Transplacental Zika virus transmission in ex vivo perfused human placentas

Abstract

A Zika virus (ZIKV) infection during pregnancy can result in severe birth defects such as microcephaly. To date, it is incompletely understood how ZIKV can cross the human placenta. Furthermore, results from studies in pregnant mice and non-human primates are conflicting regarding the role of cross-reactive dengue virus (DENV) antibodies on transplacental ZIKV transmission. Elucidating how ZIKV can cross the placenta and which risk factors contribute to this is important for risk assessment and for potential intervention strategies for transplacental ZIKV transmission. In this study we use an ex vivo human placental perfusion model to study transplacental ZIKV transmission and the effect that cross-reactive DENV antibodies have on this transmission. By using this model, we demonstrate that DENV antibodies significantly increase ZIKV uptake in perfused human placentas and that this increased uptake is neonatal Fc-receptor-dependent. Furthermore, we show that cross-reactive DENV antibodies enhance ZIKV infection in term human placental explants and in primary fetal macrophages but not in primary trophoblasts. Our data supports the hypothesis that presence of cross-reactive DENV antibodies could be an important risk factor for transplacental ZIKV transmission. Furthermore, we demonstrate that the ex vivo placental perfusion model is a relevant and animal friendly model to study transplacental pathogen transmission.

Fig 1. Efficient uptake of ZIKV immune complexes by ex-vivo perfused placentas.

1*10⁵ TCID₅₀ equivalent/mL inactivated ZIKV (ZIKV^BPL) was incubated with either flavivirus naïve serum or serum containing DENV-2 nAbs and added to the maternal circulation (MC) of the placental perfusion model and perfused for 40 or 120 minutes. A: Schematic overview of the ex vivo dual placental perfusion model. FC; fetal circulation. Created with Biorender.com. B: ZIKV RNA levels in the MC were determined every 15 minutes with RT-PCR up to 120 minutes to detect placental uptake of ZIKV^BPL. Dots represent mean ±SEM, N = 3–4 donors per condition. C: Mean fold reduction (+SEM) in ZIKV TCID₅₀ equivalent in the MC after 120 minutes of perfusion with and without different concentrations of protein G (Prot. G) to block the interaction of IgG with FcRn. N = 1–3 donors per condition. D: ZIKV RNA levels in the MC were determined every 10 minutes with RT-PCR up to 40 minutes. Dots represent mean±SEM, N = 2 donors per condition. E: ZIKV RNA was detected in tissue biopsies of the perfused placentas after 40 minutes of perfusion. Horizontal lines represent median and the 10^th and 90^th percentile cut-off. N = 2 donors per condition and 40 biopsies per condition. Data from the MC were analyzed with multiple t-tests with the Holm-Šidák correction and data from tissue lysates were analyzed with the Mann-Whitney U test. ****P<0.0001.

Fig 2. Detection of ZIKV RNA in chorionic villi of placentas perfused with ZIKV.

ZIKV RNA was detected with ISH in formalin fixed, paraffin embedded tissue of perfused placentas. Chromogenic staining for cytokeratin-7 (cyto-7) and CD163 was performed in sequential slides to detect trophoblasts and Hofbauer cells, respectively. A: Staining for ZIKV RNA, trophoblasts and Hofbauer cells in a placenta that was perfused for 40 minutes with ZIKV^BPL+DENV nAbs. B: Staining for ZIKV RNA, trophoblasts and Hofbauer cells in a placenta that was perfused for 120 minutes with ZIKV^BPL+flavivirus naïve serum. C: Staining for ZIKV RNA in a placenta that was obtained directly after birth and was not perfused. Arrows indicate positive signal for ZIKV RNA.

Fig 3. Cross-reactive flavivirus antibodies enhance ZIKV infection in placental explants.

