Differentiation-induced Posttranscriptional Control of B7-H1 in Human Trophoblast Cells

Tissue collection and processing

Normal term placentas were collected following Caesarian sections from the University of Kansas Hospital in accordance with protocols approved by the institutional Human Subjects Committee, and tissue was processed within 30 minutes of delivery. For most cell culture studies, 30-40g villous tissue was dissected, rinsed, and subjected to enzymatic digestion as described previously [17]. For studies involving analysis of monosomal and polysomal RNA, 80-120g villous tissue was dissected and processed. After mincing, placental villi were dissociated in three 30-minute stages of a digestion solution containing trypsin and DNase. The resulting cell suspension was layered over a 5-70% stepwise Percoll (Sigma) gradient and centrifuged. The trophoblast fraction was subjected to immunopurification by negative selection over magnetic bead columns coated with anti-HLA class I antibody (W6/32; BioExpress, West Lebanon, NH). Cells were routinely more than 95% pure as assessed by immunocytochemical staining for cytokeratin-7 (DAKO Corp., Carpinteria, CA).

Assessment of cytotrophoblast fusion

Trophoblast fusion was assessed similarly to previously described procedures [18]. Purified cells were seeded into 4-well glass chamber slides (Nalge-Nunc) at a density of 750 × 10⁶ cells per well. Control and EGF-treated cells were cultured for 48 hours, at which time they were stained for desmplakin to evaluate cell fusion. Trophoblast cells were rinsed in PBS, and nonspecific binding sites were then blocked for 1 hour in a solution of 50% SuperBlock buffer (Peirce, Rockford, IL), 2% donkey serum and 0.02% triton X-100. Anti-desmosomal protein (clone ZK-31, Sigma, St. Louis, MO) was added to the cells at a dilution of 1:400, and slides were incubated overnight at 4°C. Lack of binding of IgG1 control antibody confirmed specificity of desmoplakin staining. After rinsing in PBS, alexafluor 488-conjugated donkey anti-mouse IgG (eBiosciences) was added at room temperature for 1 hour. Cells were coverslipped in mounting medium (ProLong Gold with DAPI; Invitrogen). Digital images were captured at random locations within each well on a Nikon C1 Plus confocal microscope using EasyC1 software. A minimum of 100 nuclei were counted manually by ImageJ software, and data are expressed as percentage of cells containing two or more nuclei.

Protein isolation and immunoblot analysis

To study the influence of EGF and IFN-γ on B7-H1 expression by trohpoblast cells, immunopurified cells were plated overnight at a density of 7.5 × 10⁶ cells in 60 mm Petri dishes in Iscove’s modified DMEM containing 10% FBS and antibiotics. The following morning, cells were rinsed twice, treated with medium with or without 5 ng/ml EGF (Peprotech, Rocky Hill, NJ) or 100 U/ml IFN-γ (R&D Systems; Minneapolis, MN) and incubated in a 5% CO₂/95% air atmosphere for up to 144 hours. Long-term cultures (>48 hours) were replenished with fresh culture medium with or without treatments every other day. For signal transduction pathway studies, biochemical inhibitors were dissolved in DMSO (vehicle) and controls included medium alone and vehicle treatments. Cells were pretreated with each inhibitor for 1 hour prior to initiation of the 48 hour culture. Inhibitors included 20 μM tyrphostin AG1478 (EGF receptor), 20 μM PD98059 (ERK1/2 inhibitor; Sigma) and 10μM UO126 (MEK1/2 inhibitor; Sigma), 10 μM wortmannin (PI3K inhibitor; Calbiochem), 10 μM SB203580 (p38 MAPK inhibitor; Sigma) and 50 ng/ml rapamycin (mTOR inhibitor; Sigma). Wortmannin treatment induced a loss of nuclear morphology that was consistent with apoptosis by approximately 25%; other treatments did not affect cell viability.

