Schistosome Egg Antigens Elicit a Proinflammatory Response by Trophoblast Cells of the Human Placenta

Abstract

Schistosomiasis affects nearly 40 million women of reproductive age. Many of these women are infected while pregnant and lactating. Several studies have demonstrated transplacental trafficking of schistosome antigens; however, little is known regarding how these antigens affect the developing fetus and placenta. To evaluate the impact of schistosomiasis on trophoblasts of the human placenta, we isolated primary trophoblast cells from healthy placentas delivered at term. These trophoblasts were placed in culture and treated with Schistosoma japonicum soluble egg antigens (SEA) or plasma from S. japonicum-infected pregnant women. Outcomes measured included cytokine production and activation of signal transduction pathways. Treatment of primary human trophoblast cells with SEA resulted in upregulation of the proinflammatory cytokines interleukin 6 (IL-6) and IL-8 and the chemokine macrophage inflammatory protein 1α (MIP-1α). Cytokine production in response to SEA was dose dependent and reminiscent of production in response to other proinflammatory stimuli, such as Toll-like receptor 2 (TLR2) and TLR4 agonists. In addition, the signaling pathways extracellular signal-regulated kinase 1/2 (ERK1/2), Jun N-terminal protein kinase (JNK), p38, and NF-κB were all activated by SEA in primary trophoblasts. These effects appeared to be mediated through both carbohydrate and protein epitopes of SEA. Finally, primary trophoblasts cocultured with plasma from S. japonicum-infected pregnant women produced increased levels of IL-8 compared to trophoblasts cocultured with plasma from uninfected pregnant women. We report here a direct impact of SEA on primary human trophoblast cells, which are critical for many aspects of a healthy pregnancy. Our data indicate that schistosome antigens can activate proinflammatory responses in trophoblasts, which might compromise maternal-fetal health in pregnancies complicated by schistosomiasis.

INTRODUCTION

Schistosomiasis, caused by three principal species of dioecious trematodes (flatworms), currently affects over 250 million individuals, results in 1.53 million disability-adjusted life years (DALYs) lost per annum (1), and contributes to poor health and economic stagnation in areas in which schistosomiasis is endemic (2). Forty million women of child-bearing age are currently infected with schistosomes, yet the impact of schistosomiasis on the health of human pregnancy remains understudied. Strong evidence exists in rodent models for a host of deleterious outcomes from schistosome infection during pregnancy (3, 4). In humans, there is a relatively strong association between schistosome infection and an increased risk of maternal anemia (5, 6). In addition, observational studies have revealed 4 to 18% lower birth weights for babies born to infected mothers (7, 8), although clinical trials involving midgestational treatment for schistosomiasis have not demonstrated improved pregnancy outcomes (9). Schistosomiasis japonica is associated with multiple proinflammatory response-mediated morbidities in nonpregnant individuals (10, 11), and inflammation is known to result in deleterious birth outcomes, including prematurity, intrauterine growth restriction (IUGR), and low birth weight (LBW) (12–16).

We recently extended these findings to pregnancy, demonstrating that schistosomiasis is associated with elevated proinflammatory cytokines in human maternal, placental, and cord blood, as well as an increased risk for the development of acute subchorionitis at the maternal-fetal interface (17). Importantly, those women who displayed the highest levels of inflammatory markers (tumor necrosis factor alpha [TNF-α] and interleukin 1β [IL-1β] production) also had offspring with lower birth weights, underscoring the potentially detrimental effects of significant levels of inflammatory cytokines at the maternal-fetal interface (17).

Despite the importance of molecular events at the maternal-fetal interface in establishing and maintaining pregnancy, the response of the placenta to schistosomiasis has not been examined. In this study, we examined the effects of schistosome Schistosoma japonicum soluble egg antigens (SEA) on trophoblast cells of the human placenta. In addition, we evaluated cytokine production by term trophoblast cells after 24 h of culture with maternal plasma collected at 32 weeks' gestation from women with S. japonicum infection or uninfected controls. We report that SEA activates multiple signaling pathways in trophoblasts, resulting in marked upregulation of proinflammatory cytokines.

MATERIALS AND METHODS

Cell isolation and culture.

