key: cord-0310125-jqix2yj1
authors: Block, Lindsey N.; Schmidt, Jenna Kropp; McKeon, Megan C.; Bowman, Brittany D.; Wiepz, Gregory J.; Golos, Thaddeus G.
title: Zika virus impacts extracellular vesicle composition and cellular gene expression in macaque early gestation trophoblasts
date: 2021-10-07
journal: bioRxiv
DOI: 10.1101/2021.10.07.463494
sha: 6c581f686f3694c3e66cb2b5a1fd89fe4fd522a9
doc_id: 310125
cord_uid: jqix2yj1

Zika virus (ZIKV) infection at the maternal-placental interface is associated with adverse pregnancy outcomes including fetal demise and pregnancy loss. To determine how infection impacts placental trophoblasts, we utilized rhesus macaque trophoblast stem cells (TSC) that can be differentiated into early gestation syncytiotrophoblasts (ST) and extravillous trophoblasts (EVT). TSCs and STs, but not EVTs, were highly permissive to productive infection with ZIKV strain DAK AR 41524. The impact of ZIKV on the cellular transcriptome showed that infection of TSCs and STs increased expression of immune related genes, including those involved in type I and type III interferon responses. ZIKV exposure altered extracellular vesicle (EV) protein, mRNA, and miRNA cargo, regardless of productive infection. These findings suggest that early gestation macaque TSCs and STs are permissive to ZIKV infection, and that EV analysis may provide a foundation for identifying non-invasive biomarkers of placental infection in a highly translational model.

trophoblast-type responses. Macaque TSCs derived from first trimester vCTB maintain cellular 102 proliferation and can be directed to ST-or EVT-specific differentiation 12 . We previously showed 103

that TSCs differentiated to ST display features characteristic of early first-trimester, a 104 developmentally critical period before the definitive placenta has completely formed. The 105

objectives of this study were to 1) determine which macaque trophoblast cell types were 106 permissive to ZIKV infection, 2) determine the molecular and secretory impact of ZIKV infection, 107

and 3) to determine the utility of placental EVs (PEVs) to serve as readout of trophoblast infection 108 status. The cellular and EV responses associated with infection presented here provide insight 109 into how ZIKV impacts trophoblasts in the first trimester and suggests that PEV cargo may serve 110 as a readout of placental ZIKV infection. 111 112

Results 113 114

Differentiation of TSCs alters cell permissiveness to ZIKV. 115 Since maternal ZIKV infection early in pregnancy is associated with a high rate of pregnancy loss 116 and fetal malformations, an early gestation in vitro macaque TSC model was used to determine 117 which trophoblast cell types are permissive to infection. In vivo mouse 25, 26 and porcine 27 studies 118

and in vitro human studies [28] [29] [30] suggest that strains of the African lineage may be more detrimental 119

to pregnancy than Asian lineage strains, hence, an African lineage strain was utilized to inoculate 120 macaque TSCs. Preliminary time-course experiments to test multiplicities of infection (MOIs) were 121 performed for each cell type to assess viral replication through 72 hrs of culture ( Figure 1A ). The 122 amount of infectious virus detected in conditioned media from TSCs and syncytiotrophoblasts 123 grown in suspension (ST3Ds) plateaued at 60 and 48 hrs, respectively. Inoculation with a MOI of 124 5 resulted in higher levels of infectious virus produced in TSCs compared to an MOI of 1 or 10, 125

whereas viral replication was highest in ST3D's with an MOI of 10. 126 127

An increase in infectious virus was not observed during the EVT time course. Since the quantity 128 of virus did not decrease, to further verify viral replication, wells that did not contain cells but 129 contained EVT culture extracellular matrix components (Matrigel and Col IV) ("no cells") were 130 exposed to virus and the media was evaluated by plaque assay. Compared to the amount of virus 131 detected in the "no cells" wells at 60 hrs (mean 2.2 log10 plaque forming units (PFU)/ml), there 132 was ~40-fold more virus (mean 3.8 log10 PFU/ml) in the EVT samples at 60 hrs ( Figure 1A ). The 133

infectious half-life of ZIKV is ~12 hours 31 , which indicates that the EVTs did indeed release a low 134 level of infectious virus. For subsequent EVT infection experiments, cells were inoculated at an 135 MOI of 5 and cultured for 72 hrs. 136 137

To minimize cytopathic effects and maximize information provided by EVs, the length of culture 138 was chosen based on when the amount of virus being produced began to plateau, expecting that 139

EVs released during peak viral shedding would show the most impact. MOIs for the optimized 140 infection experiments were chosen based on whether there was a substantial increase in viral 141 production (see methods). Of note, an initial (i) aliquot of culture medium was collected after ZIKV 142 inoculation or mock inoculation and then a final (f) aliquot of medium was collected at the culture 143 endpoint to assess viral replication by plaque assay (Supplemental Figure 1A) . The quantity of 144 infectious virus detected in TSC and ST3D media increased significantly by ~100-fold between 145 the initial and final time points within the respective cell type (TSCi = mean 4.20 log10 ± SEM 0. 16, 146 TSCf= mean 6.35 log10 ± 0.17, ST3Di = mean 4.53 log10 ± 0.04, ST3Df= mean 6.32 log10 ± 0.07; 147 Figure 1B ). A minimal change in infectious virus was detected in the EVT samples (EVTi = mean 148 4.28 log10 ± 0.14, EVTf= mean 4.58 log10 ± 0.23; Figure 1B ) as anticipated from the data in Figure  149 1A. 150 151 Cellular ZIKV infection was validated by evaluating the presence of ZIKV E and NS2B proteins 152 within inoculated and mock-inoculated trophoblasts. Both proteins were detected via Western Blot 153 in the ZIKV exposed TSCs and ST3Ds but not in uninfected (control) cells ( Figure 1C ). Neither 154 ZIKV protein was detectable in inoculated EVT samples (Supplemental Figure 3) , supporting the 155 minimal increase of infectious virus from the plaque assays. Positive ZIKV E protein detection via 156 immunofluorescent staining further supported that TSC and ST3D are clearly permissive to ZIKV, 157 while few EVTs stained positive for ZIKV E. ( Figure 1D ). Full Western Blot gel and isotype control 158

immunofluorescence images are shown in Supplemental Figures 2 and 3 , respectively. 159 160

