The Antiviral Drug Arbidol Inhibits Zika Virus

There are many emerging and re-emerging globally prevalent viruses for which there are no licensed vaccines or antiviral medicines. Arbidol (ARB, umifenovir), used clinically for decades in several countries as an anti-influenza virus drug, inhibits many other viruses. In the current study, we show that ARB inhibits six different isolates of Zika virus (ZIKV), including African and Asian lineage viruses in multiple cell lines and primary human vaginal and cervical epithelial cells. ARB protects against ZIKV-induced cytopathic effects. Time of addition studies indicate that ARB is most effective at suppressing ZIKV when added to cells prior to infection. Moreover, ARB inhibits pseudoviruses expressing the ZIKV Envelope glycoprotein. Thus, ARB, a broadly acting anti-viral agent with a well-established safety profile, inhibits ZIKV, likely by blocking viral entry.

Measuring Zika Virus Infection. For infection of Huh7.5.1 cells, Zika virus protein expression was measured by Western blot analysis using the LI-COR Odyssey CLx imaging system. Rabbit antibodies to ZIKV E (GTX133325), NS5 (GTX133312), Capsid (GTX133317), or NS1 (GTX133304) proteins, or mouse anti-ZIKV-NS1 (GTX634159) were obtained from GeneTex. Goat-anti-Actin (sc-1616), mouse anti-vinculin (sc-73614), or mouse anti-cofilin E8 (sc-376476) antiserum (all from Santa Cruz Biotechnology) were also used to detect cellular proteins and confirm equal protein loading across all samples. Secondary antibodies to the goat, mouse, or rabbit primary antibodies were labeled with infrared dyes, permitting detection with the Odyssey CLx. Dyes designated as 800 produce green emission, while dyes designated 680 produce red emission. The secondary antibodies used were donkey anti-goat DyLight680 or donkey anti-goat DyLight800 (SA5-10090, SA5-10092, Fisher), donkey anti-rabbit DyLight680 (SA-510042, Fisher), goat anti-rabbit DyLight800 (SA-510036, Fisher), goat anti-mouse IRDye680RD (925-68070, LI-COR), and goat anti-mouse IRDye800CW (925-32210, LI-COR). For some blots, combinations of antibodies were used to facilitate multiplex detection of viral and cellular proteins. Blots were stripped using NewBlot Nitro Stripping Buffer (928-40030, LI-COR) following manufacturer's instructions. For all Western blots, Image Studio (LI-COR) software was used to obtain Western blot images, using default instrument settings of resolution at 169 μm and scan quality set to lowest. Any image manipulations were manually applied equally across the entire image and were applied equally to controls. The manipulations consisted of adjusting the brightness and contrast and/or flipping the image to obtain the proper orientation. The original Western blot images are included in the Supplementary Information. Using Image Studio software, rectangles of identical size and area were drawn around viral and cellular protein bands to obtain pixel intensities, which were exported into excel. All protein bands were normalized to cellular protein levels. ARB-treated samples were then normalized to the ethanol solvent control. Final pixel intensities were expressed as percent of ethanol control by first dividing the actin-normalized ARB treated sample pixel intensity by the actin-normalized ethanol solvent control pixel intensity. This fraction was then multiplied by 100.
For infection of A549 cells, ZIKV protein expression was measured by immunostaining and quantification of an HRP-conjugated secondary antibody using TMB substrate by measuring absorbance at 650 nm 28 . Cells were pretreated with ARB with four-fold serial dilution (40-0.001 μM) for one hour, followed by infection with ZIKV MR766 at an MOI of 1 in the presence of drug. Two-hours later, viral inocula were removed and fresh medium containing drug was added back to cell cultures. Forty-eight hours post-infection, the cells were fixed with ice-cold methanol, and washed with assay buffer (PBS with 2% nonfat milk and 0.1% Triton X-100). The cells were then incubated with anti-flavivirus group antigen E antibody (1:4000 dilution, clone D1-4G2-4-15, Sigma) for two hours at room temperature. Cells were washed three times with assay buffer and incubated with anti-mouse horse radish peroxidase (HRP) conjugated secondary antibody (1:4000 dilution, in assay buffer) for one hour at room temperature. The cells were further washed three times with assay buffer and then incubated with TMB (100 μl) (Rockland Immunochemicals, PA) for 30 min at room temperature. TMB is the substrate for HRP, which converts TMB to a blue color. The intensity of the color is proportional to the amount of viral antigen and was measured by spectrometry at 650 nm using EnVision (Perkin Elmer).
