Mammalian pregnancy is an intriguing immunological phenomenon where the semiallogeneic fetus is not rejected. Tolerance toward the fetus involves a number of mechanisms associated with modifications of the immune status of the mother. In this study, we strongly suggest a novel mechanism for fetal evasion of maternal immune attack, based on the engagement and down-regulation of the activating NK cell receptor NKG2D on PBMC by soluble MHC class I chain-related proteins A and B (collectively termed MIC). A similar immune escape pathway was previously described in tumors. We found that MIC mRNA was constitutively expressed by human placenta and could be up-regulated upon heat shock treatment. Our immunomorphologic studies showed that the MIC expression in placenta was restricted to the syncytiotrophoblast. Immunoelectron microscopy revealed a dual MIC expression in the syncytiotrophoblast: on the apical and basal cell membrane and in cytoplasmic vacuoles as MIC-loaded microvesicles/exosomes. Soluble MIC molecules were present at elevated levels in maternal blood throughout normal pregnancy and were released by placental explants in vitro. Simultaneously, the cell surface NKG2D expression on maternal PBMC was down-regulated compared with nonpregnant controls. The soluble MIC molecules in pregnancy serum were able to interact with NKG2D and down-regulate the receptor on PBMC from healthy donors, with the consequent inhibition of the NKG2D-dependent cytotoxic response. These findings suggest a new physiological mechanism of silencing the maternal immune system that promotes fetal allograft immune escape and supports the view of the placenta as an immunoregulatory organ.

The survival of the semiallogeneic fetus is a fascinating unresolved question in reproductive and transplantation immunology. The fetus itself is never in direct contact with the uterine tissues and maternal blood. Instead, fetally derived epithelial cells, the trophoblasts, invade the pregnant uterine mucosa, the decidua, and form the placenta. Two structural interfaces exist in the placenta, one between the invading extravillous trophoblast and the decidua, and a second, between the syncytiotrophoblast (STB)3 and the maternal blood. In early pregnancy, two layers of trophoblast cells cover the placental villi, an inner layer of cytotrophoblast (CTB) adhering to the basal membrane and an outer layer of STB, derived from the fusion of CTB. The placenta contains both paternal and maternal genes, but the paternal ones are preferentially expressed (1), making recognition and rejection of the placenta by the maternal immune system a potential risk. However, the immunological mechanisms that govern immune responses directed against “non-self”, including MHC, seem to be bypassed during pregnancy.

Murine studies suggest that maternal innate and adaptive immunity is transiently tolerant to fetal histocompatibility Ags during mammalian gestation (2, 3, 4). Several hypothetical mechanisms are proposed to explain this tolerance. The extravillous trophoblast expresses an unusual combination of MHC class I molecules: the classical HLA-C and nonclassical HLA-G and -E that could locally modulate NK cell function (5). The placenta expresses IDO, an enzyme that inhibits T lymphocyte proliferation by catalyzing the degradation of tryptophan (4). Subsets of regulatory T cells operate in human and murine pregnancy (6, 7). Expression of FasL by trophoblast (8, 9) may further protect the fetus by inducing apoptosis of activated maternal T lymphocytes. In addition, a possible complement attack is avoided by placental expression of the complement regulatory proteins (10). Besides local mechanisms, the global influence of pregnancy on systemic maternal immunity is well recognized. Pregnancy can modulate the course of autoimmune diseases (11). Moreover, animal experiments have shown that the maternal immune system specifically tolerates paternally derived tumor cells (3). Thus, fetal tolerance is highly complex and uses a variety of mechanisms that modulate the maternal immune system at the local and systemic level.

Recently, a new immune escape mechanism has been described for tumors based on the interactions of the human stress-inducible MHC class I chain (MIC)-related proteins A and B (MICA/B) molecules and their cognate receptor NKG2D (12, 13). MICA and B glycoproteins are related cell surface molecules and are both highly polymorphic. The genes encoding these molecules contain heat shock response elements in their promoters (14, 15). MIC molecules are ligands of the NK-cell activating receptor NKG2D expressed on the surface of human NK, γδ T, and CD8+αβ T cells (16). Normally, MIC proteins are constitutively expressed in low levels on epithelial cells in the gut and thymus, endothelial cells, and fibroblasts (14, 15), but become up-regulated or expressed de novo in stressed conditions, such as in tumors (12, 17), viral and bacterial infections (18, 19, 20), and in autoimmune conditions (21). Although related to MHC class I Ags, MIC molecules and their murine homologs RAE/H60 do not present Ags, but serve as signals of cellular stress, and trigger a range of immune effector functions such as cytotoxicity and cytokine production (14, 22). The cross-linking of the “inducible-self” MIC molecules with the NKG2D receptor enables immune cells to recognize and attack stressed cells without the need of MHC class I expression or Ag recognition (23). Thus, the MIC/NKG2D interaction is an effective mechanism for immunosurveillance.

Several reports have shown that soluble form of MIC (sMIC) proteins is released from some types of tumors in the serum of cancer patients and affect NKG2D function by acting as competitive inhibitors that block the recognition of membrane-bound MIC molecules and down-regulate the NKG2D surface expression on NK and T cells (24, 25, 26, 27). Thus, MIC molecules are not only involved in boosting cytotoxicity but also in its suppression. In this respect, they can be regarded as truly immunomodulatory molecules of the cytotoxic response.

Prompted by these findings, we set out to examine whether MIC proteins are expressed in human placenta, and whether a similar escape mechanism based on interaction between sMIC molecules and NKG2D receptor operates in pregnancy.