To determine if the MOIs chosen for inoculation induced a cytopathic effect, cellular LDH was 161 assayed in each cell type. Vero cells were used as a positive control as they are highly permissive 162

to ZIKV, and cytopathic effects have been observed following ZIKV infection ( Figure 1E ). If ZIKV 163

induced cell death, less LDH would be detected in infected cell media compared to uninfected 164 control media resulting in a ZIKV:Control ratio less than one. A ratio of less than one was not 165

consistently detected in trophoblast cultures, whereas a decrease in the LDH ratio was observed 166

in Vero cells at 48 hrs that continued through 86 hrs. These data suggest that the MOI and viral 167 strain chosen for this study did not result in trophoblast cell death during the culture period. 168 169 170 171 ) in infected TSCs and  193 ST3Ds, respectively, compared to controls (Figure 2A, 2B) . Figure 2C log10 ± SEM 0) in EVTs exposed to ZIKV (EVT-Zs) ( Figure 2D ). 200 201

Integrated Pathway Analysis (IPA) was performed on the TSC ( Figure 2E ) and ST3D ( Figure 2F ) 202 data sets. IPA identified "role of hypercytokinemia/hyperchemokinemia in the pathogenesis of 203

Influenza" to be highly upregulated and the "coronavirus pathogenesis pathway" to be 204 downregulated by ZIKV infection in TSCs and ST3Ds. The "role of PKR in interferon induction 205 and antiviral response" was another top canonical pathway upregulated by ZIKV infection in TSCs 206

(TSC-Z) ( Figure 2E ). Of the top five most significant diseases or functions identified by IPA in the 207 TSC-Z cell mRNA data, four were related to viral replication and three were predicted to have 208 decreased activation (data not shown). 209 210

The top canonical pathways upregulated by ZIKV infection in ST3D (ST3D-Z) cells included 211 "interferon signaling", "activation of IRF by cytosolic pattern recognition receptors", "role of pattern 212 recognition receptors in recognition of bacteria and viruses", "systemic lupus erythematosus in B 213 cell signaling pathway", and "retinoic acid mediated apoptosis signaling" ( Figure 2F ). The 214 "LXR/RXR activation" was downregulated in ST3D-Z cells. Diseases and function IPA of the 215 ST3D-Z data predicts increased activation of the "antiviral response" and maturation of immune 216 cells along with decreased replication of many viruses and "viral infection" (data not shown). 217 218

The expression of seven of the genes involved in the antiviral response pathway (PARP14,  219  ISG20, DHX58, INFL1, DDX58, IFNB1 , and OASL) and two genes involved in the anti-apoptosis 220 pathway (MT1E and MT1X) 32, 33 were validated with qRT-PCR ( Figure 2G ). The trends in 221 expression observed with qRT-PCR agreed with those identified by poly(A)-seq analysis 222 (Supplemental Table 3 ). TSC-Z or ST3D-Z samples had significantly increased expression of six 223 and seven genes, respectively, with only minor elevations in PAPR14, DHX58, IFNL1, DDX58, 224 IFNB1, and OASL genes in EVT-Z ( Figure 2G) 

ZIKV infection altered the cellular miRNAome.

Placental miRNAs are temporally expressed throughout gestation and have critical roles in 250 trophoblast differentiation and function 34, 35 , thus the miRNAome was profiled to assess alterations 251 in expression relative to infection. miRNA-seq identified expression of 326 miRNAs in TSCs, 335 252 miRNAs in ST3Ds, and 317 miRNAs in EVTs. In the TSCs, one miRNA (mml-miR-663) was 253 significantly decreased and none were significantly increased in infected versus control cells 254

( Figure 3A ). Although ST3Ds were highly infected, no miRNAs were significantly impacted by 255 ZIKV exposure. Despite modest productive infection of EVTs, 52 miRNAs were significantly 256 increased and six significantly decreased in infected versus control EVTs ( Figure 3B ). 257 258

To determine if ZIKV exposure resulted in similar trends of altered miRNA expression across 259 trophoblast cell types, miRNAs significantly impacted in one cell type were compared to the other 260 cell types (data not shown). Only miR-122a-5p trended towards decreased expression in all three 261 cell types exposed to ZIKV (TSC 1.2 log2FC; ST3D 0.4 log2FC, and EVT 5.9 log2FC) with only a 262 change in EVT expression statistically significant. Overall, cell types maintained a distinct miRNA 263 profile regardless of ZIKV exposure ( Figure 3C ). 264 265 266 267 268 To assess the impact of infection on trophoblast function, secretion of hormones, cytokines, 278 chemokines, and growth factors within conditioned cell culture media was assayed. CG and 279

progesterone are two key hormones in the recognition and maintenance of pregnancy. There 280 were no differences in CG or progesterone secretion between infected and control samples for 281 TSC, ST3D, or EVTs (Supplemental Figure 4A ). TSCs secreted minimal CG or progesterone and 282