For ZIKV infection of primary vaginal and cervical epithelial cells, Western blots were used to measure ZIKV proteins as described above. Progeny virus production was also measured by plaque forming unit (PFU) assays after diluting supernatants at least 1:100. ZIKV RNA genome copies were also quantified by digital droplet RT-PCR with a Biorad QX200 system. RNA from cell cultures was extracted using the Qiagen RNeasy kit according to manufacturer's instructions. An equivalent amount of RNA for all conditions was reverse transcribed using random hexamers, diluted 1:5, and used as template for digital droplet PCR using Biorad ddPCR supermix for probes, according to manufacturer's instructions. Copies of ZIKV RNA were normalized to the cellular housekeeping gene RPP30. PCR primers and probe for ZIKV were ZKV F (5p-CCGCTGCCCAACACAAG), ZKV R (5p-CCACTAACGTTCTTTTGCAGACAT), ZKV probe (5p-AGCCTACCTTGACAAGCAGTCAGACACTCAA).
Zika pseudovirus particles. ZIKV pseudovirus was made by harvesting culture supernatant of HEK-293T cells co-transfected with a DNA-launched WNV replicon expressing EGFP 33 (kindly provided by Theodore Pierson, National Institute of Allergies and Infectious Diseases, MD, USA) and a plasmid expressing ZIKV C-PrM-E (kindly provided by Oscar Burrone and Jose Luis Slon Campos, International Centre for Genetic Engineering and Biotechnology, Trieste, Italy) as described 34 . Vero cells were pretreated with ARB for 2 h prior to pseudovirus infection, harvested 24 h after infection, fixed with paraformaldehyde and analyzed by flow cytometry on a BD LSR Fortessa. EGFP data were derived from gating on live cells, based on forward/side scatter characteristics. The percentage of dead cells was comparable between all samples and generally less than 10%. Cytoxicity Testing. Cytotoxicity of ARB on Huh7.5.1, Vero, and primary human vaginal or cervical epithelial cells was evaluated by measuring cellular ATP levels with a commercial kit (ATPlite assay, Perkin Elmer). Cytotoxicity of ARB on A549 cells was measured using the CellTiter-Glo reagent (Promega). Figure 1 depicts the structure of ARB, an indole-based compound. Figure 2 demonstrates that ARB inhibits the infection of Vero cells by the Ugandan (MR766) isolate of ZIKV, as indicated by decreased viral protein synthesis. Asian lineage virus isolates circulating in the world including Puerto Rican (PRVABC59), Brazilian (KX811222.1), Cambodian (FSS13025), Panamanian (PA259249), and Mexican (MEX2-81) isolates were also inhibited by ARB (Supplemental Figure S1).

Results
We also measured the cytotoxicity profile of ARB on Vero cells in the presence and absence of ZIKV infection. Zika infection is cytolytic to Vero cells and killed many cells after a 3-day infection. When the magnitude of cellular ATP reduction in ARB-treated cells was compared to cells treated with ethanol (the solvent for ARB), ARB treatment prevented the reduction in ATP levels (i.e. cell death) during Zika virus infection in Vero cells (Fig. 3A). Microscopically, Zika virus-infected cultures had clear evidence of reduced cell numbers and cytopathic effect (CPE; Fig. 3B, middle panel) as compared to non-infected (i.e. mock) cultures (Fig. 3B, left panel). ARB treatment protected against virus-induced CPE (Fig. 3B, right panel). The data suggest that ARB partially protects cells from ZIKV-induced cell death. These results were not unique to Vero cells as ARB-treated human hepatoma Huh7.5.1 cells were also protected from ZIKV-induced CPE and exhibited reduced ZIKV protein production (Supplemental Figure S2).
