Human placental syncytiotrophoblasts restrict Toxoplasma gondii vertical transmission at two distinct stages and induce CCL22 in response to infection

Toxoplasma gondii is a major source of congenital disease worldwide, but the cellular and molecular factors associated with its vertical transmission are largely unknown. In humans, the placenta forms the key interface between the maternal and fetal compartments and forms the primary barrier that restricts the hematogenous spread of microorganisms. Here, we utilized primary human trophoblast (PHT) cells isolated from full-term placentas and human mid-gestation chorionic villous explants to determine the mechanisms by which human trophoblasts restrict and respond to T. gondii infection. We show that placental syncytiotrophoblasts, multinucleated cells that are in direct contact with maternal blood, restrict T. gondii infection at distinct stages of the parasite lytic cycle—at the time of attachment and also during intracellular replication. Utilizing comparative RNAseq transcriptional profiling, we also show that human placental trophoblasts at both mid- and late-stages of gestation induce the chemokine CCL22 in response to T. gondii infection, which relies on the secretion of parasite effector(s). Collectively, our findings provide new insights into the mechanisms by which the human placenta restricts the vertical transmission of T. gondii at early and late stages of human pregnancy, and demonstrate the existence of at least two interferon-independent pathways that restrict T. gondii access to the fetal compartment. Significance statement Toxoplasma gondii is a major source of congenital disease worldwide and must breach the placental barrier to be transmitted from maternal blood to the developing fetus. The events associated with the vertical transmission of T. gondii are largely unknown. Here, we show that primary human syncytiotrophoblasts, the fetal-derived cells that comprise the primary placental barrier, restrict T. gondii infection at two distinct stages of the parasite life cycle and respond to infection through the induction of the chemokine CCL22. Collectively, our findings provide important insights into the mechanisms by which human syncytiotrophoblasts restrict T. gondii infection at early and late stages of human pregnancy and identify the placental-enriched signaling pathways induced in response to infection.


103
Toxoplasma gondii is a major source of congenital disease, with ~200,000 global cases 104 of congenital toxoplasmosis reported each year (1). In the majority of instances (~80%), in utero 105 infections by T. gondii result in a range of severe birth defects, including ocular disease and 106 developmental delays, and can also result in fetal death (2). However, despite the clear impact of 107 T. gondii infections on fetal health, the mechanisms by which the parasite is transmitted from the 108 maternal bloodstream into the fetal compartment are largely unknown.

109
In eutherian organisms, the placenta serves as the sole source of gas, nutrient, and waste

117
In general, the pathways that exist in the human placenta to limit the vertical transmission 118 of microbes are poorly defined. Our previous studies in primary human trophoblast (PHT) cells, 119 which were focused on viral pathogens, have identified at least two potent antiviral pathways that 120 restrict viral replication in trophoblasts (3,4). However, these pathways do not appear to be 121 relevant during infection with non non-viral pathogens, including T. gondii (5). While studies in 122 placental explants suggest that the SYN layer is not permissive to T. gondii infection (6), the 123 mechanistic basis for SYN resistance is incompletely understood, as is whether the SYN layer 124 mounts any innate defense in response to parasite exposure. Moreover, while placental explant 125 models are useful in their recapitulation of placental structure, they are limited in their capacity to 126 dissect trophoblast cell type-specific pathways that might exist to limit T. gondii infection.

127
In this study, we interrogated the trophoblast cell-type specificity of T. gondii infection 128 5 utilizing PHT cells isolated from full-term placentas and identified two cellular mechanisms that 129 mediate SYN-specific resistance to T. gondii infection. In addition to discovering that SYNs

148
We found that PHT cells isolated from full term placentas exhibited reduced susceptibility to T. 149 gondii infection when compared to primary human foreskin fibroblast (HFF) cells (Supplemental 150 Figure 1A, 1B). These data are consistent with our previous work demonstrating that PHT cells 151 exhibit reduced susceptibility to infection by the three major types of T. gondii in North America

152
and Europe compared to non-placental cells (7). Importantly, human trophoblast cell lines 153 (including BeWo, HTR8, and JEG-3 cells) were unable to recapitulate this restrictive phenotype 154 6 and were permissive to parasite infection (Supplemental Figure 1C). In addition, this phenotype 155 was specific to PHT cultures as primary placental fibroblasts were as permissive to infection as 156 HFF cells (Supplemental Figure 1D).

