Targeting Innate Immunity for Antiviral Therapy through Small Molecule Agonists of the RLR Pathway

ABSTRACT

The cellular response to virus infection is initiated when pathogen recognition receptors (PRR) engage viral pathogen-associated molecular patterns (PAMPs). This process results in induction of downstream signaling pathways that activate the transcription factor interferon regulatory factor 3 (IRF3). IRF3 plays a critical role in antiviral immunity to drive the expression of innate immune response genes, including those encoding antiviral factors, type 1 interferon, and immune modulatory cytokines, that act in concert to restrict virus replication. Thus, small molecule agonists that can promote IRF3 activation and induce innate immune gene expression could serve as antivirals to induce tissue-wide innate immunity for effective control of virus infection. We identified small molecule compounds that activate IRF3 to differentially induce discrete subsets of antiviral genes. We tested a lead compound and derivatives for the ability to suppress infections caused by a broad range of RNA viruses. Compound administration significantly decreased the viral RNA load in cultured cells that were infected with viruses of the family Flaviviridae, including West Nile virus, dengue virus, and hepatitis C virus, as well as viruses of the families Filoviridae (Ebola virus), Orthomyxoviridae (influenza A virus), Arenaviridae (Lassa virus), and Paramyxoviridae (respiratory syncytial virus, Nipah virus) to suppress infectious virus production. Knockdown studies mapped this response to the RIG-I-like receptor pathway. This work identifies a novel class of host-directed immune modulatory molecules that activate IRF3 to promote host antiviral responses to broadly suppress infections caused by RNA viruses of distinct genera.

IMPORTANCE Incidences of emerging and reemerging RNA viruses highlight a desperate need for broad-spectrum antiviral agents that can effectively control infections caused by viruses of distinct genera. We identified small molecule compounds that can selectively activate IRF3 for the purpose of identifying drug-like molecules that can be developed for the treatment of viral infections. Here, we report the discovery of a hydroxyquinoline family of small molecules that can activate IRF3 to promote cellular antiviral responses. These molecules can prophylactically or therapeutically control infection in cell culture by pathogenic RNA viruses, including West Nile virus, dengue virus, hepatitis C virus, influenza A virus, respiratory syncytial virus, Nipah virus, Lassa virus, and Ebola virus. Our study thus identifies a class of small molecules with a novel mechanism to enhance host immune responses for antiviral activity against a variety of RNA viruses that pose a significant health care burden and/or that are known to cause infections with high case fatality rates.

INTRODUCTION

RNA viruses pose a significant public health problem worldwide and are a frequent cause of emerging and reemerging viral infections. There has been an increased incidence of disease caused by arthropod-borne members of the Flaviviridae in recent decades. West Nile virus (WNV) infections were on the decline from 2008 to 2011, while 2012 saw a sudden increase in the incidence of WNV infection that resulted in death in about 8.8% of cases, and an unprecedented 50.8% of the reported cases involved neuroinvasive disease (1). The World Health Organization reported an incidence of 50 million to 100 million new cases of dengue virus (DV) infection yearly that included 500,000 cases of dengue hemorrhagic fever and 22,000 deaths, mostly among children (2, 3). With about 40% of the world's population being at risk of DV infection and with the increased incidence of morbidity and mortality from both WNV and DV infections, these pathogens are emerging viruses of public health concern that call for effective therapy. Another member of the Flaviviridae, hepatitis C virus (HCV), is transmitted through the blood. Given the 3 million to 4 million new cases of HCV infection each year and given that about 150 million people are chronically infected with HCV and at risk for developing liver cirrhosis or liver cancer, HCV is a major virus of global public health concern (4). With the development of recent therapies, including direct-acting antiviral (DAA) drugs, control of HCV appears to be an achievable goal in the next decade, but drug resistance and the need to treat disparate HCV genotypes remain concerns with the prolonged use of these DAAs (5). Additionally, the exorbitant cost of these DAAs against HCV make them unaffordable for most patients, inviting the design of alternative and more economical therapies to control HCV. The 2014 Ebola virus (EBOV) outbreak resulted in the widespread transmission of this member of the Filoviridae through the West African countries of Guinea, Liberia, and Sierra Leone and localized cases in Nigeria, Mali, Spain, Senegal, the United Kingdom, Italy, and the United States. With a total of 11,298 deaths being reported as of August 2015, this epidemic is the deadliest and largest EBOV outbreak in recorded history and has only recently been contained with the help of foreign aid and experimental drugs and treatments (6). Other RNA viruses that have emerged to cause a significant health care burden or high case fatality rates in humans include influenza A virus (IAV), respiratory syncytial virus (RSV), Nipah virus (NiV), and Lassa virus (LASV). Incidences of real-time epidemics and the emergence of these pathogenic RNA viruses impart an urgent need for novel therapeutic interventions.

DAA therapies that target viral gene products and interfere with the viral life cycle are well characterized and widely used as therapeutic intervention strategies (7). The DAA therapy approach is highly effective and works with a high specificity against the targeted virus, though it does have disadvantages. RNA viruses have inherently unstable genomes that rapidly mutate, creating quasispecies, while the selection pressure on viral genes created by DAAs naturally promotes viral escape through resistance mutations. Viral mutation creates a problem for the long-term use of DAAs, as the mutations accumulated by the viral genomes eventually select for drug-resistant virus strains that then render the DAA ineffective. The use of DAAs also requires the precise identification of the target specific to that particular virus or subset of viruses, which precludes their use to control outbreaks of newly emergent virus strains that have yet to be identified or characterized.

Cellular proteins known as pathogen recognition receptors (PRRs), including the RIG-I-like receptors (RLRs) and Toll-like receptors, function to detect viral RNA and signal an innate immune response essential for limiting and controlling viral infections in the host (8–12). Upon recognition and engagement of viral pathogen-associated molecular patterns (PAMPs), including viral nucleic acid and other macromolecules, PRRs signal downstream to activate transcription factors, including interferon (IFN) regulatory factor 3 (IRF3), NF-κB, and others, which in turn induce the expression of many innate immune and antiviral genes and the production of antiviral gene products, proinflammatory cytokines, chemokines, and IFNs. These products act in concert to suppress and control virus infection. Synthetic and natural ligands that artificially induce PRR signaling and activate the innate immune response in cells and in mice show a proof of concept that triggering PRR pathways for innate immune induction can provide protection against virus infection (13, 14). Signaling convergence to activate IRF3 is a common theme shared among PRRs (12). Thus, the use of therapeutic agonists that can stimulate IRF3 activation may therefore be an effective strategy to control virus infections by promoting or modulating host innate immune antiviral responses. Such agents would have the added benefit of being broadly effective against different viruses. These immune modulating agents would exert a broad immune response that could easily be overcome by an individual virus or the accumulation of viral mutations.

We previously reported on a high-throughput cell-based screening approach (which uses the Kineta AViiD Discovery platform) which can be used to identify small, drug-like molecules that drive IRF3 activation and innate immune antiviral activity (15). Among the compounds identified in this screen were a group of isoflavone compounds that exhibited antiviral activity against HCV and IAV in cultured cells (15) and a second group of benzothiazole compounds that suppressed RSV infections in cultured cells (data not shown). Here, we report on the identification of a third family of small molecule compounds consisting of hydroxyquinolines that can activate IRF3 and induce antiviral genes in treated cells. Prophylactic and therapeutic treatment of cell cultures significantly decreased the viral RNA load within cells infected with a range of pathogenic RNA viruses and concomitantly suppressed the production of infectious virus. Our study thus identifies immune modulating molecules that function as innate immune agonists of IRF3 activation to mediate broad-spectrum antiviral activity.

MATERIALS AND METHODS

Cell lines and viruses.

