Screening of Natural Extracts for Inhibitors against Japanese Encephalitis Virus Infection

The mosquito-borne Japanese encephalitis virus (JEV) causes serious illness worldwide that is associated with high morbidity and mortality. Currently, there are no effective drugs approved for the treatment of JEV infection. Drug-repurposing screening is an alternative approach to discover potential antiviral agents.


RESULTS
Optimization of HCS assay conditions. The HCS assay-optimized conditions contain cell density, infective dose, and assay endpoint. The assay conditions for HCS were optimized as follows: cell density of 10,000 cells per 96-well plate, a multiplicity of infection (MOI) of 0.8, and an endpoint of 24 h postinfection. The signal-to-basal (S/B) ratio, coefficient of variation (CV), and Z factor were 2,177, 8.9%, and 0.72, respectively, which demonstrated that the assay was robust and reproductive for the large-scale screening of novel antiviral compounds against JEV infection.
HCS assay of drug library screening. To reveal novel targets and inhibitors against JEV infection, an immunofluorescence assay (IFA)-based HCS assay was performed by screening a library of 1,034 natural extracts for the anti-JEV activities ( Fig. 1A; Table S1). As shown in Fig. 1B and C, 851 compounds were selected for the second screening because of their solubility in dimethyl sulfoxide (DMSO) at room temperature and absence of any obvious cytotoxicity on Vero cells. Compounds (50 M) that exerted a Ͼ90% inhibition against JEV were defined as prime candidates; based on this criterion, 23 compounds (2.22%) were selected. A screening to reconfirm the results was then carried out using these prime candidates over a broader concentration range (3.125 to 50 M). Eight hit compounds (0.77%) were selected based on their concentration-dependent inhibitory effects and a cell viability of Ͼ80% ( Fig. 2A and B). Among the 8 compounds, ouabain and digoxin were subjected to further investigation because of their high selective indexes (SI, which is defined as 50% cytotoxic concentration [CC 50 ]/50% inhibitory concentration [IC 50 ]) and similar mechanisms (namely, they are cardiac glycoside compounds and act as inhibitors of Na ϩ /K ϩ -ATPase) of Ͼ1,917 and Ͼ969.9, respectively.

Antiviral effects of ouabain and digoxin in different cell lines and viruses.
For AT31 infection of Vero cells, digoxin (392 nM) and ouabain (196 nM) displayed inhibition rates Ͼ90%; the ability to block the virus reduced with a decrease in concentration (Fig. 3A). Similarly, SA14 reproduction could be greatly inhibited when the concentration of digoxin and ouabain reached 392 nM and 98 nM, respectively, Simultaneously, the fluorescence signal increased with diminishing drug concentrations (Fig. 3B).
Under the same conditions of infection, ouabain and digoxin robustly inhibited AT31 and SA14 virus production in a plaque assay, with a reduction of approximately 4 to 6 log units at the highest concentration ( Fig. 3C and E). A sharp decrease in JEV RNA levels was also detected with a quantitative reverse transcription PCR (qRT-PCR) assay. In particular, the attenuated RNA levels of the SA14 strain in the high-dose-, middle-dose-, and low-dose-treated groups were all above 40%, indicating a strong inhibition of viral replication ( Fig. 3D and F).
Further, the inhibitory effect of two compounds on AT31 and SA14 infection was also evaluated in Huh-7 cells and U251 cells. In line with JEV-infected Vero cells, plaque assays were performed to detect the viral titer, and a dose-dependent anti-JEV effect was similarly observed in the two cell lines (Fig. 3G to J).
Ouabain and digoxin inhibit JEV infection during viral RNA synthesis. To determine the cellular process that these drugs target, a time-of-addition experiment was performed (Fig. 4A). As shown in Fig. 4B, neither digoxin nor ouabain showed suppression of JEV in virucidal treatments or pretreatments. However, ouabain showed a robust inhibition of JEV during treatment. Most importantly, when added posttreatment, both ouabain and digoxin exerted full inhibitory effects on JEV infection, suggesting that viral replication was the main stage at which these drugs showed inhibitory activity.
