Identification of Piperazinylbenzenesulfonamides as New Inhibitors of Claudin-1 Trafficking and Hepatitis C Virus Entry

Abstract

Hepatitis C virus (HCV) infection causes 500,000 deaths annually, in association with end-stage liver diseases. Investigations of the HCV life cycle have widened the knowledge of virology, and here we discovered that two piperazinylbenzenesulfonamides inhibit HCV entry into liver cells. The entry of HCV into host cells is a complex process that is not fully understood but is characterized by multiple spatially and temporally regulated steps involving several known host factors. Through a high-content virus infection screening analysis with a library of 1,120 biologically active chemical compounds, we identified SB258585, an antagonist of serotonin receptor 6 (5-HT6), as a new inhibitor of HCV entry in liver-derived cell lines as well as primary hepatocytes. A functional characterization suggested a role for this compound and the compound SB399885, which share similar structures, as inhibitors of a late HCV entry step, modulating the localization of the coreceptor tight junction protein claudin-1 (CLDN1) in a 5-HT6-independent manner. Both chemical compounds induced an intracellular accumulation of CLDN1, reflecting export impairment. This regulation correlated with the modulation of protein kinase A (PKA) activity. The PKA inhibitor H89 fully reproduced these phenotypes. Furthermore, PKA activation resulted in increased CLDN1 accumulation at the cell surface. Interestingly, an increase of CLDN1 recycling did not correlate with an increased interaction with CD81 or HCV entry. These findings reinforce the hypothesis of a common pathway, shared by several viruses, which involves G-protein-coupled receptor-dependent signaling in late steps of viral entry.IMPORTANCE The HCV entry process is highly complex, and important details of this structured event are poorly understood. By screening a library of biologically active chemical compounds, we identified two piperazinylbenzenesulfonamides as inhibitors of HCV entry. The mechanism of inhibition was not through the previously described activity of these inhibitors as antagonists of serotonin receptor 6 but instead through modulation of PKA activity in a 5-HT6-independent manner, as proven by the lack of 5-HT6 in the liver. We thus highlighted the involvement of the PKA pathway in modulating HCV entry at a postbinding step and in the recycling of the tight junction protein claudin-1 (CLDN1) toward the cell surface. Our work underscores once more the complexity of HCV entry steps and suggests a role for the PKA pathway as a regulator of CLDN1 recycling, with impacts on both cell biology and virology.

Keywords: hepatitis C virus; piperazinylbenzenesulfonamide; protein kinase A; tight junction protein; virus entry; virus-host interaction.

FIG 1

Identification of the 5-HT6 antagonist SB258585 as a new inhibitor of the HCV life cycle. (A to C) An HCS screen with a library of 1,120 chemical compounds was performed at three different concentrations (1 μM, 10 μM, and 20 μM). Dot plots show the distributions of the populations on the bases of cell number and percentage of HCVcc JFH1 (genotype 2a)-infected cells at the indicated concentrations. Each dot represents one compound. Green lines represent the average values corresponding to DMSO control wells. Red lines represent the selected cutoffs determining the positive hits (light red squares). (D) Summary of the positive hits selected according to the cutoffs for each concentration. The red circle indicates compounds confirmed as positive hits for at least two concentrations. (E) Graph showing percentages of HCVcc infection corresponding to the wells treated with chemical compounds targeting 5-HT6 in a more or less specific way, normalized to the mean for the DMSO control wells at each concentration. The green line corresponds to the compound SB258585 hydrochloride. (F) Wells corresponding to DMSO and SB258585 for each concentration of the screen. Nuclei are shown in blue, and HCV E1 staining is shown in green.

FIG 2

SB258585 inhibits HCV infection, modulating PKA in a 5-HT6-independent manner. (A) Huh-7 cells were treated with SB258585 at different concentrations for 2 h prior to, during, or after HCVcc JFH1 incubation, following the schematized kinetics. Infection at 30 h postinfection was quantified by immunofluorescence assay. (B) Huh-7 cells were treated with the drug for 2 h, followed by 28 h of rest. An MTS assay was performed in order to evaluate cell toxicity. (C) Quantification of 5-HT6 mRNA levels in 17 tissues from human biopsy specimens by qRT-PCR. a.u., arbitrary units. (D) Huh-7 cells were treated for 2 h with DMSO, H89 (10 μM), SB258585 (100 μM), or SB399885 (100 μM). A representative Western blot (n = 3) and relative quantification of the total phosphorylation of PKA substrates normalized to the loading control (β-tubulin) are presented. Results are presented as means ± SEM (n = 3) in panels A, B, and D. One-way (B and D) or two-way (A) analysis of variance (ANOVA) followed by the Dunnett or Bonferroni posttest was performed for statistical analysis. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.001.

