ABSTRACT
Antiviral therapies that impede virus entry are attractive because they act on the first phase of the infectious cycle. Drugs that target pathways common to multiple viruses are particularly desirable when laboratory-based viral identification may be challenging, e.g., in an outbreak setting. We are interested in identifying drugs that block both Ebola virus (EBOV) and Lassa virus (LASV), two unrelated but highly pathogenic hemorrhagic fever viruses that have caused outbreaks in similar regions in Africa and share features of virus entry: use of cell surface attachment factors, macropinocytosis, endosomal receptors, and low pH to trigger fusion in late endosomes. Toward this goal, we directly compared the potency of eight drugs known to block EBOV entry with their potency as inhibitors of LASV entry. Five drugs (amodiaquine, apilimod, arbidol, niclosamide, and zoniporide) showed roughly equivalent degrees of inhibition of LASV and EBOV glycoprotein (GP)-bearing pseudoviruses; three (clomiphene, sertraline, and toremifene) were more potent against EBOV. We then focused on arbidol, which is licensed abroad as an anti-influenza drug and exhibits activity against a diverse array of clinically relevant viruses. We found that arbidol inhibits infection by authentic LASV, inhibits LASV GP-mediated cell-cell fusion and virus-cell fusion, and, reminiscent of its activity on influenza virus hemagglutinin, stabilizes LASV GP to low-pH exposure. Our findings suggest that arbidol inhibits LASV fusion, which may partly involve blocking conformational changes in LASV GP. We discuss our findings in terms of the potential to develop a drug cocktail that could inhibit both LASV and EBOV.
IMPORTANCE Lassa and Ebola viruses continue to cause severe outbreaks in humans, yet there are only limited therapeutic options to treat the deadly hemorrhagic fever diseases they cause. Because of overlapping geographic occurrences and similarities in mode of entry into cells, we seek a practical drug or drug cocktail that could be used to treat infections by both viruses. Toward this goal, we directly compared eight drugs, approved or in clinical testing, for the ability to block entry mediated by the glycoproteins of both viruses. We identified five drugs with approximately equal potencies against both. Among these, we investigated the modes of action of arbidol, a drug licensed abroad to treat influenza infections. We found, as shown for influenza virus, that arbidol blocks fusion mediated by the Lassa virus glycoprotein. Our findings encourage the development of a combination of approved drugs to treat both Lassa and Ebola virus diseases.
INTRODUCTION
Lassa virus (LASV) is an enveloped ambisense RNA virus belonging to the Arenaviridae. As the most clinically significant member of this large family, LASV is a major pathogen in West Africa, where it infects an estimated 300,000 people each year. LASV has also been responsible for a number of imported cases of Lassa hemorrhagic fever (LHF) in Europe and North America in recent years (1). The 2018 outbreak of LHF in Nigeria was particularly severe, with over 430 confirmed positive cases and a case fatality rate of ∼25% (2). Classic symptoms of acute LHF include malaise, headache, fever, vomiting, respiratory distress, facial edema, and hemorrhaging of mucosal surfaces (3). Even in fatal cases, however, patients may not present with redolent hemorrhagic fever symptoms, complicating diagnosis (4).
The only antiviral treatment option for LHF is the guanosine analogue ribavirin. There are a substantial number of contraindications and adverse effects associated with ribavirin, and its efficacy in clinical trial settings remains controversial and underevaluated. Furthermore, while ribavirin is effective against other hemorrhagic fever arenaviruses, it has limited efficacy against filoviruses. Thus, current guidelines recommend ribavirin only after high-risk exposures to LASV (5). Given the partial geographic overlap between Ebola virus (EBOV) and LASV in West Africa and similar clinical presentation in early infection stages, it would be advantageous to have a common therapeutic effective against both viruses (6, 7).
Promising new compounds against LASV have been identified (6, 8–17), but the limited geographical endemicity of LASV, its inefficient person-to-person transmission, and low reinfection rates make the prospect of collecting adequate clinical trial data on new drugs challenging. Thus, a practical approach to more expeditiously grow the arsenal of drugs against these highly pathogenic viruses is to screen approved drugs for antiviral activity. When this strategy was employed, many FDA-approved compounds with repurposing potential were identified that showed inhibitory effects against EBOV (18–24). Many of these are thought to act upon the entry stages of EBOV infection. A similar recent screen revealed FDA-approved drugs with potential activity against LASV (8).
Viral entry inhibitors are valuable as therapeutics, since blocking infection early in the life cycle will reduce cellular and tissue damage associated with the replication of incoming viruses and the production of viral progeny. LASV employs several key features in common with EBOV for its entry: (i) it is internalized into the endocytic pathway by a macropinocytotic-like process after initial contact with surface receptors/attachment factors, (ii) low pH is needed to trigger fusion, and (iii) an endosomal, cholesterol binding receptor promotes endosomal escape (Lamp1 for LASV and NPC1 for EBOV) (12, 25–32). Hence, for this study, we selected eight low-molecular-weight drugs shown to inhibit EBOV entry and directly compared their inhibitory activities against LASV and EBOV. Five of these drugs have FDA approval (amodiaquine, clomiphene, niclosamide, sertraline, and toremifene), one is licensed abroad (arbidol), and two have been evaluated in clinical trials (apilimod and zoniporide).
