Zika Virus Causes Acute Infection and Inflammation in the Ovary of Mice Without Apparent Defects in Fertility

Abstract

Background

Zika virus (ZIKV) has become a global concern because infection of pregnant mothers was linked to congenital birth defects. Zika virus is unique from other flaviviruses, because it is transmitted vertically and sexually in addition to by mosquito vectors. Prior studies in mice, nonhuman primates, and humans have shown that ZIKV targets the testis in males, resulting in persistent infection and oligospermia. However, its effects on the corresponding female gonads have not been evaluated.

Methods

In this study, we assessed the effects of ZIKV on the ovary in nonpregnant mice.

Results

During the acute phase, ZIKV productively infected the ovary causing accumulation of CD4⁺ and virus-specific CD8⁺ T cells. T cells protected against ZIKV infection in the ovary, as higher viral burden was measured in CD8^−/− and TCRβδ^−/− mice. Increased cell death and tissue inflammation in the ovary was observed during the acute phase of infection, but this normalized over time.

Conclusions

In contrast to that observed with males, minimal persistence and no long-term consequences of ZIKV infection on ovarian follicular reserve or fertility were demonstrated in this model. Thus, although ZIKV replicates in cells of the ovary and causes acute oophoritis, there is rapid resolution and no long-term effects on fertility, at least in mice.

Although ZIKV can be transmitted sexually and target the testis in males for chronic infection, its effects on the female gonads have not been studied. We evaluated the short- and long-term effects of ZIKV infection in the ovaries of mice.

Zika virus (ZIKV) is an arthropod-transmitted flavivirus that was first isolated in 1947 in Uganda [1]. Historically, ZIKV infections occurred sporadically in Africa and Asia without significant morbidity beyond a mild febrile illness. In 2007, a ZIKV outbreak occurred on Yap Island in the Pacific Ocean, with more than half of the inhabitants infected, many of which experienced rash, fever, and joint or muscle pain [2]. This was followed by a larger ZIKV outbreak in 2013 in French Polynesia with approximately 20 000 reported cases [3]. Zika virus then was introduced into Brazil in late 2013 or early 2014 [4], with epidemic spread occurring throughout the Americas in 2015 [5]. In addition to mosquito transmission, ZIKV can be spread sexually between men and women as well as vertically from a pregnant mother to her developing fetus [5–9]. Infection in utero can cause congenital malformations including microcephaly [5, 10–12].

Evidence of sexual transmission prompted exploration of the male and female reproductive tracts as reservoirs for ZIKV. Several groups have demonstrated in mice that ZIKV exhibits tropism for the male reproductive tract and causes inflammation [13–15] that leads to structural damage and breakdown of the blood-testis barrier. Zika virus ribonucleic acid (RNA) has been identified in sperm, with long-term viral shedding in human semen that can persist for up to 6 months and is associated with genitourinary symptoms and oligospermia [13, 16–18]. In females, ZIKV RNA has been detected in cervical mucus and vaginal secretions at time points after clearance from the blood [19, 20]. Infectious ZIKV and RNA also have been detected in the ovary of female mice and nonhuman primates [21–24]. One in vitro fertilization (IVF) human case described oocytes that were positive for ZIKV RNA [25]. Given the impact of ZIKV infection on male reproductive tissues, we investigated its effects on the ovary in greater detail.

The ovary is resistant to infection by most pathogens due to the presence of a blood-follicle barrier (BFB). The BFB comprises vascular endothelium, a subendothelial basement membrane, the thecal interstitium, the follicular basement membrane, and the membrane granulosa [26]. The BFB is permeable to small proteins and has charge-selectivity [27]. The permeability of the BFB also varies throughout folliculogenesis [28]. Even with these protective mechanisms, the ovary still is vulnerable to infection by some viruses. Mumps virus (MuV), a paramyxovirus, can be gonadotoxic to both the ovary and testis. Mumps virus causes acute inflammation of the ovary (oophoritis) in a small percentage (5%–7%) of infected postpubertal women [29–31]. Acute oophoritis is not always recognized clinically, but it can present with pelvic pain, fever, and vomiting [31]. A proportion of women with MuV oophoritis develop premature ovarian insufficiency due to early follicle depletion. Although there are reports that oophoritis can be caused by cytomegalovirus infection, it is a rare clinical entity [32, 33].