ST3Ds and EVTs secreted significantly more CG than TSCs, in agreement with our previous 283 report 12 . 284 285

Cytokines and growth factors associated with infection and the inflammatory response were also 286 quantified (Supplemental Figure 4B , 4C). ITAC (CXCL11) was significantly upregulated in infected 287

ST3Ds compared to control (a similar trend was observed in CXCL11 mRNA expression, Figure  288 2B), and was more highly expressed in infected ST3Ds compared to infected TSCs. Otherwise, 289 no other significant differences in cytokine or growth factor secretion were observed between 290 infected and control cells. 291 292

Regardless of infection status, there were significant differences in the secretion of several 293 cytokines and growth factors between trophoblast cell types (Supplemental Figure 4B ). ST3D and 294

EVTs expressed significantly more IL-1RA and bNGF than TSCs. IL-6 was significantly 295 upregulated in control EVTs compared to control ST3Ds. FGF-2, VEGF-A, and VEGF-D 296 expression was significantly increased in ST3D cells compared to TSCs or EVTs. Conversely, 297 some cytokines and growth factors showed no significant differences across cell type or infection 298 status (Supplemental Figure 4C ). 299 300

Characterization and impact of ZIKV infection on EVs 301

Trophoblast-secreted EVs were isolated from ZIKV and control cell conditioned media to assess 302 changes in their physical properties and cargo following ZIKV exposure. EV samples were 303 characterized by Zetaview NTA (Figure 4A ), and a consistent trend towards increased particle 304 size was observed in TSC-Z and ST3D-Z EVs compared to their controls ( Figure 4A ). This trend 305 also was seen with EVT samples, which were not widely infected. No differences in particle 306 concentration were observed in any cell type, although Figure 4A shows that TSCs tended to 307 release fewer EVs. TEM imaging verified the appropriate shape and size of isolated EVs ( Figure  308 4B). 309 310

To determine if ZIKV proteins were packaged into EVs, EV preparations were assessed for E and 311 NS2B proteins by Western Blot. The ZIKV E protein was readily detected in the TSC-Z and ST3D-312 Z EV samples with little detected in the EVT-Z EV samples and none in any controls ( Figure 4C ; 313

Supplemental Figure 3D ). Since EVTs were modestly infected with ZIKV, it is not surprising that 314 the E protein was not detected. The presence of two proteins commonly enriched in EVs, CD9 315 and HSP70, was also determined. On this blot CD9 was only detected in the TSC-C and TSC-Z 316

EVs; however, on another blot, EV samples (those submitted for mass spectrometry) from all 317 three cell types were positive for CD9 (Supplemental Figure 3E ). Calnexin, an ER resident protein 318 not expected to be present within EVs, was confirmed to be absent from all EV samples, and, as 319 expected, calnexin was observed in cell samples along with CD9 ( Figure 1C ). Interestingly, a 320 higher molecular weight cross-reactive band was seen in the TSC-Z sample. Calnexin was not 321 detected in EVs by mass spectroscopy so it is unlikely that the cross-reactive protein is calnexin; 322 also of note is that this band is slightly larger than that observed in the cell lysates (Supplemental 323 Figure 3D ). 324 325

To assess whether the ZIKV E protein associated with EVs, EVs were stained with a fluorescently 326

labeled ZIKV E antibody and analyzed by Zetaview NTA. EV preparations and concentrated ZIKV 327 stock (positive control) were analyzed under scatter and fluorescent modes ( Figure 4D and 4E, 328 respectively). The average number of particles at a given particle size were calculated and are 329 depicted as a single histogram. Zika virions are ~ 50 nm in size 36 and a shift to the left was 330 observed in the "ZIKV stock" histogram compared to the EV preparations, indicative of the 331 abundant presence of smaller particles which are potentially ZIKV virions ( Figure 4D ) (larger 332 particles in this preparation may represent Vero cell-secreted EVs). The fluorescence histogram 333 of EVs isolated from ZIKV cultured trophoblasts, termed "ZIKV" in the figure, showed that the size 334 of particles ranged from ~50-300 nm ( Figure 4E ). The detection of ZIKV E protein in particles 335 larger in size than virions suggests that the E protein was associated with EVs. Of note, minimal 336 background fluorescence was detected in the control EV samples. To further support these data, 337

the number of particles detected in the size range of ZIKV (~50 nm) were quantified (reanalysis 338 of the 24 samples shown in Figure 4A ) and the percentage of particles between 50 and 54 nm 339 was calculated ( Figure 4F ). ZIKV ST3D and EVT EV samples contained fewer "ZIKV-sized" 340 particles than controls, again suggesting that Zika virions were not abundant in the EV samples. 341

Overall, these data indicate that ZIKV E protein is a component of EVs. 342 343 344 345 and control) were pooled and analyzed with DAVID. The mass spectrometry results showed 357 remarkably high enrichment of the "extracellular exosome" cellular component ( Figure 5A ). Other 358 highly enriched cellular components included focal adhesion, as well as membrane, plasma 359 membrane, and cytosol, as expected. Based on the top 1000 genes identified in the eight EV 360 poly(A)-seq samples, the most enriched cellular component also was "extracellular exosome" 361 ( Figure 5B ). Lastly, IPA on all proteins detected in the TSC, ST3D, and EVT EV mass 362 spectroscopy data show that ~50% of proteins isolated were cytoplasmic with ~20-25% 363 associated with the plasma membrane across all cell types ( Figure 5C ). Enzymes, transporter 364 proteins, kinases, transcription regulators, translation regulators, transmembrane receptors, and 365

peptidases were some of the more abundantly detected protein types ( Figure 5D ). Altogether, 366 these data support the authenticity of these trophoblast-produced EV preparations. 367 368 369 370 371 