A&B: Term human placental explants were infected with 1.0x10⁵ TCID₅₀/mL ZIKV alone or ZIKV that was preincubated with flavivirus naïve serum (ZIKV+control) or serum containing DENV-2 neutralizing antibodies (ZIKV+DENV nAbs), both in a dilution of 1:250, or with a humanized panflavirus monoclonal antibody (hu4G2, 1μg/mL). ZIKV titers were determined in supernatants (A) and ZIKV RNA levels were determined with RT-PCR in tissue lysates (B). C&D: Term human placental explants were pre-treated with FcyR blocking antibodies or protein G was added to ZIKV–DENV nAbs immune complexes. Subsequently, the explants were infected with either 1.0x10⁵ TCID₅₀/mL ZIKV or ZIKV+DENV nAbs. ZIKV titers were determined in supernatants at two dpi (C) and ZIKV RNA levels were determined in tissue lysates with RT-PCR at two dpi (D). N = 3–4 donors per condition, on average 12 explants per donor. Horizontal lines in the violin plots represent median and the 10^th and 90^th percentile cut-off. Statistical significance was determined using the Kruskal-Wallis test followed by Dunn’s post hoc test. * P<0.05, **P<0.01, ****P<0.0001.

Fig 4. ZIKV infects Hofbauer cells and trophoblasts in placental explants in presence DENV nAbs.

A-D: Representative pictures of ZIKV infection in Hofbauer cells (CD163) and trophoblasts (Cyto-7), in term human placental explants infected with ZIKV that was pre-incubated with human serum containing DENV-2 nAbs (A&B) or a humanized pan-flavivirus monoclonal antibody (hu4G2, C&D). Arrows indicate cells stained with either Cyto-7 or CD163 that correspond with cells in which ZIKV RNA was detected. E-G: Staining for ZIKV RNA in placental explants infected with ZIKV+flavivirus naïve serum (E), only ZIKV (F) or uninfected (G). All tissues were stained six days after (mock) infection.

Fig 5. Primary Hofbauer cells are permissive for ZIKV infection and ADE of ZIKV infection.

Hofbauer cells (HBCs) and trophoblasts were isolated from term human placentas and infected with ZIKV+flavivirus naive serum (ZIKV+control) or ZIKV+DENV nAbs at an MOI of 0.5 in presence or absence of FcγR blocking antibodies and protein G, for 48 hours. A: Confocal laser scanning microscopy image of ZIKV+DENV nAbs infected HBCs and uninfected HBCs. B: Confocal laser scanning microscopy image of ZIKV+DENV nAbs infected trophoblasts and uninfected trophoblasts. HBCs were visualized by fluorescent staining for CD68, trophoblasts for cytokeratin-7 (Cyto-7), ZIKV by staining for ZIKV envelope protein (ZIKV-E) and nuclei with Hoechst 33342 staining (DNA). C: Percentage of infection of HBCs and trophoblasts was determined with confocal laser scanning microscopy. Bars represent mean+SEM. Significance was determined with a Student’s T-test. D&E: ZIKV titers were determined in supernatants of HBCs and trophoblasts. Bars represent median+95%CI. Significance was determined using the Kruskal-Wallis test followed by Dunn’s post hoc test, comparing ZIKV+DENV nAbs without block to the other conditions. F-I: Cytokines were determined in the supernatants of HBCs with a multiplex bead-based assay. Each dot represents one value of experiments performed in triplicate/quadruplicate, lines represent mean ±SEM. Significance was determined using one-way ANOVA with Dunnett’s post hoc test. N = 2–3 donors per condition for all experiments. *P<0.05, ** P<0.01, ***P<0.001, ****P<0.0001.

Fig 6. Proposed mechanism of ADE of ZIKV infection in the term human placenta.

Left panel: In absence of cross-reactive DENV antibodies, ZIKV crosses the placental barrier less efficiently than in presence of cross-reactive flavivirus antibodies through a mechanism that is not fully elucidated yet. Right panel: In presence cross-reactive DENV antibodies, ZIKV immune complexes can be transported across the syncytiotrophoblasts layer through FcRn-mediated transcytosis. In the villus core, some non-neutralized complexes are taken up by the perivascular located HBCs through FcγRII, after which ZIKV can replicate in these cells. To reach the fetal circulation, ZIKV needs to subsequently cross the fetal endothelial barrier, possibly by infecting these cells. IVP; intervillous space, FcRn; neonatal Fc-receptor, STBs; syncytiotrophoblasts, CTBs; cytotrophoblasts, HBC; Hofbauer cell, FcγRII; Fcγ-receptor II. Created with Biorender.com.