We characterized ZIKV infection and the ensuing immune response in the ovary using a murine model, and we found that it can be acutely infected with either Asian or African strains. By 7 days postinfection (dpi), CD4⁺ and CD8⁺ T cells had infiltrated into the ovary causing oophoritis, with ZIKV-specific CD8⁺ T cells being detected. Infection and inflammation was associated with an increase in number of terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)-positive, dying cells during the acute stage of infection. In contrast to results in the testis, excessive and persistent tissue damage was not seen in the ovaries of ZIKV-infected mice, and persistent viral RNA was not detected in most animals at 90 dpi. The resolution of infection and tissue injury was associated with no detrimental long-term impact on fertility or ovarian reserve. Thus, even though the ovary is vulnerable to acute ZIKV infection and inflammation, long-term effects on fertility were not observed, at least in mice.

MATERIALS AND METHODS

Ethics Statement

This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocols were approved by the Institutional Animal Care and Use Committee at the Washington University School of Medicine (Assurance number A3381-01, Protocol 20180234). Inoculations were performed under anesthesia induced and maintained with ketamine hydrochloride and xylazine, and all efforts were made to minimize animal suffering.

Viruses

Zika virus strains Dakar 41525 (Senegal, 1984) and PRVABC59 (Puerto Rico, 2015) were provided by the World Reference Center for Emerging Viruses and Arboviruses at the University of Texas Medical Branch (R. Tesh, S. Weaver, and K. Plante). The ZIKV Dakar 41525 isolate was passaged in Rag1^−/− mice to create a more pathogenic variant [34]. Zika virus H/PF/2013 (French Polynesia, 2013) was provided by the Arbovirus Branch of the Centers for Disease Control and Prevention with permission (X. de Lamballerie). All virus stocks were propagated in mycoplasma-free Vero cells (ATCC) and titrated by focus-forming assay, as described previously [35].

Mouse Infection Experiments

Wild-type (WT) and congenic CD8^−/− and TCRβδ^−/− C57BL/6 mice were purchased commercially (000664, 002665, 002122, respectively; Jackson Laboratories) and bred at Washington University in a pathogen-free facility. Female mice (8-week-old) were inoculated by a subcutaneous route in the footpad with 10⁶ focus-forming units (FFU) of ZIKV in a volume of 50 μL. One day before virus inoculation, mice were treated with 1 mg (ZIKV Dakar 41525) or 2 mg (ZIKV PRVABC59 and ZIKV H/PF/2013) of anti-Ifnar1 mouse monoclonal antibody ([mAb] MAR1-5A3; Leinco Technologies) via intraperitoneal injection. Ovaries were collected and processed as described below at 7, 14, 21, and 90 dpi. For long-term (90 dpi) studies, 1 mg of anti-Ifnar1 and 10³ FFU of ZIKV Dakar 41525 were used to enhance survival rates.

Viral RNA Burden

Mice were euthanized on specific days. Ovaries were homogenized with a Bullet Blender Storm 24 homogenizer instrument (Next Advance) for 1 to 2 minutes in 300 μL Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (FBS). Homogenates were clarified by centrifugation at 6000 ×g for 5 minutes and stored at −80°C. The RNA was extracted using an Applied Biosystems 5× MagMax RNA 96 viral isolation kit (Thermo Scientific) on a Kingfisher duo prime extraction machine (Thermo Scientific). Zika virus RNA levels were determined by one-step quantitative reverse transcription (qRT)-quantitative polymerase chain reaction (PCR) using an Applied Biosystems TaqMan RNA-to-Ct 1-step kit (Thermo Scientific) on an ABI 7500 Fast Instrument using standard cycling conditions: reverse transcription - 48°C, 15 minutes, 95°C, 10 minutes, followed by 40 cycles of 95°C, 15 seconds, 60°C, 1 minute. Published primer and/or probe sets for qRT-PCR of ZIKV were used (Asian/American ZIKV strains [36]; Dakar African strain [13]). As described and validated previously [8], viral burden was expressed on a log₁₀ scale as viral RNA equivalents per gram or microliter after comparison with a standard curve produced using serial dilutions of ZIKV RNA for each strain.

