Eur. J. Immunol. 1990. 20: 2485-2490
Franqois Lef&vreo, Franqoise Martinat-BottP, Michel Guillomotv, Kame1 Zouariv Bernard Charleyo and Claude La BonnardikreO Laboratoire de Virologie et Immunologie MolCculairesO and Station de Physiologie Animale, Unit6 d’Endocrinologie de I’Embryonv, lnstitut National de la Recherche Agronomique, Jouy-en-Josas and Station de Recherches sur la Physiologie de la Reproduction, Institut National de la Recherche Agronomiquen, Monnaie
Spontaneous IFN-y expression by porcine trophectoderm early in gestation
2485
Interferon-gamma gene and protein are spontaneously expressed by the porcine trophectoderm early in gestation The nature and the source of the antiviral activity found in the reproductive tract of pregnant gilts early in gestation were analyzed. Two antigenically distinct antiviral activities were found in uterine flushings and in supernatants of conceptus-conditioned culture medium between days 12 and 20 of gestation, using Madin Darby bovine kidney cells and vesicular stomatitis virus as a challenge in the antiviral bioassay. One component was antigenically identified as interferon-y (IFN-y). Northern blot analysis of conceptus poly(A)+ RNA with a human IFN-y cDNA probe revealed two mRNA of 1.3 and 1.4 kb. In addition, immunoprecipitation of metabolically labeled conceptus secretory proteins with an antiserum raised against purified porcine rIFN-y resulted in four bands in sodium dodecyl sulfate-polyacrylamide gel electrophoresis, with molecular mass 18.5 to 24.5 kDa. Pre-electrophoresis incubation of the immunoprecipitate with glycopeptidase F, which removes N-linked carbohydrates, yielded a single band of 16.5 kDa. Finally, staining of ultrathin sections by indirect immunofluorescence using the same antiserum to rIFN-y revealed that all cells of extra-embryonic trophectoderm contained intensely fluorescent granules in their apical cytoplasm. Neither endoderm nor embryonic cells stained positive. These results clearly show that IFN-y, known so far as a T o r NK cell-derived lymphokine, is spontaneously and intensively secreted by the porcine trophectoderm, an embryonic tissue not related to the hematopoietic lineage. They also suggest that the implanting conceptus, at least in the porcine species, could play an active role in immune interactions with the mother.
1 Introduction IFNy is a lymphokine produced by activated T lymphocytes and NK cells [l]. It exerts many immunoregulatory activities such as enhancement of NK and Tcell-mediated cytotoxicity, B cell differentiation, MHC antigen expression and M@ activation [2]. However, it can also induce biological responses in nonimmune cells, such as antiviral and antiproliferative effects and expression of MHC antigens [2, 31. Many hypotheses have been put forward to explain the paradoxical survival of the semiallogenic conceptus in immunocompetent maternal tissues during gestation. Most of them assume the existence of specific interactions between the conceptus and the local maternal immune system [4-61. A growing number of cytokines appear to be spontaneously expressed at the maternal-fetal interface during gestation (recently reviewed in [7]) and IFN are among these cytokines. In the human and murine species, IFN expression in placental or fetal tissues was reported only in postimplantation or late stages of gestation when the maternal-fetal relationships are well established and
[I 86091 Correspondence: Frangois Leftivre, Laboratoire de Virologie et Immunologie MolCculaires, Institut National de la Recherche Agronomique, Domaine de Vilvert, F-78350 Jouy-en-Josas, France Abbreviation: MDBK cells: Madin Darby bovine kidney cells 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
placental formation is achieved [8-121. In ruminants and pigs, by contrast, proteins with antiviral activity appear to be transiently expressed in early gestation, during the period of implantation in the uterus [13-161, suggesting that IFN, in these species, are involved in the initiation of maternal-fetal immune interactions or in other types of intercellular communications between the conceptus and the maternal organism. In ruminants, these proteins were found to be IFN-a of the class I1 subfamily expressed by the trophoblast of the conceptus and acting as embryonic signals for maintenance of pregnancy (reviewed in [17]). In the pig, a clear-cut antiviral activity has also been found in the uterine lumen and in the culture medium of conceptuses between days 12 and 17 of gestation ([15, 161 and *) but these antiviral compounds have not been completely characterized [15] and their biological role remains unknown in this species. In this report,we show that during the implantation period, which occurs in pigs between days 13 and 18 of gestation [18, 191, the uterine lumen of pregnant sows contains at least two distinct antiviral proteins. Surprisingly, one of them is an IFN-y and appears to be a conceptus product secreted by the trophectoderm, the external monolayer of the trophoblast.This finding represents the first description of an IFN-y production by non-lymphoid cells and suggests that the conceptus allograft plays an active role in early interactions with the maternal immune system before placental formation.
* La Bonnardikre et al., in press. 0014-2980/90/1111-2485$3.50+ .25/0
2486
Eur. J. Immunol. 1990.20: 2485-2490
F. Lefevre, E Martinat-BottC, M. Guillomot et al.
2 Materials and methods 2.1 Animal breeding and conceptus collection Large white (LW) and Chinese Meishan (MS) gilts were observed for estrus and were naturally inseminated. Day "zero" of gestation was defined as the first day of estrus. Pregnant gilts were slaughtered on the appropriate day of gestation, the uteruses were immediately isolated and uterine horns flushed with 20 ml of sterile PBS as previously described [15]. Elongated conceptuses were collected for further treatment. Flushing fluid was centrifuged 10 min at 1400 x g and the SN stored at -70°C for subsequent antiviral assays and neutralization tests.
2.2 In vitro culture of conceptus tissues Pooled conceptuses from the same animal were rinsed twice in PBS before introduction in the culture medium. Nonlabeled cultures were performed at 37 "C for 24 h in serumfree B2 medium [20] supplemented with aprotinin at 100 pg/ml (about 0.5 g of tissues in 12 ml of medium) and culture SN was treated as above. For metabolic labeling of conceptus secretory products, the culture was performed in the same conditions except that methionine-free MEM supplemented with FCS (2%, v/v), aprotinin (100 pg/ml) and [35S]methionine (50 pCi/ml= 1.85 MBq/ml, 800 Ci/mmole, Amersham-France, Les Ulis, France) was used. The culture medium was then centrifuged for 20min at 10000 x g, concentrated fivefold by vacuum dialysis and stored in aliquot fractions at -70°C for subsequent immunoprecipitation analysis.
2.3 Antiviral activity and seroneutralization Antiviral activity was quantified by inhibition of the cytopathic effect of vesicular stomatitis virus on Madin Darby bovine kidney (MDBK) cells as previously described [21], and expressed in laboratory standard units (SU). Seroneutralizations were assayed by adding the specified antisera (final dilution 1/100) in the titration medium [21]. A sheep anti-human leukocyte IFN serum (anti-HuIFNLe) and a rabbit serum directed against porcine rIFN-y (PorIFN-y) made in Escherichia coli and purified to homogeneity [22] were used.
2.4 Procedure for immunopurification of IFN-y activity One hundred microliters of protein A-Sepharose beads (Pharmacia, Uppsala, Sweden) was coated with Ig from rabbit anti-PorIFN-y serum or control serum (preimmune serum from the same rabbit), rinsed with 20 mM Tris-HC1, pH 7.5,lOO mM NaCl and incubated overnight at 4°C with 2.5 ml of conceptus culture SN. Beads were rinsed extensively with the same buffer and Ig-bound proteins were eluted by three successiveincubations in 100 pI of Tris-HC1, pH7.5, 1.5 M NaCl, 50% (v/v) propylene glycol. The antiviral activity detected in the eluate obtained from anti-PorIFN-y Ig was tested for acidic lability and neutralization by the anti-HuIFN-Le serum. Acidic treatment was performed by lowering the pH of the sample to 2.0 with 1 M HCl. After 2 h at 0 ° C pH was adjusted to 7.0 with 1 M NaOH .
