Malaria in pregnancy regulates P‐glycoprotein (P‐gp/Abcb1a) and ABCA1 efflux transporters in the Mouse Visceral Yolk Sac

Abstract Malaria in pregnancy (MiP) induces intrauterine growth restriction (IUGR) and preterm labour (PTL). However, its effects on yolk sac morphology and function are largely unexplored. We hypothesized that MiP modifies yolk sac morphology and efflux transport potential by modulating ABC efflux transporters. C57BL/6 mice injected with Plasmodium berghei ANKA (5 × 105 infected erythrocytes) at gestational day (GD) 13.5 were subjected to yolk sac membrane harvesting at GD 18.5 for histology, qPCR and immunohistochemistry. MiP did not alter the volumetric proportion of the yolk sac's histological components. However, it increased levels of Abcb1a mRNA (encoding P‐glycoprotein) and macrophage migration inhibitory factor (Mif chemokine), while decreasing Abcg1 (P < 0.05); without altering Abca1, Abcb1b, Abcg2, Snat1, Snat2, interleukin (Il)‐1β and C‐C Motif chemokine ligand 2 (Ccl2). Transcripts of Il‐6, chemokine (C‐X‐C motif) ligand 1 (Cxcl1), Glut1 and Snat4 were not detectible. ABCA1, ABCG1, breast cancer resistance protein (BCRP) and P‐gp were primarily immunolocalized to the cell membranes and cytoplasm of endodermic epithelium but also in the mesothelium and in the endothelium of mesodermic blood vessels. Intensity of P‐gp labelling was stronger in both endodermic epithelium and mesothelium, whereas ABCA1 labelling increased in the endothelium of the mesodermic blood vessels. The presence of ABC transporters in the yolk sac wall suggests that this fetal membrane acts as an important protective gestational barrier. Changes in ABCA1 and P‐gp in MiP may alter the biodistribution of toxic substances, xenobiotics, nutrients and immunological factors within the fetal compartment and participate in the pathogenesis of malaria‐induced IUGR and PTL.


| INTRODUC TI ON
The harmful effects of malaria in pregnancy (MiP) include high rates of maternal anaemia and death, as well as spontaneous abortion, fetal intrauterine growth restriction (IUGR), preterm labour (PTL), low birthweight, foetal/neonatal demise and impaired postnatal cognitive and neurosensory development. [1][2][3] MiP is characterized by adherence of plasmodium falciparum-infected erythrocytes to specific syncytiotrophoblast glycosaminoglycans, a condition referred to as "placental malaria". 3,4 In response, the placental barrier may undergo various adaptive changes that comprise activation and migration of immune cells, alteration of cytokine and chemokine output, 2,5 intervillositis, decreased enrichment of syncytiotrophoblast microvilli and lower expression of placental glucose (GLUT1), amino acid (SNAT1, SNAT2, Cat1, Lat1 and 4F2hc), solute carrier (SLC) drug uptake (Oatp2b1, Oct3, Ent1 and Ent2) and nutrient/drug efflux ABC transport systems (ABCA1, BCRP and P-gp). [6][7][8][9][10][11] The ATP-binding cassette (ABC) transport system is highly expressed in different trophoblastic lineages. [12][13][14] Its disruption has been associated with IUGR, PTL, preeclampsia and chorioamnionitis. 13,[15][16][17] In this context, reports from different groups have demonstrated that it participates in the adaptive trophoblastic responses to maternal infection. 9,12,14,[16][17][18] Comprised of 50 transporters divided into seven sub-classes ranging from ABCA through ABCG, 19 these transmembrane transporters are expressed in different cell types, including those from biological barriers. 12 They are responsible for the efflux of diverse endogenous and exogenous substrates, from one side to the other of the plasma membrane. 20 There are a multitude of endogenous substrates including lipids, phospholipids, cytotoxic oxysterols, amino acids, steroid hormones, folate, metabolites and pro-inflammatory cytokines and chemokines. Exogenous substrates comprise environmental toxins (bisphenol A, ivermectin, pesticides) and clinically relevant drugs (antibiotics, antiretrovirals, antidepressants and synthetic glucocorticoids). ABC transporters are major components of the immune response and mediate nutrient transfer and fetal protection against drugs and environmental toxins that may be circulating in the maternal blood. 12,20,21 The lipid transporters, ABCA1 and ABCG1 (encoded by the ABCA1 and ABCG1 genes, respectively), and the multidrug resistance transporters, P-glycoprotein (P-gp/ABCB1) and breast cancer-related protein (BCRP/ABCG2) are amongst the best-described ABC transporters in the syncytial barrier. ABCA1, P-gp and BCRP are predominantly expressed in the apical membrane of the syncytiotrophoblast, and efflux their substrates from the fetal side into the maternal circulation, whereas ABCG1 is localized in the syncytiotrophoblast's basolateral membrane and efflux cholesterol into the fetal circulation. 12 Plasmodium berghei ANKA (PBA) infection is largely used to mimic malaria infection in mice. In pregnant BALB/c mice, PBA decreases placental Abca1, Abcb1a and Abcb1b (both encoding P-gp in rodents) and Abcg2. 11 Similarly, in pregnant C57BL/6 dams, PBA leads to IUGR, PTL and impairs the placental expression of ABCA1/Abca1, P-gp/Abcb1b and BCRP/Abcg2, 9 suggesting that MiP has the potential to increase fetal exposure to a range of clinically relevant substrates capable of profoundly impacting fetal outcome, via dysfunction of ABC transporters in the placental barrier.
The human yolk sac is the major hematopoietic site 22 and is the source of primordial germ cells in early pregnancy. 23 It is estimated that human yolk sac is viable until the 49th day of gestation. 24 Emerging evidence suggests the yolk sac is an important nutrient exchange site between the coelomic fluid and the fetal capillaries present in its mid-mesodermal layer; or alternatively, it delivers nutrients to the embryo primitive gut via the vitelline duct. [24][25][26] In mice, the visceral portion of the yolk sac involves the embryo amniotic membrane and is functional throughout pregnancy, mediating the transport of critical substances from the mother to the fetus, acting as a syncytiotrophoblast equivalent throughout intrauterine development. 27,28 It is known that human 24 and mouse 24,29 yolk sac express different ABC transport carriers, suggesting that this membrane acts together with the placenta and other fetal membranes to form an efficient protective barrier during gestation.
Despite the importance of the yolk sac for embryo and fetal development, no studies have investigated the effects of MiP on yolk sac morphology and efflux transport expression. In the present study, we hypothesized that MiP, in a murine model of malaria-induced IUGR and PTL, modifies the yolk sac morphology and efflux transport potential, via modulation of key ABC efflux transporters. Improved knowledge as to how the yolk sac responds to MiP may improve the understanding of the mechanisms by which malaria impacts pregnancy outcome.