The compound we investigated in most detail was the anti-influenza drug arbidol (umifenovir), which was developed and is currently used as an antiviral. Arbidol has been reported to have inhibitory effects on a diverse array of viruses, including DNA and RNA viruses as well as capsid- and membrane-enclosed viruses (33–36). Studies aimed at determining the mechanism of action of arbidol implicate a number of possible antiviral effects, including several steps of entry as well as later phases of the infectious cycle (36). The principal inhibitory effect of arbidol on influenza virus, for which it has been a licensed treatment in China and Russia for many years, appears to be during a late stage of entry, when influenza virus fuses with an endosomal membrane. While arbidol can bind directly to influenza hemagglutinin (HA) and inhibit its ability to transition to an activated conformation (37–39), it is not yet clear whether this is its sole or primary mechanism of anti-influenza activity or if arbidol also impairs fusion by intercalation into the viral or target membrane, thereby rendering the membrane less yielding for fusion (35).
RESULTS
Comparison of the potency of small-molecule inhibitors against LASV GP- and EBOV GP-mediated entry.Enveloped viruses that are endocytosed rely on their glycoproteins (GPs) to mediate the entire entry process, from attachment to the cell surface to fusion within endosomal membranes. In this study, we directly compared the effects of eight drugs that block EBOV entry for their effects on LASV GP-mediated entry. To do this, we used murine leukemia viruses (MLV) carrying a luciferase reporter and pseudotyped with either LASV or EBOV GP. Drug dosing ranges were determined by establishing the concentration of each drug needed to elicit a near-total inhibition of infection. The remaining doses in each set were 2-fold serial dilutions. A mock (vehicle-only) treatment was included as an anchor point in each series to assess the extent of inhibition in treated cells.
Representative direct comparative dose-response curves for LASV and EBOV for each of the eight drugs are presented in Fig. 1. Each drug was tested in parallel against LASV GP- and EBOV GP-MLV pseudoviruses in three to five independent experiments. Table 1 reports the average ratio of the 50% inhibitory concentration (IC50) value against LASV GP divided by that for EBOV GP, analyzed in parallel, for each of the eight drugs tested. These ratios indicated that the IC50 values against LASV GP-MLV pseudoviruses for five drugs (zoniporide, amodiaquine, niclosamide, apilimod, and arbidol) were either approximately the same as or lower than the corresponding values for EBOV GP-MLV pseudoviruses, indicating similar or enhanced potency against LASV GP-mediated infection. For three drugs, clomiphene, sertraline, and toremifene, the IC50 values for LASV were ∼3- to 6-fold greater than those for EBOV, indicating that these drugs are more potent against EBOV. It is noteworthy that these three drugs are cationic amphiphilic drugs (CADs), which may be especially active against EBOV (19, 20, 23, 40).
Representative dose response curves for eight low-molecular-weight drugs against LASV GP- and EBOV GP-mediated MLV pseudovirus infection. BSC-1 cells were pretreated with the indicated dose of the indicated drug for 1 h and then infected with MLV pseudoviruses encoding luciferase. Luciferase signals were measured 24 h later and normalized to the maximal signal from a triplicate set of mock-treated cells. Data points indicate the average percent inhibition from triplicate wells. Error bars represent the SDs. The red horizontal dashed line indicates 50% inhibition. Each dose-response comparison was conducted 3 to 5 times, with similar results.
Comparative ability of drugs to block infections by LASV and EBOV GP-MLV pseudovirusesa
For the remainder of the study we focused on arbidol, for reasons outlined in the introduction. Pécheur and colleagues reported an EC50 of 5.8 µM in Vero cells for arbidol against the New World arenavirus Tacaribe virus (34); to the best of our knowledge, this is the only published evaluation of the efficacy of arbidol against an arenavirus. Using MLV pseudoviruses, we found that in addition to inhibiting entry mediated by LASV GP (Fig. 1), arbidol inhibited entry mediated by the GPs of two other arenaviruses, those of lymphocytic choriomeningitis virus (LCMV) and Junin virus (Fig. 2A). We also found, using MLV pseudoviruses, that arbidol is somewhat more potent against LASV GP-mediated infection than against infection mediated by influenza virus HA from the WSN (H1N1) strain (Fig. 2B).
Comparative effects of arbidol on infection by MLV pseudoviruses bearing LASV or other viral glycoproteins: LASV, LCMV, and Junin GP (A) and LASV GP and influenza virus HA (B). MLV pseudoviruses bearing LASV GP, LCMV GP, Junin GP, or WSN influenza HA and NA were prepared as described in Materials and Methods. BSC-1 cells were pretreated with the indicated concentrations of arbidol and then processed and analyzed for infection as described in the legend to Fig. 1. Data in panel A are the averages ± SEMs from three experiments, each performed with triplicate samples. Data in panel B represent the averages ± SDs from triplicate samples from one experiment.