2.5 Northern blot analysis and in vitro translation of conceptus mRNA Total RNA was isolated by the guanidinium isothiocyanatecesium chloride method from porcine conceptuses obtained between day 14 and 20 of pregnancy as described in Sect. 2.1. A poly(A)+ RNA-enriched fraction was purified by one cycle of oligo(dT) cellulose affinity chromatography. Ten micrograms of each sample was analyzed by Northern blot. Electrophoresis was carried out on a 1.5% agarose-methylmercury hydroxide gel and blotted onto nitrocellulose [23]. The probe was the 0.56 kb Dde I fragment of human IFN-y cDNA [24] encompassing three quarters of the coding region and 0.25 kb of the 3' non-coding region. Hybridization was performed at 37 "C in a standard hybridization buffer containing 30% formamide and blots were washed at 58 "C in 2 x SSC, 0.1% SDS prior to autoradiography without an intensifying screen. Poly(A)+-enriched RNA (2 pg) were translated in rabbit reticulocyte lysate (Promega Biotec, Madison, WI) in the presence of ["S]methionine according to the manufacturer's recommendations.
2.6 Immunoprecipitation Immunoprecipitations were performed on labeled mRNA translation products and secreted polypeptides prepared as described in Sects. 2.5 and 2.2, respectively. All procedures for immunoprecipitations, PAGE and fluorography were as described [25]. Each gel line was loaded with material immunoprecipitated from 400 000 (in v i m translation products) or 250 000 (conceptus secretory products) trichloracetic acid-precipitable cpm. For glycopeptidase F treatment, immunoprecipitates were eluted for 2 min at 100°C in 0.1 M 2-ME, 1% SDS, diluted ten times with 70 mM sodium acetate, pH 7.5, 10 mM EDTA, 0.7% NP40 (w/v), incubated overnight with 3.6 U/ml of glycopeptidase F (Boehringer Mannheim, Mannheim, FRG) and then concentrated by acetone precipitation before electrophoresis.
2.7 Immunohistochemistry Porcine conceptuses were collected as described in Sect. 2.1, then immediately fixed for 5 h at room temperature with 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2, washed in the same buffer without paraformaldehyde and embedded in paraffin. All procedures for tissue sections and indirect immunofluorescence are described elsewhere [26]. Before use, rabbit sera and FITC-labeled goat anti-rabbit IgG were diluted 1/50 and 1/200, respectively.
3 Results 3.1 Two types of IFN are present in uterine flushings and conceptus cultures An IFN-like antiviral activity was detected in uterine flushings from pregnant gilts between days 12 and 20 of gestation and in the SN from 24 h cultures of conceptuses
Eur. J. Immunol. 1990. 20: 2485-2490
Spontaneous IFN-y expression by porcine trophectoderm early in gestation
2487
pH 2 sensitive, similar t o lymphocyte-derived IFN-y from
Figure 1. Characterization of an IFN-y antiviral activity present in the uterine lumen of pregnant gilts during implantation and in the SN of implanting conceptus culture. (a) Neutralization of the antiviral activity in uterine flushings of day 16-pregnantsows using a sheep anti-HuIFN-Le serum (anti-IFN-Le) and a rabbit antiPorIFN-y serum (anti-IFN-y) [22], individually and in combination. (b) Similar analysis in the SN of day 16 porcine conceptus cultures. Values presented are the mean f SEM of five samples from different animals.
collected during the same period ([4, 51 and *). To characterize this antiviral activity, seroneutralization experiments were performed.We found that the antiviral activity present in both types of samples was partially, but significantly, neutralized by two different antisera, o n e directed against HuIFN-Le and the other against PorIFN-y [22], (Fig. 1A and b). A mixture of both antisera abolished the antiviral activity almost completely (> 95%). Moreover, it was possible t o purify part of the antiviral activity released in conceptus cultures by immunoaffinity on immobilized Ig from the anti-PorIFN-y serum (Fig. 2). In contrast, parallel experiments using control serum gave no detectable antiviral activity (Fig. 2). This purified antiviral activity was
* La Bonnardikre et al.,
other species [27] and it was not neutralized by antiHuIFN-Le serum, clearly showing that each antiserum neutralized distinct antiviral proteins in conceptus culture SN and probably also in corresponding uterine flushings. T h e anti-PorIFN-y serum used in this study was highly monospecific since it was raised against a n electrophoretically pure recombinant protein [22]. This serum neutralized the antiviral effect of PorLFN-y and the antiviral activity secreted by lectin-stimulated porcine lymphocytes [22], but did not neutralize the antiviral activity of porcine IFN-Le, porcine IFN-fi [22] and PorIFN-a (data not shown). Thus, these results indicate that during the implantation period, the same two types of antiviral proteins are spontaneously expressed in viva in the uterine lumen and released in vitro by conceptuses. O n e is antigenically related to human leukocyte-derived IFN and t h e other is an IFN-y. Our subsequent experiments were focused on the characterization of this IFN-y.
3.2 Conceptus mRNA contain 1FN-y-related transcripts and encode IFN-y-related proteins To establish that this PoIFN-y activity was actually synthesized by the conceptus, rather than taken from uterine secretions then released into the culture medium, we examined the conceptuses for the presence of IFN-y m R N A by Northern blot analysis. Using a heterologous human IFN-y cDNA probe, two distinct m R N A species of about 1.3 and 1.4 k b in length were detected in conceptus tissues between days14 and 20 of pregnancy (Fig. 3).
DAYS OF GESTATION
kb
14 15 16 17 20
in press.
4.4 -
1.4 -
U
Treatment
none none - 1
Antiserum used for immunopurification
ctr.
PH2
y;e,
I
anti- 1FN-y
Figure 2. Immunopurification of IFN-y antiviral activity released by day 16 porcine conceptus in culture. Immunopurification was performed on the SN of a conceptus culture using immobilized Ig from rabbit anti-PorIFN-y or control (ctr.) sera. No detectable antiviral activity was present in the eluate from control serum Ig. The antiviral activity detected in the eluate from anti-IFN-y serum Ig was p H 2 sensitive and not neutralized by the anti-human IFN-Le serum.
0.24 Figure 3. Identification of mRNA cross-hybridizing to human IFN-y cDNA in porcine conceptuses during the implantation period. Poly(A)+-enriched fractions (10 pg per lane) from conceptus tissues sampled at the appropriate day of gestation were analyzed by Northern blot and hybridized with a human IFN-y cDNA probe under low stringency. The two RNA transcripts detected (1.3 and 1.4 kb) are indicated by arrows. Position of RNA molecular mass standards is indicated.
Eur. J. Immunol. 1990. 20: 2485-2490
F. Lefi?vre. F. Martinat-Bottt, M. Guillomot et al.
2488
Synthesis of both mRNA appeared to be maximal on days 14 to 15 and decreased until day 20. Moreover, the anti-PorIFN-y serum specifically immunoprecipitated a single protein species with an apparent molecular mass of 18.5 kDa from the in vitro translation products of day 15 conceptus mRNA (Fig. 4). This protein is in the size range reported, according to cDNA sequences, for lymphocytederived IFN-y preprotein of man, cattle, mouse and rat [28]. Thus, our results indicate that conceptus mFWA can encode a protein having the antigenicity and the size of an IFN-y preprotein.
sylated mature protein has a predicted size of 16-17 kDa [28]; however, the occurrence of N-glycosylation sites in sequences might result in several secreted forms of higher and discrete apparent molecular mass.These forms are well characterized in natural human IFN-y [27]. Accordingly, these results demonstrate that the porcine conceptus synthesizes during the implantation period an IFN-y that exhibits biochemical and structural similarities with lymphocyte-derived homologues from other species.
