Dampening of IFN-γ-Inducible Gene Expression in Human Choriocarcinoma Cells Is Due to Phosphatase-Mediated Inhibition of the JAK/STAT-1 Pathway¹

Trophoblast cells (TBCs)³ form the blastocyst-derived component of the placenta and perform multiple functions that are critical for successful pregnancy. In the placentas of humans and rodents, the trophoblast layer forms a protective barrier surrounding the developing embryo/fetus and containing the only cells derived from the blastocyst that are directly exposed to maternal blood (1). Thus, TBCs play a significant role in preventing immune rejection reactions against the developing semiallogeneic conceptus by the maternal immune system, in part by producing a variety of soluble and membrane-bound immunosuppressive molecules (2, 3). Human TBCs also express a unique repertoire of MHC Ags on their cell surface (4). Extravillous TBCs express the MHC class Ia Ag HLA-C, and the nonpolymorphic MHC class Ib Ags HLA-E and HLA-G on their cell surface, whereas syncytiotrophoblast cells express only soluble HLA-G (1, 4). These MHC Ags are believed to regulate the activities of uterine NK cells (1, 4). Lastly, the expression of MHC class Ia (HLA-A and HLA-B) (5) and class II Ags (5, 6) is silenced in all human TBCs, both constitutively and in response to IFN-γ. The lack of these classical MHC Ags on human TBCs may be critical for preventing transplant rejection reactions against the semiallogeneic conceptus by the maternal immune system (3, 5).

The proinflammatory cytokine IFN-γ plays important roles in diverse cellular processes, including activating innate and adaptive immune responses against pathogens and tumors, inhibiting cell proliferation, and inducing apoptosis (7, 8). Activation of adaptive immune responses by IFN-γ is in part due to transcriptional induction of genes encoding MHC class I and class II Ags, invariant chain, HLA-DM/H2-DM, transporters associated with Ag processing (TAPs) and the immunoproteasome subunits LMP-2, LMP-7, and LMP-10 (7). Induction of apoptosis and cell-cycle arrest occurs through activation of caspase and p21 gene expression, respectively (7, 8). Interestingly, IFN-γ has been detected in the placentas of several mammals, including humans and mice (9, 10, 11, 12, 13, 14, 15, 16). In mice, placental IFN-γ is secreted primarily by uterine NK cells (17), and studies using mice deficient for IFN-γ, IFN-γR1, or NK cells demonstrated that this cytokine is essential for the pregnancy-induced remodeling of the uterine vasculature and proper formation and maintenance of the decidual layer of the placenta during pregnancy (17). Although IFN-γ is present in the placenta, neither human nor rodent TBCs express MHC class II genes, due to silencing of expression of the CIITA (18, 19, 20), the master regulator of constitutive and IFN-γ-inducible MHC class II gene transcription (21, 22). Moreover, primary term human cytotrophoblast cells (cTBCs) and human trophoblast-derived choriocarcinoma cells are resistant to apoptosis activated by IFN-γ alone (23, 24). These properties of TBCs may be essential for the maintenance of successful pregnancy. Despite these observations, several studies collectively suggest that TBCs have the capacity to respond to IFN-γ. Primary human TBCs and the choriocarcinoma cell lines Jar and JEG-3 express cell surface IFN-γ receptors (IFN-γR) (6, 15, 25). Furthermore, the expression of LMP-7, TAP-1, TAP-2, guanylate-binding protein (GBP), IFN regulatory factor 1 (IRF-1), and B7-H1 is up-regulated in both primary human TBCs and choriocarcinoma cell lines following exposure to IFN-γ (19, 20, 26, 27, 28, 29). Taken together, these studies suggest that responses of TBCs to IFN-γ may be selective. However, to date a comprehensive examination of IFN-γ signal transduction has not been reported in TBCs.

IFN-γ-mediated activation of Ag presentation gene expression, induction of apoptosis, and inhibition of cell proliferation are regulated by the JAK/STAT pathway (7, 8). Binding of IFN-γ to its cognate receptor results in activation of the receptor-associated kinases JAK-1 and JAK-2, which phosphorylate the intracellular domain of the IFN-γR1 (7, 8). Monomers of the transcription factor STAT-1, which are localized in the cytoplasm, interact with the phosphorylated IFN-γR1 and are subsequently phosphorylated on tyrosine residue 701 by the JAKs (8, 30). Tyrosine phosphorylation of STAT-1 leads to its homodimerization, translocation to the nucleus, and transcriptional activation of multiple different genes that contain a IFN-γ-activating sequence in their promoters, such as the gene encoding the transcription factor IRF-1 (7, 8). IRF-1 can subsequently activate transcription of caspase genes involved in initiating apoptosis, the p21 gene that inhibits cell growth, and/or the genes encoding MHC class Ia Ags, TAP-1, TAP-2, and the immunoproteasome subunits LMP-2, LMP-7 and LMP-10, all of which are required for effective immune responses to pathogens and tumors (7, 31). Furthermore, STAT-1 and IRF-1 cooperate with the ubiquitously expressed transacting factor upstream stimulatory factor 1 (USF-1) to activate transcription from the IFN-γ-inducible CIITA promoter IV (32, 33). CIITA subsequently activates transcription of the MHC class II genes (21, 22).