Flow Cytometric Analysis

Ovaries were digested as previously described [37]. Tissues were incubated with Dispase for 15 minutes at 37°C, followed by 0 minutes at 37°C incubation with collagenase D and DNase I (all enzymes; Roche). After digestion, tissues were passed through a 70-μm filter before staining. Cells were maintained on ice in phosphate-buffered saline (PBS) supplemented with 2% FBS and 1 mM ethylenediaminetetraacetic acid. After blocking Fcγ receptors with anti-CD16/32 (clone 93; eBioscience), ZIKV tetramer staining was performed at 37°C for 20 minutes in PBS with 2% FBS. Major histocompatibility complex class I tetramers (D^b) were prepared with ZIKV E protein peptide (residues 294 to 302; ABI Peptides) by the Center of Human Immunology and Immunotherapy Programs at Washington University and conjugated to streptavidin-allophycocyanin (APC) (Invitrogen). Cells also were stained for viability (Fixable Viability Dye eFluor 506; eBioscience) and surface antigens CD45 (fluorescein isothiocyanate [FITC]), CD3 (BUV395), CD19 (APC-Cy7), CD4 (BV605), and CD8 (PE-Cy7) at 4°C for 1 hour. All antibodies were purchased from BioLegend. Cell counts were performed by adding Precision count beads (BioLegend) to samples before data acquisition. Samples were processed on an LSRII flow cytometer and analyzed using FlowJo software X10.0.7.

Histology and Immunohistochemistry

Tissues were collected at necropsy and fixed overnight in 4% paraformaldehyde (PFA) in PBS. Subsequently, 3-μm-thick ovarian sections were processed for histology by hematoxylin and eosin staining. For immunohistochemistry, the tissue sections were incubated with rabbit monoclonal anti-CD4 (product no. ab183685, Abcam) or anti-CD8 (product no. 98941T, Cell Signaling Technology). After washing, slides were incubated with secondary antibody, ImmPRESS HRP Antirabbit IgG (product no. MP-7451, Vector Laboratories), for 1 hour and mounted. Immunostaining was analyzed by brightfield microscopy.

Viral RNA In Situ Hybridization

The RNA in situ hybridization (ISH) was performed using RNAscope 2.5 (Advanced Cell Diagnostics) according to the manufacturer’s instructions. The PFA-fixed, paraffin-embedded tissue sections were deparaffinized by incubating for 60 minutes at 60°C. Endogenous peroxidases were quenched with H₂O₂ for 10 minutes at room temperature. Slides were boiled for 15 minutes in RNAscope Target Retrieval Reagents and incubated for 30 minutes in RNAscope Protease Plus before probe hybridization. The probe targeting ZIKV RNA was designed and synthesized by Advanced Cell Diagnostics (catalog no. 467871). Positive (targeting plr2a gene) and negative (targeting bacterial gene dapB) control probes also were obtained from Advanced Cell Diagnostics (catalog nos. 312471 and 310043, respectively). Tissues were counterstained with Gill’s hematoxylin and visualized using brightfield microscopy.

TUNEL Assay

The DeadEnd Fluorometric TUNEL Kit (product no. G3250, Promega) was used according to the manufacturer’s instructions. Tissues were visualized using a NanoZoomer. ImageJ software was used to quantify the intensity and area of fluorescent puncta.

Follicle Counts

Zika virus-infected mice were euthanized at 90 dpi. Ovaries were collected at necropsy and fixed overnight in 4% PFA in PBS. Subsequently, 3-μm-thick serial sections were stained with hematoxylin and eosin. Differential counts of primordial, primary, secondary, and antral follicles were performed in a blinded manner under ×60 magnification, counting every fifth section.