Cells of the Huh7, PH5CH8, HeLa, HEK293, Vero, and MDCK cell lines were cultured in Dulbecco's modified Eagle's medium (DMEM; Cellgro) supplemented with 10% fetal bovine serum (FBS; HyClone), l-glutamine, sodium pyruvate, and nonessential amino acids. Cells of the THP-1 (human monocytic) cell line were cultured in complete RPMI (Invitrogen) supplemented with 10% heat-inactivated FBS. THP-1 cells were differentiated in complete RPMI supplemented with 40 nM phorbol myristate acetate (PMA) for 24 h prior to their use in experiments. Normal primary human umbilical vein endothelial cells (HUVECs) were obtained from Clonetics, Inc. (San Diego, CA), and maintained in endothelial cell growth medium supplemented with human recombinant epidermal growth factor, hydrocortisone, fetal bovine serum, gentamicin, and bovine brain extract (Clonetics). WNV isolate TX 2002 (WNV-TX) was described previously (16), and the titer was determined by a standard plaque assay on Vero cells. DV serotype 2 (DV2) was a gift from Alec J. Hirsch. DV2 was grown on cells of the C636 mosquito cell line, and the titer was determined by standard plaque assay on Vero cells. Cell culture-adapted HCV was produced from pJFH-1 carrying the JFH-1 genotype 2A infectious clone and in vitro transcribed as described previously (17). The HCV titer was determined by a standard focus-forming assay on Huh7 cells. EBOV strain Zaire (strain Kikwit), LASV (strain Josiah), NiV (strain Malaysia), IAV (H3N2 strain A/Udorn/72), RSV (strain A2; ATCC VR-1540), and Sendai virus (SeV; Cantell strain; Charles River) were used in infections.

Antibodies.

The following primary antibodies were used for detection by immunoblotting: rabbit anti-RIG-I (antibody 969; raised in rabbit against a RIG-I CARD peptide sequence) (18), rabbit anti-MDA5 (Enzo Life Sciences), rabbit anti-HCV NS5A (a gift from Jin Ye), rabbit anti-IFIT1 (antibody 972; raised in rabbit against an IFIT1 peptide sequence), rabbit anti-Mx1 (Cell Signaling), rabbit anti-MAVS (Novus Biologicals), rabbit anti-IRF7 (Cell Signaling), rabbit anti-IRF3 (a gift from Michael David), rabbit anti-IRF3 phosphoserine 396 (Cell Signaling), mouse antiactin (EMD Millipore), mouse anti-IAV NP (Chemicon), and mouse anti-RSV fusion protein (EMD Millipore). Horseradish peroxidase (HRP)-conjugated secondary antibodies were obtained from Jackson ImmunoResearch. Alexa Fluor 488-conjugated secondary antibodies were obtained from Thermo Fisher Scientific.

Compounds.

Compounds were synthesized by Life Chemicals Inc. and Kineta, Inc., solubilized in 100% dimethyl sulfoxide (DMSO), and kept frozen as 10 mM stocks. The compounds were stored in small aliquots to prevent multiple freeze-thaws and were stepwise diluted to reach the desired concentration in 0.5% DMSO for all treatments (15).

RT-PCR quantitation of viral RNA or innate immune genes.

Cells were harvested in RLT buffer (Qiagen), and total cellular RNA was purified using an RNeasy kit (Qiagen). cDNA was synthesized from the purified RNA by both random and oligo(dT) priming using an iScript cDNA synthesis kit (Bio-Rad). Intracellular WNV and DV2 RNA levels were measured using the SYBR green method (Applied Biosystems) on a reverse transcription (RT) machine (7300 RT-PCR; Applied Biosystems) and the relative quantitation method. The samples were normalized by subtracting the threshold cycle values of GAPDH (glyceraldehyde-3-phosphate dehydrogenase). The fold induction of viral RNA or innate immune genes over the levels of induction for either mock-infected cells or DMSO-treated control cells was calculated. Primer sequences are available upon request. The HCV RNA for the standard curve was generated by an in vitro transcription assay using the pSJ plasmid that carries the full-length JFH-1 genotype 2a clone (19). A standard curve of HCV RNA levels was generated by serially diluting the in vitro-transcribed HCV RNA from 10⁷ to 10⁰ copies/μl. HCV RNA levels were measured using a TaqMan assay (EZ RT-PCR core reagents; Applied Biosystems) through the absolute quantitation method. The HCV standard curve was used to extrapolate the number of copies of HCV RNA per microliter present in each sample. The sequences of the RT-PCR primers used are available upon request.

Immunoblot analyses.

Cells were trypsinized and collected in radioimmunoprecipitation assay buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with okadaic acid (Calbiochem) and a cocktail of protease inhibitors (Sigma) and phosphatase inhibitors (Calbiochem). Following cell lysis, nuclear material was removed by centrifugation at 15,000 × g for 10 min at 4°C. Cell lysates were quantified by the bicinchoninic acid method (Thermo Fisher Scientific), analyzed on a denaturing Tris-HCl–polyacrylamide gel, and transferred onto nitrocellulose membranes. Cellular and viral proteins of interest were detected by immunoblot analysis using the specific primary antibodies described earlier and HRP-conjugated secondary antibodies (Jackson Laboratory) and visualized by chemiluminescence on X-ray film, and the band intensity was quantitated using ImageJ software (see the National Institutes of Health ImageJ website at http://imagej.nih.gov/ij/ and reference 20).

Infection studies with WNV, DV2, and HCV.

HEK293 or Huh7 cells were used for WNV infection, and Huh7 cells were used for HCV and DV2 infection. For experiments in which cells were treated prior to virus infection, cells were pretreated with either 0.5% DMSO or the compound indicated below over a range of concentrations for 24 h and then infected with the respective virus at a multiplicity of infection (MOI) of 1. The virus inoculum was removed 2 h after infection and replaced with complete medium without any treatment. For experiments in which cells were treated after virus infection, cells were infected with HCV or WNV at an MOI of 0.1 or DV2 at an MOI of 1. The virus inoculum was removed 2 h after infection and replaced with complete medium supplemented with 0.5% DMSO or the indicated compound over a range of concentrations. Cell pellets were collected in RLT buffer (Qiagen) at 24 and 48 h postinfection for RNA extraction and RT-PCR quantitation of viral RNA. The cell culture supernatant was collected from infected cells at 24 and 48 h after infection, and infectious virus particles were quantitated by conventional plaque or focus-forming assays. For time-of-addition studies, cells were infected with HCV or WNV at an MOI of 0.1 or DV2 at an MOI of 1. The virus inoculum was removed at the times after infection indicated below and replaced with complete medium supplemented with 0.5% DMSO or the compound indicated below. Cell pellets and cell culture supernatants were collected from infected cells at 48 h postinfection for the quantitation of viral RNA by RT-PCR and the quantitation of infectious virus particles by conventional plaque or focus-forming assays. Infectious WNV or DV2 particles were measured by a standard plaque assay using an agar overlay on Vero cells. Virus plaques were counted on day 2 postinfection for WNV using neutral red staining and day 8 postinfection for DV2 using crystal violet staining. Infectious HCV particle titers were determined on Huh7.5 cells. Virus foci were detected using a rabbit polyclonal antibody directed against the HCV NS5A protein in combination with a horseradish peroxidase-conjugated secondary antibody, visualized using a Vector VIP peroxidase substrate kit (Vector Laboratories), and counted by microscopy (21).

Infection studies with EBOV, NiV, and LASV.

All EBOV, NiV, and LASV infections were done in the biosafety level 4 facility at the University of Texas Medical Branch (UTMB) at Galveston, TX. HUVECs were seeded in 24-well plates and allowed to grow to confluence. Cells were treated with 0.5% DMSO or KIN1408 at 1 or 5 μM for 22 h, at which point the medium was replaced with fresh medium supplemented with 0.5% DMSO or KIN1408 at 1 or 5 μM. Two hours later (24 h after the initial treatment), cells were infected with EBOV at an MOI of 0.5, NiV at an MOI of 0.1, or LASV at an MOI of 0.01 or were mock infected (treated with medium alone). The virus inoculum was removed at 1 h postinfection and replaced with medium containing 0.5% DMSO or KIN1408 at 1 or 5 μM. The cell culture supernatant was collected at 96 h postinfection from EBOV-infected cells or at 24 and 48 h postinfection from NiV- and LASV-infected cells and analyzed by a conventional plaque assay on Vero E6 cells with a limit of detection of 25 PFU/ml.

Infection studies with IAV or RSV.

HEK293 cells were infected with IAV at an MOI of 0.1. HeLa cells were infected with RSV at an MOI of 0.1. At 2 h postinfection, the cells were treated with 0.5% DMSO or the compound indicated below over a range of concentrations. At 24 h after IAV infection or 48 h after RSV infection, the cell culture supernatant was collected and analyzed by a conventional focus-forming assay on MDCK cells for IAV and HeLa cells for RSV. Briefly, cells were fixed using a 1:1 methanol-acetone solution, and virus-infected cell foci were detected using a fluorescein isothiocyanate (FITC)-coupled monoclonal antibody directed against the IAV NP protein or a mouse monoclonal antibody directed against the 47- to 49-kDa fusion protein of RSV in combination with an FITC-conjugated secondary antibody. Virus-infected foci were quantified using a Thermo Fisher Scientific ArrayScan VTI fluorescent microscope imager, and the number of focus-forming units (FFU) per milliliter of cell culture supernatant was calculated.

