Identification of In Vitro Inhibitors of Monkeypox Replication

ABSTRACT Monkeypox virus (MPXV) infections in humans have historically been restricted to regions of endemicity in Africa. However, in 2022, an alarming number of MPXV cases were reported globally, with evidence of person-to-person transmission. Because of this, the World Health Organization (WHO) declared the MPXV outbreak a public health emergency of international concern. The supply of MPXV vaccines is limited, and only two antivirals, tecovirimat and brincidofovir, approved by the U.S. Food and Drug Administration (FDA) for the treatment of smallpox, are currently available for the treatment of MPXV infection. Here, we evaluated 19 compounds previously shown to inhibit different RNA viruses for their ability to inhibit orthopoxvirus infections. We first used recombinant vaccinia virus (rVACV) expressing fluorescence (mScarlet or green fluorescent protein [GFP]) and luciferase (Nluc) reporter genes to identify compounds with antiorthopoxvirus activity. Seven compounds from the ReFRAME library (antimycin A, mycophenolic acid, AVN-944, pyrazofurin, mycophenolate mofetil, azaribine, and brequinar) and six compounds from the NPC library (buparvaquone, valinomycin, narasin, monensin, rotenone, and mubritinib) showed inhibitory activity against rVACV. Notably, the anti-VACV activity of some of the compounds in the ReFRAME library (antimycin A, mycophenolic acid, AVN-944, mycophenolate mofetil, and brequinar) and all the compounds from the NPC library (buparvaquone, valinomycin, narasin, monensin, rotenone, and mubritinib) were confirmed with MPXV, demonstrating their inhibitory activity in vitro against two orthopoxviruses. IMPORTANCE Despite the eradication of smallpox, some orthopoxviruses remain important human pathogens, as exemplified by the recent 2022 monkeypox virus (MPXV) outbreak. Although smallpox vaccines are effective against MPXV, access to those vaccines is limited. In addition, current antiviral treatment against MPXV infections is limited to the use of the FDA-approved drugs tecovirimat and brincidofovir. Thus, there is an urgent need to identify novel antivirals for the treatment of MPXV infection and other potentially zoonotic orthopoxvirus infections. Here, we show that 13 compounds, derived from two different libraries, previously found to inhibit several RNA viruses, also inhibit VACV. Notably, 11 compounds also displayed inhibitory activity against MPXV.

libraries against orthopoxviruses, which were identified as having antiviral activity against different RNA viruses, we used our previously described fluorescent protein (Scarlet or GFP)-and Nluc-expressing rVACV (rVACV Nluc/Scarlet and rVACV Nluc/GFP) ( Fig. 1A) (54). We have shown that expression levels of reporter genes from these rVACVs can be used as an accurate surrogate for viral infection (54). First, we validated the use of reporter-expressing rVACV (rVACV Nluc/Scarlet and rVACV Nluc/GFP) by assessing the inhibitory activity of tecovirimat, an FDA-approved antiviral against poxviruses (38)(39)(40)(41). We used 3-fold serial dilutions (starting at 50 mM) of tecovirimat to treat human A549 cells infected with rVACV Nluc/Scarlet or rVACV Nluc/GFP, and viral infection was assessed based on fluorescent and luciferase expression ( Fig. 1B and C). Tecovirimat exhibited a dose-dependent inhibitory effect in Scarlet (Fig. 1B, left) or GFP (right) expression levels. Tecovirimat half-maximal effective concentrations (EC 50 s) for rVACV Nluc/Scarlet and Nluc/GFP were similar, 0.031 and 0.026 mM, respectively, regardless of which fluorescence reporter was used for quantifications (Fig. 1C, left). Likewise, EC 50 s calculated by assessing Nluc activity in the cell culture supernatants from rVACV Nluc/Scarlet-and rVACV Nluc/GFP-infected cells were also similar (0.047 and 0.033 mM, respectively) ( Fig. 1C, right). Selectivity index (SI) values based on Scarlet or GFP levels were also similar (.1,612.9 and .1,923.1, respectively) and comparable to those based on Nluc from either rVACV Nluc/Scarlet or rVACV Nluc/GFP (.1,063.8 and .1,515.2, respectively). To further validate the use of bireporter rVACV to assess compound inhibitory activity, we conducted a viral titer reduction assay. For this we infected A549 cells at low (0.01) and high (3) multiplicities of infection (MOI) and treated them with 10-fold serial