Placental samples were collected in accordance with protocols approved by the institutional review boards at Women and Infants and Rhode Island Hospitals. Cytotrophoblasts (CTBs) were isolated from normal, healthy placentas (i.e., from gestations free of any identified complications) delivered via elective cesarean section at term as described previously (18). Briefly, villous tissue was dissected away from the basal plate and major blood vessels and subjected to enzymatic digestion with DNase I (type IV; Sigma Chemical, St. Louis, MO) and trypsin (Invitrogen, Carlsbad, CA). The resulting single-cell suspension was size fractionated by application to a Percoll density gradient. To ensure a homogenous population of CTBs, the cells underwent negative selection using an antibody recognizing human leukocyte antigen A (HLA-A), -B, and -C (W6/32; eBioscience, San Diego, CA) with a magnetically labeled secondary antibody (Miltenyi Biotec GmBH, Bergisch Gladbach, Germany), and the cells were found to be >98% pure by cytokeratin 7 staining (data not shown). Purified cells were cultured in Iscove's modified Dulbecco's medium with 10% fetal bovine serum (FBS), 1% l-glutamine, 1% penicillin-streptomycin-amphotericin B and allowed to form syncytia in vitro by incubation in a humidified chamber for 96 h prior to stimulation with SEA. All experiments were conducted on cells that were first allowed to differentiate to better approximate the syncytiotrophoblast layer found in direct contact with maternal blood in the human placenta.

Positive (lipopolysaccharide [LPS] and zymosan; Sigma Chemical) and negative (Sj68, a highly purified recombinant schistosome protein comprising amino acids [aa] 20 to 176 of GenBank accession no. CAX72484.1) controls were added for the final 24 h of a 5-day culture period. Cells used in the studies of signaling pathways were cultured for 4 days in the presence of 10% fetal bovine serum (FBS) and then switched to serum-free medium. Pharmacological inhibitors of extracellular signal-regulated kinase 1/2 (ERK1/2) (UO126; Sigma Chemical), p38 mitogen-activated protein (MAP) kinase (SB202190; Sigma), Jun N-terminal protein kinase (JNK) (SP600125; Sigma), and NF-κB (JSH-23; Calbiochem, Merck KGaA, Darmstandt, Germany) were added (10 μM) 1 h prior to the addition of SEA (25 μg/ml for 24 h). To ensure cell viability after addition of the inhibitor, 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) viability assays (Sigma) were performed (data not shown).

Antigen preparation.

Schistosome eggs were collected from rabbit livers infected with Schistosoma japonicum. Soluble egg antigen (SEA) was prepared under endotoxin-free conditions according to standard procedures (19). In brief, 7 to 8 weeks after Schistosoma japonicum cercarial exposure, infected rabbits (∼2,500 cercariae/rabbit) were perfused, and their livers were collected and rinsed with LPS-free phosphate-buffered saline (PBS). Liver homogenate was filtered, washed, and centrifuged over a Percoll-0.25 M sucrose gradient. The purified eggs were washed and homogenized via mortar and pestle in PBS for 20 min. The homogenate was ultracentrifuged with the resulting supernatant and frozen at −80°C. Preparations were evaluated for contaminating endotoxin using an FDA standard Limulus amebocyte lysate (LAL) assay (Acila GmbH). Endotoxin levels for all SEA preparations used were <6 IU/mg protein, which is at least 1,000-fold lower than levels that have been shown to influence human trophoblast cells (5).

We subjected schistosome SEA to a variety of treatments in order to evaluate the relative contribution of carbohydrate and peptide epitopes to cytokine stimulation. The disruption of terminal saccharide rings was achieved by treatment with sodium m-periodate (20 mM) in sodium acetate buffer (100 mM) for 45 min at 25°C in the dark. Aldehydes produced by this reaction were reduced to primary alcohols by the addition of sodium borohydride (50 mM) for 30 min at 25°C in the dark. Mock-treated SEA was diluted in sodium acetate buffer without the addition of sodium m-periodate and with the addition of water instead of sodium borohydride. Finally, all samples were dialyzed against PBS prior to use.

In addition, SEA was subjected to protein degradation by incubation with the serine protease proteinase K (50 μg/ml) for 15 min at 37°C. Reactions were stopped by the addition of phenylmethylsulfonyl fluoride (PMSF) (5 mM). As a negative control, an aliquot of SEA was denatured by heating to 95°C for 10 min. Modified SEA (25 μg/ml) was used to stimulate primary trophoblast cells as outlined above.

Cytokine assays.