Healthy women undergoing elective termination of pregnancy (8–14 wk of gestation) and at normal delivery donated placenta specimens after ethical committee approval and informed consent. The following Abs were used: affinity purified goat anti-MICA/B Ab (E-16) against a peptide mapping near the N terminus of MICA (Santa Cruz Biotechnology); anti-MICA mAbs from Immatics, AMO1 (specific for α1 and α2 domains of MICA), and BAMO3 (specific for the α3 domain of MICA and MICB), anti-NKG2D mAb, clone 149810 (R&D Systems), and clone 1D11 (BD Biosciences). Isotype-matched Ig control DAK-GO1 was purchased from DakoCytomation. Recombinant MICA was purchased from Immatics. The human MIC-expressing leukemia cell line K562 (28) was used as a target in cytotoxicity assays.

Early and term placenta samples were fixed in 4% paraformaldehyde for 2 h, washed in PBS with 3.5% sucrose and 1% fish gelatin, and snap-frozen in liquid nitrogen. Cryosections (5 μm) were stained using goat anti-MICA/B Ab as described (9). Normal goat serum was used in all blocking steps. Parallel sections were stained after pretreatment with blocking peptide, according to the manufacturer’s instructions. Additional negative control sections were included omitting the primary Ab and using goat serum instead.

Early and term placenta samples were fixed in 4% paraformaldehyde (Merck) in 0.1 M phosphate buffer for 4 h, washed in the same buffer containing 7% sucrose and 0.05% saponin (Merck) at 4°C overnight, and snap-frozen. Cryosections were processed by the indirect immunoperoxidase method (9) using goat anti-MICA/B Ab. Ultrathin sections were examined without additional staining. Sections incubated with PBS instead of specific Ab were used as negative controls.

PBMC from pregnant women and nonpregnant donors were isolated by Lymphoprep (Nycomed) gradient centrifugation.

First trimester trophoblast cells were isolated following the procedure described in Ref.9 and used for total RNA preparation. For stress induction of MICA and MICB mRNA, the isolated cells were incubated at 42°C for 0, 30, and 60 min.

First trimester placenta samples (11–12 wk of gestation, n = 3) were cultured in vitro using a standard technique (29). Briefly, chorionic villi were washed with sterile RPMI 1640 and cut into small pieces of ∼5 mg of wet weight. Two such pieces were cultured in individual tissue culture inserts (10-mm diameter, 0.4-μm polycarbonate mesh; Nunc) in 1.5 ml of culture medium (RPMI 1640, 10% FBS, and antibiotics). The explant preparations were performed in triplicates to assess reproducibility within the same placental sample. Cultures were maintained at 37°C in 5% CO2 and humidified air. Because syncytiotrophoblast viability in cultured human placental explants is limited to 24 h (29), supernatants were harvested after 24 h of culture. Each supernatant was centrifuged at 400 × g and the presence of sMIC molecules was assayed by ELISA.

Total RNA and cDNA were obtained as described (6) and amplified on ABI PRISM 7700 by TaqMan Gene Expression Plate (I) protocol (Applied Biosystems). Primer and probe sequences, detecting all allelic forms of MICA and MICB, were as follows: forward primer for both MICA/B, 5′-TGGAACACAGCGGGAATCA-3′; specific reversed primers for MICA, 5′-TGACTCTGAAGCACCAGCACT-3′, and MICB, 5′-TGACTCTGAAGCACCAGCG-3′; and MICA/B probe, 5′-CACTCACCCTGTGCCCTCTGGGAA-3′. 18S rRNA was used as an endogenous control. Cycling conditions: 50°C for 2 min and 95°C for 10 min, followed by 40 cycles of 95°C for 5 s, and 60°C for 1 min. The MIC mRNA expression was presented as quantities related to a reference (=1) of MIC mRNA expression in PMA-ionomycin-stimulated PBMC (30).

ELISA plates (Nunc) were coated with goat anti- MICA/B Ab at a 1/100 dilution overnight. After blocking and washing, standards of rMICA, test sera and PBS were incubated for 1 h at room temperature. After washing, a mixture of anti-MICA/B mAbs AMO1 and BAMO3 was applied for 1 h followed by incubation with peroxidase-conjugated rabbit anti-mouse IgG (DakoCytomation) at 1/1000 for 1 h. Washed plates were developed by the TMB Peroxidase Substrate System (KPL), and absorbance was measured at 450 nm. The same ELISA procedure was used to determinate sMIC concentration in culture medium of early placenta explants.

NKG2D expression on PBMC from pregnant and control women was analyzed by flow cytometry with anti-NKG2D mAb (clone 149810) and quantified by comparing the mean fluorescence intensity (MFI) of the NKG2D staining. For blocking experiments of NKG2D expression, PBMC from healthy donors were incubated in 5% CO2 at 37°C for 24 h in the presence of pregnancy serum, pregnancy serum preincubated with anti-MICA/B Ab (10 μg/ml), serum from nonpregnant women, and medium alone. After incubation, the cells were stained for NKG2D as described above. In parallel, anti-CD3 staining was performed as a control of the specificity of the NKG2D blocking. A part of the cells incubated with pregnancy serum was fixed and permeabilized before staining. The fixation was done in 1.5% paraformaldehyde and 0.1% saponin for 20 min at room temperature. After washing, the aldehyde groups were blocked with 50 mM glycine in PBS for 15 min at room temperature. After centrifugation, the cells were resuspended in PBS with 0.1% saponin and allowed to permeabilize for 15 min at room temperature. Appropriate isotype controls were used. Five individual pregnancy sera from the first (n = 2) and second (n = 3) gestational trimester with the highest concentrations of sMIC as estimated by ELISA were tested in the blocking experiments.

Cell-mediated cytotoxicity was measured by Live/Dead Viability Cytotoxicity kit (Molecular Probes/Invitrogen Life Technologies). K562 target cells were labeled with the green fluorescence stain DiOC18 for 20 min at 37°C. After labeling, the target cells were mixed with the effector cells at E:T ratio of 20:1 or 40:1, counterstained by propidium iodine (PI), pelleted and incubated in 96-well plates at 37°C in a humidified 5% CO2 incubator for 2 h. Two-color scatter plot analysis separated live target cells (DiOC18+, PI) and membrane-compromised target cells (DiOC18+, PI+). The percent-specific lysis was calculated by a standard formula according to the manufacturer’s instructions. For blocking of the NKG2D receptor, the effector cells were preincubated with pregnancy serum, normal human serum, rMICA, or blocking anti-NKG2D mAb (clone 1D11) at a concentration of 10 μg/ml. For blocking of the surface MIC molecules, the target cells were incubated with goat anti-human MICA/B Ab.