To confirm this hypothesis, Huh-7 cells electroporated with the JEV replicon were treated with the indicated concentration of ouabain and digoxin; we revealed the Our results indicated that both drugs inhibited JEV RNA synthesis in a dose-dependent manner, while neither drug inhibited the initial translation of replicon RNA, confirming that these drugs inhibit JEV infection at the replication stage.
JEV inhibition with ouabain and digoxin occurs via Na ؉ /K ؉ -ATPase. Na ϩ /K ϩ -ATPase is the only target of cardiac glycosides that has been found to date. To confirm whether JEV inhibition occurs via blockade of Na ϩ /K ϩ -ATPase, the infected cells were treated with DMSO and two compounds in the presence of increasing NaCl and KCl, and two compounds inhibited JEV in a dose-responsive manner. Furthermore, at the same concentration of digoxin, JEV inhibition reduced with the decrease of NaCl (Fig.  5A) but with an increase of KCl (Fig. 5B). Under the same experimental conditions, the results of ouabain action were consistent with digoxin ( Fig. 5C and D), which indicated that the JEV-inhibiting effect of both compounds is positively correlated with extracellular NaCl but inversely correlated with KCl. These results confirm that both ouabain and digoxin exert the antiviral effect by targeting the Na ϩ /K ϩ -ATPase.
Ouabain protects against JEV infection-induced lethality in vivo. As ouabain exhibited a stronger inhibitory activity on JEV infections, we further examined the protective effect of ouabain on JEV-induced lethality by using the BALB/c mouse. As anticipated, mice in the JEV-infected, vehicle-treated group started to show symptoms To further relate these protective effects to viral load and histopathological changes in mouse brains, the viral titer was determined and mouse brain sections were collected and assayed at days 7, 10, and 21 postinfection. The results indicated that, during the progression of the disease, ouabain treatment significantly reduced viral load in infected mice compared to untreated infected mice. Notably, no plaques formed in the ouabain-treated group on day 21 postinfection (Fig. 6B). Similarly, apparent damage in the brain, including meningitis, perivascular cuffing, and vacuolar degeneration, was observed in the JEV-infected and vehicle-treated group on day 7 postinfection, while ouabain treatment remarkably alleviated these phenomena (Fig. 6C). These results indicate that alleviation of histopathological changes is accompanied by a reduction in the viral load as well as a reduction in mortality rate, further confirming the curative effects of ouabain on viral encephalitis. The viral loads of brain, blood, and spleen were also evaluated at the early stage of infection. As shown in Fig. 6D, on day 1 postinfection, viral titers in the brain were almost undetectable, and ouabain exhibited little effect on viral titers in the brain at the preclinical stage. Accordingly, ouabain had little effect on decreasing viral titers in the blood and spleen at the earlier time points, which was mostly due to that the detectable viral genome copy numbers were too low to be comparable. Notably, titers in the blood decreased sharply from day 1 to day 3 postinfection, which was in line with the characteristic viremia caused by JEV. These results indicate that ouabain could alleviate histopathological changes and reduce viral loads in the brain, thus protecting mice against JEV-induced lethality and confirming the curative effects of ouabain on viral encephalitis.  FGIN-1-27, an anxiolytic drug that targets the peripheral benzodiazepine receptor, reduced the JEV infection in vitro (15). Drug screening and repurposing has become a very useful approach for identifying antiviral drugs, as it explores novel molecular targets to study virus pathogenesis.

DISCUSSION
To address the urgent need for anti-JEV therapy, we presented a library of natural extracts to check for the ability to inhibit JEV infection. Our high-content screening assay design could identify compounds that inhibit JEV viral entry, translation, and RNA synthesis. In this study, eight hit compounds with SI indexes greater than 10 were found to exert inhibitory effects on JEV. Among these eight compounds, some were previously reported to possess a wide spectrum of pharmacological effects, including antiviral activity. Moreover, some compounds, such as lycorine, emodin, and procyanidin, have been proven to be effective in inhibiting flavivirus or HCV infections via different mechanisms (16)(17)(18)(19)(20). These results demonstrate that our HCS assay was effective and credible.