Arbidol blocks authentic LASV infection.To evaluate the efficacy of arbidol against authentic LASV, we performed LASV (Josiah) plaque reduction assays under biosafety level 4 (BSL4) conditions, testing the effects of concentrations of arbidol up to 40 µM. Cells were pretreated with arbidol for 1 h and then infected with LASV (Josiah) in the continued presence of arbidol for 24 h. In the first of three experiments the IC50 was ∼5 to 10 µM and the maximum inhibition was 98% (Fig. 3A); in the second, the IC50 was ∼20 µM and the maximum inhibition was 100% (Fig. 3B). In a third experiment, testing only 20 µM arbidol, the percent inhibition was 74% (Fig. 3C). The average percent inhibition caused by 20 µM arbidol from the three experiments was 72.5% (Fig. 3D). By visual inspection 40 µM arbidol had no effect on Vero cell monolayers for up to 5 days (data not shown). We note that the apparent IC50 for arbidol versus authentic LASV (Fig. 3) is higher than that seen with MLV pseudoviruses bearing LASV GP (Fig. 1 and 2).
Arbidol inhibits authentic LASV infection. Duplicate wells of Vero 76 cells were pretreated with the indicated concentration of arbidol (or vehicle or 25 mM NH4Cl) for 1 h and then infected with LASV (Josiah strain) virus at an MOI of 0.01. Following a 1-h binding period at 4°C, unadsorbed virus was removed and the cells were incubated for 24 h in the presence of drug in a 37°C CO2 incubator. Culture supernatants were harvested and serially diluted 10-fold in fresh medium, and then titers were determined on Vero 76 cells by a 96-h plaque assay. Results in panels A to C are the average titers from duplicate wells. The values in panel D indicate the average normalized infection in samples treated with 20 μM arbidol (±SD) from the experiments shown in panels A to C. **, P < 0.01.
Arbidol blocks LASV GP-mediated fusion.We next asked if arbidol impairs LASV GP-mediated fusion, as it does for other viruses (33, 35, 38, 39, 41). Given that optimal LASV fusion requires the endosomal protein Lamp1 (26, 31, 42), we used cells expressing Lamp1 at the plasma membrane (pmLamp) as fusion targets. Cell-cell fusion (CCF) was then induced between cocultured effector cells (expressing LASV GP at their surface) and target cells (expressing Lamp1 at their surface) by briefly exposing the cells to low pH, as described previously (31). To assess the effects of arbidol, effector cells (expressing LASV GP) were pretreated for 1 h with the indicated concentration of arbidol, cocultured with pmLamp1-expressing target cells, and then triggered to fuse by brief exposure to pH 5 (all in the continued presence of arbidol). The efficiency of CCF was then determined by measuring the activity of the luciferase reporter that is functionally restored upon cytoplasmic mixing of fused cells (43). As seen in Fig. 4A, CCF by LASV GP (at pH 5.0) was suppressed by 20 μM and 40 μM arbidol. Based on findings in parallel experiments (Fig. 4B), arbidol appeared more potent at impeding LASV-GP than influenza virus HA-mediated CCF, consistent with its somewhat stronger effect on LASV GP- compared to influenza virus HA-MLV pseudovirus infection (Fig. 2B).
Arbidol suppresses LASV GP-mediated cell-cell fusion (CCF). Effector cells were generated by transfecting HEK293T/17 cells with plasmids encoding DSP1-7 (the N-terminal split luciferase plasmid) and either LASV GP (A) or WSN influenza HA and NA (B). Target cells were generated by transfecting HEK293T/17 cells with plasmids encoding DSP8-11 (the C-terminal split luciferase plasmid) and pmLamp1. For the experiments, effector cells were preloaded with a luciferase substrate and then pretreated for 1 h with the indicated concentration of arbidol or 10% ethyl alcohol (EtOH; mock control). Effector cells were then cocultured with target HEK293T/17 cells (in the continued presence of arbidol or 10% EtOH) for 3 h at 37°C. At this time the cultures were pulsed with pH 5 buffer for 5 min at 37°C, reneutralized, and then returned to the 37°C CO2 incubator for 1 h, at which time the luminescent signal was measured. The data represent the normalized luminescent signals (relative to that of the mock-treated controls) from three experiments, each performed with triplicate samples. Error bars indicate SDs. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
As a complement to the CCF study (Fig. 4), we employed an assay involving forced fusion at the plasma membrane (FFPM) and assessed fusion of LASV GP-vesicular stomatitis virus (VSV) pseudoviruses with the surface of cells expressing pmLamp1 (i.e., with Lamp1 at the cell surface), as previously described (31). As seen in Fig. 5A, arbidol suppressed LASV-GP-mediated FFPM, with strong and complete inhibition seen with 20 and 40 μM doses, respectively. The experiment shown in Fig. 4A was conducted with a pulse at pH 5.0. As seen in Fig. 5B, 40 μM arbidol strongly inhibited LASV GP-mediated FFPM at both pH 5.0 and pH 5.5.