Examination of Meiotic Spindles

Metaphase II (MII) oocytes from uninfected and ZIKV-infected mice at 90 dpi were fixed with 4% PFA for 30 minutes, permeabilized (0.5%, v/v, Triton X-100) for 20 minutes, and blocked (1%, w/v, bovine serum albumin-supplemented PBS) for 1 hour at room temperature. Samples were incubated with anti-α-tubulin FITC antibody (Sigma-Aldrich) and counterstained with 4’,6-diamidino-2-phenylindole (DAPI) for 1 hour. After washing, MII oocytes were mounted in VECTASHIELD media and analyzed by fluorescence microscopy. Alignment of spindles and chromosomes in each oocyte was categorized as normal or abnormal in a blinded manner.

Fertility Studies

To induce superovulation, mice were stimulated after intraperitoneal injection of pregnant mare serum gonadotropin (5 IU) and human chorionic gonadotropin (5 IU) (Harbor-UCLA Research Education). Thirteen hours after human chorionic gonadotropin injection, oviducts were collected and the ampulla punctured to release ovulated oocytes. In vitro fertilization was performed as previously described [38]. Sperm from uninfected control 12-week-old C57BL/6 males was used.

Data Analysis

All data were analyzed with GraphPad Prism software. For viral burden studies, the log₁₀-transformed titers were analyzed by Kruskal-Wallis one-way analysis of variance (ANOVA) with a multiple comparisons posttest. Immune cell infiltrates, TUNEL staining, and IVF analysis were analyzed by Mann-Whitney test. A Kruskal-Wallis ANOVA with multiple comparison posttest correction was used for the longitudinal analysis of TUNEL staining. A two-way ANOVA was used to analyze ovary follicle counts. A P value of <.05 indicated statistically significant differences.

RESULTS

Zika Virus Exhibits Tropism for the Ovary

We performed a longitudinal infection study in the ovaries of 8-week-old female WT C57BL/6J mice after subcutaneous inoculation of a mouse-adapted African ZIKV strain (Dakar 41525) in the hind footpad [34]. Because ZIKV does not efficiently antagonize type I interferon (IFN) signaling in mice compared with humans [39], animals were treated with a single dose of anti-Ifnar1 blocking mAb to facilitate infection and dissemination, as described previously [13, 35]. We detected ZIKV RNA (10⁴ –10⁶ FFU equivalents per gram) in the ovaries of most animals at 7 dpi (Figure 1A). Levels waned over time, but viral RNA was still present in the ovaries of the majority of mice at 21 dpi. Viral RNA levels were higher in the ovary than in serum at all time points (Figure 1A). To confirm these results using more contemporary strains, mice were treated with anti-Ifnar1 mAb and inoculated with ZIKV isolated from French Polynesia (H/PF/2013) or Puerto Rico (PRVABC59). At 7 dpi with these strains, approximately 10⁵ FFU equivalents per gram of viral RNA was detected (Figure 1B and C). To determine the regions in the ovary targeted by ZIKV, we performed ISH for viral RNA. This staining demonstrated the presence of ZIKV RNA in the stromal and follicular compartments at 7 and 14 dpi (Figure 1D). Viral RNA was not detected by ISH at 21 dpi, but this may be due to the sensitivity of the assay.

Figure 1.

Zika virus (ZIKV) infection in the ovary. (A–C) Eight-week-old female C57BL/6J mice were inoculated subcutaneously with 10⁶ focus-forming units (FFU) of Dakar 41525 (A), French Polynesia H/PF/2013 (B), or Puerto Rico PRVABC59 (C) strains of ZIKV. One (Dakar) or 2 (French Polynesia and Puerto Rico) mg of anti-Ifnar1 monoclonal antibody was given 24 hours before infection. Viral RNA in the serum and ovary was measured by reverse-transcription quantitative polymerase chain reaction. Dashed lines indicate limits of detection of the assay. Results are pooled from 2 or 3 experiments, and each data point represents an individual mouse. Zika virus Dakar 41525: day 7, n = 15; day 14, n = 13; and day 21, n = 11. Zika virus French Polynesia H/PF/2013: day 7, n = 10; and ZIKV Puerto Rico PRVABC59: day 7, n = 10. Bars indicate median values (Kruskal-Wallis one-way analysis of variance; *, P < .05 and ***, P < .001). (D) Viral RNA in situ hybridization staining was performed on ovary sections from ZIKV Dakar 41525-infected animals. Representative images: day 7, n = 15; day 14, n = 13; and day 21, n = 11 are shown. 4× scale bar = 500 μm; 10× scale bar = 200 μm; and 40× scale bar = 50 μm.

