Combinations of approved oral nucleoside analogues confer potent suppression of alphaviruses in vitro and in vivo

Abstract

Background: Alphaviruses, including chikungunya virus (CHIKV), pose a significant global health threat, yet specific antiviral therapies remain unavailable.

Methods: We evaluated combinations of three oral directly acting antiviral drugs (sofosbuvir (SOF), molnupiravir (MPV), and favipiravir (FAV)), which are approved for other indications, against CHIKV, Semliki Forest virus (SFV), Sindbis virus (SINV), and Venezuelan Equine Encephalitis virus (VEEV) in vitro and in vivo. We assessed antiviral efficacy in human skin fibroblasts and liver cells, as well as in a mouse model of CHIKV-induced arthritis.

Findings: In human skin fibroblasts, synergistic antiviral effects were observed for combinations of MPV + SOF and FAV + SOF against CHIKV, and for FAV + SOF against SFV. In human liver cells, FAV + MPV conferred additive to synergistic activity against VEEV and SINV, while SOF synergized with FAV against SINV. In mice, MPV improved CHIKV-induced foot swelling and reduced systemic infectious virus titres. Combination treatment with MPV and SOF significantly reduced swelling and infectious virus titres compared to monotherapies of each drug. Sequencing of CHIKV RNA from joint tissue revealed that MPV caused dose-dependent increases in mutations in the CHIKV genome. Upon combination therapy of MPV with SOF, the number of mutations was significantly lower compared to monotherapy with several higher doses of MPV.

Interpretation: Combining these approved oral nucleoside analogues confers potent suppression of multiple alphaviruses in vitro and in vivo with enhanced control of viral genetic evolution in face of antiviral pressure. These drug combinations may ultimately lead to the development of potent combinations of pan-family alphavirus inhibitors.

Funding: This work was supported by a PhD fellowship granted to S.V. by the Research Foundation - Flanders (FWO) (11D5923N). L.D.C. was also supported by Research Foundation - Flanders (FWO) PhD fellowship (11L1325N). Dr. Polyak and Schiffer are partially supported by R01AI121129.

Keywords: Chikungunya virus (CHIKV); Mosquito-borne viruses; alphaviruses; antiviral drugs; drug combination therapy; nucleoside analogues.

Figure 1.. Antiviral efficacy of single compounds against alphaviruses in human cells.

Dose-response activity of (a,d) EIDD-1931 (MPV), (b,e) SOF and (c,f) FAV against cytopathic effect induced by CHIKV, SFV, VEEV, RRV or SINV (AR86 and HRsp strain), quantified in (a-c) skin fibroblasts or (df) Huh7 cells by the MTS or ATP method at 72 hours post infection. Data represent percentage of virus inhibition (left y-axis), or cell viability (right y-axis) compared to virus or cell control samples, respectively, and is shown as mean values ± standard deviation from at least three independent experiments.

Figure 2.. Antiviral efficacy of combination treatment against CHIKV in human skin fibroblasts.

Antiviral efficacy of combinations of EIDD-1931 (MPV) and FAV, EIDD-1931 (MPV) and SOF or FAV and SOF against CHIKV in human skin fibroblasts. Inhibition of virus-induced CPE was quantified by the MTS method at 72 hours post infection. Data were analysed and visualized using SynergyFinder based on the Bliss independence model. (a) Overall Bliss Synergy score averaged over all dose combination measurements of the matrix, representing the percentage excess response compared to the expected responses. Bliss Synergy scores lower than −10, between −10 and 10 or larger than 10 indicate combinations which are likely to be antagonistic, additive or synergistic, respectively. (b) Maximum Bliss synergy score representing the most synergistic 3-by-3 window in the full dose-response matrix. (a-b) Data represent mean scores and standard deviations from multiple independent experiments for each drug combination. (c-e) Top plots present dose-response matrices showing percentage virus inhibition for each combination. MSA is indicated within white squares. Bottom plots show three-dimensional maps highlighting the areas of synergy (red) and antagonism (green) across the full dose response matrix for each combination. Data in panels a and b indicate the mean of 5-7 independent experiments (with each experiment denoted by a dot). Data in panels c-e comprise representative SF3·0 plots from single experiments. MSA, most synergistic area.