Arbidol inhibits LASV-GP-mediated forced fusion at the plasma membrane (FFPM). COS7 cells were transfected with plasmids encoding pmLamp1 and firefly luciferase (FLUC). Roughly 24 h later, the cells were pretreated with the indicated dose of arbidol (or mock treated) for 1 h. At this time, VSV pseudoviruses bearing LASV GP and encoding Renilla luciferase (RLUC) were allowed to bind at 4°C for 1 h. The cells were then pulsed for 5 min at 37°C with prewarmed buffers at either pH 5 (A) or the indicated pH (B), and then the buffer was replaced with complete medium containing NH4Cl to prevent infection via the normal endocytic route. After 12 to 18 h at 37°C, RLUC (an indicator of infection via FFPM) and FLUC (to standardize transfected cell numbers) were measured. The ratio of Renilla to firefly luciferase (RLUC/FLUC) was then normalized to RLUC/FLUC ratio for the mock-treated cells (A) or directly plotted (B). Data are from a representative experiment performed with quintuplicate samples. Error bars indicate SD. ***, P < 0.001; ****, P < 0.0001. Each experiment was repeated one time, with similar results.
Effects of arbidol on LASV GP1 dissociation.In the case of influenza, arbidol stabilizes HA (the fusion protein) such that the pH dependence for its fusion-inducing conformation change is shifted by 0.2 to 0.3 U in the more acidic direction (37, 39). Stabilization of HA is considered part of the mechanism of arbidol against influenza virus (38, 39, 44, 45). Since two independent assays (CCF and FFPM) showed that arbidol impairs the fusion activity of LASV GP, we next asked whether it impairs a conformational change in GP1 required for fusion activation. Upon exposure to low pH, LASV GP undergoes structural rearrangements, one of the earliest being dissociation of GP1, the receptor binding subunit, from GP2, the fusion subunit. This early change is thought to license subsequent changes that allow the fusion loop (in GP2) to access the target membrane and then to permit GP2 to fold back into a trimer of hairpins, which brings the viral and endosomal membranes into intimate contact leading to their fusion (46–49). Experiments using isolated LASV GP1/GP2 captured on beads showed that in this system, dissociation of the 44-kDa GP1 subunit occurs optimally at 37°C and half maximally at pH ∼6.4 at 37°C following a 1-min low-pH pulse (data not shown). We therefore treated LASV GP1/GP2 immobilized on beads with either 0 or 40 µM arbidol and then exposed the beads to buffers of different pH values for 1 min at 37°C. As seen in Fig. 6A, the presence of 40 μM arbidol shifted the pH dependence for GP1 dissociation by ∼0.5 U in the more acidic direction, suggesting that, as for influenza HA, arbidol can stabilize LASV GP. If arbidol stabilizes LASV GP, then it might delay GP1 dissociation in this system. To test this idea, we again captured LASV GP1/GP2 on beads, pretreated the samples with 0 or 40 μM arbidol, treated the beads at pH 6.5 and 37°C in the presence of arbidol, and then took samples from 0 to 5 min and assayed them for GP1 dissociation. As seen in Fig. 6B, arbidol appeared to introduce an ∼30-s lag, thereby slowing GP1 dissociation.
Arbidol impairs LASV GP1 dissociation from GP2. (A) Flag-tagged LASV GP from cell lysates of Lamp1 KO HEK293T/17 cells was immobilized on anti-Flag (M2) magnetic beads, which were then treated with 0 or 40 μM arbidol for 1 h at 4°C. The beads were then subjected to a pulse for 1 min at 37°C at the indicated pH in the presence or absence of arbidol. The extent of GP1 dissociation was then determined by Western blot analysis of GP1 in the supernatant and bead-bound fractions. Data are the averages from four experiments. Error bars represent SEMs. (B) Flag-tagged LASV GP, prepared and immobilized as for panel (A), was pretreated with 0 or 40 μM arbidol. The beads were then exposed to pH 6.5 at 37°C for the indicated time, in the presence or absence of arbidol, and the extent of GP1 dissociation was determined as for panel A. Data are the averages from five experiments. Error bars represent SEMs.
DISCUSSION
Drugs that block LASV and EBOV GP-mediated entry with similar potencies.We began this study by comparing the ability of eight drugs to inhibit LASV and EBOV GP-mediated infection. All eight are orally available, room temperature-stable small molecules that block EBOV entry (18–20, 22, 23, 40, 50) and target entry processes also used by LASV (20–23, 51–53). Six are approved for clinical use and two are in advanced clinical testing. These collective features offer practical advantages (e.g., net costs and ease of transport and delivery) compared to novel drugs, many of which are designed in a “one drug-one bug” approach (54).
Five drugs showed similar potencies against LASV and EBOV GP-mediated entry (Table 1). Zoniporide, an inhibitor of the plasma membrane Na+/H+ exchanger, blocks infection by the arenavirus LCMV by thwarting macropinocytotic uptake of viral particles (52). As both LASV (30) and EBOV (55, 56) are taken into cells by macropinocytosis, we expected and found zoniporide to have similar activities against both. Two drugs that impair endosome acidification, needed for the entry of both viruses (25, 49)—amodiaquine (an antimalarial) and niclosamide (an anthelminthic)—also showed similar potencies against these two viral GPs in our direct comparative tests. Both have been shown, albeit not in direct comparative studies, to inhibit many pathogens that enter cells by endocytosis (21–23, 57–60). The fourth drug with similar activity against both viruses is apilimod, which inhibits PIKfyve, an enzyme required for late endosome maturation (51, 53, 61) and EBOV entry (27). Apilimod blocks EBOV trafficking to late endosomes (51, 53), which serve as portals for both EBOV and LASV (25, 26, 62–64). The fifth drug with similar potency against LASV GP- and EBOV GP- mediated entry is arbidol (Umifenovir), a synthetic antiviral approved and used in Russia and China to combat influenza and shown to have broad-spectrum antiviral activity (33, 34, 36, 38, 39).