In this report, quantitative RT-PCR was used to demonstrate that the expression of several IFN-γ-responsive genes is reduced in the human choriocarcinoma lines Jar and JEG-3 relative to HeLa epithelial cells. The dampening of IFN-γ-inducible gene expression in human choriocarcinoma cells correlates with compromised tyrosine phosphorylation of JAK-2 and STAT-1, and reduced expression of IRF-1. Impaired phosphorylation of STAT-1 and reduced IRF-1 expression were also observed in primary human cTBCs. Furthermore, treatment of choriocarcinoma cells with the tyrosine phosphatase inhibitor pervanadate (PV) enhanced IFN-γ-inducible JAK and STAT-1 phosphorylation and IFN-γ-inducible gene expression. Taken together, these studies suggest that IFN-γ signaling is inhibited in human TBCs by tyrosine phosphatase(s) that target the JAKs. This phenomenon may have important implications for successful mammalian pregnancy, as well as choriocarcinoma tumor survival.

Jar and JEG-3 choriocarcinoma, HeLa cervical carcinoma, 2fTGH fibrosarcoma cells, and Jar/2fTGH and JEG-3/2fTGH stable hybrids were cultured as previously described (34). BG-9 primary human foreskin fibroblasts were cultured as described by Lin et al. (35). Human IFN-γ was purchased from PBL Biomedical Laboratories and used at concentrations ranging from 100 to 1000 U/ml. Sodium orthovanadate (S6508 450243), hydrogen peroxide (31642), and bovine liver catalase (C1345) were purchased from Sigma-Aldrich.

Activation of sodium orthovanadate was performed as described by Gordon (36). PV solution (1 mM) was generated as described by Huyer et al. (37). Excess H₂O₂ was removed by incubation for 5 min with 100 μg/ml bovine liver catalase. The PV solution was used on cells at a final concentration of 100 μM within 5 min of generation.

Cytotrophoblast cells were purified as previously described by Petroff et al. (27). In brief, ∼40 g of term villous placental tissue were finely minced and subjected to three 30-min stages of digestion in a solution of trypsin and DNase. The resulting cell suspensions were layered over a discontinuous 5–70% Percoll gradient (Sigma-Aldrich) and centrifuged. The cell layer located between the densities of 1.053 and 1.060 was collected and resuspended in culture medium (IMDM containing 10% FBS, 100 μg/ml streptomycin, and 100 U/ml penicillin), and further subjected to negative selection using anti-HLA class I Ab (clone W6/32; American Type Culture Collection) coupled to magnetic microbeads (Miltenyi Biotec). The purity of the cTBCs used in these studies ranged from 96 to 99% based on immunostaining with cytokeratin-7 Abs (clone OV-TL; DakoCytomation).

RNA was isolated using TRIzol (Invitrogen Life Technologies) as specified by the manufacturer. Reverse transcriptase reactions were performed on 2 μg of total RNA using Superscript II RT (Invitrogen Life Technologies) as described previously (19, 34). The primers and annealing temperatures used for RT-PCR analysis were described previously and are as follows: CIITA (38), GBP (39), IRF-1 (18), LMP-2, LMP-7 (40), and GAPDH (26). Quantitative RT-PCR was performed in duplicate for each sample on 1/100 of the total cDNA using an iCycler (Bio-Rad) instrument and SYBR Green master mix (Bio-Rad) as directed by the manufacturer. Standard curves were generated using plasmids containing the respective cDNAs for each gene examined. Expression of IFN-γ-inducible genes was normalized to GAPDH gene expression, and data are represented as the ratio of the relative cDNA copy number of each of these genes per copy number of GAPDH. The results were statistically analyzed using Student’s t test.

HeLa, Jar, and JEG-3 cells (∼10⁶) were harvested by trypsinization and stained for 30 min at 4°C with PE-conjugated Abs to human IFN-γR1 (GIR-94; BioLegend) or PE-conjugated human IFN-γR2 (2HUB-159; BioLegend) at concentrations suggested by the manufacturer. Background staining was established using isotype control Abs sc-2868 for IFN-γR1 and sc-2875 for IFN-γR2. The cells were subsequently washed with 2% BSA/0.1% sodium azide in PBS and fixed with 2% paraformaldehyde diluted in PBS. Cell surface fluorescence was analyzed with a FACS-440 instrument (BD Biosciences) using CellQuest and WinMDI software. The maximum fluorescence intensity values were obtained using WinList version 5.0 (Verity Software House).