dilutions of tecovirimat (starting concentration of 100 mM). Cell culture supernatants were collected at 24, 48, and 72 hpi. Then, extracellular viral titers and Nluc activity (rVACV Nluc/Scarlet or rVACV Nluc/GFP) were determined from the cell culture supernatants. We found a dose-dependent decrease in extracellular viral titers of rVACV Nluc/Scarlet, rVACV Nluc/GFP, and MPXV (Fig. 1D). Tecovirimat acts primarily by inhibiting the activity of VACV protein VP37, which is required for virus egress. This results in decreased production of extracellular virus and viral spread without affecting gene expression, reflected in cell culture supernatants having Nluc activity even in the presence of the highest concentration (100 mM) of tecovirimat (Fig. 1E). The decrease in extracellular viral titers in high-MOI infections in the presence of tecovirimat presumably reflects inhibition of extracellular virus formation by the drug. Nonetheless, the antiviral activities of tecovirimat obtained with rVACV Nluc/Scarlet and rVACV Nluc/GFP and with MPXV were consistent with those reported in the literature for VACV (60) and MPXV (38,39,43,61,62). These results demonstrated that rVACV Nluc/Scarlet or rVACV Nluc/GFP can be used to determine accurately the inhibitory properties of compounds against orthopoxvirus based on fluorescent or luciferase expression.
Effect of selected compounds from the ReFRAME and NPC libraries on MPXV multiplication. VACV and MPXV are both orthopoxviruses and share similar genetic and biological properties. We therefore examined whether compounds with anti-VACV activity from the ReFRAME and NPC libraries were also able to inhibit MPXV infection. To assess the ability of the compounds in the ReFRAME and NPC libraries to inhibit MPXV, we used a focus-forming reduction assay (FFRA). We found that five of the compounds tested from the ReFRAME library ( Fig. 4A and Table 3) and all of the compounds from the NPC library ( Fig. 4B and Table 4) also had inhibitory activity against MPXV. As anticipated, benzimidazole ( Fig. 4A and B) did not inhibit MPXV infection, whereas tecovirimat ( Fig. 4A and B) potently inhibited MPXV (63). As with VACV, antimycin A, AVN-944, and brequinar from the ReFRAME library (Table 3) and valinomycin, rotenone, and mubritinib from the NPC library (Table 4) were the compounds with the strongest inhibitory activity. However, similar to the results with VACV, the SI values of the compounds differed based on the CC 50 values obtained from the MTT and XTT toxicity assays (Tables 3 and 4). These results demonstrated that a set of compounds from Representative images are shown. Bars, 100 mm. Magnification, Â20. To quantify inhibition of viral replication, Scarlet (red squares) or GFP (green triangles) expression levels were quantified at 24 hpi using a fluorescent microplate reader (C, left) or Nluc expression (C, right) was assessed using a luminometer for rVACV Nluc/Scarlet (red squares) or rVACV Nluc/GFP (green triangles). The EC 50 of tecovirimat was calculated using sigmoidal doseresponse curves. The CC 50 of tecovirimat was determined using an MTT assay kit. The SI was calculated by dividing the CC 50 by the EC 50 . The percent viral inhibition was normalized to non-tecovirimat-treated controls. The dotted lines indicate 50% inhibition. Data are means and SD of viral inhibition from quadruplicate wells (n = 4). (D and E) Viral titers (D) and Nluc activity (E). Human A549 cells (2 Â 10 4 cells/well, 96-well plates, in triplicate) were infected with an MOI of 0.01 (top) or 3 (bottom) of rVACV Nluc/Scarlet, rVACV Nluc/GFP, or MPXV for 1 h and incubated with 10-fold serial dilutions (starting concentration of 100 mM) of tecovirimat. At 24, 48, and 72 hpi, tissue culture supernatants were collected, and viral titers were determined by plaque assay in CV1 cells. Additionally, Nluc activity (E) in cell culture supernatants was determined using a luminometer for rVACV Nluc/Scarlet (left) or rVACV Nluc/GFP (right). Mock-infected cells and cells infected in the absence of drug were included as internal controls. Data are means and SD of viral inhibition from triplicate wells (n = 3).
the ReFRAME and NPC libraries we previously identified as having antiviral activity against RNA viruses also inhibit the DNA viruses VACV and MPXV, expanding their potential as inhibitors against a diverse collection of RNA and DNA viruses.

