key: cord-284646-fhruiw23
authors: Jaeger, Anna S.; Weiler, Andrea M.; Moriarty, Ryan V.; Rybarczyk, Sierra; O'Connor, Shelby L.; O'Connor, David H.; Seelig, Davis M.; Fritsch, Michael K.; Friedrich, Thomas C.; Aliota, Matthew T.
title: Spondweni virus causes fetal harm in Ifnar1(-/-) mice and is transmitted by Aedes aegypti mosquitoes
date: 2020-05-24
journal: Virology
DOI: 10.1016/j.virol.2020.05.005
sha: 
doc_id: 284646
cord_uid: fhruiw23

Spondweni virus (SPONV) is the most closely related known flavivirus to Zika virus (ZIKV). Its pathogenic potential and vector specificity have not been well defined. SPONV has been found predominantly in Africa, but was recently detected in a pool of Culex quinquefasciatus mosquitoes in Haiti. Here we show that SPONV can cause significant fetal harm, including demise, comparable to ZIKV, in a mouse model of vertical transmission. Following maternal inoculation, we detected infectious SPONV in placentas and fetuses, along with significant fetal and placental histopathology, together suggesting vertical transmission. To test vector competence, we exposed Aedes aegypti and Culex quinquefasciatus mosquitoes to SPONV-infected bloodmeals. Aedes aegypti could efficiently transmit SPONV, whereas Culex quinquefasciatus could not. Our results suggest that SPONV has the same features that made ZIKV a public health risk.

Zika virus (ZIKV) was originally isolated over seventy years ago, and was thought to cause a 14 mild, self-limiting, febrile illness (DICK et al., 1952; SIMPSON, 1964) . Not until the outbreak 15 in the Americas in 2015 and 2016 was ZIKV identified as a cause of significant adverse 16 pregnancy outcomes (Johansson et al., 2016; Melo et al., 2016) . Before the definition of congenital Zika syndrome (CZS) in 2016, gestational arbovirus infection was not associated with birth defects. Spondweni virus (SPONV) is the closest known relative to ZIKV, but 19

whether SPONV is an emerging threat to pregnant women and their babies is unknown. It 20 was previously thought that SPONV was geographically confined to Africa and caused only 21 mild disease in rare human infections, reminiscent of the consensus around ZIKV in the 22 decades following its discovery, but recent data suggest that it may be spreading beyond 23

Africa (White et al., 2018). SPONV may therefore be poised to harm pregnancies in new, 24 immunologically naive populations. To do this, SPONV would need to fulfill two major 25 criteria: it would need to be vertically transmitted and cause fetal harm, and be transmitted 26 between humans by the urban mosquito vector Aedes aegypti, which is associated with 27 large-scale outbreaks of related arboviruses. 28

The first identification of SPONV was thought to have occurred in 1955 in South Africa 29 (Theiler and Downs, 1973; Wolfe et al., 1982) . However, it was later recognized that SPONV 30 was in fact isolated three years earlier in Nigeria, but was misidentified at the time as a 31 strain of ZIKV because of serological cross-reactivity (HADDOW et al., 1964; SIMPSON, 1964; 32 DRAPER, 1965) . Serological cross-reactivity with ZIKV and other flaviviruses likely still 33 confounds accurate diagnostics today. As a result, only six well-documented clinical cases of 34 DRAPER, 1965) . It is likely that many infections have gone unrecognized-serosurveys have 36 detected evidence of SPONV infection in 10 countries throughout Sub-Saharan Africa 37 (Kokernot et al., 1965a; Kokernot et al., 1965b; Brottes et al., 1966; Ardoin et al., 1976; Wolfe 38 SPONV caused fetal harm, similar to what is observed from ZIKV infection in this model. 69

