Hydroxychloroquine-inhibited dengue virus is associated with host defense machinery

Abstract

Hydroxychloroquine (HCQ) is an antimalarial drug also used in treating autoimmune diseases. Its antiviral activity was demonstrated in restricting HIV infection in vitro; however, the clinical implications remain controversial. Infection with dengue virus (DENV) is a global public health problem, and we lack an antiviral drug for DENV. Here, we evaluated the anti-DENV potential of treatment with HCQ. Immunofluorescence assays demonstrated that HCQ could inhibit DENV serotype 1-4 infection in vitro. RT-qPCR analysis of HCQ-treated cells showed induced expression of interferon (IFN)-related antiviral proteins and certain inflammatory cytokines. Mechanistic study suggested that HCQ activated the innate immune signaling pathways of IFN-β, AP-1, and NFκB. Knocking down mitochondrial antiviral signaling protein (MAVS), inhibiting TANK binding kinase 1 (TBK1)/inhibitor-κB kinase ɛ (IKKɛ), and blocking type I IFN receptor reduced the efficiency of HCQ against DENV-2 infection. Furthermore, HCQ significantly induced cellular production of reactive oxygen species (ROS), which was involved in the host defense system. Suppression of ROS production attenuated the innate immune activation and anti-DENV-2 effect of HCQ. In summary, HCQ triggers the host defense machinery by inducing ROS- and MAVS-mediated innate immune activation against DENV infection and may be a candidate drug for DENV infection.

FIG. 1.

Hydroxychloroquine (HCQ) attenuates dengue virus infection in vitro. (A) A549, Hepa1-6 and WS-1 cells were pretreated with HCQ as indicated for 24 h before DENV-2 inoculation. Immunofluorescent assay (IFA) of DENV-2 infection with anti-DENV-2 NS3 antibody (left panels), and DAPI staining of nuclei (right panels) and quantification. (B) Plaque-forming assays of viral titers in DENV-2-infected A549 cells treated with HCQ as described in (A). (C) MTT assay of viability of A549, Hepa1-6, and WS-1 cells at 24 h after HCQ treatment. Data are mean±SD from 3 determinations. (D) A549 cells pretreated with 50 μM HCQ, 1 μM Bafilomycin A1 (Baf-A1), or 500 U type I IFN-α-2a for 24 h were infected with DENV-2 at multiplicity of infection (MOI)=1; IFA of infectivity with anti-DENV-2 NS3 antibody (left panels) and DAPI nuclear staining (right panels) and quantification. (E) A549 cells pretreated with 50 and 100 μM HCQ for 24 h were infected with DENV serotype 1–4 (MOI=5) for 24 h. Anti-NS3 (green fluorescence) and DAPI (blue florescence) stained images were merged to show DENV-infected cells, and quantification. *P≤0.05, **P≤0.01, ***P≤0.005 versus untreated control group.

FIG. 2.

HCQ pretreatment reduces DENV-2 infectivity. (A) A549 cells were treated with HCQ (50 μM) before, during and after viral adsorption (upper panel). A549 cells were mock-infected (a) or infected with DENV-2 for 24 h (b) or pretreated with HCQ for 18 h before DENV-2 infection (c), during DENV-2 adsorption for 2 h (d), or after DENV-2 adsorption for 24 h (e). IFA of DENV-2 infection with anti-DENV-2 NS3 antibody (left panels), and DAPI staining of nuclei (right panels, h–n) (combined HCQ treatment is shown: f and g) and (B) quantification of IFA infectivity results. Data are mean±SD of 3 observation fields. (C) Plaque-forming assay of DENV-2 titration with treatment as described in (A).

FIG. 3.

HCQ induces the activities of IFN regulatory factor 3 (IRF3), NFκB, and c-Jun. (A–C) Dual luciferase assay of IFN-β, NFκB, and AP-1 luciferase reporters. A549 cells (3×10⁵) transfected with Luc reporter (0.2 μg) and pRL-TK (0.02 μg) were treated with HCQ for 24 h at the indicated doses. Data are mean±SD from 3 independent tests. *P≤0.05, **P≤0.01 versus untreated group. (D) Immunoblot assay of levels of the indicated proteins in cell lysates from A549 cells treated or not with HCQ (50 μM). (E) IFA of NFκB p65 nuclear translocation with anti-NFκB p65 antibody (left panels), and DAPI staining of nuclei (right panels) in A549 cells (3×10⁵) with HCQ (50 and 100 μM) or polyI:C (1 μg) stimulation for 24 h, and (F) quantification. Data are mean±SD of 5 observation fields. ***P≤0.005 versus untreated group.

FIG. 4.