Immunofluorescent assays for nuclear translocation of IRF3.

HEK293, PH5CH8, or Huh7 cells were treated with complete DMEM supplemented with 0.5% (vol/vol) DMSO with a small molecule (KIN1400, KIN1408, and KIN1409 at 5, 10, or 20 μM), SeV at 100 hemagglutinating units (HAU)/ml, or 0.5% DMSO alone. Cells were fixed in 4% paraformaldehyde and stained with an anti-IRF3 rabbit polyclonal antibody, an Alexa Fluor 488-conjugated secondary antibody, and either 4′,6′-diamidino-2-phenylindole (DAPI) or Hoechst dye for the nuclei. Images were taken using a Nikon Eclipse Ti microscope with a 60× oil immersion lens and NIS-Elements software. Images were additionally visualized and quantitated for a nuclear IRF3 signal using a Cellomics ArrayScan VTI fluorescent microscope imager (Thermo Fisher Scientific).

Cell viability assays.

HEK293 or Huh7 cells were treated with complete DMEM supplemented with 0.5% (vol/vol) DMSO with a small molecule (KIN1400, KIN1408, and KIN1409 at 5, 10, or 20 μM), a combination of tumor necrosis factor alpha (TNF-α) plus cycloheximide (CHX), or 0.5% DMSO alone. For cell survival analysis, the bioreduction of the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) tetrazolium compound was measured using a CellTiter 96 Aqueous One Solution cell proliferation assay (Promega) following the manufacturer's instructions, and Sytox green (Thermo Fisher Scientific) uptake was visualized and quantitated using an IncuCyte Zoom live cell imaging system (Essen Biosciences).

Microarray analysis.

Differentiated THP-1 cells were treated with complete RPMI supplemented with 0.5% (vol/vol) DMSO with a small molecule (KIN1400, KIN1408, and KIN1409 at 10, 2.5, or 0.625 μM), beta interferon (IFN-β) at 100 U/ml, SeV at 100 HAU/ml, or 0.5% DMSO alone. Additional cells were transfected with HCV PAMP poly(U/UC) (pU/UC) at 2 μg/ml or HCV X region RNA (XRNA) at 2 μg/ml with a TransIT-mRNA transfection kit (Mirus Bio LLC). Cells were harvested at 20 h in RLT buffer. RNA was purified using Qiagen RNeasy kits and submitted to Labcorp, Seattle, WA (previously called Covance), for microarray analysis using Agilent SurePrint G3 human genome microarrays (version 2). Array data were processed by the M. Gale lab using R (version 3.2.1)/Bioconductor (version 3.1) (see the website of the R Development Core Team [https://www.r-project.org] and reference 22). Raw data were quantile normalized followed by linear modeling using the limma package (version 3.24.15) (23). Genes with significant changes in expression following treatment were defined by those with a >2-fold increase or decrease in expression over that for the DMSO or XRNA controls with a Benjamini-Hochberg-corrected P value of <0.01. The gene expression heat map was clustered using Spearman correlation distances, and Gene Ontology biological processes enriched in lists of genes mapping to the various clusters were determined using the DAVID (version 6.7) program (24, 25). Genes with predicted IRF7 binding sites according to the UCSC Genome Browser database were identified using the Enrichr program (26). Finally, genes mapping to the Reactome Homo sapiens interferon alpha/beta signaling pathway (R-HAS-909733) were identified using the InnateDB database (27).

Microarray data accession number.

The microarray data discussed in this publication have been deposited in the NCBI Gene Expression Omnibus (GEO) database (28) and are accessible through GEO series accession number GSE74047 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE74047).

RESULTS

KIN1400 triggers IRF3-dependent innate immune antiviral genes and IFN-β expression.

A cell-based screen using a small molecule diversity library was conducted, In that screen, we identified compounds that can activate RLR-dependent IRF3 activation leading to IFIT1, IFIT2, or IFN-β transcription (15). We have previously reported the identification of two classes of small molecules with antiviral activity: a group of isoflavone-like compounds that suppressed HCV and IAV growth in cultured cells (15) and a second group of small molecules with a benzothiazole core, represented by KIN1000 (Fig. 1A), that suppressed RSV growth in cultured cells. Here, we report on the identification of a third class of IRF3-activating compounds that is represented by KIN1400, which has a completely distinct chemical structure with a hydroxyquinoline at its core (Fig. 1B).

FIG 1.

IRF3-activating small molecule compounds KIN1000 and KIN1400 induce differential expression of innate immune genes. (A, B) Structures of KIN1000, which belongs to the benzothiazole family (A), and KIN1400, which belongs to the hydroxyquinoline family (B). (C) Immunoblot analyses of total protein lysates collected from PMA-differentiated THP-1 cells treated with 20 μM compound for 20 h. The levels of protein expression by the RIG-I, MDA5, IFIT1 (ISG56), and Mx1 genes relative to the level of tubulin expression were determined using the respective antibodies. Shown here are representative images from one of three independent experiments. (D) PMA-differentiated THP-1 cells were treated with 1.25, 5, or 20 μM compound for 20 h, and the total cellular RNA was purified. The levels of expression of innate immune genes RIG-I (DDX58), IFIT1 (ISG56), IFIT2 (ISG54), OAS3, Mx1, IFITM1, IFN-β, IFN-λ2/3, and IL-6 were measured by RT-PCR, were normalized to the level of GAPDH expression, and are expressed as the fold induction over that achieved with 0.5% DMSO treatment from three independent experiments. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05.

To define the innate immune signaling pathways induced by KIN1400, we first surveyed whether KIN1400 can induce the expression of IRF3-dependent antiviral genes. We treated macrophage-like human THP-1 cells with 0.5% DMSO or 20 μM KIN1000 or KIN1400 in 0.5% DMSO and evaluated gene expression. As a control, cells were treated with exogenous IFN-β (25 or 50 IU/ml), infected with Sendai virus (SeV; a potent activator of RIG-I-dependent signaling), or mock infected. KIN1400 treatment induced the expression of the innate immune genes RIG-I, MDA5, IFIT1, and Mx1 compared to the level of expression for the DMSO control, as determined by immunoblot analysis (Fig. 1C). By RT-PCR, we confirmed that KIN1400 treatment induced RIG-I (DDX58), IFIT1, and Mx1 expression in a dose-dependent fashion (Fig. 1D). Additionally, KIN1400 treatment also induced the expression of the innate immune genes IFIT2, IFITM1, and OAS3. On comparison of the two compounds dose for dose, KIN1400 appeared to be more potent than KIN1000 at inducing overall innate immune gene expression. We note that KIN1400 treatment induced little if any expression of IFN-β, type III IFN (IFN-λ2/3), or the proinflammatory cytokine interleukin-6 (IL-6) compared to the level of expression for the DMSO-treated control. Despite having identified KIN1400 and KIN1000 in similar screens that selected for small molecules that activate RLR-dependent signaling, the two compounds induced innate immune gene expression but little expression of type I and III IFNs, suggesting that they induce transcriptional activity that is distinct from that induced by IFN treatment alone and also different from that induced by RLR agonist RNA, which potently induces type I IFN expression (12, 13).

KIN1400 suppresses WNV in cultured cells.

We tested the compounds KIN1000 and KIN1400 for their ability to protect cultured cells from WNV infection. The compounds were added to HEK293 cells for 24 h before the cells were infected with WNV at a multiplicity of infection (MOI) of 1. Total cellular RNA was collected 24 h after infection and evaluated by real-time quantitative PCR (qPCR) for intracellular WNV RNA levels relative to the level of cellular GAPDH expression. We also assessed the cell culture supernatant for infectious WNV particles by a standard plaque assay. At 20 μM, KIN1000 had modest effects on intracellular WNV RNA levels in comparison to the effects of the control DMSO treatment (Fig. 2A). In contrast, KIN1400 suppressed intracellular WNV RNA to levels greater than even the level achieved with 100 IU/ml IFN-β treatment. Additionally, we show that KIN1400 suppression of WNV RNA levels is dose dependent, with 2 μM KIN1400 being sufficient to achieve a 50% or greater inhibition of WNV RNA levels (Fig. 2B). The pretreatment of cells with KIN1400 also led to a significant log reduction in the infectious WNV titer at 24 and 48 h postinfection compared to that for cells that were treated with DMSO (Fig. 2C). These data show KIN1400 to be effective at controlling WNV infection in cultured cells when the compound is administered prior to infection.