We used the two-tailed Student t test for statistical analyses. A value of p < 0.05 was considered significant.

Isolated trophoblast cells of early pregnancy constitutively expressed mRNAs for MICA and MICB, and could be up-regulated by heat shock treatment in a time-dependent manner as detected by real-time RT-PCR (Fig. 1, A and C). Term placenta samples expressed mRNA for MICA at a similar level as early placenta, while MICB mRNA was expressed at lower levels (Fig. 1,B). To confirm the constitutive gene expression at the protein level, we stained samples from early and term placenta with goat Ab that recognize both MICA and MICB. To our knowledge, there are no commercial mAbs presently available that work in immunohistochemistry and discriminate between MICA and MICB. A positive staining was revealed in STB of early and term placenta (Fig. 2). Additionally, a weak staining of the wall of the blood vessels in the chorionic villi was observed (data not shown).

FIGURE 1.

Constitutive expression (□) and heat shock up-regulation (▪) of MICA and MICB mRNAs in human placenta. Quantitative real-time RT-PCR for MICA and MICB mRNA expression in freshly isolated and heat-shocked trophoblast cells from five early (A) and four term (B) placenta samples. C, Kinetic experiments showing time-dependent up-regulation of MICA and MICB mRNAs from early placenta (▦).

FIGURE 1.

Constitutive expression (□) and heat shock up-regulation (▪) of MICA and MICB mRNAs in human placenta. Quantitative real-time RT-PCR for MICA and MICB mRNA expression in freshly isolated and heat-shocked trophoblast cells from five early (A) and four term (B) placenta samples. C, Kinetic experiments showing time-dependent up-regulation of MICA and MICB mRNAs from early placenta (▦).

Close modal
FIGURE 2.

Expression of MIC proteins in early and term placenta. MICA/B-positive staining of STB in early (A) and term (C) placenta. Insets: STB staining at a higher magnification (long arrow). CTB does not stain for MICA/B (short arrow). Negative controls: early (B) and term (D) placenta. Original magnification: A–D, ×400; insets, ×800.

FIGURE 2.

Expression of MIC proteins in early and term placenta. MICA/B-positive staining of STB in early (A) and term (C) placenta. Insets: STB staining at a higher magnification (long arrow). CTB does not stain for MICA/B (short arrow). Negative controls: early (B) and term (D) placenta. Original magnification: A–D, ×400; insets, ×800.

Close modal

We determined the precise location of MIC molecules by immunoelectron microscopy (Fig. 3). There were no significant differences in the intensity of immunostaining and distribution pattern of MIC molecules in different villi from one placenta or among the three individual early placentas examined. The presence of MIC proteins was restricted to STB, confirming previous data that isolated CTB cells did not express MIC molecules (31). STB revealed a bipolar labeling along the apical and basal surfaces (Fig. 3, A–C). It was previously shown that apical MIC expression in polarized epithelial cells is associated with MICA*008/A5.1 allelic combination, i.e., the most common in several populations (32). A Swedish investigation determined that the frequency of this MIC allelic combination was 71% (33). Basolateral MIC staining has been described earlier in the intestinal epithelium (34). Additionally, numerous vacuoles in the syncytioplasm were stained for MIC molecules (Fig. 3,C). These vacuoles were ∼1.5–2 μm in size and had variable cell locations (Fig. 3, C–E). Some of them fused with the apical plasma membrane and formed an opening to the intervillous space (Fig. 3,D), others were in intimate contact with the nuclei (Fig. 3,E). At high magnification, the vacuoles had a multivesicular ultrastructure, displaying numerous stained intraluminal microvesicles of 60- to 90-nm diameter (Fig. 3,F). In some areas of the syncytioplasm, the rough endoplasmic reticulum (RER) and the Golgi complexes were stained (Fig. 3 G). Similar results were obtained when term placenta was stained (data not shown). Negative controls, omitting anti-MIC Abs revealed no detectable MIC staining (data not shown). Two conclusions could be drawn from our electron microscopy studies: 1) MIC molecules were synthesized and expressed on the apical and basal surfaces of STB; and 2) MIC molecules were also present in numerous syncytioplasmic vacuoles, structurally similar to multivesicular bodies/secretory lysosomes (35, 36).

FIGURE 3.

Immunoelectron microscopy of early placenta stained by an indirect immunoperoxidase method without additional counterstaining. A, Low-power micrograph of STB and adjacent CTB. The electron-dense reaction product stains the apical microvillous surface (arrows) and the basal membrane (arrowheads) of STB. B, Micrograph of a basal part of STB showing positive staining of the basal membrane (arrowheads) and cytoplasmic processes (arrows) those penetrate between two CTB cells (∗). BL, basal lamina. C, Micrograph demonstrating positive staining of the apical microvillous surface membrane (arrowheads) and cytoplasmic vacuoles (arrows). D, Micrograph of an apical portion of STB showing two labeled cytoplasmic vacuoles (arrows), one of which (thick arrow) forms an opening to the intervillous space (∗). Arrowheads indicate stained microvilli. E, Micrograph of a perinuclear area of STB demonstrating a number of positively stained cytoplasmic vacuoles (arrows). N, nucleus. F, High-power micrograph of a cytoplasmic vacuole, containing a granular material and numerous MICA/B-positive internal microvesicles (arrowheads). G, High-power micrograph demonstrating MICA/B staining of the Golgi complex (arrows, ∗). Arrowheads point to labeled ribosomes of the RER. Magnification: A, ×3,500; B, ×9,100; C, ×8,900; D, ×15,700; E, ×15,400; F, ×17,000; G, ×17,500.