Acute Zika Virus Infection Is Associated With Cell Death in the Ovary

To assess the possible pathological consequences of ZIKV infection in the ovary, we performed a TUNEL assay to detect dying cells. We observed an increase in TUNEL-positive cells at 7 dpi in the ovary compared with uninfected animals (Figure 2A), the same time point when we detected peak viral burden (Figure 1). Whereas TUNEL-positive cells were identified in the follicular compartment in infected and uninfected ovaries, consistent with the physiological process of follicular degeneration, they were uniquely present in the stromal compartment of ovaries of ZIKV-infected mice (Figure 2B).

Figure 2.

Cell death in ovary after Zika virus (ZIKV) infection. Female, 8-week-old C57BL/6J mice were inoculated with 10⁶ focus-forming units of ZIKV Dakar 41525. Mice were given 1 mg of anti-Ifnar1 monoclonal antibody 1 day before infection. At 7 days postinfection, ovaries were removed and TUNEL staining was performed. Quantification of TUNEL-positive staining per millimeter of ovary was performed on samples from 7, 14, and 21 dpi (A). Each data point represents counts from an individual mouse, and median levels are shown. Naive: n = 19; day 7, n = 10; and day 14, n = 11; and day 21, n = 8 (Kruskal-Wallis one-way analysis of variance; *, P < .05). TUNEL staining in follicular granulosa cells is shown (green puncta and white arrows). TUNEL staining in stromal compartment is shown (green puncta and yellow arrows). (B). Representative images are shown. Scale bar = 0.5 mm. n.s., not significant.

T-Cell Infiltration Into Zika Virus-Infected Ovaries

To begin to determine whether the adaptive immune response contributed to cell death in ZIKV-infected ovaries, we quantified inflammatory cell infiltrates by flow cytometry. At 7 dpi, there was an increase in the percentage and number of CD4⁺ and CD8⁺ T cells in the ovary (Figure 3A–C). Zika virus-specific CD8⁺ T cells also were detected in the ovaries of infected but not uninfected mice (Figure 3B and D). To extend these findings, immunostaining for CD4 and CD8 was performed. Compared with uninfected controls, ovaries from ZIKV-infected mice showed infiltrating CD4⁺ and CD8⁺ T cells at 7 dpi (Figure 4), and these were present in both stromal and follicular compartments. To assess the contribution of T cells to the pathogenesis of ZIKV infection in the ovary, we inoculated CD8^−/− (lacking CD8⁺ T cells) and TCRβδ^−/− (lacking both CD4⁺ and CD8⁺ T cells) with ZIKV after a similar treatment with anti-Ifnar1 mAb. Because higher viral burden was detected in both to CD8^−/− and TCRβδ^−/− mice compared with WT mice, these adaptive immune cells contribute to the control ZIKV infection in the ovary (Figure 5A). Because the level of cell death in the ovaries of TCRβδ^−/− mice was not statistically different than WT mice at 7 dpi (Figure 5B) despite the higher viral burden, acute injury in the ovary may be a function of both viral burden and the subsequent T-cell response; indeed, we would have anticipated greater cell death in the ovaries of T cell-deficient than WT mice due to the higher level of ZIKV infection. Greater levels of ZIKV RNA were detected in the stromal and follicular compartments of the ovaries of CD8^−/− and TCRβδ^−/− mice compared with WT animals (Figure 5C).

Figure 3.