Mass spectroscopy analysis on EV proteins 378

Proteomic data analysis revealed that many proteins were significantly differentially detected in 379 TSC (191 total: 180 up, 11 down), ST3D (75 total: 63 up, 12 down), and EVT (128 total: 78 up, 380 50 down) EVs released from ZIKV-exposed cells ( Figure 6A -C). Interestingly, unsupervised 381

clustering of protein expression revealed that samples do not group by infection status. 382 383

IPA was performed to predict the disease and biological functions and canonical pathways that 384 the EV proteins may have a role in. Top diseases and biological functions predicted for TSC-Z 385 EV proteins included decreased functions associated with death (necrosis, apoptosis) as well as 386

increased cellular organization and infection ( Figure 6D upper graph). The topmost significant 387 predicted canonical pathways increased in TSC-Z EVs were tRNA charging, EIF2 signaling, 388 purine nucleotide biosynthesis, actin cytoskeleton signaling, Rho GTPase signaling, and RhoA 389 signaling ( Figure 6D lower graph). In comparison, ST3D-Z EV proteins were associated with 390 decreased death, lifespan, and apoptosis with enrichment for infection and viral infection ( Figure  391 6E upper graph), and ZIKV infection was predicted to impact various other pathways but 392 directional shifts were not identified ( Figure 6E lower graph). Decreased degranulation of various 393 cells and decreased adhesion of immune cells were predicted for EVT-Z EV proteins ( Figure 6F  394 upper graph). EVT-Z EVs had increased NAD, Protein Kinase A, actin cytoskeleton, and 14-3-3 395 mediated signaling, whereas ferroptosis and p70S6K signaling pathways were decreased ( Figure  396 6F lower graph). 397 398

To determine if EV protein cargo was shared among the three cell types, proteins significantly 399 differentially detected by mass spectroscopy were compared. This comparison revealed two 400 proteins common among the three cell types, eight proteins between TSC and EVT, two proteins 401 between ST3D and EVT, and 13 proteins between TSC and ST3D ( Figure 6G ). 

The impact of ZIKV on EV poly(A) and miRNA cargo 428

A total of 24 EV samples were submitted for Poly(A)-seq but cDNA libraries could only be 429 prepared for eight samples (3 TSC-C, 1 TSC-Z, 2 ST3D-C, and 2 ST3D-Z). A total of 31,459 430 transcripts were identified of which 2,618 transcripts were detected in all eight EV samples. 431

Conversely, 3,898 and 9,490 transcripts were specific to either ZIKV or control EV samples, 432

respectively. When comparison was restricted to transcripts detected in a majority of samples 433 regardless of cell type origin (3/5 control or 2/3 ZIKV-exposed; Figure 7A) control, respectively) are shown in Figure 7B . 442 443 miRNA-seq on EVs 444

A total of three miRNAs were detected at significantly different levels between ZIKV exposed and 445

control TSC EVs ( Figure 7C ). mml-miR-1249 was significantly increased by ZIKV exposure and 446 mml-miR-100-5p and mml-let-7d were significantly decreased. mml-miR-19b and mml-miR-19a-447 3p were significantly decreased in the ST3D-Z EV samples while none were increased ( Figure  448 7D). Only mml-miR-122a-5p was significantly increased in the EVT-Z EVs and none were 449 significantly decreased ( Figure 7E ). 450 451

To determine if ZIKV exposure resulted in similar trends of altered EV miRNA cargo across 452 different cell types, miRNAs significantly impacted in one cell type were compared to the other 453 cell types ( Figure 7F ). mml-miR-100-5p was decreased in all three cell types exposed to ZIKV 454

and was the only miRNA with similar trends observed among the three different cell types. IFNs) transcript abundance increased after ZIKV exposure. Provocatively, IFNL1 mRNA was 500 detected in EVs from ZIKV exposed cells but not control. The different responses between human 501 ESC-derived trophoblasts and the rhesus TSC-derived trophoblasts in the current study suggests 502 these two cell types represent different developmental stages. It should be noted that no changes 503 in IFN concentrations measured by Luminex assay were detected in response to ZIKV infection 504 and is potentially due to either post-transcriptional mRNA degradation or suboptimal cross-505 reactivity or sensitivity of the Luminex assay with the macaque proteins. 506 507 miRNAs also are important for trophoblast antiviral responses 46 . Interestingly, the predominant 508 changes in miRNA expression were observed in EVTs, the cell type that appeared to control/limit 509 ZIKV infection compared to TSCs or ST3Ds. This raises questions as to whether these cells were 510 able to respond more readily and whether changes in miRNA expression were involved in 511 controlling the infection. mml-miR-122a-5p was decreased in all three trophoblast cell types. 512

Interestingly, EVT-Zs had significantly decreased quantities of cellular mml-miR-122a-5p, while 513 significantly increased quantities of the miRNA were detected in EVs of EVT-Z. mml-miR-122a-514 5p inhibits cellular proliferation, migration, and invasion in Pancreatic ductal adenocarcinoma cells 515 47 . Although the functional significance in trophoblast function remains to be investigated, the 516 consistent trend towards decreased presence of mml-miR-122a-5p in the three trophoblast cell 517 types raises questions as to its role in viral infection. 518 519