The other three drugs—clomiphene, toremifene, and sertraline—showed ∼3- to 6-fold greater activity against EBOV GP- than LASV GP-mediated entry. They are CADs that block EBOV infections (19, 20, 22, 23, 40) and cause cholesterol accumulation in late endosomes, mimicking effects of dysfunctional NPC1, the EBOV receptor (27, 28, 65). Since LASV also enters cells through late endosomes (31), CADs may interfere with LASV entry due to general impairments of late endosome function. The enhanced activity of toremifene and sertraline against EBOV (Table 1) may be because they can bind to both EBOV GP, as shown by thermal stability assays and X-ray crystallography (66, 67), and the viral and/or endosomal membrane (35, 39, 45, 66).
Mechanisms by which arbidol may block LASV entry and infection.Arbidol showed roughly equivalent potencies against LASV and EBOV GP-mediated pseudovirus infection (Fig. 1 and Table 1), blocked infection by authentic LASV (Fig. 3), and was somewhat more effective at blocking LASV GP- versus influenza virus HA-mediated fusion and entry (Fig. 2B and Fig. 4B). Three mechanisms have been proposed for the action of arbidol against influenza: direct binding to HA (Kd [dissociation constant], 47 μM for binding to PR8 influenza virus HA) (38, 44) and, as proposed for other viruses, binding to the viral and/or target membrane so as to reduce fusion fitness (34, 35). By binding to and stabilizing HA, arbidol shifts the pH threshold for fusion-inducing conformational changes by ∼0.2 to 0.3 U in the more acidic direction (37, 38). This is noteworthy, as even such small changes in fusion pH can significantly impact influenza virus infectivity (67). Similarly, we found that arbidol shifts the pH dependence for LASV GP1 dissociation from GP2 by ∼0.5 U (in the more acidic direction). This event is thought to unclamp the LASV fusion subunit (GP2), analogous to the unclamping of influenza HA2 and HIV gp41. Hence, in addition to its likely general fusion-impairing effects caused by intercalation into the viral and/or endosomal membrane (35, 45, 66), it is plausible that arbidol additionally affects the stability of LASV GP in a manner similar to its effects on influenza HA (37–39). And, as reviewed elsewhere (36), arbidol may also affect steps upstream or downstream of fusion.
Therapeutic potential of drugs with similar potencies against LASV and EBOV.As reasoned in the introduction, a long-term goal is to identify a drug cocktail that inhibits both LASV and EBOV. Our focus is on orally available, room temperature-stable drugs that target processes used in common for LASV and EBOV entry into cells. In this study, we identified five drugs that target discrete steps of entry and show approximately equal potencies against LASV and EBOV GP-mediated entry: the macropinocytosis inhibitor zoniporide, the endosome acidification inhibitors amodiaquine and niclosamide, the trafficking inhibitor apilimod, and the fusion inhibitor arbidol.
While zoniporide is not approved, it has advanced to phase II clinical trials for treatment of cardiovascular diseases. As an alternate, the FDA-approved drug aripiprazole (trade name, Abilify) may have utility. Aripiprazole blocks EBOV infection and synergizes with other entry inhibitors (20, 23). Preliminary data suggest that it blocks EBOV particle internalization as well as LASV GP-mediated infection (laboratory of J. M. White, unpublished data). The endosome acidification inhibitors amodiaquine and niclosamide are orally available and FDA approved to treat malaria and helminthic diseases, respectively. For both, the maximum serum concentration (Cmax) is within range of the IC50 for anti-EBOV/LASV activity (20–23, 57, 68, 69). Newer amodiaquine derivatives (59) or synergistic drug pairs containing amodiaquine or niclosamide and another agent could lower the needed doses (23). And while the trafficking inhibitor apilimod is well tolerated and a potent antagonist of EBOV and LASV (51, 53, 70; also this study), in an early test, it did not protect mice from lethal EBOV challenge (see reference 23), likely due to its inhibition of interleukin 12/23 (IL-12/23) production (reference 61 and K. Rogers, L. Stunz, O. Shtanko, L. Mallinger, J. M. White, M. Schmidt, S. Varga, N. Butler, G. Bishop, and W. Maury, submitted for publication).
