Novel Antiviral Efficacy of Hedyotis diffusa and Artemisia capillaris Extracts against Dengue Virus, Japanese Encephalitis Virus, and Zika Virus Infection and Immunoregulatory Cytokine Signatures

Currently, there are no specific therapeutics for flavivirus infections, including dengue virus (DENV) and Zika virus (ZIKV). In this study, we evaluated extracts from the plants Hedyotis diffusa (HD) and Artemisia capillaris (AC) to determine the antiviral activity against DENV, ZIKV, and Japanese encephalitis virus (JEV). HD and AC demonstrated inhibitory activity against JEV, ZIKV, and DENV replication and reduced viral RNA levels in a dose–responsive manner, with non-cytotoxic concentration ranging from 0.1 to 10 mg/mL. HD and AC had low cytotoxicity to Vero cells, with CC50 values of 33.7 ± 1.6 and 30.3 ± 1.7 mg/mL (mean ± SD), respectively. The anti-flavivirus activity of HD and AC was also consistent in human cell lines, including human glioblastoma (T98G), human chronic myeloid leukemia (K562), and human embryonic kidney (HEK-293T) cells. Viral-infected, HD-treated cells demonstrated downregulation of cytokines including CCR1, CCL26, CCL15, CCL5, IL21, and IL17C. In contrast, CCR1, CCL26, and AIMP1 were elevated following AC treatment in viral-infected cells. Overall, HD and AC plant extracts demonstrated flavivirus replication inhibitory activity, and together with immunoregulatory cytokine signatures, these results suggest that HD and AC possess bioactive compounds that may further be refined as promising candidates for clinical applications.


Introduction
Over a third of the global population are at risk of flavivirus infection [1][2][3]. In the recent decade, flaviviruses have caused several global epidemics with high morbidity and mortality rates, in which a majority of these epidemics were caused by dengue virus (DENV), Zika virus (ZIKV), and Japanese encephalitis virus (JEV) [4,5]. Flaviviruses are encoded by a positive-sense single-stranded RNA that can immediately hijack the ribosome for translation [4,6]. The vast majority of flavivirus infections involve an intermediate host prior to transmission to humans through the biting by arthropods, such as mosquitoes in the case of DENV, ZIKV, and JEV [4,7].
Dengue is the most prevalent mosquito-borne, extensively spread, and highly endemic viral infectious disease throughout tropical and sub-tropical regions [8,9]. There are four antigenically distinct serotypes of DENV, designated as DENV-1, DENV-2, DENV-3, and DENV-4; all cause a spectrum of symptoms [10,11]. The most severe clinical syndrome can manifest in the form of dengue shock syndrome (DSS), with the most prominent clinical presentation being plasma leakage with hypovolemic shock and multi-organ failure [12][13][14]. Dengue is endemic in more than 100 countries and the incidence of dengue has significantly Roswell Park Memorial Institute 1640 (RPMI, Gibco, Grand Island, NY, USA) Medium supplemented with 15% FCS and 100 units/mL penicillin. Human chronic myeloid leukemia (K562) cell lines were grown in RPMI containing 10% FCS and 100 units/mL penicillin. Human embryonic kidney (HEK-293T) cell lines were cultured in Dulbecco's Modified Eagle Medium (DMEM, Gibco, Paisley, UK), supplemented with 10% FCS and 100 units/mL penicillin. All the FCS was heat-inactivated without antibiotics, and all cell lines were cultured at 37 • C in a humidified atmosphere of 5% CO 2 .

Cell Cytotoxicity Assay and CC 50 Determination
The cytotoxicity of herbs was determined by using MTT cell proliferation assay kit (Cayman Chemical, Michigan, MI, USA), following the manufacturer's protocol. Briefly, cells were seeded in a 96-well plate at a density of 2 × 10 4 cells/well. After 24 h, cells were overlaid with media containing various concentrations (0.1-100 mg/mL) of herb or EMEM supplemented with 10% FCS and were cultured for an additional four days at 37 • C in a humidified atmosphere of 5% CO 2 . A total of 10 µL of MTT reagent was added into the cell culture supernatant. The MTT-reagent-treated cells were cultured at 37 • C in a humidified atmosphere of 5% CO 2 . Four hours later, 100 µL of crystal dissolving solution was added into the cell culture supernatant and the mixture was incubated for 18 h at 37 • C in a humidified atmosphere of 5% CO 2 . The absorbance of each sample was measured at 570 nm using a Synergy microplate reader (BioTek). Cytotoxicity was determined with the following equation using absorbance: cell viability (%) = [ OD value of treated cells − mean OD of wells without cells]/[ mean OD value of untreated cells − mean OD of wells without cells] ×100% (OD represents optical density). Half maximal cytotoxic concentration (CC 50 ) value was calculated using GraphPad Prism. Each assay was conducted in double replicates and three experimental repeats.

Virus Infection
Vero 9013 cells (2 × 10 4 cells/well) were cultured in a 96-well plate at 37 • C in a humidified atmosphere of 5% CO 2 for 24h. Subsequently, the cells were infected with a ten-fold serial dilution of virus stock at multiplicity of infection (MOI) values of 0.1, 0.01, and 0.001, respectively. After 1 h incubation, cells were washed twice with EMEM and supplemented with 200 µL fresh EMEM/10% FCS and cultured at 37 • C in a humidified atmosphere of 5% CO 2 . At 24, 48, 72, and 96 h post infection (h p.i.), the cell culture supernatants were harvested, and viral titer (PFU) was determined via plaque assay.