Whole cell and nuclear extracts were prepared as previously described (41) using radioimmunoprecipitation assay buffer (1× PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 150 mM NaCl, and 0.1% SDS) or lysis buffer composed of: 20 mM HEPES (pH 7.9); 420 mM NaCl; 1.5 mM MgCl₂; 0.2 mM EDTA; 25% glycerol (v/v); 10 mM sodium molybdate; 1.0 mM DTT, 50 μg/ml pepstatin, 25 μg/ml aprotinin, 25 μg/ml leupeptin, and 1.0 mM PMSF. Protease inhibitors and DTT were added to the lysis buffer just before use. Radioimmunoprecipitation buffer was used in the analyses of primary cytotrophoblast and fibroblast cells. Abs against IRF-1 (sc-497), IRF-2 (sc-498), OCT-1 (sc-232), STAT-1 (sc-417), phosphotyrosine 701-STAT-1 (sc-7988), phosphoserine 727-STAT-1 (sc-16570-R), and USF-1 (sc-229) were purchased from Santa Cruz Biotechnology. Protein extracts (15–30 μg) were fractionated on 7–8% polyacrylamide gels and transferred to nitrocellulose membranes (Schleicher & Schuell Microscience) using a semiwet transfer apparatus (Bio-Rad). For detection of STAT-1, IRF-1, IRF-2, USF-1, OCT-1, and heat shock cognate 70 (HSC70), membranes were blocked overnight in 5% milk/PBS-0.075% Tween 20 and incubated with primary Ab for 1 h at room temperature using the following Ab concentrations: IRF-1 (100 ng/ml for cell lines, 250 ng/ml for cTBCs), IRF-2 (1 μg/ml), STAT-1 (200 ng/ml), USF-1 (333 ng/ml); OCT-1 (250 ng/ml), and HSC70 (1/20,000 monoclonal 3a3 ascites) (42). The membranes were washed four times with PBS-0.075% Tween 20 at room temperature, incubated with HRP-conjugated secondary Ab (100 ng/ml; Promega) for 45 min, and washed five times with PBS-0.075% Tween 20. For examination of phosphorylated proteins, membranes were blocked with 3% BSA/TBS/0.1% Tween 20, incubated with primary phospho-specific Abs (phosphotyrosine 701-STAT-1 (200 ng/ml) or phosphoserine 727-STAT-1 (200 ng/ml)), washed with TBS/0.075% TBS, and incubated with secondary Ab as described above. Signals were detected using the SuperSignal West Pico Chemiluminescent Substrate (Pierce) as described by the manufacturer and subsequent exposure to Kodak Scientific Imaging film (Kodak). Relative levels of transcription factors were determined by dilution analysis of the cell extracts, followed by Western blot analysis. Signals were quantitated using a Molecular Dynamics Computing Densitometer model 300S.

HeLa, Jar, and JEG-3 cells (∼1.2 × 10⁷) were treated with 1000 U/ml IFN-γ for various times as indicated. Whole cell extracts (WCE) were isolated using lysis buffer composed of 20 mM Tris (pH 8.0), 150 mM NaCl, 1% Nonidet P-40 (v/v), 50 mM NaF, 1 mM sodium orthovanadate, 50 μg/ml pepstatin, 50 μg/ml aprotinin, 50 μg/ml leupeptin, and 1 mM PMSF. Aliquots (50 μl) were removed from each sample before immunoprecipitation to use in Western blot analysis for assessment of STAT-1 tyrosine phosphorylation. The extracts were precleared by overnight incubation at 4°C with 30 μl of protein A-agarose beads (Upstate Biotechnology), followed by immunoprecipitation with an isotype-matched Ab for 1 h. Immunoprecipitations were performed by incubating extracts for 4–16 h at 4°C with 1 μg of Abs to JAK-1 or JAK-2 followed by 3 h at 4°C with 30 μl of protein A-agarose beads. The immunoprecipitates were washed three times with lysis buffer, eluted from the beads, and subjected to Western blot analysis using Abs to JAK-1 (sc-277), JAK-2 (sc-294), phosphotyrosine 4G10 (100 ng/ml; Upstate Biotechnology), or phospho-JAK-2-Y¹⁰⁰⁷/Y¹⁰⁰⁸ (sc-16566-R; Santa Cruz Biotechnology and 44-426G, 200 ng/ml; BioSource International) as described above.

Previous studies demonstrated that the expression of a subset of IFN-γ-inducible genes is enhanced in TBCs in response to IFN-γ, but these studies did not precisely distinguish whether there were quantitative differences in IFN-γ induction of these genes in TBCs compared with other cell types, such as fibroblasts and epithelial cells. To address this possibility, RNA was isolated from Jar and JEG-3 choriocarcinoma cells and HeLa epithelial cells cultured for 24 h in 0, 500, or 1000 U/ml IFN-γ and subjected to quantitative RT-PCR using primers for IRF-1, GBP, LMP-2 and LMP-7. Relative levels of expression of these genes were normalized to GAPDH mRNA expression. Fig. 1 demonstrates that although basal expression of IRF-1 and GBP mRNAs is comparable in HeLa, Jar, and JEG-3 cells, IFN-γ-induced expression of these genes is clearly significantly reduced in Jar and JEG-3 cells when compared with HeLa cells. Specifically, IRF-1 mRNA expression is 4.8-fold lower (500 U/ml IFN-γ) to 6.7-fold lower (1000 U/ml IFN-γ) in Jar vs HeLa cells and 4-fold lower in JEG-3 cells. Similarly, GBP mRNA expression is reduced ∼33-fold and 8-fold in IFN-γ-treated Jar and JEG-3 cells vs HeLa cells, respectively. Moreover, LMP-2 mRNA was expressed constitutively in all three cell lines, but following exposure to IFN-γ was also lower in Jar (∼30-fold reduction) and JEG-3 (∼6-fold reduction) cells compared with HeLa cells (Fig. 1). Lastly, basal expression of LMP-7 mRNA was substantially higher in HeLa cells compared with Jar and JEG-3, and was enhanced following exposure to IFN-γ. LMP-7 mRNA expression was also up-regulated by IFN-γ in Jar and JEG-3 cells, but the absolute levels remained substantially lower than those in HeLa cells (Fig. 1). GAPDH mRNA was expressed at comparable levels in all three cell lines in the absence or presence of IFN-γ, demonstrating that the differences in gene expression were not due to differences in RNA quantitation or integrity (Fig. 1). Similar trends of gene expression were observed in two experiments using 100 U/ml IFN-γ (data not shown). Taken together, these results indicate that although choriocarcinoma cells clearly respond to IFN-γ by up-regulating transcription of a number of genes, both the absolute levels of expression and the fold induction are significantly lower compared with HeLa epithelial cells.