3.4 IFN-y protein is expressed by trophectoderm cells 3.3 IFN-y protein is expressed by the conceptus To establish that the mature form of this protein was secreted by conceptuses, the secretory products from isolated day 15 conceptuses were metabolically labeled with [35S]methionineand subjected to immunoprecipitation analysis using the anti-PorIFN-y serum. Four major labeled polypeptides of apparent molecular mass 18.5, 20.5,23 and 24.5 kDa were specifically recognized (Fig. 4). When treated with glycopeptidase F (an enzyme removing N-linked carbohydrates from glycoproteins, [29]) the four protein species were resolved into a single 16.5-kDa band (Fig. 4), strongly suggesting that they represented different N-glycosylated forms of the same polypeptide. These obervations are consistent with previous findings: in the four mammalian species mentioned above, IFN-y unglyco-
To identify the cellular origin of this IFN-y production, tissue sections from days 15 to 17 conceptuses were subjected to indirect immunofluorescence staining with antiPorIFN-y serum. The specific fluorescence was localized within the apical perinuclear cytoplasm of almost all cells of the trophectoderm (Fig. 5). No cell with the distinctive morphology of lymphocytes or other blood cells could be seen in this monolayer, ruling out the possibility of an IFN-y expression from ectopic maternal lymphoid cells.The other constituents of the extra-embryonic membranes (i.e. parietal endoderm and mesoderm) and the embryonic zone
IN VlTRO CULTURE TRANSLATION SUPERNATANT
' antiM r(K)
' ' anti -
anti?
IFN-y ctr.s. IFN-'I ctr.s. IFN-y -gl.F +gl.F
67.0-
43.0-
30.0-
20.1 -
14.4-
Figure 4. Immunoprecipitation of IFN-y from the in vilro mRNA translation products and secreted polypeptides of day 15 porcine conceptuses. Conceptus poly(A)+ RNA were translated in rabbit reticulocyte lysate in the presence of [35S]methionine and translation products were analyzed by immunoprecipitation using rabbit anti-PorIFN-y (anti-IFN-y) and control preimmune (ctr. s.) sera. Porcine conceptuses were cultured for 24 h in the presence of [35S]methionine and the SN was analyzed in the same way. Immunoprecipitated polypeptides were electrophorescd before (- g1.F) and after (+ g1.F) treatment with glycopeptidase F. Rabbit control preimmune serum did not precipitate significant amounts of labeled material. Position of protein molecular mass standards is indicated.
(bl Figure 5. Detection of immunoreactive IFN-y in trophectoderm of a day 16 porcine conceptus. (a) Indirect immunofluorescence staining of a trophoblast section using rabbit anti-PorLFN-y serum ( x 340). The immunoreactive product was located only in the apical cytoplasm of trophectoderm cells and absent from the endodcrm layer (arrowed). (b) Similar analysis performed on a section of the cephalic region of the embryo ( x 85). (c) Control experiment performed on a section adjacent to (a) and treated with the rabbit preimmune control serum ( x 340). Similar results were obtained with day 15 and 17 conceptuses.
Eur. J. Immunol. 1990. 20: 2485-2490
Spontaneous IFN-y expression by porcine trophectoderm early in gestation
(Fig. 5) were clearly not labeled, indicating that IFN-y was expressed only by the trophectoderm. Maternal endometrium was also negative (not shown).
4 Discussion Our results clearly show that two distinct types of antiviral activity are spontaneously expressed at the maternal-fetal interface during early pregnancy in pigs. The first type appeared to be antigenically related to HuIFN-Le, which is known to contain mainly IFN-a subspecies and traces of other antiviral cytokines [30].Thus, although it remains to be precisely identified, this antiviral activity is probably IFN-a related, as also suggested by another recent report [ 151. However, further studies are required to determine whether this IFN-like activity is actually a conceptus product. The amount released in vitro by isolated conceptuses appeared to be lower than the amount found in vivo (Fig. 1) so the possibility that it was in fact a maternal product taken nonspecifically from uterine secretions by conceptus tissues and released in vitro cannot be totally excluded, although it seems unlikely. The second type is clearly a conceptus-derived IFN-y expressed by almost all the cells of the trophectoderm. This is the first report of an IFN-y synthesis by cells that do not belong to the hematopoietic lineage. The identity of porcine trophoblastic and immune I F N y was not proven by the present data, so it is possible that each protein is encoded by distinct genes. Moreover, our results show that two distinct IFN-y-related mRNA species are expressed by the conceptus.The in vitro translation experiment indicates that both mRNA probably encode the same preprotein, so their duality could be most simply explained by transcription from a single gene expressing alternatively spliced transcripts or bearing two functional polyadenylation signals. However, their expression from two genes encoding antigenically related preproteins of identical size cannot be excluded. Thus, the actual number of IFN-y genes in the pig species remains to be elucidated. If there is a single gene, as in all mammalian species studied so far [28], it will be interesting to study how this gene is differentially regulated in trophectoderm cells and lymphocytes. It has recently been shown that, during implantation, the trophoblast of ruminant conceptus secretes proteins with antiviral activity belonging to the class I1 subfamily of IFN-a [31-341.These IFN-a are supposed to play an active role in maternal recognition of pregnancy by prolonging the life-span of the corpus luteum on the ovary, since they have such an effect when they are locally administered in the uterus of cyclic ruminant females [35, 361. It is possible that the porcine IFN-a-related antiviral activity mentioned above represents the homologue of these ruminant proteins; however, maternal recognition of pregnancy is mediated in the pig by distinct mechanisms so their biological role during early pregnancy is probably different. Moreover, unlike in the pig, no IFN-y seems to be expressed by the ruminant conceptus as all the antiviral activity detected in uterine flushings or in conceptus culture can be neutralized by antisera directed against the purified conceptus-derived IFN-a (J. Martal, personal communication). Also, we have preliminary data indicating the absence of IFN-y-related mRNA in sheep conceptus during implantation (data not shown). Thus, spontaneous expres-
2489
sion of IFN-y by trophoblast early in gestation is probably not a general phenomenon in mammals and could be an original feature of the pig. This physiological IFN-y expression seems to be temporally regulated as no significant antiviral activity is secreted by conceptuses outside the day-12 to -20 period ([15, 161 and *). During this period, trophoblastic tissues represent the major part of the conceptus and establish intimate structural and functional relationships with the endometrium [19,37]. The fact that the porcine embryonic trophectoderm expresses an important immunoregulatory molecule suggests that the conceptus allograft plays an active role in early interactions with the maternal immune system before placental formation. Recent studies in rodents and humans indicate that paternal MHC antigens and other fetal alloantigens are expressed by the trophoblast or surrounding fetal tissues during pregnancy and elicit a maternal immune response [4]. There is much evidence to indicate that the deleterious effects of this immune response on the fetus are selectively impaired by local immunoregulatory events leading to local immune suppression and inhibition of fetal allograft rejection [4-61. Porcine trophoblastic IFN-y could act on the local maternal immune system to induce such specific mechanisms. The study of intrauterine immunosuppression has been recently initiated in species with epithelial placentation, such as the pig for which the presence of intrauterine suppressive cells has been reported in early pregnancy [38]. Conceptus-derived IFN-y could be involved directly or indirectly in the recruitment of these cells. Moreover, a specific recruitment of NK cells to the endometrial stroma and a significant reduction of intraepithelial lymphocytes in the endometrium were also recently reported during early pregnancy in the pig [39,40] and could be related to an IFN-y expression. Alternatively, one of the main biological properties of IFN-y is its ability to induce MHC class I and class I1 antigen expression [3]. No data are available about MHC antigen expression at the maternal-fetal interface during implantation in the pig. In the human species, the absence of classical polymorphic MHC class I and class I1 molecule expression on trophoblast has been reported during gestation [41,42]. This fact was often interpreted as functional since expression of paternal class I antigen would presumably risk maternal rejection of the fetus. However, several lines of evidence indicate that less polymorphic class I molecules are expressed by trophoblast cells in several species including man [43-451. It has been suggested that these molecules might have different functions from those encoded by polymorphic class I MHC loci and could have a role in enabling the fetus to escape recognition by the maternal immune system [46] or in inducing specific suppression of the maternal immune response [42]. Whether IFN-y is the physiological autocrine inducer of such MHC molecules on the porcine trophoblast will be an interesting question to solve. However, the possibility that this IFN-y induces classical MHC class I and class I1 antigens on trophoblast or maternal tissues must also be investigated. Spontaneous and temporally regulated expression of IFNy by the trophoblast in early pregnancy has an important
* La Bonnardi2re
et al., in press.