Figure 3.. Antiviral efficacy of combination treatment against SFV in human skin fibroblasts.

Antiviral efficacy of combinations of EIDD-1931 (MPV) and FAV, EIDD-1931 (MPV) and SOF or FAV and SOF against SFV in human skin fibroblasts. Inhibition of virus-induced CPE was quantified by the MTS method at 72 hours post infection. Data were analysed and visualized using SynergyFinder based on the Bliss independence model. (a) Overall Bliss Synergy score averaged over all dose combination measurements of the matrix, representing the percentage excess response compared to the expected responses. Bliss Synergy scores lower than −10, between −10 and 10 or larger than 10 indicate combinations which are likely to be antagonistic, additive or synergistic, respectively. (b) Maximum Bliss synergy score representing the most synergistic 3-by-3 window in the full dose-response matrix. (a-b) Data represent mean scores and standard deviations from multiple independent experiments for each drug combination. (c-e) Top plots present dose-response matrices showing percentage virus inhibition for each combination. MSA is indicated within white squares. Bottom plots show three-dimensional maps highlighting the areas of synergy (red) and antagonism (green) across the full dose response matrix for each combination. Data in panels a and b indicate the mean of 3-5 independent experiments (with each experiment denoted by a dot). Data in panels c-e comprise representative SF3·0 plots from single experiments. MSA, most synergistic area.

Figure 4.. Antiviral efficacy of combination treatment against SINV and VEEV in Huh7 cells.

Antiviral efficacy of combinations of EIDD-1931 (MPV) and FAV, EIDD-1931 (MPV) and SOF or FAV and SOF against (a) VEEV TC83, (b) SINV AR86, (c) SINV HRsp in Huh7 cells. Inhibition of virus-induced CPE was quantified by the ATP method at 72 hours post infection. Data were analysed and visualized using synergyfinder 3·0 based on the bliss independence model. left panels: overall Bliss Synergy score averaged over all dose combination measurements of the matrix, representing the percentage excess response compared to the expected responses. Bliss Synergy scores lower than −10, between −10 and 10 or larger than 10 indicate combinations which are likely to be antagonistic, additive or synergistic, respectively. Right panels: maximum Bliss synergy score representing the most synergistic 3-by-3 window in the full dose-response matrix (MSA). Data in panels a-c indicate the mean scores and standard deviations of 2-4 independent experiments (with each experiment denoted by a dot). MSA; most synergistic area.

Figure 5.. Dose-dependent efficacy of MPV against CHIKV infections AG129 mice.

AG129 mice (n=5 per group) were treated orally with different single doses of MPV (10, 50, 100 mg/kg) and infected subcutaneously with CHIKV (100 PFU) in the left hind foot. (a) Median and standard deviation of percentage swelling of the infected foot relative to the contralateral foot, as measured daily using a digital calliper. (b-d) Infectious virus titres in tissues, quantified by end-point titrations on Vero cells. (e-g) Viral RNA levels in tissues, quantified by qRT-PCR. (b,e) Virus loads in blood collected by submandibular (day 2 pi) or cardiac (day 3 pi) puncture. (c,f) Virus loads in the left, infected ankle joint on day 3 pi. (d,g) Virus loads in the right, contralateral ankle joint on day 3 pi. Individual data points are shown, with solid lines representing median values. Statistical significance was assessed with (a,b) two-way repeated measures ANOVA with Tukey’s correction or (c-g) Kruskal-Wallis test with Dunn’s multiple comparisons test (*, p<0·05; **, p<0·01; ***, p<0·005; ****, p<0·0001). (b-d) Dotted lines represent the LOQ; (e-g) dotted lines represent the LOD. Gc, genome copies; TCID50, tissue culture infectious dose 50; pi, post infection; LOQ, limit of quantification; LOD, limit of detection.

