Differential Activation of Fetal Hofbauer Cells in Primigravidas Is Associated with Decreased Birth Weight in Symptomatic Placental Malaria

Background Placental malaria is a leading global cause of low birth weight neonates, especially in first-time mothers. To better understand the role of innate immunity in placental malaria, we investigated the relationships between histopathological markers of placental malaria, fetal and maternal macrophage responses, and perinatal outcomes in a cross-sectional case control study of pregnant women presenting with symptomatic malaria at the time of delivery. Results Primigravidas showed increased hemozoin deposition in placental villi (p=0.02), syncytiotrophoblasts (p=0.01), and fetal Hofbauer cells (p=0.01). The percentage of hemozoin-positive villi negatively correlated with infant birth weight (regression coefficient [b] = -0.03 kg decrease in birth weight per % increase in hemozoin-positive villi, p=0.035). Malaria-infected placentas showed a twofold increase in Hofbauer cells (p<0.001) and maternal macrophages (p<0.001). Placental malaria was associated with a threefold increase in the percentage of M2 maternal macrophages (19.2% vs 6.4%, p=0.01). Primigravidas showed a significant decrease in the Hofbauer cell M2-percentage in placental malaria (92.7% vs. 97.0%, p=0.04), which was predictive of infant birth weight (b=0.08 kg increase in birth weight per % increase in M2 Hofbauer cells, p=0.001). There was no association between maternal macrophage response and infant birth weights. Conclusions Placentas with malarial infection had increased numbers of fetal Hofbauer cells in the villous stroma and maternal macrophages in the intervillous space. In primigravidas, decreased anti-inflammatory M2-type Hofbauer cells were predictive of lower birth weight. M2-type maternal macrophages were increased in placental malaria, but there was no association with gravidity or birth weight. These results suggested a protective role of M2 Hofbauer cells in fetal growth restriction.


Background
Each year, approximately 85 million pregnant women worldwide are at risk of Plasmodium falciparum malaria infection [1]. Maternal malaria (particularly in primigravidas) leads to poor perinatal outcomes, including low birth weight (BW), fetal growth restriction, and preterm labor [2,3].
Placental malaria (PM) is characterized by the accumulation of P. falciparum-infected red blood cells in the intervillous spaces of the placenta. Deposits of hemozoin (HZ, a malarial by-product) are found in several locations-within the intervillous space, maternal monocytes, the villous cores, and perivillous fibrin deposits [2]. HZ in maternal monocytes circulating within the intervillous space has been consistently associated with low BW [2]. Previously, we found that syncytiotrophoblast denudation was a feature of chronic PM, resulting in direct exposure of the villous cores to maternal blood [4]. Syncytiotrophoblast denudation was associated with villous HZ deposition and maternal monocyte infiltration. HZ stimulates syncytiotrophoblast recruitment of 2 Malaria Research and Treatment monocytes in vitro [5] and triggers other inflammatory pathways [6]. The placental inflammatory response has been associated with poor fetal growth [7][8][9], possibly through dysfunctional vascularization or nutrient transport [2,10,11].
Surprisingly little is known about the role of the fetal Hofbauer cell (HBC), the placental resident macrophage, in PM. HBCs are tissue macrophages localized within the stromal core of the placental villus and are the developing fetus's primary immune cells at the maternal-fetal interface. The functional phenotypes of macrophages are diverse and dynamic, but are broadly classified into two groups: the "classically activated," M1-type, proinflammatory macrophage and the "alternatively activated," anti-inflammatory, M2-type macrophage. M1 macrophages form the first line in host defense against a variety of bacteria, protozoa, and viruses, as well as in antitumor immunity. In contrast, the M2 subset is characterized by diverse immunosuppressive activity and includes wound-healing macrophages, IL-10-secreting regulatory macrophages, and tumor-associated macrophages that suppress antitumor immunity. Highly adaptive to their local microenvironment, macrophages have been shown to switch their programming from one functional phenotype to another in response to local signals and are thought to play an important role in maintaining the balance between inflammation (and associated tissue destruction) and the restoration of tissue homeostasis after infection and local injury [12]. Although studies are sparse, HBCs play an anti-inflammatory role in the placenta and may contribute to maternal-fetal tolerance, and most studies have found predominately M2 HBCs in normal placentas [13][14][15]. Present early in gestation, HBCs are predominately in the M2 anti-inflammatory phenotype and are thought to maintain maternal-fetal tolerance [16][17][18]. In pathologic states, HBCs are implicated in villitis of unknown etiology, where destructive villous inflammation is associated with poor perinatal outcomes such as preterm labor and intrauterine growth restriction [19].
Here, we characterized the fetal and maternal macrophage responses to PM at the maternal-fetal interface and investigated their association with placental HZ deposits. We also explored the correlations with perinatal outcomes.