Vector competence experiments showed that Ae. aegypti could transmit SPONV when 70 exposed to bloodmeal titers that approximate physiological titers, while Cx. quinquefasciatus nonpregnant, mixed sex 6-to 11-week-old mice lacking type I interferon signaling (Ifnar1 -/-) Ar94 (this is the only strain used in these studies, so it will be referred to hereafter as 86 SPONV); or 10 2 PFU of the highly pathogenic African-lineage ZIKV strain DAK AR 41524 87 (ZIKV-DAK) (Jaeger et al., 2019) . Since contemporary SPONV isolates from Haiti do not exist, 88

we used the only available low-passage isolate, SPONV strain SA Ar94. This strain is 98.8% 89 nucleotide identical with the SPONV genome recovered from mosquitoes in Haiti 90 (Genbank:MG182017). Serum was collected at 2, 4, and 6 days post-inoculation (dpi) to 91 confirm infection and determine the replication kinetics of SPONV in nonpregnant Ifnar1 -/-92 mice. We also collected and tested serum at 7, 14, and 21 days from mice surviving SPONV 93 inoculation, because sustained vRNA loads were observed with the Ifnar1-blocking mAb 94 model (Salazar et al., 2019) . SPONV viral titer in the serum peaked at 4 dpi (Fig. 1a) , and in 95 surviving animals there was no detectable viremia at 7, 14, or 21 dpi. Higher serum titers 96 were observed in animals inoculated with the lowest dose of SPONV (10 2 PFU). We 97 postulate that this could be the result of higher inoculating doses causing a rapid initial rise 98 in viremia, which in turn induces a more robust immune response, leading to more rapid 99 clearance of virus from the serum, but confirmation will require further studies. ZIKV-DAK 100 viremia also peaked at 4 dpi and reached significantly higher titers at 4 dpi than either 101 PFU of SPONV or 10 2 PFU ZIKV-DAK. Based on our preliminary experiments with SPONV in 120 nonpregnant animals, and the results from our past studies (Jaeger et al., 2019), we chose 121 this dose to minimize the potential confounding impacts of maternal illness on fetal 122 outcomes. We collected serum samples from dams at 2 and 4 dpi to confirm maternal 123 infection. All dams were productively infected, with detectable viremia for all groups by 4 124 dpi (Fig. 2a) . ZIKV-DAK replicated to significantly higher titers at 4 dpi as compared to 125 SPONV (Student's t-test p-value = 0.0008, t = 5.641, df = 7). Dams were monitored daily pregnancies and with uninfected counterparts. In general, fetuses appeared either grossly 135

At the time of necropsy, we observed high rates of resorption from both ZIKV-DAK-and 137 SPONV-infected pregnancies. Resorption rates from ZIKV-DAK-and SPONV-infected 138 pregnancies were not significantly different (ZIKV-DAK: 76.92% vs. SPONV: 68.29%, Fisher's 139 exact test, p = 0.457). Resorption rates for both SPONV and ZIKV-DAK were significantly 140 higher than PBS-inoculated controls (p <0.0001). Despite significantly higher maternal 141 viremia observed at 4 dpi with ZIKV-DAK-infected dams, the fact that resorption rates did 142 not significantly differ between the two groups indicates that both ZIKV-DAK and SPONV 143 have a propensity to harm the developing fetus that is independent of the amount of 144 replication in maternal blood. Surprisingly, and in contrast to the results described by 

To further characterize the range of pathogenic outcomes of congenital SPONV infection 150 and to assess differences between models, we repeated experiments by treating dams with inoculation with ZIKV-DAK or SPONV (Sheehan et al., 2006) . This model has been used 153 previously for assessing both ZIKV and SPONV pathogenesis during pregnancy, but does confirm infection, and all dams were productively infected with SPONV or ZIKV-DAK 159 following treatment with either dose of mAb (Fig. 2d) . Maternal viremia did not significantly 160 differ between treatment groups (SPONV/1mg vs. SPONV/2mg: p=0.996; ZIKV/1mg vs.

ZIKV/2mg: p=0.35; one-way ANOVA with Tukey's correction for multiple comparisons). ZIKV-

DAK titers, however, were significantly higher than SPONV titers (SPONV/1mg vs. ZIKV/1mg: 163 p=0.04; SPONV/2mg vs. ZIKV/2mg: p=0.006). Next, adhering to our previously established 164 experimental timeline, dams were necropsied on E14.5 to assess and compare fetal 165 outcomes. At the time of necropsy, we observed no significant resorption from either ZIKV 166 or SPONV infected pregnancies, after either dose of mAb (Fig. 2e) , consistent with the 167 results described by Salazar et al. observed after E6.5 virus challenge and E13.5 or E18.5 dam sacrifice (Salazar et al., 2019) . Resorption rates from ZIKV-DAK-and SPONV-infected 169 pregnancies were not significantly different (Fisher's exact test, p>0.06 for all comparisons).