One explanation for the relative dampening of IFN-γ-inducible gene expression in Jar and JEG-3 cells compared with HeLa cells is that choriocarcinoma cells express reduced numbers of IFN-γR. To address this possibility, Jar, JEG-3, and HeLa cells were subjected to flow cytometric analysis using Abs specific to the IFN-γR1 and IFN-γR2 chains, respectively. As shown in Fig. 2, cell surface expression of both the IFN-γR1 and IFN-γR2 chains was detected on Jar and JEG-3 cells, at levels comparable to HeLa cells. Therefore, reduced numbers of cell surface IFN-γR on choriocarcinoma cells do not account for the dampening of IFN-γ-inducible gene expression.

Binding of IFN-γ to its receptor results in activation of JAK-1 and JAK-2 through the phosphorylation of tyrosine 1022/1023 on JAK-1 and tyrosine 1007/1008 on JAK-2, respectively (43, 44). To determine whether this step in the JAK/STAT-1 pathway is intact in choriocarcinoma cells, expression and phosphorylation of the JAKs were assessed. WCE were prepared from equal numbers of HeLa, Jar, and JEG-3 cells cultured for 0, 0.25, 1, and 3 h in 1000 U/ml IFN-γ and subjected to immunoprecipitation using Abs to either JAK-1, JAK-2, or isotype-matched Abs. Immunoprecipitated products were subsequently subjected to Western blot analysis with Abs to JAK-1, JAK-2, P-Y¹⁰⁰⁷/Y¹⁰⁰⁸-JAK-2, or phosphotyrosine. As shown in Fig. 3,A, JAK-1 is expressed in Jar and JEG-3 cells, but expression is 2.4-fold lower relative to HeLa cells, whereas JAK-2 expression is comparable among the three lines. Tyrosine phosphorylation of JAK-1 was detected within 0.25 h and sustained for up to 3 h after exposure to IFN-γ in all three cell lines (Fig. 3,A). However, although we observed an average 10.6-fold increase of phosphorylated JAK-1 in HeLa cells at all time points following IFN-γ treatment, only an average 2.5-fold increase was detected in both Jar and JEG-3 cells exposed to IFN-γ at the same time points (Fig. 3,B). Phosphorylated JAK-1 was also observed in all three lines after 6-h IFN-γ treatments (data not shown). Importantly, the induction of JAK-2 phosphorylation was significantly lower in Jar and JEG-3 cells (average, 1.3- and 1.5-fold increases, respectively) vs HeLa cells (average, 10.9-fold increase) at all time points following exposure to IFN-γ (Fig. 3,B), whether Abs to phosphotyrosine (Fig. 3 A) or P-Y¹⁰⁰⁷/Y¹⁰⁰⁸ JAK-2 (data not shown) were used in Western blot analysis. Immunoprecipitations with isotype-matched Abs demonstrated the specificity of the JAK-1 and JAK-2 Abs (data not shown). These studies suggest that phosphorylation and therefore activation of the JAKs, particularly JAK-2, is compromised in human choriocarcinoma cells.