Despite changes in the miRNAome and transcriptome, the EVTs displayed modest levels of ZIKV 520 replication, which is unlike findings with other in vitro models 37, 38, 48 . The low level of EVT infection 521 in the current study is an important finding as endovascular EVTs that migrate into and remodel 522 maternal spiral arteries are in direct contact with maternal blood, circulating cells of the maternal 523

immune system, and potentially ZIKV within maternal circulation. published protocols 12 were followed with some minor modifications as described. TSCs were 594 differentiated into either EVTs or ST3D aggregates (Supplemental Figure 1A) . TSCs were grown 595 in 13 ml media, EVTs in 15 ml media, and ST3Ds in 10 ml media (Supp Figure 1B) 621 alanine -valine), and 3790 (missense, alanine -valine) were identified in this ZIKV stock. The 622 frequencies of these substitutions were all below 16%. The stock concentration (4.6 x 10 7 PFU/ml) 623 was determined by plaque assays that were run in triplicate, as previously described 58 . 624 625

Trophoblast ZIKV Inoculations 626

Duration of culture and MOI were first optimized prior to generating experimental infection 627 replicates ( Figure 1A ). Both ZIKV-infected cells and uninfected controls underwent the same 628 processing except uninfected cells were exposed to ZIKV-free Vero cell conditioned media that 629 was collected alongside the ZIKV stock (mock infected). The MOI used was determined based 630 on the quantity of virus calculated from the Vero cell plaque assay described above. 631 632

For infection of the TSCs and EVTs, one flask of cells from each line was lifted and counted just 633 prior to infection. For TSCs and EVTs, media were aspirated, and 1 ml of inoculum was added to 634 the flask. Cultures were incubated at 37°C and rocked gently every 15 mins for 1 hr. The inoculum 635 was removed, and the cells were washed once with PBS followed by addition of fresh medium. 636

The number of ST3D "cells" (at this point they are aggregates) present was based on the number 637 of TSCs added to the flask three days prior (Supplemental Figure 1B) . To infect the ST3Ds that 638

were grown in suspension, aggregates/cells were pelleted by centrifugation at 500 x g for 3 mins, 639

the supernatant was removed, and the cells were resuspended in 300 µl of inoculum. Aggregates 640

were incubated in the ZIKV inoculum for 2 hrs at 37°C and gently every 30 mins. Since ST3Ds 641 are grown in suspension, a larger volume of inoculum was used, and the length of inoculation 642 was extended to account for this increase. Next, 2 ml of PBS was added, the cells were re-pelleted 643 by centrifugation as above, and the supernatant was removed. A volume of 10 ml ST3D medium 644 was added to the cells, and they were then transferred back into a T75 flask. 645 646

Once fresh media were added back for all cell types, a 500 µl aliquot was immediately removed 647 to serve as a baseline of initial viral titer in the culture. The aliquot was spun at 500 x g for 5 mins 648

to remove any cellular debris and a plaque assay was conducted. For EVTs, Growth Factor 649

Reduced Matrigel (0.5%; Corning, Cat #354230) was supplemented to the cultures after this step. 650

EVTs were cultured for a day longer than previously reported 12 (72 hrs total) to extend the 651 duration of ZIKV infection. An MOI of 5, 10, and 5 were used to infect the TSCs, ST3Ds, and 652

EVTs, respectively. Inoculated and mock-inoculated control TSCs were cultured for 60 hrs, while 653

ST3Ds and EVTs were cultured for 48 and 72 hrs, respectively. 654 655

Sample collection: cells and media 656

At the end of the culture period, TSCs were lifted, EVTs were scraped off the flasks, and ST3Ds 657

were pelleted prior to aliquoting cells for RNA, DNA, and protein isolation (Supplemental Figure  658 1A). Conditioned media collected from all flasks were spun at 500 x g for 5 mins to remove dead 659 cells and debris, pooled, and then aliquoted and frozen back at -80°C for EV isolation, hormone 660 analysis, Luminex assay, or plaque assay. 661 662 EV isolation 663

To isolate EVs, 20 ml of medium was placed onto a concentration column (Vivaspin 20; Sartorius, 664

Swedesboro, NJ, USA, Cat # 1208L91) and spun for 60 mins at 3,000 x g at room temperature 665 (RT). For EVT EV samples, due to a high density of Matrigel in the conditioned media, the media 666

were first filtered using a 0.22 um filter (Millipore Sigma, Cat # SLGP033RS) and then spun for 667 60 -90 mins. The concentrated sample was passed through a size exclusion column (Izon,  668 Medford, MA, USA, Cat # SP5, serial #1000788) according to the manufacturer's protocol and 669 the 1.5 ml flow through was collected. The sample was then concentrated using an Amicon 670 concentration column (Millipore Sigma, Cat # UFC801024) for 45-60 min at 3,000 x g at RT. Eight 671

replicates of 20 ml media volumes were processed for TSC and EVT EV isolation and then 672 combined into one sample. Only six replicates of ST3D media were combined. This sample was 673 Detection of ZIKV via immunofluorescence 707