FIG 2.

KIN1400 inhibits WNV infection in cultured cells. (A) HEK293 cells were treated with 0.5% DMSO, 20 μM KIN1000 or KIN1400, or 100 IU/ml IFN-β for 24 h prior to WNV infection at an MOI of 1. WNV RNA levels in the cells were measured by RT-PCR at 24 h after infection and calculated as a percentage of the WNV RNA levels detected in DMSO-treated control cells. Statistical significance was determined by one-way analysis of variance with Dunnett's multiple-comparison test. **, P < 0.001; ns, not significant (P ≥ 0.5). (B) HEK293 cells were pretreated with 0.5% DMSO or a titration (0.2, 2, 10, or 20 μM) of KIN1000 or KIN1400 for 24 h before WNV infection at an MOI of 1. WNV RNA levels in the cells were measured by RT-PCR at 24 h after infection and calculated as a percentage of the WNV RNA levels detected in DMSO-treated control cells. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, not significant (P ≥ 0.5). (C) HEK293 cells were pretreated with 0.5% DMSO, 20 μM KIN1000 or KIN1400, or 100 IU/ml IFN-β for 24 h before they were infected with WNV at an MOI of 1. The cell culture supernatant was collected at 24 or 48 h after infection, and the number of infectious virus particles was determined by plaque assay on Vero cells. The results shown are the average number of infectious virus particles obtained at 24 or 48 h after infection and were calculated as the number of PFU per milliliter of cell culture supernatant. Statistical significance was determined by two-way analysis of variance with Bonferroni posttests. ***, P < 0.001; ns, not significant (P ≥ 0.5). (D) HEK293 cells were infected with WNV (MOI, 0.1) for 2 h and then treated with 0.5% DMSO, 20 μM KIN1000 or KIN1400, or 100 IU/ml IFN-β. WNV RNA levels in the cells were measured by RT-PCR at 24 h after infection and calculated as a percentage of the WNV RNA levels detected in DMSO-treated control cells. Statistical significance was determined by one-way analysis of variance with Dunnett's multiple-comparison test. **, P < 0.001; ns, not significant (P ≥ 0.5). (E) HEK293 cells were infected with WNV (MOI, 0.1) for 2 h and then treated with 0.5% DMSO or a dose of 0.2, 2, 10, or 20 μM compound. WNV RNA levels in the cells were measured by RT-PCR at 24 h after infection and calculated as a percentage of the WNV RNA levels detected in DMSO-treated control cells. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.001; ns, not significant (P ≥ 0.5). Rel, relative; 1000, KIN1000; 1400, KIN1400.

KIN1000 and KIN1400 were also tested for their ability to inhibit WNV when the compounds were administered postinfection. In this case, WNV-infected HEK293 cells (MOI, 0.1) were treated with compounds 2 h after infection was established. Total cellular RNA and cell culture supernatant were collected 24 h after infection and evaluated as described above. When administered postinfection, IFN-β treatment resulted in a moderate suppression of intracellular WNV RNA levels compared to the level of suppression achieved with DMSO treatment (Fig. 2D). This finding is consistent with reports that the genome of the strain of WNV used here (strain TX 2002) encodes viral proteins that can antagonize and overcome type I IFN-induced JAK-STAT signaling (16). Importantly, KIN1400 but not KIN1000 was able to suppress WNV RNA levels. When administered postinfection, KIN1400 suppression of WNV RNA levels was also dose dependent, with a concentration of between 2 and 10 μM being required to achieve 50% suppression (Fig. 2E). Our data show that KIN1400 is a compound that exhibits antiviral activity against WNV infection whether it is administered before or after WNV exposure.

KIN1400 suppresses DV infection.

We evaluated the compounds for their ability to inhibit DV, another member of the Flaviviridae family of the genus Flavivirus related to WNV. In this case, 20 μM each compound was administered to human hepatoma Huh7 cells 24 h prior to infection with DV serotype 2 (DV2) at an MOI of 1. KIN1000 treatment resulted in a moderate reduction in intracellular DV2 RNA levels by real-time qPCR compared to those achieved with the control DMSO treatment, whereas treatment with 100 IU/ml IFN-β or KIN1400 suppressed viral RNA levels (Fig. 3A). Although all KIN compounds suppressed DV2 RNA levels in cells in a dose-dependent fashion, KIN1400 was by far more effective than KIN1000, demonstrating the production of a greater than 50% decrease in viral RNA levels at a 2 μM concentration (Fig. 3B). Remarkably, this response corresponded to an almost 4-log-unit drop in the level of infectious DV2 particle production at 24 h postinfection (Fig. 3C). A 4-log-unit drop in the amount of infectious virus was also achieved with IFN-β treatment relative to the amount achieved with the control DMSO treatment. When administered at 4 h postinfection, only KIN1400 effected any appreciable reduction in intracellular DV2 RNA levels, with the effects at 20 μM being equivalent to or better than those achieved with 100 IU/ml IFN-β treatment (Fig. 3D). Furthermore, a similar level of suppression was achievable with just 2 μM KIN1400 (Fig. 3E). Taken together, our data identify KIN1400 to be a compound that inhibits the flaviviruses WNV and DV2 in cultured cells whether it is administered prophylactically prior to infection or therapeutically at times postinfection.

FIG 3.

KIN1400 inhibits DV2 infection in cultured cells. (A) Huh7 cells were pretreated with 0.5% DMSO, 20 μM KIN1000 or KIN1400, or 100 IU/ml IFN-β for 24 h before DV2 infection at an MOI of 1. DV2 RNA levels in the cells were measured by RT-PCR at 24 h after infection and calculated as a percentage of the DV2 RNA levels detected in DMSO-treated control cells. Statistical significance was determined by one-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.001; **, P < 0.01. (B) Huh7 cells were pretreated with a dose of 2, 10, or 20 μM KIN1000 or KIN1400 for 24 h before DV2-infection at an MOI of 1. DV2 RNA levels in the cells were measured by RT-PCR at 24 h after infection and calculated as a percentage of the DV2 RNA levels detected in DMSO-treated control cells. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.001; **, P < 0.01. (C) Huh7 cells were pretreated with 0.5% DMSO or 20 μM KIN1000 or KIN1400 for 24 h before they were infected with DV2 (MOI, 1). Cell culture supernatant was collected at 24 or 48 h after infection, and the number of infectious virus particles was measured by plaque assay on Vero cells. The results shown are the average number of infectious virus particles obtained at 24 or 48 h after infection and were calculated as the number of PFU per milliliter of cell culture supernatant. Statistical significance was determined by two-way analysis with Dunnett's multiple-comparison test. ***, P < 0.001. (D) Huh7 cells were infected with DV2 at an MOI of 0.1 for 2 h and then treated with 0.5% DMSO, 20 μM compound, or 100 IU/ml IFN-β. DV2 RNA levels in the cells were measured by RT-PCR at 24 h after infection and calculated as a percentage of the DV2 RNA levels detected in DMSO-treated control cells. Statistical significance was determined by one-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.001; ns, not significant (P ≥ 0.5). (E) Huh7 cells were infected with DV2 (MOI, 0.1) for 2 h and then treated with 0.5% DMSO or a dose of 2, 10, or 20 μM compound. DV2 RNA levels in the cells were measured by RT-PCR at 24 h after infection and calculated as a percentage of the DV2 RNA levels detected in DMSO-treated control cells. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.001; *, P < 0.05; ns, not significant (P ≥ 0.5). (F, G) Huh7 cells were infected with DV2 (MOI, 0.1) prior to treatment with 0.5% DMSO, 20 μM KIN1400, or 100 IU/ml IFN-β at 1, 2, 4, 8, and 24 h after infection. (F) DV2 RNA levels in the cells were measured by RT-PCR at 48 h after infection and calculated as a percentage of the DV2 RNA levels detected in DMSO-treated control cells. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ****, P < 0.0001; **, P < 0.01. (G) The cell culture supernatant was collected 48 h after infection, and the number of infectious virus particles was determined by plaque assay on Vero cells. The results shown are the average number of infectious virus particle obtained at 48 h after infection and were calculated as the number of PFU per milliliter of cell culture supernatant. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ****, P < 0.0001.