Transcriptional activation of genes regulated by the JAK/STAT-1 pathway requires phosphorylation of tyrosine residue 701 on STAT-1 by the JAKs (8, 30). In addition, phosphorylation of serine 727 on STAT-1α plays a role in maximizing the transcriptional capacity of STAT-1 through interactions with histone acetyltransferases (45). The reduced levels of JAK phosphorylation in Jar and JEG-3 cells exposed to IFN-γ prompted us to examine the expression and phosphorylation status of STAT-1 in these cells. Because activation of STAT-1 in response to IFN-γ occurs very rapidly, and in many cell types is transient, particularly at low concentrations of IFN-γ (30), a kinetic analysis of STAT-1 tyrosine phosphorylation and IRF-1 protein expression was performed in IFN-γ-treated Jar and JEG-3 cells. WCE were isolated from HeLa, Jar, and JEG-3 cells treated with IFN-γ for various times from 1 to 24 h and subjected to Western blot analyses using Abs to STAT-1, tyrosine-phosphorylated STAT-1 (P-Y⁷⁰¹-STAT-1), and IRF-1. Basal STAT-1 expression was comparable among the three cell lines (Fig. 4). Over the 24-h time course, total STAT-1 levels increased ∼4–5-fold in HeLa cells treated with IFN-γ as previously described (46). However, only a ∼1.5- to 2-fold increase of total STAT-1 was observed in IFN-γ-treated Jar and JEG-3 cells. Importantly, high levels of P-Y⁷⁰¹-STAT-1 were detected within 1 h in IFN-γ-treated HeLa cells, and the levels were sustained through 24 h (Fig. 4). In contrast, substantially reduced levels of P-Y⁷⁰¹-STAT-1 were detected in Jar and JEG-3 cells treated with IFN-γ for 1 h (at least ∼7-fold lower compared with HeLa cells), and they decreased further by 3 h and remained low thereafter. P-Ser⁷²⁷-STAT-1 was comparable, albeit slightly reduced, in IFN-γ-treated Jar and JEG-3 cells relative to HeLa cells (data not shown).

Accumulation of IRF-1 protein was detectable within 1 h in HeLa cells, peaked at 3 h (∼10-fold induction relative to untreated HeLa), and remained at maximal levels through 24 h (Fig. 4). Conversely, IRF-1 protein reached peak levels by 3 h in IFN-γ-treated Jar and JEG-3 cells, but was markedly decreased within 6 h (∼7-fold lower than HeLa cells), and remained low throughout the remainder of the time course. The maximal levels of IRF-1 protein detected in Jar and JEG-3 cells at 3 h after initiation of IFN-γ treatment were still ∼2- to 4-fold lower compared with HeLa cells at the same time point. These results are consistent with the reduced levels of IRF-1 mRNA detected in IFN-γ-treated Jar and JEG-3 cells vs HeLa cells (Fig. 1). Lastly, both the basal and IFN-γ-induced expression of IRF-2 was dramatically lower in Jar or JEG-3 cells compared with HeLa cells at all time points (Fig. 4). The levels of USF-1 and HSC70 were comparable at all time points in Jar and JEG-3 cells vs HeLa cells, indicating that the differences in P-Y⁷⁰¹-STAT-1, STAT-1, IRF-1, and IRF-2 expression were specific (Fig. 4). Similar results for P-Y⁷⁰¹-STAT-1, IRF-1, and IRF-2 were obtained when cells were treated with 100 U/ml IFN-γ (data not shown). Collectively, these studies indicate that the magnitude and duration of STAT-1 Y⁷⁰¹ phosphorylation and IRF-1 expression are significantly reduced in IFN-γ-treated Jar and JEG-3 cells compared with HeLa cells.

The transient phosphorylation of STAT-1 Y⁷⁰¹ and relatively low levels of IRF-1 expression in Jar and JEG-3 cells exposed to IFN-γ could be a reflection of: 1) the normal trophoblastic phenotype, 2) the fact that the cells are transformed, or 3) adaptation to tissue culture. To distinguish these possibilities, Western blot analysis was used to examine STAT-1 phosphorylation and IRF-1 expression in human villous cTBCs freshly isolated from term placentas. As observed in Jar and JEG-3 cells, basal STAT-1 expression was clearly detectable in human cTBCs, but was only very weakly up-regulated following a 24-h IFN-γ treatment (Fig. 5,A). Importantly, STAT-1 Y⁷⁰¹ phosphorylation was very low in primary cTBCs treated for 3 h with IFN-γ and decreased to basal levels by 24 h (Fig. 5,A). Moreover, the levels of IRF-1 protein detected in IFN-γ-treated human villous cTBCs were significantly lower than in HeLa cells (Fig. 5,A). IRF-2 expression was not detected in WCE from two preparations of primary cTBCs (data not shown). Comparable results were observed using 100 U/ml IFN-γ (data not shown). An alternative explanation for our results is that the sustained STAT-1 Y⁷⁰¹ phosphorylation observed in HeLa cells is unique to this cell line and not reflective of a normal cell type. However, high level, sustained STAT-1 phosphorylation was observed in BG-9 primary human foreskin fibroblasts exposed to IFN-γ, and this correlated with high levels of IRF-1 expression (Fig. 5,B). Similar results were observed with 2fTGH fibrosarcoma cells (Fig. 5,B and see Fig. 6). Taken together, these studies strongly suggest that transient phosphorylation of STAT-1 and dampening of IRF-1 expression in response to IFN-γ are normal trophoblast phenotypes and not due to transformation or adaptation to tissue culture.