To evaluate cells exposed to ZIKV by immunocytochemistry, additional cells were infected and 708 cultured as described above. For staining, TSCs were cultured for 60 hrs on col IV coated 709 coverslips (5 µg/ml), rinsed in PBS, and fixed with 4% PFA for 10 mins. Previous work showed 710 that EVTs did not grow well on glass coverslips. Thus, after culture, EVTs were lifted, rinsed in 711 PBS, fixed in 4% PFA, and cytospun at 1000 x g for 1 min onto coverslips. ST3Ds were cultured 712 for 48 hrs, then transferred to wells containing col IV-coated coverslips (5 µg/ml) and allowed to 713 attach for 2 hrs. The cells were rinsed with PBS and fixed in 4% PFA for 10 mins. After fixation, 714 all coverslipped cells were rinsed twice with PBS and stored at 4°C in PBS. For immunostaining, 715 the cells were permeabilized for 10 mins with 0.1% Triton-100 (Millipore Sigma, cat #T-9284) and 716 then blocked for 10 mins with Background Punisher (Biocare Medical, Pacheco, CA, USA, Cat 717 #BP974H). For antibody details please see Supplemental Table 1 . Cells were incubated with 718 either primary specific or rabbit IgG isotype control antibody diluted in DaVinci Green Diluent 719 (Biomedical Care, PD900M) for 1 hr at RT, washed 3 times at 5 mins each with 0.1% Tween Tris-720 buffered saline (TBST), and then exposed to secondary rabbit antibody for 45 mins at RT. Finally, 721

the cells were stained with DAPI for 5 mins, washed in Milli-Q water, and then coverslips were 722 adhered to slides using ProLong Diamond Mountant (Fisher Scientific, Cat # P36961). The 723 following day the sides were imaged using a Nikon confocal microscope and Elements software 724 (Nikon, Tokyo, Japan). 725 726

Lactate Dehydrogenase (LDH) Apoptosis Assay 727

To determine if ZIKV induced cell death, an LDH (Cytotox96 non-radioactive cytotoxicity assay,  728 Promega, Madison, WI, Cat #G1780) time course was completed on each cell type. Infection of 729

Vero cells was done as a positive control. For this assay, additional cells were inoculated at the 730 same MOI and length of duration as previously stated and shown in Supplemental Figure 1A . 731 732

TSCs (100,000 cells/well) were seeded into 24-well col IV coated plates (Corning, Cat #3527). 733

The following day they were exposed to ZIKV or mock infected. EVTs were plated (100,000 734 cells/well) in collagen IV coated 24-well plates and differentiated in the wells following the 735 previously stated media changes. ST3Ds were differentiated in non-adherent T25 flasks, and after 736 they were exposed to ZIKV (day 3 of differentiation, as previously stated) they were transferred 737 (three-quarters of a T25 flask/well) to non-adherent 24-well plates (Eppendorf, Hamburg, 738

Germany, Cat #0030 722.019). Vero cells were plated in 24-well plates and then exposed to ZIKV 739

at an MOI of 1, 5, or 10 for 1 hr the following day. Cells were plated such that three ZIKV-exposed 740

and three control wells were quantified at each time point. 741 742

After the ZIKV exposure (as previously determined for each cell type), cells were rinsed once with 743 PBS and then an LDH assay was performed. For the LDH assay, media were removed, the cells 744 rinsed with PBS, and then 100 µl of 1X lysis solution was added. Plaque assay 752

To determine the quantity of virus in the conditioned media, plaque assays on Vero cells were 753 conducted as previously reported 58 . Samples were assessed in duplicate (EVT) or triplicate (TSC 754 and ST3D). 755 756

Hormone quantification 757

Monkey chorionic gonadotropin (mCG) and progesterone assays were performed as previously 758

reported 58 . Samples were run in duplicate and unconditioned media were analyzed to determine 759 background mCG and progesterone. The lower limit of detection for the mCG and progesterone 760 assays are 0.1 ng/ml and 10 pg/ml, respectively. 761 762

Total RNA isolation from cells 763

Total RNA was isolated from cells using an RNeasy kit (Qiagen, Cat #74104) following kit 764 recommendations with modifications. A volume of 700 µl of Qiazol was added to the cell pellet 765

and then frozen at -80°C until RNA extraction. The protocol was followed with the same minor 766

adaptations as with the miRNeasy kit. A 15 min DNAse treatment was performed on the column 767 using RNase-Free DNase (Qiagen, Cat # 79254) prior to washing the column. 768 769

Total RNA isolation from EVs 770

Each EV sample was diluted in 5 volumes Qiazol and incubated for 5 mins at RT prior to freezing 771 at -80°C until RNA extraction. The miRNeasy serum/plasma kit (Qiagen, Cat #217184) was used 772 to isolate total RNA from EVs following the manufacturer's protocol with minor adaptations. The 773

Qiazol with Chloroform was overlaid onto phase maker tubes (ThermoFisher, Cat #A33248) and 774 then spun at 4°C at 16,000 x g for 15 mins. In addition, two RPE washes of 500 µl were applied 775 prior to RNA elution. All samples were eluted in 30 µl RNAse free water and the initial elution was 776 placed back on the column and reeluted to increase RNA concentration. Sample concentration 777

and purity was determined with the NanoDrop One (ThermoFisher, Cat # ND-ONE-W). 778 779

Poly(A)RNAseq 780

RNA integrity was checked with Agilent Technologies 2100 Bioanalyzer. Poly(A) tail-containing 781 mRNAs were purified using oligo-(dT) magnetic beads with two rounds of purification. After 782 purification, poly(A) RNA was fragmented using a divalent cation buffer in elevated temperature. 783