FIGURE 3.

Immunoelectron microscopy of early placenta stained by an indirect immunoperoxidase method without additional counterstaining. A, Low-power micrograph of STB and adjacent CTB. The electron-dense reaction product stains the apical microvillous surface (arrows) and the basal membrane (arrowheads) of STB. B, Micrograph of a basal part of STB showing positive staining of the basal membrane (arrowheads) and cytoplasmic processes (arrows) those penetrate between two CTB cells (∗). BL, basal lamina. C, Micrograph demonstrating positive staining of the apical microvillous surface membrane (arrowheads) and cytoplasmic vacuoles (arrows). D, Micrograph of an apical portion of STB showing two labeled cytoplasmic vacuoles (arrows), one of which (thick arrow) forms an opening to the intervillous space (∗). Arrowheads indicate stained microvilli. E, Micrograph of a perinuclear area of STB demonstrating a number of positively stained cytoplasmic vacuoles (arrows). N, nucleus. F, High-power micrograph of a cytoplasmic vacuole, containing a granular material and numerous MICA/B-positive internal microvesicles (arrowheads). G, High-power micrograph demonstrating MICA/B staining of the Golgi complex (arrows, ∗). Arrowheads point to labeled ribosomes of the RER. Magnification: A, ×3,500; B, ×9,100; C, ×8,900; D, ×15,700; E, ×15,400; F, ×17,000; G, ×17,500.

Close modal

The expression of MIC molecules by placenta implied the presence of STB-derived sMIC molecules in the maternal circulation. To test this suggestion, we performed a capture ELISA for sMIC proteins (Fig. 4,B) using sera from pregnant women in the first (8–11 wk, n = 12), second (17–20 wk, n = 11), and third trimester (33–38 wk, n = 7); and nonpregnant women (n = 5). Soluble MIC molecules were present in all pregnancy sera from all trimesters (n = 30) in significantly higher levels compared with normal controls (Fig. 4,B). The measured quantity of sMIC proteins was in the range from 2,100 to 20,000 pg/ml with a mean value of 6,100, 10,000, and 4,700 pg/ml in the first, second, and third trimester, respectively. The highest sMIC levels in the controls were 32 pg/ml with a mean value of 26 pg/ml (Fig. 4 B). Thus, we could conclude that significantly elevated levels of sMIC proteins are present in the serum of pregnant women throughout pregnancy.

FIGURE 4.

NKG2D on PBMC of pregnant women is down-regulated by sMIC proteins. A, Immunoflow cytometry of NKG2D-stained PBMC from pregnant women shows significantly decreased MFI compared with nonpregnant controls (n = 9, p = 0.004). B, Capture ELISA for sMICA/B with pregnancy sera from first (8–11 wk, n = 12,), second (17–20 wk, n = 11), and third trimesters (33–38 wk, n = 7); and nonpregnant sera (n = 5) shows significantly increased amounts of sMICA/B compared with controls. ROD, relative OD.

FIGURE 4.

NKG2D on PBMC of pregnant women is down-regulated by sMIC proteins. A, Immunoflow cytometry of NKG2D-stained PBMC from pregnant women shows significantly decreased MFI compared with nonpregnant controls (n = 9, p = 0.004). B, Capture ELISA for sMICA/B with pregnancy sera from first (8–11 wk, n = 12,), second (17–20 wk, n = 11), and third trimesters (33–38 wk, n = 7); and nonpregnant sera (n = 5) shows significantly increased amounts of sMICA/B compared with controls. ROD, relative OD.

Close modal

To confirm that the sMIC molecules present in the serum of pregnant women are placenta-derived, explants from early normal pregnancy placentae of approximately the same gestational age (11–12 wk) were cultured for 24 h. The supernatant from the individual explant cultures was collected and analyzed for presence of sMIC by ELISA. The results from the three placental samples tested are summarized in Table I. As can be seen, all individual placentae released sMIC molecules at similar levels.

Table I.

Release of sMIC molecules into first trimester placental explant culture media

Sample No.No. of explantsApproximate Wet Weight/Explant (mg)sMIC Concentration (ng/ml)
10 15.5 ± 0.9 
10 25.4 ± 8.0 
10 20.0 ± 4.8 
Sample No.No. of explantsApproximate Wet Weight/Explant (mg)sMIC Concentration (ng/ml)
10 15.5 ± 0.9 
10 25.4 ± 8.0 
10 20.0 ± 4.8 

Given the previous finding of sMIC-induced down-regulation of NKG2D on effector cells (24, 27), we next investigated the expression of NKG2D receptor on PBMC from pregnant women (n = 9). The MFI of NKG2D expression was markedly reduced by ∼55% (p = 0,005, Fig. 4 A) compared with nonpregnant controls. The decrease was not selective for a particular subset of PBMC, but was observed on NK, γδ T, and CD8+αβ T cells (not shown), indicating a general down-regulation of NKG2D.

