Trophoblasts are fetal epithelial cells that form an interface between mother and offspring. To evaluate their anti-inflammatory capacity, we tested the hypothesis that trophoblasts deactivate neutrophils using single-cell assays. Several biophysical (Ca2+ and NAD(P)H oscillation frequency) and physiological (oxidant production) markers of activated neutrophils revert to a nonactivated phenotype as activated cells make contact with trophoblasts. Indistinguishable results were obtained using syncytiotrophoblasts and in experiments using trophoblasts and neutrophils from the same mother to recapitulate the semiallogeneic system. These changes suggest reduced hexose monophosphate shunt (HMS) activity. We discovered that two metabolic regulatory points, glucose transport and HMS enzyme trafficking, are affected by trophoblasts. This restriction in HMS activity deactivates neutrophils, thereby limiting oxidative DNA damage within trophoblasts.
Successful pregnancy remains one of the great paradoxes of immunology: the fetus survives despite its semiallograft status. One likely site of immunoregulation is the placenta, which forms an interface such that the maternal immune system interacts directly with the placenta, not the fetus. The principal placental cell type interacting with the maternal immune system is the trophoblast. Although the mechanisms allowing the immune system to spare the fetus are not understood, it is becoming clear that trophoblasts are important participants. Trophoblasts exhibit some characteristics in common with immune cells, such as their ability to migrate, phagocytose, and invade tissues (1, 2, 3). These cells also express complement regulatory proteins, cytokine receptors, TLRs, FcRn, and proteases (4, 5, 6, 7, 8, 9). The expression of complement receptor-1-related protein y is important in the maintenance of rodent pregnancy due to its ability to inhibit the deposition of complement at the materno-fetal interface (9). Trophoblasts do not express conventional HLA class I Ags, but express HLA-G, which promotes tolerance (10, 11). In addition to these rather passive defense strategies to deflect immune responses, trophoblasts may also actively modify the mother’s immune response. For example, it is widely known that trophoblasts produce and secrete cytokines, largely Th2 cytokines, that promote the Th2-biased phenotype observed in pregnant women (12). The Th1 to Th2 shift during pregnancy is thought to be crucial for normal pregnancy, because proinflammatory responses associated with Th1 cytokines have been associated with spontaneous abortion and preterm labor (12). Other active mechanisms of maternal leukocyte regulation are suggested by studies showing that the conceptus is protected from destruction by activated macrophages (13) and that neutrophil activation is depressed during pregnancy (14).
To better understand neutrophil activation in pregnancy, we have analyzed the phenotypic properties of peripheral blood cells from pregnant women. We have recently shown that neutrophils from pregnant women decrease their production of oxidants by translocating enzymes of the hexose monophosphate shunt (HMS)3 from the plasma membrane to the centrosome, thereby limiting the availability of glucose-6-phosphate, which is produced at the cell periphery, and decreasing HMS activity and oxidant production (14, 15, 16). In this study we use the opposite approach to characterize neutrophil regulation in pregnancy; we activate cells from nonpregnant individuals in vitro, then watch as they interact with trophoblasts. Using highly sensitive single-cell assays to follow these intercellular interactions, we show that neutrophils rapidly deactivate upon trophoblast contact. Our results suggest that trophoblasts accomplish this by influencing both glucose transport and the intracellular location of HMS enzyme complexes, which reduces the neutrophil’s ability to make reactive oxygen metabolites (ROMs) and damage neighboring cells.
Materials and Methods
Peripheral blood was collected from nonpregnant and pregnant women after written informed consent was provided. Placental samples were obtained after childbirth, with written informed consent. These studies were approved by the Institutional Review Board of the participating institutions. The nonpregnant group consisted of women in the secretory phase of the menstrual cycle who were not taking oral contraceptives and who had no history of acute or chronic inflammatory conditions. The pregnant group was made up of women with normal pregnancies. Eligible patients were approached at the Detroit Medical Center (Detroit, MI).
2TMRos) fluorescein-conjugated N-formyl-Nle-Leu-Phe-Nle-Tyr-Lys (Fl-FNLPNTL), H2-2′,7′dichlorofluorescein (CDCF), and 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (2-NBDG) were obtained from Molecular Probes. Rabbit anti-glucose-6-phosphate dehydrogenase (G-6-PDase) and goat anti-lactate dehydrogenase polyclonal Ab were obtained from Chemicon International. Anti-transaldolase was obtained from Nordic Immunological Laboratories.
