| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, Kansas 66160
ABSTRACT
Modulation of the maternal immune system by the placenta is a mechanism by which the fetus ensures its own survival in a genetically foreign environment. The immunoinhibitor CD274 (also called B7-H1 or PD-L1) is highly expressed in the placenta, positioned to interact with maternal leukocytes. Further, immunoblot analysis of first- and second-trimester placental lysates showed that CD274 expression is low in the first trimester but dramatically rises around the onset of the second trimester. As this coincides with the expected onset of maternal blood flow to the placenta and a corresponding rise in local oxygen tension, we explored the possibility that oxygen regulates CD274 expression in trophoblast cells by culturing term trophoblast cells under oxygen concentrations similar to those found in vivo. Indeed, CD274 protein levels paralleled the in vivo situation: expression increased with rising oxygen concentrations. Furthermore, downregulation of CD274 mRNA by low oxygen was rapid, occurring within 4–12 h. We conclude that oxygen is a potential mediator of CD274 expression in vivo such that it is induced coincidentally on exposure of fetal tissues to maternal blood. Further, the regulation of this immunomodulator by oxygen may implicate its alteration during and involvement in the pathogenesis of complications of pregnancy such as preeclampsia.
immunology, placenta, pregnancy, syncytiotrophoblast, trophoblast
INTRODUCTION
The requirement for maternal immunological tolerance of the fetal placenta is underscored by the fact that this tissue lies in direct apposition to maternal blood lymphocytes and decidual macrophages, dendritic cells, and natural killer cells. Trophoblast cells, as the anatomical interface between the fetus and mother, possess the means to bias the maternal immune system toward immunological acceptance of pregnancy. The mechanisms by which lymphocyte tolerance toward trophoblast cells is established, however, have not been fully elucidated.
B7 family molecules are type I transmembrane proteins belonging to the immunoglobulin superfamily. In concert with their CD28 family receptors, the B7s are key regulators of the adaptive immune response. To date, there are seven proteins classified within the B7 family and four within the CD28 family [1]. CD274 (also called B7-H1 and PD-L1) was the first of these novel B7 family proteins to be discovered [2, 3]. Several lines of in vitro and in vivo experimental evidence strongly suggest that CD274 is a negative regulator of T and B cells and that its expression in peripheral organs may be of key importance in mediating tolerance of lymphocytes to self-antigens. CD274 inhibits proliferation and cytokine production by activated T cells in vitro [1, 3, 4]. In mice, targeted null-mutagenesis of the CD274 receptor, CD279 (also called PD-1), results in autoimmune disorders involving CD274-expressing organs [5–7]. In addition, genetic deletion of CD274 causes accumulation of chronically activated lymphocytes in peripheral organs and increased susceptibility to experimental autoimmune diseases [8, 9]. CD274 expression by cancer cells appears also to mediate immunoevasion by inducing apoptosis of tumor-specific T cells [10]. Recently, CD274 was found to have a obligatory role in prevention of T cell-mediated attack of allogeneic fetuses in the murine knockout model, confirming the critical nature of its expression at the maternal-fetal interface [11].
We have mapped the expression of B7 family proteins within the human placenta and found CD274 to be expressed in the syncytiotrophoblast and extravillous trophoblast throughout gestation [12, 13]. In examining expression of CD274 across gestation, we observed that tissue levels of CD274 are highest in the second and third trimesters as compared to the first trimester. During the first trimester, blood flow to the placenta via the spiral arteries is severely restricted because of obstruction of these vessels by migrating trophoblast cells [14]. On physiologic transformation of these arteries, which occurs around the beginning of the second trimester, these trophoblastic plugs disappear, allowing the unrestricted flow of oxygenated maternal blood to the intervillous space. A major consequence of arterial transformation is that the first-trimester placenta resides in an oxygen environment that is low relative to that of second and third trimesters (10–15 mm Hg or 1%–2% oxygen vs. 40–60 mm Hg or 6%–10% oxygen, respectively) [15, 16]. Furthermore, failure of arterial transformation is believed to result in chronic hypoxia or ischemic damage to the placenta and leads to clinical onset of preeclampsia, a life-threatening condition affecting as many as 8% of all pregnancies [17]. In this study, we performed a more detailed analysis of the timing of the increase in CD274 expression in the placenta and examined whether oxygen could be a mediator of this change.