| Animals
An animal model of severe experimental malaria, which recapitulates many of the features of the human malarial disease in pregnancy, ABCA1 and P-gp in MiP may alter the biodistribution of toxic substances, xenobiotics, nutrients and immunological factors within the fetal compartment and participate in the pathogenesis of malaria-induced IUGR and PTL.

K E Y W O R D S
ABCA1, ABCG1, Breast Cancer Resistance Protein (BCRP), malaria in pregnancy (MiP), P-glycoprotein (P-gp), Plasmodium berghei ANKA, yolk sac including placental malarial, IUGR, low birthweight and PTL, was used as previously described. 9 This model resulted in a 20% rate of PTL in infected dams; however, all analyses were undertaken in the yolk sac of fetuses born at GD18.5 (ie term). Briefly, C57BL/6 female mice (8 to 10 weeks old) were housed at room temperature (22°C), with a light cycle of 12-hours light/12-hours dark and were mated. Animals were maintained on a commercial Nuvilab ® CR1 (Nuvilab, PR, Brazil) chow and water ad libitum. Mating was confirmed by the presence of a vaginal plug and was defined as gestational day (GD) 0.5. Pregnant dams were injected intraperitoneally (ip) with 5 × 10 5 erythrocytes infected with Plasmodium berghei ANKA (PBA group) or with saline (control group) at GD13.5. Killing was performed at GD18.5, using a pentobarbital overdose (300 mg/kg ip), followed by maternal and fetal decapitation. Six PBA-infected dams from our previous cohort, 9 which exhibited peripheral parasitaemia (approximately 16% of infected erythrocytes) at GD18.5, were included in this study. Approval was ob-

| Volumetric proportion assessment of mouse visceral yolk sac
The visceral yolk sac of PBA-infected and control animals (n = 6/group) was dissected and immediately immersed in buffered 4% paraformaldehyde phosphate solution (24 hours). After fixation, samples were dehydrated in an ethanol series, diaphanized in xylene and embedded in paraffin wax (Histopar, SP, Brazil). Sections (4 μm) were stained with haematoxylin and eosin (H&E). Image acquisition and analysis were performed on a Zeiss Axiolab 1 photomicroscope, coupled with a CCD camera and computer running Zeiss AxioVision (Carl Zeiss, NY, EUA).
The histomorphological analysis was undertaken using the Fiji ImageJ