The poly(A) RNA cDNA sequencing library was prepared following Illumina's TruSeq-stranded-784 mRNA sample preparation protocol. Quality control analysis and quantification of the sequencing 785 library were performed using Agilent Technologies 2100 Bioanalyzer High Sensitivity DNA Chip. 786

Pair-end sequencing reads of 150 bp reads were generated on an Illumina NovaSeq 6000 787 sequencing system. 788 789

For cell poly(A)seq data, sequencing reads were filtered to remove adaptors and primer 790 sequences and to remove sequences with a quality score lower than 20 60 . The cleaned 791 sequencing reads were aligned to the reference genome (GCF_003339765.1_Mmul _10) using 792 the HISAT2 package 61 . Multiple alignments with a maximum of two mismatches were allowed for 793 each read sequence (up to 20 by default). Transcript abundance estimation and differential 794 expression analysis of aligned reads of individual samples were assembled using StringTie 62 . 795

Transcriptomes from all samples were then merged to reconstruct a comprehensive 796 transcriptome using a proprietary Perl script designed by LC Sciences. Following transcriptome 797 reconstruction, raw read counts were filtered, normalized, and differential expression determined 798

with DESeq2 63 and edgeR 64, 65 . Genes were considered differentially expressed if they were 799 called by both DESeq2 and edgeR. The glmLRT test was used for edgeR differential expression 800

analysis. EdgeR normalized values were used to produce the heatmaps. 801 802

For EV poly(A)seq data, sequences were filtered and adaptors removed with Cutadapt 60 . Salmon 803 66 was used to obtain transcripts estimates. The mapping rate for the eight sequencable samples 804 was between 3.5-12% for all except one, which had a mapping rate of 86%. Due to the low quality 805

we chose to assess purely based on transcript presence or absence in control/ZIKV EVs. EVs 806 released by control TSCs and ST3Ds (five total samples) were combined and compared to EV 807 data obtained from ZIKV exposed TSCs and ST3Ds (three total samples). Due to limited sample 808 size, differential expression analysis was not performed. 809 810 miRNA-seq 811

The total RNA quality and quantity was assessed using a Bioanalyzer 2100 (Agilent, CA, USA TBST. Blots were incubated in primary antibody (at the concentrations specified in Supplemental 857 Table 1 ) for 1 hr on a rocker in 5% milk at 37°C. The blots were then washed in 1X TBST three 858 times for 5 mins each at RT and then exposed to either mouse or rabbit secondary antibody for 1 859 hr on a rocker in 5% milk at 37°C. The blots were washed again in TBST three times for 5 mins 860 each at RT and then exposed to Immobilon Crescendo HRP (Millipore Sigma, Cat # WBLUR0100) 861

for identification. Strict principles of parsimony were applied for protein grouping. Chromatograms 890

were aligned for feature mapping and area-based quantification using unique and razor peptides. 891

Normalization was performed on total peptide amount and scaling on all averages. 892

To quantify secreted immunomodulatory proteins and growth factors, conditioned media were 894

analyzed with the Protcarta 37-plex (ThermoFisher, Cat #: EPX370-40045-901) as previously 895 described 58 Cell pellets were frozen at -80°C until DNA extraction. DNA was extracted using a Qiagen 903

FlexiGene kit (Qiagen, Cat # 51206) according to the "Isolation of DNA from cultured cells" 904 protocol. Two to four samples were extracted per sample type and DNA was dissolved in 100-905 500 µl of the kit's FG3 buffer. If the sample was viscous after the heat incubation, it was sonicated 906 for 3-6 sec on ice. DNA was quantified using a Nanodrop One spectrophotometer. 907 908

Quantification and statistical analysis 909

Data are represented as the mean ± standard error of the mean (SEM) of four biological 910

replicates. For Zetaview NTA, each sample was processed at least thrice and the CV was 911 calculated. Zetaview NTA data were analyzed in GraphPad Prism using a Kruskal-Wallis test and 912

post-hoc Dunn's correction. All secretion data were normalized to the total DNA calculated for 913 that sample set. Statistical analysis of secretion data was performed using GraphPad Prism 9.0 914 (GraphPad Software) by first log transforming the data prior to conducting a one-way ANOVA test 915 with a post-hoc Bonferroni correction applied (p < 0.05). 916 917

For the mass spectrometry, Poly(A)seq, and miRNAseq data, significance was determined with 918 an adjusted p-value < 0.05 and 1 < log2 fold change < -1. For heatmaps, the rows were organized 919 by hierarchical clustering using agglomerative clustering with Ward's minimum variance method 920 and the Euclidean distance metric. 921 922

Data availability: Poly(A)-seq and miRNAseq data have been deposited at GEO 923 (GSE185113 and awaiting accession ID, respectively) and are publicly available as of the date of 924 publication. The EV poly(A)-seq data was deposited and is available at GEO (GSE185291). Mass 925 Supplemental Table 3 . qRT-PCR, DESeq2, and edgeR log fold change (log2FC) and significance (p-value/adjusted p-value) comparisons 

Characterization of Placental Infection by Zika Virus in Humans: A Review of the Literature

Zika Virus Infection -After the Pandemic

Probable Sexually Transmitted Zika Virus Infection in a Pregnant Woman

Association between Zika virus and fetopathy: a prospective cohort study in French Guiana

Prolonged Detection of Zika Virus RNA in Pregnant Women

Zika virus infection in pregnant rhesus macaques causes placental dysfunction and immunopathology

Prolonged Maternal Zika Viremia as a Marker of Adverse Perinatal Outcomes

Zika Virus Infection with Prolonged Maternal Viremia and Fetal Brain Abnormalities