HCQ enhances mitochondrial antiviral signaling protein (MAVS) protein expression and its transactivity. (A) Immunoblot assay of endogenous MAVS level in A549 cells treated with 50 μM HCQ for 24 h. (B) Dual luciferase assay of IFN-β, ISRE, NFκB, and AP-1 luciferase reporters. A549 cells (1×10⁵) transfected with Luc reporter (0.5 μg) and pRL-TK (0.04 μg), plus empty vector or pcDNA3.1-V5-MAVS (0.8 μg) treated with HCQ (50 μM) for 24 h. (C) Immunoblot assay with anti-V5 antibody of ectopic V5-MAVS protein level in luciferase assay in (B). (D) Immunoblot assay of MAVS protein level in A549 cells (2×10⁵) transfected with MAVS siRNA (siMAVS) or negative control siRNA (siNC) for 48 h. β-actin was a internal loading control. (E) MAVS and control siRNA-transfected A549 cells were treated with HCQ for 24 h, then infected with DENV-2 at MOI=1. IFA of DENV-2-infected cells with anti-NS3 antibody (green color) and DAPI staining of nuclei (blue color). DENV-2 infectivity (F) and plaque assay of viral titration (G) of (E). (H) RT-qPCR analysis of MAVS mRNA expression in shLacZ or shMAVS stably expressed A549 cells (2×10⁵) treated with HCQ (50 μM) for 24 h. (I and J) IFA of nuclear translocation of IRF3 and NFκB p65 in A549 cells described in (H). Right panels, estimated rate of IRF3 and NFκB p65 nuclear translocation. Data are mean±SD from at least 3 independent tests, *P≤0.05; **P≤0.01; ***P≤0.005.

FIG. 5.

Blocking type I IFN signaling reduces HCQ anti-DENV-2 activity. (A) A549 cells (2×10⁵) were incubated with anti-type I IFN receptor antibody (1:100) for 8 h before HCQ treatment. After 24 h of 50 μM HCQ treatment, cells were infected with DENV-2 at MOI=1. Quantification of DENV-2 infectivity (left panel). Plaque-forming assay of DENV titers (right panel). (B) J774A.1 cells were incubated with anti-type I IFN receptor antibody (1:100) for 8 h before 30 μM HCQ incubation for 24 h, then infected with DENV-2 (MOI=10) plus anti-E antibody (1:100). Quantification of DENV-2 infectivity (left panel) and representative immunofluorescence images (right panels). TBK1/IKKɛ inhibitors BX795 (C) and Amlexanox (D) were co-incubated with 30 μM HCQ as indicated in A549 cells; then, cells were infected with DENV-2 (MOI=1) for 24 h. Quantification of DENV-2 infectivity (left panel) and immunofluorescence images (right panels) are shown. The data are mean±SD from 3 independent experiments, *P≤0.05; **P≤0.01.

FIG. 6.

HCQ induces reactive oxygen species (ROS) generation. (A) ROS were detected by DCFH-DA assay in the A549 cells (2×10⁵) with HCQ treatment or not for 24 h. DCF signal in A549 cells with H₂O₂ (3%) treatment for 5 min was a positive control. (B) Quantification of ROS production (DCF fluorescence) in HCQ- (left panel) or H₂O₂-treated A549 cells (2×10⁴) (right panel). Data are mean±SD (N=6) *P≤0.05; ***P≤0.005. (C) Immunoblots of H₂O₂-treated A549 cells (1×10⁶) with indicated antibodies. (D) A549 cells (2×10⁴) were pretreated with 10 mM N-tert-Butyl-(-phenylnitrone (PBN) for 2 h, then HCQ (50 μM) for 24 h. ROS were detected by DCFH-DA assay. Data are mean±SD (N=6) ***P≤0.005. (E) Left panels, A549 cells (2×10⁵) were pretreated with PBN (10 mM) for 6 h, then HCQ (50 μM) for 24 h. Cells were infected with DENV-2 at MOI=1. IFA of DENV-2 infectivity at 24 h with anti-NS3 antibody (green color) and DAPI staining of nuclei (blue color). Estimated DENV-2 infectivity in the right panel. Data are mean±SD (N=3) ***P≤0.005. (F) Immunoblots of MAVS, phospho-IRF3, and total IRF3 protein levels and loading control β-actin in A549 cells (5×10⁵) pretreated with PBN (10 mM) for 6 h, then HCQ (50 μM) for 24 h. (G) Dual luciferase assay of IFN-β, NFκB, and ISRE luciferase reporters. A549 cells (1×10⁵) cotransfected with Luc reporters (0.6 μg) and pRL-TK (0.06 μg), plus empty vector or pcDNA3.1-V5-MAVS (0.8 μg) were treated with PBN (10 mM) for 6 h, then HCQ (50 μM) for 24 h. Data are mean±SD from 3 independent experiments. **P≤0.001; ***P≤0.005.