It is possible that the differences in outcomes in these two models may be due to the closely related to both ZIKV and SPONV and it is not known to cause adverse pregnancy 179 outcomes in humans. To examine whether maternal DENV-2 infection is sufficient to induce 180 fetal resorption, we s.c. inoculated pregnant dams on E7.5 with 7.5 X 10 4 PFU of DENV-2. 181

Prior to studies in pregnant animals we confirmed that this route and dose would result in 182 productive infection in nonpregnant animals (Fig. 3a) . All dams were productively infected 183 with DENV-2 with detectable vRNA loads at 2 and 4 dpi (Fig. 3a) . Importantly, fetuses 184 continued to develop as examined on E14.5, and rates of resorption were not significantly 185 exact test, p = 0.665) (Fig. 3b) . These observations confirm that fetal harm was specifically 187 associated with ZIKV-DAK and SPONV infection, but because DENV-2 infected mice do not 188

show clinical signs, we cannot exclude the possibility that the more severe fetal outcomes 189 To begin to understand the potential for SPONV to be vertically transmitted, a subset of 194 placentas and fetuses were collected for plaque assay at time of necropsy from all virus 195 treatment groups. From the Ifnar1 +/tissues, infectious virus was detected in 100% of ZIKV-

DAK placentas and fetuses screened (Fig. 2c) . Virus was detected in all but one SPONV 197 placenta and 35% of fetuses (Fig. 2c) . Viral titers were significantly higher in SPONV 198 placentas than their corresponding fetuses (one-way ANOVA with Tukey's multiple 199 comparisons; p < 0.0001), as were ZIKV-DAK placenta viral titers as compared to ZIKV-DAK 200 fetuses (p = 0.006). In addition, ZIKV placenta and fetal viral titers were significantly higher 201 than SPONV titers (p < 0.0001). 202 placental tissues from dams treated with anti-Ifnar1 mAb (Fig. 2f) . Antibody dose did not 204 affect the viral titer present in fetuses or placentas after either SPONV-or ZIKV-DAK-205 inoculation (p > 0.9 for all comparisons; one-way ANOVA with Tukey's multiple 206 comparisons). In general, fetal and placenta tissue titers were significantly higher in ZIKV-207 DAK challenge groups as compared to SPONV challenge groups, with a more significant 208 difference in placenta tissue titers than fetal tissue titers (Fig. 2f) . Of note, infectious SPONV 209 was detected in fetuses from both mAb treatment groups, which is in contrast to the placental tissues from DENV-2 infected pregnancies were also screened for infectious virus 215 via plaque assay. Infectious virus was not detected in any of the screened fetal and placental 216 tissues, further suggesting the specificity of fetal harm to ZIKV and SPONV (Fig. 3c) . 217

To better understand the impact of in utero SPONV exposure, tissues from the developing 218 Ifnar1 +/placenta and fetus were evaluated microscopically. In PBS-and DENV-inoculated fetal blood spaces (Fig. 4) . In contrast, ZIKV-DAK-and SPONV-inoculated dams displayed 221 varying degrees of placental pathology with severe effects predominantly observed in the 222 the labyrinth zone, including vascular injury involving maternal and/or fetal vascular spaces, 223 infarction (obstructed blood flow), necrosis, apoptosis, and hemorrhage (Fig. 4) . Overall, the 224 severity of the vascular injury in the labyrinth zone was similar between ZIKV-DAK and 225

In the fetuses, there was no significant microscopic pathology from PBS-and DENV-227 inoculated dams. In contrast, fetuses from ZIKV-DAK-and SPONV-inoculated dams 228 demonstrated varying degrees of pathology. In fetuses from the SPONV-inoculated dams, 229 fetal injury was evident as mild pulmonary inflammation and mild to moderate segmental 230 necrosis of the brain and spinal cord (Fig. 5) . These data provide indirect evidence that 231 vertical transmission did occur. Pathologic findings were more widespread and severe in 232 fetuses from ZIKV-DAK-inoculated dams and included severe necrosis and inflammation of 233 the lung, liver, kidney, brain, and spinal cord. 234