To test whether sMIC molecules but not other soluble factors were responsible for the reduced NKG2D surface expression on PBMC of pregnant women, we isolated PBMC from healthy donors (n = 5) and assessed the NKG2D expression before and after incubation with individual pregnancy serum, normal serum, or individual pregnancy serum, pretreated with MICA/B Ab. The results of one representative experiment are shown in Fig. 5. NKG2D surface expression declined 8- to 10-fold after incubation with pregnancy serum compared with PBMC incubated with normal serum or untreated PBMC (Fig. 5, A and C). In contrast, the surface staining of CD3 remained unchanged. When pregnancy serum was pretreated with MICA/B Ab, the NKG2D expression was not down-regulated (Fig. 5,B), suggesting that sMIC molecules were the main factor that reduced cell surface NKG2D expression. Pregnancy sera with the highest sMIC concentration were chosen from first and second gestational trimester and used individually in the blocking experiments. There was a tendency of higher down-modulation of NKG2D when the PBMC were treated with serum samples with higher sMIC concentration (data not shown). To determine whether NKG2D was internalized or just masked by the soluble ligand, a part of the PBMC that was incubated with pregnancy serum was fixed, permeabilized, and analyzed for the presence of intracellular NKG2D. The staining, reflecting the total NKG2D expression, showed MFI comparable or slightly reduced to that of freshly isolated PBMC, indicating that the cell surface down-modulation of NKG2D was caused by sMIC-induced endocytosis (Fig. 5 D). These results are in contrast to the previously published data showing strong reduction of total NKG2D expression after ligand engagement in T cell clones (24) but in line with the results of Ogasawara et al. (37) demonstrating that the total amount NKG2D remained unchanged. The reason for these discrepancies cannot be ascertained from our experiments.

FIGURE 5.

Pregnancy serum down-modulates NKG2D expression on PBMC from healthy donors by ligand-induced endocytosis. Green-colored histograms represent NKG2D staining of untreated PBMC and gray-colored histograms represent negative control. Filled red histograms represent PBMC incubated with (A) pregnancy serum, (B) pregnancy serum, blocked with MICA/B Ab, (C) nonpregnant serum, and (D) PBMC incubated with pregnancy serum, fixed, and permeabilized before staining. Blue-colored histograms show anti-CD3 staining.

FIGURE 5.

Pregnancy serum down-modulates NKG2D expression on PBMC from healthy donors by ligand-induced endocytosis. Green-colored histograms represent NKG2D staining of untreated PBMC and gray-colored histograms represent negative control. Filled red histograms represent PBMC incubated with (A) pregnancy serum, (B) pregnancy serum, blocked with MICA/B Ab, (C) nonpregnant serum, and (D) PBMC incubated with pregnancy serum, fixed, and permeabilized before staining. Blue-colored histograms show anti-CD3 staining.

Close modal

Next, we assessed whether the NKG2D down-regulation by sMIC on PBMC had functional consequences for their killing ability (Fig. 6). The MIC-expressing target cells K562 were used at T:E ratio of 40:1 and 20:1. The PBMC were used directly, or after pretreatment with sMIC containing pregnancy serum, normal serum, NKG2D blocking mAb, or rMICA. Additionally, the target cells were pretreated with anti-MICA/B Ab to see whether blocking of MIC molecules could affect the specific cytotoxicity. In the presence of pregnancy serum, the lysis of target cells was substantially decreased. The reduction of the specific lysis was similar to that seen when the effector cells were pretreated with anti-NKG2D mAb or with rMICA, or when MICA/B Ab blocked the MIC molecules on K562 cells. Normal human serum did not reduce cytotoxicity. These results show that down-regulation of NKG2D by sMIC molecules in pregnancy serum impairs the in vitro cytotoxic ability of PBMC.

FIGURE 6.

Down-regulation of the in vitro killing ability of PBMC from healthy donors by pregnancy serum. Cytotoxicity against K562 cells was tested in an E:T ratio of 40:1 and 20:1 with untreated PBMC or in the presence of normal serum, pregnancy serum, anti-NKG2D mAb, rMICA, and anti-MICA/B Ab. Pregnancy serum significantly decreased the cytotoxicity of untreated effector cells to a degree comparable to treatment with anti-NKG2D mAb, rMICA, or anti-MICA/B Ab as measured by percentage of lysis. Normal serum did not affect cytotoxicity.

FIGURE 6.

Down-regulation of the in vitro killing ability of PBMC from healthy donors by pregnancy serum. Cytotoxicity against K562 cells was tested in an E:T ratio of 40:1 and 20:1 with untreated PBMC or in the presence of normal serum, pregnancy serum, anti-NKG2D mAb, rMICA, and anti-MICA/B Ab. Pregnancy serum significantly decreased the cytotoxicity of untreated effector cells to a degree comparable to treatment with anti-NKG2D mAb, rMICA, or anti-MICA/B Ab as measured by percentage of lysis. Normal serum did not affect cytotoxicity.

Close modal

Four main points summarize the results presented here: 1) MIC molecules are constitutively transcribed in human placenta throughout normal pregnancy and sMIC molecules are released from in vitro cultured placental explants. 2) The MIC protein expression is restricted to STB, and exists in two forms–as a membranal form at the apical and basal surfaces, and as microvesicles in cytoplasmic vacuoles. 3) The cell surface expression of NKG2D receptors is significantly down-regulated on PBMC from pregnant women. 4) Elevated levels of sMIC molecules are present in pregnancy sera and are able to down-regulate the NKG2D receptor and impair the cytotoxic function of PBMC from healthy donors. Taken together, these results suggest a novel mechanism for immune evasion of the fetal allograft through fetal MIC and maternal NKG2D interactions.

The cell surface expression of MIC molecules on STB exposes the fetoplacental unit to maternal cytotoxic attack. It would make little sense, from the point of view of the fetus, to display molecules on the placenta that would identify it as abnormal and provoke rejection by the maternal immune system unless it gives the fetus an advantage. The permanent secretion/shedding of MIC molecules that we found in pregnancy explains this paradox. A similar mechanism of sMIC shedding is used by MIC-expressing tumors as immune escape strategy (24, 25, 26, 27, 38). It has been shown that sMIC molecules shed by these tumors block NKG2D and impair the cytotoxic function of PBMC in cancer patients. In analogy with tumors, sMIC molecules present in pregnancy serum were able to down-regulate the NKG2D surface expression on PBMC and their killing ability. Placenta may be compared with a tumor; highly proliferative and invasive trophoblast cells penetrate the decidua and the uterine vasculature to build the placenta. These cells, like tumor cells, use similar molecular mechanisms for growth and migration. In both instances, changes in expression of adhesion molecules, proto-oncogene products and matrix-digestive enzymes occur, ensuring a rich blood supply (39, 40). Thus, human placenta can be regarded as a physiological counterpart of a tumor, and might use similar immune escape strategies. Alternatively, tumors may use evolutionary ancient immune evasion mechanism primarily used by placenta. Except in tumors, expression of cell surface and sMIC has been observed in a variety of pathological conditions such as viral and bacterial infections, inflammation, and autoimmunity (18, 19, 20, 21). To our knowledge, pregnancy is so far the only example of a normal physiological condition that takes advantage of this adverse phenomenon and uses it to promote fetal allograft survival. However, the cost of using this escape mechanism might be the impairment of systemic maternal immune responses due to down-regulation of the NKG2D receptor and might be accountable for the observed fact that pregnant women are more sensitive to infections.