Acute oophoritis after Zika virus (ZIKV) infection. Eight-week-old female C57BL/6J mice were inoculated with 10⁶ focus-forming units of ZIKV Dakar 41525. Mice were given 1 mg of anti-Ifnar1 monoclonal antibody 1 day before infection. At 7 days postinfection, ovaries were removed, and cells were processed by flow cytometry to enumerate the percentage and number of CD4⁺ T cells (A and B), CD8⁺ T cells (B and C), and major histocompatibility complex class I tetramer-positive, ZIKV-specific CD8⁺ T cells (B and D). Results are pooled from 2 or 3 experiments, and each data point represents an individual mouse. Naive, n = 10; ZIKV-infected, n = 10. Bars indicate median levels (Mann-Whitney test; ****, P < .0001).

Figure 4.

CD4⁺ and CD8⁺ T-cell infiltration into the ovary of Zika virus (ZIKV)-infected mice. Eight-week-old female C57BL/6J mice were inoculated with 10⁶ focus-forming units of ZIKV Dakar 41525. Mice were given 1 mg of anti-Ifnar1 monoclonal antibody 1 day before infection. Immunohistochemistry analysis of uninfected and ZIKV-infected ovaries for CD4⁺ and CD8⁺ T cells (brown) at 7 days postinfection. Representative images of naive (n = 7) and ZIKV-infected (n = 10) are shown. 4× scale bar = 500 μm and 40× scale bar = 50 μm.

Figure 5.

Zika virus (ZIKV) infection and cell death in the ovaries of T cell-deficient mice. Wild-type (WT), CD8^−/−, and TCRβδ^−/− mice were inoculated with 10⁶ focus-forming units (FFU) of ZIKV Dakar 41525 as described above. At 7 days postinfection (dpi), viral RNA burden was measured in ovaries by quantitative reverse-transcription polymerase chain reaction (A). Dashed lines indicate limits of detection of the assay. Results are pooled from 2 or 3 experiments, and each data point represents an individual mouse. WT, n = 5; CD8^−/−, n = 11; and TCRβδ^−/−, n = 9. Bars indicate median levels (Kruskal-Wallis one-way analysis of variance; *, P < .05 and **, P < .01). TUNEL staining of ovaries from WT and TCRβδ^−/− mice (B). Quantification of TUNEL-positive staining per millimeter of ovary was performed. Each data point represents an individual mouse, and median levels are shown. WT, n = 15 (data are pooled from Figure 2 and performed in parallel in this experiment), TCRβδ^−/− n = 9 (Mann-Whitney test). (C) Viral RNA in situ hybridization staining was performed on ovary sections from ZIKV-infected WT, CD8^−/−, and TCRβδ^−/− mice at 7 dpi. Representative images: WT, ZIKV n = 3; WT, uninfected n = 2; CD8^−/−, ZIKV n = 6; CD8^−/−, uninfected n = 2; and TCRβδ^−/−, ZIKV n = 6; TCRβδ^−/−, uninfected n = 2; 10× scale bar = 200 μm and 40× scale bar = 50 μm. n.s., not significant.

Impact of Zika Virus Infection on Long-Term Ovarian Function

Given that ZIKV caused acute oophoritis, we explored its potential for long-term consequences. Female mice were given anti-Ifnar1 mAb followed by ZIKV inoculation, and animals were followed through 90 dpi, at which point ovaries were collected for measurement of viral RNA and histological analysis. At 90 dpi, persistent ZIKV RNA was largely absent in the ovary, with only 3 of 14 samples positive at low levels (Figure 6A). Consistent with this data, inflammation and tissue damage were not detected in sections of the ovary by histological analysis (Figure 6B). Follicle counts that assess ovarian reserve were similar between ZIKV-infected and uninfected age-matched mice (Figure 6C). To assess the impact on ZIKV on female fertility, a separate set of mice were superovulated at 90 dpi, and their oocytes were tested for morphology and IVF capacity. Metaphase oocytes were stained with an anti-α-tubulin antibody to visualize meiotic spindles, which in healthy oocytes are barrel-shaped, and chromosomes were stained with DAPI. Spindle and chromosome morphology appeared normal in oocytes from ZIKV-infected mice compared with controls (Figure 7A). We also observed no significant differences in oocyte yield, fertilization rate, or blastulation rate after IVF between oocytes harvested from ZIKV-infected mice or age-matched uninfected controls (Figure 7B–D). Thus, even though ZIKV infection in the ovary during the acute phase caused oophoritis and cell death, no durable deleterious consequences on ovarian reserve or fertility were observed.