Protein from cultured cells was collected by lysis in RIPA buffer (1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate) containing protease inhibitors (100 μg/ml phenylmethylsulfonic acid, 10 μg/ml aprotinin, 10 μg/ml leupeptin). Protein (10 or 20 μg) was electrophoresed under denaturing conditions and transferred to nitrocellulose membranes. Membranes were blocked in 3% nonfat milk/TBS and probed with goat anti-human B7-H1 antibody (100 ng/ml) (R&D Systems, Minneapolis, MN). To verify equal loading of samples, membranes were stripped and re-probed with rabbit anti-actin (Sigma cat. no. A-5060). After incubating with the corresponding horseradish peroxidase-labeled secondary antibody (Jackson ImmunoResearch, West Grove, PA), bound antibodies were detected using a chemiluminescent detection system (Pierce Biotechnology, Rockford, IL) and autoradiography. Normalization of B7-H1 expression to actin was performed by taking the ratio of densitometric intensities using Gel-Pro image analysis software (Silver Spring, MD).

RNA Extraction and RT-PCR

To study the effects of EGF on B7-H1 mRNA expression, 7.5 × 10⁶ trophoblast cells were plated in 60 mm Petri dishes and treated the following day for up to 48 hours. Total cellular RNA was extracted by addition of Trizol reagent (Invitrogen, Carlsbad, CA), following the manufacturer’s instructions. Total RNA was pretreated with DNase I (Sigma) and quantified by spectrophotometry. One μg RNA was reverse transcribed using MMLV reverse transcriptase and oligo dT primers (Invitrogen). Real time PCR was performed using an ABI Prism 7900 System (Applied Biosystems, Foster City, CA). Primers used included B7-H1 (5’-tca atg ccc cat aca aca aa-3’ (Fwd) and 5’-cga agt cat ctg gac aag c-3’ (Rev), and the housekeeping gene β-actin (Applied Biosystems). Each sample was assayed in triplicate and was compared to arbitrary values assigned to a standard curve generated from serial dilutions of placental RNA to obtain relative abundance of amplified product. These values were then normalized to those of β-actin products.

Analysis of fractionated RNA was performed by culturing 180-240 × 10⁶ trophoblast cells divided equally between 100 mm Petri dishes (45-60 × 10⁶/dish) and treated with either medium alone or 5ng/ml EGF. After 36 hours, cells were lysed as described [19] with several modifications. Briefly, cells were incubated (10 min) in media containing 10ug/mL cycloheximide, scraped and resuspended in 1mL PBS containing 100ug/ml cycloheximide and 1mM PMSF. A 50μl aliquot was placed in Trizol reagent for extraction of total RNA. The remaining cell suspension was pelleted and then lysed in low-salt buffer (LSB) (20mM Tris, 10mM NaCl, 3mM MgCl₂) with 100U/ml RNAse Inhibitor (Invitrogen) on ice. Detergent buffer (1.2% Triton N-101 in LSB) (250μl) was added and the cells were homogenized on ice using a glass dounce homogenizer. The lysate was centrifuged (4°C, 10,000×g, 1 min) and the supernatant transferred to a tube containing 100ul of LSB with 10mg heparin/ml and 1.5M NaCl. The suspension was loaded directly onto a 11.5-ml continuous sucrose gradient (15-50% w:v in LSB with 25U/ml RNAse Inhibitor), and centrifuged at 36,000 × g for 4h. Each sucrose/RNA gradient was run through a UA6 UV detector/chart recorder (Teledyne Isco) and a running trace (250nm) was recorded as 250ul fractions were collected using the Foxy Jr. Fraction Collector (Teledyne Isco). RNA fractions corresponding to the 40S, 60S, monosomal, and individual polysomal (disomic, trisomic, etc.) were then combined to generate 11 fractions, and RNA was isolated and examined by real time RT-PCR as described above. Total RNA was examined for quantity and quality on a Dual Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA).

Plasmid construction and luciferase assays

Two putative B7-H1 promoters corresponding to 1.069 kb upstream of exon 1 (UTR), and 1.2 kb upstream of the translation start codon (Int1) were analyzed for transcription factor binding sites using Transcription Factor Search online software (http://www.cbrc.jp/research/db/TFSEARCH.html) [20]. Human genomic DNA (Applied Biosystems) was amplified by PCR using primers to these regions containing cloning sites (underlined). Primers were: UTR-Fwd, 5’ – gca act cga ggt ggt caa gct cat – 3’; UTR-Rev: 5’ – cag aga tct tgc cct aat tat g – 3’; Int1-Fwd: 5’ – tga ctc gag gtg gct caa gcc tgt a – 3’, and (Int1-Rev: 5’ – ctc tcg agc cca aag aaa ggg tgt a – 3’. These regions were cloned into the promoterless luciferase expression vector pGL3 (Promega) using XhoI/BglII (to create B7H1.UTR) and XhoI (B7H1.Int1).