2490
E Lefkvre, E Martinat-BottC, M. Guillomot et al.
biological significance, although its occurrence in other mammalian species remains to be explored. It indicates that the early conceptus, though devoid of adult lymphoid organs [47], is an immunologically active allograft emitting specific signals, possibly to communicate with the maternal immune system. It also suggests that developmentally controlled expression of lymphokines by the trophoblast modulates the maternal immune response against the conceptus. We thank C. De Vaureix, R . Procureur and t?Despris for excellent technical assistance; Dr. C. Chany (INSERM, Paris) for the anti-human IFN-Le serum; Dr. W Fiers (Ghent, Belgium) for the human IFN-y cDNA; Dr. B. Delmas for discussions; and Drs. J. Jarvis and H . Laude for critical reading of the manuscript. Received June 5, 1990.
5 References 1 Vilcek, J., Gray, P.W., Rinderknecht, E. and Sevastopoulos, C. G., Lymphokines 198.5. 11: 1. 2 De Maeyer, E. and De Maeyer-Guignard, J., Interferons and Other Regulatory Cytokines, John Wiley and Sons, New York 1988, p. 144. 3 Rosa, F. M., Cochet, M. M. and Fellous, M., in Gresser, I. (Ed.), Interferon, vol. 7, Academic Press, London 1986, p. 47. 4 Hunziker, R. D. and Wegmann, T. G., CRC Crit. Rev. Immunol. 1986. 6: 245. 5 Chaouat, G., J. Reprod. Immunol. 1987. 10: 179. 6 Wegmann, T. G., Immunol. Lett. 1988. 17: 297. 7 Hunt, S. H., J. Reprod. Immunol. 1989. 10: 1. 8 Duc-Goiran, I?, Robert-Galliot, B., Lopez, J. and Chany, C., Proc. Natl. Acad. Sci. USA 1985. 82: 5010. 9 Chard, T., Craig, P. H., Menabawey, M. and Lee, C., Br. J. Obstet. Gynaecol. 1986. 93: 1145. 10 Fowler, A. K., Reed, C. D. and Giron, D. J., Nature 1980.286: 266. 11 Baker, D. J. and Nieder, G. L., J. Reprod. Fertil. 1990. 88: 307. 12 Toth, F. D., Juhl, C. N0rskov-Lauritsen, N., Mosborg Petersen, P. and Ebbesen, P., J. Reprod. Immunol. 1990. 17: 217. 13 Pontzer, C. H., Torres, B. A., Vallet. J. L., Bazer, F. W. and Johnson, H. M., Biochem. Biophys. Res. Commun. 1988.152: 801. 14 Betteridge, K. J., Derbyshire, J. B., Rorie, R.W., Scodvas, J. M. and Johnson,W. H., J. Reprod. Fertil. (abstract series) 1988. I : 21. 15 Cross, J. C. and Roberts, R. M., Biol. Reprod. 1989. 40: 1109. 16 Mirando, M. A., Harney, J. P., Beers, S., Pontzer, C. H.,Torres, B. A., Johnson, H. M. and Bazer, F.W., J. Reprod. Fertil. 1990. 88: 197. 17 Roberts, R. M., J. Interferon Res. 1989. 9: 373. 18 King, G. J., Atkinson, B. A. and Robertson, H. A., J. Reprod. Fertil. [Suppl.] 1982. 31: 17.
Eur. J. Immunol. 1990.20: 2485-2490
19 Dantzer,V., Anat. Embryol. 1985. 172: 281. 20 Menezo,Y., C.R. Acad. Sci. Paris [D] 1976. 282: 1967. 21 La Bonnardiere, C., Laude, H. and Berg, K., Ann. Inst. PasteurlVirol. 1986. 137E: 171. 22 Charley. B., McCullough, K. and Martinod, S., Vet. Immunol. Immunopathol. 1988. 19: 95. 23 Sambrook, J., Fritsch, E. F. and Maniatis, T., Molecular Cloning: a Laboratory Manual, 2nd Edn., Cold Spring Harbor Laboratory Press, New York 1989. 24 Devos, R., Cheroutre. H., Taya,Y., Degrave, W. ,Van Heuverswyn, H. and Fiers, W., Nucleic Acids Res. 1982. 10: 2487. 25 Laude, H., Chapsal, J. M., Gelfi, J., Labiau, S. and GrosClaude, J., J. Gen. Virol. 1986. 67: 119. 26 Guillomot, M., Michel, C., Gayc, P., Charlier, M.,Trojan, M. and Martal. J., Biol. Cell 1990, in press. 27 Langer. J. A. and Pestka, S., Pharmacol. Ther. 1985. 27: 371. 28 Weissman, C. and Weber, H., Prog. Nucleic Acid Res. Mol. Biol. 1986. 33: 251. 29 Tarentino, A. L., Gomez, C. M. and Plummer, T. H., Jr., Biochemistry 1985. 24: 4665. 30 De Maeyer, E. and De Maeyer-Guignard. J., Interferons and Other Regulatory Cytokines, John Wiley and Sons, New York 1988, p. 42.5. 31 Imakawa, K., Anthony, R. V.. Kazemi, M., Marotti, K. R., Polirzq, H. F. and Roberts, R. M., Nature 1987. 330: 377. 32 Charpigny, G., Reinaud, P.. Huet, J. C., Guillomot, M., Charlier, M., Pernollet, J. C. and Martal, J., FEBS Lett. 1988. 228: 12. 33 Charlier, M., Huc, D., Martal, J. and Gaye, P.. Gene 1989. 77: 341. 34 Imakawa, K.. Hansen, T. R., Malathy, P.-V., Anthony, R.V., Polites, H.V., Marotti, K. R. and Roberts, R. M., Endocrinology 1989. 3: 127. 35 Vallet, J. L., Bazer, F. W., Fliss, M. F.V. and Thatcher,W. W., J. Reprod. Fertil. 1988. 84: 493. 36 Thatcher, W. W., Hansen, P. J., Gross, T. S., Helmer, S. D., Plante, C. and Bazer, F.W., J. Reprod. Fertil. [Suppl.] 1989.37: 91. 37 Bazer, F. W. and Roberts, R. M., J. Exp. Zool. 1983. 228: 373. 38 Croy, B. A., Wood, W. and King, G. J., J. Immunol. 1987.139: 1088. 39 Croy, B. A., Waterfield, A., Wood, W. and King, G. J., Cell. Immunol. 1988. 115: 471. 40 King, G. J., J. Reprod. Immunol. 1988. 14: 41. 41 Redman, C. W. G.. McMichael, A. J., Stirrat, G. M., Sunderland, C. A. and Ting, A., Immunology 1984. 52: 457. 42 Bulmer. J. N. and Johnson, P. M., Placenta 1985. 6: 127. 43 Ellis, S. A., Palmer, M. S. and McMichael, A. J., J. Immunol. 1990. 144: 713. 44 Stern, P. L., Beresford, N., Friedman, C. I., Stevens, V. C., Risk, J. M. and Johnson, P. M., J. Immunol. 1987. 138: 1088. 45 MacPherson,T. A . , Hon-Nerng, H., Kunz, H.W. and Gil1,T. J., Transplantation 1986. 41: 392. 46 Head, J. R., Drake, B. L. and Zuckermann, F. A., A m . J. Reprod. Immunol. Microbiol. 1987. 15: 12. 47 Gill, T. J., 111 and Repetti, C. F., A m . J. Pathol. 1979. 95: 463.
