment.

and then centrifuged at 600 g for 10 minutes to elim­inate cell debris. The supernatants were stored at - 70° C until assayed for PGE2 • The cells were recov­ered from the wells with trypsin. More than 90% of the

cells excluded trypan blue. Measurement of PGE2_ PGE. was measured by ra­

dioimmunoassay directly from the media. The concen­trations of PGE2 in the media incubated in the same conditions but without amnion cells were used as blanks. PGE2 formation was calculated per nanogram of amnion cell deoxyribonucleic acid (DNA); DNA anal­ysis was done as described. l5 The different treatments with cytokines or TGF-13 did not significantly affect the DNA content of the wells.

July 1992 Am J Obstet Gynecol

Statistics. The test used was the Wilcoxon signed­rank test.

Results

Effect of TGF-p on PGE. production by amnion cells stimulated with IL-l or TNF. Human amnion cells were treated with IL-Ia (6.8 ngl ml), IL-ll3 (5 ng/ml), TNF-a (50 ng/ml), TNF-13 (10 ng/ml), or vehicle in the presence or absence of human TGF-l3l (10 ng/ml). Because amnion cells do not metabolize PGE2 , the concentration of PGE. in the medium at the end of the 42-hour incubation represents the total PGE2

released during that time.s For each cytokine the PGE. production was lower in the presence ofTGF-13 in each of the six separate experiments performed (p = 0.028) (Table I). The decrease caused by TGF-j3 in the basal PGE2 production was 27% ± 9%. However, in cells stimulated by IL-l or TNF, the decrease in total PGE2

production occurring in the presence of TGF-l3l ranged from 71 % ± 9% to 77% + 4% (Table I).

Effect of TGF -p on stimulation of amnion cell PGE. production by combination ofIL-l and TNF. Amnion cells were incubated for 42 hours with IL-113 (5 ng/ml), IL-la (6.8 ng/ml), or TNF-a (5 J-lg/ml) in the presence or absence of human TGF-l3l (10 ng/ml). IL-l acts synergistically with TNF to stimulate amnion cell PGE2

production.9 Nevertheless, the degree of inhibition due to TGF-l3l was approximately the same as was obtained in cells treated with only one cytokine (Table II). Similar results were obtained with TNF-13 (10 ng/ml) instead of TNF-a (results not shown).

Effect of TGF-Il as a function of IL-l and TNF con­centration. Amnion cells were incubated with different concentrations of IL-113 (0 to 500 ng/ml) or TNF-a (0 to 5 J-lg/ml) with or without TGF-131 (10 ng/ml). The PGE. release was assayed after incubation for 42 hours. IL-113 and TNF-a stimulated the PGE2 produc­tion in a concentration-dependent manner; the satu­rating concentrations for IL-113 and TNF-a were 0.5 and 50 ng/ml, respectively_ At each concentration of IL-ll3 and ofTNF-a, TGF-131 inhibited the PGE2 pro­duction as seen in Fig. I.

Effect of different concentrations of TGF-pl, TGF­P2, or TGF-pl,2 on PGE. production by amnion cells. Cells were treated for 42 hours with IL-113 (5 ng/ml) or TNF (50 ng/ml), together with different concentra­tions (0 to 10 ng/ml) of porcine TGF-I3I, TGF-132, or TGF-I3I,2. The potencies of TGF-j3I, TGF-j32, and TGF-IH,2 in inhibiting the PGE2 production were sim­ilar at each concentration tested (Fig. 2). In each case maximal inhibition was obtained with TGF-13 concen­trations 2:0.1 ng/ml (Fig. 2). The effects of porcine TGF-131 were indistinguishable from those of human TGF-131 (data not shown).

In amnion cells not treated with cytokines, TGF-I3I,

Volume 167 Number I

TGF-I3, PGE2 production, and labor 225

Table II. Effect of TGF-fjl on amnion cell PGE2 production stimulated by combination of IL-l and TNF

PGE2 production

Without TGF-f31 I With TGF-f31 Inhibition caused by TGF-f31 (%)

Control IL-la (6.8 ng/ml) IL-ll3 (5 ng/ml) TNF-a (50 ng/ml)

0.046 ± 0.002 1.55 ± 0.035 1.83 ± 0.029 2.83 ± 0.034

0.029 ± 0.001 0.237 ± 0.024 0.195 ± 0.013 0.312 ± 0.025 1.678 ± 0.051

40.0 ± 1.7 84.7 ± 1.6 89.3 ± 0.8 89.0 ± 0.8 81.3 ± 0.8 80.2 ± 1.2

IL-la (6.8 ng/ml) + TNF-Ct (50 ng/ml) IL-ll3 (5 ng/ml) + TNF-Ct (50 ng/ml)

9.006 ± 0.323 8.671 ± 0.312 1.71 ± 0.077

Amnion cells were treated with indicated cytokines or cytokine combinations in presence or absence of human TGF-131 (l0 ng/ ml). Total release ofPGE2 (picograms per nanogram DNA) during 42-hour incubation was measured. Data are results (means ± SE, n = 4) of one representative experiment.