Primary trophoblast cells (n = 12 trophoblast preparations from distinct placentas) were treated with SEA (25 μg/ml) or medium alone for 24 h. Cytokine assays were performed on culture media collected at the end of the treatment period. IL-1β, IL-6, gamma interferon (IFN-γ), tumor necrosis factor alpha (TNF-α), IL-4, Il-5, IL-10, IL-13, IL-12, IL-8, and IL-2 were measured with a bead-based platform (BioPlex; Bio-Rad, Hercules, CA) using a sandwich antibody-based assay as described previously (20). In addition, chemokines, including CCL3 (macrophage inflammatory protein 1α [MIP-1α]), CCL18, and CCL2 (monocyte chemotactic protein 1 [MCP-1]), were measured using a similar bead-based approach. As expected, many of these cytokines were not produced at detectable levels, reflecting the nonimmune origin of the syncytiotrophoblast.

Signaling assays.

Primary trophoblasts were collected and allowed to differentiate as described above. Cultures were treated with SEA (25 μg/ml) or medium alone for 24 h, 30 min, 15 min, or 5 min. Cells were then harvested using a cell lysis kit (Bio-Rad) in accordance with the manufacturer's instructions. Briefly, cells were washed and then lysed in the presence of protease inhibitors by shaking for 20 min at 4°C. Cellular debris was cleared by centrifugation at 4,500 × g and 4°C for 20 min. The resulting whole-cell lysates were quantitated using a standard bicinchoninic acid (BCA) assay (Thermo Scientific, Rockford, IL) to ensure that all samples were within the working range of the assay. Total and phosphorylated protein levels for the signaling molecules ERK1/2, JNK, p38 MAP kinase (MAPK), Akt, and IκBα were assessed using the BioPlex phosphoprotein and total protein detection kits (Bio-Rad), according to the manufacturer's instructions, on a bead-based analyzer (BioPlex; Bio-Rad). Data were analyzed as the ratio of phosphorylated to total protein for a given signaling molecule.

Progesterone assay.

Progesterone was measured in the culture media from trophoblasts that had been in culture for 5 days, with SEA exposure (25 μg/ml) for the final 24 h. Hormone levels were measured using a progesterone enzyme immunoassay (EIA) kit (Cayman Chemical, Ann Arbor, MI) according to the manufacturer's instructions. All samples were run in duplicate at a 1:100 dilution.

Human plasma assays.

Term trophoblasts (described above) were cultured for 4 days before being cultured for 24 h in serum-free media with the addition of 10% plasma collected from pregnant women at 32 weeks' gestation. All plasma samples were collected from women residing in Leyte, the Philippines, an area where schistosomiasis is endemic. The study population was the same as has been described elsewhere (17). From this larger population, we selected 9 women infected with schistosomiasis and 8 uninfected women matched for socioeconomic status (SES), coinfections (Ascaris lumbricoides, Trichuris trichiura, and hookworm), gravida, parity, gestational age, body mass index (BMI), smoking status, and maternal age (Table 1). Socioeconomic status was calculated from a detailed questionnaire that we previously validated in this study population (17) and is reported as a composite score.

Table 1.

Demographic data for selected plasma samples from pregnant women collected at 32 weeks of gestation, matched on a number of potential confounders

Potential confoundera	In Schistosoma japonicum-infected samples (n = 9)	In Schistosoma japonicum-uninfected samples (n = 8)
Coinfections (n)
Ascaris lumbricoides	8	7
Trichuris trichiura	8	8
Hookworm	7	6
SES (mean [95% CI])	12.90 (10.03, 15.77)	14.22 (10.90, 17.53)
Smoking status (n)
Yes	0	0
No	9	8
Gestational age (wk) (mean [95% CI])	38.67 (36.18–41.16)	38.88 (37.93–39.82)
BMI (kg/m2) (mean [95% CI])	22.49 (20.38–24.60)	22.39 (19.02–25.75)
Maternal age (yr) (mean [95% CI])	28.43 (25.11–31.74)	32.64 (27.28–38.00)
Parity (mean [95% CI])	2.56 (0.97–4.15)	3.63 (2.15–5.10)
Gravida (mean [95% CI])	3.78 (2.46–5.10)	4.89 (3.30–6.45)

^aSES, socioeconomic status; BMI, body mass index; CI, confidence interval.

Following the 24 h of incubation with maternal plasma, trophoblast culture supernatants were collected and analyzed for cytokine production as described above.

Statistical analysis.

All data are reported as means plus or minus the standard errors of the mean (SEM). Data analysis was performed using JMP 10 (SAS Institute, Cary, NC). All data were evaluated using matched-pair Wilcoxon signed-rank analysis, with experiments performed on trophoblast preparations from unique placentas (n = number of distinct placentas used for each experiment). Statistical significance was considered to be indicated by a P value of <0.05. SEA manipulation experiments (see Fig. 4) were analyzed using matched-pair analysis, with a one-sided P value of <0.05 considered significant.