Franqois Lef&vreo, Franqoise Martinat-BottP, Michel Guillomotv, Kame1 Zouariv Bernard Charleyo and Claude La BonnardikreO Laboratoire de Virologie et Immunologie MolCculairesO and Station de Physiologie Animale, Unit6 d’Endocrinologie de I’Embryonv, lnstitut National de la Recherche Agronomique, Jouy-en-Josas and Station de Recherches sur la Physiologie de la Reproduction, Institut National de la Recherche Agronomiquen, Monnaie
Spontaneous IFN-y expression by porcine trophectoderm early in gestation
2485
Interferon-gamma gene and protein are spontaneously expressed by the porcine trophectoderm early in gestation The nature and the source of the antiviral activity found in the reproductive tract of pregnant gilts early in gestation were analyzed. Two antigenically distinct antiviral activities were found in uterine flushings and in supernatants of conceptus-conditioned culture medium between days 12 and 20 of gestation, using Madin Darby bovine kidney cells and vesicular stomatitis virus as a challenge in the antiviral bioassay. One component was antigenically identified as interferon-y (IFN-y). Northern blot analysis of conceptus poly(A)+ RNA with a human IFN-y cDNA probe revealed two mRNA of 1.3 and 1.4 kb. In addition, immunoprecipitation of metabolically labeled conceptus secretory proteins with an antiserum raised against purified porcine rIFN-y resulted in four bands in sodium dodecyl sulfate-polyacrylamide gel electrophoresis, with molecular mass 18.5 to 24.5 kDa. Pre-electrophoresis incubation of the immunoprecipitate with glycopeptidase F, which removes N-linked carbohydrates, yielded a single band of 16.5 kDa. Finally, staining of ultrathin sections by indirect immunofluorescence using the same antiserum to rIFN-y revealed that all cells of extra-embryonic trophectoderm contained intensely fluorescent granules in their apical cytoplasm. Neither endoderm nor embryonic cells stained positive. These results clearly show that IFN-y, known so far as a T o r NK cell-derived lymphokine, is spontaneously and intensively secreted by the porcine trophectoderm, an embryonic tissue not related to the hematopoietic lineage. They also suggest that the implanting conceptus, at least in the porcine species, could play an active role in immune interactions with the mother.
1 Introduction IFNy is a lymphokine produced by activated T lymphocytes and NK cells [l]. It exerts many immunoregulatory activities such as enhancement of NK and Tcell-mediated cytotoxicity, B cell differentiation, MHC antigen expression and M@ activation [2]. However, it can also induce biological responses in nonimmune cells, such as antiviral and antiproliferative effects and expression of MHC antigens [2, 31. Many hypotheses have been put forward to explain the paradoxical survival of the semiallogenic conceptus in immunocompetent maternal tissues during gestation. Most of them assume the existence of specific interactions between the conceptus and the local maternal immune system [4-61. A growing number of cytokines appear to be spontaneously expressed at the maternal-fetal interface during gestation (recently reviewed in [7]) and IFN are among these cytokines. In the human and murine species, IFN expression in placental or fetal tissues was reported only in postimplantation or late stages of gestation when the maternal-fetal relationships are well established and
[I 86091 Correspondence: Frangois Leftivre, Laboratoire de Virologie et Immunologie MolCculaires, Institut National de la Recherche Agronomique, Domaine de Vilvert, F-78350 Jouy-en-Josas, France Abbreviation: MDBK cells: Madin Darby bovine kidney cells 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
placental formation is achieved [8-121. In ruminants and pigs, by contrast, proteins with antiviral activity appear to be transiently expressed in early gestation, during the period of implantation in the uterus [13-161, suggesting that IFN, in these species, are involved in the initiation of maternal-fetal immune interactions or in other types of intercellular communications between the conceptus and the maternal organism. In ruminants, these proteins were found to be IFN-a of the class I1 subfamily expressed by the trophoblast of the conceptus and acting as embryonic signals for maintenance of pregnancy (reviewed in [17]). In the pig, a clear-cut antiviral activity has also been found in the uterine lumen and in the culture medium of conceptuses between days 12 and 17 of gestation ([15, 161 and *) but these antiviral compounds have not been completely characterized [15] and their biological role remains unknown in this species. In this report,we show that during the implantation period, which occurs in pigs between days 13 and 18 of gestation [18, 191, the uterine lumen of pregnant sows contains at least two distinct antiviral proteins. Surprisingly, one of them is an IFN-y and appears to be a conceptus product secreted by the trophectoderm, the external monolayer of the trophoblast.This finding represents the first description of an IFN-y production by non-lymphoid cells and suggests that the conceptus allograft plays an active role in early interactions with the maternal immune system before placental formation.
* La Bonnardikre et al., in press. 0014-2980/90/1111-2485$3.50+ .25/0
2486
Eur. J. Immunol. 1990.20: 2485-2490
F. Lefevre, E Martinat-BottC, M. Guillomot et al.
2 Materials and methods 2.1 Animal breeding and conceptus collection Large white (LW) and Chinese Meishan (MS) gilts were observed for estrus and were naturally inseminated. Day "zero" of gestation was defined as the first day of estrus. Pregnant gilts were slaughtered on the appropriate day of gestation, the uteruses were immediately isolated and uterine horns flushed with 20 ml of sterile PBS as previously described [15]. Elongated conceptuses were collected for further treatment. Flushing fluid was centrifuged 10 min at 1400 x g and the SN stored at -70°C for subsequent antiviral assays and neutralization tests.
2.2 In vitro culture of conceptus tissues Pooled conceptuses from the same animal were rinsed twice in PBS before introduction in the culture medium. Nonlabeled cultures were performed at 37 "C for 24 h in serumfree B2 medium [20] supplemented with aprotinin at 100 pg/ml (about 0.5 g of tissues in 12 ml of medium) and culture SN was treated as above. For metabolic labeling of conceptus secretory products, the culture was performed in the same conditions except that methionine-free MEM supplemented with FCS (2%, v/v), aprotinin (100 pg/ml) and [35S]methionine (50 pCi/ml= 1.85 MBq/ml, 800 Ci/mmole, Amersham-France, Les Ulis, France) was used. The culture medium was then centrifuged for 20min at 10000 x g, concentrated fivefold by vacuum dialysis and stored in aliquot fractions at -70°C for subsequent immunoprecipitation analysis.
2.3 Antiviral activity and seroneutralization Antiviral activity was quantified by inhibition of the cytopathic effect of vesicular stomatitis virus on Madin Darby bovine kidney (MDBK) cells as previously described [21], and expressed in laboratory standard units (SU). Seroneutralizations were assayed by adding the specified antisera (final dilution 1/100) in the titration medium [21]. A sheep anti-human leukocyte IFN serum (anti-HuIFNLe) and a rabbit serum directed against porcine rIFN-y (PorIFN-y) made in Escherichia coli and purified to homogeneity [22] were used.
2.4 Procedure for immunopurification of IFN-y activity One hundred microliters of protein A-Sepharose beads (Pharmacia, Uppsala, Sweden) was coated with Ig from rabbit anti-PorIFN-y serum or control serum (preimmune serum from the same rabbit), rinsed with 20 mM Tris-HC1, pH 7.5,lOO mM NaCl and incubated overnight at 4°C with 2.5 ml of conceptus culture SN. Beads were rinsed extensively with the same buffer and Ig-bound proteins were eluted by three successiveincubations in 100 pI of Tris-HC1, pH7.5, 1.5 M NaCl, 50% (v/v) propylene glycol. The antiviral activity detected in the eluate obtained from anti-PorIFN-y Ig was tested for acidic lability and neutralization by the anti-HuIFN-Le serum. Acidic treatment was performed by lowering the pH of the sample to 2.0 with 1 M HCl. After 2 h at 0 ° C pH was adjusted to 7.0 with 1 M NaOH .
