Broad-Range Antiviral Activity of Hydrogen Sulfide Against Highly Pathogenic RNA Viruses

Abstract

Hydrogen sulfide is an important endogenous mediator that has been the focus of intense investigation in the past few years, leading to the discovery of its role in vasoactive, cytoprotective and anti-inflammatory responses. Recently, we made a critical observation that H2S also has a protective role in paramyxovirus infection by modulating inflammatory responses and viral replication. In this study we tested the antiviral and anti-inflammatory activity of the H2S slow-releasing donor GYY4137 on enveloped RNA viruses from Ortho-, Filo-, Flavi- and Bunyavirus families, for which there is no FDA-approved vaccine or therapeutic available, with the exception of influenza. We found that GYY4137 significantly reduced replication of all tested viruses. In a model of influenza infection, GYY4137 treatment was associated with decreased expression of viral proteins and mRNA, suggesting inhibition of an early step of replication. The antiviral activity coincided with the decrease of viral-induced pro-inflammatory mediators and viral-induced nuclear translocation of transcription factors from Nuclear Factor (NF)-kB and Interferon Regulatory Factor families. In conclusion, increasing cellular H2S is associated with significant antiviral activity against a broad range of emerging enveloped RNA viruses, and should be further explored as potential therapeutic approach in relevant preclinical models of viral infections.

Figures

Figure 1. GYY4137 inhibits replication of influenza viruses in vitro.

(a) MDCK cells were infected with influenza A viruses H1N1, H3N2 or influenza B at MOI of 0.01 and treated with 5 or 10 mM of GYY4137 1 h after infection. Cell supernatants and cell pellets were collected at 24 h p.i. to measure viral titers. (b) MDCK cells were infected with influenza virus H1N1 at MOI of 0.01 and treated with 10 mM of GYY4137 at 1, 3 or 6 h p.i. Cell supernatants and cell pellets were collected at 24 h p.i. to measure viral titers. The values represent the means of logarithmically-transformed titer values, error bars – standard error of means (SEM), n = 3; *p < 0.05; **p < 0.01; ***p < 0.001 treated versus untreated samples when compared using one-way ANOVA with Tukey post-hoc test.

Figure 2. Effect of GYY4137 treatment on influenza viral RNA and protein expression.

A549 cells were infected with influenza virus H1N1 at MOI of 1 and treated with 10 mM GYY4137 1 h p.i. Cells were harvested at the indicated time points after infection for total RNA and protein extraction. (a) Real-time PCR for H1N1 specific genes. Solid lines – H1N1 only, dashed lines – H1N1-infected cells treated with 10 mM GYY4137. The values represent fold change (2−ΔΔ(Ct)) of target viral RNA compared to the 3 h p.i. timepoint, after normalization using 18S RNA. Values represent arithmetic means, error – standard error of mean (SEM), n = 3; comparison are between H1N1-infected and GYY4137-treated samples at each time point *p < 0.05; **p < 0.01; ***p < 0.001 using separately for each time point. (b) Western Blot assay for H1N1 proteins. 15 μg of protein of cell lysates were used per lane. Membranes were probed using rabbit polyclonal anti-whole H1N1 antibody, and were stripped and reprobed with anti-human β-actin antibody for loading control.

Figure 3. GYY4137 inhibits replication of highly pathogenic enveloped RNA viruses in vitro.

(a–d) Viral titers and viral RNA expression of EBOV-eGFP, CCHFV 10200, RVFV ZH501 and RSSEV were assessed in Vero cells untreated or treated with GYY4137. Cells were infected with the above viruses at MOI of 1 and treated one hour later with 10 mM GYY4137 for EBOV or 5 mM for the other viruses. Cells and supernatants were collected at indicated time p.i. to extract RNA to assess viral RNA expression and to determine viral titers, respectively. Squares are virus-only samples, triangles are GYY4137-treated samples. For RNA expression data the values represent fold changes (2−ΔΔ(Ct)) of target RNA levels over control samples using delta-delta Ct method. 18S RNA was used as housekeeping control RNA. The values represent mean and SEM, n = 3; *p < 0.05; **p < 0.01; ***p < 0.001 treated versus untreated samples on the same day when compared using Mann-Whitney test. (e) Immunofluorescent and brightfield images of Vero cells infected with EBOV-GFP with or without GYY4137 at 6 day p.i. (f) Immunofluorescent staining of Vero cells infected with CCHFV, RVFV or RSSEV with or without GYY4137 using virus-specific antibodies at the indicated time points.

Figure 4. GYY4137 decreases influenza-induced cytokine and chemokine production.

Confluent monolayers of A549 cells were infected with H1N1 at MOI of 1 and treated with 10 mM GYY4137 at 1 h p.i. The concentration of cytokines (a) and chemokines (b) was determined in cell supernatants collected 48 h later using Bio-Plex assay. Values represent arithmetic means, error – standard error of mean (SEM), n = 3; comparison between H1N1-infected and GYY4137-treated samples **p < 0.01; ***p < 0.001 treated versus untreated samples using one-way ANOVA with Tukey post-hoc analysis.

Figure 5. GYY4137 decreases EBOV-induced cytokine and chemokine production.

Confluent monolayers of Vero cells were infected with EBOV and treated with 10 mM GYY4137 as described above. At 4 and 6 day p.i. the supernatant was collected and γ-irradiated. The concentration of cytokines and chemokines was determined using Bio-Plex system. Values represent arithmetic means, error bars – standard error of mean (SEM), n = 3; *p 

Figure 6. GYY4137 inhibits activation of influenza-induced signaling pathways.

Confluent monolayers of A549 cells were infected with H1N1 at MOI of 1 and treated with 10 mM GYY4137 at 1 h p.i. Cells were collected at 48 h p.i. for nuclear protein extraction. Western Blot assays were performed using 25 μg of protein per lane of nuclear fraction. Membranes were probed either with rabbit polyclonal anti-IRF3 (a) or rabbit polyclonal anti-p65 antibody (b). Membranes were stripped and reprobed with anti-human β-actin antibody for loading control. The relative density of the bands for IRF-3 (a) and p65 (b) was calculated by normalizing to β-actin using ImageJ software. The images were cropped to show the bands of interest only. Full length blots are provided in Figs S1 and S2. Brightness and contrast were linearly adjusted for the whole image to enhance visibility.