Figure 6.

Long-term effects of Zika virus (ZIKV) infection on the ovary. Eight-week-old female C57BL/6J mice were inoculated with 10³ focus-forming units (FFU) of ZIKV Dakar 41525. Mice were given 1 mg of anti-Ifnar1 monoclonal antibody 1 day before infection. At 90 days postinfection (dpi), mice were superovulated, and ovaries were assessed for viral RNA by quantitative reverse-transcription polymerase chain reaction and compared with 7 dpi (A). Dashed lines indicate limit of detection of the assay. Results are pooled from 2 or 3 experiments, and each data point represents an individual mouse. Median levels are shown (day 7, n = 20; day 90, n = 14). (B) Hematoxylin and eosin staining was performed at 90 dpi on age-matched uninfected and ZIKV-infected mice. Representative images of 10 mice are shown, 10× scale bar = 200 μm. Follicle counts were performed on ovary sections (C). Naive, n = 3; ZIKV, n = 4. n.s., not significant.

Figure 7.

Impact of Zika virus (ZIKV) on long-term ovarian function and fertility. Eight-week-old female C57BL/6J mice were inoculated with 10³ focus-forming units of ZIKV Dakar 41525. Mice were given 1 mg of anti-Ifnar1 monoclonal antibody 1 day before infection. At 90 days postinfection, mice were superovulated and in vitro fertilization was performed. (A) Representative images of 4 mice of metaphase II oocyte meiotic spindles collected from uninfected and ZIKV-infected ovaries. Red arrows, spindles (α-tubulin); yellow arrows, chromatid pairs. Scale bar = 10 μm. (B–D) Numbers of oocytes (B), blast rate (the number of blastocysts counted 4 days after fertilization compared with the total number of fertilized cells) (C), and fertility rate (the total number of cells that divided 1 day after fertilization compared with the total number of recovered oocytes) (D) were measured. Naive, n = 9; ZIKV n = 14. Results are pooled from 2 experiments, and each data point represents an individual mouse. Median levels are shown (Mann-Whitney test). n.s., not significant.

DISCUSSION

Our study demonstrates that ZIKV has tropism for the ovary in a mouse model, which mirrors the findings of ZIKV in the male gonad, and corroborates previous findings of viral RNA or infectious virus in tissue homogenates from the ovary [13–15, 21–24]. Localization of ZIKV RNA in the ovarian granulosa cells, thecal cells, and ovarian vessels was visualized previously in sections from more immune-deficient AG129 mice (lacking IFN-α/β/γ signaling) infected by sexual transmission [22]. We visualized ZIKV RNA by ISH in cells of the follicular compartment, indicating ineffectiveness of the BFB in preventing access of the virus to the follicular space. Consistent with this observation, ZIKV RNA was isolated from the follicular fluid of oocytes of humans undergoing IVF [25]. Thus, despite the BFB and the ability of ovarian stromal cells to detect viral pathogen-associated molecular patterns [40], ZIKV gained entry and replicated within this immune-privileged site through an undefined mechanism.