Very little is known about the intracellular localization or trafficking of MIC molecules (34). Our electron microscopy data throw some light on these events. The localization of MIC proteins in the RER and Golgi complex indicates their site of synthesis and maturation in STB. However, the major intracellular protein fraction was localized in numerous cytoplasmic vacuoles filled with MIC-positive microvesicles. There are two options for the identity of the MIC-containing vacuoles: they may either represent endocytosed microvesicle-associated complexes or newly synthesized MIC molecules in the so-called secretory lysosomes (41). We favor the latter explanation for the following reasons: 1) the multivesicular body-like structure of the vacuoles and the size of the microvesicles classify them as secretory lysosomes, i.e., organelles used for temporary storage of secreted proteins (41). 2) We used anti-MICA/B Ab directed to the extracellular domain of MIC molecules, thus, because the Ab stained intracellularly, they would be recognizing newly synthesized molecules that come directly from the synthetic pathway. If the vacuoles contained processed MIC proteins, i.e., endocytosed from the surface, the extracellular domain would have been cleaved by matrix metalloproteinases (MMP; Ref.25), and we would have failed to stain the microvesicles. Therefore, the MIC-loaded microvesicles represent storage of newly synthesized protein in secretory lysosomes. The vacuoles were localized at different levels in the syncytioplasm, some reaching out and fusing with the apical surface membrane of STB. It has been shown that secretory lysosomes can fuse with the plasma membrane, resulting in release of the internal microvesicles into the external milieu. The released vesicles are then called exosomes (42). Convincing data indicate that intracellular microvesicles and exosomes are one and the same entity (43). Thus, we can suggest a release of MIC-containing exosomes from the placental villi to the maternal blood. This suggestion is supported by a recent report that MIC molecules were found in exosomes isolated from the growth medium of tumor cells and were able to impair cytotoxicity of NKG2D-positive effector cells (44). Exosome excretion is not only a well-known feature of tumor cells (45), but also is a feature of the placenta, and is proven to be the form by which FasL is secreted by STB in the maternal circulation (8, 9). The alternative to exosomal MIC release is MMP-mediated shedding that has been proposed as a source of sMIC molecules in cancer patients (26). Because human placenta expresses high amounts of a variety of MMP (46), we cannot exclude the possibility that surface-cleaved MIC could also be a part of the ELISA-detected sMIC molecules. Thus, the serum MIC moiety in pregnant women might constitute a heterogeneous population of MMP-cleaved-, and exosome-released MIC molecules. Experiments are in progress to elucidate this assumption.

The immune escape mechanism described here concerns human normal pregnancy. An obvious question is how the MIC/NKG2D system works in pathological pregnancies threatened by abortion. We could up-regulate MIC mRNA expression in isolated villous trophoblast cells in vitro by thermal stress. Pathological conditions that stress the pregnancy should up-regulate the expression of MIC molecules, and more sMIC should be released and down-modulate NKG2D-mediated cytotoxicity. At the same time, TNF-α and IL-15, known to induce NKG2D overexpression on T and NK cells (21), are produced in high amounts by placenta in pathologic pregnancies and abortions (47). Recently, it was shown in rheumatoid arthritis patients that synoviocyte-derived sMIC molecules failed to induce down-modulation of NKG2D due to the opposing effect of abundant amounts of TNF-α and IL-15 in inflamed synovia (21). A similar dysregulation of the MIC/NKG2D system may prevail in pathologic pregnancies. In contrast, a normally proceeding pregnancy in rheumatoid arthritis patient will supply additional amounts of circulating sMIC molecules, and thus partly overcome the effect of the synovial TNF-α and IL-15 on NKG2D overexpression. Moreover, high amounts of TGF-β1, known to down-regulate NKG2D (48), are expressed in decidua (6). These events could partially explain why the course of rheumatoid arthritis ameliorates during pregnancy and deteriorates after delivery. Investigation of the MIC/NKG2D system in abnormal pregnancies and other reproductive disorders may give new insights in their pathogenesis and the possibility to design suitable immune therapy.

The immune escape mechanism described here clearly differs from that of tumors. In pregnancy, the counterparts in the system, the MIC ligands and the NKG2D receptor, come from two different individuals; i.e., an allogeneic situation exists, while in tumors the system is syngeneic. In this respect, the sMIC-NKG2D interaction-based immune escape mechanism in pregnancy may be of particular interest to transplantation immunology, and open novel possibilities to modulate cytotoxic responses and promote transplant survival.

We gratefully acknowledge the donors and colleagues at the Department of Obstetrics and Gynecology at Umeå University Hospital for kindly providing samples.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by grants from the Swedish National Cancer Research Foundation Cancerfonden (4565-B03-03XAC), and the Cancer Research Foundation of Norrland, Sweden (AMP 01-247).

3

Abbreviations used in this paper: STB, syncytiotrophoblast; CTB, cytotrophoblast; MIC, MHC class I chain; MICA/B, MIC-related proteins A and B; sMIC, soluble MIC; MFI, mean fluorescence intensity; PI, propidium iodide; RER, rough endoplasmic reticulum; MMP, matrix metalloproteinase.