because SPONV RNA was detected in a pool of Cx. quinquefasciatus in Haiti, we compared 237 the relative abilities of Ae. aegypti and Cx. quinquefasciatus from Florida to transmit SPONV 238 in the laboratory. SPONV titers in naturally infected hosts-to which feeding mosquitoes 239 might be exposed in nature-are undefined. Therefore, we conducted our experiments with 240 blood meal titers ranging from ~10 6 -10 8 PFU/ml. We considered these doses to be 241 physiologically relevant based on studies with DENV (50% mosquito infectious doses = 242 10 5.68 -10 7.21 viral cDNA copies/ml) (Duong et al., 2015) and ZIKV (50% mosquito infectious 243 doses = 10 6.1 -10 7.5 PFU/ml) (Ciota et al., 2017) . To assess vector competence, mosquitoes 244 were exposed to viremic bloodmeals via water-jacketed membrane feeder maintained at Table 1 ). Ae. aegypti that had been exposed to We speculate that the difference in outcomes between these two models could be due to displayed the most severe histologic phenotype that corresponded with higher placenta and 329 fetus titers in both pregnancy models (Fig. 2) . SPONV histopathology was more Following inoculation with SPONV, ZIKV, or PBS, mice were sacrificed at E14.5. Tissues were 506 carefully dissected using sterile instruments that were changed between each mouse to 507 minimize possible cross contamination. For all mice, each organ/neonate was evaluated 508 grossly in situ, removed with sterile instruments, placed in a sterile culture dish, and further 509 processed to assess viral burden and tissue distribution or banked for future assays. Briefly, 510 uterus was first removed, and then dissected to remove each individual conceptus (i.e, fetus 511 and placenta when possible). Fetuses and placentas were either collected in PBS 512 supplemented with 20% FBS and penicillin/streptomycin (for plaque assays) or fixed in 4% 513 PFA or 10% Neutral Buffered Formalin for imaging. We characterized an embryo as in the 514 resorption process if it met the following criteria: significant growth retardation compared to 515 litter mates and controls accompanied by clearly evident developmental delay, i.e., 516 morphology was ill defined; or visualization of a macroscopic plaque in the uterus (Flores et 517 al., 2014) . 518

Tissues were fixed in 4% paraformaldehyde for 24 hours and transferred into cold, sterile 520 DPBS until alcohol processed and embedded in paraffin. Paraffin sections (5 μm) were 521 stained with hematoxylin and eosin (H&E). Pathologists were blinded to gross pathological findings when tissue sections were evaluated microscopically. The degree of pathology at the maternal-fetal interface was rated on a scale of 0-4: 0 -no lesions (normal); 1 -mild 524 changes (1-2 focal lesions or 10-15% of zone involved); 2 -mild to moderate changes (3-4 525 focal lesions or 10-15% of zone involved); 3 -moderate to severe changes (4-6 focal lesions 526 or 15-25% of zone involved); 4 -severe (>6 focal lesions or >25% of zone involved). The 527

final score was dependent upon the greater of two parameters (# of lesions or % zone 528 involved). This was an identical scoring system to what we reported previously (Jaeger et al., 529 2019). The final scores were determined as a consensus score of two independent 530

pathologists. For each zone in the placenta (myometrium, decidua, junctional zone, labyrinth, 531 and chorionic plate/membranes) a 'General' overall score was determined, a score for the 532 amount of 'Inflammation', and a score for direct 'Vascular Injury'. The 'General' score was 533 based on an interpretation of the overall histopathologic findings in each placenta, which 534 included features of necrosis, infarction, apoptosis, hemorrhage, thrombosis, mineralization, 535 vascular injury, and inflammation. The 'Inflammation' score quantified the amount of 536 inflammation in that layer. The 'Vascular Injury' score assessed vascular wall injury (fibrinoid 537 necrosis, endothelial swelling), dilatation of the vessels or spaces, necrosis, loss of vascular 538 lumen diameter, and intraluminal thrombi. The myometrial layer representing the uterine 539 therefore meaningful comparisons between strains could not be assessed. The decidual layer 541 (maternal in origin), the junctional zone composed of fetal giant cells and 542 spongiotrophoblast, and the labyrinth layer (the critical layer for gas and nutrient exchange 543 between the fetal and maternal vascular systems) were scored. Since the percentage of 544 injured/pathologic labyrinth zone is a predictor of poor fetal outcome, we also 545 independently scored the labyrinth zone based only on the percentage of fetal and maternal 546 vascular injury/loss using the following scoring system: 0-5%-0 (background); 5-15%-1 547 (mild); 15-30%-2 (moderate); 30-50%-3 (moderate to severe); and >50%-4 (severe). 548