We previously generated stable hybrids between 2fTGH fibrosarcoma cells (which express CIITA in response to IFN-γ) and Jar or JEG-3 cells and demonstrated that silencing of IFN-γ-inducible CIITA gene transcription in human choriocarcinoma cells is genetically recessive (34). To determine whether the ability of the stable hybrids to express CIITA in response to IFN-γ correlates with sustained tyrosine phosphorylation of STAT-1, Western blot analysis was performed on WCE prepared from 2fTGH clone 3, Jar clone 3, and Jar/2fTGH stable hybrids grown for 24 h in the absence or presence of IFN-γ. As expected, tyrosine 701 phosphorylation of STAT-1 was observed in the IFN-γ-treated 2fTGH clone 3, but not Jar clone 3 (Fig. 6,A). Sustained STAT-1 Y⁷⁰¹ phosphorylation was also observed in all of the stable Jar/2fTGH hybrids (Fig. 6,A). In addition, IFN-γ-inducible IRF-1 expression in the stable hybrids was similar to that observed in 2fTGH clone 3 and elevated relative to Jar clone 3 (Fig. 6,A). Similar results were observed for both basal and IFN-γ-induced expression of IRF-2 in the Jar/2fTGH hybrids (Fig. 6,A). The differences in STAT-1 tyrosine phosphorylation and IRF-1/IRF-2 expression between the stable hybrids and Jar clone 3 were not due to variations in extract integrity or quantitation, because OCT-1, HSC70 (Fig. 6,A), and USF-1 (data not shown) expressions were all comparable. Lastly, quantitative RT-PCR analyses demonstrated that, following exposure to IFN-γ, expression of the IRF-1, GBP, and LMP-7 mRNAs in the stable hybrids was similar to 2fTGH clone 3 and significantly higher than Jar clone 3 (Fig. 6 B). Identical results were obtained in the analyses of stable hybrids between JEG-3 and 2fTGH cells (data not shown). Collectively, these studies suggest that transient phosphorylation of STAT-1 and the general dampening of IFN-γ-inducible gene expression is recessive in human choriocarcinoma cells.

Our collective results suggest that dampening of IFN-γ-inducible gene expression in human choriocarcinoma cells is due, at least in part, to compromised JAK tyrosine phosphorylation and the consequential reduction in the level of activated STAT-1. Previous studies show that JAK activity is negatively regulated in various cell types by at least two distinct mechanisms: 1) dephosphorylation by protein tyrosine phosphatases (PTPs) (47, 48) or 2) inhibition of JAK phosphorylation and catalytic activity by suppressors of cytokine signaling 1 (SOCS-1) (48). To determine whether PTPs play a role in inhibiting JAK phosphorylation in IFN-γ-treated choriocarcinoma cells, Jar cells were treated for 0, 0.25, 1, and 3 h with IFN-γ alone and in combination with the tyrosine phosphatase inhibitor PV. Tyrosine phosphorylation of JAK-1 and JAK-2 was assessed by immunoprecipitation/Western blot analyses as described for Fig. 3. In IFN-γ-treated HeLa cells, we observed significant JAK-1 and JAK-2 phosphorylation at all time points examined (Fig. 7,A). As previously observed, phosphorylation of JAK-1 and JAK-2 was very low to undetectable at all time points in Jar cells exposed to IFN-γ alone (Fig. 7,A). In contrast to previous studies on HeLa cells (49), treatment of Jar cells with PV alone had only minor effects on JAK-1 phosphorylation, but no apparent effect on JAK-2 (Fig. 7,A). However, simultaneous treatment of Jar cells with IFN-γ and PV resulted in robust phosphorylation of JAK-1 and JAK-2 (average, 7.2- and 8.7-fold increases, respectively, relative to treatment with IFN-γ alone) at all time points examined (Fig. 7,A). The enhancement of phosphorylated JAKs in Jar cells exposed to IFN-γ/PV was not due to increases in total JAK-1 or JAK-2 protein expression (Fig. 7,A). Jar cells were subsequently treated with IFN-γ and/or PV for 0, 3, and 16 h and harvested for isolation of WCE and RNA. Western blot analysis demonstrated that the observed increase in phosphorylated JAKs correlated with a significant enhancement of P-Y⁷⁰¹-STAT-1 and increased IRF-1 expression in Jar cells (Fig. 7,B). IFN-γ/PV treatment had little effect on either USF-1 or HSC70 expression, demonstrating that the effect of the treatment was selective (Fig. 7,B). Consistent with the enhancement of IRF-1 protein, quantitative RT-PCR demonstrated that the combined IFN-γ/PV treatment of Jar cells significantly increased the expression of IRF-1 mRNA (2.8- and 7.3-fold enhancements, respectively, at 3 and 16 h compared with IFN-γ treatment alone; Fig. 7,C). Similar enhancements of GBP (25.6-fold) and LMP-7 (11.8-fold) mRNA expression were observed in Jar cells exposed for 16 h to IFN-γ/PV relative to IFN-γ alone (Fig. 7,C). Prolonged exposure of Jar cells to PV alone had little to no effect on phosphorylation of STAT-1 Y⁷⁰¹ (Fig. 7,B) or IFN-γ-inducible gene expression (Fig. 7 C). Lastly, combined IFN-γ/PV treatment had no apparent effect on CIITA or MHC class I mRNA expression (data not shown). Taken together, these results strongly suggest that IFN-γ signaling is inhibited in human choriocarcinoma cells by protein tyrosine phosphatase(s) that prevent productive phosphorylation and, thus, activation of the JAKs.