FIG 3

Piperazinylbenzenesulfonamides inhibit a late step of HCV entry for all the major HCV genotypes. (A) Huh-7 cells were treated for 3 h with SB258585 at different concentrations, concomitant with HCVpp JFH1 or RD114pp incubation. At 48 h postinfection, cells were lysed and luciferase activity measured in order to quantify the infection. (B) Huh-7 cells were incubated for 2 h with SB258585 at different concentrations, along with a GFP-expressing adenovirus. The percentage of infection was determined by quantifying GFP-positive cells at 24 h postinfection. (C) Primary human hepatocytes (PHHs) were treated for 2 h with SB258585 at different concentrations, in addition to infection with HCVcc JFH1. Infection was quantified by qRT-PCR at 24 h postinfection. (D) Huh-7.5 cells were treated for 3 h with SB258585, in addition to infection with JFH1-based HCV genotype 1 to 6 recombinant viruses with strain-specific core-NS2. Infection was quantified by immunofluorescence assay at 48 h postinoculation. Error bars represent standard deviations (SD). (E) After 1 h of viral attachment at 4°C, Huh-7 cells were shifted to 37°C and treated with SB258585 at the indicated concentrations, following 15-min kinetics for 2 h. Proteinase K (50 μg/ml) and bafilomycin A (25 nM) were used as controls for early and late entry steps, respectively. Infection was quantified at 30 h postinfection by immunofluorescence assay, and the value for each time point was normalized to that for the corresponding DMSO condition. (F) Chemical structures of SB258585 and SB399885. (G) Huh-7 cells were treated with SB399885 at different concentrations for 2 h prior to, during, or after HCVcc JFH1 incubation, following the schematized kinetics. Infection was quantified at 30 h postinfection by immunofluorescence assay. (H) Huh-7 cells were treated with SB399885 for 2 h, followed by 28 h of rest. An MTS assay was performed in order to evaluate cell toxicity. (I) After 1 h of viral attachment at 4°C, Huh-7 cells were shifted to 37°C and treated with SB399885 at the indicated concentrations, following 15-min kinetics for 2 h. Proteinase K (50 μg/ml) and bafilomycin A (25 nM) were used as controls for early and late entry steps, respectively. Infection was quantified at 30 h postinfection by immunofluorescence assay, and the value for each time point was normalized to that for the corresponding DMSO condition. Results are presented as means ± SEM (n = 3 [A to D, G, and H] and n = 2 [E and I]). PHH results are presented as means for triplicates ± SEM for two independent experiments. One-way (B, C, and H) or two-way (A and G) ANOVA followed by the Dunnett or Bonferroni posttest was performed for statistical analysis. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.001; ns, nonsignificant.

FIG 4

SB258585 alters CLDN1 recycling, causing its intracellular accumulation. (A) Huh-7 cells were treated for 2 h with DMSO or increasing concentrations of SB258585. Cell surface expression of CD81 and CLDN1 was analyzed by immunofluorescence assay. Images were taken using a Zeiss LSM-880 microscope and a 63× objective. (B) Pearson correlation coefficients (PCCs) were calculated for cell surface ROIs for at least 40 different cells for each condition. (C) Huh-7 cells were treated for 2 h with DMSO or increasing concentrations of SB258585, and CLDN1 expression was analyzed by flow cytometry. Curves from a representative experiment are shown. Mean fluorescence intensities (MFI) relative to that for the DMSO-treated condition are also presented. (D) Huh-7 cells were treated for 2 h with DMSO or increasing concentrations of SB399885, and CLDN1 expression was analyzed by flow cytometry. (E) Huh-7 cells were incubated with SB258585 (100 μM) for the indicated periods. CLDN1 present at the cell surface was quantified by flow cytometry. (F) Huh-7 cells were treated for 2 h with SB258585 (100 μM). The drug was then removed and replaced by DMEM for the indicated times. Cytometry analyses were performed to quantify CLDN1 at the cell surface. For panels D to F, mean fluorescence intensities relative to those for the DMSO-treated condition are shown. (G) Huh-7 cells were treated for 2 h with DMSO, SB258585 (100 μM), SB399885 (100 μM), or H89 (10 μM). The total quantity of CLDN1 was assessed by Western blotting. β-Tubulin was used as a loading control. (H) Huh-7 cells were treated for 2 h with DMSO, SB258585 (100 μM), or H89 (10 μM). CLDN1 subcellular localization was determined by immunofluorescence assay after membrane permeabilization. Images were taken with a 63× objective. (I) TGN46 was stained concomitantly with CLDN1, and PCCs were calculated for intracellular CLDN1-TGN46 colocalization for >35 cells for each condition. (J) After surface biotinylation, Huh-7 cells were incubated at 37°C with DMSO or SB258585 (100 μM) for the indicated times. Biotin remaining at the cell surface was cleaved by use of glutathione. The amount of internalized CLDN1 was determined by Western blotting after pulldown of biotin-labeled proteins with streptavidin-agarose beads. A representative Western blot (n = 3) is presented. All results are presented as means ± SEM (n = 3). One-way ANOVA (B to E and I) or two-way ANOVA (F) followed by the Dunnett or Bonferroni posttest was performed for statistical analysis. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.001. Bars = 30 μm.