TGF-fj2, and TGF-fjl,2 also appeared equipotent, but the degree of inhibition was smaller than that obtained in cells stimulated with cytokines (Fig. 2).

Comment In this report we show for the first time that TGF-~

strikingly opposes the IL-l- or TNF-stimulated PGE2 production by human amnion cells. It also suppressed the synergistic PGE2-stimulatory effect of the combi­nation of IL-l with TNF. The basal PGE, production by amnion cells was also inhibited by TGF-[3 but to a smaller extent. In According to the present results, TGF­[3 may playa role in limiting or terminating the pros­taglandin production induced by proinflammatory cy­tokines in the setting of intrauterine infection.

At all concentrations tested the effects of TGF -fjl, TGF-[32, and TGF-[31,2 on amnion cell PGE, produc­tion were similar. TGF-fjl and TGF-[32 differ from each other by 32-amino-acid substitutions and have 71 % sequence identity. In many cell systems TGF-~1 and TGF-fj2 are biologically equipotent; however, they show marked differences in certain cells. 10

In contrast to its effects on amnion cells, TGF-[3 has a stimulatory effect on the production of prostaglan­dins in many cells. Thus it stimulates the PGE, pro­duction in neonatal mouse calvaria l7 and in lung fibro­blasts,'8 as well as prostacyclin production in human vascular endothelial cells. 19 In addition, TGF-13 potently synergizes with IL-113 and TNF-a in stimulating the PGE2 production in rat renal mesangial cells.20

TGF-fj-like activity is present in many cells at the fetomaternal interface.]2 TGF-fjl antibody reacts with cells in the chorion and with extracellular matrix com­ponents in the placenta and uterus.'1 TGF-131 messen­ger ribonucleic acid is present in the placenta, decidua, and placental membranes.]' A TGF-fj2-like factor has been described in murine amniotic fluid. 22 High levels of TGF-fj2 messenger ribonucleic acid are present in the uterus of the pregnant mouse; they fall rapidly as birth approaches. 22

Besides the observed effect of TGF-fj on PGE, syn-

thesis, there is additional evidence suggesting the im­portance of TGF-fj to the maintenance of pregnancy. TGF-fj is a potent immunosuppressant both in vivo and in vitro."' Murine decidual cells secrete an immuno­suppressive TGF-132-1ike molecule!4 TGF-13 may be essential in the maintenance of pregnancy by prevent­ing rejection of the fetus."

TGF-fj induces the accumulation of extracellular ma­trix by several mechanisms. It stimulates the synthesis of many extracellular matrix components.'5 It also en­hances the adhesion of extracellular matrix on cells!6 In addition, TGF-13 up-regulates inhibitors of extra­cellular matrix-degrading enzymes'"28 and inhibits en­zymes (such as collagenase) that degrade extracellular matrix components.'" The effect ofTGF-13 on cell adhe­sion receptors may be of importance in the attachment of the fetoplacental unit to the uterine wall. Because of its effects on extracellular matrix, TGF-fj has the po­tential of inhibiting processes required for both term and preterm labor. These include softening and dila­tation of the cervix, which is associated with a decrease in collagen content and an increase in collagenolytic activity!9 The biologic activity of TGF-fj and its regu­lation during pregnancy remain to be clarified.

From the current results we propose that TGF-fj plays an important role in the modulation of processes that regulate the initiation of preterm labor. In cases of intrauterine inflammation, increased production of cytokines threaten pregnancy. The continuation of pregnancy may depend on the balance of the activities of labor-promoting agents, such as IL-I and TNF, and of agents, such as TGF-I3, that favor the continuation of pregnancy. It remains to be seen whether TGF-13 has therapeutic potential in the prevention of preterm labor.

REFERENCES 1. Romero R, Avila C, Brekus CA, Morotti R. The role of

systemic and intrauterine infection in pre term parturi­tion. Ann N Y Acad Sci 1991;622:355-75.

2. Romero R, Brody DT, Oyarzun E, et al. Infection and

226 Bry and Hallman

labor. III. Interleukin-l: a signal for the onset of partu­rition. AM] OBSTET GYNECOL 1989;160:1117-23.

3. Casey ML, Cox SM, Beutler B, Milewich L, MacDonald PC. Cachectinltumor necrosis factor-a formation in hu­man decidua. J Clin Invest 1989;83:430-6.

4. Romero R, Manogue KR, Mitchell MD, et al. Infection and labor. IV. Cachectin-tumor necrosis factor in the am­niotic fluid of women with intraamniotic infection and preterm labor. AM] OBSTET GYNECOL 1989;161:336-41.

5. Romero R, Durum S, Dinarello CA, Oyarzun E, Hobbins ]C, Mitchell MD. InterIeukin-l stimulates prostaglandin biosynthesis by human amnion. Prostaglandins 1989;37: 13-22.

6, Romero R, Mazor M, Wu YK, Avila C, Oyarzun E, Mitchell MD. Bacterial endotoxin and tumor necrosis factor stim­ulate prostaglandin production by human decidua. Pros­taglandins Leukot Essent Fatty Acids 1989;37:183-6.