| qPCR
Samples of the visceral yolk sac (n = 6/group) were immersed in RNA Later solution (Invitrogen, MA, USA) and frozen at −20°C until further processing. Extraction of total RNA was undertaken using TRIzol (Invitrogen) with homogenization conducted using a TissueLyser LT (Qiagen, Hilden, Germany). Concentration and purity of samples were determined by spectrophotometric analysis (Implen nanophotometer, Munich, Germany). Total RNA (1 μg) was reverse-transcribed into cDNA, using the High Capacity cDNA Reverse Transcription kit (Applied Byosystems, CA, USA), according to the manufacturer's instructions. cDNA (4 μL) was added to the primer-containing mix (intron-exon spanning) of each gene of interest (Table 1), as well as with the fluorescent DNA interlayer Eva Green (Biotium, CA, USA), in a total volume of 6 μL. The quantification of each gene of interest relative to the reference genes, Ywhaz, Ppib and Gapdh (table 1), was performed in the QuantStudio Real-Time PCR (Applied Biosystems, CA, USA), with the following amplification cycles: initial denaturation at 50°C for 2 minutes and then at 95°C for 10 minutes, followed by 40 cycles of denaturation at 95°C for 15 seconds. Annealing was performed at 60°C for 30 seconds and extension at 72°C for 45 seconds. Differences in mRNA gene expression were calculated according to the 2 −ΔΔCT method, 31 and the assay was considered acceptable when its efficiency ranged from 95% to 105%.

| Statistical method
Values are expressed as mean ± standard error of the mean (SEM) and were analysed using PRISM program (GraphPad Software Inc, CA, USA). Statistical assessment of volumetric dimensions, qPCR and immunohistochemistry data was undertaken in the yolk sac of fetuses exhibiting the closest placental weight to the mean weight of all placentae from each litter. Thus, "n" represents the number of litters. 9,33,34 After confirming the non-parametric distribution of the data and exclusion of outliers (Grubbs' test), differences between PBA-infected and control groups were assessed by non-parametric Mann-Whitney test, when comparing two variables, and by Kruskal-Wallis test, followed by Dunn's post-test, when comparing more than two variables. Statistical significance was considered when P < 0.05.

| Malaria does not affect visceral yolk sac histomorphological parameters
The wall of the visceral yolk sac in both PBA-infected and control animals exhibited its three typical layers: an outer (uterine-facing) endodermic epithelium, a blood vessel-enriched mesodermal mid-layer and an inner (amnion facing) mesothelial layer. There were no visible differences in gross morphology of these layers between infected and control groups ( Figure 1A,B). Additionally, MiP did not affect the volumes of the histological components of the yolk sac, comparing PBA-infected and control groups ( Figure 1A). Of note, constitutive endodermic epithelium comprised approximately 66% of the total  approximately 5% of the total number of yolk sac cells ( Figure 1C).

| Malaria in pregnancy modifies P-gp and Abca1 protein in the yolk sac
We next evaluated the effects of PBA infection on protein localization and expression (semi-quantitative) of the ABC transporter that exhibited altered gene expression following infection. P-gp was enriched in the apical membrane and cytoplasm of the outer endodermic epithelium and inner mesothelial cells. The endothelium of blood vessels, localized in the mesodermal mid-layer, exhibited faint or no P-gp labelling, whereas the mesodermal layer connective tissue was completely negative for P-gp ( Figure 3A,B). In PBA-infected pregnancies, distribution of yolk sac P-gp was similar to controls; however, semi-quantitative analysis revealed increased P-gp protein in the plasma membranes of the endodermic epithelium and mesothelium (P < 0.05), with no changes in the total area of P-gp immunolabelling ( Figure 3F,G).
ABCG1 protein was detected in both the endodermic epithelium and the endothelium of mesodermal vessels of the yolk sac, with lower levels of ABCG1 detected in the mesothelium. No ABCG1 was detected in the mesodermal layer connective tissue ( Figure 4A,B).
There were no differences in ABCG1 immunostaining area and intensity between PBA-infected and control pregnancies ( Figure 4F,G).

