The small molecules AZD0530 and dasatinib inhibit dengue virus RNA replication via Fyn kinase

Abstract

In this study, we characterized the antiviral mechanism of action of AZD0530 and dasatinib, two pharmacological inhibitors of host kinases, that also inhibit dengue virus (DV) infection. Using Northern blot and reporter replicon assays, we demonstrated that both small molecules inhibit the DV2 infectious cycle at the step of steady-state RNA replication. In order to identify the cellular target of AZD0530 and dasatinib mediating this anti-DV2 activity, we examined the effects of RNA interference (RNAi)-mediated depletion of the major kinases known to be inhibited by these small molecules. We determined that Fyn kinase, a target of both AZD0530 and dasatinib, is involved in DV2 RNA replication and is probably a major mediator of the anti-DV activity of these compounds. Furthermore, serial passaging of DV2 in the presence of dasatinib led to the identification of a mutation in the transmembrane domain 3 of the NS4B protein that overcomes the inhibition of RNA replication by AZD0530, dasatinib, and Fyn RNAi. Although we observed that dasatinib also inhibits DV2 particle assembly and/or secretion, this activity does not appear to be mediated by Src-family kinases. Together, our results suggest that AZD0530 and dasatinib inhibit DV at the step of viral RNA replication and demonstrate a critical role for Fyn kinase in this viral process. The antiviral activity of these compounds in vitro makes them useful pharmacological tools to validate Fyn or other host kinases as anti-DV targets in vivo.

Fig 1

Dual Src-Abl kinase inhibitors AZD0530 and dasatinib inhibit DV2 infection at a postentry step. (A) Chemical structures of AZD0530 and dasatinib. (B) Huh7 cells were mock infected or infected with DV2 at MOI of 1 then treated with a range of AZD0530 and dasatinib concentrations. At 24 h posttreatment, compound cytotoxicity under each condition was measured using a cell viability assay that quantitates the presence of ATP. RLU, Renilla light units. The values were plotted relative to DMSO-treated samples. The values for the concentrations that lead to 90% cytotoxicity (CC₉₀) and 90% inhibition (EC₉₀) were calculated using the nonlinear fit variable slope model (GraphPad Software). (C) The antiviral activity of different concentrations of AZD0530 and dasatinib against DV2 was measured by titration of the infectious particles that had been secreted to culture supernatants at 24 h postinfection/treatment of Huh7 cells. FFU, focus-forming unit. The results were plotted relative to the average of the DMSO-treated samples, 35.10⁴ FFU/ml, which was set at 100. The error bars represent the standard deviation of the assay (titrations were done in duplicate). The potency of each compound was assessed by performing a nonlinear regression fit: the concentration that caused a 90% reduction in virus titer (EC₉₀) is indicated, and the R² value was calculated to show the strength of the regression. (D) Cells were treated with DMSO, 10 μM AZD0530, or 5 μM dasatinib for 24 h before (−24 to 0 h) or after (0 to 24 h) DV2 infection. The results of a representative experiment out of three repeats are shown. Quantification of infectious virus in the supernatants by focus forming assay revealed that inhibition was equivalent when the cells were pretreated with those drugs, which supports the idea that they target a host factor. Asterisks indicate that the differences between experimental samples and DMSO-treated control samples are statistically significant when compared using the unpaired t test (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; not significant [ns], P > 0.05). (E) HEK293T cells were transfected with D2.CprME and the WNV replicon plasmid to produce RVPs. RVPs were collected from the supernatants and used to infect Huh7 cells, followed by treatment with DMSO, 10 μM AZD0530, 5 μM dasatinib, or 1 μM MPA. Intracellular Renilla luciferase activity was quantified at 3, 6, and 12 h posttreatment as a measure of successful RVP entry. The results of a representative experiment out of two repeats are shown. No significant inhibition was observed for the AZD0530-, dasatinib-, and MPA-treated samples at times prior to genome replication. This suggested that these small molecules do not inhibit DV2 entry and genome release. (F) Huh7 cells were infected with DV2 at an MOI of 1, then treated with DMSO, 10 μM AZD0530, or 5 μM dasatinib for 24 h, and then analyzed by immunofluorescence staining for the presence of the DV2 core (C) protein. The results of a representative experiment out of two repeats are shown. The images were taken at ×200 magnification. Comparable numbers of core-positive cells are visible for all treatments, suggesting that AZD0530 and dasatinib do not affect the early steps of DV2 infection.

