Feasibility and biological rationale of repurposing sunitinib and erlotinib for dengue treatment

Abstract

There is an urgent need for strategies to combat dengue virus (DENV) infection; a major global threat. We reported that the cellular kinases AAK1 and GAK regulate intracellular trafficking of multiple viruses and that sunitinib and erlotinib, approved anticancer drugs with potent activity against these kinases, protect DENV-infected mice from mortality. Nevertheless, further characterization of the therapeutic potential and underlying mechanism of this approach is required prior to clinical evaluation. Here, we demonstrate that sunitinib/erlotinib combination achieves sustained suppression of systemic infection at approved dose in DENV-infected IFN-α/β and IFN-γ receptor-deficient mice. Nevertheless, treatment with these blood-brain barrier impermeable drugs delays, yet does not prevent, late-onset paralysis; a common manifestation in this immunodeficient mouse model but not in humans. Sunitinib and erlotinib treatment also demonstrates efficacy in human primary monocyte-derived dendritic cells. Additionally, DENV infection induces expression of AAK1 transcripts, but not GAK, via single-cell transcriptomics, and these kinases are important molecular targets underlying the anti-DENV effect of sunitinib and erlotinib. Lastly, sunitinib/erlotinib combination alters inflammatory cytokine responses in DENV-infected mice. These findings support feasibility of repurposing sunitinib/erlotinib combination as a host-targeted antiviral approach and contribute to understanding its mechanism of antiviral action.

Keywords: Antivirals; Dengue virus; Drug repurposing; Kinase inhibitors; Virus-host interactions.

Figure 1.. Sunitinib/erlotinib combination protects mice from severe dengue-associated mortality but not from paralysis in a high inoculum model.

(A and B) Weight loss (A) and mortality (B) of DENV-infected AG-B6 mice treated once daily for 5 days with sunitinib/erlotinib (SM/E) combination (30 mg/kg each) or vehicle (n=8 per treatment group). Administration was intraperitoneal and started at inoculation. In A, shown are means±SD. Dotted line in B represents first onset of paralysis. Mice were euthanized soon after the onset of paralysis. C. DENV titer in spleen of AG-B6 mice measured by plaque assay at various time points following once-daily administration of SM/E combination or vehicle for 5 days (n=4 per treatment group). Shown are means±SEM. D. DENV titer in brain, spleen, liver and serum measured by standard plaque assays in AG-B6 mice treated once daily for 5 days with SM/E or vehicle on day 11 postinfection (when the first mouse showed early signs of paralysis, n=6 per treatment group). Representative experiments out of two conducted are shown. Dotted lines represent limit of detection (100 PFU (C) and 10 PFU (D)). *P < 0.05, ***P < 0.001 relative to vehicle control by log-rank test (B) and two-way ANOVA with Bonferroni multiple comparisons test (C). ND, not detected; NA, not applicable.

Figure 2.. Sunitinib/erlotinib combination delays paralysis onset in a low-inoculum dengue model.

(A and B) Mortality (A) and cumulative percentage of paralysis (B) in AG-B6 mice inoculated with a low inoculum of DENV (3×10³ PFU) and treated once daily for 5 days with sunitinib/erlotinib (SM/E) combination (30 mg/kg each) or vehicle (n = 10 per treatment group, per experiment). Administration was intraperitoneal and started at inoculation. Data pooled from two independent experiments is shown. Dotted line represents first onset of paralysis. Mice were euthanized soon after the onset of paralysis. C. DENV load in the brain of AG-B6 mice on day 8–12 postinoculation with a low DENV titer and once daily 5-day treatment with SM/E combination (n = 13) or vehicle (n = 5). *P < 0.05 relative to vehicle control by log-rank test (A, B) and nonparametric Mann-Whitney-Wilcoxon test (C). GE, genome equivalent; GAPDH, Glyceraldehyde 3-phosphate dehydrogenase.

Figure 3.. Lower dose combinations of sunitinib/erlotinib are effective in vivo.

Weight loss (A) and mortality (B) of DENV-infected AG-B6 mice treated once daily for 5 days with sunitinib/erlotinib (SM/E) combination at the indicated doses (mg/kg) or vehicle (n=8 per treatment group, per experiment). Administration was by gavage and started at inoculation. Data pooled from three independent experiments is shown. In A, shown are means±SD. P by log-rank test are relative to vehicle control or between the various tested groups as indicated (B). N.S, non-significant.

Figure 4.. Ex vivo antiviral activity of sunitinib and erlotinib in human primary dendritic cells.

Cell viability (blue) and dose response of DENV infection (black) to sunitinib (left) and erlotinib (right) measured by alamarBlue assays and luciferase assays, respectively, 72 hours after infection of primary human MDDCs. Data are plotted relative to vehicle control. Shown is a representative experiment with cells from a single donor, out of 3 independent experiments conducted with cells derived from 3 donors, each with 5 biological replicates; shown are means ± SD.

Figure 5.. AAK1, but not GAK, gene expression correlates with intracellular DENV abundance.

AAK1 and GAK transcript levels in bulk (A) and single cell (B-C) DENV infected (MOI=0 (control), 1, 10) Huh7 samples at various time points postinfection, measured via qRT-PCR (A) and virus-inclusive single-cell RNA-Seq (B-C). *P < 0.05, **P < 0.01 in A are relative to vehicle control normalized to TBP (TATA-binding protein) by two-way ANOVA with Tukey’s multiple comparisons test. B. Scatter plots depicting the correlation between vRNA content (viral reads, X axis) and AAK1/GAK gene expression (Y axis) during DENV infection. Each dot represents a single cell, color-coded based on the MOI. ρ represents Spearman correlation coefficients between abundances of dengue vRNA and of AAK1/GAK mRNA from the same cell. C. Correlation values measured for AAK1 (green) and GAK (orange) at various time points postinfection.

Figure 6.. AAK1 and GAK are molecular targets underlying the antiviral effect of sunitinib and erlotinib.

A. Expression of the indicated proteins in T-REx 293 cells induced with doxycycline to express AAK1 or GAK or uninduced control cells. B-C. DENV infection normalized to cellular viability in these cells via luciferase and alamarBlue assays, respectively, 72 hours following infection with a luciferase reporter DENV and treatment with sunitinib (B) or erlotinib (C). Shown is a representative experiment out of two conducted, each with 5 replicates. *P<0.05, **P<0.01 relative to induced or uninduced cells treated with DMSO by one-way ANOVA with Fisher’s least significant difference test. Means ± SD are shown.

Figure 7.. Sunitinib/erlotinib combination alters inflammatory cytokines levels in the serum of dengue infected mice.

Cytokines levels measured by Luminex assay in AG-B6 mice 48 or 72 hours postinfection with DENV or in uninfected mice following once-daily administration of vehicle or sunitinib/erlotinib (SM/E) combination. Shown are box plots of the concentrations of the indicated cytokines. The box horizontal lines indicate the first, second (median) and third quartiles. Outliers above the upper whisker (Q3+1.5xIQR) or below the lower whisker (Q1–1.5xIQR) are labeled with black lines. n = 8 per infected treatment group and 2–3 per uninfected group treated with vehicle or sunitinib/erlotinib combination, respectively. *P < 0.05 and ***P < 0.001 relative to untreated infected mice by 2-way ANOVA with Tukey’s multiple comparisons test. IQR, interquartile range; NS, Non-Significant.