To define the kinetics of the KIN1400 antiviral action, we conducted a time-of-addition study to determine the maximum delay in the treatment of cells with KIN1400 after DV2 infection that can be used and still achieve antiviral activity. For this purpose, cells infected with DV2 at an MOI of 1 were treated with 20 μM KIN1400 at 1, 2, 4, 8, and 24 h postinfection. We then evaluated DV2 RNA levels in cells and measured the number of infectious viral particles in the cell culture supernatant at 48 h postinfection. As a control, DV2-infected cells were similarly treated with DMSO or IFN-β at 100 IU/ml. KIN1400 at 20 μM and IFN-β at 100 IU/ml were both effective at reducing both viral RNA levels in cells (Fig. 3F) and infectious DV2 particle production (Fig. 3G) when administered at all the time points tested up to 24 h postinfection. Remarkably, treatment with KIN1400 imposed an approximately 2-log-unit drop in infectious virus production that was sustained through the 48-h culture period (Fig. 3G). Our data indicate that KIN1400 can inhibit viral replication and infectivity even when administered at times approaching the time of peak viral replication of 24 h, thus demonstrating that KIN1400 is an antiviral agent that may be developed to treat two different flaviviruses of public health concern.

KIN1400 inhibits HCV replication.

Having shown that KIN1400 inhibits WNV and DV, two mosquito-borne members of the Flaviviridae and the genus Flavivirus, we asked whether the compounds would have antiviral activity against other viruses of the Flaviviridae. For this reason, we tested the compounds for their efficacy in inhibiting HCV, a member of the Flaviviridae but belonging to the genus Hepacivirus. When administered to Huh7 cells at 20 μM and 24 h prior to infection, KIN1000 and KIN1400 variably suppressed HCV (JFH1, genotype 2A) RNA levels, whereas KIN1400 suppressed HCV RNA to levels approaching those achieved with 100 IU/ml IFN-β treatment (Fig. 4A). KIN1400 also had a dose-dependent effect in reducing HCV RNA levels (Fig. 4B) that correlated with reductions in the amount of infectious HCV particles released into the cell culture supernatant (Fig. 4C). Importantly, when administered at 2 h postinfection, only treatments with 100 IU/ml IFN-β or KIN1400 significantly inhibited HCV RNA levels in cells compared to the levels achieved with the control DMSO treatment (Fig. 4D). The effect of KIN1400 on HCV was dose dependent, with KIN1400 demonstrating a 50% effective concentration (EC₅₀) of <2 μM when it was administered 24 h before infection (Fig. 4B), and we estimated an EC₅₀ of ∼2 to 5 μM when it was administered after infection (Fig. 4E). Thus, KIN1400 demonstrates antiviral activity to control infections caused by viruses of the Flaviviridae family.

FIG 4.

KIN1400 inhibits HCV (JFH-1) infection. (A) Huh7 cells were pretreated with 0.5% DMSO, 20 μM KIN1000 or KIN1400, or 100 IU/ml IFN-β for 24 h before they were infected with HCV (MOI, 1). The HCV RNA levels in the cell culture supernatant collected 24 h after infection were measured by the TaqMan RT-PCR assay and calculated as the amount of HCV RNA (in number of copies per milliliter) using a standard curve. Statistical significance was determined by one-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.0001. (B) Huh7 cells were pretreated with 0.5% DMSO or KIN1000 or KIN1400 at a dose of 0.2, 2, or 20 μM for 24 h before infection with HCV (MOI, 1). HCV RNA levels in the cell culture supernatant collected 24 h after infection were measured by the TaqMan RT-PCR assay and calculated as the amount of HCV RNA (number of copies per milliliter) using a standard curve. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.001; **, P < 0.01; ns, not significant (P ≥ 0.5). (C) Huh7 cells were pretreated with 0.5% DMSO, 20 μM KIN1000 or KIN1400, or 100 IU/ml IFN-β for 24 h before they were infected with HCV (MOI, 1). The number of infectious virus particles in the cell culture supernatant collected 48 h after infection was determined by a focus-forming assay on Huh7 cells. The results shown are the average number of focus-forming units (FFU) per milliliter of cell culture supernatant. Statistical significance was determined by one-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.0001. (D) Huh7 cells were infected with HCV (MOI, 0.1) for 2 h and then treated with 0.5% DMSO, 20 μM compound, or 100 IU/ml IFN-β. HCV RNA levels in the cell culture supernatant collected 24 h after infection were measured by the TaqMan RT-PCR assay and calculated as the amount of HCV RNA (number of copies per milliliter) using a standard curve. Statistical significance was determined by one-way analysis of variance with Dunnett's multiple-comparison test. **, P < 0.001; ns, not significant (P ≥ 0.5). (E) Huh7 cells were infected with HCV (MOI, 0.1) for 2 h and then treated with the compound at a dose of 0.2, 2, 10, or 20 μM. HCV RNA levels in the cell culture supernatant collected 24 h after infection were measured by the TaqMan RT-PCR assay and calculated as the amount of HCV RNA (number of copies per milliliter) using a standard curve. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. *, P < 0.5; ns, not significant (P ≥ 0.5).

KIN1400 induces innate antiviral immunity through a MAVS-IRF3 axis.

To confirm that IRF3 is required for KIN1400 induction of innate immune gene expression and antiviral activity, we assessed HEK293 cells that transiently express an IRF3 mutant that is dominant negative for signaling (IRF3ΔN) (29) for their ability to express IFIT2, an antiviral gene and target of IRF3 (30), after treatment with KIN1400. Cells expressing IRF3ΔN failed to induce IFIT2 expression in response to KIN1400, whereas vector-transfected control cells were able to induce IFIT2 expression in response to KIN1400 (Fig. 5A). Consistent with previous findings, IRF3ΔN expression also suppressed the potent expression of IFIT2 that is induced by SeV infection (31). These data verify that KIN1400 signals through IRF3 to drive innate immune gene expression. We additionally assessed the requirement for MAVS, the essential adaptor protein of the RLR pathway (12), in KIN1400 induction of innate immune gene expression. For this purpose, we created a Huh7 cell line in which MAVS was targeted for deletion by the clustered regularly interspaced short palindromic repeat (CRISPR)–CRISPR-associated protein 9 (Cas9) system. We confirmed by immunoblot analysis that MAVS expression is lacking in MAVS-knockout (KO) Huh7 cells, whereas it is present in control Huh7 cells (Fig. 5B). In these cells, KIN1400 retained its ability to induce IFIT1 and IFIT2 expression in wild-type control cells but not in the MAVS-KO cells (Fig. 5C), suggesting that KIN1400 signals through MAVS and, thus, the RLR pathway to drive innate immune gene expression. Taken together, our data show KIN1400 to be a small molecule that can drive innate immune gene expression in a MAVS- and IRF3-dependent manner.

FIG 5.