Fig 4.

The ERK1/2 and NF-κB signaling pathways are the dominant mediators of the proinflammatory cascade initiated by SEA in trophoblasts. Pharmacological inhibition of the JNK, p38, and ERK1/2 MAPK pathways as well as the NF-κB pathway prior to SEA (25 μg/ml) exposure for 24 h. (A, C, E, and G) IL-6 production after SEA with the addition of either vehicle (DMSO) or specific inhibitors. (B, D, F, and H) IL-8 production from the same experiments. (A and B) Inhibition of the ERK1/2 signaling pathways (n = 6). (C and D) Inhibition of the p38 MAPK signaling pathway (n = 6). (E and F) Inhibition of the JNK signaling pathway (n = 5). (G and H) Inhibition of the NF-κB signaling pathway (n = 6 individual trophoblast preparations from distinct placentas). All data are reported as means ± SEM. Unique letters denote statistically significant differences at P values of <0.05.

RESULTS

Schistosome egg antigens stimulate proinflammatory cytokine release from trophoblast cells.

In order to examine the direct effect that schistosome infection might exert on the trophoblast cells of the placenta, we treated term trophoblast cells that had formed syncytia in vitro with schistosome SEA in culture. After 24 h of exposure to SEA, the media were collected and analyzed for a variety of cytokines, including interleukin 1β (IL-1β), IL-6, gamma interferon (IFN-γ), TNF-α, IL-4, IL-5, IL-10, IL-13, IL-12, IL-8, and IL-2. Of these, secretion of IL-6 and IL-8 was increased 6.0-fold (P = 0.03) and 2.0-fold (P < 0.01), respectively, after 24 h of treatment with SEA (Fig. 1A and B). Both of these proinflammatory cytokines are known to play important roles at the maternal-fetal interface, and IL-8 in particular can also act as a chemokine to attract immune cells (21, 22).

Fig 1.

The proinflammatory cytokines IL-6 and IL-8 and the chemotactic cytokine MIP-1α are upregulated by SEA in trophoblast cells. Trophoblast cells isolated from healthy pregnancies at term formed syncytia in vitro for 4 days before being treated with SEA (25 μg/ml) for 24 h. Media from all treatment conditions were collected and measured for cytokine expression. (A) IL-6 production is upregulated 6.0-fold in trophoblast cells exposed to SEA (n = 12, P = 0.03). (B) IL-8 production is increased 2.0-fold after SEA treatment (n = 12, P < 0.01). (C) MIP-1α production is increased 2.8-fold following SEA treatment in trophoblasts (n = 12 individual trophoblast preparations from distinct placentas, P = 0.03).

Of the other cytokines measured, most were undetectable (IL-1β, IFN-γ, TNF-α, IL-5, IL-13, and IL-12), while others did not differ across treatment groups (IL-4, IL-10, and IL-2). Most of these (IL-1β, IFN-γ, TNF-α, IL-12, IL-4, IL-10, and IL-2) have been reported to be expressed by human trophoblasts (23–26), although in a number of different treatment/culture paradigms as well as cytokine measurement techniques. The diversity among these reports, ours included, likely explains the divergence in cytokine production by trophoblast cells in culture. Interestingly, the anti-inflammatory cytokine IL-10 was not significantly altered following SEA treatment (data not shown). This is particularly important given that chronic schistosome infection is typically associated with a systemic anti-inflammatory Th2-type (high IL-10) response in other tissues (27).

In addition to assessing the levels of cytokines in the culture media following SEA treatment, we also evaluated the secretion of select chemokines, including MCP-1 and MIP-1α. Of these, MIP-1α was significantly upregulated (2.8-fold, P = 0.03) by primary trophoblasts exposed to SEA for 24 h in vitro compared to the corresponding control cells that received media alone (Fig. 1C). These data suggest that, in addition to augmenting the proinflammatory cytokine response, trophoblasts also respond to SEA by releasing chemokines, which might ultimately influence the immune cell milieu at the maternal-fetal interface.

Cytokine production in trophoblast cells is SEA dose dependent.

Previous dose-ranging experiments performed in our laboratory using SEA and placental explant cultures showed an optimal dose of SEA to be 25 μg/ml (data not shown). To determine the optimal SEA concentration for stimulation of purified trophoblasts, we stimulated isolated trophoblast cells that had been cultured for 4 days with various levels of SEA (2.5, 10, 25, and 50 μg/ml). With as little as 10 μg/ml SEA, evaluation of the proinflammatory cytokines IL-6, IL-8, and MIP-1α showed significant upregulation, which continued to rise as the level of SEA increased (Fig. 2A, C, and E). These dose response data are comparable to those reported for studies of SEA stimulation of professional immune cells (28–31) and informed our selection of an SEA dose of 25 μg/ml for 24 h.