MATERIALS AND METHODS
Tissue Acquisition
Term placentas with no associated pathologies of pregnancy were collected after caesarian sections and first- and second-trimester placentas were collected following elective pregnancy termination. Tissues were obtained upon the receipt of informed consent when exempt status was not approved and in accordance with protocols approved by the University of Kansas Medical Center's Human Subjects Committee (HSC #9028 to Dr. Joan Hunt and HSC #3037 to Dr. Margaret Petroff). For extraction of RNA and protein, villous placenta was snap-frozen in liquid nitrogen and stored at –80°C. For immunohistochemistry, histological specimens were prepared by embedding in TBS freezing medium (Fisher Scientific, St. Louis, MO).
Western Blot Analysis
Protein from placental tissue and 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 concentration was determined using the DC Assay (Bio-Rad, Hercules, CA). Protein (10–25 µg) was electrophoresed under denaturing conditions and transferred to a supported nitrocellulose membrane (Schleicher & Schuell, Keene, NH). The membranes were blocked in 3% nonfat milk/TBS and probed with 1 µg/ml goat anti-human CD274 (R&D Systems, Minneapolis, MN). Specificity of this antibody was confirmed using Jar cells, which do not express endogenous CD274 [13], that were transfected with either empty vector (pcDNA3.1; Invitrogen, Carlsbad, CA) or vector containing the coding sequence for human CD274 (unpublished data). To verify equal loading of proteins, membranes were stripped and reprobed with a rabbit antibody recognizing all isoforms of actin (1:5000; Sigma cat. no. A-5060). Additionally, blots were incubated with goat IgG to document the absence of nonspecific immunoglobulin binding. Incubations with primary antibodies were carried out overnight at 4°C. After stringent washing and probing with the corresponding anti-goat or anti-rabbit horseradish peroxidase-labeled secondary antibody (Jackson ImmunoResearch, West Grove, PA), bound antibodies were detected using the enhanced chemiluminescent detection system (Pierce Biotechnology, Rockford, IL) and autoradiography. Densitometric intensities were obtained using Gel-Pro image analysis software (Silver Spring, MD) and were normalized to ß-actin (ACTB).
Immunohistochemistry
For immunoperoxidase staining, 10-µm tissue sections were cut from frozen first and second trimester placentas using a cryostat. Nonspecific immunoglobulin binding was blocked in 10% normal horse serum. Primary antibody (anti-CD274, clone 2H1, 6.7 µg/ml) [2] or its isotype specific control at the same concentration (IgG1; Pharmingen) was incubated with the tissue sections or cells for 1 h at room temperature. After the addition of secondary antibody (biotinylated horse anti-mouse IgG; Vector Laboratories, Burlingame, CA), the samples were depleted of endogenous peroxidases by incubating in 0.5% H202/methanol. Reactivity was detected using the streptavidin-peroxidase and AEC reagent sets (Zymed, San Francisco, CA), and tissues were lightly counterstained in Mayer hematoxylin. Positive staining was viewed microscopically as a reddish coloration.
Trophoblast Isolation, Purification, and Culture
Term villous cytotrophoblast cells were isolated from term placentas as described [18]. Briefly, after removal of basal plate tissue, approximately 40 g of villous placental tissue were collected, finely minced, and dissociated by three 15-min stages in HBSS, 0.5% trypsin (Invitrogen), and 30 U/ml DNase (Sigma). The resulting cell suspension was layered over a 5%–70% discontinuous Percoll (Sigma) gradient and centrifuged at 2000 x g for 20 min. The cells migrating between the densities of 35% and 50% Percoll were collected and subjected to immunopurification by negative selection over columns consisting of magnetic microbeads coupled to the HLA-class I antibody, W6/32 (ATTC, Manassas, VA). Cells collected in this way are routinely >95% pure as assessed by flow cytometric and immunohistochemical analysis for cytokeratin-7 (clone OVTL12/30; DAKO Corp., Carpenteria, CA) [19].