Pathway mapping and structure-activity relationship of KIN1400 and analogs. (A) HEK293 cells were transfected with either a vector control plasmid or a plasmid that ectopically expressed IRF3ΔN (a dominant negative IRF3 mutant that abolishes signaling). Cells were then treated with 0.5% DMSO or 0.5 μM KIN1400 or infected with 100 HAU/ml SeV for 20 h. Total cellular RNA was extracted, and the level of expression of the innate immune gene IFIT2 (ISG54), normalized to the level of GAPDH expression, was measured by RT-PCR and expressed as the fold induction over that for the respective DMSO-treated control. Statistical significance was determined by multiple t tests. *, P = 7.8 × 10⁻⁶ and 3.6 × 10⁻⁶ for KIN1400 and SeV, respectively; ns, not significant (P ≥ 0.5). (B) Immunoblotting of control Huh7 cells and Huh7 cells with CRISPR-Cas9-induced deletion of MAVS confirms the reduction in the level of MAVS expression at the protein level. Shown here are representative images from one of three independent experiments. (C) Control or MAVS-KO Huh7 cells were treated with 0.5% DMSO or 20 μM KIN1400 for 20 h. Total cellular RNA was extracted, and the levels of expression of innate immune genes IFIT1 (ISG56) and IFIT2 (ISG54) were measured by RT-PCR. Statistical significance was determined by multiple t tests compared with the results obtained with KIN1400. *, P = 4.6 × 10⁻⁵ and 3.9 × 10⁻⁷ for IFIT1 and IFIT2, respectively; ns, not significant (P ≥ 0.5). WT, wild type. (D) KIN1400 and the analogs share an aminothiazole core with either a side chain substitution of benzene, as shown in the structures of KIN1407 and KIN1408, or a side chain substitution of phenyl, as shown in the structures of KIN1409 and KIN1410. (E) PH5CH8 cells were treated with 0.5% DMSO or 10 μM compound or infected with 100 HAU/ml SeV for 4 h. Cells were fixed in paraformaldehyde and stained for immunofluorescence microscopy for the detection of IRF3 (green, Alexa Fluor 488) and nuclei (DAPI). Shown here are representative images from one of three independent experiments. Bar = 50 μm. (F) Huh7 cells were treated with 5, 10, or 20 μM KIN1400 or one of its analogs for 20 h. Cells were stained for immunofluorescence microscopy, and the relative amounts of nuclear IRF3 were measured by the use of an ArrayScan fluorescent microscope imager and plotted over the concentration of the compound (in micromolar). Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. **, P < 0.01; *, P < 0.1. (G) Huh7 cells were treated with 5, 10, 25, or 50 μM KIN1400 or an analog or with 0.5% DMSO or were infected with EMCV (MOI, 0.1) for 24 h, and MTS tetrazolium bioreduction was measured using a CellTiter 96 aqueous cell proliferation kit (Promega). HEK293 cells were treated with 5, 10, or 20 μM KIN1400 or an analog, 0.5% DMSO, or a combination of TNF-α and CHX over the course of 36 h, and Sytox green dye uptake was measured using an IncuCyte live cell imaging system (Essen Biosciences). The statistical significance of data from three independent experiments was calculated using two-way analysis of variance with Dunnett's multiple-comparison test. ****, P < 0.0001; ns, not significant (P ≥ 0.5). OD₄₉₀, optical density at 490 nm. (H) THP-1 cells were differentiated with PMA for 30 h and treated with 0.5% DMSO, KIN1400, or an analog at a dose of 5, 10, or 20 μM, infected with SeV at 50 HAU/ml, or mock infected for 20 h. The cell lysate was visualized by immunoblot analysis for expression of various innate immune genes, phosphorylated IRF3 (IRF3-P) as a measurement of IRF3 activation, and actin as a control. The intensity of the bands was quantitated by ImageJ software, and the level of expression of each gene relative to that obtained with DMSO treatment was calculated. The results showed that KIN1400 and its analogs induce IRF3-P expression over a range of 1.6- to 3.6-fold of the level achieved with DMSO treatment, whereas SeV infection induces IRF3-P expression >10-fold. Shown here are representative images from one of three independent experiments. (I) Huh7 cells were infected with DV2 (MOI, 1) for 2 h and then treated with compound at a dose of 2, 10, or 20 μM. Total cellular RNA was extracted from cells collected at 48 h postinfection, and DV2 RNA levels were measured by RT-PCR and calculated as a percentage of the DV2 RNA levels detected in DMSO-treated control cells. Statistical significance was determined by one-way analysis of variance with Dunnett's multiple-comparison test. ***, P < 0.0001.

Analogs KIN1408 and KIN1409 retain activities similar to the activity of the KIN1400 parent.

Structural analogs of KIN1400 were designed to identify chemical modifications that can potentially increase drug potency through improved antiviral activity and/or solubility. As many as 10 analogs were synthesized, but of these, only 4 retained the ability to activate IRF3 (data not shown). All four analogs consist of an aminothiazole core with either a benzene group (Fig. 5D, KIN1407 and KIN1408), like the KIN1400 parent compound (Fig. 1B), or a phenyl group (Fig. 5D, KIN1409 and KIN1410). The analogs also differed in their side chain groups to include structures such as trifluoromethyl, difluoromethoxy, fluoro, bromo, or methoxy groups (Fig. 5D). The ability of the KIN1400 analogs to promote IRF3 nuclear translocation, as a measurement of IRF3 activation, was compared with that of KIN1400. When administered to PH5CH8 cells (Fig. 5E) or Huh7 cells (liver epithelial cells; data not shown) at 10 μM, the analogs were sufficient to induce IRF3 translocation into the nucleus, as was the case for the positive control of SeV infection. Measurements obtained with the ArrayScan fluorescent microscope imager showed that KIN1400 and analogs KIN1407, KIN1408, KIN1409, and KIN1410 drove IRF3 translocation into the nucleus in dose-dependent manner (Fig. 5F), confirming that the analogs retain an ability to activate IRF3 similar to that of their parent compound, KIN1400. Among the analogs tested, KIN1408 and KIN1409 appeared to be able to induce IRF3 activation at a level equivalent to that for the KIN1400 parent. In contrast, KIN1407 and KIN1410 consistently induced less IRF3 activation in cultures of Huh7 cells than equivalent doses of KIN1400. By separate assays that measured cell viability by MTS tetrazolium bioreduction and Sytox green uptake, we showed that HEK293 and Huh7 cells treated with KIN1408, KIN1409, and the KIN1400 parent at concentrations up to 50 μM demonstrated viability that was not significantly different from that of DMSO-treated control cells after up to 36 h of compound treatment (Fig. 5G and data not shown). As expected, cells infected with the rapidly lytic encephalomyocarditis virus (EMCV) or treated with a combination of tumor necrosis factor alpha (TNF-α) and cycloheximide (CHX) demonstrated significantly lower levels of viability by the two methods.

By immunoblot analysis, we demonstrated that KIN1408 induces the expression of the innate immune genes MDA5, RIG-I, Mx1, IRF7, and IFIT1 in a dose-dependent fashion to levels similar to those to which they were induced by KIN1400 (Fig. 5H). Although KIN1409 also induced the expression of these innate immune genes, the level of induction at each equivalent dose was generally weaker than that observed with KIN1400 and KIN1408. Compound treatment also induced weak but dose-dependent activation of IRF3, as measured by IRF3 phosphorylation (IRF3-P; bands indicated by asterisks in Fig. 5H). Quantitation of the band intensity using ImageJ software showed that KIN1400 induced up to 2.4-fold the levels of IRF3-P compared to the levels achieved in DMSO-treated cells, whereas KIN1408 and KIN1409 induced up to 1.6- and 3.6-fold the levels of IRF3-P, respectively, compared to the levels achieved in DMSO-treated cells. SeV-infected cells, which were included as a positive control, potently induced innate immune gene expression and IRF3 phosphorylation (up to 10.5-fold the levels of IRF3-P compared to those achieved in DMSO-treated control cells).

We next evaluated the analogs for their ability to suppress DV2 infection in cultured cells. Huh7 cells were infected with DV2 at an MOI of 1, and the compounds were administered to cells at the concentrations indicated below at 24 h postinfection. Total cellular RNA was collected at 48 h postinfection and assessed by RT-PCR for DV2 RNA levels relative to those achieved with the control DMSO treatment. As with KIN1400 (see above), treatment with all four analogs resulted in a dose-dependent reduction in DV2 RNA levels (Fig. 5I). Our data thus identify two analog molecules, KIN1408 and KIN1409, that have the potency to drive IRF3 activation to induce innate immune gene expression and that concomitantly suppress DV2 RNA to levels similar to those achieved with the KIN1400 parent.

KIN1408 and KIN1409 induce a selective gene expression profile.

We note that without further formulation, KIN1408 and KIN1409 appeared to be more readily soluble than the KIN1400 parent or other analogs (data not shown). We therefore chose to evaluate these two analogs in gene expression analyses and to compare the results obtained to those obtained with the KIN1400 parent compound. Macrophage-like THP-1 cells were treated with KIN1400, KIN1408, or KIN1409 at 0.625, 2.5, or 10 μM for 24 h. Total cellular RNA was collected 20 h after compound treatment and analyzed for genome-wide gene expression by microarray analysis. As controls, cells were treated with 0.5% DMSO (vehicle) or 100 IU/ml IFN-β. We also included cells that had been infected with 100 HAU/ml SeV as additional controls. For further comparison, we included an analysis of cells that had been transfected with a RIG-I-specific and MAVS-dependent agonistic RNA from the HCV genome, poly(U/UC) (pU/UC) RNA, or negative control HCV XRNA, which cannot activate RIG-I signaling (14, 21). Differential gene expression was defined as at least a 2-fold change in expression and a Benjamini-Hochberg-corrected P value of <0.01 compared to the results obtained with the appropriate negative control: XRNA for pU/UC RNA and DMSO alone for all other samples.

