Obatoclax, Saliphenylhalamide, and Gemcitabine Inhibit Influenza A Virus Infection*

Background: Novel options should be developed for treatment of IAV infections. Results: Obatoclax, saliphenylhalamide, and gemcitabine target host factors and inhibit IAV and several other viruses infections. Conclusion: These compounds represent potent antiviral agents. Significance: These compounds could be exploited in treatment of severe viral infections. Influenza A viruses (IAVs) infect humans and cause significant morbidity and mortality. Different treatment options have been developed; however, these were insufficient during recent IAV outbreaks. Here, we conducted a targeted chemical screen in human nonmalignant cells to validate known and search for novel host-directed antivirals. The screen validated saliphenylhalamide (SaliPhe) and identified two novel anti-IAV agents, obatoclax and gemcitabine. Further experiments demonstrated that Mcl-1 (target of obatoclax) provides a novel host target for IAV treatment. Moreover, we showed that obatoclax and SaliPhe inhibited IAV uptake and gemcitabine suppressed viral RNA transcription and replication. These compounds possess broad spectrum antiviral activity, although their antiviral efficacies were virus-, cell type-, and species-specific. Altogether, our results suggest that phase II obatoclax, investigational SaliPhe, and FDA/EMEA-approved gemcitabine represent potent antiviral agents.

Influenza A viruses (IAVs) 2 are enveloped viruses with segmented negative-sense single-stranded RNA genome. The rapid accumulation of mutations in IAV genomes enables the emerging viruses to evade the immunity developed after vaccination or previous infections with IAV and cause yearly epidemics and major pandemics with high morbidity and a great number of severe and fatal cases.
Human IAVs invade the respiratory tract causing acute respiratory infections. Ocular involvement is also common. IAVs replicate in lung and retinal epithelial cells as well as in dendritic cells, type II pneumocytes, and alveolar macrophages (1)(2)(3). The replication cycle starts when the hemagglutinin (HA) proteins of the IAV virion recognize cell surface receptors. Then the IAV particle is taken up and transported via the endocytic pathway. Acidification of the endosome by cellular vacuolar ATPase (vATPase) and the virion interior by viral matrix protein 2 (M2) triggers HA-mediated fusion of viral and endosomal membranes, and the uncoating of viral ribonucleoproteins (vRNPs) from the lipid envelope and matrix protein M1 shell. The vRNPs that contain PB1, PB2, and PA polymerase subunits, nucleoprotein (NP), and genomic viral RNAs (vRNAs) are imported to the nucleus, from which the viral genes direct the production of new viral components. These assemble into new virus particles at the plasma membrane which bud and release from the cell to infect other cells.
To ensure efficient production of infectious viral particles, IAV subverts endocytosis, transcription, translation, apoptosis, and other cellular functions (4). These functions could be perturbed by small molecules to inhibit IAV infection. Here, we evaluated a set of small molecules that affect cellular functions implicated in IAV infection to find novel and validate known IAV inhibitors, to provide novel insight into the biological process of IAV-host cell interaction, and to establish a foundation for the development of host-directed antivirals. These antivirals could overcome the problem of emerging resistance to available antiinfluenza drugs that target viral neuraminidase and M2. Table 1 summarizes compounds used in this study and their suppliers. Saliphenylhalamide (SaliPhe) was synthesized by Jef De Brabander's group as described in Ref. 5. 10 mM compounds in 100% dimethyl sulfoxide (Sigma-Aldrich) were stored at Ϫ80°C until use.
Compound Validation against WT IAV Strains-Selected compounds were further validated against PR8 and WSN IAV strains in RPE cells or in human PBMC-derived macrophages. Efficacy and cytotoxicity testing were performed in 96-well plate format. Typically, 40,000 RPE cells or 600,000 PBMCs were seeded per well and grown for 24 h or 7 days, respectively, in appropriate cell growth media. RPE cells and PBMC-derived macrophages were washed twice with phosphate-buffered saline (PBS) and covered with appropriate VGM. The compounds were added to the cells in 3-fold dilutions at seven different concentrations starting from 10 M. No compounds were added to the control wells. The cells were infected with the viruses at m.o.i. 3 or 10, or noninfected. When a virus-induced cytopathic effect was observed, cell viability was measured using a CTG assay.