To obtain a low-oxygen environment, trophoblast cells were first plated at ambient oxygen overnight. Cells were rinsed and replenished with fresh cell culture medium (Iscove modified DMEM) and transferred to a humidified 37°C chamber containing 2%, 5%, or 8%O2/5% CO2/balance nitrogen (BioSpherix Ltd, Redfield, NY). Alternatively, for ambient (
21%) oxygen conditions, cells were cultured in a standard humidified 5% CO2/95% air incubator. All culture media were preequilibrated to the appropriate oxygen concentration before addition to the cells. Culture medium was assessed for dissolved oxygen using a blood gas analyzer (Radiometer America, Westlake, OH) that showed oxygen levels in the media similar to theoretical oxygen levels, with some deviation (10 mm Hg) due to introduction of oxygen by electrode sensors into the non-oxygen-buffered solution [20]. Cells were cultured for 4–72 h under these conditions, at which time total cellular RNA or protein lysates were obtained. Alternatively, viability was assessed by trypsinization and trypan blue exclusion.
Some experiments were carried out using the chemical hypoxia mimetic CoCl2 (Sigma). Trophoblast cells were again plated overnight and replenished with either medium alone or 50–200 µM CoCl2. As described previously, cells were cultured for 4–72 h and harvested for RNA or protein analysis.
CGB ELISA
Human chorionic gonadotropin-ß (CGB) was measured in 25-µl volumes of cell culture supernatant using a commercial ELISA kit (DRG Instruments, Inc., Mountainside, NJ) according to the manufacturer's instructions.
Reverse Transcription and Real-Time PCR
Total cellular RNA was extracted using Trizol reagent (Invitrogen) according to the manufacturer's protocol, pretreated with DNaseI (Sigma) to prevent amplification of genomic DNA, and quantified by spectrophotometry. RNA (500 ng) was reverse transcribed using MMLV reverse transcriptase and oligo-dT primers in 40-µl reactions (Invitrogen). Of this, 1 µl was subjected to real time RT-PCR for 40 cycles in an ABI Prism 7900 Sequence Detection System (Applied Biosystems, Foster City, CA) using SYBR Green PCR mix. CD274 primers used were: 5'-tca atg ccc cat aca aca aa-3' (Fwd) and 5'-cga agt cat ctg gac aag c-3' (Rev); ß-actin (ACTB) primers were purchased from Applied Biosystems. Both sets were used at 300 nM final concentration. Standard curves were generated by reverse transcription and amplification of placental RNA in serial dilutions from 1:2 to 1:500 and were run alongside experimental samples in each of the PCR amplifications. To verify specificity of the product, amplified products were subjected to melting curve analysis as well as electrophoresis, and product sequencing was performed to verify identity. Negative controls included substitution of water for cDNA. Threshold cycle (Ct) values obtained from each reaction was determined and compared to those of the standard curve to obtain relative abundance of amplified product. These values were then normalized to those of ACTB products.
Statistical Analysis
Graphical data are expressed as means ± SEM. Normalized immunoblot and real-time RT-PCR data were analyzed by either two-tailed paired t-test or two-way analysis of variance blocked on placental sample to eliminate patient-to-patient variability. Where indicated, data were natural log (ln)-transformed to obtain homogeneity of variance. Statistically significant differences were analyzed by Student-Newman-Keuls post hoc test. Differences were considered significant when P < 0.05.
RESULTS
Trophoblast CD274 Expression in First- and Second-Trimester Placentas
We previously reported that CD274 protein expression in the placenta varies across gestation. Specifically, CD274 expression was lowest in the first trimester, higher in second-trimester placentas, and remained high in the third trimester [13]. To more precisely time the induction of CD274 expression, we subjected eight first-trimester placental samples (Gestational Weeks 5–12) and six second-trimester samples (Weeks 13–19) to immunoblot analysis using goat-anti-human CD274. Figure 1 shows that all first-trimester placentas except for the 12-wk sample contained low levels of CD274 protein relative to all the second-trimester samples. Thus, CD274 protein expression in the placenta increases between Weeks 10 and 12 of gestation.