Zika virus infection caused cell damage and death in the ovary during the acute phase. Some of the features are similar to MuV infection, which causes granulosa cell apoptosis [41]. We detected immune cell infiltrates in the ovary in response to ZIKV infection, consistent with mild oophoritis. Although small numbers of T cells are observed in the stromal and thecal layers throughout the menstrual cycle, most surround atretic follicles [42, 43]. We observed higher numbers of T lymphocytes in ZIKV-infected ovaries, with CD4⁺ and CD8⁺ T cells apparent in the stroma and the follicle. In a healthy ovarian follicle, immune cells are absent from the basement membrane and follicular space [44, 45]. In murine models of autoimmune-mediated premature ovarian insufficiency, there is an intense lymphocytic infiltrate of thecal cells and higher levels of degenerating follicles [46]. In women with autoimmune premature ovarian insufficiency, activated CD4⁺ T cells are observed throughout the ovaries [47, 48]. Although the mechanistic basis for cell death in the ZIKV-infected ovary remains unclear, it could be mediated by both viral-infection and immune cell response. Indeed, T cell-deficient mice (CD8^−/− and TCRβδ^−/−) showed similar TUNEL staining in the ovary compared with WT mice despite higher levels of virus infection.

Autoimmune premature ovarian insufficiency and MuV infection can cause enough damage to deplete the follicular pool and lead to ovarian failure. However, we did not detect decreased follicle counts in our mice after recovery from ZIKV infection, and they had similar responses to ovarian stimulation compared with controls. It remains possible that ZIKV has an effect on ovarian reserve in only a subset of animals, which we could not detect with the cohort of animals tested. There are other limitations inherent to the ZIKV mouse model. Mice are not innately susceptible to ZIKV infection, necessitating treatment with the anti-Ifnar1 mAb before infection and the use of a mouse-adapted ZIKV strain for the bulk of our experiments. Zika virus infection studies in newly generated human STAT2 knock-in, more immunocompetent mice [34] with contemporary ZIKV isolates are planned to address this limitation. We also did not synchronize the phases of the estrous cycle of the mice at the time of infection because we did not want to bias the results to a particular phase. The ovary undergoes many cyclic dynamic changes, with variations in immune cell populations and steroid hormone production. There is evidence that steroid hormones can modulate susceptibility to infection and the inflammatory response in female hosts [49, 50]. A study of ZIKV in a nonhuman primate model demonstrated that treatment with a progestin can increase susceptibility to infection [24]. Thus, the hormonal phase of the estrous cycle could impact acute ZIKV infection in the ovary and potentially affect the level of acute and long-term damage.

Given the discovery of ZIKV RNA in sperm from infected animals [13], we also plan to determine whether ZIKV RNA is present in follicular fluid or the oocyte itself in mice. Because human oocytes can harbor ZIKV [25], understanding the implications for potential embryonic and fetal infection, both with spontaneous conceptions and those using assisted reproductive technology, is necessary.

CONCLUSIONS

In summary, our studies in mice show that ZIKV infects the ovary and causes an acute oophoritis characterized by T-cell infiltration and cell death. Nonetheless, this acute inflammatory process did not appear to have long-term deleterious consequences for fertility in the model we used. Indeed, only a small minority of ovaries showed evidence of persistent ZIKV RNA at 90 dpi, and the levels detected are of uncertain significance. Longitudinal studies in ZIKV-infected nonhuman primates and women are needed to corroborate these findings and conclude definitively that there is a lack of effect of ZIKV on ovarian function and reserve.

Notes

Author contributions. E. A. C., S. M. S., D. E. B., M. S. D., and K. H. M. designed the experiments. E. A. C., S. M. S., D. E. B., V. S., and J. G. performed the experiments. E. A. C., S. M. S., M. E. S.-W., M. S. D., and K. H. M. performed data analysis. A. O., J. H., and S. P. provided key reagents or experimental advice. E. A. C., S. M. S., D. E. B., M. S. D., and K. H. M. wrote the initial draft of the manuscript with other authors providing editorial comments.

Financial support. This work was funded by grants from the National Institutes of Health (NIH) (R01 AI073755, R01 AI127828, and R01 HD091218 [to M. S. D.] and R01 HD083895 [to K. H. M.]) and an NIH Shared Instrumentation Grant (S10 RR0227552) to the Hope Center for Neurological Disorders at Washington University.

Potential conflicts of interest. M. S. D. is a consultant for Inbios and Atreca and is on the Scientific Advisory Board of Moderna. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.