The human choriocarcinoma cell lines JEG-3, Jar or BeWo (ATCC, Manassas, VA) were grown to 50% confluence in 12-well tissue culture plates. pGL3-basic (negative control), pGL3-B7H1(UTR) or PGL3-B7H1(Int1) were transfected using Lipofectamine reagent (Invitrogen) with some modifications [21]. Following transfection, the cells were treated with medium alone, EGF (5 ng/ml) or IFN-γ (100 U/ml) for 8 hours. Luciferase assays were performed using the Dual Luciferase Reporter Assay System (Promega) according to the manufacturer’s protocol. Transfection studies were conducted in triplicate in 3-4 separate experiments, and data are expressed as luciferase activity (relative light units [RLU]) of each promoter region, and normalized to that of Renilla.

Statistical analysis

Experimental data were analyzed by two-way analysis of variance at α = 0.05, using the Tukey post-hoc test where differences were found. For time-course and polysomal mRNA analyses, two-way ANOVA was performed using repeated measures so that treatment effect from all time points or fractions, respectively, could be analyzed simultaneously. A significant (P < 0.05) interaction indicated that the effects of the treatment on mRNA levels was dependent on the time point or fraction analyzed. Data are expressed as means ± SEM, and asterisks denote statistically significant differences (P < 0.05) from the indicated control.

Time-course analysis of EGF-stimulated B7-H1 expression in trophoblast cells

We previously showed that the syncytiotrophoblast abundantly expresses B7-H1 protein in vivo, and that trophoblast cells syncytialized in vitro with EGF express higher levels of this protein than undifferentiated control cells [12]. To gain insight into the mechanism of EGF-induced B7-H1 protein induction, we performed a time-course analysis of B7-H1 mRNA and protein. Unstimulated levels of B7-H1 protein remained constant over the course of a 6-day culture period (P > 0.05), whereas EGF-induced protein doubled by 24 hours and further increased to nearly 3-fold by 48 hours. EGF-induced protein levels remained two- to three-fold above untreated control cells throughout the culture period high throughout the 6-day culture period (Figure 1A and B).

B7-H1 can be regulated at the level of transcription by cytokines such as IFN-γ [22]. To determine whether increased transcription of the B7-H1 gene might also account for EGF-induced elevation of protein levels in trophoblast cells, a short-term time course was performed in which trophoblast cells were treated with EGF for 0-24 hours and analyzed by real-time RT-PCR. EGF significantly upregulated B7-H1 mRNA expression as early as 4 hours after the onset of treatment, and message levels remained high at levels for the duration of the 24 hour culture (Figure 1C). In untreated control cells, B7-H1 mRNA levels did not increase before 24 hours of culture (P < 0.05). Thus, EGF treatment resulted in an accelerated increase in B7-H1 mRNA expression in comparison to untreated control cells.

B7-H1 upstream regions do not respond to EGF

Given the rapid induction of mRNA, we next tested the hypothesis that this growth factor could regulate B7-H1 at the transcriptional level. Two potential promoter regions of the B7-H1 gene were each cloned into a promoterless vector such that promoter activity of the cloned regions would drive luciferase expression. These two regions, depicted in Figure 2A and termed B7H1.UTR and B7H1.Int1, were each tested for their responsiveness to EGF. JEG-3 cells, which, like primary trophoblast cells, express B7-H1 constitutively, and Jar cells, which lack B7-H1 even when treated with proinflammatory cytokines, were transiently transfected with vectors containing these regions. IFN-γ, a positive control, induced B7H1.UTR promoter activity in both JEG-3 and Jar cells, and moderately induced B7-H1.Int1 promoter activity in JEG-3, but not Jar cells (Figure 2C-F). In contrast, EGF failed to stimulate activity of either of the two promoters, regardless of whether JEG-3 or Jar cells were used, despite its ability to upregulate the protein in JEG-3 cells (Figure 2B). Similar results were obtained with BeWo cells; these cells exhibited a 3.6-fold increase (P = 0.002) and 2.5-fold increase (P = 0.001) in luciferase activity in response to IFN-γ over untreated controls when transfected with B7H1.UTR or B7H1.Int1, respectively, whereas they did not support EGF-stimulated activity of either putative B7-H1 promoters (P > 0.05).