Figure 6.. Dose-dependent efficacy of SOF against CHIKV infections in AG129 mice.

AG129 mice (n=5 per group) were treated orally with different single doses of SOF (40, 80 mg/kg) and infected subcutaneously with CHIKV (100 PFU) in the left hind foot. (a) Median and standard deviation of percentage swelling of the infected foot relative to the contralateral foot, as measured daily using a digital calliper. (b-d) Infectious virus titres in tissues, quantified by end-point titrations on Vero cells. (e-g) Viral RNA levels in tissues, quantified by qRT-PCR. (b,e) Virus loads in blood collected by submandibular (day 2 pi) or cardiac (day 3 pi) puncture. (c,f) Virus loads in the left, infected ankle joint on day 3 pi. (d,g) Virus loads in the right, contralateral ankle joint on day 3 pi. Individual data points are shown, with solid lines representing median values. Statistical significance was assessed with (a,b) two-way repeated measures ANOVA with Tukey’s correction or (c-g) Kruskal-Wallis test with Dunn’s multiple comparisons test (ns, p>0·05). (b-d) Dotted lines represent the LOQ; (e-g) dotted lines represent the LOD. Gc, genome copies; TCID50, tissue culture infectious dose 50; pi, post infection; LOQ, limit of quantification; LOD, limit of detection.

Figure 7.. Combined efficacy of MPV with SOF against CHIKV infections in AG129 mice.

AG129 mice (n=5 per group) were treated orally with either single doses of MPV (10 mg/kg), SOF (80 mg/kg) or a combination of MPV and SOF (10 + 80 mg/kg) and infected subcutaneously with CHIKV (100 PFU) in the left hind foot. (a) Median and standard deviation of percentage swelling of the infected foot relative to the contralateral foot, as measured daily by using a digital calliper. (b-d) Infectious virus titres in tissues, quantified by end-point titrations on Vero cells. (e-f) Viral RNA levels in tissues, quantified by qRT-PCR. (b) Virus loads in blood collected by submandibular (day 2 pi) or cardiac (day 3 pi) puncture. (c,e) Virus loads in the left, infected ankle joint on day 3 pi. (d,f) Virus loads in the right, contralateral ankle joint on day 3 pi. Individual data points are shown, with solid lines representing median values. Statistical significance was assessed with (a,b) two-way repeated measures ANOVA with Tukey’s correction or (c-f) Kruskal-Wallis test with Dunn’s multiple comparisons test (*, p<0·05; **, p<0·01; ***, p<0·005). (b-d) Dotted lines represent the LOQ; (e-f) dotted lines represent the LOD. Gc, genome copies; TCID50, tissue culture infectious dose 50; pi, post infection; LOQ, limit of quantification; LOD, limit of detection.

Figure 8.. CHIKV mutation profiles of MPV- and SOF-treated mice.

Sequencing was performed on viral RNA isolated from contralateral ankle joints of CHIKV-infected AG129 mice (n=5 per condition) that were treated with vehicle, MPV (10, 50, 100 mg/kg), SOF (80 mg/kg) or a combination of MPV (10 mg/kg) and SOF (80 mg/kg). (a) Mutation profile of viral RNA across treatment groups. The colour intensity represents the mutation count per genome for each specific mutation type. (b) Total mutation count in viral RNA normalized to the total number of reads per sample. Data is represented as boxplots, showing individual data points with solid lines indicating the median values. Statistical significance was assessed with a one-sided Wilcoxon test with Benjamini-Hochberg correction. *, p<0·05. (c) Bar plot showing the fold change in the mean count of each mutation type for each treatment group (n=5) relative to the control group (vehicle treatment). The dotted line indicates no change in mutation count compared to the vehicle-treated control group.