Fig 2.

Trophoblast cells respond to SEA in a dose-dependent manner. Trophoblast cells isolated from healthy pregnancies at term were allowed to form syncytia in vitro for 4 days before being treated with SEA for 24 h. Media from all treatment conditions were collected and measured for cytokine expression. (A) IL-6 expression is increased with higher doses SEA (n = 7, P = 0.03). (B) IL-6 production in trophoblasts responds to SEA in a specific manner (n = 6, P = 0.03). (C) IL-8 production by trophoblasts increases with increasing doses of SEA exposure (n = 7, P = 0.02). (D) SEA stimulates IL-8 production in trophoblasts at a level on par with the TLR4 and TLR2 agonists LPS and zymosan, respectively (n = 6, P = 0.03). (E) The chemokine MIP-1α is also dose responsive to SEA (n = 7 individual trophoblast preparations for distinct placentas, P = 0.02). nt, no treatment.

Importantly, the trophoblast responses detected in our experiments are specific to SEA, as exposure of trophoblast cells to an irrelevant schistosome protein, Sj68 (25 μg/ml for 24 h), did not elicit any cytokine response (Fig. 2B and D). This result is in contrast to that observed for the positive controls of lipopolysaccharide (LPS), a TLR4 agonist and known stimulant of proinflammatory cytokines, and zymosan (TLR2 agonist) (Fig. 2B and D). Of note, despite being recognized stimulants of proinflammatory cytokines in isolated term trophoblast cells, both LPS and zymosan (32) showed high variability between placental preparations in their stimulation potential, whereas SEA, at lower doses, is much more consistent in its proinflammatory stimulation.

Trophoblast cells respond to SEA through a variety of signaling pathways.

SEA has been shown to activate members of the MAPK signaling cascades in other cell types (33–35). In order to determine intracellular signaling pathways that are stimulated by SEA in trophoblast cells, we performed a bead-based assay measuring the activated (i.e., phosphorylated) and total forms of the signaling molecules: Akt, c-Jun N-terminal kinase (JNK), extracellular signal-regulated kinase 1/2 (ERK1/2), p38 mitogen-activated protein kinase (p38 MAPK), and nuclear factor of κ light polypeptide gene enhancer in B-cells inhibitor α (IκBα). Within 15 min of SEA stimulation, all three MAPK signaling pathways evaluated (JNK, ERK1/2, and p38 MAPK) were activated above the media-only control, with a significant elevation in ERK1/2 as early as 5 min after treatment (Fig. 3A to C). As expected, these responses were transient, as all three signaling molecules returned to control levels within 30 min of SEA stimulation.

Fig 3.

SEA stimulates a variety of signaling pathways in trophoblast cells. Trophoblast cells isolated from healthy pregnancies at term formed syncytia in vitro for 4 days before being treated with SEA. Whole-cell lysates from all treatment conditions were collected and measured for expression of total and phosphorylated forms of signaling molecules. ERK1/2 (A), JNK (B), p38 MAPK (C), and IκBα (D) activity in trophoblast cells is upregulated within 15 min of exposure to SEA (n = 5 individual trophoblast preparations from distinct placentas).

In addition, IκBα was activated in a transient manner, with a significant increase in phosphorylation 15 min post-SEA exposure. Taken together, these results suggest that SEA acts through a variety of signaling molecules to influence the biology of the trophoblast cell. In contrast, Akt was not significantly altered by treatment with SEA (data not shown), demonstrating the specificity of the MAPK stimulation seen after SEA exposure.

We next used pharmacological inhibitors to specifically block the activity of each signaling pathway (UO126, ERK1/2; SB202190, p38 MAPK; SP600125, JNK; JSH23, NF-κB) and evaluated the trophoblast response to SEA. In most cases, pretreatment with an inhibitor of a specific signaling pathway prior to 24-h SEA exposure resulted in an intermediate level of cytokine production compared to the level in either the vehicle control alone or vehicle control plus SEA (Fig. 4A, B, and E through H), with the notable exception of SB202190, the inhibitor of the p38 MAPK pathway. Pretreatment with SB202190 lowered the levels of IL-6 and IL-8 production compared to pretreatment with dimethyl sulfoxide (DMSO) (vehicle), but not to the point of statistical significance (Fig. 4C and D). Of note, inhibition of no single pathway resulted in complete abolishment of the proinflammatory response to SEA, suggesting that multiple pathways might be activated simultaneously to lead to IL-6 and IL-8 production by trophoblasts in response to SEA.