Participant Recruitment and Tissue
Sampling. The patients and samples in this study were described previously [4]. Briefly, placental biopsies were obtained from singleton deliveries at the Kingasani Maternity Hospital, Kinshasa, Democratic Republic of the Congo. Participants were categorized as malaria-negative controls (n=17) or malariapositive cases based on the presence of malaria symptoms and peripheral blood smear. Malaria-negative women received antenatal care at the Kingasani Maternity Hospital and intermittent preventative therapy (IPT) per standard of care (two doses of sulfadoxine pyrimethamine for HIVnegative women and three doses for HIV-positive women). Malaria-negative women had no symptoms of malaria upon presentation to the labor and delivery unit, and peripheral smear confirmed the absence of peripheral malaria. Malariapositive women did not receive antenatal care at Kingasani Maternity Hospital and presented at this institution at the time of delivery with symptoms of malaria; their IPT status was unknown. Malaria-positive women had symptoms such as fever, headache, and joint and muscle pain. Diagnosis of malaria was confirmed with peripheral blood smear test at the time of delivery. Exclusion criteria for this study were common pregnancy complications such as hypertension, preeclampsia, chorioamnionitis, or maternal anemia. The following maternal and neonatal data was collected for all participants: maternal age, gestational age, gravidity, and infant birth weight. Other demographic variables were unknown.
Placental biopsies collected were obtained within 20 minutes of delivery and transferred to 10% neutral buffered formalin (VWR) 10:1 (ml:gram wet weight), made fresh to avoid crystallization artifact. To account for heterogeneity across the maternal-fetal interface, five full-depth biopsies were collected from the center (n=2) and periphery (n=3). Biopsies were fixed for 24 h, transferred to 70% ethanol, paraffin embedded, and sectioned (5 m).
For double staining experiments, sections were first incubated with mouse anti-CD163 (an M2 marker) (10D6, 1:150, Pierce), followed by biotin-conjugated donkey anti-mouse, ABC peroxidase, and the SG substrate (Vector Laboratories). Sections were washed, reblocked, and incubated with mouse anti-CD68 (a pan-monocyte/macrophage marker) (KP1, 9.4 g/ml, Zymed). Binding was detected as above and the color developed with AEC substrate. For each sample, six randomly chosen fields representing the entire depth of the biopsy were scored for the number of CD163+/CD68+ (M2) and CD163-/CD68+ (M1) cells in the intervillous (maternal) or intravillous (fetal) compartments. Each sample was evaluated for the number of immunopositive cells of both phenotypes by a blinded scorer on two independent occasions. The percentage of M2 macrophages was calculated as the mean number of (CD163+/CD68+ cells)/(total CD68+ cells) x100.

Statistical Analyses.
Data were analyzed with the Welch's t-test, Wilcoxon rank sum test, or logistic regression as appropriate. Histopathologic correlations were determined by Spearman's rank order test with Bonferroni's correction. Linear regression was performed to detect associations with perinatal outcomes. P-values of <0.05 were considered statistically significant. Analyses were performed in Stata 13.