Isolation of placental tissue and cells
Placental villi, which are covered with syncytiotrophoblasts, were dissected from placental tissue, washed thoroughly with HBSS, then used immediately for experimentation. Primary trophoblast cultures were obtained as previously described (4
Neutrophils were placed on glass coverslips, incubated with reagents as described below, then fixed as previously described (14). Briefly, cells were fixed with 2% paraformaldehyde, permeabilized with 1% Brij-58, then fixed a second time with 2% paraformaldehyde at room temperature for 20 min. Cells were washed with HBSS, labeled with 1 μg of FITC-conjugated Abs at 4°C for 30 min., then washed again with HBSS at room temperature.
Cells were observed using an Axiovert fluorescence microscope (Zeiss) with mercury illumination interfaced to a computer using Scion image processing software (17). A narrow bandpass-discriminating filter set (Omega Optical) was used with excitation at 485/22 nm and emission at 530/30 nm for FITC, and excitation at 540/20 nm and emission at 590/30 nm for tetramethylrhodamine isothiocyanate. Long-pass dichroic mirrors of 510 and 560 nm were used for FITC and tetramethylrhodamine isothiocyanate, respectively. The fluorescence images were collected with intensified, charge-coupled device cameras (Princeton Instruments and Q Imaging).
Ca2+ and NAD(P)H oscillations
Ca2+ and NAD(P)H oscillations were monitored at 37°C as previously described (18, 19, 20). Indo 1 was used to detect Ca2+ transients, whereas unlabeled cells were used to measure NAD(P)H oscillations. For NAD(P)H autofluorescence, a 365WB50 excitation filter, a 400-nm long-pass dichroic mirror, and a 450AF58 emission filter were used. In the case of indo 1 detection, an emission filter at 405DF43 was used. A cooled, high sensitivity photomultiplier tube in a D104 detection system (Photon Technology International) attached to a Zeiss microscope was used. Data were analyzed using Felix software (Photon Technologies). Calcium data were smoothed using a Savitzky-Golay function with a buffer size of seven.
Detection of ROMs
Pericellular ROM release from single cells was monitored as previously described (21). Briefly, adherent neutrophils were surrounded in 2% gelatin containing 100 ng/ml H2TMRos (Molecular Probes). ROMs, especially H2O2, released by the cells entered the gelatin matrix, where they oxidized H2TMRos to TMRos, which was detected by fluorescence microscopy. Experiments to detect intracellular ROMs were performed as previously described (22).
Detection of DNA damage
Trophoblasts were grown overnight on glass coverslips. For ROM detection, trophoblasts were labeled with H2TMRos as described above. Neutrophils in suspension were mixed with 50 nM FMLP, then placed on subconfluent trophoblast cultures for 1.5 h at 37°C. Coverslips were gently washed, followed by collection of differential interference contrast (DIC) and TMRos images. Coverslips were removed, then fixed with 3.7% paraformaldehyde. After cells were permeabilized with 70% ethanol for 30 min, TUNEL staining was performed as recommended by the manufacturer (Upstate Biotechnology). Coverslips were then remounted on slides, and the same regions were photographed. Images of TMRos and TUNEL staining were pseudocolored and overlaid using Image-Pro Plus (Media Cybernetics). For quantitation of DNA damage, 1.2 × 105 neutrophils were added to subconfluent trophoblasts at a 2:1 ratio.
Single-cell features of activated neutrophils
Two important features of neutrophil activation are changes in Ca2+ signaling (23, 24, 25, 26) and metabolism, especially activation of the HMS and the accompanying respiratory burst (15, 18). Fig. 1⇓ illustrates the relationships between neutrophil physiology and metabolic dynamics. Unstimulated spherical neutrophils displayed low-amplitude oscillations, with a period of ∼3 min. (Fig. 1⇓, A and B), which probably reflects the dynamic properties of phosphofructokinase. When cells were morphologically polarized, which is generally associated with cell migration, metabolic oscillations with a period of ∼20 s were observed (Fig. 1⇓, C and D). After activation with 100 nM FMLP, oscillations with a period of ∼10 s were observed (Fig. 1⇓, E and F), which may be due to the diversion of carbon away from the glycolytic pathway to the shunt. Parallel changes in the Ca2+ signaling pathway have also been observed (18). These noninvasive readouts were used to probe the physiological status of neutrophils in the presence of trophoblasts.