Fig 2

AZD0530 and dasatinib inhibit DV2 RNA replication. (A and B) Huh7 cells were infected with DV2 at an MOI of 1 and then treated with DMSO, 10 μM AZD0530, or 5 μM dasatinib for 24 h. The results of a representative experiment out of two repeats are shown. (A) The accumulation of DV2 genomic RNA (gRNA) and subgenomic flaviviral RNA (sfRNA) was analyzed in cellular extracts by Northern blotting. Staining of rRNA served as a loading control. The signal intensity corresponding to the gRNA was quantified (AU, arbitrary units) as a measure of steady-state viral RNA abundance, and the averaged values were plotted as a percentage of the average value for the DMSO-treated control samples. Treatment with AZD0530 or dasatinib is associated with decreased steady-state accumulation of DV2 RNAs. (B) The cell lysates were analyzed for the presence of DV2 C, E, NS3, and NS5 proteins, as well as for the presence of GAPDH (glyceraldehyde-3-phosphate dehydrogenase) by Western blotting. Treatment with AZD0530 or dasatinib is associated with decreased steady-state expression of E, C, NS3, and NS5. (C) Huh7 cells were electroporated with an in vitro-transcribed reporter DV2 replicon and treated with DMSO, 10 μM AZD0530, 5 μM dasatinib, or 1 μM MPA at 24 h postelectroporation. The results of a representative experiment out of three repeats are shown. Intracellular firefly luciferase activity (LUC) was quantified at 72 h postelectroporation as a measure of successful RNA replication. All drugs successfully inhibited DV2 RNA replication. Asterisks indicate that the differences between experimental samples and the DMSO-treated control samples are statistically significant when compared using the unpaired t test (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; nonsignificant [ns], P > 0.05). (D) Huh7 cells were electroporated with an in vitro transcribed reporter DV2(GVD) replicon, which is unable to replicate its RNA, and then treated with DMSO, 10 μM AZD0530, 5 μM dasatinib, or 1 μM MPA immediately postelectroporation. The results of a representative experiment out of two repeats are shown. Intracellular firefly luciferase activity (LUC) was quantified at 24 h postelectroporation as a measure of RNA translation. None of the drugs inhibited translation of the input DV2 replicon RNA. The differences between experimental samples and the DMSO-treated control samples are not statistically significant using the unpaired t test (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; nonsignificant [ns], P > 0.05).

Fig 3

Dasatinib but not AZD0530 inhibits DV2 secretion of viral particles. Huh7 cells were first transfected with pcDNA3.1-D2.VLP, then treated with DMSO, 10 μM AZD0530, or 5 μM dasatinib, and then analyzed 24 h posttreatment. The results of a representative experiment out of three repeats are shown. The cell lysates were analyzed by Western blotting for DV2 E and GAPDH as a loading control. The VLPs released to the supernatants were purified and analyzed by Western blotting with an antibody against the DV2 E protein; Coomassie blue staining served as a loading control (Coom). Notably, there was a significant reduction in the release of VLPs from cells treated with dasatinib despite the intracellular accumulation of wild-type levels of E protein.

Fig 4

A mutation in DV2 NS4B reduces the sensitivity of DV2 RNA replication to AZD0530 and dasatinib. (A) DV2 was serially passaged in Huh7 cells in the presence of DMSO or 5 μM dasatinib. After each passage, the virus in the supernatants was titered by FFA, and the progressive emergence of a dasatinib-resistant viral population was observed. (B) Membrane topology of the DV2 NS4B protein (31). The NS4B protein is anchored in the membrane through three transmembrane domains (TMD) and the 2K peptide at its C-terminal end. The sequences for dengue virus serotypes 1, 2, 3, and 4 (DV1, DV2, DV3, and DV4) NS4B TMD3 (residues 101 to 129) were aligned using CLUSTAL W2. Conserved (*) and semiconserved (:.) amino acids are indicated below the alignment. The residue DV2(NS4B-T108) is boxed. (C) Huh7 cells were infected with DV2 or DV2(NS4B-T108I) at MOI of 1 and then treated with DMSO, 10 μM AZD0530, 5 μM dasatinib, or 1 μM MPA. The results of a representative experiment out of three repeats are shown. Quantification of infectious virus in the supernatants at 24 h postinfection was evaluated by FFA. The presence of the NS4B-T108I mutation fully compensated for the AZD0530 or dasatinib treatments but did not confer resistance to MPA inhibition. The statistical significance of the differences between experimental samples and DMSO-treated control samples was calculated for each virus using the unpaired t test (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; nonsignificant [ns], P > 0.05). (D) Huh7 cells were electroporated with in vitro-transcribed reporter DV2 or DV2(NS4B-T108I) replicons. The results of a representative experiment out of three repeats are shown. Intracellular firefly luciferase activity (LUC) was quantified at 6, 24, 48, and 72 h postelectroporation as a measure of steady-state RNA replication. The DV2(NS4B-T108I) replicon achieved higher steady-state replication of its RNA relative to the wild type. The statistical significance of the differences between the wild-type and the DV2(NS4B-T108I) replicons at each time point was calculated using the unpaired t test (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; nonsignificant [ns], P > 0.05). (E) Huh7 cells were electroporated with in vitro-transcribed reporter DV2 or DV2(NS4B-T108I) replicons and then treated with DMSO, 10 μM AZD0530, 5 μM dasatinib, or 1 μM MPA at 24 h postelectroporation. The results of a representative experiment out of three repeats are shown. Intracellular firefly luciferase activity (LUC) was quantified at 72 h postelectroporation as a measure of steady-state RNA replication. DV2(NS4B-T108I) RNA replication was successfully inhibited by MPA but only partially inhibited by AZD0530 and dasatinib. The statistical significance of the differences between experimental samples compared to DMSO-treated control samples was calculated for each replicon by using the unpaired t test (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; nonsignificant [ns], P > 0.05).