For each virus-compound pair we calculated the quantitative activity and toxicity score (ATS), where A is the maximal antiviral activity (with a virus), T is the toxic side effects (without a virus), M is the concentration level of maximal activity (over the whole dose range), and B refers to the base-line concentration level (i.e. the lowest dose). In case the maximal activity A M is Ͼ100%, we set A M ϭ 100, and if M ϭ B (i.e. the maximal activity is reached already at the lowest dose), we set ATS ϭ 0. Using these constraints, ATS varies between Ϫ100 and ϩ100, where negative values indicate excessive toxicity and highest positive values indicate most potent compounds.
For other viruses the titration procedure slightly differed from the one described above. Echo6 virus was titered on A549 cells, and both VGM and AM contained 0.4 g/ml TPCK-trypsin. SINV, SFV, HSV-1, and VACV were titered on Vero-E6 cells, and VGM was supplemented with 5% FBS, 2 mM L-glutamine, and 50 units/ml PenStrep in DMEM, and TPCK-trypsin was omitted. BUNV was titered on Vero-E6 cells, and supernatants were diluted in PBS containing 2% newborn calf serum (Invitrogen), AM contained 0.6% Avicel, and 2% newborn calf serum in minimal essential medium, and cells were incubated with AM for 3 days. Titers of HSV-1 were determined by infecting 12-well plates of B-Vero cells with serial dilutions of supernatants in DMEM containing 7% heat-inactivated fetal calf serum (FCS; Invitrogen) and 20 g/ml human immunoglobulin G (Baxter). Cells were fixed with methanol for 3 min and stained with 0.1% crystal violet in 2% ethanol. The degree of inhibition mediated by a compound was calculated as a ratio between virus titers in nontreated and compound-treated cells.
Immunofluorescense-Compound-treated or nontreated RPE cells were infected with WSN IAV at m.o.i. 30 on ice for 1 h. Cells were washed twice with ice-cold VGM, overlaid with the media with or without compound, and incubated at 37°C for 1-4 h. Cells were fixed with 4% paraformaldehyde (in PBS). PBS with 1% BSA and 0.1% Triton X-100 was used for blocking and permeabilization of the fixed cells and for dilution of antibodies. NP and M1 of WSN were stained with corresponding rabbit polyclonal antibodies (1:1000; from I. J. laboratory), and the secondary antibody was Alexa Fluor 488 goat anti-rabbit IgG (HϩL) (1:1000, Invitrogen Molecular Probes). Mcl-1 was stained with anti-human MCL-1 (1:100; clone 22/Mcl-1; BD Transduction Laboratories). Secondary antibody was Alexa Fluor 594 goat anti-mouse IgG (1:2000; Invitrogen). Nuclei were counterstained with DAPI. Images were captured with Nikon 90i microscope and processed with NIS Elements AR software. Note, obatoclax produces autofluorescence (absorbance peak, 490 nm; emission peak, 550 nm (17)). Immunoblots

Eight Host-directed Compounds Inhibit Infection of Human
Nonmalignant Cells with Reporter IAV Strain-A total of 201 host-directed chemical compounds have been preselected for this study (supplemental Table S1). These include known IAV inhibitors, such as gemfibrozil and TOFA (lipid metabolism), 3-methyladenine and wortmannin (autophagy), 17-AAG and bortezomib (protein quality control), fluorouracil and cytarabine (de novo nucleotide biosynthesis). Notably, the antiviral efficacy of these compounds has been demonstrated mainly in cancer cell lines. Our drug set also contains functional analogs of known IAV inhibitors, such as erlotinib and gefitinib (analogs of PD153035), 17-DMAG (analog of 17-AAG), carfilzomib (analog of bortezomib), leflunomide and gemcitabine (analog of fluorouracil), and a pool of potential IAV inhibitors that target the cell cycle, apoptotic pathways, and other functions that could be essential for IAV infection of a host cell.
We have tested the set of 201 compounds in nonmalignant human RPE cells at different concentrations against infection with reporter IAV strain (PR8-GFP). This strain expresses the NS1 1-116 -GFP fusion which allows monitoring antiviral efficacy of the compounds by measuring GFP fluorescence and identification of small molecules that target stages of IAV infection ranging from IAV attachment to viral protein synthesis and quality control (9). Compound cytotoxicity testing was performed in parallel to calculate selectivity indexes (Fig. 1A). We identified five novel (obatoclax, gemcitabine, carfilzomib, NVP-AUY922, and 17-DMAG) and validated three known (SaliPhe, bortezomib, and 17-AAG (9, 19, 20)) IAV inhibitors that selectively suppressed IAV-mediated GFP expression at noncytotoxic concentrations in dose-dependent manner in RPE cells (Fig. 1, B and C).