2.5 Northern blot analysis and in vitro translation of conceptus mRNA Total RNA was isolated by the guanidinium isothiocyanatecesium chloride method from porcine conceptuses obtained between day 14 and 20 of pregnancy as described in Sect. 2.1. A poly(A)+ RNA-enriched fraction was purified by one cycle of oligo(dT) cellulose affinity chromatography. Ten micrograms of each sample was analyzed by Northern blot. Electrophoresis was carried out on a 1.5% agarose-methylmercury hydroxide gel and blotted onto nitrocellulose [23]. The probe was the 0.56 kb Dde I fragment of human IFN-y cDNA [24] encompassing three quarters of the coding region and 0.25 kb of the 3' non-coding region. Hybridization was performed at 37 "C in a standard hybridization buffer containing 30% formamide and blots were washed at 58 "C in 2 x SSC, 0.1% SDS prior to autoradiography without an intensifying screen. Poly(A)+-enriched RNA (2 pg) were translated in rabbit reticulocyte lysate (Promega Biotec, Madison, WI) in the presence of ["S]methionine according to the manufacturer's recommendations.
2.6 Immunoprecipitation Immunoprecipitations were performed on labeled mRNA translation products and secreted polypeptides prepared as described in Sects. 2.5 and 2.2, respectively. All procedures for immunoprecipitations, PAGE and fluorography were as described [25]. Each gel line was loaded with material immunoprecipitated from 400 000 (in v i m translation products) or 250 000 (conceptus secretory products) trichloracetic acid-precipitable cpm. For glycopeptidase F treatment, immunoprecipitates were eluted for 2 min at 100°C in 0.1 M 2-ME, 1% SDS, diluted ten times with 70 mM sodium acetate, pH 7.5, 10 mM EDTA, 0.7% NP40 (w/v), incubated overnight with 3.6 U/ml of glycopeptidase F (Boehringer Mannheim, Mannheim, FRG) and then concentrated by acetone precipitation before electrophoresis.
2.7 Immunohistochemistry Porcine conceptuses were collected as described in Sect. 2.1, then immediately fixed for 5 h at room temperature with 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2, washed in the same buffer without paraformaldehyde and embedded in paraffin. All procedures for tissue sections and indirect immunofluorescence are described elsewhere [26]. Before use, rabbit sera and FITC-labeled goat anti-rabbit IgG were diluted 1/50 and 1/200, respectively.
3 Results 3.1 Two types of IFN are present in uterine flushings and conceptus cultures An IFN-like antiviral activity was detected in uterine flushings from pregnant gilts between days 12 and 20 of gestation and in the SN from 24 h cultures of conceptuses
Eur. J. Immunol. 1990. 20: 2485-2490
Spontaneous IFN-y expression by porcine trophectoderm early in gestation
2487
pH 2 sensitive, similar t o lymphocyte-derived IFN-y from
Figure 1. Characterization of an IFN-y antiviral activity present in the uterine lumen of pregnant gilts during implantation and in the SN of implanting conceptus culture. (a) Neutralization of the antiviral activity in uterine flushings of day 16-pregnantsows using a sheep anti-HuIFN-Le serum (anti-IFN-Le) and a rabbit antiPorIFN-y serum (anti-IFN-y) [22], individually and in combination. (b) Similar analysis in the SN of day 16 porcine conceptus cultures. Values presented are the mean f SEM of five samples from different animals.
collected during the same period ([4, 51 and *). To characterize this antiviral activity, seroneutralization experiments were performed.We found that the antiviral activity present in both types of samples was partially, but significantly, neutralized by two different antisera, o n e directed against HuIFN-Le and the other against PorIFN-y [22], (Fig. 1A and b). A mixture of both antisera abolished the antiviral activity almost completely (> 95%). Moreover, it was possible t o purify part of the antiviral activity released in conceptus cultures by immunoaffinity on immobilized Ig from the anti-PorIFN-y serum (Fig. 2). In contrast, parallel experiments using control serum gave no detectable antiviral activity (Fig. 2). This purified antiviral activity was
* La Bonnardikre et al.,
other species [27] and it was not neutralized by antiHuIFN-Le serum, clearly showing that each antiserum neutralized distinct antiviral proteins in conceptus culture SN and probably also in corresponding uterine flushings. T h e anti-PorIFN-y serum used in this study was highly monospecific since it was raised against a n electrophoretically pure recombinant protein [22]. This serum neutralized the antiviral effect of PorLFN-y and the antiviral activity secreted by lectin-stimulated porcine lymphocytes [22], but did not neutralize the antiviral activity of porcine IFN-Le, porcine IFN-fi [22] and PorIFN-a (data not shown). Thus, these results indicate that during the implantation period, the same two types of antiviral proteins are spontaneously expressed in viva in the uterine lumen and released in vitro by conceptuses. O n e is antigenically related to human leukocyte-derived IFN and t h e other is an IFN-y. Our subsequent experiments were focused on the characterization of this IFN-y.
3.2 Conceptus mRNA contain 1FN-y-related transcripts and encode IFN-y-related proteins To establish that this PoIFN-y activity was actually synthesized by the conceptus, rather than taken from uterine secretions then released into the culture medium, we examined the conceptuses for the presence of IFN-y m R N A by Northern blot analysis. Using a heterologous human IFN-y cDNA probe, two distinct m R N A species of about 1.3 and 1.4 k b in length were detected in conceptus tissues between days14 and 20 of pregnancy (Fig. 3).
DAYS OF GESTATION
kb
14 15 16 17 20
in press.
4.4 -
1.4 -
U
Treatment
none none - 1
Antiserum used for immunopurification
ctr.
PH2
y;e,
I
anti- 1FN-y
Figure 2. Immunopurification of IFN-y antiviral activity released by day 16 porcine conceptus in culture. Immunopurification was performed on the SN of a conceptus culture using immobilized Ig from rabbit anti-PorIFN-y or control (ctr.) sera. No detectable antiviral activity was present in the eluate from control serum Ig. The antiviral activity detected in the eluate from anti-IFN-y serum Ig was p H 2 sensitive and not neutralized by the anti-human IFN-Le serum.
0.24 Figure 3. Identification of mRNA cross-hybridizing to human IFN-y cDNA in porcine conceptuses during the implantation period. Poly(A)+-enriched fractions (10 pg per lane) from conceptus tissues sampled at the appropriate day of gestation were analyzed by Northern blot and hybridized with a human IFN-y cDNA probe under low stringency. The two RNA transcripts detected (1.3 and 1.4 kb) are indicated by arrows. Position of RNA molecular mass standards is indicated.
Eur. J. Immunol. 1990. 20: 2485-2490
F. Lefi?vre. F. Martinat-Bottt, M. Guillomot et al.
2488
Synthesis of both mRNA appeared to be maximal on days 14 to 15 and decreased until day 20. Moreover, the anti-PorIFN-y serum specifically immunoprecipitated a single protein species with an apparent molecular mass of 18.5 kDa from the in vitro translation products of day 15 conceptus mRNA (Fig. 4). This protein is in the size range reported, according to cDNA sequences, for lymphocytederived IFN-y preprotein of man, cattle, mouse and rat [28]. Thus, our results indicate that conceptus mFWA can encode a protein having the antigenicity and the size of an IFN-y preprotein.
sylated mature protein has a predicted size of 16-17 kDa [28]; however, the occurrence of N-glycosylation sites in sequences might result in several secreted forms of higher and discrete apparent molecular mass.These forms are well characterized in natural human IFN-y [27]. Accordingly, these results demonstrate that the porcine conceptus synthesizes during the implantation period an IFN-y that exhibits biochemical and structural similarities with lymphocyte-derived homologues from other species.
