Zika Virus infection of rhesus macaques leads to viral persistence in multiple tissues

Abstract

Zika virus (ZIKV), an emerging flavivirus, has recently spread explosively through the Western hemisphere. In addition to symptoms including fever, rash, arthralgia, and conjunctivitis, ZIKV infection of pregnant women can cause microcephaly and other developmental abnormalities in the fetus. We report herein the results of ZIKV infection of adult rhesus macaques. Following subcutaneous infection, animals developed transient plasma viremia and viruria from 1-7 days post infection (dpi) that was accompanied by the development of a rash, fever and conjunctivitis. Animals produced a robust adaptive immune response to ZIKV, although systemic cytokine response was minimal. At 7 dpi, virus was detected in peripheral nervous tissue, multiple lymphoid tissues, joints, and the uterus of the necropsied animals. Notably, viral RNA persisted in neuronal, lymphoid and joint/muscle tissues and the male and female reproductive tissues through 28 to 35 dpi. The tropism and persistence of ZIKV in the peripheral nerves and reproductive tract may provide a mechanism of subsequent neuropathogenesis and sexual transmission.

Fig 1. Cohort design and timeline used for ZIKV infection of rhesus macaques.

Graphical depiction of the cohorts and timeline used for the infection of Rhesus monkeys with ZIKV. Blood and urine were collected daily for the first 10 days post infection and then at 14, 21, 28, and 35 dpi. Cohorts 1 and 2 and 3 were taken to necropsy at 28, 7, and 35 dpi, respectively. Cohort 1 was used to examine ZIKV dose effect and viral persistence. Cohorts 2 and 3 was used to determine viral tissue tropism during acute infection as well as at later times after resolution of serum viremia. Monkey identification number, age and sex as well as the infectious dose of ZIKV used to infect the animals are listed for each cohort.

Fig 2. Viral loads in plasma and urine from ZIKV-infected rhesus macaques.

One-step qRT-PCR was used to measure ZIKV RNA loads in the plasma (top panels) and urine (bottom panels) from each animal at indicated days pi and represented as copies per milliliter of fluid. (A) Cohort 1 and 3: Animal 24961(1x10⁴ ffu)-red lines; 25147 (1x10⁵ ffu)-blue lines; 25421 (1x10⁶ ffu)-black lines; 26021 (1x10⁵ ffu)-green lines; 26023 (1x10⁵ ffu)-purple lines. (B) Cohort 2: Animal 24504 (1x10⁵ ffu)-red lines; and Animal 27679 (1x10⁵ ffu)-blue lines. 1/10^th of total RNA extracted from 100 μl plasma or urine was used in each reaction. Approximate limit of detection at 1e4 genomes/ml is based on a detection limit of ~100 genomes in each reaction (S1 Fig) as indicated by the dotted line. Asterisks indicate infectious virus co-cultured from plasma harvested on day shown.

Fig 3. Viral loads in the tissues following necropsy of ZIKV-infected rhesus macaques.

One-step qRT-PCR was used to measure ZIKV RNA loads in the tissues of animals in Cohort 2 (A), Cohort 1 (B), and Cohort 3 (C). Total RNA was generated using the Trizol method on precleared samples following bead beating. Approximately 80 different tissues were assessed for the presence of viral RNA. Shown are the tissues with positive detection in at least one of the animals per cohort. Arrows indicate samples in which virus was successfully co-cultured from tissue homogenate. Approximate limit of detection at 1e4 genomes/ml is based on a detection limit of ~100 genomes in each reaction (S1 Fig) as indicated by the horizontal line. (D) Paraffin sections of sciatic nerve cut in cross section were hybridized with ZIKV specific chromogenic probe (red) and counterstained with hematoxylin (blue). Nerve fibers (NF) show a normal distribution within the endoneurium surrounded by perineurium (PN) and perineurial adventia (PA). Hybridization for ZIKV was robust but limited to the PA region. Original magnification was 50X.

Fig 4. Infected cell types in spleen and axillary lymph nodes of ZIKV-infected rhesus macaques.