|
To examine whether the increase in trophoblast CD274 expression accounts for the rise in total placental CD274 from first to second trimester, we compared CD274 expression in two first-trimester and two second-trimester placentas by immunohistochemistry (Fig. 2). Although the staining intensity varied within each of the samples, overall intensity of CD274 immunoreactivity was higher in the 16- and 19-wk placentas as compared to the 8- and 9-wk samples. Since no other cell types within the placental tissue of either trimester exhibited CD274 immunoreactivity, the increase in total placental CD274 expression (Fig. 1) is due to an increase in expression by trophoblast cells.
|
Trophoblast CD274 Regulation by Oxygen
Because the changes in placental expression of CD274 corresponded with the approximate expected timing of the onset of maternal blood flow to the placenta, we next examined whether oxygen could regulate the protein. Purified term villous trophoblast cells were cultured in the absence or presence of increasing concentrations of the hypoxia mimetic CoCl2. Figure 3A depicts an immunoblot of trophoblast cells cultured in 0–200 µM CoCl2 for 72 h. CD274 protein decreased with increasing concentrations of CoCl2, while expression of the structural protein actin was unchanged. Therefore, decreasing the amount of available oxygen in the media resulted in a reduction of CD274 protein in trophoblast cells. To assess the possibility that changes in mRNA abundance accompany the oxygen-dependent changes in protein abundance, we performed a similar experiment using semiquantitative real-time RT-PCR. Figure 3B shows that, indeed, CoCl2 reduced CD274 mRNA abundance in a dose-dependent manner.
|
To determine whether these results could be duplicated under varying concentrations of environmental oxygen, we next cultured purified trophoblast cells in either 5% or ambient (21%; standard cell culture) oxygen conditions. Figure 4, A and B, shows the results from three preparations of trophoblast cells cultured under these conditions. In all three experiments, there was a tendency for a reduction in the level of CD274 protein by 72 h; in one of the three, this reduction was already apparent by 24 h.
|
Physiologic concentrations of oxygen within the intervillous space during the first and second trimesters are approximated to be 1%–2% and 8%–10%, respectively [15]. Therefore, trophoblast cells were cultured in 2% and 8% oxygen to mimic the first and second trimesters, respectively, and 21% oxygen. As can be seen in Figure 4, C and D, the increase from 2% to 8% oxygen resulted in a rise in CD274 expression (P < 0.05). The increase in oxygen from 8% to 21%, however, failed to further increase CD274 protein (P > 0.05); instead, there was a tendency for a reduction of CD274 expression at 21% in comparison to 8%. These results show that CD274 protein levels in trophoblast cells are sensitive to physiologic concentrations of oxygen.
From these studies, it could be seen that changes in CD274 protein due to oxygen occur within a period of 72 h. We wondered whether changes in the abundance of CD274 mRNA preceded this time period. Trophoblast cells were first subjected to culture for 0–72 h in the presence of either culture medium alone or 100 µM CoCl2. Figure 5A shows that, while CD274 expression did not change over time with no treatment, CoCl2 caused a significant reduction in CD274 mRNA abundance by 12 h following exposure (P < 0.05).
|
We next investigated whether physiological concentrations of oxygen could also effect this rapid change in CD274 expression. Trophoblast cells were cultured under 2%, 8%, or 21% oxygen for 0–72 h. Culture of the cells in 2% oxygen caused a significant (P < 0.05) reduction of CD274 mRNA in comparison to both 8% and 21% oxygen by 4 h of exposure (Fig. 5B). Under this condition, CD274 mRNA expression remained low throughout the 72-h culture period. Cells cultured in 8% oxygen also exhibited a reduction in CD274 mRNA expression; however, this reduction was delayed in comparison to those cells in 2% oxygen until about 48 h.