1
Georgiades, P., M. Watkins, G. J. Burton, A. C. Ferguson-Smith.
2001
. Roles for genomic imprinting and the zygotic genome in placental development.
Proc. Natl. Acad. Sci. USA
98
:
4522
-4527.
2
Sacks, G., I. Sargent, C. Redman.
1999
. An innate view of human pregnancy.
Immunol. Today
20
:
114
-118.
3
Tafuri, A., J. Alferink, P. Moller, G. J. Hammerling, B. Arnold.
1995
. T cell awareness of paternal alloantigens during pregnancy.
Science
270
:
630
-633.
4
Mellor, A. L., D. H. Munn.
2000
. Immunology at the maternal-fetal interface: lessons for T cell tolerance and suppression.
Annu. Rev. Immunol.
18
:
367
-391.
5
Moffett-King, A..
2002
. Natural killer cells and pregnancy.
Nat. Rev. Immunol.
2
:
656
-663.
6
Nagaeva, O., L. Jonsson, L. Mincheva-Nilsson.
2002
. Dominant IL-10 and TGF-β mRNA expression in γδT cells of human early pregnancy decidua suggests immunoregulatory potential.
Am. J. Reprod. Immunol.
48
:
9
-17.
7
Aluvihare, V. R., M. Kallikourdis, A. G. Betz.
2004
. Regulatory T cells mediate maternal tolerance to the fetus.
Nat. Immunol.
5
:
266
-271.
8
Abrahams, V. M., S. L. Straszewski-Chavez, S. Guller, G. Mor.
2004
. First trimester trophoblast cells secrete Fas ligand which induces immune cell apoptosis.
Mol. Hum. Reprod.
10
:
55
-63.
9
Frängsmyr, L., V. Baranov, O. Nagaeva, U. Stendahl, L. Kjellberg, L. Mincheva-Nilsson.
2005
. Cytoplasmic microvesicular form of Fas ligand in human early placenta: switching the tissue immune privilege hypothesis from cellular to vesicular level.
Mol. Hum. Reprod.
11
:
35
-41.
10
Morgan, B. P., C. H. Holmes.
2000
. Immunology of reproduction: protecting the placenta.
Curr. Biol.
10
:
R381
-R383.
11
Ostensen, M., P. M. Villiger.
2002
. Immunology of pregnancy-pregnancy as remission inducing agent in rheumatoid arthritis.
Transplant. Immunol.
9
:
155
-160.
12
Diefenbach, A., D. H. Raulet.
2002
. The innate immune response to tumors and its role in the induction of T-cell immunity.
Immunol. Rev.
188
:
9
-21.
13
Raulet, D. H..
2003
. Roles of the NKG2D immunoreceptor and its ligands.
Nat. Rev. Immunol.
3
:
781
-790.
14
Bahram, S..
2000
. MIC genes: from genetics to biology.
Adv. Immunol.
76
:
1
-60.
15
Groh, V., S. Bahram, S. Bauer, A. Herman, M. Beauchamp, T. Spies.
1996
. Cell stress-regulated human major histocompatibility complex class I gene expressed in gastrointestinal epithelium.
Proc. Natl. Acad. Sci. USA
93
:
12445
-12450.
16
Bauer, S., V. Groh, J. Wu, A. Steinle, J. H. Phillips, L. L. Lanier, T. Spies.
1999
. Activation of NK cells and T cells by NKG2D, a receptor for stress-inducible MICA.
Science
285
:
727
-729.
17
Groh, V., R. Rhinehart, H. Secrist, S. Bauer, K. H. Grabstein, T. Spies.
1999
. Broad tumor-associated expression and recognition by tumor-derived γδT cells of MICA and MICB.
Proc. Natl. Acad. Sci. USA
96
:
6879
-6884.
18
Groh, V., R. Rhinehart, J. Randolph-Habecker, M. S. Topp, S. R. Riddel, T. Spies.
2001
. Costimulation of CD8αβ T cells by NKG2D via engagement by MIC induced on virus-infected cells.
Nat. Immunol.
2
:
255
-260.
19
Das, H., V. Groh, C. Kuiji, M. Sugita, C. T. Morita, T. Spies, J. F. Bukowski.
2001
. MICA engagement by human Vγ2Vδ2 T cells enhances their antigen-dependent effector function.
Immunity
15
:
83
-93.
20
Tieng, V., C. Le Bouguenec, L. du Merle, P. Bertheau, P. Desreumaux, A. Janin, D. Charron, A. Toubert.
2002
. Binding of Escherichia coli adhesin AfaE to CD55 triggers cell-surface expression of the MHC class I-related molecule MICA.
Proc. Natl. Acad. Sci. USA
99
:
2977
-2982.
21
Groh, V., A. Bruhl, H. El-Gabalawy, J. L. Nelson, T. Spies.
2003
. Stimulation of T cell autoreactivity by anomalous expression of NKG2D and its MIC ligands in rheumatoid arthritis.
Proc. Natl. Acad. Sci. USA
100
:
9452
-9457.
22
Girardi, M., D. E. Oppenheim, C. R. Steele, J. M. Lewis, E. Glusac, R. Filler, P. Hobby, B. Sutton, R. E. Tigelaar, A. Hayday.
2001
. Regulation of cutaneous malignancy by γδ T cells.
Science
294
:
605
-609.
23
Diefenbach, A., D. H. Raulet.
2001
. Strategies for target cell recognition by natural killer cells.
Immunol. Rev.
181
:
170
-184.
24
Groh, V., J. Wu, C. Yee, and T. Spies, T. 2002. Tumor-derived soluble MIC ligands impair expression of NKG2D and T-cell activation. Nature 419: 734–738.
25
Salih, H. R., H. G. Rammensee, A. Steinle.
2002
. Cutting edge: down-regulation of MICA on human tumors by proteolytic shedding.