EGF regulates B7-H1 through the PI3 Kinase/mTOR pathway

To gain a better understanding of the mechanisms of EGF-induced B7-H1 protein expression, we next analyzed the role of components of intracellular signaling cascades. As shown in Figure 3A, the EGF receptor tyrosine kinase inhibitor AG1478 completely blocked B7-H1 protein stimulation by EGF, confirming the requirement for tyrosine phosphorylation of the EGF receptor.

mTOR, which can be regulated by several different signal transduction pathways, is a key mediator of translation. To examine a potential role for mTOR, we tested the effect of rapamycin on EGF-induced B7-H1 protein expression. Figure 3B shows that rapamycin blocked the ability of EGF to upregulate B7-H1 in trophoblast cells (P < 0.05). To determine the likely regulator of EGF-induced mTOR activity, we next tested the ERK1/2 inhibitor PD98059 and the PI3K inhibitor wortmannin on EGF-treated trophoblast cells (Figure 3B and C). Only wortmannin prevented the rise in B7-H1 protein, indicating that the PI3K-Akt but not the ERK1/2-MEK1/2 pathway controls mTOR activation in this EGF-induced response.

EGF is well-known to induce differentiation trophoblast cells as measured by syncytialization. EGF-mediated syncytialization has previously been shown to be dependent upon the p38 MAPK pathway [23]. To evaluate whether the B7-H1-inducing effect of EGF is direct or secondary to differentiation, we measured the degree of syncytialization of trophoblast cells in the presence an absence of rapamycin and SB203580, an inhibitor of p38 kinase. Treatment with EGF for 48 hours induced a morphological transformation of the cells, from small aggregates of 5-15 cells to larger aggregates of cells (Figure 4A). In addition, a modest but consistent increase in syncytialization was observed as a result of EGF treatment, from approximately 15% to about 25% in EGF-treated wells, although this level did not reach statistical significance (P = 0.134; Figure 4B). Treatment of cells with the p38 MAPK inhibitor, SB203580 neither prevented this modest increase in syncytialization, nor did it interfere with EGF-mediated induction of B7-H1; two experiments using trophoblast cells from different placentas revealed 1.6- and 3.7-fold increases in B7-H1 expression following EGF treatment. In contrast, rapamycin, which inhibited B7-H1 expression (Figure 3B), appeared to promote EGF-induced syncytialization.

EGF induces an increase in the proportion of translatable B7-H1 mRNA

The identification of the mTOR pathway in EGF-induced B7-H1 expression prompted us to next evaluate whether EGF operates through a post-transcriptional gene regulatory mechanism to upregulate B7-H1. We therefore examined whether induction of B7-H1 protein expression by EGF is accompanied by a shift in the proportion of B7-H1 mRNA being actively translated. In 3 separate experiments, primary cytotrophoblast cells were treated with either medium alone or EGF, lysed, and subjected to density gradient fractionation and analysis of mRNA within each fraction. The majority of β-actin mRNA migrated with the heavier polysomal RNA fractions, regardless of treatment (Figure 5). There was no significant fraction × treatment interaction (P = 0.478), revealing that EGF had no effect on β-actin mRNA distribution among polysomal RNA. Conversely, B7-H1 mRNA exhibited a markedly different distributional pattern in response to EGF treatment. In untreated control cells B7-H1 mRNA was biphasically distributed within the monosomal fraction (fraction 5) and the heavier polysomal fractions (fractions 6-11; Figure 5). Upon treatment with EGF, however, the proportion of total B7-H1 mRNA present in the monosomal fraction shifted to the polysomal fractions (Figure 5). This was reflected by a significant fraction × treatment interaction (P = 0.043).