Among the fusion inhibitors tested, arbidol emerges as a candidate for future consideration, as it shows similar activities against LASV and EBOV and appears, in our assays, somewhat more potent against LASV- and EBOV GP- versus influenza HA-mediated fusion and entry. Arbidol is approved and used in China and Russia against influenza and has shown strikingly broad antiviral activity (33, 36, 71). A single standard human dose (200 mg) of arbidol yields a Cmax lower (33, 72–74) than our preliminary indication (Fig. 3) of its IC50 against authentic LASV. We note, however, that the estimated IC50s for arbidol against authentic LASV (Fig. 3) and influenza (39, 45, 75) are similar (roughly ∼10 μM; variable for different influenza virus strains). As standard dosing of arbidol for influenza is 200 mg three (72) or four (75) times per day, and as arbidol has a long half-life (36), the net (cumulative) Cmax with multiple daily dosing is expected to be significantly higher (M. Paine, personal communication). Hence, as standard dosing of arbidol is clinically beneficial against influenza (see, for example, reference 75), it may have utility against LASV and EBOV. However, the only modest reduction in titer seen (Fig. 3) coupled with experience with another hemorrhagic fever virus (76) argues against proposing arbidol as a stand-alone therapeutic. A tolerated higher dose of arbidol (74), a new arbidol derivative (39, 44), or a combination of arbidol with another drug (23, 77, 78) might lower the dose needed to be in line with attainable antiviral efficacy. With respect to a potential drug cocktail including arbidol, it is interesting that amantadine, rimantadine, ribavirin, and ribamidil have been reported to enhance the activity of arbidol against influenza (79).
We envision that an orally available, room temperature-stable, approved drug or cocktail of approved drugs could be rapidly deployed for treatment or prophylaxis against (suspected) cases of LASV and EBOV, especially in regions around the globe that are challenged in terms of resources, infrastructure, and accessibility. Such a drug, or drug cocktail, might be valuable before a definitive diagnosis has been made, concurrent with vaccination (e.g., of health care workers and/or first responders to outbreaks), and/or during the setup of ring vaccination. Here we have identified several drugs with these attributes that show similar degrees of inhibition of LASV and EBOV entry. Moreover, three of them—amodiaquine, niclosamide, and arbidol—inhibit multiple enveloped viruses (33, 36, 57, 60, 71, 80–83). Hence, a drug or drug cocktail containing these drugs could inhibit multiple enveloped viruses that enter cells through the endocytic pathway (49).
MATERIALS AND METHODS
Chemicals and cell culture.Dulbecco’s modified Eagle’s medium (DMEM), phenol-red free DMEM, Opti-MEM (OMEM), sodium pyruvate, antibiotic/antimycotic, trypsin-EDTA (0.05%), phenol-red free trypsin-EDTA (0.5%), and neutral red (NR) were from Thermo Fisher Scientific. Phosphate-buffered saline (PBS) was from Corning. Cosmic Calf serum (CCS), fetal bovine serum (FBS), and supplemented calf serum (SCS) were from HyClone. Fibronectin was from Millipore. Polyethylenimine (PEI) and nonenzymatic cell dissociation solution were from Sigma. Lipofectamine 2000 was from Invitrogen. Toremifene citrate was from Selleck Chemicals. Zoniporide, amodiaquine, niclosamide, and clomiphene citrate were from Sigma. Apilimod was from Axon MedChem. Sertraline hydrochloride was from Toronto Research Chemicals. Arbidol was synthesized commercially, and the purity and structure of the product were confirmed as described previously (34).
HEK293T/17 and BSC-1 cells were from the ATCC. Lamp1 KO HEK293T/17 cells (clone 1D4) were described by Hulseberg et al. (31). COS7 cells were from the ATCC and a kind gift from Douglas DeSimone at the University of Virginia. BHK-21 cells were from the ATCC and a kind gift from James Casanova at the University of Virginia. Vero 76 cells were from the U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID). HEK293T/17 and BSC-1 cells were maintained in DMEM containing 10% CCS. BHK21 cells were maintained in DMEM containing 10% SCS. COS7 and Lamp1 KO HEK293T/17 cells were maintained in DMEM containing 10% FBS, 1% sodium pyruvate, and1% antibiotic/antimycotic. Vero 76 cells were maintained in Corning DMEM with 10% Gibco FBS, 1% penicillin-streptomycin, 1% l-glutamine, and 1% sodium pyruvate.
Plasmids and virus.The pCMV-LASV-GPC Josiah strain plasmid was from F. L. Cosset (Université de Lyon, France) via Gregory Melikian (Emory University); the LASV-GPC-Flag pCC421 Josiah strain was from Jason Botten (University of Vermont). VSV-G plasmid was from Michael Whitt (University of Tennessee); pDisplay-EBOV-GPΔ Mayinga strain was from Erica Saphire (Scripps Research Institute). pTG-Luc plasmid was from Jean Dubuisson (Centre National de la Recherche Scientifique, Lille, France) via Gary Whittaker (Cornell University), and pCMV-Gag-Pol plasmid was from Jean Millet and Gary Whittaker (Cornell University) and Jean Dubuisson. Gag-βlaM plasmid was made by James Simmons (University of Virginia). The pcDNA3-luciferase (firefly) plasmid was from Addgene. The DSP1-7 and DSP8-11 plasmids were from Naoyuki Kondo (Kansai Medical University, Japan). The WSN HA and neuraminidase (NA) plasmids were from Gary Whittaker (Cornell University). Plasmids encoding LCMV GP and Junin virus GP were from Jack Nunberg (University of Montana).