157
PHT cells isolated from full-term placentas spontaneously fuse to form SYNs during their 158 culture period (~72hrs), with some retaining a mononuclear CYT phenotype. Therefore, to 159 determine whether the lack of PHT cell infection occurred in a cell-type specific manner, we 160 infected PHT cells with YFP-tagged T. gondii (RH strain) and quantified parasite growth 161 specifically in CYTs versus SYNs. These studies revealed dramatic differences in the 162 susceptibility of SYNs and CYTs to T. gondii infection-whereas CYTs were permissive to 163 infection, SYNs were highly resistant ( Figure 1A, left). Furthermore, we observed that parasites 164 within SYNs replicated to a lesser degree, as indicated by a highly significant reduction in total 165 cell area occupied by parasites ( Figure 1A). Since T. gondii replicates within a parasitophorous

169
Importantly, fusion of BeWo cells with forskolin, which induces syncytin-mediated fusion (7), was 170 not sufficient to confer resistance to T. gondii infection (Supplemental Figure 1E), supporting 171 that this phenomenon is specific to primary cells.

172
Transmission electron microscopy (TEM) revealed that whereas parasite growth and PV 173 morphology were normal in mononucleated cells within the preparation (which are likely CYTs 174 but could also be rare contaminating placental fibroblasts), SYN-internalized parasites were found 175 within PVs containing host cell cytoplasmic contents indicative of a loss of vacuole integrity 176 ( Figure 1B). Moreover the parasites within these PVs contained more vacuoles of minimal 177 electron density and poorly defined organelles ( Figure 1B). This phenotype is reminiscent of 178 drug-induced death that we observed previously after treatment with a benzodioxole-containing

186
The primary mechanisms for cell-autonomous immunity to T. gondii are driven by the effector 187 cytokine interferon γ (IFNγ). However, we found that uninfected and T. gondii-infected PHT cells 188 had low levels of IFNγ transcript (Figure 2A) and that culture supernatants were devoid of 189 secreted IFNγ protein (Figure 2A, 2B). Importantly while the expression of GBP1 and GBP2 as 190 well as other innate immunity-related factors (e.g., NOS1,2 and IDO) have comparatively higher 191 transcript levels in PHT cells (Figure 2A), none of these well-characterized IFNγ-driven host 192 effector proteins were uniquely expressed in PHTs (Figure 2A). These findings suggest that the

202
In addition to the intracellular control of parasite replication, it is possible that SYNs are

215
SYNs, which were easily distinguishable using DIC based upon the number, size, and clustering 216 of their nuclei. Using this approach, we found that there were significantly fewer parasites overall 217 (i.e., uninvaded and invaded) that were associated with SYNs compared to CYTs (normalized for 218 cell area; p=0.010; Figure 2D). However, the percentage of invasion events (of all total parasite 219 associations) was nearly identical between SYNs and CYTs, demonstrating that while there is a 220 significant defect in parasite attachment to and/or association with, SYNs, there is no obvious  significantly induced by T. gondii infection. In particular, we found that CCL22, a chemokine 246 known to be expressed constitutively during pregnancy(12, 13) and that has also been found to 247 increase during miscarriage(12) was induced by >400 fold in infected PHT cells based on RNAseq 248 ( Figure 3C), which was confirmed in independent PHT preparations using RT-qPCR (>1000-fold; 249 Figure 3D, left) and at the protein level by ELISA on infected PHT supernatants (from ~25 pg/mL 250 in mock controls to >500 pg/mL in infected cells; Figure 3D, right). Heat-killed T. gondii failed to 251 induce CCL22 release from PHT cells, indicating that production of this chemokine requires live 252 parasites and suggesting that the CCL22 response requires parasite invasion.