We further tested ARB against ZIKV infection of human A549 cells and asked whether overnight pre-treatment with ARB provides better antiviral effects against ZIKV as compared to when ARB is added 1 hour before infection. Figure 4 demonstrates that ARB inhibits ZIKV at a concentration that inhibits ZIKV infection by 50 percent, IC 50  is known to block entry of HCV and EBOV 16,23 , we performed time of addition experiments (Fig. 5). Addition of ARB to cells 24 hours before ZIKV infection caused the greatest suppression of infection, while adding ARB 1 hour before, at the same time as, and 1 hour after infection also robustly suppressed infection. Addition of ARB to cells at 24 hours post-infection was not as potent as the other times of ARB addition, but still suppressed virus infection by about 50%. The data indicate that ARB is most effective at suppressing ZIKV infection when cells are pretreated with the drug, suggesting that ARB inhibits an early step in the ZIKV lifecycle. In addition, ARB might also impact later steps in the virus life cycle.  To focus on virus entry mediated by the ZIKV Envelope glycoprotein, we generated pseudovirus particles with Zika virus Envelope and Capsid packaging a West Nile virus replicon encoding an EGFP reporter 34 . Infection of Vero cells with the pseduovirus particles yielded EGFP positive cells, and entry was blocked by Bafilomycin A1, a known inhibitor of endosomal acidification (Fig. 6A). The infectivity of these particles on Vero cells was prevented in a dose-dependent manner by pretreatment of cells with ARB (Fig. 6B). ARB was most effective when the drug was present throughout the infection process, including during pre-treatment of cells, during virus inoculation, and when the inoculum was removed and replaced with fresh medium (Fig. 6C, ARB treated sample labeled as "pre-Rx"). In contrast, ARB was less effective at suppressing pseudovirus infection, when the drug was added to cells after the virus inoculum was removed (Fig. 6C, ARB treated sample labeled as "post-Rx"). These data reinforce the concept that ARB blocks an early step in the ZIKV life cycle, and may also have post-entry effects.
ARB efficacy in primary cell lines was evaluated by infecting primary human vaginal and cervical epithelial cells with ZIKV MR766 in the presence and absence of ARB pretreatment. Figure 7A shows that primary vaginal epithelial cells (HVE2) and primary endocervical (ENDO) and ectocervical (ECTO) cells were robustly infected  by ZIKV. ARB treatment (20 μM) resulted in significant suppression of ZIKV protein synthesis (Fig. 7A,B). The yield of infectious virus from ARB-treated cells that were infected by ZIKV was reduced by 10, 10, and 100 fold in HVE2, ENDO, and ECTO cells, respectively (Fig. 7C). ARB also inhibited infection of cells by ZIKV strain 1848, an Asian lineage isolate (Fig. 7D). Critically, the dose of ARB used in these studies (20 μM) was not toxic to either cell type (Fig. 7E). Finally, ZIKV RNA expression in primary vaginal and cervical epithelial cells from different donors infected by MR766 and 1848 virus isolates was also suppressed by ARB (Supplemental Figure S3).

Discussion
We show herein that ARB caused dose-dependent inhibition of multiple ZIKV isolates of both the African and Asian lineages in multiple cell lines, including primary human vaginal and cervical epithelial cells. ARB protected against ZIKV-induced CPE, and multiple lines of data suggest that ARB potentially blocks multiple steps in the ZIKV lifecycle, with a major effect on virus entry into cells.
How does ARB inhibit so many enveloped viruses? ARB is an indole-based molecule ( Fig. 1) able to form supramolecular arrangements through aromatic stacking interactions with selective amino-acid residues of proteins (phenylalanine, tyrosine, tryptophan) 22 . As such, ARB may impair several steps in the life cycle of viruses including virus attachment to cells, fusion of viral and cellular membranes during virus entry 17,21,22,[35][36][37] , clathrin-mediated endocytosis 16 , or virus replication on intracellular membranes, such as membranous webs 20,38 . However, the bulk of the data coalesce on a model whereby ARB targets viral glycoproteins to prevent the fusion of viral membranes with endosomal membranes during virus entry: (1) the crystal structure of ARB bound to the Influenza virus hemagglutinin (HA) fusion protein shows that ARB prevents HA membrane fusion by blocking low pH-induced conformational changes in the protein 39 ; (2) ARB inhibits HCV glycoprotein-mediated fusion of viral membranes with model endosomal membranes 11,19-22 ; (3) ARB inhibits entry of EBOV and ZIKV pseudoparticles mediated by EBOV-GP 23 and ZIKV-E viral fusion proteins (this report); (4) ARB resistant viruses selected in vitro have mutations in viral fusion proteins 17,40 ; (5) several drugs were recently found to bind to a pocket in the native EBOV GP 41 . Clearly, additional work is required on this interesting compound, including structure-guided optimization of ARB against specific virus families 42 .