DISCUSSION
The MPXV outbreak in early 2022 was declared a public health emergency of international concern by the WHO, with over 30,286 confirmed cases and 38 deaths in the United States as of March 2023 (86,746 cases and 112 deaths worldwide) (1-3). Alarmingly, the current MPXV outbreak exhibited uncommon person-to-person transmission through prolonged exposure or sexual contact with infected individuals (1, 5-7). Despite two vaccines being available for prevention of MPXV infection, insufficient supplies have restricted their accessibility to those in high-risk populations, including immunocompromised individuals and men who have sex with other men. Only two antiviral drugs, tecovirimat and brincidofovir, previously approved by the U.S. FDA for the treatment of smallpox, are available for the treatment of MPXV infection, highlighting the and tecovirimat (K) were included as negative and positive controls, respectively. Mock-infected cells and cells infected in the absence of drug were included as internal controls. At 24 hpi, inhibition of rVACV Nluc/Scarlet and rVACV Nluc/GFP viral replication was evaluated by quantifying fluorescent Scarlet or GFP (top graphs) and Nluc (bottom graphs) expression using a fluorescent microplate reader and a luminometer, respectively. The EC 50 for each compound was calculated using sigmoidal dose-response curves with GraphPad Prism. The dotted lines indicate 50% inhibition. Data are means and SD of viral inhibition from quadruplicate wells (n = 4). At the same time point (hpi), Scarlet (rVACV Nluc/Scarlet, top) and GFP (rVACV Nluc/GFP, bottom) fluorescent expression in infected cells in the absence (0 mM) or in the presence of the indicated concentrations of the compounds (maximum concentration and EC 50 ) was visualized using a fluorescence microscope. Representative images are shown. Bars, 100 mm. Magnification, Â20. urgent medical need to identify novel antivirals for the treatment of MPXV and other orthopoxvirus infections.
In this study, we first demonstrated the feasibility of using rVACV expressing both fluorescent proteins (Scarlet and GFP) and luciferase (Nluc) to easily identify compounds with inhibitory activity against VACV. Next, we used rVACV Nluc/Scarlet and rVACV Nluc-GFP to test the anti-VACV activity of compounds from the ReFRAME and NPC libraries previously reported to have antiviral activity against RNA viruses (55)(56)(57)(58)(59). Using both fluorescent Scarlet or GFP and Nluc readouts, we showed that several of the compounds in the ReFRAME and NPC libraries we tested had inhibitory activity against VACV and that assessing their inhibitory activity using fluorescent-protein (Scarlet or GFP) or luciferase (Nluc) expression resulted in similar EC 50 s and SI values.
The NPC library compounds valinomycin, rotenone, and mubritinib also exhibited potent anti-VACV activity. Other compounds with inhibitory activity against VACV from the NPC library included buparvaquone, narasin, and monensin. Valinomycin has been described to have properties against tumors (72), bacteria (73), fungi (74), and viruses (75,76), by acting as a potassium ionophore that increases potassium ion permeability of the mitochondrial inner membrane, thereby inhibiting oxidative phosphorylation (77). Monensin is also a selective ionophore that facilitates the transmembrane exchange of sodium ions for protons, which leads to the inhibition of intracellular transport of Golgi apparatus-associated proteins. Monensin has been shown to have anticancer (78) and antiviral properties (79)(80)(81)(82), including VACV inhibition (83,84). Rotenone has been extensively studied and validated as a complex I inhibitor of the mitochondrial electron transport chain (85). Rotenone toxicity in humans is moderate, but it can be used therapeutically at doses posing very low safety risks and has been studied as a potent antiviral (86). Importantly, six of the compounds from the ReFRAME library and all of the compounds from the NPC library found to have inhibitory activity against VACV were confirmed to also inhibit MPXV. Some of the differences observed in the compounds' inhibitory activities against VACV and MPXV likely reflect different readouts to assess their inhibitory activity. The ability to use fluorescence or luciferase reporter genes and the lower level of biocontainment required to work with VACV support the use of the bireporter expressing rVACV Scarlet/Nluc and rVACV GFP/Nluc in high throughput screening (HTS) format to identify compounds with inhibitory activity against orthopoxviruses. Importantly, the compounds from the ReFRAME and NPC libraries we tested are FDA approved for use in humans, which should facilitate the process of bringing them into the clinic. Inhibition of rVACV Nluc/Scarlet or rVACV Nluc/GFP viral replication was evaluated by quantifying Scarlet or GFP (top graphs) and Nluc (bottom graphs) expression at 24 hpi using fluorescent and luciferase microplate readers, respectively. The EC 50 for each compound was calculated using sigmoidal doseresponse curves with GraphPad Prism. The dotted lines indicate 50% inhibition. Data are means and SD of viral inhibition from quadruplicate wells (n = 4). At the same time point (hpi), Scarlet (rVACV Nluc/Scarlet; top) and GFP (rVACV Nluc/GFP; bottom) fluorescent expression in infected cells in the absence (0 mM) or presence of the indicated concentrations of the compounds (maximum concentration and EC 50 ) was visualized using a fluorescence microscope. Representative images are shown. Bars, 100 mm. Magnification, Â20.