Trophoblast contact alters Ca2+ signaling in activated neutrophils
In view of the trophoblast’s potential immunoregulatory capacity, we have tested the hypothesis that trophoblasts directly regulate the activation status of neutrophils. We therefore examined the trophoblast’s ability to affect the neutrophil’s metabolic and signal transduction apparatus. We studied Ca2+ signaling using microfluorometry of single activated neutrophils (100 nM FMLP at 37°C) before and after contact with JEG-3 trophoblasts (Fig. 2⇓, A–C). Previous studies from this group have shown that unstimulated, adherent neutrophils display Ca2+ spikes at ∼20-s intervals, whereas FMLP-activated neutrophils exhibit Ca2+ spikes at 10-s intervals (18, 19), which parallels the metabolic changes noted in Fig. 1⇑. Morphologically polarized and activated neutrophils also displayed Ca2+ spikes at ∼10-s intervals when mixed with trophoblasts (Fig. 2⇓). At the time scale used in Fig. 2⇓C, the spikes appeared featureless, because the number of data points per spike was small; when data were collected over a short time scale (a large number of pixels per spike), there was a shoulder at short times and a decay after the peak at longer time periods. However, the important point is that the high-frequency oscillations returned to normal (∼20-s period) after neutrophil-trophoblast contact. Fig. 2⇓C illustrates the shift in Ca2+ spike frequency, whereas A and B document the contact of the neutrophil with the trophoblast. Similar results were obtained for both JEG-3 and JAR (data not shown) trophoblast cell lines. However, the HTR-8/SVneo trophoblasts, a model of invasive trophoblasts, did not demonstrate neutrophil deactivation (data not shown). Furthermore, activated neutrophils did not undergo deactivation when contact was made with HT1080 fibrosarcoma cells (data not shown). Thus, Ca2+ oscillations return to a nonactivated profile after contact with trophoblasts.
Because FMLP is known to promote Ca2+ signals and neutrophil activation, we reasoned that trophoblast-mediated neutrophil deactivation could be accomplished by dissociation of the activating stimulus from the cell. To test this idea, we used the fluorescent FMLP analog Fl-FNLPNTL to stimulate neutrophils. Although Fl-FNLPNTL-labeled neutrophils underwent deactivation after contact with trophoblasts (data not shown), the fluorescence intensity did not noticeably change in intensity or distribution. Therefore, the immunoregulatory effects of trophoblasts on neutrophils cannot be explained by ligand dissociation.
Trophoblasts deactivate neutrophil metabolic oscillations
The change in Ca2+ signaling suggests that trophoblast-neutrophil contact may affect neutrophil functions. Because previous studies have shown that Ca2+ signals and metabolic oscillations correlate with one another (18), we examined metabolic oscillations during interactions between activated neutrophils and trophoblasts. Metabolic oscillations were monitored using the autofluorescence of NAD(P)H, which provides a noninvasive tool to monitor cells. Neutrophils were first activated with FMLP. Activated cells were then added to subconfluent trophoblasts. Subconfluence was chosen so that some neutrophils would be in contact with the trophoblasts, whereas other neutrophils (internal controls) would not be in contact with trophoblasts. This allowed the study of metabolic oscillations of trophoblast-attached and nonattached neutrophils within a small area of a microscope slide. Fig. 2⇑D shows a nonattached activated neutrophil, and the left side of Fig. 2⇑F shows its metabolic oscillations; high frequency NAD(P)H oscillations indicated that the HMS is activated (14, 18, 20). However, when this activated neutrophil made contact with a trophoblast (Fig. 2⇑E), the metabolic oscillations decreased in frequency and returned to the normal, nonactivated profile (Fig. 2⇑F). Thus, it is not sufficient to have trophoblasts near an activated neutrophil (Fig. 2⇑D); the activated neutrophil appears to require physical contact with a trophoblast to deactivate. As negative controls, we cocultured neutrophils with tumor cells (HT1080 fibrosarcoma). Importantly, tumor cell contact switched NAD(P)H from the low to the high frequency (activated) mode (data not shown), just the opposite of the trophoblast’s effects. Hence, trophoblast contact regulates certain aspects of neutrophil metabolism.