Despite the fact that the proinflammatory cytokine IFN-γ, which plays critical roles in regulating immune responses, apoptosis and cell proliferation, is produced in the human placenta during normal pregnancy, a comprehensive and quantitative analysis of the IFN-γ signaling pathway in human TBCs has not been reported. In this study, we demonstrate that the expression of multiple IFN-γ-inducible genes is significantly reduced in human choriocarcinoma cells relative to epithelial or fibroblast cells. These results suggest that there is a general dampening of IFN-γ signal transduction in choriocarcinoma cells. Consistent with this proposal, the IFN-γR, JAK-1, and JAK-2 are present in Jar and JEG-3 cells, but JAK phosphorylation, particularly JAK-2, is inhibited due to the activity of PTPs. Furthermore, a corresponding inhibition of STAT-1 Y⁷⁰¹ phosphorylation and reduced IRF-1 expression are also observed in Jar and JEG-3 cells exposed to IFN-γ, as well as primary human cTBCs isolated from term placentas.

The regulation of cellular responses to IFN-γ is very complex and is mediated by the equilibrium between the activities of the JAKs and STAT-1, and a number of negative regulatory molecules, which include SOCS-1, protein inhibitor of activated STAT (PIAS), and PTPs (48). SOCS-1 down-regulates IFN-γ responses by interacting with JAK-2 and either inhibiting kinase activity and/or promoting JAK-2 degradation (48). PIAS1 directly blocks IFN-γ-inducible transcription by inhibiting STAT-1 DNA-binding activity, while PIASy inhibits STAT-1 transactivation by a mechanism that has not yet been elucidated (48). PTPs fine tune or attenuate responses to a wide array of cytokines and growth factors, including IFN-γ, by dephosphorylating JAK and/or STAT family members (48, 50). However, rather than merely fine tuning, PTPs appear to play a key role in suppressing the initiation of IFN-γ signaling in human choriocarcinoma cells, based on our observations that: 1) phosphorylation of JAK-2 and STAT-1 is compromised in IFN-γ-treated Jar and JEG-3 cells, even at times as early as 0.25 h; 2) there is a substantial dampening of IFN-γ-inducible gene expression in the choriocarcinoma cell lines relative to epithelial or fibroblast cells, even at high doses of IFN-γ; and 3) simultaneous treatment of Jar and JEG-3 cells with IFN-γ and the PTP inhibitor PV results in sustained, high-level phosphorylation of the JAKs and STAT-1 and a significant enhancement of expression of multiple IFN-γ-inducible genes.

Although our data strongly support the concept that PTPs play a central role in inhibiting JAK phosphorylation in IFN-γ-treated human choriocarcinoma cells, the precise molecular basis for the PTP-mediated dampening of IFN-γ responses in Jar and JEG-3 cells vs HeLa and 2fTGH cells is currently unclear. The observation that JAK-1 and JAK-2 are only weakly phosphorylated even at early time points (0.25 h) of IFN-γ treatment in choriocarcinoma cells suggests that the PTP(s) is constitutively active. Furthermore, the fact that high levels of STAT-1 phosphorylation and IFN-γ-inducible gene expression are observed in stable hybrids of 2fTGH fibrosarcoma and choriocarcinoma cells suggests that PTP-mediated dampening of IFN-γ responses is recessive in TBCs. One mechanism to explain these results is that TBCs express a novel and/or trophoblast-specific tyrosine phosphatase(s), the expression of which is extinguished in the stable hybrids. Alternatively, a ubiquitously expressed tyrosine phosphatase(s) may control JAK phosphorylation in TBCs, but the activity of the PTP(s) may be differentially regulated in TBCs vs HeLa and 2fTGH cells. Several PTPs have previously been shown to attenuate IFN-γ signaling by dephosphorylating JAK-1 and/or JAK-2, and these include CD45, T cell protein tyrosine phosphatase (TcPTP), Src homology region 2 domain-containing phosphatase (SHP) 1, SHP-2, and PTP-1B (48). CD45 and SHP-1 are primarily expressed in cells of hemopoietic origin and are therefore unlikely to inhibit IFN-γ responses in human TBCs. In contrast, TcPTP, SHP-2, and PTP-1B may potentially function in this capacity in TBCs, for they are all ubiquitously expressed, and have all been detected in the placenta (51).