In summary, in this study, we demonstrated that some compounds from the ReFRAME and NPC libraries previously identified to have antiviral activity against different RNA virus families also inhibit VACV and MPXV, demonstrating their potential for the treatment of different viral infections. However, despite having been shown for many years to inhibit viral replication in cell culture systems, some of the identified hit compounds have never been proved to have clinical antiviral activity. Future studies in validated animal models of either VACV or MPXV infection (38,63) will be necessary to assess the antiviral activity of the compounds in vivo and to test their therapeutic potential. It should be noted that the compounds we tested target host factors, which will pose a high genetic barrier to the emergence of viral escape mutants, a common event with antivirals directly targeting a viral protein activity. This is illustrated by tecovirimat resistance, which is the result of mutations in the viral gene encoding the target protein F13 (44). In addition, these results open the possibility of combination therapy using compounds targeting viral factors (e.g., tecovirimat) and host factors (e.g., ReFRAME and NPC compounds) for the treatment of orthopoxvirus infections.

MATERIALS AND METHODS
Biosafety. Experiments with MPXV were performed at biosafety level 3 (BSL3) containment laboratories at Texas Biomedical Research Institute (TX Biomed) and were approved by the Institutional Biosafety Committee (IBC) at TX Biomed. All the researchers involved in studies using MPXV were vaccinated with the modified vaccinia virus Ankara JYNNEOS vaccine.
Viruses. The two reporter-expressing rVACVs expressing fluorescent proteins (GFP or Scarlet) and luciferase (Nluc), rVACV Nluc/GFP and rVACV Nluc/Scarlet, respectively, were previously described and characterized (54). We used both fluorescent-expressing rVACVs to assess the feasibility of using both GFP and Scarlet coupled with fluorescence microscopy and plate readers to identify compounds with inhibitory activity (54). We selected Nluc because it is a secreted small luciferase that allow us to detect bioluminescence activity in cell culture supernatants and is therefore used in longitudinal studies (54

Inhibitors of MPXV Microbiology Spectrum
Resources repository and propagated in CV-1 cells. Virus infections were conducted in DMEM containing 2% FBS and 1% PSG.