J. Immunol.
169
:
4098
-4102.
26
Salih, H. R., H. Antropius, F. Gieseke, S. Z. Lutz, L. Kantz, H. G. Ramensee, A. Steinle.
2003
. Functional expression and release of ligands for the activating immunoreceptor NKG2D in leukemia.
Blood
102
:
1389
-1396.
27
Doubrovina, E. S., M. M. Doubrovin, E. Vider, R. B. Sisson, R. J. O′Reilly, B. Dupont, Y. M. Vyas.
2003
. Evasion from NK cell immunity by MHC class I chain-related molecules expressing colon adenocarcinoma.
J. Immunol.
171
:
6891
-6899.
28
Romanski, A., G. Bug, S. Becker, M. Kampfmann, E. Seifried, D. Hoelzer, O. G. Ottmann, T. Tonn.
2005
. Mechanisms of resistance to natural killer cell-mediated cytotoxicity in acute lymphoblastic leukemia.
Exp. Hematol.
33
:
344
-352.
29
Siman, C. M., C. P. Sibley, C. J. P. Jones, M. A. Turner, S. L. Greenwood.
2001
. The functional regeneration of syncytiotrophoblast in cultured explants of term placenta.
Am. J. Physiol.
280
:
R1116
-R1122.
30
Molinero, L. L., M. B. Fuertes, M. V. Girart, L. Fainboim, G. A. Rabinovich, M. A. Costas, N. W. Zwirner.
2004
. NF-κB regulates expression of the MHC class I-related chain A gene in activated T lymphocytes.
J. Immunol.
173
:
5583
-5590.
31
Menier, C., B. Riteau, E. D. Carosella, N. Rouas-Freiss.
2002
. MICA triggering signal for NK cell tumor lysis is counteracted by HLA-G1-mediated inhibitory signal.
Int. J. Cancer
100
:
63
-70.
32
Stephens, H. A..
2001
. MICA and MICB genes: can the enigma of their polymorphism be resolved?.
Trends Immunol.
22
:
378
-385.
33
Ghaderi, M., P. Hjelmstrom, G. Hallmans, F. Wiklund, P. Lenner, J. Dillner, C. B. Sanjeevi.
1999
. MICA gene polymorphism and the risk to develop cervical intraepithelial neoplasia.
Hum. Immunol.
60
:
970
-973.
34
Suemizu, H., M. Radosavljevic, M. Kimura, S. Sadahiro, S. Yoshimura, S. Bahram, H. Inoko.
2002
. A basolateral sorting motif in the MICA cytoplasmic tail.
Proc. Natl. Acad. Sci. USA
99
:
2971
-2976.
35
Denzer, K., M. J. Kleijmeer, H. F. Heijnen, W. Stoorvogel, H. J. Geuze.
2000
. Exosome: from internal vesicle of the multivesicular body to intercellular signaling device.
J. Cell Sci.
113
:
3365
-3374.
36
Fevrier, B., G. Raposo.
2004
. Exosomes: endosomal-derived vesicles shipping extracellular messages.
Curr. Opin. Cell Biol.
16
:
415
-421.
37
Ogasawara, K., J. A. Hamerman, H. Hsin, S. Chikuma, H. Bour-Jordan, T. Chen, T. Pertel, C. Carnaud, J. A. Bluestone, L. L. Lanier.
2003
. Impairment of NK cell function by NKG2D modulation in NOD mice.
Immunity
18
:
41
-51.
38
Raffaghello, L., I. Prigione, I. Airoldi, M. Canoriano, I. Levreri, C. Gambimi, D. Pende, A. Steinle, S. Ferrone, V. Pistoia.
2004
. Downregulation and/or release of NKG2D ligands as immune evasion strategy of human neuroblastoma.
Neoplasia
6
:
558
-568.
39
Rama, S., A. J. Rao.
2003
. Regulation of growth and function of the human placenta.
Mol. Cell. Biochem.
253
:
263
-268.
40
Red-Horse, K., Y. Zhou, O. Genbacev, A. Prakobhol, R. Foulk, M. McMaster, S. J. Fisher.
2004
. Trophoblast differentiation during embryo implantation and formation of the maternal-fetal interface.
J. Clin. Invest.
114
:
744
-754.
41
Blott, E. J., G. M. Griffiths.
2002
. Secretory lysosomes.
Nat. Rev. Cell Biol.
3
:
122
-131.
42
Farsad, K..
2002
. Exosomes: novel organelles implicated in immunomodulation and apoptosis.
Yale J. Biol. Med.
75
:
95
-101.
43
Taylor, D. D., C. Gercel-Taylor.
2005
. Tumor-derived exosomes and their role in cancer-associated T-cell signaling defects.
Br. J. Cancer
92
:
305
-311.
44
Clayton, A., Z. Tabi.
2005
. Exosomes and MICA-NKG2D system in cancer.
Blood Cells Mol. Dis.
34
:
206
-213.
45
Whiteside, T. L..
2005
. Tumor-derived exosomes or microvesicles: another mechanism of tumor escape from the host immune system?.
Br. J. Cancer
92
:
209
-211.
46
Bischof, P., A. Meisser, A. Campana.
2000
. Paracrine and autocrine regulators of trophoblast invasion—a review.
Placenta
21
:
S55
-S60.
47
Chegini, N., C. Ma, M. Roberts, R. S. Williams, B. A. Ripps.
2002
. Differential expression of interleukins (IL) IL-13 and IL-15 throughout the menstrual cycle in endometrium of normal fertile women and women with recurrent spontaneous abortion.
J. Reprod. Immunol.
56
:
93
-110.
48
Lee, J. C., K. M. Lee, D. W. Kim, D. S. Heo.
2004
. Elevated TGF-β1 secretion and down-modulation of NKG2D underlies impaired NK cytotoxicity in cancer patients.
J. Immunol.
172
:
7335
-7340.