The stock of LASV (Josiah strain) used was generated by infecting Vero E6 cells in complete Eagle’s minimal essential medium (EMEM; VWR) containing 10% FBS (HyClone) and 2% l-glutamine (Thermo Fisher Scientific). Infected cells were incubated at 37°C and 5% CO2. Virus-containing cell culture supernatant was harvested 3 days postinoculation, clarified at 10,000 × g and 4°C for 10 min, and frozen at −80°C. All work with native LASV was conducted in a BSL4 containment suite with personnel in positive-pressure encapsulating suits following appropriate institutional standard operating procedures.
Antibodies and immunoprecipitation reagents.For Western blotting, the mouse anti-LASV-GP L52-134-23A was from USAMRIID and anti-mouse IR680RD was from Licor. For LASV GP bead capture, anti-Flag M2 magnetic beads were from Sigma.
Pseudovirus production.To produce VSV pseudoviruses, 1 × 106 BHK-21 cells were seeded in each of multiple 10-cm2 dishes. Cells in each dish were transfected with 12 μg of plasmid encoding LASV-GPC using PEI. The following day, cells were infected with 40 μl (per dish) of VSV-ΔG helper virus (from plaque eluates of known titer) encoding Renilla luciferase (diluted in serum-free media) for 1 h at 37°C. After infection, cells were washed extensively with cold PBS and incubated overnight in complete DMEM. Supernatants containing pseudoviruses were collected, clarified, and pelleted through a 20% sucrose-HM (20 mM HEPES, 20 mM morpholineethanesulfonic acid [MES], 130 mM NaCl [pH 7.4]) cushion. The pellet was resuspended in 10% sucrose-HM. VSV-ΔG helper virus was produced following the same procedure, by infecting VSV-G transfected cells with eluate from VSV-ΔG plaques.
For MLV pseudoviruses, HEK293T/17 cells were seeded in 10-cm2 dishes. The following day, cells were transfected with 6 μg of total DNA using a 2:1:1:1 ratio of pTG-Luc:pCMV Gag-Pol:Gag-βlam:glycoprotein. At 48 h posttransfection, virus-containing medium was harvested, clarified, pelleted through a 20% sucrose-HM cushion, resuspended in 10% sucrose-HM, and stored at −80°C in single-use aliquots.
LASV plaque reduction assay.Vero 76 cells were seeded on 6-well plates. At full confluence, cells in duplicate wells were pretreated with the indicated concentration of arbidol, 25 mM NH4Cl, or 10% ethanol vehicle for 1 h at 37°C. Cells were then infected with LASV at a multiplicity of infection (MOI) of 0.01 for 1 h in the presence of the indicated concentration of arbidol or vehicle. Cells were washed twice to remove unbound virus and incubated for 24 h at 37°C in drug-containing medium. Supernatants were harvested and 10-fold serial dilutions were made to infect ∼90% confluent monolayers of Vero 76 cells in 6-well plates for 1 h at 37°C, with rocking every 15 min. A primary overlay consisting of a 1:1 mixture of 1.6% SeaKem agarose (Lonza) and 2× Eagle’s basal medium (Thermo Fisher Scientific) supplemented with 20% FBS and 8% Glutamax (Thermo Fisher Scientific) was then added on top of the infected cells and allowed to solidify. Cells were incubated for 4 days at 37°C, followed by addition of a secondary overlay consisting of a 1:1 mixture identical to the above but with the addition of 8% NR (final concentration was 4%). Plaques were counted the following day. Plaque counts were averaged from duplicate wells and then multiplied by the dilution factor to establish the starting titer of input supernatants.
Biosafety.All manipulations involving live LASV were performed in a biosafety level 4 containment suite at USAMRIID with personnel wearing positive-pressure protective suits fitted with HEPA filters and umbilical-fed air. USAMRIID is registered with the Centers for Disease Control Select Agent Program for the possession and use of biological select agents and toxins and has implemented a biological surety program in accordance with U.S. Army regulation AR 50-1, “Biological Surety.”
Pseudovirus infection assay.BSC-1 cells were seeded on white 96-well plates (1.5 × 104 cells/well). The following day, cells were pretreated with drugs (or mock) for 1 h in OMEM and then, while maintaining the presence of drug, were infected with an input of EBOV GP- and LASV GP-pseudoviruses adjusted to achieve roughly equivalent RLU signals in the mock-treated samples. After 24 h at 37°C and 5% CO2, the cells were lysed with Britelite reagent (PerkinElmer) and luminescence was measured. IC50s were determined and statistical analysis of all data was performed using GraphPad Prism 7 (GraphPad Software, Inc.): log(agonist) versus response-variable slope (four parameters) constrained to bottom = 0.