FFRA. Confluent monolayers of human A549 cells (96-well plate format, 4 Â 10 4 cells/well, in quadruplicate) were infected with 200 PFU/well of MPXV for 1 h at 37°C. After virus absorption, the virus inoculum was removed, and cells were incubated with infection medium containing 3-fold serial dilutions (starting concentration of 50 mM for all compounds except for azoxystrobin, buparvaquone, tryptanthrin, and OSU-03012 [starting at 450 mM]; mycophenolic acid, mycophenolate mofetil, and azaribine [starting at 150 mM]; and antimycin A, narasin, monensin, valinomycin and AVN-944 [starting at 1.85 mM]) of the indicated compounds and 1% Avicel (Sigma-Aldrich). Mock-infected cells and cells infected in the absence of drug were included as internal controls. At 24 hpi, cells were fixed in 10% neutral buffered formalin for 24 h and then permeabilized with 0.5% Triton X-100 in phosphate-buffered saline (PBS) for 10 min at room temperature (RT). Then, cells were blocked with 2.5% bovine serum albumin (BSA) in PBS for 1 h, followed by immunostaining with an anti-VACV A33R polyclonal antibody (BEI Resources, NR-628), Vectastain ABC kit, and DAB (3,39-diaminobenzidine) peroxidase substrate kit (Vector Laboratories), based on the manufacturer's recommendations. Viral infections were determined based on the number of plaques present in each of the 96well plates using an ImmunoSpot plate reader, as previously described (57,58,90,91). Similar assays have been described in the literature to assess the antiviral activities of compounds (92,93). Experiments were conducted using technical quadruplicates. Microsoft Excel was used to calculate the mean and SD of viral inhibition from quadruplicate wells (n = 4). Nonlinear regression curves and EC 50 s were determined using sigmoidal dose response curves in GraphPad Prism (version 9).
Viral titer reduction assays. Confluent monolayers of human A549 cells (24-well plate format, 2 Â 10 5 cells/well, triplicates) were infected with rVACV Nluc/Scarlet, rVACV Nluc/GFP, or MPXV at an MOI of 0.01 or 3 for 1 h at 37°C. After virus absorption, the virus inoculum was removed, and cells were incubated with infection medium containing 10-fold serial dilutions of tecovirimat (starting concentration of 100 mM). Mock-infected cells and cells infected in the absence of drug were included as internal controls. At 24, 48, and 72 hpi, cell culture supernatant was collected, and Nluc activity was determined by adding Nano-Glo luciferase substrate (Promega) and quantified in a microplate reader (rVACV Nluc/ Scarlet or rVACV Nluc/GFP). Viral titers in the cell culture supernatants were determined by plaque assay in CV-1 cells. Briefly, cells were infected with 10-fold serially diluted supernatants for 1 h. After viral adsorption, cells were overlaid with medium containing 1% Avicel (Sigma-Aldrich). Cells were fixed in 10% neutral buffered formalin for 24 h and then permeabilized with 0.5% Triton X-100 in PBS for 10 min at room temperature. Next, cells were blocked with 2.5% BSA in PBS for 1 h and immunostained with an anti-VACV A33R polyclonal antibody (BEI Resources, NR-628) and developed with an anti-rabbit immunoglobulin Vectastain ABC kit and a DAB peroxidase substrate kit (Vector Laboratories), following the manufacturer's recommendations. Experiments were conducted using technical triplicates. Microsoft Excel was used to calculate the mean and SD of viral inhibition from triplicate wells (n = 3). Cell viability assays. MTT (CellTiter 96 nonradioactive cell proliferation assay; Promega) and XTT (cell viability and proliferation assay; Sigma-Aldrich) assays were used to determine A549 cell viability as previously described (58). Briefly, confluent monolayers of human A549 cells (96-well plate format, 4 Â 10 4 cells/well, in quadruplicate) were incubated with 100 mL of infection medium containing 3-fold dilutions of compounds (starting concentration of 50 mM for all compounds except for azoxystrobin, buparvaquone, tryptanthrin, and OSU-03012 [starting at 450 mM]; mycophenolic acid, mycophenolate mofetil, and azaribine [starting at 150 mM]; and antimycin A, narasin, monensin, valinomycin, and AVN-944 [starting at 1.85 mM]) or with 0.1% DMSO as a vehicle control. Then, plates were incubated at 37°C with 5% CO 2 for 48 h, treated with 15 mL of dye solution for the MTT assay or 50 mL of XTT labeling reagent for the XTT assay, and incubated at 37°C for an additional 4 h. Stop solution was then added to the MTT assay to halt the reaction. Absorbance (570 nm) in each of the wells was measured using a microplate reader (BioTek Synergy). Cell viability was determined as the percentage of values for DMSO vehicle-treated cells. Nonlinear regression curves and the CC 50 at 48 h were determined using GraphPad Prism software (version 9).
Statistical analysis. GraphPad Prism software (version 9) was used for data analysis. CC 50 and EC 50 s were calculated using sigmoidal dose-response curves, and the selective index (SI) of each compound was determined by dividing the CC 50 by the EC 50 s. Significance was determined by standard unpaired Student's t test.