Cultured term villous trophoblast cells differentiate morphologically and endocrinologically into syncytiotrophoblast-like cells over time, an event that is prevented by low concentrations of oxygen [21–24]. In addition, increases in trophoblast CD274 protein accompany differentiation in vivo and in vitro [13]. We therefore wished to determine whether the oxygen-dependent decline in CD274 expression was simply an effect secondary to the effects of oxygen on trophoblast differentiation. Trophoblast differentiation was assessed by time course analysis of human chorionic gonadotropin (CGB) production under various oxygen concentrations. As shown in Figure 6, CGB secretion by cells cultured in 8% and 2% oxygen was low or absent, confirming that hypoxia prevented trophoblast differentiation. In contrast, CGB secretion was detectable in the media by 24–48 h, when cells were cultured in ambient oxygen conditions, suggesting that these cells had begun to differentiate. Thus, the oxygen-dependent decline in CD274 mRNA preceded apparent cytotrophoblast differentiation as measured by CGB secretion.
|
DISCUSSION
These results document for the first time the ability of oxygen to modulate expression of the trophoblast immunoinhibitory molecule CD274. Trophoblast CD274 is differentially expressed across gestation in a manner correlating with exposure to maternal blood and lymphocytes. Exposure of term trophoblast cells to oxygen was found to be associated with a rapid induction of CD274 expression at both the RNA and the protein level. Thus, we have identified a novel mechanism by which immune molecules in the placenta can be regulated.
By both immunohistochemical analysis and flow cytometric analysis, we have shown that all populations of trophoblast cells express the protein throughout pregnancy, including syncytiotrophoblast and the underlying, less differentiated villous cytotrophoblast cells of early and late pregnancy [12, 13] (unpublished data). In addition, our initial studies detected a lower amount of CD274 in whole tissue lysates of first-trimester as compared to second- and third-trimester placentas [13]. In the current study, we expanded the number of placentas analyzed to include the time period bracketing the transition between the first and second trimesters. In addition to confirming our previously published results, we found that the increase in CD274 expression occurs rather abruptly at the onset of the second trimester, corresponding to the expected period of physiological transformation of the maternal uterine spiral arteries.
Maximization of the maternal-fetal interface is accomplished by continuous sprouting and branching of chorionic villi such that trophoblastic surface area expands. Trophoblast cells proliferate early in gestation, and the total numbers of trophoblast cells steadily increase as gestation progresses [25]. Therefore, greater contribution of trophoblast cells to total tissue composition in the second trimester (Fig. 1) could account for the overall rise in CD274 protein at this time. If this were this the case, however, one would expect a gradual rather than abrupt rise in placental CD274. Furthermore, concomitant analysis of first- and second-trimester tissue sections by immunohistochemistry suggests that CD274 expression by first-trimester trophoblast cells is less than that of second-trimester cells. Thus, even on a per cell basis, CD274 expression appears to be lower in the first than in the second trimester.
Because the rise in trophoblast expression of CD274 coincides with the expected rise in placental oxygen partial pressure [15], we investigated oxygen as a potential regulator of CD274 expression, using term trophoblast cells as our model. Consistent with the postulate that oxygen induces expression of CD274 in vivo, we found that lowering the oxygen concentration, either chemically or environmentally, also reduced cellular levels of the protein. We further examined the possibility that physiologically relevant concentrations of oxygen mimic CD274 expression patterns seen in vivo. Intervillous oxygen concentrations reside near 2% and 8% in the first and second trimesters, respectively [15]. When trophoblast cells were cultured under these conditions, we consistently found that the most profound changes in CD274 protein levels occurred between 2% and 8% oxygen in a manner that paralleled the in vivo rise in CD274 expression in the second trimester. The change in CD274 expression that occurred between 8% and 21% oxygen were, in contrast, inconsistent; CD274 expression was either the same in the two conditions or reduced from 8% to 21% oxygen. The changes in CD274 expression were not due to cytotoxicity, as viability remained unchanged under the varying levels of oxygen (data not shown).