The neutrophil deactivation events were not specific for FMLP-activated neutrophils. Several neutrophil-activating agents, including LPS, TNF-α, and IL-8, activate neutrophil metabolism, as judged by an increase in oscillation frequency (15, 18). When LPS-, TNF-α-, and IL-8-activated neutrophils were incubated with subconfluent JEG-3 cells, their metabolic oscillations were found to shift from ∼10 to ∼20 s (data not shown). Hence, multiple neutrophil-activating stimuli may be regulated by trophoblasts.
Trophoblasts diminish ROM production by activated neutrophils
We next sought to determine the physiological consequences of trophoblast contact with neutrophils. We chose to monitor ROM production because of its role in inflammatory processes and its connection with NAD(P)H oscillations (14, 20). We again evaluated single cells before and during neutrophil contact, because test tube assays of ROM production would not be informative, as many nonattached cells would produce ROMs. Using subconfluent trophoblast cultures, we labeled neutrophils with the nonfluorescent molecules HE, CDCF, or dihydrorosamine using procedures established previously (20). These molecules were converted into the highly fluorescent molecules ethidium, fluorescein, and TMRos, respectively, upon exposure to ROMs. Labeled, but nonfluorescent, neutrophils were then cocultured with subconfluent trophoblasts. FMLP (100 nM) was added to activate the neutrophils. Fig. 3⇓ shows two representative experiments using two probes for ROM detection, HE and H2-CDCF. The addition of FMLP to these cocultures caused extensive conversion of HE and H2-CDCF to their fluorescent products in neutrophils not adherent to trophoblasts. This was true for neutrophils only 1 μm away from a trophoblast cell. However, neutrophils bound to trophoblasts were dim in these experiments. The presence of unbound/bright and bound/dim neutrophils indicated that this was not a nonspecific artifact of the assay, such as slow oxidation of the labels by air. Similar results were obtained with two different trophoblast cell lines (JEG-3 and JAR).
Because the experiments shown in Fig. 3⇑ measure the amount of probe oxidation inside neutrophils, we next tested ROM deposition outside neutrophils using other methods that detect ROM release (21). Neutrophils were FMLP activated before addition to trophoblasts. Including H2TMRos in the gelatin matrix surrounding trophoblasts and neutrophils monitored ROM release. In this experiment, a neutrophil approaching a clump of trophoblasts was identified (Fig. 2⇑G). The kinetics of ROM release were measured by microfluorometry, which showed a substantial rate reduction during the experiment (Fig. 2⇑I). When the sample was imaged after acquiring microfluorometry data, the neutrophil was found to be in contact with a trophoblast (Fig. 2⇑H). Our studies indicate that contact with a trophoblast down-regulates neutrophil inflammatory activity by a mechanism involving a reduction in Ca2+ and NAD(P)H oscillations frequencies, which is consistent with HMS inhibition and reduced ROM production.
Neutrophil deactivation is observed for trophoblasts and neutrophils from the same patients
The above experiments have focused on neutrophil interactions with a trophoblast cell line. The interpretations of this work may be limited by the facts that 1) this is a transformed cell line and may not be representative of normal trophoblasts; and 2) the trophoblasts and neutrophils are from dissimilar genetic backgrounds. We therefore performed a series of experiments using trophoblasts and neutrophils from the same patient. After delivery by patients undergoing cesarean sections, trophoblasts were obtained from placenta. Before discharge on day 3, blood was drawn from the patient. Experiments indicated that at least 95% of the patients’ neutrophils could be activated on day 3 with 100 nM FMLP, as observed for cells from nonpregnant individuals. Fig. 4⇓ shows a series of experiments with trophoblasts and neutrophils derived from the same mother that parallel those shown in Fig. 2⇑. Fig. 4⇓ shows Ca2+ responses, NAD(P)H oscillations, and ROM production by neutrophils in response to binding semiallogeneic trophoblasts from the same individual. The left sides of Fig. 4⇓, C, F, and I, indicate that the neutrophils were activated by FMLP treatment. Fig. 4⇓, A, B, D, E, G, and H, shows images of cells before and after neutrophil binding to trophoblasts. In all cases, the frequencies of Ca2+ spikes and NAD(P)H oscillations and the amount of ROMs produced were reduced immediately after trophoblast binding. Thus, the normal semiallogeneic biological system displays the same properties as those described above for the trophoblast cell line.