SEA activates trophoblasts via both carbohydrate and peptide epitopes.

SEA is a heterogeneous mix of potential antigens. In order to broadly classify the nature of the relevant epitopes, we evaluated the impact of periodate and protease treatment on SEA-induced trophoblast activation. Sodium meta-periodate oxidatively cleaves between vicinal diols of sugar residues on glycoproteins, thus altering carbohydrate-associated antigenicity (36), while proteinase K, which has broad specificity, destroys peptide epitopes.

Pretreatment of SEA with either periodate or proteinase K attenuated the IL-6 and IL-8 responses of stimulated trophoblasts (Fig. 5), suggesting that the syncytiotrophoblast responds to both carbohydrate and peptide epitopes in SEA.

Fig 5.

The proinflammatory effects of SEA on trophoblast cells involve both carbohydrate and protein epitopes. SEA was manipulated in a manner that disturbed carbohydrate (periodate) or peptide (proteinase) epitopes prior to treatment of trophoblasts. Heat-denatured SEA was included as a negative control. Secretion of the proinflammatory cytokines IL-6 (A) and IL-8 (B), normally stimulated by SEA exposure, was disturbed with SEA manipulation. mock, SEA subjected to dilution and incubations in parallel to periodate-treated SEA, without the addition of periodate. Matched-pair analysis with a one-sided t test; a P value of <0.05 was considered significant (n = 5 individual trophoblast preparations from distinct placentas).

Exposure to SEA might enhance hormone production by trophoblast cells.

Progesterone levels in the culture media were assayed following 24-h exposure to SEA (25 μg/ml) in vitro. Although not statistically significant, with n = 8 placental preparations there was a trend for increased levels of progesterone in the culture media from cells that had been exposed to SEA for the final 24 h of the 5-day culture period (Fig. 6). This is an intriguing finding that warrants further investigation.

Fig 6.

Progesterone production increases in trophoblasts exposed to SEA in vitro. Progesterone levels in the culture media following 24-h SEA exposure (25 μg/ml) were measured by an EIA. Although not statistically significant, a trend toward increased levels of progesterone was observed in those cells treated with SEA (n = 8 individual trophoblast preparations from distinct placentas, P = 0.06).

Factors present in human plasma from infected pregnant women cause increased proinflammatory cytokine production by term trophoblast cells.

A subset of plasma samples collected at 32 weeks' gestation was selected from a larger study performed to examine the impact of schistosomiasis on pregnancy (17). Primary term trophoblast cells from North American women with healthy pregnancies were placed in culture and allowed to differentiate for 4 days. The media were then replaced with serum-free media containing 10% plasma from individual subjects infected with schistosomiasis, an uninfected matched control, or 10% FBS (negative control). All media were collected after 24 h and assayed for IL-6 and IL-8 levels. IL-8 was significantly upregulated in those cells exposed to plasma from women with an active schistosome infection compared to cells exposed to plasma from their uninfected counterparts (Fig. 7), with IL-6 also showing a trend toward an increased response to plasma from infected women (data not shown). IL-8 and IL-6 levels in the plasma alone ranged from undetectable to more than 100-fold lower than the lowest levels in any of the trophoblast cultures (data not shown). Therefore, the maternal plasma itself is not a significant contributor of IL-8 or IL-6 production in these trophoblast cultures.

Fig 7.

Plasma from pregnant women with schistosomiasis causes increased IL-8 production by human trophoblast cells. Term trophoblasts from North American control women were cultured for 4 days prior to culture for 24 h in media supplemented with plasma samples (10% final concentration). Plasma samples were collected at 32 weeks' gestation from women with and without schistosomiasis residing in an area of the Philippines where schistosomiasis is endemic. IL-8 production is significantly elevated in trophoblasts that were exposed to plasma from women infected with schistosomiasis. Data are expressed as the median fold increase in IL-8 levels over the no-plasma controls for each trophoblast preparation (n = 6 individual trophoblast preparations from distinct placentas) (n = 17 plasma samples, P = 0.04).

DISCUSSION

Schistosomiasis is responsible for significant morbidity in low- and middle-income countries and significantly contributes to the global burden of anemia, undernutrition, and hepatic fibrosis (2). Although praziquantel (PZQ) effectively treats schistosomiasis, PZQ remains an FDA pregnancy class B drug, and thus schistosome-infected women in some regions where schistosome infection is endemic, such as the Philippines, are excluded from treatment during pregnancy and lactation.