Results
Although there was a trend toward decreased rates of preterm labor and increased rates of low BW and small for gestational age in patients with malaria, in our cases, PM was not associated with differences in maternal age, gestational age, or infant BW (Table 1). Patients with malaria were less likely to be multigravida (odds ratio [OR] 0.2, confidence interval [CI] 0.0-0.9, p=0.04).
First, we characterized the HZ-containing villous cells and examined their association with perinatal outcomes. Fetal and maternal macrophages were identified according to their histological location. Experiments that colocalized KRT7 (a trophoblast marker) and CD68 (a panmacrophage marker) revealed that HZ is found within syncytiotrophoblasts (Figures 1(g)-1(h)), HBCs, and maternal macrophages in PM samples (Figures 2(h)-2(g)). As expected, no HZ was identified in the control placentas (Figures 1(c) and 2(c)). In primigravidas, compared to multigravidas, there were increased percentage of HZ-positive villi (p=0.02) and increased numbers of HZ-positive syncytiotrophoblasts (p=0.01) and HBCs (p=0.01, Figure 3(c)). Linear regression showed a negative correlation between the percentage of HZ-positive villi and infant BW (regression coefficient [b] = -0.03 kg decrease in BW per % increase in HZ-positive villi, p=0.035, Figure 3(d)). Correlation analysis also showed a statistically significant association between the percentage of HZ-positive villi and birth weight (Spearman correlation coefficient r= -0.39, p=0.024). No associations were found between HZ localization and gestational age, preterm labor, low BW, or small for gestational age (data not shown).
Linear regression was performed to determine associations between perinatal outcomes and the numbers and phenotypes of HBCs and maternal macrophages. Increasing M2-percentage in HBCs was highly predictive of BW in primigravidas (b=0.08 kg increase in BW per % increase in M2 HBCs, p=0.001, Figure 4(c)), but not in multigravidas (b=-0.02 kg, p=0.45), nor with all patients (b=0.003 kg, p=0.89). Upon stratification by malaria, we also found a significant association with HBC M2% in primigravidas infected with malaria (b=0.10 kg, p=0.004), but not in multigravidas (b= -0.08 kg, p=0.31). There were no associations between M2-percentage and BW in maternal macrophages (data not shown).
We next explored potential associations between HZ localization and the fetal and maternal inflammatory responses (numbers and M2-percentages). Correlations were analyzed for PM and gravidity (Table 2). HZ-positive HBCs correlated with the percentage of HZ-positive syncytiotrophoblasts (Spearman's correlation coefficient [r]=0.85, p<0.001). The maternal macrophage M2-percentage was associated with HZ-positive syncytiotrophoblasts (r=0.7, p=0.02). Interestingly, the maternal M2-percentage was associated with HZ-positive HBCs (r=0.88, p=0.03) in primigravidas, but not multigravidas, suggesting that HZ-positive HBCs may influence maternal macrophage polarization in a gravidity-dependent manner. We found no associations between villous HZ localization and HBC numbers nor phenotype. Further, there were no correlations found between numbers of maternal macrophages and HBCs, nor M2-percentages.