Villous syncytiotrophoblasts deactivate neutrophils
The experiments described above have focused on cytotrophoblasts. To determine whether syncytiotrophoblasts exhibit these same properties, placental villi were studied. Because placental villi must be studied immediately, it was not possible to study villi and leukocytes from the same normal patient, as her leukocytes would have been refractory to FMLP activation. Fig. 5⇓ shows representative studies of neutrophil interactions with this tissue. Again, the left sides of Fig. 5, C, F, and I, clearly show that the cells were activated before contact with syncytiotrophoblasts. Fig. 5⇓, A, B, D, E, G, and H, confirms that neutrophils made contact with the tissue under study. After contact, all three cell parameters changed, as described above for cell lines and the natural semiallogeneic systems. This suggests that the ability to deactivate neutrophils is shared among multiple types of trophoblasts.
Trophoblasts influence metabolic control points of neutrophils
As mentioned above, enhanced Ca2+ signaling can lead to activation of the neutrophil’s HMS (23, 24, 26). Importantly, inhibition of the HMS of activated neutrophils or other cell types using 6-aminonicotinamide can lead to changes in Ca2+ signaling (27) (our unpublished observations). Given the parallel reduction in Ca2+ signaling, NAD(P)H oscillation frequency, and ROM release noted above for several different types of trophoblasts and the correspondence of these parameters with HMS activity, we hypothesized that trophoblasts influence one or more metabolic control points preceding the HMS. It is known that glucose transport is a rate-controlling step in metabolism (28) and that enhanced glucose transport is required for neutrophil activation (29, 30, 31, 32). To test glucose transport at the single-cell level, we used the fluorescent glucose analog 2-NBDG (33). To illustrate the properties of neutrophil activation using this method, resting and FMLP-activated neutrophils were incubated with 2-NBDG at 25 μg/ml in HBSS for 35 min at 37°C, then washed (Fig. 6⇓, A–E). All cells were labeled with 2-NBDG. However, activated neutrophils were brighter than unstimulated cells (Fig. 6⇓, B vs D). The fluorescence intensity (in arbitrary units based upon pixel intensity under identical conditions) was 47 ± 9 for unstimulated cells and 144 ± 14 for stimulated cells (p < 0.001). As expected based upon previous studies (29, 30, 31, 32), glucose uptake was enhanced by cell activation.
To evaluate the effect of neighboring trophoblasts on neutrophil glucose transport, subconfluent JEG-3 trophoblasts were grown on microscope coverslips. To avoid confounding neutrophil glucose uptake with trophoblast-mediated glucose uptake, trophoblast membrane fragments, prepared by osmotic lysis, were used. FMLP-activated neutrophils were added to adherent lysed trophoblast fragments for 20 min at 37°C to allow cell-cell contact to take place. 2-NBDG at 25 μg/ml in HBSS was then added for 35 min at 37°C. This time was chosen because 2-NBDG cellular fluorescence is linear at this incubation time. Coverslips were washed four times with HBSS, then examined microscopically. A representative experiment is shown in Fig. 6⇑, F–H. Fig. 6⇑F shows that some FMLP-activated neutrophils, prepared as described above, were in contact with the glass substrate, whereas others were touching cell fragments derived from trophoblasts. The fluorescence of 2-NBDG is shown in Fig. 6⇑G. Unbound neutrophils were brightly labeled with 2-NBDG (intensity, 141 ± 18), whereas neutrophils in contact with trophoblast fragments were comparatively dim (intensity, 48 ± 16), as indicated by the arrows. To control for possible nonspecific effects, we evaluated the ability of HT1080 fragments to influence 2-NBDG uptake by FMLP-activated neutrophils (Fig. 6⇑, I–K). Both HT1080-bound and unbound neutrophils were labeled brightly with 2-NBDG (Fig. 6⇑J). Interestingly, activated neutrophils bound to HT1080 fibrosarcoma cells were brighter than unbound activated cells (235 ± 20 vs 148 ± 23), suggesting that glucose transport regulation may also be important in other immunological regulatory pathways. Therefore, trophoblast fragments can affect the amount of 2-NBDG taken up by neutrophils.