The top two compounds, FDA-approved Na ϩ /K ϩ -ATPase inhibitors ouabain and digoxin, are cardiac glycosides with similar chemical structures and have been used for the treatment of cardiac arrhythmias and hypotension for more than 200 years. Recently, ouabain and digoxin have been proven to inhibit different kinds of viruses, including enveloped viruses such as coronaviruses, nonenveloped viruses such as reoviruses, DNA viruses such as human cytomegalovirus, positive-sense RNA viruses such as chikungunya virus, and negative-sense RNA viruses such as lymphocytic choriomeningitis virus (LCMV) (21)(22)(23)(24)(25). Notably, we have tried to select drug-resistant variants by serial passaging of JEV using increasing concentrations of digoxin and ouabain, respectively. However, no adaptive mutant was found after 25 passages with either drug. This result suggested that both drugs might exert the antiviral effects by targeting the cellular protein other than the viral protein, making the barrier to resistance more difficult to overcome.
Cardiac glycoside acts via inhibiting the sodium-potassium ion pump, leading to changes in the intracellular concentration of sodium, potassium, and calcium, which have been shown to play essential roles in many cellular biosynthetic signaling and vesicular sorting pathways (26). In this study, ouabain exhibited therapeutic effects on JEV infection in an adult mouse model by decreasing viral loads and alleviating pathological injuries in the brain, which significantly improved the survival rate of JEV-infected mice. We proposed two mechanisms that may contribute to the antiviral effects. First, ouabain may block JEV infection by inducing the cellular stress response. Ouabain treatment in mammalian cells causes the interaction between the inositol 1,4,5-trisphosphate (IP3) receptor and Na ϩ /K ϩ -ATPase to induce calcium oscillations and further activate calcium-dependent transcription factors, such as the NF-B and activator protein (27). Some evidence suggested that cardiac glycosides mediate inflammatory processes via the activation of the phosphoinositide 3-kinase/Akt pathway and the Src/mitogen-activated protein kinase pathway (28,29). These signaling pathways also cause the activation of NF-B. Many of the genes expressed were elicited by activation of NF-B, including innate immune response, growth, and differentiation, which stimulate an antiviral state that might block JEV infection in the end (30). Second, JEV could replicate in the central nervous system (CNS), and infections result in the breakdown of the blood-brain barrier (BBB) along with an influx of inflammatory cells; the breakdown of BBB caused by JEV infection may have allowed BBB-nonpermissive ouabain to enter the brain, bind the murine ATPase ␣2 and ␣3 isoforms (31), and inhibit viral replication in neurons.
Together, these findings provide valuable information that may be used in future clinical trials examining the effects of ouabain on JEV infection. Moreover, our results indicate that ATPase is a promising pharmacological target in JEV infection. administered with ouabain at 3 mg/kg of body weight or a vehicle control every day for 21 consecutive days. Mice were sacrificed at day 1, 3, 7, 10, or 21, depending on the experimental design. Blood, spleen, and brain samples were collected for plaque assay, qRT-PCR, and histopathology investigation (Vectra ss; PerkinElmer).
The viral burden was determined by plaque assay. The samples were weighed and homogenized in 300 l of PBS. The homogenates were clarified by centrifugation (2,000 ϫ g at 4°C) for 15 min and then diluted serially prior to infection of BHK-21 cells. The viral titer was adjusted for sample volume and tissue weight to calculate the titer as PFU/g. Tissue samples and blood from JEV-infected mice were extracted with the RNAprep pure tissue kit (Tiangen; catalog no. DP431) and RNAprep pure blood kit (Tiangen; catalog no. DP433), respectively. JEV RNA levels were determined by one-step qRT-PCR on ABI Real-Time PCR systems using standard cycling conditions. The viral burden was calculated on a standard curve produced using serial 10-fold dilutions of plasmids carrying the infectious clone.
For histopathology investigation, brain tissues were collected from the mice and fixed in 4% paraformaldehyde; the hematoxylin-eosin staining protocol was followed for tissue staining.
Statistical analysis. Student's t test and log rank tests were used to evaluate the statistical significance of differences. A value of P Ͻ 0.05 was considered statistically significant.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, XLSX file, 0.1 MB.