Discussion
In this study, we characterized the cellular localization of HZ in placental villi and assessed the expression of the M2 marker CD163 in association with the fetal and maternal macrophage response in patients with symptomatic placental malaria at the time of delivery. We also investigated the relationships between these histopathologic markers of PM and inflammation, gravidity, and birth outcomes. Poor fetal growth is a hallmark of malaria infection in pregnancy, particularly in first-time mothers. We found a graviditydependent HBC response to PM and evidence that a shift in the M2 balance away from the alternative pathway is associated with increased infant BW. In primigravidas, the    fetuses had a decrease in anti-inflammatory M2-percentage of HBCs with PM, compared to gravidity-matched controls. Although the phenotypic shift was subtle, it was highly predictive of infant BW, implicating a protective effect of M2 HBCs on fetal growth. This fetal inflammatory response could alter maternal-fetal tolerance and lead to disturbances in placental function. This phenomenon was not observed in multigravidas, perhaps due to partial maternal immunity from prior exposure to PM that decreased exposure to malarial antigens and tempered the overall inflammatory response which can be related to a previous study that described trained innate immunity induced by P. falciparum [23]. These findings suggest that fetal inflammation may be a novel mechanism for the increased susceptibility of primigravidas to fetal growth restriction in PM. Consistent with prior reports [24], we found increased numbers of maternal macrophages during PM infection. Furthermore, our phenotypic analysis showed an increased maternal M2-percentage with PM, but no difference was found when stratified by gravidity, although there was a trend toward increased M2-percentage in primigravidas. There was no correlation between maternal macrophage numbers nor phenotype and clinical outcomes. These results suggested that the fetal HBC response may be a more specific marker of adverse outcomes in this condition. However, the likely complex interactions between maternal and fetal macrophage responses warrant further investigation.
Intriguingly, both fetal and maternal macrophages are strongly recruited to the site of infection, but they were polarized in opposite directions. In the malaria-infected primigravida group, we found decreased M2 polarization of HBCs and a trend toward an increase in M2 maternal macrophages. These findings suggested distinct mechanisms of macrophage stimulation and polarization in fetal versus maternal macrophages. In the setting of decreased fetal antiinflammatory M2 HBCs against PM, it is possible that the increased maternal M2 response is an attempt to restore the maternal-fetal tolerance that is essential for pregnancy maintenance. Alternatively, the altered fetal M2 response could be that of a naïve immune system reacting to a more significant pathogen challenge in the primigravida. Future studies of the regulatory pathways that are involved in the distinct fetal and maternal innate immune responses may provide additional insights into this hypothesis. Our study also examined the associations between HZ deposition, gravidity, and perinatal outcomes. We characterized the cellular localization of villous HZ and found an increase in HZ-positive syncytiotrophoblasts and HBCs in primigravidas compared to multigravidas. HZ-positive syncytiotrophoblasts were associated with a maternal M2 response, which is consistent with prior studies [5]. However, there was no association with the fetal macrophage response, suggesting that HBCs are activated through a distinct, yet unknown, pathway.
The primary limitation of our study was the small sample size, which limited our power to detect associations between our histopathologic findings and clinical outcomes. Conclusions regarding preterm labor or fetal growth restriction may be confounded by the lack of confirmed obstetrical dating, which is often a significant limitation of PM studies in endemic settings. Furthermore, the participants with malaria did not receive prenatal care, so other potential contributors to birth outcomes and risk of PM could not   Figure 4: In primigravidas, activation of fetal macrophages was associated with decreased birth weight. (a) Increased numbers of fetal CD68+ HBCs (left) and CD68+ maternal macrophages (right) were found in PM (gray) as compared to uninfected controls (white). Upon stratification by gravidity, there were increased HBCs and maternal macrophages irrespective of gravidity. Data is shown as median, interquartile range, and adjacent values. P-values were calculated with the Wilcoxon rank sum test. (b) Phenotypic characterization of HBCs (left) and maternal macrophages (right) showed a decrease in HBC M2 percentage in primigravidas with PM, but not multigravidas, as compared to uninfected controls. Overall maternal macrophages showed an increase in M2 polarization with PM. (c) Linear regression analysis showed that the HBC M2 percentage was highly predictive of birth weight in primigravidas, but not multigravidas or all gravidities. Abbreviations: PG, primigravidas; MG, multigravidas; HPF, high-powered field; HBC, Hofbauer cell; PM, placental malaria.
be assessed in this study. As such, other pregnancy complications apart from those diagnosed during labor cannot be excluded. Strengths of the study include the prospectively recruited cases and confirmation of active malaria infection by symptoms, peripheral blood smear, and placental histology.

Conclusions
This study identified a new potential mechanism of poor fetal growth in PM. We found that the fetal HBC response to PM was gravidity-dependent and characterized by a shift in primigravidas of the M2 percentage away from the 8 Malaria Research and Treatment Spearman's rank correlation coefficients with Bonferroni's correction (r) and p-values are indicated for each pairwise comparison between % villi with HZ (HZ+ villi), mean numbers of STB with HZ per 400x HPF (HZ+STB), mean numbers of HBC with HZ per 400x HPF (HZ+ HBC), mean numbers of HBC per 400x HPF (#HBC), M2 percentage of HBC (HBC M2%), mean numbers of maternal macrophage (MM) per 400x HPF (#MM), and M2 percentage of MM (MM M2%). Data is shown for the total of all malaria-positive patients (n=17) and stratified by gravidity (primigravidas n=9, multigravidas n=8). Statistically significant values, defined as p<0.05, are shown in bold. Abbreviations: HZ, hemozoin; STB, syncytiotrophoblast; HBC, Hofbauer cell; MM, maternal macrophage; HPF, high-powered field immune-tolerant M2 phenotype, suggesting activation. This phenotypic shift in HBCs was highly predictive of decreased BW and implicated the fetal inflammatory response as a mediator of susceptibility to poor outcomes in PM.