Cell subpopulations were isolated by positive selection magnetic bead separation from lymphocytes isolated from the spleen and axillary lymph nodes at 28 dpi. For macrophage and T cell isolation, CD14-microbeads were used to isolate macrophages from total spleen and axillary lymph node lymphocytes, and then anti-CD3 was used to isolate T cells from macrophage depleted flow-through. For B cell and DC isolation, total splenocytes or axillary lymph node lymphocytes were first positively selected for CD20⁺ B cells and the depleted fraction was bound to CD1c microbeads to isolate DCs. To increase purity of the isolated cell populations all positively selected samples were eluted after primary selection and then re-bound to a second fresh column. Depiction of isolated cell populations as characterized by flow cytometry is shown in S4 Fig. Total RNA was isolated from the positively selected cell fractions and quantified. One-step qRT-PCR was used to measure ZIKV RNA loads in each of the cell fractions isolated from animals at 28dpi (A). Viral RNA loads were highest in the macrophage and B cell fractions, with consistently less vRNA detected in DCs and rarely in the T cells. (B) Serial paraffin sections of axillary lymph node tissue from animal #25421 were hybridized with ZIKV specific chromogenic probe (red) or Influenza specific chromogenic probe (red) and counterstained with hematoxylin (blue) showed strong positive staining for ZIKV in the lymphoid follicles. Original magnification was 50X.

Fig 5. ZIKV-infection induces robust innate immune cell activation.

Total peripheral blood mononuclear cells from all time points were stained with fluoroflore-conjugated antibodies directed against the cellular markers CD3, CD8, CD11c, CD14, CD16, CD169 and HLA-DR in order to assess changes in the activation of A) monocyte/macrophages; B) myeloid dendritic cells; C) other dendritic cells; and D) NK cells. Multi-color flow cytometry was used to visualize the stained cells. The percentage of activated cells (CD169⁺) was calculated using FlowJo and the data was graphed in GraphPad Prism v6 software.

Fig 6. Rhesus cytokine and chemokine production in response to ZIKV infection and block of NF-kB signaling in Rhesus fibroblasts.

A 29-plex-cytokine/chemokine/growth factor magnetic bead assay was performed on plasma from rhesus monkeys at all time points post infection. Cytokine analysis revealed changes in only A) IL-RA; B) MCP-1-CCL2; C) IP-10-CXCL10; and D) ITAC-CXCL11. Reporter assay showing induction of NF-κB-dependent (E, F) or interferon stimulated response element (ISRE)-dependent (G) LUC expression in fibroblasts infected for 56h with ZIKV at MOI = 5ffu/cell. Luminescence was measured 8h after treatment with 60μg/mL poly(I:C) (E) 100ng/mL human IL-1β (F) or 5,000 units/ml IFNβ1 (G). Values displayed are average fold changes (three replicates) of stimulated versus untreated cells ±SD.

Fig 7. Rhesus adaptive immune cell proliferative responses following ZIKV infection.

Total peripheral blood mononuclear cells were analyzed by flow cytometry for the presence of T and B cell proliferative responses following infection. T cells were identified by staining with antibodies directed against the cellular markers CD3, CD4, CD8β, CD95, CD28, CD127 and for intracellular levels of Ki67 (proliferation marker) to assess changes the proliferation of A) CD4⁺ central memory T cells; B) CD8⁺ central memory T cells; C) CD4⁺ effector memory T cells; and D) CD8⁺ effector memory T cells. B cells were stained with antibodies directed against CD3, CD20, CD27, IgD and HLA-DR as well as Ki67 in order to compare the proliferative responses in E) naïve B cell; F) memory B cells; and G) marginal-zone like B cells. The percentage of actively proliferating cells (Ki67⁺) was calculated using FlowJo and the data was graphed in GraphPad Prism v6 software.

Fig 8. Detection of anti-ZIKV antibody responses in Rhesus plasma.

Rhesus macaques infected with ZIKV were analyzed for the presence of antibodies directed against ZIKV-PRABC59 by ELISA using whole virus as capture antigen with an HRP-conjugated anti-Rhesus IgM (A) or IgG (B) secondary antibody. C. Sera from indicated animals obtained pre-infection (d0) or at terminal bleed (d28 or 35 pi) were tested for neutralizing activity via plaque reduction neutralization titer (PRNT) assay. D. Fold dilution giving 50% reduction in infectious titer for each serum sample.