Clinical success of ARB for ZIKV infection will require sufficient drug levels in target cells and tissues, without toxicity. The recommended oral dose of ARB for treatment of influenza in humans is 200 mg three times daily 43 , but single doses of up to 800 mg have been administered without adverse effects 44 . The maximal plasma concentration (C max ) and area under the concentration-time curve (AUC) after the standard single ARB dose in humans is approximately 0.9-1.5 µM and 4-6 µM/hr [43][44][45] , which is in the range of in vitro antiviral activity of ARB against many viruses including influenza virus, EBOV, ZIKV, and HCV. Granted, some viruses show inhibition by ARB in vitro at doses very close to the plasma exposure (EBOV, IC 50 = 2.7 µM 23 ), while some viruses showing inhibition at higher in vitro doses (ZIKV in A549 cells, IC 50 = 11 µM; this report). It is possible that the dose of ARB for inhibiting a particular virus depends primarily on how tightly the compound engages viral glycoproteins 42 . An 800 mg dose elicits a C max and AUC of 4 µM and 24 µM/hr 44 , which is well within the range of ARB inhibition of most viruses in vitro. PK studies in humans indicate plasma C max of ARB is reached within 60-90 minutes, with a half-life (t 1/2 ) of 17-21 hours in Russian subjects 11 , with possible faster clearance in Chinese subjects (t 1/2 = 6-7 hours) in some 45,46 , but not all studies 44 . In mouse studies, doses up to 600 mg/kg/day have been studied without toxicity 13,[47][48][49] , and doses of 90 and 180 mg/kg/day show significant suppression of Influenza H1N1 viral loads, pathology, and mortality 50 . Moreover, allometric scaling 51 provides non-linear extrapolation of the recommended daily human ARB dose (600 mg) to an oral mouse dose of 120 mg/kg, which is well within the tolerability and efficacious doses of ARB. ARB appears to be safe and without teratogenic effects in pregnant women 52,53 . Thus, in many human and animal studies, ARB has a reasonable pharmacokinetic profile and is well tolerated (reviewed in 11 ). Sexual transmission of ZIKV is well established 10,54 . It is also plausible to speculate that sexually transmitted virus may increase the risk of infection of the placenta and fetus in pregnant women, as compared to mosquito borne transmission. Therefore, infection and replication in the female genital tract are important to consider during assessment of any potential anti-ZIKV therapeutics. We show here that ARB strongly impairs ZIKV in primary cells of the female genital tract and reduces the production of progeny virions. This suggests that ARB would have high activity in cells likely to be targeted during sexual transmission of ZIKV and may be able to prevent local viral replication and systemic spread.
ARB is one example of a drug that should be considered for drug repurposing for virus outbreaks against ZIKV and Ebola virus. Other studies aimed at repurposing of existing drugs have found that bortezomib and mycophenolic acid (MPA) strongly inhibit ZIKV 55 . However, these drugs are immunosuppressive and potentially teratogenic 55 , so their utility in treating ZIKV infection in pregnant women is untenable. Since ARB appears to be safe for pregnant women 52,53 , repurposing the drug with other anti-ZIKV drugs may provide urgently needed options for clinical management of ZIKV disease and its consequences on human health. The image is a composite of the following antibody probings: a blot was first probed with rabbit anti-ZIKV-NS1 (detected with anti-rabbit 800 labeled secondary antibodies) and mouse anti-vinculin antiserum (detected with anti-mouse 680 labeled secondary antibodies). A separate blot of the same protein samples was probed with rabbit anti-ZIKV-E (detected with anti-rabbit 680 labeled secondary antibodies). The same blot was then stripped and reprobed with mouse anti-vinculin antiserum (detected with anti-mouse 680 labeled secondary antibodies) and rabbit anti-ZIKV-Capsid (detected with anti-rabbit 800 labeled secondary antibodies). (B) Quantitation of ZIKV protein intensity in ethanol or ARB-treated cells. (C) Infectious virus production by cells treated with ethanol versus ARB. Culture supernatants were harvested at 72 hours post infection, diluted 1:10,000, and titered on Vero cells. Results are expressed as plaque forming units per ml (PFU/ ml). (D) ARB inhibits infection of primary human vaginal and cervical epithelial cells by ZIKV strain 1848, an Asian lineage isolate. Cells were infected and treated as in panel A. The image is a composite of the following antibody probings: a blot was first probed with rabbit anti-ZIKV-E (detected with anti-rabbit 800 labeled secondary antibodies) and mouse anti-vinculin antiserum (detected with anti-mouse 680 labeled secondary antibodies). The blot was then stripped and reprobed with rabbit anti-ZIKV-Capsid antiserum (detected with anti-rabbit 800 labeled secondary antibodies). Expression of ZIKV E and Capsid proteins are denoted, as well as expression of the cellular protein Vinculin. (E) Cytotoxicity profile of ARB in primary human vaginal and ectocervical cells. Cells were treated with ethanol (EtOH) or 1, 5, 10, 20, 30, 60 μM ARB for 72 hours before viability was measured using ATPlite assay. Numbers represent fold change relative to ethanol treated triplicate cultures. Error bars represent standard deviation. Original blot images are shown in Supplemental Information.