Interestingly, glucose levels, growth factors, and environmental stress modulate the intracellular localization of SHP-2 and TcPTP (52, 53, 54), and therefore the specific target proteins subject to control by these PTPs. Thus, differences in the intracellular localization of a ubiquitously expressed PTP in Jar and JEG-3 cells vs HeLa cells, 2fTGH cells, and Jar/2fTGH stable hybrids may account for the differential effects of PTP activity on JAK phosphorylation in these cells. Alternatively, the enzymatic activity of a ubiquitous PTP may be differentially regulated in TBCs vs epithelial or fibroblast cells. Signaling through the BCR and in response to multiple cytokines and growth factors is attenuated by PTPs that are constitutively active (55, 56, 57). In these cases, binding of ligand to receptor results in the generation of the reactive oxygen species (ROS) hydrogen peroxide, which transiently inactivates PTP activity by reversible oxidation of an invariant cysteine residue in the PTP catalytic site (58). This results in amplification of the initial kinase activation and generation of a robust and sustained signal (55, 56, 57). Importantly, the concentration of the hydrogen peroxide produced appears to regulate the magnitude and duration of the corresponding signal (57). Furthermore, the receptors, kinases, and PTPs are in close physical proximity, and this appears to control the specificity of the cellular response to the receptor signal (57). Although a role for ROS-mediated inhibition of PTP activity has not been directly demonstrated in IFN-γ signaling, two pieces of evidence strongly support this possibility: 1) the catalytic sites of PTPs involved in regulating cytokine signaling are highly conserved (58) and 2) PTPs previously shown to down-regulate IFN-γ responses in some cell types are inactivated by ROS in response to other stimuli (56, 57). Thus, in response to IFN-γ, ubiquitously expressed PTP(s) may be inhibited by a ROS-mediated mechanism in cells that display robust and sustained phosphorylation of JAKs and STAT-1, such as HeLa, 2fTGH, and Jar/2fTGH stable hybrids. In contrast, robust activation of the JAKs may not occur in IFN-γ-treated human TBCs due to insufficient inactivation of the negative regulatory PTP(s) by ROS. Future studies will address the mechanisms involved in regulating PTP activity in Jar and JEG-3 cells.

The demonstration that STAT-1 Y⁷⁰¹ phosphorylation and IRF-1 expression are significantly reduced in both magnitude and duration in human TBCs provides a straightforward explanation for the dampening of IFN-γ-inducible gene expression in these cells. Specifically, there are limiting amounts of the requisite transcription factors available to activate IFN-γ-inducible gene transcription. This interpretation is consistent with the fact that PV significantly enhanced both IFN-γ-induced STAT-1 tyrosine phosphorylation and transcription of the IRF-1, GBP, and the LMP genes. The limiting amounts of P-Y⁷⁰¹ STAT-1 and IRF-1 are also very likely to contribute to the inability of human TBCs to express the CIITA gene, based on previous studies showing that transient phosphorylation of STAT-1 is not sufficient to activate CIITA transcription in IFN-γ-treated human retinoblastoma cells (59). However, based on the fact that CIITA expression was not detected in either Jar or JEG-3 cells treated simultaneously with IFN-γ and PV, additional levels of control are also clearly involved in silencing CIITA transcription in human TBCs. Thus, CIITA transcription is likely to be silenced in human TBCs by multiple overlapping mechanisms, which may include repressive chromatin structure at the CIITA promoter IV.

In addition to controlling immune responses by transcriptional activation of the MHC and LMP genes, IRF-1 also mediates IFN-γ-induced activation of apoptosis and inhibition of cell proliferation through up-regulation of caspase and p21 expression, respectively (7, 8). Both primary human TBCs and choriocarcinoma cell lines are resistant to IFN-γ-mediated apoptosis (23, 24). However, pretreatment of Jar or JEG-3 cells with IFN-γ significantly enhanced apoptosis induced by the chemotherapeutic drug etoposide (23). Importantly, sequential treatment of Jar cells with IFN-γ and etoposide resulted in an additive increase of IRF-1 expression relative to treatment with either agent alone, and this increase in IRF-1 was required for enhanced apoptosis (23). Thus, PTP-mediated dampening of IFN-γ-inducible signaling may play a role in preventing IFN-γ-mediated apoptosis in TBCs by inhibiting maximal accumulation of IRF-1.

The observation that IFN-γ is present in the placentas of humans and mice (9, 10, 11, 12, 13, 14, 15, 16) suggests that TBCs may be exposed to this proinflammatory cytokine during normal pregnancy. In fact, recent studies suggest that normal pregnancy is a state of controlled inflammation (60). Sargent et al. (60) have hypothesized that excess production of IFN-γ plays a pivotal role in pathologies of pregnancy such as preeclampsia. Thus, rigorous control of IFN-γ production may be critical for normal pregnancy. Our studies showing that the levels of activated STAT-1 and IRF-1 are significantly reduced in both human term cTBCs and choriocarcinoma cells exposed to IFN-γ suggests that this is a phenotype of normal human TBCs. Based on our collective results, we propose that the dampening of IFN-γ signal transduction in human TBCs by protein tyrosine phosphatase(s) contributes to successful pregnancy by inhibiting responses to IFN-γ that could result in placental damage. Furthermore, stringent regulation of the PTP(s) that control trophoblastic responses to IFN-γ may be an important adaptation to the inflammatory microenvironment at the implantation site of the conceptus. Lastly, dampening of IFN-γ-responsiveness in choriocarcinoma cells may contribute to tumor survival for the same reasons. Future studies will address these hypotheses.

We thank Dr. Kailash Chadha for providing BG-9 primary human foreskin fibroblast cells. We are grateful to Drs. Willis Li, Edith Lord, and Richard K. Miller, and Kelly Cycon for critical reading of this manuscript.

The authors have no financial conflict of interest.