Photomicrographs were obtained using a bright light microscope Olympus BX43 and 549

Olympus BX46 (Olympus Inc., Center Valley, PA) with attached Olympus DP72 digital camera 550 (Olympus Inc.) and Spot Flex 152 64 Mp camera (Spot Imaging), and captured using 551 commercially available image-analysis software (cellSens DimensionR, Olympus Inc. and spot 552 software 5.2). 553

All mosquitoes used in this study were maintained at the University of Minnesota, Twin 555

Cities as described (Christensen and Sutherland, 1984) Vector competence studies 564 Mosquitoes were exposed to SPONV-or ZIKV-infected bloodmeals via water-jacketed 565 membrane feeder maintained at 36.5 °C (Rutledge et al., 1964) . Bloodmeals consisted of 566 defibrinated sheep blood (HemoStat Laboratories, Inc.) and fresh virus supernatant, yielding 567 infectious bloodmeal titers ranging from ~10 6 -10 8 PFU/ml. Bloodmeal titer was determined 568 after feeding. Infection, dissemination, and transmission rates were determined for individual 569 mosquitoes and sample sizes were chosen using long established procedures (Aliota et al., Fisher's exact test was used to determine differences in rates of normal vs. abnormal 582 concepti. 583

Virus stock sequence data have been deposited in the Sequence Read Archive (SRA) with 585 accession codes pending. The authors declare that all other data supporting the findings of 586 this study are available within the article. 587 compared to PBS controls (Fisher's exact test). ***p < 0.0005; **p < 0.005. 827 --------

Culex 589 pipiens and Aedes triseriatus Mosquito Susceptibility to Zika Virus

Wolbachia Reduces Transmission of Zika virus by Aedes aegypti

The wMel Strain of Wolbachia Reduces Transmission of Chikungunya 595 Virus in Aedes aegypti

Louis encephalitis virus in mice

Epidemiologic study of arboviruses in 599 the Arba-Minch district of Ethiopia

Isolations of arboviruses in the Lagos area of

Nigeria, and a survey of antibodies to them in man and animals

Human to mosquito transmission of 606 dengue viruses

Antiviral immunity backfires: Pathogenic effects of 608 type I interferon signaling in fetal development

Exsheathment and Midgut Penetration 610 in Aedes aegypti

Effects of Zika Virus Strain and Aedes Mosquito Species on 613

Zika virus. I. Isolations and 615 serological specificity

Infection via 622 mosquito bite alters Zika virus tissue tropism and replication kinetics in rhesus macaques

Asymptomatic 626 humans transmit dengue virus to mosquitoes

Evidence of 630 vertical transmission and co-circulation of chikungunya and dengue viruses in field 631 populations of Aedes aegypti (L.) from Guerrero

Early detection and 634 staging of spontaneous embryo resorption by ultrasound biomicroscopy in murine 635 pregnancy

A Diversified Portfolio

Virol 3, vi-viii

Genomic epidemiology reveals 650 multiple introductions of Zika virus into the United States

Dengue, Urbanization and Globalization: The Unholy Trinity of the 652 21(st) Century

Genetic Characterization of Spondweni and Zika Viruses and 655 Susceptibility of Geographically Distinct Strains of Aedes aegypti

Culex quinquefasciatus (Diptera: Culicidae) to Spondweni Virus

Distinguishing between Zika and Spondweni 659 viruses

A Gorilla Adenovirus-Based Vaccine against

Zika Virus Induces Durable Immunity and Confers Protection in Pregnancy

Zika virus infection in pregnant rhesus macaques causes placental 673 dysfunction and immunopathology

Protective Efficacy of Nucleic Acid Vaccines Against Transmission of Zika Virus During 682 Pregnancy in Mice