F I G U R E 3 P-glycoprotein (P-gp) is localized in distinct cellular barriers of the murine yolk sac and is up-regulated by malaria in pregnancy (MiP). A-E, Representative immunohistochemistry (IHC) images from murine yolk sac sections of control (A) and (B)
Plasmodium berghei ANKA-infected dams at GD18.5. C-D: yolk sac (C) and placental (D) negative controls. E, Placental positive control showing P-gp immunoreactivity in labyrinthine and junctional zone cells. F-G: Semiquantitative evaluation of the area (F) and intensity (G) of P-gp immunolabelling. Arrowheads = endodermic epithelium; thin arrows = mesothelium; * = mesodermal blood vessels. Statistical differences were tested by Mann-Whitney test. *P < 0.05. Data are presented as mean ± SEM (n = 6/group). Magnification bars represent 100 µm A disconnect between mRNA and protein expression patterns for some ABC transporters has been previously reported in the placenta. 16,35 As such, we conducted immunohistochemical analysis of ABCA1 lipid and BCRP drug transporters in the mouse yolk sac.
ABCA1 was primarily localized to the endodermic epithelium, with heterogeneous intensity. In addition, the cytoplasm of some epithelial cells was positive for ABCA1. The mesothelium and the endothelium of mesodermal vessels were also stained, but with less intensity compared to the endodermic epithelium ( Figure 5A,B). No ABCA1 staining was observed in the connective tissue. Semi-quantitative analysis showed that both area and intensity of ABCA1 staining were increased in the endothelium of the mesodermal blood vessels (P < 0.05) in MiP, with no changes in the endodermic epithelium or in the mesothelium (Figure 5F,G).
BCRP staining was localized to the endodermic epithelium and in the endothelium of mesodermal vessels, with lower levels in the mesothelium regions of the yolk sac. Semi-quantitative analysis revealed no effect of MiP on BCRP in any regions of the yolk sac ( Figure 6A,B). As predicted, mouse placentae-positive controls exhibited intense ABCA1, ABCG1, BCRP and P-gp staining in  Previously, we have demonstrated that MiP impairs the placental expression of ABCA1/Abca1, P-gp/Abcb1b and BCRP/Abcg2, which were accompanied by increased placental levels of Cxcl1 and Ccl2 mRNA and up-regulation of maternal Il1-β, Il-6, Cxcl1 and Ccl2. 9 In the previous study, we provided evidence that MiP has the potential to increase foetal exposure to drugs and environmental toxins, F I G U R E 5 ABCA1 staining is increased in the endothelium of the mesodermal blood vessels in malariainfected dams. A-E, Representative immunohistochemistry (IHC) images from murine yolk sac sections of control (A) and (B) Plasmodium berghei ANKAinfected dams at GD18.5. C-D: yolk sac (C) and placental (D) negative controls.

| D ISCUSS I ON
(E) Placental positive control showing ABCA1 immunoreactivity in labyrinthine and junctional zone cells. F-G: Semiquantitative evaluation of the area (F) and intensity (G) of ABCA1 immunolabelling. Arrowheads = endodermic epithelium; thin arrows = mesothelium; * = mesodermal blood vessels. Statistical differences were tested by Mann-Whitney test. *P < 0.05. Data are presented as mean ± SEM (n = 6/group). Magnification bars represent 100 µm via down-regulation of major placental drug and nutrient ABC efflux transporters. 9 Moreover, other studies have demonstrated the sensitivity of placental P-gp and BCRP to infection. C57BL/6 mice exposed to sublethal (foetal) LPS in mid-pregnancy, exhibited impaired placental P-gp activity, 18 while polyinosinic:polycytidylic acid (PolyI:C, a viral mimic) had no effect on placental P-gp activity. 36 Human first-trimester placentae exposed to lipopolysaccharide (LPS, a bacterial antigen) exhibited decreased P-gp/ABCB1 and BCRP/ABCG2 levels. 37 In contrast, human chorioamnionitis in 2nd trimester, often resulting from polymicrobial infection, 38 resulted in decreased P-gp levels but elevated BCRP expression. 16 Further, extravillous trophoblast (HTR8/SVneo)-like cells treated with LPS and single-stranded RNA (ssRNA, another viral mimic) showed a profound down-regulation of BCRP/ABCG2. 14 Together, these studies demonstrate that the nature of the infective stimuli determines the type of the drug efflux transporter response in different trophoblast lineages.
Notably, in the present study, we did not observe substantial up-regulation of pro-inflammatory factors in the yolk sac following has the potential to disrupt the human yolk sac protective barrier and thus impact early pregnancy outcome. This highlights the importance of investigating the human yolk sac response to infection.

ACK N OWLED G EM ENTS
We thank Carlos Henrique da Silva for technical assistance and Drs.
Cristina Fonseca Guatimosim and Patricia Massara Martinelli for assistance in imaging analysis and acquisition.

CO N FLI C T O F I NTE R E S T S
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.