In mouse models, schistosomiasis is associated with strikingly poor pregnancy outcomes, even at low infection intensity. CBA/J mice infected with approximately 15 cercariae of Schistosoma mansoni produce 66% fewer viable litters than uninfected controls, a result attributable to increased rates of abortion (20% versus 1%), maternal death (5% versus 0%), and infanticide (42% versus 23%) (3). In addition, pup weight at 2 weeks was significantly lower in pups born to infected mothers (3). In separate studies, heavier (40 cercariae/mouse) and chronic infections (4) of C57BL/6 mice resulted in similar profoundly adverse birth outcomes, including increases in maternal death rate, abortion, and infanticide from infected pregnancies. Interestingly, these data were not replicated in a study involving exposure to S. japonicum cercariae 1 to 2 weeks postconception (37). These data suggest that either (i) different schistosome species (S. mansoni versus S. japonicum) influence the health of the pregnancy differently or (ii) schistosome infection is most detrimental if the exposure and initial response to infection occur prior to successful breeding.

As might be expected given the discordance in placental structure, the effect of schistosomiasis on pregnancy outcomes appears to be somewhat species specific. Pigs infected with ∼9,000 cercariae of S. japonicum early in gestation (week 4) produce offspring that appear to be smaller and weaker and fail to thrive, while pigs infected prior to insemination or late in gestation produce normal offspring (38). To our knowledge, no studies have examined the specific effect of schistosome infection on the placenta itself in either animal models or in humans. Reports suggest that maternal schistosomiasis in humans is associated with an increased risk of maternal anemia (5, 6), lower birth weight (7, 8), and risk of low birth weight (LBW) and preterm delivery (8, 39). In addition, we recently demonstrated that pregnant women infected with S. japonicum had increased proinflammatory cytokine levels in maternal peripheral, placental, and cord blood (17).

Two observational studies, as well as a recently completed randomized controlled trial (RCT) in Uganda, have addressed the role of schistosomiasis in the pathogenesis of adverse pregnancy outcomes in humans (7–9). A large (n = 592) observational study in which S. japonicum infection was examined in women in China revealed lower birth weights among firstborn children from women with schistosomiasis than among uninfected women; however, no adjustment was made for potential confounders such as maternal nutritional status and socioeconomic status. A separate case-control study of Schistosoma haematobium-infected pregnant women in Ghana revealed differences in birth weight of infants born to infected versus uninfected women, although this was significant only in premature deliveries (n = 8 S. haematobium-infected women) (8). The results from this study were also likely influenced by selection bias, with infected women being recruited only if they felt poorly enough to present to the hospital for care. This possible bias, coupled with the low sample size in the single stratum in which effects were demonstrated, makes the results of this study difficult to interpret.

Finally, a recent RCT conducted in Uganda addressed the efficacy of PZQ and/or albendazole given during the second or third trimester (9). Treatment with praziquantel was not associated with decreased risk of maternal anemia or low birth weight compared to placebo. An RCT nearing completion in the Philippines (ClinicalTrials.gov, registered study number NCT00486863) will address the efficacy of praziquantel given earlier (12 to 16 weeks' gestation) and in the context of infection with S. japonicum, which is considered the more virulent species.

In this study, we examined for the first time the impact of schistosome infection on trophoblast cells of the human placenta. The human placenta is composed of a multitude of cell types, with the syncytiotrophoblast layer comprising the functional unit responsible for nutrient, waste, and gas exchange between mother and fetus. Although isolated reported studies have demonstrated direct trafficking of schistosome worms and whole eggs to the placental compartment in humans, this event appears to be relatively rare and is difficult to assess (40). Exposure and transfer of soluble worm and egg antigens across the placenta, however, is well documented and occurs in the majority of human pregnancies complicated with schistosome infection (41). We therefore utilized schistosome SEA in our experiments as a starting point for examination of the effect that schistosome infection might have on trophoblast cells. As a follow-up experiment, we also placed trophoblast cells in culture with plasma from pregnant women infected with schistosomiasis.

We previously reported that schistosome infection results in a proinflammatory cytokine response detectable in both maternal and fetal blood at term (17). Herein, we report results that establish that the trophoblast cells of the placenta contribute, in part, to this proinflammatory response. In addition to direct production of proinflammatory cytokines, trophoblasts collected at term and allowed to form syncytia in vitro produce increased amounts of the chemokine MIP-1α in response to SEA, suggesting that trophoblast cells might alter the immune cell repertoire at the maternal-fetal interface in response to SEA exposure. Schistosome egg antigen has been previously shown to induce MIP-1α production by macrophages during granuloma formation in a murine model (42). In this study, we extended this paradigm to the placenta, with implications for an altered immune cell environment at the maternal-fetal interface.