Cell-cell fusion (CCF) assay.Effector (HEK293T/17) cells were seeded on 6-well plates (6.75 × 105 cells/well). Target (HEK293T/17) cells were seeded on fibronectin-coated opaque white 96-well plates (3.5 × 104 cells/well). Effector cells were transfected with 1 µg/well of GP plasmid and 1 µg/well of DSP1-7 plasmid. Target cells were cotransfected with 33 ng/well of pmLamp1 and 33 ng/well of DSP8-11 plasmid. Cells were transfected using Lipofectamine 2000 according to the manufacturer’s instructions. Twenty-four hours posttransfection, effector cells were loaded with EnduRen luciferase substrate (Promega) (60 µM in complete DMEM) for 2 h at 37°C. Effector cells were then rinsed with PBS and lifted with nonenzymatic cell dissociation solution. Effector cells were resuspended in complete DMEM and 1 × 105 effector cells were overlaid onto each well of target cells (96-well plate). Cells were cocultured for 3 h. At this time, a low pH pulse was applied with fusion buffer (100 mM NaCl, 15 mM HEPES, 15 mM succinate, 15 mM MES, 2 mg/ml of glucose) adjusted to pH 5.0 for 5 min at 37°C. The pH was reneutralized by replacing the fusion buffer with complete DMEM, and the cells were returned to 37°C for 1 h before measuring luciferase activity.
Forced fusion at the plasma membrane (FFPM) assay.COS7 cells were seeded in 6-well plates (4 × 105 cells/well). Approximately 24 h postseeding, the cells were transfected with 1 µg of plasmid encoding firefly luciferase using Lipofectamine 2000 according to the manufacturer’s instructions. Approximately 24 h posttransfection, the cells were washed, lifted, and reseeded at 1.5 × 104 cells/well on fibronectin-coated opaque white 96-well plates. The day after reseeding, cells were chilled on ice for 15 min and LASV-GP VSV-luciferase (Renilla) pseudoviruses, an amount determined to reach a target signal of at least 1 × 106 RLUs in a standard infection assay, were added to cells in quintuplicate in serum-free DMEM. Pseudoviruses were bound to the cells by centrifugation (250 × g, 1 h, and 4°C). Cells were returned to ice and washed once with cold PBS. Fusion was triggered by applying a pulse of prewarmed low pH fusion buffer (as in CCF assay) for 5 min at 37°C adjusted to the indicated pH values. Cells were returned to ice, and the fusion buffer was replaced with complete DMEM containing 40 mM NH4Cl (to block virus entry via the normal endocytic route). Sixteen hours later, luciferase activity was measured using the Dual-Glo luciferase assay system (Promega) according to the manufacturer’s instructions using a Promega GloMax luminometer. The ratio of Renilla luciferase activity (an indicator for pseudovirus infection) over firefly luciferase activity (to account for the number of cells) was calculated to assess viral GP-mediated fusion with the plasma membrane.
GP1 dissociation assay.Lamp1 KO HEK293T/17 cells were seeded in 6-well plates (6.25 × 105 cells/well). The following day, cells were transfected with 1 µg of LASV-GPC-Flag using PEI. At 48 h posttransfection, the cells were lysed with NETI buffer (150 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, 0.5% Igepal) at pH 8. After clearing cell debris (centrifugation for 15 min at 21,000 × g), the lysate was incubated with anti-Flag M2 magnetic beads (that had been prewashed twice in NETI buffer [pH 8]) for 1 h at 4°C. Arbidol was then added to the bead-plus-lysate mixture as indicated, and the samples were incubated for an additional hour at 4°C. For pH-dependent dissociation experiments, beads with captured LASV GP and pretreated with or without arbidol were then pulled over on a magnetic stand and quickly washed with cold NETI buffer (without arbidol) at the desired pH. The cold NETI buffer was then replaced with prewarmed NETI buffer at the same pH with or without arbidol, as indicated. Samples were incubated at 37°C for 1 min. For time-dependent dissociation experiments, beads with captured LASV GP and pretreated with or without arbidol as described above were quickly washed with cold NETI buffer at pH 6.5. The cold pH 6.5 buffer was then replaced with prewarmed NETI buffer at pH 6.5 with or without arbidol. Samples were then incubated at 37°C for 0.5, 1.0, 2.5, or 5.0 min. The “0 min” samples were treated with prewarmed buffer and then placed immediately on ice after buffer addition. At the end of the incubation period, the samples were immediately placed on the magnetic rack and supernatants collected. Proteins were then eluted from the residual beads using 100 mM glycine (pH 3.5) for 15 min at 25°C with constant vortexing, and the eluted samples were neutralized by the addition of 1 M Tris-HCl (pH 8.5). Supernatant and bead samples were then analyzed by SDS-PAGE and Western blotting, with a primary antibody against LASV GP1. The signal intensity of the GP1 bands in the supernatant and corresponding bead samples was measured using ImageJ. Percent GP1 dissociation was calculated as the signal intensity of the GP1 band in the supernatant divided by the summed signal intensity of the GP1 bands in the supernatant and bead samples.
ACKNOWLEDGMENTS
We thank Aura Garrison (USAMRIID) for her generous gift of the LASV stock used in the plaque reduction assays. We also thank Mary Paine, Washington State University, for helpful discussions on how to calculate drug accumulation indices.
S.J.P. dedicates this manuscript to the late Delsworth G. Harnish, who inspired a love for virology and arenaviruses in particular.
This work was supported by NIH RO1 AI114776 (to J.M.W.).
FOOTNOTES
- Received 6 December 2018.
- Accepted 22 January 2019.
- Accepted manuscript posted online 30 January 2019.
- Copyright © 2019 American Society for Microbiology.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