FIG 5

SB258585 does not alter surface localization of the other main HCV entry factors. (A) Huh-7 cells were treated for 2 h with DMSO or increasing concentrations of SB258585. Surface biotinylation followed by biotin immunoprecipitation was performed, and a representative Western blot for OCLN is shown. The quantity of biotinylated OCLN was determined from Western blots by use of Fiji (n = 3). (B to E) Huh-7 cells were treated for 2 h with DMSO or increasing concentrations of SB258585. CD81 (B), EGFR (C), SRB1 (D), and E-cadherin (E) expression levels were analyzed by flow cytometry. Curves from a representative experiment are shown. Mean fluorescence intensities relative to those for the DMSO-treated condition are presented. All results are presented as means ± SEM (n = 3). One-way ANOVA followed by the Dunnett posttest was performed for statistical analysis. ns, nonsignificant.

FIG 6

Inhibition of the PKA signaling pathway downregulates HCV entry and CLDN1 cell surface localization. (A) Huh-7 cells were treated for 2 h with H89 at different concentrations. An MTS assay was performed at 30 h posttreatment in order to evaluate the cell toxicity of the compound. (B) Huh-7 cells were treated for 3 h with H89 at different concentrations, along with infection by HCVpp JFH1 and RD114pp. At 48 h postinfection, cells were lysed and luciferase activity measured in order to quantify the infection. (C) Huh-7 cells were treated with H89 at different concentrations, in accordance with the schematized kinetics, along with HCVcc JFH1 infection. Infection was quantified at 30 h postinfection by immunofluorescence assay. (D) After 1 h of viral attachment at 4°C, Huh-7 cells were shifted to 37°C and treated with H89 (10 μM), following 15-min kinetics for 2 h. Proteinase K (50 μg/ml) and bafilomycin A (25 nM) were used as controls for early and late entry steps, respectively. Infection was quantified at 30 h postinfection by immunofluorescence assay, and the value for each time point was normalized to that for the corresponding DMSO condition. (E) Huh-7 cells were treated for 2 h with DMSO or H89 (10 μM). Cell surface CLDN1 was analyzed by flow cytometry. Curves from a representative experiment and mean fluorescence intensities relative to that for the DMSO-treated condition are shown. Results are presented as means ± SEM (n = 3 [A to C and E] and n = 2 [D]). Two-tailed Student's t test (E) or one-way (A) or two-way (B and C) ANOVA followed by the Dunnett or Bonferroni posttest was performed for statistical analysis. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.001.