We plotted all genes differentially expressed in at least one treatment on a heat map (Fig. 6A). Gene clusters were identified through hierarchical clustering using the Spearman correlation as a distance measure and then classified by the most highly enriched gene ontology biological process meeting the criteria of a Benjamini-Hochberg-corrected P value of <0.05. The patterns of gene expression across doses demonstrated that the three compounds induced comparable changes in gene expression. Additionally, many genes regulated by the KIN compounds were also differentially expressed during IFN-β treatment, pU/UC RNA transfection, or SeV infection, consistent with the shared induction of innate immune antiviral responses. Finally, the parental KIN1400 compound primarily differed from the analogs in its induction and repression of sets of genes not overlapping with genes responsible for innate immune antiviral responses and not enriched in genes responsible for any particular biological process category that may indicate functional utility. Importantly, the KIN1408 and KIN1409 analog compounds were highly comparable to the parental KIN1400 compound in their induction of innate immune system- and inflammatory response-related genes, supportive of their comparable signaling profiles. To highlight innate immune genes, we plotted those genes either that were predicted to have IRF7 binding sites within their promoters (Fig. 6C) or that mapped to the reactome interferon alpha/beta signaling pathway (R-HAS-909733) (Fig. 6D) to specifically examine genes known to be involved in the interferon response. Again, we found highly comparable levels of activation of genes responsible for the innate immune profiles by the three compounds. These results indicate that KIN1408 and KIN1409 activate the same antiviral gene signatures as the parental compound, KIN1400.

FIG 6.

Genomics analysis of KIN1400, KIN1408, and KIN1409 treatment. PMA-differentiated THP-1 cells were treated with 10, 2.5, and 0.625 μM compounds KIN1400, KIN1408, and KIN1409 for 20 h. (A) Heat map of the union of differentially expressed genes across treatments. Differential gene expression was defined as at least a 2-fold change in expression and a Benjamini-Hochberg-corrected P value of <0.01 compared to the result for the appropriate negative control (XRNA for pU/UC RNA and DMSO for all other samples). Expression levels not meeting the cutoff thresholds were set to zero for the visual identification of differential expression. Gene clusters were identified by hierarchical clustering using the Spearman correlation as a distance measure and then classified by the most highly enriched gene ontology biological process meeting a Benjamini-Hochberg-corrected P value of <0.05. (B) Heat map of genes whose promoters are predicted by the UCSC Genome Browser database to contain IRF7 binding sites. (C) Heat map of genes mapping to the Reactome interferon alpha/beta signaling pathway (R-HAS-909733). (D) Two-dimensional principal-component analysis shows patterns among the gene expression profiles across SeV, IFN-β, HCV pU/UC RNA, KIN1400, KIN1408, and KIN1409 treatments. The microarray data in this figure are accessible through GEO series accession number GSE74047 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE74047). (E) PMA-differentiated THP-1 cells were treated with 0.5% DMSO or each compound at a of dose of 1.25, 5, or 20 μM or infected with SeV at 25 HAU/ml for 20 h. Total cellular RNA was extracted and evaluated by RT-PCR for the levels of expression of innate immune genes, the IFN gene, and the IL-6 gene relative to the level of expression of the GAPDH gene. Shown here is the fold induction of the indicated genes compared to that for the DMSO-treated control. Statistical significance was determined by two-way analysis of variance with Dunnett's multiple-comparison test. ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05. 1400, 1408, and 1409, KIN1400, KIN1408, and KIN1409, respectively.

Two-dimensional principal-component analysis was used to identify the patterns of the gene expression profiles among the treatments (Fig. 6D). The expression profiles achieved with the KIN1400 family of compounds showed a dose-dependent change across the three doses tested, and the three compounds grouped together at each dose, demonstrating that these compounds induce very similar patterns of gene induction across their dose ranges. Furthermore, the expression profile of cells treated with the KIN compounds was distinct from that of cells treated with IFN-β, transfected with HCV pU/UC RNA, or infected with SeV, indicating the ability to induce the novel manipulation of the RLR innate immune-induced transcriptional response via these small molecule compounds. The results of gene expression profiling by microarray analysis were validated by real-time qPCR, which confirmed the KIN compound induction of innate immune gene expression, in particular, the genes for Mx1, RIG-I (DDX58), IFIT1, IFIT2, and IFITM1 (Fig. 6E). Dose for dose, KIN1408 induced the expression of the innate immune genes to levels equivalent to those induced by the KIN1400 parent and, in addition, the weak but detectable expression of IFN-β and IFN-λ2/3. Taken together, our data suggest that these hydroxyquinoline small molecules activate RLR signaling to inhibit infection with members of the Flaviviridae in a manner distinct from that of IFN treatment or PAMP RNA, which activate RIG-I signaling.

KIN1400 analogs exhibit broad-spectrum antiviral activity.

To assess whether the KIN compounds can control other pathogenic RNA viruses, we evaluated the KIN1400 parent and analogs KIN1408 and KIN1409 for their ability to suppress IAV (a member of the Orthomyxoviridae) or RSV (a member of the Paramyxoviridae) in cultured cells. A titration of KIN1400, KIN1408, and KIN1409 on IAV-infected cells (MOI, 0.1) demonstrated ∼1.5- to 2-log-unit decreases in the numbers of infectious viral particles produced at 24 h postinfection (Fig. 7A). Titrations of KIN1400 and its analogs were additionally tested for their ability to control RSV-infected HeLa cells (MOI, 0.1). We showed a 1-log-unit decrease in the numbers of infectious virus particles produced in the supernatant collected at 48 h postinfection when cells were treated with KIN1400, KIN1408, or KIN1409 at a concentration of 3 μM or higher (Fig. 7B).

FIG 7.

Broad-spectrum antiviral effect of KIN1400 analogs. (A) HEK293 cells were infected with influenza A virus H3N2 strain A/Udorn/72 at an MOI of 0.1 and then treated postinfection with the indicated compound or DMSO. The supernatants were collect at 24 h, titers were determined on MDCK cells. The MDCK cells were immunostained with FITC-coupled influenza A virus NP monoclonal antibody, and the numbers of focus-forming units per milliliter were measured on an ArrayScan fluorescent microscope imager. ***, P < 0.001. (B) HeLa cells were infected with RSV ATCC VR-1540 (A2 strain) at an MOI of 0.1, followed by postinfection treatment with either the indicated compound or 0.5% DMSO. The supernatants were collected at 48 h, titers were determined on HeLa cells, the HeLa cells were immunostained with antibody to the 47- to 49-kDa fusion protein of RSV as the primary antibody and goat anti-mouse FITC-conjugated secondary antibody, and the number of focus-forming units per milliliter was measured on the ArrayScan fluorescent microscope imager. Statistical significance was tested by two-way analysis of variance with Bonferroni posttests. (C) Antiviral activity of KIN1408 against EBOV. Normal primary HUVECs were treated with 5 and 1 μM KIN1408 or culture medium for 22 h and infected with EBOV (MOI, 0.5) for 1 h, and then EBOV was replaced with medium containing the indicated compound. The cell culture supernatants were collected at 96 h after infection and clarified by centrifugation, and the number of infectious virus particles was measured by plaque assay on Vero cells. The results shown are the average number of infectious virus particles per milliliter of cell culture supernatant calculated at 96 h and were analyzed for statistical significance by two-way analysis of variance with Bonferroni posttests. ***, P < 0.001. (D) Normal primary HUVECs were treated with 5 and 1 μM KIN1408 or culture medium for 22 h and infected with NiV (MOI, 0.1) for 1 h, and then NiV was replaced with medium containing the compound. The cell culture supernatants were collected at 24 and 48 h after infection, and the number of infectious virus particles was measured by plaque assay on Vero cells. The results shown are the average number of infectious virus particles per milliliter cell culture supernatant calculated at 24 and 48 h and were analyzed for statistical significance by two-way analysis of variance with Bonferroni posttests. *, P < 0.05; **, P < 0.01. (E) Normal primary HUVECs were treated with 5 and 1 μM KIN1408 or culture medium for 22 h and infected with LASV (MOI, 0.01) for 1 h, and then LASV was replaced with medium containing the compound. The cell culture supernatants were collected at 24 and 48 h after infection, and the number of infectious virus particles was measured by plaque assay on Vero cells. The results shown are the average number of infectious virus particles per milliliter of cell culture supernatant calculated at 24 and 48 h and were analyzed for statistical significance by two-way analysis of variance with Bonferroni posttests. *, P < 0.05; **, P < 0.01.