We have previously shown that the translocation of HMS enzymes, such as G-6-PDase, 6-phosphogluconate dehydrogenase, and transaldolase, to the centrosome is a hallmark of neutrophils from pregnant women (14, 15, 16). By physically separating the HMS from its substrate G-6-P, which is formed at the plasma membrane, the ability of the shunt to produce NADPH is restricted. To determine whether trophoblasts influence the metabolic microcompartmentalization of neutrophils, neutrophils were cocultured with trophoblasts for various periods of time, then fixed and stained with anti-G-6-PDase, anti-transaldolase, or anti-LDH Abs (14, 15, 16). G-6-PDase and transaldolase are participants in the oxidative and nonoxidative arms of the HMS. The immunofluorescence micrographs in Fig. 7⇓ show that G-6-PDase, transaldolase, and LDH were found largely at the periphery of neutrophils after a 4-h coculture period with trophoblasts. However, ∼25% of neutrophils stained with anti-G-6-PDase and anti-transaldolase displayed specific labeling of the centrosome after a 16-h coculture period with trophoblasts. Longer incubation times led to progressively higher percentages of centrosomal accumulation of HMS enzymes (data not shown). LDH, which is not a component of the HMS, was not relocated during trophoblast coculture. Hence, trophoblasts can promote the redistribution of HMS enzymes within neutrophils, which probably contributes to the diminished levels of ROM production found for pregnancy neutrophils after stimulation.
Neutrophil deactivation protects trophoblasts
The preceding paragraphs have identified a unique anti-inflammatory mechanism of trophoblasts as well as its underlying biochemical regulatory circuits. To explore the physiological ramifications of this new pathway, we examined the ability of FMLP-treated neutrophils to deposit ROMs and elicit DNA damage in neighboring trophoblasts. Trophoblasts were labeled with H2TMRos, as described above, then incubated with FMLP-treated neutrophils at 37°C for 1.5 h. Although little TMRos fluorescence could be found in JEG-3 cells, considerable fluorescence could be found in HTR8/SVneo cells (Fig. 8⇓). Because ROMs are known to damage DNA, we next evaluated DNA damage using the TUNEL protocol for labeling strand breaks. TUNEL staining was not observed for JEG-3 cells, indicating that FMLP-treated neutrophils did not affect these cells, although they contacted these trophoblasts. However, TUNEL staining was observed when HTR8/SVneo cells were cocultured with FMLP-treated neutrophils (Fig. 8⇓D). Although both JEG-3 and HTR8/SVneo cells had associated neutrophils (69 ± 9 vs 74 ± 8%, respectively), only 3.8 ± 2.9% of JEG-3 cells were stained by the TUNEL method compared with 60 ± 12% of HTR8/SVneo cells (p < 0.001). However, it remains possible that these differences might be accounted for by something other than the neutrophil’s activation state. JEG-3 and HTR8/SVneo cells did not display TUNEL labeling in the absence of ROMs (Fig. 9⇓, A–D), but did display TUNEL staining after incubation with 0.005% H2O2 for 2 h at 37°C (Fig. 9⇓, E–H). Thus, both cell types displayed DNA damage after exposure to ROMs. As similarly noted above (Fig. 8⇓), DNA damage was detected in HTR8/SVneo cells after incubation with FMLP-activated neutrophils for 80 min at 37°C (Fig. 9⇓L), but not in parallel experiments with JEG-3 cells (Fig. 9⇓J). This effect, however, appeared to be specific for activation of the neutrophil’s HMS, because it was inhibited by 6-aminonicotinamide. Fig. 9⇓P shows that pretreatment of neutrophils with 10 mM 6-aminonicotinamide for 30 min, followed by coincubation with HTR8/SVneo in the presence of 5 mM 6-aminonicotinamide for 2 h at 37°C blocked DNA damage in HTR8/SVneo cells. We next sought to confirm that H2O2 is capable of entering JEG-3 and HTR8/SVneo cells with equal efficacy. Fig. 9⇓, Q and R, shows JEG-3 and HTR8/SVneo cells, respectively, treated with various amounts of H2O2. Individual cells were exposed to 0 (trace 1), 0.001% (trace 2), 0.005% (trace 3), and 0.01% (trace 4) H2O2, then followed for several minutes to observe the kinetics of TMRos formation. As this figure shows, no significant differences in the rate of TMRos formation in these cells were observed. Hence, differences in trophoblast DNA damage are not likely to be explained by differences in ROM entry or ROM-mediated DNA damage, but appear to be closely linked with the ability of trophoblasts to affect the neutrophil’s HMS, inasmuch as 6-aminonicotinamide affects the ability of neutrophils to damage HTR8/SVneo cells, and HMS activity is depressed by JEG-3 cells. We therefore suggest that neutrophil deactivation protects JEG-3 cells from neutrophil-mediated DNA damage during cell-cell contact.