Placenta: the forgotten organ

Development of the human placenta

Derivation of Human Trophoblast Stem Cells

Placenta-derived macaque trophoblast stem cells: differentiation to syncytiotrophoblasts and extravillous trophoblasts reveals phenotypic reprogramming

Zika Virus Infection in Pregnant Women in Rio de Janeiro

The association between Zika virus infection and microcephaly in Brazil 2015-2017: An observational analysis of over 4 million births

Pregnancy Outcomes after ZIKV Infection in French Territories in the Americas

Factors associated with the development of Congenital Zika Syndrome: a case-control study

Teratogen update: Zika virus and pregnancy

Placental exosomes: A proxy to understand pregnancy complications

Application of exosomes as liquid biopsy in clinical diagnosis

Hypoxia-induced changes in the bioactivity of cytotrophoblast-derived exosomes

Exosomes and Their Role in the Life Cycle and Pathogenesis of RNA Viruses

Arthropod EVs mediate dengue virus transmission through interaction with a tetraspanin domain containing glycoprotein Tsp29Fb

Animal Models to Study Placental Development and Function throughout Normal and Dysfunctional Human Pregnancy

Miscarriage and stillbirth following maternal Zika virus infection in nonhuman primates

Zika viruses of African and Asian lineages cause fetal harm in a mouse model of vertical transmission

Recent African strains of Zika virus display higher transmissibility and fetal pathogenicity than Asian strains

The African strain of Zika virus causes more severe in utero infection than Asian strain in a porcine fetal transmission model

African and Asian strains of Zika virus differ in their ability to infect and lyse primitive human placental trophoblast

Vulnerability of primitive human placental trophoblast to Zika virus

Zika Virus Replicates in Proliferating Cells in Explants From First-Trimester Human Placentas, Potential Sites for Dissemination of Infection

Zika Virus Is Not Uniquely Stable at Physiological Temperatures Compared to Other Flaviviruses

Increased metallothionein gene expression, zinc, and zincdependent resistance to apoptosis in circulating monocytes during HIV viremia

Metallothioneins in Normal and Cancer Cells

Placental miRNAs in feto-maternal communication mediated by extracellular vesicles

Pregnancy-associated miRNA-clusters

Zika Virus Structure, Maturation, and Receptors

ZIKA virus reveals broad tissue and cell tropism during the first trimester of pregnancy

Zika Virus Targets Different Primary Human Placental Cells, Suggesting Two Routes for Vertical Transmission

The interferon stimulated gene 20 protein (ISG20) is an innate defense antiviral factor that discriminates self versus non-self translation

Viral subversion of the host protein synthesis machinery

Zika Virus Infects Early-and Midgestation Human Maternal Decidual Tissues, Inducing Distinct Innate Tissue Responses in the Maternal-Fetal Interface

Type I interferons instigate fetal demise after Zika virus infection

Type III Interferons Produced by Human Placental Trophoblasts Confer Protection against Zika Virus Infection

An evolutionary NS1 mutation enhances Zika virus evasion of host interferon induction

Zika virus antagonizes interferon response in patients and disrupts RIG-I-MAVS interaction through its CARD-TM domains

The role of trophoblastic microRNAs in placental viral infection

MicroRNA-122-5p inhibits cell proliferation, migration and invasion by targeting CCNG1 in pancreatic ductal adenocarcinoma

The Mechanism of the Zika Virus Crossing the Placental Barrier and the Blood-Brain Barrier

Increased Resistance of Breast, Prostate, and Embryonic Carcinoma Cells against Herpes Simplex Virus in Three-Dimensional Cultures

Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells

Role of Host Cell Secretory Machinery in Zika Virus Life Cycle

The biology of extracellular vesicles: The known unknowns

The immunoproteasome and viral infection: a complex regulator of inflammation

HCV and flaviviruses hijack cellular mechanisms for nuclear STAT2 degradation: Up-regulation of PDLIM2 suppresses the innate immune response

Human trophoblast stem cells: Real or not real?

Human placenta and trophoblast development: key molecular mechanisms and model systems

Growth and adaptation of Zika virus in mammalian and mosquito cells

Embryotoxic impact of Zika virus in a rhesus macaque in vitro implantation modeldagger

Isolation and characterization of exosomes from cell culture supernatants and biological fluids

Cutadapt removes adapter sequences from high-throughput sequencing reads

HISAT: a fast spliced aligner with low memory requirements

StringTie enables improved reconstruction of a transcriptome from RNA-seq reads

Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2

edgeR: a Bioconductor package for differential expression analysis of digital gene expression data

Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation

Salmon provides fast and bias-aware quantification of transcript expression

Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method

Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines

Omics analyses and biochemical study of Phlebiopsis gigantea elucidate its degradation strategy of wood extractives

 TGA GCT GCG TGT GG  GTA CAT GGC TGG GGT GTT GA  130  INFL1 ATC GTG GTG CTG GTG ACT TT  TTG AGT GAC TCT TCC AAG GCA  170  ISG20 TGC TGT GCT GTA CGA CAA GT  CCA AGC AGG CTG TTC TGG AT  339  OASL CAT CGT GCC TGC CTA CAG AG  GAC CTG GCT TTC ACA TAC TGC T  219  DDX58 CTG GTT CCG TGG CTT TTT GG  CCC CTT AGT GGA GCA AAT CTG T  238  IFNB1 CTC CTG TTG TGC TTC TCC ACT  AAT GCA GCG TCC TCC TTC