Gestational Stage and IFN-λ Signaling

Zika and the Risk of Microcephaly

Survey 689 for antibodies against arthropod-borne viruses in the sera of indigenous residents of

Studies on 692 arthropod-borne viruses of Tongaland. VIII. Spondweni virus, an agent previously unknown, 693 isolated from Taeniorhynchus (Mansonioides) uniformis

Survey for 695 antibodies against arthropod-borne viruses in the sera of indigenous residents of the Caprivi 696 Strip and Bechuanaland Protectorate

Quantitative genetics of Aedes aegypti vector competence for 698 dengue viruses: towards a new paradigm

Discovery and full genome characterization of two highly 702 divergent simian immunodeficiency viruses infecting black-and-white colobus monkeys 703 (Colobus guereza

A Mouse Model of Zika Virus Pathogenesis

Shared and Distinct 708 Functions of Type I and Type III Interferons

Zika virus: a report on three cases of human infection during an 710 epidemic of jaundice in Nigeria

Pathogenesis and sexual 712 transmission of Spondweni and Zika viruses

Further isolations of the arboviruses 714 from mosquitoes collected in Tongaland, South Africa, 1960-1968

Isolation 717 of Spondweni virus from four species of culicine mosquitoes and a report of two laboratory 718 infections with the virus

The interaction of glutamic and aspartic acids with 720 excitatory amino acid receptors in the mammalian central nervous system

Congenital Zika Virus Infection: Beyond Neonatal Microcephaly

Increased yellow fever virus infection and dissemination rates in Aedes 735 aegypti mosquitoes orally exposed to freshly grown virus

Virus Infection during Pregnancy in Mice Causes Placental Damage and Fetal Demise

Zika virus: following the path of 742 dengue and chikungunya

Host and viral features of human 747 dengue cases shape the population of infected and infectious Aedes aegypti mosquitoes

Zika virus-related neurotropic flaviviruses infect human placental explants and cause 751 fetal demise in mice

No evidence of Zika, dengue, or chikungunya virus 754 infection in field-caught mosquitoes from the Recife Metropolitan Region

Zika Virus 757 and Birth Defects--Reviewing the Evidence for Causality

Vaccine Mediated Protection Against Zika Virus-Induced Congenital Disease

Studies on the feeding response of 765 mosquitoes to nutritive solutions in a new membrane feeder

Dengue and Zika Virus Cross-Reactive

Human Monoclonal Antibodies Protect against Spondweni Virus Infection and Pathogenesis 770 in Mice

Neutralizing human antibodies 774 prevent Zika virus replication and fetal disease in mice

Blocking monoclonal antibodies specific for mouse IFN-alpha/beta receptor subunit 1 778 (IFNAR-1) from mice immunized by in vivo hydrodynamic transfection

ZIKA VIRUS INFECTION IN MAN

The arthropod-borne viruses of vertebrates. An 783 account of theRockefeller Foundation virus program

Zika virus and the nonmicrocephalic fetus: why we should still worry

Emerg 790

Spondweni virus infection in a 792 foreign resident of Upper Volta

The incidence of arthropod-794 borne viruses in a population of culicine mosquitoes in Tongaland

Type I interferons instigate fetal demise after Zika virus infection. Sci 799 Immunol 3, 800 analyzed data and drafted the manuscript

O. developed and performed the deep sequencing pipeline

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work

The authors acknowledge the University of Minnesota, Twin Cities BSL3 Program for facilities 802 and Neal Heuss for support. We thank Natalie Benett for her contribution in mosquito

The authors declare no competing financial interests. 816 collected from mice during three independent replicates at 2, 4, and/or 6 days post 820 inoculation and titered via plaque assay. Assay limit of detection was 100 PFU. Viremia 821 peaked at 4dpi for all virus groups, with ZIKV-DAK replicating to significantly higher titers at 822 4dpi than SPONV (one-way ANOVA). ****p < 0.0001; ***p < 0.0005; **p < 0.006 (b) Survival 823 curves of six-to eleven-week old Ifnar1 -/mice s.c. inoculated with 10 3 PFU of SPONV, 10 2 824 PFU of SPONV, 10 2 PFU ZIKV-DAK, or a PBS control. SPONV 10 3 : n=14; SPONV 10 2 : n=13, 825