Our goal in these experiments was to examine the possibility that trophoblasts directly regulate the activation status of nearby neutrophils. To unravel these cell-cell interactions, it was necessary to use single-cell assays to investigate the properties of each cell. Single-cell microscope-based technologies allow us to identify trophoblast-neutrophil events and characterize their chemical properties over time. To model trophoblast-mediated cellular immunoregulation in vitro, we used trophoblasts grown on microscope slides combined with neutrophils stimulated with FMLP. Multiple single-cell parameters of activation reveal that trophoblast contact deactivates activated neutrophils. This finding is consistent with previous studies suggesting that crude cell fractions, plasma membrane vesicles, and glycocalyx components of trophoblasts inhibit the allogeneic reactivity of lymphocytes (34, 35, 36). A previous study has suggested that the trophectoderm deflects the injurious potential of LPS-activated macrophages (13). Unfortunately, the mechanisms responsible for these immunoregulatory processes have remained obscure. In this study we show that trophoblasts are capable of a contact-dependent deactivation of neutrophils via a mechanism involving at least two metabolic “choke” points preceding the HMS: glucose transport and HMS enzyme trafficking, which lessen the ROM burden and DNA damage of neighboring trophoblasts.
There is broad agreement that Ca2+ signaling plays a central role in neutrophil activation (19, 23, 24, 25, 26). It seemed likely, therefore, that trophoblast-mediated immunomodulation of neutrophils would be accompanied by alterations in Ca2+ signaling. We have shown that the rapid periodic Ca2+ spikes of activated adherent neutrophils are normalized to the profiles seen for unstimulated neutrophils after trophoblast contact. It is not surprising that perturbation of an activating signal causes downstream effects on cell activation. Because other neutrophil functions, such as degranulation, are also regulated by Ca2+ signaling, it is likely that additional cell functions are depressed by trophoblasts.
Neutrophil activation leads to metabolic changes, such as NADPH production, that are required for ROM production by NADPH oxidase. Trophoblast contact alters the dynamics of NAD(P)H production by neutrophils and, in parallel, ROM production. Furthermore, kinetic studies show that the rate of ROM release by activated neutrophils is dramatically reduced by contact with trophoblasts. Our results have shown that fewer ROMs accumulate in and around neutrophils associated with trophoblasts. Not surprisingly, fewer ROMs accumulate within trophoblasts capable of deactivating neutrophils. Moreover, trophoblast DNA damage was not detected in trophoblasts capable of deactivating neutrophils. Hence, one important consequence of trophoblast contact is depression of neutrophil-mediated ROM production and DNA damage in neighboring cells, which would minimize potential harm to the placenta and conceptus by inflammatory cells. Although most placental immunobiology studies are restricted to cell lines and/or cytotrophoblasts, we have also examined syncytiotrophoblasts, cytotrophoblasts, and neutrophils from the same mother. As identical regulatory events and ROM production changes were observed using multiple trophoblast types, the findings are likely to be of physiological relevance.
An immunoregulatory model based on metabolic control
Our studies suggest the model of trophoblast-mediated inflammatory regulation shown in Fig. 10⇓. Under normal conditions, neutrophil stimuli enhance Ca2+ signaling events (23, 24, 26) that accompany an increase in glucose uptake (32), which, in turn, is required for ROM production (Fig. 10⇓). Inhibitors of G-6-PDase (37) or glucose transport (U. Kummer, J. Zobeley, K. Naxerova, J. C. Brasen, R. Fahmy, A. L. Kindzelskii, A. R. Petty, and H. R. Petty, manuscript in preparation) block HMS activation despite the presence of activating signals. In the absence of glucose or enhanced glucose uptake, there is no HMS-driven respiratory burst (29, 30, 31, 32). During neutrophil activation, Ca2+ signals contribute to an increase in the affinity of the glucose transporter and an increase in glucose flux. This and perhaps other chemical events lead to HMS activation. Once activated, the HMS draws down G-6-P levels, causing the increase in glycolytic oscillation frequency noted above (Kummer et al., manuscript in preparation). Enhanced NADPH levels support ROM production. When neutrophils are attached to trophoblast membrane fragments, glucose uptake is significantly lower than that observed for neighboring activated neutrophils, thus limiting G-6-P, NADPH, and ROM production. The reduction in glucose transport is sufficient to shut down the HMS, thus causing normalization of the oscillatory frequency due to the return of normal carbon flux through glycolysis. Hence, trophoblast-mediated alterations in glucose transport affect downstream biochemical pathways, such as ROM production. Moreover, these trophoblast-mediated changes in neutrophil activation do not appear to be restricted to FMLP, because other neutrophil activators, such as LPS and TNF-α, can be deactivated by trophoblasts.