3.4 IFN-y protein is expressed by trophectoderm cells 3.3 IFN-y protein is expressed by the conceptus To establish that the mature form of this protein was secreted by conceptuses, the secretory products from isolated day 15 conceptuses were metabolically labeled with [35S]methionineand subjected to immunoprecipitation analysis using the anti-PorIFN-y serum. Four major labeled polypeptides of apparent molecular mass 18.5, 20.5,23 and 24.5 kDa were specifically recognized (Fig. 4). When treated with glycopeptidase F (an enzyme removing N-linked carbohydrates from glycoproteins, [29]) the four protein species were resolved into a single 16.5-kDa band (Fig. 4), strongly suggesting that they represented different N-glycosylated forms of the same polypeptide. These obervations are consistent with previous findings: in the four mammalian species mentioned above, IFN-y unglyco-
To identify the cellular origin of this IFN-y production, tissue sections from days 15 to 17 conceptuses were subjected to indirect immunofluorescence staining with antiPorIFN-y serum. The specific fluorescence was localized within the apical perinuclear cytoplasm of almost all cells of the trophectoderm (Fig. 5). No cell with the distinctive morphology of lymphocytes or other blood cells could be seen in this monolayer, ruling out the possibility of an IFN-y expression from ectopic maternal lymphoid cells.The other constituents of the extra-embryonic membranes (i.e. parietal endoderm and mesoderm) and the embryonic zone
IN VlTRO CULTURE TRANSLATION SUPERNATANT
' antiM r(K)
' ' anti -
anti?
IFN-y ctr.s. IFN-'I ctr.s. IFN-y -gl.F +gl.F
67.0-
43.0-
30.0-
20.1 -
14.4-
Figure 4. Immunoprecipitation of IFN-y from the in vilro mRNA translation products and secreted polypeptides of day 15 porcine conceptuses. Conceptus poly(A)+ RNA were translated in rabbit reticulocyte lysate in the presence of [35S]methionine and translation products were analyzed by immunoprecipitation using rabbit anti-PorIFN-y (anti-IFN-y) and control preimmune (ctr. s.) sera. Porcine conceptuses were cultured for 24 h in the presence of [35S]methionine and the SN was analyzed in the same way. Immunoprecipitated polypeptides were electrophorescd before (- g1.F) and after (+ g1.F) treatment with glycopeptidase F. Rabbit control preimmune serum did not precipitate significant amounts of labeled material. Position of protein molecular mass standards is indicated.
(bl Figure 5. Detection of immunoreactive IFN-y in trophectoderm of a day 16 porcine conceptus. (a) Indirect immunofluorescence staining of a trophoblast section using rabbit anti-PorLFN-y serum ( x 340). The immunoreactive product was located only in the apical cytoplasm of trophectoderm cells and absent from the endodcrm layer (arrowed). (b) Similar analysis performed on a section of the cephalic region of the embryo ( x 85). (c) Control experiment performed on a section adjacent to (a) and treated with the rabbit preimmune control serum ( x 340). Similar results were obtained with day 15 and 17 conceptuses.
Eur. J. Immunol. 1990. 20: 2485-2490
Spontaneous IFN-y expression by porcine trophectoderm early in gestation
(Fig. 5) were clearly not labeled, indicating that IFN-y was expressed only by the trophectoderm. Maternal endometrium was also negative (not shown).
4 Discussion Our results clearly show that two distinct types of antiviral activity are spontaneously expressed at the maternal-fetal interface during early pregnancy in pigs. The first type appeared to be antigenically related to HuIFN-Le, which is known to contain mainly IFN-a subspecies and traces of other antiviral cytokines [30].Thus, although it remains to be precisely identified, this antiviral activity is probably IFN-a related, as also suggested by another recent report [ 151. However, further studies are required to determine whether this IFN-like activity is actually a conceptus product. The amount released in vitro by isolated conceptuses appeared to be lower than the amount found in vivo (Fig. 1) so the possibility that it was in fact a maternal product taken nonspecifically from uterine secretions by conceptus tissues and released in vitro cannot be totally excluded, although it seems unlikely. The second type is clearly a conceptus-derived IFN-y expressed by almost all the cells of the trophectoderm. This is the first report of an IFN-y synthesis by cells that do not belong to the hematopoietic lineage. The identity of porcine trophoblastic and immune I F N y was not proven by the present data, so it is possible that each protein is encoded by distinct genes. Moreover, our results show that two distinct IFN-y-related mRNA species are expressed by the conceptus.The in vitro translation experiment indicates that both mRNA probably encode the same preprotein, so their duality could be most simply explained by transcription from a single gene expressing alternatively spliced transcripts or bearing two functional polyadenylation signals. However, their expression from two genes encoding antigenically related preproteins of identical size cannot be excluded. Thus, the actual number of IFN-y genes in the pig species remains to be elucidated. If there is a single gene, as in all mammalian species studied so far [28], it will be interesting to study how this gene is differentially regulated in trophectoderm cells and lymphocytes. It has recently been shown that, during implantation, the trophoblast of ruminant conceptus secretes proteins with antiviral activity belonging to the class I1 subfamily of IFN-a [31-341.These IFN-a are supposed to play an active role in maternal recognition of pregnancy by prolonging the life-span of the corpus luteum on the ovary, since they have such an effect when they are locally administered in the uterus of cyclic ruminant females [35, 361. It is possible that the porcine IFN-a-related antiviral activity mentioned above represents the homologue of these ruminant proteins; however, maternal recognition of pregnancy is mediated in the pig by distinct mechanisms so their biological role during early pregnancy is probably different. Moreover, unlike in the pig, no IFN-y seems to be expressed by the ruminant conceptus as all the antiviral activity detected in uterine flushings or in conceptus culture can be neutralized by antisera directed against the purified conceptus-derived IFN-a (J. Martal, personal communication). Also, we have preliminary data indicating the absence of IFN-y-related mRNA in sheep conceptus during implantation (data not shown). Thus, spontaneous expres-
2489
sion of IFN-y by trophoblast early in gestation is probably not a general phenomenon in mammals and could be an original feature of the pig. This physiological IFN-y expression seems to be temporally regulated as no significant antiviral activity is secreted by conceptuses outside the day-12 to -20 period ([15, 161 and *). During this period, trophoblastic tissues represent the major part of the conceptus and establish intimate structural and functional relationships with the endometrium [19,37]. The fact that the porcine embryonic trophectoderm expresses an important immunoregulatory molecule suggests that the conceptus allograft plays an active role in early interactions with the maternal immune system before placental formation. Recent studies in rodents and humans indicate that paternal MHC antigens and other fetal alloantigens are expressed by the trophoblast or surrounding fetal tissues during pregnancy and elicit a maternal immune response [4]. There is much evidence to indicate that the deleterious effects of this immune response on the fetus are selectively impaired by local immunoregulatory events leading to local immune suppression and inhibition of fetal allograft rejection [4-61. Porcine trophoblastic IFN-y could act on the local maternal immune system to induce such specific mechanisms. The study of intrauterine immunosuppression has been recently initiated in species with epithelial placentation, such as the pig for which the presence of intrauterine suppressive cells has been reported in early pregnancy [38]. Conceptus-derived IFN-y could be involved directly or indirectly in the recruitment of these cells. Moreover, a specific recruitment of NK cells to the endometrial stroma and a significant reduction of intraepithelial lymphocytes in the endometrium were also recently reported during early pregnancy in the pig [39,40] and could be related to an IFN-y expression. Alternatively, one of the main biological properties of IFN-y is its ability to induce MHC class I and class I1 antigen expression [3]. No data are available about MHC antigen expression at the maternal-fetal interface during implantation in the pig. In the human species, the absence of classical polymorphic MHC class I and class I1 molecule expression on trophoblast has been reported during gestation [41,42]. This fact was often interpreted as functional since expression of paternal class I antigen would presumably risk maternal rejection of the fetus. However, several lines of evidence indicate that less polymorphic class I molecules are expressed by trophoblast cells in several species including man [43-451. It has been suggested that these molecules might have different functions from those encoded by polymorphic class I MHC loci and could have a role in enabling the fetus to escape recognition by the maternal immune system [46] or in inducing specific suppression of the maternal immune response [42]. Whether IFN-y is the physiological autocrine inducer of such MHC molecules on the porcine trophoblast will be an interesting question to solve. However, the possibility that this IFN-y induces classical MHC class I and class I1 antigens on trophoblast or maternal tissues must also be investigated. Spontaneous and temporally regulated expression of IFNy by the trophoblast in early pregnancy has an important
* La Bonnardi2re
et al., in press.