FIG 7

PKA activation increases CLDN1 localization at the cell surface, without increasing HCV entry or CLDN1-CD81 colocalization. (A to D) Huh-7 cells were transfected with pcDNA or pcDNA-PRKACA for 48 h. Cells were treated for 2 h with DMSO or forskolin (20 μM), as indicated. (A) A representative Western blot (n = 3) is presented in order to show the phosphorylation of PKA substrates. β-Tubulin was used as a loading control, and PKA transfection was checked by HA immunoblotting. (B) Cytometry analysis was performed to quantify CLDN1 cell surface localization and PKA-HA transfection. Mean fluorescence intensities relative to that for the DMSO-treated condition are shown (n = 3). (C) CLDN1 and CD81 were analyzed by immunofluorescence assay, and PCCs were calculated after confocal image acquisition for cell surface ROIs for at least 40 different cells for each condition. (D) Transfected cells were treated for 2 h with DMSO or forskolin (20 μM), together with HCVcc infection. Infection was quantified at 30 h postinfection by immunofluorescence assay. (E to I) Huh-7 cells were transfected with pcDNA or pcDNA-5-HT6 for 48 h. (E) A representative Western blot (n = 2) and relative quantification of the total phosphorylation of PKA substrates normalized to the loading control (β-tubulin) are presented. HA–5-HT6 transfection was checked by HA immunoblotting. (F and G) Cell surface CLDN1 and HA–5-HT6 were immunolabeled and quantified by flow cytometry. Mean fluorescence intensities relative to that for the DMSO-treated condition are shown. For panel G, transfected Huh-7 cells were treated for 2 h with SB258585 (100 μM) before labeling for flow cytometry analyses. (H) CLDN1 and CD81 immunostaining followed by confocal microscopy allowed for the calculation of PCCs for cell surface ROIs for at least 40 different cells for each condition. (I) Transfected Huh-7 cells were infected with HCVcc for 2 h. Infection was quantified at 30 h postinfection by immunofluorescence assay. (J to L) Huh-7 cells were transfected with pcDNA or pcDNA-CLDN1 for 48 h. Cell surface CLDN1 was immunolabeled and quantified by flow cytometry. (J) Mean fluorescence intensities relative to that for the DMSO-treated condition are shown. (K) CLDN1 and CD81 immunostaining followed by confocal microscopy allowed for the calculation of PCCs for cell surface ROIs for at least 40 different cells for each condition. (L) Transfected cells were infected with HCVcc at 48 h posttransfection. Infection was quantified at 30 h postinfection by immunofluorescence assay. Results are presented as means ± SEM (n = 3 [C, D, H, and I] or n = 4 [B, F, G, and J to L] independent experiments). Two-tailed Student's t test (E, F, and H to L) or two-way ANOVA (B, D, and G) followed by the Bonferroni posttest was performed for statistical analysis. *, P < 0.05; ***, P < 0.001; ns, nonsignificant.

FIG 8

An increase of CLDN1 localization at the cell surface is not sufficient to increase HCV entry upon either CD81 overexpression or EGF stimulation. (A) Huh-7 cells were transfected with pcDNA, HA–5-HT6, or CD81-YFP alone or in combination, as indicated. At 48 h posttransfection, cells were infected with HCVcc JFH1. At 30 h postinfection, cells were fixed and the infection rate was determined by IFA. (B) After 2 h of starvation, Huh-7 cells were kept nonstimulated or stimulated for 1 h with EGF (1 μg/ml). Western blotting was performed to verify the activation of EGFR through phosphorylation. β-Tubulin was used as a loading control. (C) Huh-7 cells transfected for 48 h with pcDNA, HA–5-HT6, or CD81-YFP, alone or in combination, after 2 h of starvation were treated for 1 h with EGF (1 μg/ml) and infected for 2 h with HCVcc JFH1 concomitantly with EGF (1 μg/ml). At 30 h postinfection, cells were fixed and the infection rate was determined by IFA. Results are presented as means ± SEM (n = 2 [A and C]). One-way (A) or two-way (C) ANOVA followed by the Dunnett or Bonferroni posttest was performed for statistical analysis. ns, nonsignificant.

FIG 9

Deletion of the CLDN1 C-terminal region does not alter HCV sensitivity to SB258585 and H89. Flow cytometry (A) and Western blotting (B) were performed on CLDN1 CRISPR/Cas9 and control pX459 cells in order to evaluate CLDN1 silencing. OCLN was used as a loading control for the Western blot. (C and D) CLDN1 CRISPR/Cas9 cells were transfected with pcDNA 3.1, wild-type CLDN1 (CLDN1 WT), or CLDN1 ΔCter. At 48 h posttransfection, cells were labeled for CLDN1 at the cell surface and analyzed by flow cytometry (C) or infected with HCVcc JFH1 (D). At 30 h postinfection, cells were fixed and the infection rate was determined by IFA. Results are presented as means ± SEM (n = 3). Two-way ANOVA (D) followed by the Bonferroni posttest was performed for statistical analysis. ****, P < 0.001.