In addition to this acute regulatory pathway, trophoblasts are capable of promoting the retrograde transport of the neutrophil’s HMS enzymes on microtubules. HMS enzymes form a macromolecular complex that includes G-6-PDase, 6-PGDase, and transaldolase (14, 15, 16). Because HMS complex trafficking to the centrosome requires microtubule-based motility, it requires hours instead of seconds to be accomplished. The metabolic microcompartmentalization of the HMS at the centrosome also limits the availability of G-6-P to the HMS by allowing G-6-P to be irreversibly metabolized at the cell periphery by glycolytic enzymes (14). Although the regulation of glucose transport may account for the rapid changes in cell activation at the materno-fetal interface, the systemic changes in peripheral blood neutrophil activation may be more closely related to HMS metabolic microcompartmentalization. Thus, we have identified two new mechanisms that allow trophoblasts to actively regulate the inflammatory potential of neutrophils.
The present study of trophoblasts offers insight into neutrophil properties during pregnancy. It is widely known that ROM production by peripheral blood neutrophils from pregnant women cannot be fully activated (14, 15, 16, 38, 39, 40). Because diminished neutrophil activation lasts for ∼2 h after trophoblast contact, and the placental surface area allows large volumes of maternal blood to pass villous syncytiotrophoblasts, it is possible that the placental interface, through contact with circulating neutrophils, is capable of regulating their propensity to activate. Hence, trophoblast to neutrophil interactions may contribute to the systemic properties previously observed for normal maternal neutrophils, including HMS enzyme trafficking (14).
Placental oxidative damage plays a central role in preeclampsia, diabetic pregnancy, and miscarriage (41). Neutrophils from pre-eclamptic patients produce enhanced levels of ROMs (42), which probably contribute to oxidative stress (43). Although syncytiotrophoblasts derived from normal pregnant women deactivate neutrophil ROM production in this study, syncytiotrophoblasts from patients with severe preeclampsia stimulate neutrophil superoxide production (44). This suggests that there is a fundamental biological difference between trophoblasts from normal and pre-eclamptic pregnancies. Enhanced placental oxidative stress may account for the DNA damage found in cytotrophoblasts from pre-eclamptic patients (45, 46). Because large numbers of leukocytes come into contact with trophoblasts, the aberrant regulation of ROM production may contribute to disease mechanisms.
Although we have discovered a novel cellular mechanism regulating neutrophil activation, some details have not yet been established. It seems likely that a component of the trophoblast’s plasma membrane participates in neutrophil deactivation, because both intact and lysed cells possess this activity. Although the ligand(s) and receptor(s) mediating this process are unknown, it has not escaped our attention that they will probably form important new lead compounds in drug discovery to manage inflammatory diseases.
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 National Institute of Allergy and Infectious Diseases, the National Multiple Sclerosis Society, and the Intramural Program of the National Institute of Child Health and Human Development.
↵2 Address correspondence and reprint requests to Dr. Howard R. Petty, Department of Ophthalmology and Visual Sciences, 1000 Wall Street, University of Michigan Medical School, Ann Arbor, MI 48105. E-mail address:
↵3 Abbreviations used in this paper: HMS, hexose monophosphate shunt; DIC, differential interference contrast; Fl-FNLPNTL, fluorescein-conjugated N-formyl-Nle-Leu-Phe-Nle-Tyr-Lys; H2TMRos, dihydrotetramethylrosamine; 2-NBDG, 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose; ROM, reactive oxygen metabolite; CDCF, 2′,7′-dichlorofluorescein; HE, hydroethidine; G-6-PDase, glucose-6-phosphate dehydrogenase.
- Received July 13, 2005.
- Accepted December 29, 2005.
- Copyright © 2006 by The American Association of Immunologists