Fig 5

DV2 RNA replication is sensitive to RNAi-mediated depletion of Fyn kinase. (A to C) Huh7 cells were transfected with 100 nM nontargeting (NT) siRNA or siRNAs targeting Frk, Fyn, Lyn, Src, or Yes mRNA. At 48 h posttransfection, the cells were infected with DV2 at an MOI of 1 for 24 h. The results of a representative experiment out of two repeats are shown. (A) The expression of each targeted kinase at the time of DV2 infection (t0) or 24 h postinfection (DV2) was analyzed in cell lysates by Western blotting with GAPDH as a loading control. (B) RNAi-mediated depletion of Fyn kinase led to a significant inhibition in the release of infectious virus to the supernatants as quantified by FFA. The statistical significance of the differences between experimental samples and the NT siRNA-treated controls was calculated using the unpaired t test (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; nonsignificant [ns], P > 0.05). (C) The accumulation of DV2 genomic RNA (gRNA) and subgenomic flaviviral RNA (sfRNA) was analyzed in cellular extracts by Northern blotting. Staining of rRNA served as a loading control. Viral RNA accumulation was quantified in each experiment (AU, arbitrary units), and the averaged value was plotted relative to the average value of the DMSO-treated controls. RNAi-mediated depletion of Fyn kinase is associated with decreased steady-state accumulation of DV2 RNAs. (D) Huh7 cells were transfected with 100 nM nontargeting (NT) siRNA or siRNAs targeting Fyn mRNA. At 48 h posttransfection, the cells were electroporated with the in vitro-transcribed reporter DV2 replicon. The results of a representative experiment out of three repeats are shown. Intracellular firefly luciferase activity (LUC) was quantified at 6, 24, 48, and 72 h postelectroporation as a measure of successful RNA replication. Steady-state RNA replication of the DV2 replicon was lower in cells depleted of the Fyn kinase compared to cells treated with the nontargeting siRNA control. Asterisks indicate that the differences to NT siRNA treated samples are statistically significant at each time point when evaluated by using the unpaired t test (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; nonsignificant [ns], P > 0.05).

Fig 6

Fyn kinase RNAi mediates anti-DV2 activity through a pathway inhibited by AZD0530 and dasatinib. (A) Huh7 cells were transfected with 100 nM nontargeting (NT) siRNA or siRNAs targeting Fyn mRNA. At 48 h posttransfection, the cells were infected with DV2 at an MOI of 1 and then treated with DMSO, 10 μM AZD0530, or 5 μM dasatinib for 24 h. The results of a representative experiment out of two repeats are shown. The expression of Fyn kinase at the time of DV2 infection (t0) or at 24 h postinfection (lines with DMSO, AZD0530, or dasatinib treatment) was analyzed in cell lysates by Western blotting with GAPDH as a loading control. The release of infectious virus to the supernatants was quantified by FFA. DV2 inhibition mediated by AZD0530 and dasatinib was not additive to the inhibition mediated by Fyn depletion. (B) Huh7 cells were transfected with 100 nM nontargeting (NT) siRNA or siRNAs targeting Fyn mRNA. At 48 h posttransfection, the cells were infected with DV2 or DV2(NS4B-T108I) at an MOI of 1. The results of a representative experiment out of two repeats are shown. Quantification of infectious virus in the supernatants at 24 h postinfection was evaluated by FFA. The presence of the NS4B-T108I mutation conferred partial resistance to inhibition mediated by Fyn depletion. Asterisks indicate the differences that are statistically significant by the unpaired t test when compared to the control samples treated with the NT siRNA plus DMSO (***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05; nonsignificant [ns], P > 0.05).