Three Host-directed Compounds Inhibit Infection of Human Nonmalignant Cells with Wild Type IAV Strains-The identified compounds have been demonstrated to target cellular vATPase, proteasome, heat shock protein 90 (HSP90), ribonucleotide reductase, or induced myeloid leukemia cell differentiation protein Mcl-1 (5,17,21,22). Therefore, we next selected representative inhibitors for each of the 5 targets and tested them against two wild type IAV strains (PR8 and WSN) in human RPE cells and PBMC-derived macrophages. The effect of the drugs on cell survival and virus progeny production was examined. Potent inhibitors of Mcl-1, ribonucleotide reductase, and vATPase were found to exhibit dose-dependent antiviral activity which could be expressed in positive ATS. Inhibitors of proteasomes showed a moderate effect, and HSP90 inhibitors were ineffective at the selected concentrations, which could be expressed in neutral or negative ATS, respectively (Fig. 2, A and B). These results suggest that cellular Mcl-1, ribonucleotide reductase, and vATPase could represent essen-tial host factors for wild type IAV infection. The proteasome could also be involved, but the therapeutic window was too narrow to allow a final conclusion. By contrast, HSP90 is not required by wild type, but is exploited by the mutant IAV strain containing an altered NS1, indicating that full-length NS1 could perform HSP90 function during IAV infection.
Virus-, Cell Type-, and Species-specific Antiviral Activity of Three Compounds-We next tested whether obatoclax, gemcitabine, and SaliPhe inhibit seasonal H3N2 IAV and InfB infections. Supplemental Fig. S1 shows that obatoclax and SaliPhe, by contrast to gemcitabine, rescued RPE cells from infections with these viruses. This result indicates that obatoclax and SaliPhe could possess anti-IAV and anti-InfB activity, whereas gemcitabine activity could be limited to certain influenza strains.
We next tested whether these compounds are able to rescue cancer cells from IAV infection. We found that obatoclax and SaliPhe protected human adenocarcinomic alveolar basal epithelial (A549) cells from the challenge with WSN virus and that the antiviral effect of this compound was achieved at lower concentrations than those needed to mediate cancer cell death (supplemental Fig. S1). Interestingly, gemcitabine was unable to rescue IAV-infected A549 cells, suggesting that gemcitabine activity could be cell type-specific. We also tested whether obatoclax, gemcitabine, and SaliPhe possess species-specific antiviral activity and found that obatoclax, by contrast to gemcitabine and SaliPhe, was less effective in MDCK cells than in RPE or A549 ( Fig. 2A and  supplemental Fig. S1).
We next explored whether these compounds display antiviral activity against several other viruses of different families. Compound efficacy and cytotoxicity testing was performed against BUNV, MeV, SINV, SFV, Echo6, HSV-1, and VACV in dedicated cell lines (supplemental Fig. S2). Further, we titered the viruses produced by cells nontreated or treated with effective but noncytotoxic concentration of the compound. Table 1 shows the degrees of inhibition of virus production mediated by a , the compound did not affect virus production and survival of infected cells at indicated concentrations. b ND, not determined because the compound did not protect cells from IAV-mediated death (see Fig. 7).
the compounds. We found that obatoclax and SaliPhe efficiently (Ͼ2 logs reduction in virus titers) attenuated infections of BUNV and SINV, whereas gemcitabine suppressed SINV and HSV-1. Altogether, these results suggest that obatoclax, SaliPhe, and gemcitabine possess virus-, cell type-, and species-specific anti-viral activities, indicating that different viruses could possess differential preferences for cellular targets of obatoclax, SaliPhe, and gemcitabine upon infection.
Obatoclax, SaliPhe, and Gemcitabine Inhibit Virus Entry-A time-of-compound-addition experiment in macrophages demonstrated that obatoclax, gemcitabine, and SaliPhe inhibited early steps of IAV infection (Fig. 3A). An immunofluorescence experiment monitoring the viral NP at 4 hpi revealed that obatoclax and SaliPhe treatment prevented nuclear accumulation of vRNPs which was observed in nontreated and gemcitabine-treated RPE cells (Fig. 3B). Obatoclax, SaliPhe, and gemcitabine treatment affected subsequent vRNA transcription and replication, viral protein synthesis, and infectious viral particle production (Fig. 3, C-E). Thus, these three compounds inhibit entry of IAV into the cells.