2490
E Lefkvre, E Martinat-BottC, M. Guillomot et al.
biological significance, although its occurrence in other mammalian species remains to be explored. It indicates that the early conceptus, though devoid of adult lymphoid organs [47], is an immunologically active allograft emitting specific signals, possibly to communicate with the maternal immune system. It also suggests that developmentally controlled expression of lymphokines by the trophoblast modulates the maternal immune response against the conceptus. We thank C. De Vaureix, R . Procureur and t?Despris for excellent technical assistance; Dr. C. Chany (INSERM, Paris) for the anti-human IFN-Le serum; Dr. W Fiers (Ghent, Belgium) for the human IFN-y cDNA; Dr. B. Delmas for discussions; and Drs. J. Jarvis and H . Laude for critical reading of the manuscript. Received June 5, 1990.
5 References 1 Vilcek, J., Gray, P.W., Rinderknecht, E. and Sevastopoulos, C. G., Lymphokines 198.5. 11: 1. 2 De Maeyer, E. and De Maeyer-Guignard, J., Interferons and Other Regulatory Cytokines, John Wiley and Sons, New York 1988, p. 144. 3 Rosa, F. M., Cochet, M. M. and Fellous, M., in Gresser, I. (Ed.), Interferon, vol. 7, Academic Press, London 1986, p. 47. 4 Hunziker, R. D. and Wegmann, T. G., CRC Crit. Rev. Immunol. 1986. 6: 245. 5 Chaouat, G., J. Reprod. Immunol. 1987. 10: 179. 6 Wegmann, T. G., Immunol. Lett. 1988. 17: 297. 7 Hunt, S. H., J. Reprod. Immunol. 1989. 10: 1. 8 Duc-Goiran, I?, Robert-Galliot, B., Lopez, J. and Chany, C., Proc. Natl. Acad. Sci. USA 1985. 82: 5010. 9 Chard, T., Craig, P. H., Menabawey, M. and Lee, C., Br. J. Obstet. Gynaecol. 1986. 93: 1145. 10 Fowler, A. K., Reed, C. D. and Giron, D. J., Nature 1980.286: 266. 11 Baker, D. J. and Nieder, G. L., J. Reprod. Fertil. 1990. 88: 307. 12 Toth, F. D., Juhl, C. N0rskov-Lauritsen, N., Mosborg Petersen, P. and Ebbesen, P., J. Reprod. Immunol. 1990. 17: 217. 13 Pontzer, C. H., Torres, B. A., Vallet. J. L., Bazer, F. W. and Johnson, H. M., Biochem. Biophys. Res. Commun. 1988.152: 801. 14 Betteridge, K. J., Derbyshire, J. B., Rorie, R.W., Scodvas, J. M. and Johnson,W. H., J. Reprod. Fertil. (abstract series) 1988. I : 21. 15 Cross, J. C. and Roberts, R. M., Biol. Reprod. 1989. 40: 1109. 16 Mirando, M. A., Harney, J. P., Beers, S., Pontzer, C. H.,Torres, B. A., Johnson, H. M. and Bazer, F.W., J. Reprod. Fertil. 1990. 88: 197. 17 Roberts, R. M., J. Interferon Res. 1989. 9: 373. 18 King, G. J., Atkinson, B. A. and Robertson, H. A., J. Reprod. Fertil. [Suppl.] 1982. 31: 17.
Eur. J. Immunol. 1990.20: 2485-2490
19 Dantzer,V., Anat. Embryol. 1985. 172: 281. 20 Menezo,Y., C.R. Acad. Sci. Paris [D] 1976. 282: 1967. 21 La Bonnardiere, C., Laude, H. and Berg, K., Ann. Inst. PasteurlVirol. 1986. 137E: 171. 22 Charley. B., McCullough, K. and Martinod, S., Vet. Immunol. Immunopathol. 1988. 19: 95. 23 Sambrook, J., Fritsch, E. F. and Maniatis, T., Molecular Cloning: a Laboratory Manual, 2nd Edn., Cold Spring Harbor Laboratory Press, New York 1989. 24 Devos, R., Cheroutre. H., Taya,Y., Degrave, W. ,Van Heuverswyn, H. and Fiers, W., Nucleic Acids Res. 1982. 10: 2487. 25 Laude, H., Chapsal, J. M., Gelfi, J., Labiau, S. and GrosClaude, J., J. Gen. Virol. 1986. 67: 119. 26 Guillomot, M., Michel, C., Gayc, P., Charlier, M.,Trojan, M. and Martal. J., Biol. Cell 1990, in press. 27 Langer. J. A. and Pestka, S., Pharmacol. Ther. 1985. 27: 371. 28 Weissman, C. and Weber, H., Prog. Nucleic Acid Res. Mol. Biol. 1986. 33: 251. 29 Tarentino, A. L., Gomez, C. M. and Plummer, T. H., Jr., Biochemistry 1985. 24: 4665. 30 De Maeyer, E. and De Maeyer-Guignard. J., Interferons and Other Regulatory Cytokines, John Wiley and Sons, New York 1988, p. 42.5. 31 Imakawa, K., Anthony, R. V.. Kazemi, M., Marotti, K. R., Polirzq, H. F. and Roberts, R. M., Nature 1987. 330: 377. 32 Charpigny, G., Reinaud, P.. Huet, J. C., Guillomot, M., Charlier, M., Pernollet, J. C. and Martal, J., FEBS Lett. 1988. 228: 12. 33 Charlier, M., Huc, D., Martal, J. and Gaye, P.. Gene 1989. 77: 341. 34 Imakawa, K.. Hansen, T. R., Malathy, P.-V., Anthony, R.V., Polites, H.V., Marotti, K. R. and Roberts, R. M., Endocrinology 1989. 3: 127. 35 Vallet, J. L., Bazer, F. W., Fliss, M. F.V. and Thatcher,W. W., J. Reprod. Fertil. 1988. 84: 493. 36 Thatcher, W. W., Hansen, P. J., Gross, T. S., Helmer, S. D., Plante, C. and Bazer, F.W., J. Reprod. Fertil. [Suppl.] 1989.37: 91. 37 Bazer, F. W. and Roberts, R. M., J. Exp. Zool. 1983. 228: 373. 38 Croy, B. A., Wood, W. and King, G. J., J. Immunol. 1987.139: 1088. 39 Croy, B. A., Waterfield, A., Wood, W. and King, G. J., Cell. Immunol. 1988. 115: 471. 40 King, G. J., J. Reprod. Immunol. 1988. 14: 41. 41 Redman, C. W. G.. McMichael, A. J., Stirrat, G. M., Sunderland, C. A. and Ting, A., Immunology 1984. 52: 457. 42 Bulmer. J. N. and Johnson, P. M., Placenta 1985. 6: 127. 43 Ellis, S. A., Palmer, M. S. and McMichael, A. J., J. Immunol. 1990. 144: 713. 44 Stern, P. L., Beresford, N., Friedman, C. I., Stevens, V. C., Risk, J. M. and Johnson, P. M., J. Immunol. 1987. 138: 1088. 45 MacPherson,T. A . , Hon-Nerng, H., Kunz, H.W. and Gil1,T. J., Transplantation 1986. 41: 392. 46 Head, J. R., Drake, B. L. and Zuckermann, F. A., A m . J. Reprod. Immunol. Microbiol. 1987. 15: 12. 47 Gill, T. J., 111 and Repetti, C. F., A m . J. Pathol. 1979. 95: 463.