Trophoblasts have been shown to produce a variety of cytokines, depending on the gestational age, differentiation status of the trophoblast cells, and placental environment (43–46). The data described herein are similar to the effects seen in trophoblast cells in response to other inflammatory agents, such as lipopolysaccharide (LPS) (47); however, trophoblast cells also secrete cytokines throughout a routine gestation. In the course of normal pregnancy, IL-10 is upregulated, and it is thought to play an important role in the maintenance of immune tolerance toward the developing fetus (48). It is thus somewhat surprising that we did not observe an increase in IL-10 secretion by trophoblast cells when they were exposed to SEA. This finding, however, is consistent with the proinflammatory Th1 signature that we observed both in the trophoblasts themselves (Fig. 1) and in maternal and newborn compartments following exposure to SEA (17).

The recapitulation of SEA-induced trophoblast activation by plasma from pregnant women with schistosomiasis further strengthens evidence for the association between schistosomiasis and poor pregnancy outcomes. Our data suggest that the levels of circulating schistosome-associated antigens in maternal plasma are sufficient to directly activate trophoblasts in vitro. Despite matching on measured potential confounders (geohelminths, socioeconomic status, etc.), we recognize that residual, unmeasured confounders related to both S. japonicum infection and trophoblast activation might also partly explain this relationship.

Schistosome egg antigens have been shown to stimulate both the ERK1/2 and p38 MAPK signaling pathways in dendritic cells (33, 34) and macrophages (35). Classical TLR activators, such as LPS (TLR4), Pam3cys (TLR2), and flagellin (TLR5), also activated these pathways to various degrees in dendritic cells (34), lending credence to the possibility of SEA working through a TLR-mediated pathway. Despite these findings, the ligand-receptor pairs and consequent signaling cascades initiated by SEA have not been reported in cells that are not classic immune cells (e.g., trophoblast cells). In this study, we demonstrated that SEA can robustly activate the ERK1/2 and p38 MAPK pathways, as well as the JNK MAPK pathway in trophoblast cells. In addition, the NF-κB pathway is activated in trophoblasts in response to SEA. These data are consistent with the proinflammatory response seen following SEA exposure of trophoblasts.

The schistosome egg antigen represents an admixture of many different proteins, including cytoplasmic and secretory products. Some of the most abundant products secreted from schistosome eggs have biological activity from both protein and carbohydrate epitopes (49), and antibody responses directed against SEA comprise both antiglycan and antipeptide responses (50). In addition, at least one of the major glycoproteins secreted by S. mansoni eggs, interleukin-4-inducing principle from schistosome eggs (IPSE/alpha-1), is known to stimulate cytokine responses by basophils, albeit prominently skewed toward a Th2 environment (51). Our manipulation of SEA, directed at either the glycan or peptide epitopes, resulted in reduced proinflammatory responses in trophoblast cells. Although these data suggest that both sugars and proteins in SEA play a role in initiating a proinflammatory response within trophoblast cells, they by no means preclude other possibilities, such as the importance of multiple carbohydrate structures from the same molecule initiating a proinflammatory response. More detailed analysis of the role of glycans and/or proteins in this response is warranted. In addition, neither manipulation abrogated the cytokine response to the level of untreated cells, suggesting that both components of SEA might activate the syncytiotrophoblast. These data are in keeping with the multitude of signaling pathways that we have observed to be initiated by SEA in trophoblast cells. Studies designed to specifically identify the ligand-receptor pairs responsible for SEA-induced trophoblast activation are currently ongoing.

In conclusion, we have demonstrated that antigens released from schistosome eggs and present in maternal blood activate human trophoblast cells to secrete specific proinflammatory cytokines and chemokines. Additional data suggest that SEA might influence other basic functions of the trophoblast cells, such as hormone production. However, the changes in progesterone production we observed did not reach statistical significance and thus should be interpreted with caution. We do show, however, that SEA activates a variety of signaling pathways in the trophoblast cell, including members of the MAPK family and the NF-κB pathway. These effects are mediated through both carbohydrate and protein epitopes. To our knowledge, this is the first report of a direct effect of SEA on cells of the placenta, and it represents an important initial characterization of how schistosomiasis might influence the placenta and subsequent health of the pregnancy.