The mechanisms underlying the immune control of Zika virus infection at the maternal-fetal interface

Abstract

Unlike other Flaviviruses, Zika virus (ZIKV) infection during the first trimester of pregnancy causes severe pregnancy outcomes including the devastating microcephaly and diseases associated with placental dysfunctions. We have previously reported that the maternal decidua basalis, the major maternal-fetal interface, serves as a replication platform enabling virus amplification before dissemination to the fetal compartment. However, the rate of congenital infection is quite low, suggesting the presence of a natural barrier against viral infection. Using primary cells from first-trimester pregnancy samples, we investigated in this study how the maternal decidua can interfere with ZIKV infection. Our study reveals that whether through their interactions with dNK cells, the main immune cell population of the first-trimester decidua, or their production of proinflammatory cytokines, decidual stromal cells (DSCs) are the main regulators of ZIKV infection during pregnancy. We also validate the functional role of AXL as a crucial receptor for ZIKV entry in DSCs and demonstrate that targeted inhibition of ligand-receptor interaction at the early stage of the infection is effective in drastically reducing virus pathogenesis at the maternal-fetal interface. Collectively, our results provide insights into the mechanisms through which ZIKV infection and spreading can be limited. The strategy of circumventing viral entry at the maternal-fetus interface limits virus dissemination to fetal tissues, thereby preventing congenital abnormalities.

Keywords: Zika virus; infection; inflammation; maternal-fetal interface; natural killer cells.

Figure 1

ZIKV Infection of DSCs Modulates the Expression of Key NK Cell Receptor Ligands. (A). DSCs were challenged with ZIKV at an MOI of 1. ZIKV viral protein was determined at day 3 post-infection (dpi) by flow cytometry using anti-NS3 antibody, followed by fluorochrome-conjugated secondary antibody staining. (B-D) Cell surface expression of specific NK cell receptor ligands determined at 3dpi by FACS. MOCK (white) and infected (red) DSCs. Percentage and mean fluorescence intensity (MFI) for the expression of HLA-ABC (B), HLA-E molecules (C) and NKp44 ligands, NKp44L (D). Data sets represent mean values ± SEM determined from at least five independent donors. P values are computed using paired two-tailed Student’s t test. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 2

dNK Cells Efficiently Control DSCs Infection. DSCs were infected with ZIKV at an MOI of 1 and co-cultured with autologous dNK cells at a ratio 1:5. (A-C) Viral loads are determined by qRT-PCR in cell-free culture supernatants, collected following the indicated time points d0, d3 and d5. (A) ZIKV-infected DSCs cultured alone (red line) or with autologous dNK cells added at day 0 (DSCs + dNK d0, blue line), day 1 (DSCs + dNK d1, dotted black line), day 2 (DSCs + dNK d2, dotted grey line) or day 3 (DSCs + dNK d3, solid black line) post-infection. (B) ZIKV-infected DSCs cultured alone (DSCs only, red), with autologous dNK cells in the same well (DSCs + dNK, blue) or separately by a 0.4µM porous membrane in transwell system (DSCs + dNK TW, green). (C) Fresh DSCs were infected with ZIKV (MOI 1) in the presence of UV-inactivated conditioned media collected from day 3 (CM d3) and day 5 (CM d5) co-cultures of dNK cells with uninfected (Mock) and ZIKV-infected DSCs (ZIKV). Histograms represent quantification of genome replication at day 0 (d0, black), day 3 (d3, blue) and day 5 (d5, red) post-viral challenge. Data sets are presented as mean values ± SEM from at least three independent donors. Statistical significance of differences was evaluated by repeated-measures analysis of variance with the Greenhouse and Geisser correction, and the Newman-Keuls post hoc test. *p<0.05, **p<0.01, ****p<0.0001. (D) Uninfected and ZIKV-infected DSCs plated on glass coverslips incubated with autologous dNK cells for 20 min at 37°C. Representative confocal images of conjugates with maximum intensity projection are shown. Lytic granules containing perforin (red), ZIKV-env protein (green), α-tubulin (blue) and DAPI nuclei (cyan). Scale bar, 10 µm. (E) Percentage of conjugates showing polarized perforin containing granules to the NK cell IS. Results from 5 independent conjugations were averaged. Values represent mean values ± SEM. At least 300 conjugates were analyzed in each experiment (n=3). P values are computed using paired two-tailed Student’s t test, ****p<0.0001.

Figure 3

ZIKV Infection Modulates the Decidual Environment and the Interferon Response. DSCs were infected with ZIKV at an MOI of 1. (A) Cytokines, (B) CCL-chemokines, (C) CXCL-chemokines, (D) IFN-α, IFN-β, IFN-γ, IFN-λ1 and IFN-λ2/3, quantified in the supernatant using a 42-multi-plexed cytokine assay 48 hours post-infection. (E, F) Total mRNA extracted 3 dpi and transcripts quantified using qRT-PCR and specific primers for IFNA, IFNB, IFNG, IFNL1, IFNL2/3, IFI6, IRF3, ISG20 and RSAD. Histograms showing significant differences between uninfected (white bars) and ZIKV-infected DSCs (red bars) are presented as fold change of relative mRNA expression (FC). (G) ZIKV-infected DSCs cultured with or without recombinant human IFN-γ (100ng/mL). IFN-γ was added at day 0 (d0, blue line), d1 (dotted black line), d2 (dotted grey line) or d3 (black line) post-infection. Non treated ZIKV-infected DSCs (red line) are used as positive control. Culture supernatants were collected from the different conditions following the indicated time points and viral loads were determined by qRT-PCR. All data sets are normalized and presented as mean values ± SEM from at least five independent donors. Statistical significance of differences was evaluated by repeated-measures analysis of variance with the Greenhouse and Geisser correction, and the Newman-Keuls post hoc test. ns, not significant. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 4

AXL Receptor Is Essential for ZIKV Infection of DSCs. (A) Fluorescence immunostaining for cell surface molecules, AXL, TYRO3, DC-SIGN using specific antibodies. Representative histograms gated on live cells are shown. Specific receptor staining (Pink), isotype-matched Ig control (Blue). One representative histogram out of five independent experiments is shown. (B) Representative images of maximum intensity projection of DSCs. AXL (green), DAPI nuclei (cyan). Scale bar, 20 µm. Data are representative of at least five independent experiments from five independent donors. (C) Western blot analysis of AXL expression. β-actin as the loading control, DC (dendritic cells) expressing AXL protein as positive control. (D) ZIKV-DSCs incubated with or without MYD1 decoy receptor, added during infection (T 0h) or 12 hours (T 12h) and 24 hours (T 24h) post-infection. Immunostaining performed 5 dpi in the presence of MYD1 treatment. NS3 (green), α-tubulin (blue) and DAPI nuclei (cyan). Scale bar, 20 µm. Data are representative of at least three independent experiments. (E) ZIKV replication in mock treated DSCs (red) and in DCSs that were treated with MYD1 during the infection(white) or 12 hours post-viral challenge (grey). Viral loads were determined by qRT-PCR at days 0,3 and 5 post challenge. Bar graphs represents the mean values ± SEM determined from five independent donors. P values are computed using one-way ANOVA with the Dunn post hoc test for comparisons of different experimental conditions. *p<0.05, **p<0.01. (F) Monolayers were challenged with culture supernatants collected from uninfected DSCs (MOCK) and from cells that were infected with ZIKV in the presence (MYD1) or absence (ZIKV) of MYD1. Plaques were observed at 5 dpi.