Mcl-1 Is a Novel Antiviral Target-Obatoclax is a novel anti-IAV agent, whereas SaliPhe and analogs of gemcitabine have been shown to possess anti-IAV activity (9,23). Therefore, we investigated the possible mechanism of action of obatoclax.
It was demonstrated that obatoclax preferentially binds Mcl-1 (17); however, the Mcl-1 involvement in IAV infection has not been shown. To prove that Mcl-1 is essential for IAV entry we partially silenced Mcl-1 using Mcl-1-specific siRNA (Fig. 4A), infected these cells with PR8-GFP virus and analyzed virus-mediated GFP expression. Silencing of Mcl-1 substan-tially reduced IAV-mediated GFP expression and slightly affected the viability of infected cells (Fig. 4B), suggesting that Mcl-1 is involved in IAV infection and apoptosis. Notably, Mcl-1 is up-regulated during the first hours of IAV infection (Fig. 4C) as with other viral infections (24,25). This result indicates that Mcl-1 could also be involved in host antiviral responses. An immunofluorescence experiment showed that Mcl-1 has a localization pattern similar to that of viral M1 (Fig.  4D), indicating that Mcl-1 could be involved in the recognition of IAVs.
Effect of Obatoclax, SaliPhe, and Gemcitabine on IAV-mediated Cellular Antiviral and Proinflammatory Responses-We analyzed the effect of obatoclax, SaliPhe, and gemcitabine on cytokine/chemokine production by IAV-infected and mockinfected cells. We found that all three compounds suppressed WSN-mediated production of proinflammatory IP10, CXCL1, IL6, IL8, CCL2, and CCL5 in RPE cells (supplemental Fig. S3 and Fig. 5). Surprisingly, obatoclax inhibited, whereas SaliPhe and gemcitabine stimulated, production of antiviral IFN-␤ and IFN- (Fig. 5). These results indicate that obatoclax, in contrast to SaliPhe and gemcitabine, could prevent activation of cellular antiviral responses, most probably, via suppression of Mcl-1type I IFN signaling axis.

DISCUSSION
Here, we developed a systematic approach to search for potential antiviral compounds and to provide a foundation for their further development. Using this approach we evaluated a set of carefully preselected potential and known host-directed antiinfluenza agents. We found that obatoclax, SaliPhe, and gemcitabine inhibited IAV and several other virus infections at noncytotoxic concentrations. Using these small molecules we provided novel insight into IAV-host cell interactions. In particular, the link was established between compound targets and stages of virus infection.
Obatoclax is a phase II anticancer and a novel anti-IAV agent that binds cellular Mcl-1. Silencing of Mcl-1 using Mcl-1-directed siRNA confirmed the importance of the Mcl-1-mediated pathways for efficient IAV infection.
Mcl-1 was up-regulated during IAV infection. This indicates that Mcl-1 could be implicated in host antiviral responses.
Indeed, inhibition of Mcl-1 with obatoclax suppressed the type I IFN responses in IAV-infected cells. This could be achieved via Mcl-1-mediated regulation of STAT3 signaling pathway which controls type I IFN responses (26 -28). Thus, Mcl-1 could be involved in cellular antiviral and apoptosis responses and IAV uptake.
Gemcitabine is an FDA/EMEA-approved anticancer and a novel anti-IAV agent that alters viral RNA transcription and replication. This observation is in line with previously reported findings showing that the analogs of this drug possess anti-IAV activity by targeting ribonucleotide reductase which is essential for de novo pyrimidine biosynthesis (23).
SaliPhe is an investigational anticancer and anti-IAV compound (5,9,29). It inhibits vATPase activity, and similarly to ammonium chloride, it prevents the acidification of endosomes and arrests IAV in the endocytic compartments.
In conclusion, we identified obatoclax, gemcitabine, and SaliPhe as potent inhibitors of virus-host interactions, and these results could provide a foundation for clinical development of these host-directed antivirals. These antivirals could overcome the problem of emerging resistance to available virus-directed anti-influenza drugs.