We additionally evaluated KIN1408 for its ability to suppress infections caused by several emerging viruses that cause severe to fatal disease in humans for which vaccines or treatments are not available and that are therefore classified at biosafety level 4: EBOV strain Zaire (strain Kikwit 199510621), Nipah virus (NiV; strain Malaysia 199901924), and Lassa virus (LASV; strain Josiah 057562) in cultured cells. Human umbilical vein endothelial cells (HUVECs) were pretreated with 0.5% DMSO or KIN1408 at 1 or 5 μM for 24 h before infection with EBOV at an MOI of 0.5. At 1 h postinfection, the virus inoculum was removed and replaced with medium supplemented with DMSO or KIN1408, which was kept on the cells for the remainder of the experiment. Treatment of cells with 5 μM KIN1408 was sufficient to cause a 1.5-log-unit decrease in the number of infectious virus particles compared to that achieved with treatment with DMSO at 96 h postinfection (Fig. 7C). HUVECs were similarly pretreated with 0.5% DMSO or KIN1408 at 1 or 5 μM for 24 h before infection with NiV or LASV at an MOI of 0.1. At 1 h postinfection, the virus inoculum was removed and replaced with medium supplemented with DMSO or KIN1408, which was kept on the cells for the remainder of the experiment. Treatment of the cells with 5 μM KIN1408 was sufficient to cause a 1.5-log-unit decrease in the number of NiV infectious virus particles produced at 24 and 48 h postinfection compared to the number achieved by treatment with DMSO (Fig. 7D). In contrast, treatment of cells with 5 μM KIN1408 was sufficient to cause a 4.5-log-unit decrease in the number of LASV infectious virus particles produced at 48 h postinfection compared to the number achieved by treatment with DMSO (Fig. 7E). Together, these data demonstrate that, in addition to numerous members of the Flaviviridae, the KIN1400 family of compounds can also suppress infections caused by members of the families Filoviridae (EBOV), Orthomyxoviridae (IAV), Arenaviridae (LASV), and Paramyxoviridae (NiV and RSV) in cultured cells, suggesting the strong potential for this class of molecules to be novel broad-spectrum antiviral agents that activate IRF3-dependent innate immune responses to control infections caused by different RNA viruses.

DISCUSSION

Here, we have identified from a cell-based screen (described in reference 15) a unique antiviral mechanism of action of hydroxyquinoline compounds composed of KIN1400 and derivatives that signal through MAVS to activate IRF3 and drive antiviral gene expression. These compounds induce the expression of cytokines, chemokines, and innate immune effector genes that are known to directly suppress virus infection and spread, including RIG-I, MDA5, IRF7, IFIT1, IFIT2, Mx1, OAS3, and IFITM1. The innate immune genes that are induced by the KIN1400 family of compounds function to disrupt various stages of the virus life cycle. For example, the OAS genes encode a family of 2′-5′-oligoadenylate synthetase enzymes that bind and degrade viral RNA (32, 33). The IFIT genes encode a family of interferon-inducible proteins with tetratricopeptide repeats that have been shown to disrupt viral replication and translation initiation (30, 34, 35). The Mx1 gene encodes an interferon-inducible GTPase that binds to and inactivates the nucleocapsids of RNA viruses (36). Mx1 has also been shown to disrupt the intracellular trafficking of virus particles, ribonucleocapsids, and viral nucleoproteins, thus inhibiting processes of viral entry, genome amplification, and the budding and egress of progeny viruses (37, 38). Signaling that is initiated by RIG-I and OAS can further induce the secretion of a plethora of interferons and cytokines that act in an autocrine and paracrine fashion to establish an antiviral state in cells, thus controlling viral replication and spread (39). Importantly, we demonstrate that the KIN1400 compounds inhibit the replication of a broad spectrum of viruses from the families Flaviviridae, Filoviridae, Paramyxoviridae, Arenaviridae, and Orthomyxoviridae. Thus, we characterized the KIN1400 compounds as a family of drug-like small molecules that induce a novel multicomponent antiviral response to offer a strategy for the development of broad-spectrum antiviral therapy.

In our studies, the KIN1400 compounds demonstrated the highest efficacy when they were administered prior to infection. KIN1400 derivatives KIN1408 and KIN1409 retained potency similar to that of the KIN1400 parent and exhibited better solubility and medicinal chemistry properties than the KIN1400 parent. Importantly, the two derivatives were able to inhibit viruses from numerous families even when they were administered after infection, suggesting that the drug properties of the KIN1400 compounds can be improved with additional medicinal chemistry efforts. Thus, the KIN1400 compounds have great potential for use in both prophylactic and therapeutic settings, and in cases of chronic infection, they may enhance the host innate immune response to further reduce the viral load, especially when they are used in combination with other antiviral agents.

KIN1400 was first identified from a library of drug-like small molecules that induced IFIT2 promoter activation in cell-based assays (described in reference 15). Many PRRs that detect RNA viruses, including RLRs and Toll-like receptors, induce the downstream signaling of IRF3 activation to drive IFIT2 induction. IRF3 is a transcription factor essential for host defense and is rapidly activated upon virus infection to drive the antiviral response, which includes many genes. Within PRR pathways, many factors converge on IRF3, including the TBK1 and IKKε protein kinases, the MAVS adaptor protein, and activation cofactors like CBP/p300 (9). Because signaling by KIN1400 is dependent on MAVS and IRF3, our results suggest that the KIN1400 compounds likely target factors at or above the level of MAVS in the RLR signaling pathway to drive IRF3 activation. Mueller et al. previously reported the identification of small molecule inhibitors of WNV and DV proteases (40), some of which resemble the KIN1400 hydroxyquinolines. We have evaluated the effect of KIN1400 on viral proteases and did not observe drug-mediated inhibition of protease activity. Our observations showing broad-spectrum antiviral activity strongly suggest that the compounds induce RLR-dependent signaling of the innate immune response to control virus infections. We note that we have identified additional compound structures that also activate IRF3 activation, including a benzothiazole series (KIN1000) and the previously reported isoflavone series (15). Gene expression analyses and assessments of the downstream gene activation events comparing all three compound series suggest that although these compounds commonly activate IRF3, their mechanisms of action are distinct and may reflect the action of the compounds on distinct cellular targets (15) (data not shown). Work is in progress to define the cellular target(s) of all three compound series that impart IRF3 activation.

Our studies strongly demonstrate that we can target the activation of IRF3-dependent innate immune signaling mechanisms to control virus infection and that this can be an effective strategy toward the design of new antiviral therapies. Our analyses show that the hydroxyquinolines induce gene expression patterns that are distinct from those obtained with IFN treatment and poly(U/UC) RNA transfection. We note further that the hydroxyquinolines induce low levels or no detectable levels of IFN expression, suggesting that these compounds would have an antiviral action distinct from that of traditional IFN treatments that, though effective, impart many undesirable side effects. Such compounds would have a distinct advantage over RNA PAMP therapies, which are beleaguered by issues of stability and delivery, and DAA compounds. DAAs exert strong selective pressures during therapy that result in the accumulation of mutations in the viral genome that over time cause the emergence of resistant strains that eventually render the agent ineffective (41–47). In contrast, agents that target the host immune response are predicted to have broader-spectrum antiviral activity, and because they induce a range of genes that offer pleiotropic actions to suppress virus infection, the chance of generating drug-resistant virus is nearly nonexistent. Our study shows that innate immune agonists can direct the control of infections caused by viruses of different genera and families, as shown here by KIN1400 and other similarly acting agents. The hydroxyquinolines, including KIN1400, are currently undergoing focused chemical optimization to improve aqueous solubility, metabolic and chemical stability, as well as oral bioavailability. Lead hydroxyquinolines with improved drug-like properties will be developed as broad-spectrum oral therapeutics for a variety of RNA viral pathogens.

In summary, we have identified in KIN1400 and its derivatives a new class of small molecules that can induce IRF-3-dependent innate immune signaling to control RNA virus infections.