255
Host transcriptional responses to infection with T. gondii have been shown in a variety of cell 256 types to be specific for T. gondii and are not associated with infection by one of its apicomplexan 257 relatives, Neospora caninum (14) , (15,16). Unlike T. gondii, N. caninum is not a human pathogen, 10 but causes significant mortality in cattle and dogs and is associated with congenital disease in 259 these animals (17,18). To determine the specificity of the host response to T. gondii infection in 260 PHT cells, we infected cells with T. gondii (RH-YFP) or N. caninum ) and 261 compared the cellular responses to infection using RNAseq. We found that N. caninum failed to 262 significantly induce any of the chemokine/chemokine receptor genes that were induced by

304
Next, we profiled the transcriptional changes induced by T. gondii infection of second 305 trimester chorionic villi using RNAseq to determine whether they responded to parasite infection 306 similarly to PHT cells from late gestation. We found that 172 transcripts were differentially 307 expressed in response to T. gondii infection of villous explants, with 22 of these transcripts also 12 being differentially expressed in response to infection of PHT cells, which included EGR3, EGR4 309 and CCL22 (Figure 5D, 5E). We confirmed that CCL22 was induced at the protein level by ELISA 310 in supernatants from T. gondii-infected second trimester villi ( Figure 5F). These data suggest that 311 CCL22 is specifically induced by the human placenta in response to T. gondii infection at both

329
We found that the first point of SYN-mediated restriction of T. gondii infection occurred at 330 the level of parasite association and/or attachment, which we observed both in SYNs isolated 331 from full-term placentas and from mid-gestation chorionic villi. These findings are consistent with 332 the previous work of others suggesting that attachment might be reduced in first trimester SYNs 333 (6), although this was not directly tested. Our attachment and invasion data from PHT cells provide 334 13 direct evidence that SYNs naturally restrict parasite attachment but are susceptible to invasion 335 once parasite attachment occurs, which appears to be a rare event in SYNs. It remains unclear 336 at what point of the attachment process that is altered when T. gondii associates with SYNs, but 337 a likely stage is during the early phase of gliding motility prior to the second phase of attachment 338 mediated by secretion of microneme and rhoptry organelles (22). One possibility is that when 339 parasites encounter SYNs they glide less efficiently to CYTs or placental fibroblasts, which 340 ultimately results in significantly reduced "full" attachment (mediated by microneme and rhoptry 341 secretion). Differences in membrane biochemistry in SYNs versus CYTs and fibroblasts could 342 underlie these important differences in early parasite association. However, once secondary 343 attachment occurs invasion may then proceed normally, when parasites encounter a second level 344 of resistance.

345
In mid-gestation chorionic villi, the poor association/attachment phenotype was even more 346 profound than that observed in PHT cells, with little to no parasite association with the villi 347 observed. These findings suggest that in addition to biochemical surface differences between

357
For parasites that attach to the SYN layer, our data suggest a second level of resistance 358 to infection that occurs post-invasion. Importantly, in contrast to other cells types, our data show 359 that this resistance is not mediated by IFNγ, which is not basally expressed in PHT cells or 14 induced by T. gondii infection. Furthermore, we did not find any evidence for autophagy-or 361 lysosomal-mediated degradation pathways in the intracellular restriction of parasite replication.

362
To date, all known "cell-autonomous" mechanisms of parasite killing in human cells rely on

382
In addition to resisting T. gondii infection, our data show that PHT cells robustly induce the 383 chemokine CCL22 in response to infection by a Myr-1 dependent effector secretion mechanism.

384
We do not know which cell types within the PHT preparation produce CCL22 after exposure to T.

385
gondii, but given the dependence upon successful invasion on this response a good candidate 15 cell type is the CYT rather than the SYN, although cell-specific analyses are required to address 387 this directly. Importantly, the major inflammatory responses induced in PHT cells by T. gondii 388 infection are not induced (or induced much more poorly) after infection with N. caninum, a near 389 relative of T. gondii that does not successfully infect humans or rodents, suggesting that it may 390 be a host and/or parasite adaptation that may impact disease outcome.

391
The precise role of CCL22 in human pregnancy is unknown, but maternal cells express 392 CCL22 at low levels throughout pregnancy, with increased levels associated with miscarriage 393 (12). Moreover, the induction of chemokines, including CCL22 and CCL17, are associated with 394 preterm birth in humans (30)   423 cells were isolated from healthy, term-pregnancies, and were cultured as described previously(3, 424 9). PHT cells were cultured for ~48h prior to infection to allow for SYN formation. Primary placental 425 fibroblasts were isolated and cultured as described previously (32).

447
Parasite growth curves were generated by luciferase assay (Promega) using luciferase-

467
For RT-qPCR analyses, RNA was isolated as described above and cDNA generated using

482
In some cases, imaged were adjusted for brightness and contrast using Photoshop or Fiji/Image 483 J. Image J was used for image analyses. Transmission electron microscopy was performed as 484 described previously (9).