7. Mitchell MD, Edwin S, Romero RJ. Prostaglandin biosyn­thesis by human decidual cells: effects of inflammatory mediators, Prostaglandins Leukot Essential Fatty Acids 1990;41 :35-8.

8, Casey ML, MacDonald PC, Mitchell MD. Characterization of prostaglandin formation by human amnion cells in cul­ture. Prostaglandins 1984;27:421-7.

9, Bry K, Hallman M. Synergistic stimulation of amnion cell prostaglandin E2 synthesis by interIeukin-l, tumor necro­sis factor, and products from activated human granulo­cytes. Prostaglandins Leukot Essent Fatty Acids 1991;44: 241-5.

10. Casey ML, MacDonald PC. Biomolecular processes in the initiation of parturition: decidual activation. Clin Obstet Gynecol 1988;31 :533-52.

11. Massague J. The transforming growth factor-13 family. Annu Rev Cell Bioi 1990;6:597-641.

12. Kauma S, Matt D, Strom S, Eierman D, Turner T. Inter­leukin-ll3, human leukocyte antigen HLA-DRa, and transforming growth factor-13 expression in endome­trium, placenta, and placental membranes. AM J OBSTET GVNECOL 1990;163:1430-7.

13. Sporn MB, Roberts AB. What is TGF-I3? In: Bock GR, Marsh J, eds. Clinical applications of TGF-I3. Ciba Foun­dation Symposium. Chichester: John Wiley, 1991;157: 1-6,

14. Alitalo K, Kurkinen M, Vaheri A. Extracellular matrix components synthesized by human amniotic epithelial cells in culture. Cel! 1980;19:1053-62.

15. Downs TR, Wilfinger WW. Fluorometric quantification of DNA in cells and tissue. Anal Biochem 1983;131:538-47.

16. Berchuck A, MacDonald PC, Milevich L, Casey ML. Transforming growth factor-J3 inhibits prostaglandin pro­duction in amnion and A431 cells. Prostaglandins 1989;38:453-64.

July 1992 Am J Obstet Gynecol

17. Tashjian AH, Voelkel EF, Lazzaro M, et al. ex and 13 Hu­man transforming growth factors stimulate prostaglandin production and bone resorption in cultured mouse cal­varia. Proc Natl Acad Sci USA 1985;82:4535-8.

18. Diaz A, Varga], Jimenez SA. Transforming growth fac­tor-13 stimulation of lung fibroblast prostaglandin E2 pro­duction. J Bioi Chern 1989;264: 11554-7.

19. Ristimaki A, Ylikorkala 0, Viinikka L. Effect of growth factors on human vascular endothelial cell prostacyclin production. Arteriosclerosis 1990; 10:653-7.

20. Pfeilschifter], Pignat W, Leighton], Marki F, Vosbeck K, Alkan S. Transforming growth factor 132 differentially modulates interleukin-113 and tumour-necrosis-factor-ex­stimulated phospholipase A2 and prostaglandin E2 syn­thesis in rat renal mesangial cells, Biochem ] 1990;270:269-71.

21. Thompson NL. Flanders KC, Smith]M, Ellingsworth LR, Roberts AB, Sporn MB. Expression of transforming growth factor-131 in specific cells and tissues of adult and neonatal mice. J Cell Bioi 1989;108:661-9,

22. Altman D], Schneider SL, Thompson DA, Cheng HL. Tomasi TB. A transforming growth factor 132-like im­munosuppressive factor in amniotic fluid and localization of TGF-132 mRNA in the pregnant uterus. ] Exp Med 1990; 172: 1391-401.

23. Palladino MA, Morris RE, Starnes HF, Levinson AD. The transforming growth factor-betas, A new family of im­munoregulatory molecules. Ann NY Acad Sci 1990;593: 181-7.

24. Clark DA, Flanders KC, Banwatt 0, et al. Murine preg­nancy decidua produces a unique immunosuppressive molecule related to transforming growth factor 13-2.] Im­munol 1990;144:3008-14.

25. Ignatz RA, Massague J. Transforming growth factor-13 stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. ] Bioi Chern 1986;261:4337-45.

26. Ignotz RA, Massague J. Cell adhesion receptors as targets for transforming growth factor-13 action. Cell 1987;51: 189-97.

27. Laiho M, Saksela 0, Keski-Oja J. Transforming growth factor-13 induction of type-l plasminogen activator inhib­itor. Pericellular deposition and sensitivity to exogenous urokinase, J BioI Chern 1987;262:17467-74.

28. Edwards DR, Murphy G, Reynolds]], Whitman SE, Doch­erty AJP. Transforming growth factor beta modulates the expression of COllagenase and metalloproteinase inhibitor. EMBO J 1987;6:1899-904.

29. Uldbjerg N, Ekman G, Malmstrom A, Olsson K, Ulmsten U. Ripening of the human uterine cervix related to changes in collagen, glycosaminoglycans, and collagen­olytic activity. AM] OBSTET GVNECOL 1983;147:662-6.