To determine whether oxygen has an effect on the abundance of mRNA for CD274, time course analyses were performed in which trophoblast cells were cultured for 4–72 h in either chemically or environmentally reduced oxygen atmospheres. Surprisingly, the abundance of CD274 message was reduced within 4 or 12 h following exposure to 2% oxygen or CoCl2, respectively. These results suggest that the CD274 gene may be regulated directly by oxygen-sensitive transcriptional regulators. Alternatively, reduced oxygen could have a negative impact on mRNA stability. Regardless, oxygen-mediated changes in the CD274 gene appear to be independent of the effects of oxygen on cytotrophoblast differentiation [21, 22] since the changes in CD274 expression occurred before any obvious changes in trophoblast differentiation as measured by CGB production.
Many studies of oxygen-regulated genes center on the cellular adaptations to hypoxia involving induction of genes by enzymatic sensors of low oxygen. In this study, however, we have found that hypoxia decreases expression of CD274 protein and mRNA. While this change may not relate to adaptation or survival of trophoblast cells under low oxygen per se, the in vivo consequences of oxygen stimulation of CD274 may be significant. During the first trimester, when there would be a relative paucity of maternal blood lymphocytes circulating through the placenta, CD274 may not be required. However, with the onset of maternal blood circulation in the placenta, close contact between the syncytiotrophoblast and maternal leukocytes occurs. Conceivably, oxygen-rich blood could signal the presence of lymphocytes to trophoblast, resulting in expression of immunoprotective molecules such as CD274. Taken a step further, oxygen could also signal a protective, immunoevasive mechanism for tumors cells, most of which express high levels of CD274 in contrast to their normal-tissue counterparts [10].
The mechanism by which low oxygen signals the shutdown of CD274 or, alternatively, higher oxygen triggers induction of CD274 requires further investigation. Hypoxia-inducible factors (HIF)-1 and –2 are transcription factors best understood for their actions in inducing gene expression under low oxygen. HIF1A also induces transcriptional repressors, such as deleted in esophageal cancer-1 (DEC1) [26], and therefore could mediate gene suppression indirectly. Finally, oxygen-mediated transcriptional repression has been shown to occur for other genes by HIF1A-independent mechanisms through as-yet-undefined pathways [27].
While acknowledging that a more exacting test of the ability of oxygen to regulate CD274 gene expression during early gestation will be to perform similar experiments in first-trimester trophoblast cells, our results definitively identify oxygen as a potent modulator of its expression in trophoblast cells. Furthermore, its ability to modulate CD274 expression in term trophoblast cells raises the possibility that CD274 levels are altered in clinically important conditions, such as preeclampsia and intrauterine growth retardation, in which fluctuations in placental oxygen likely occur [17]. Collectively, these results suggest that the physiological change in oxygen can result in a large difference in CD274 expression, paralleling our in vivo observations. Thus, the results have important new implications for the regulation of immune responses in the placenta.
ACKNOWLEDGMENTS
The authors thank Elza Kharatyan for her excellent technical assistance, Drs. Betsy Wickstrom and Lynn Rasmussen for assistance in obtaining informed consent for use of term placentas, and Dr. Lane Christenson for assistance with real-time PCR assays. Anti-human CD274 monoclonal antibody and the CD274 expression vector were generous gifts from Lieping Chen (Johns Hopkins University School of Medicine, Baltimore, MD). Oxygen experiments were performed within the Hypoxia core laboratory in the Institute for Maternal-Fetal Biology. Microscopy and graphic support services were provided by the University of Kansas Center for Reproductive Sciences through an NICHD grant for Specialized Cooperative Centers Program in Reproductive Research (SCCPRR) (P. Terranova, director). Histological support facilities are provided by NIH grant P30 HD02528 to the Kansas Mental Retardation and Developmental Disabilities Research Center.
FOOTNOTES
1 Supported by grants from the Lied Endowed Biomedical Pilot Research, the Kansas City Area Life Sciences Institute, the Kansas IDeA Network of Biomedical Research Excellence, and NIH grant R01 HD045611 to M.G.P. 
2 Correspondence: Margaret G. Petroff, Department of Anatomy and Cell Biology, University of Kansas Medical Center, 3901 Rainbow Blvd., Mail Stop 3038, Kansas City, KS 66160-7400. FAX: 913 588 2710; mpetroff{at}kumc.edu
Received: 11 August 2005.
First decision: 4 September 2005.
Accepted: 24 October 2005.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