Plaque Assay
The plaque assay was performed to determine the levels of infectious viral particles. Viral supernatant from infected cells was harvested on day 3 (72 h p.i.) after inoculation for JEV OH0566 and ZIKV PRVABC59 and day 4 (96 h p.i.) after inoculation for DENV-2 TL-30. Briefly, 25 µL of virus-infected cell culture was added into 225µL EMEM, followed by a ten-fold serial dilution in EMEM. A total of 100 µL of serial diluted liquid was inoculated onto monolayers of Vero cells in a 12-well plate. Next, the plates were incubated at 37 • C in a humidified atmosphere of 5% CO 2 up to 60 min, and after incubation, 2 mL of overlay  [47]. The number of plaques were counted with the naked eye and viral titer was defined as plaque-forming units per milliliter (PFU/mL).

Antiviral Assay and EC 50 Determination
Vero 9013 cells were seeded in 96-well plates (2 × 10 4 cells/well) and incubated at 37 • C in a humidified atmosphere of 5% CO 2 . After 24 h, the cells were infected with each virus at multiplicity of infection (MOI) of 0.01 and incubated for 1 h. The cells were washed with EMEM and cultured in a culture medium containing various concentrations (0.1-10 mg/mL) of herb. After 3 days (JEV OH0566 and ZIKV PRVABC59) and 4 days (DENV-2 TL-30) of incubation, the cell culture supernatants were harvested for viral progeny determination via plaque assay and quantitative detection of viral RNA copy using real-time RT-qPCR. Inhibition (%) was determined by using the following equation: Inhibition (%) = (viral titer (log 10 PFU/mL) of treated culture supernatants/mean viral titer (log 10 PFU/mL) of untreated viral-infected control culture supernatants) ×100%. Half maximal effective concentration (EC 50 ) value was calculated using GraphPad Prism. Selectivity index (SI) for herbs was determined as the ratio of CC 50 :EC 50 . Each assay was conducted in double replicates and three independent experiments.

Viral RNA Extraction
Viral RNA was extracted from 100 µL of virus-infected cell cultures using Quick Viral RNA kit (Zymo research, Irvine, CA, USA), following the manufacturer's protocol. The eluted RNA samples were either used immediately or stored at −80 • C pending analysis.

Real-Time Quantitative Reverse-Transcription Polymerase Chain Reaction
A range of ten-fold serial dilution of in vitro transcribed RNA from 10 7 to 10 3 was used to generate the standard curve [48]. Gene-specific primers and probes targeting the envelope protein for DENV-2 TL-30 [48], ZIKV PRVABC59 [49], and JEV OH0566 [50] were used (Supplementary Table S1). The viral RNA levels were defined as log10 viral genome copies per mL. PCR master mix consisted of 5 µL of RNA, 5 µL of TaqMan Fast Virus 1-Step Master Mix (Applied Biosystems, Waltham, MA, USA), 0.25 µL of 100 µM forward and reverse primer, 0.5 µL of 10 µM probe, and 9 µL of nuclease-free water. The experiment and the following real-time PCR program on ABI instrument was set up as follows: 50 • C 5 min, 1 cycle; 95 • C 20 s, 1 cycle; 95 • C 3 s, 60 • C 30 s, 40 cycles. The mixture was added to the reaction to detect the amplification of target viral RNA (QuantStudio 7 Flex, Thermo Fisher Scientific, Waltham, MA, USA). The real-time RT-qPCR results were analyzed with the QuantStudio™ Real-Time PCR Software ver. 1.1 and the amplification plots were reviewed for baseline and threshold value correction. All isolated RNA and synthetic RNA were stored in −80 • C.

Time of Drug Addition Assay
Vero 9013 cells at a density of 2 × 10 4 cells per well were seeded on 96-well plates and incubated for 24 h at 37 • C in a humidified atmosphere of 5% CO 2 . Cells were infected with each virus at MOI 0.01 and treated with each herb (5 mg/mL) at (1) one hour prior to infection, (2) simultaneously, or (3)  with fresh medium containing each herb (5 mg/mL). After an additional 3 days for ZIKV PRVABC59 and JEV OH0566, or 4 days for DENV-2 TL-30, the cell culture supernatants were harvested for viral progeny determination via plaque assay. The time of incubation for each virus was determined as the optimal time for the detection of virus in the supernatant (data not shown). The pre-and co-treated cells were not washed, and there remains a possibility of unbound virus in the supernatant; as such, the virus mixture in the corresponding untreated viral-infected control supernatant was not removed, for synchronizing the infection in the same condition. However, the unbound virus in the post-treated group was removed; accordingly, the corresponding untreated viral-infected control cells were washed by fresh medium to remove the unbound virus.

RT 2 Profiler PCR Arrays
Total RNA from T98G cells was isolated using Qiagen RNeasy Mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The concentration of eluted RNA samples was measured using Qubit®RNA BR Assay Kits (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). cDNA synthesis was carried out using 3.0 µg total RNA sample as a template with the RT 2 First Strand Kit (Qiagen, Maryland, MD, USA). Human Inflammatory Cytokines and Receptors RT 2 Profiler PCR Arrays (PAHS-011ZC-12; Qiagen, Maryland, USA) were used in the present study. A 96-well plate which contained 84 target genes, a housekeeping gene panel to normalize array data (HK1-5), a genomic DNA control (GDC), replicate reverse-transcription controls (RTC), and replicate positive PCR controls (PPC) was used. Custom RT 2 Profiler PCR Arrays were performed on the StepOnePlus quantitative PCR cycler (Thermo Fisher Scientific, Waltham, MA, USA) using the RT 2 SYBR Green qPCR Mastermix (Qiagen, Maryland, ME, USA), with a thermal cycling condition of 95 • C for 10 min followed by 40 cycles of 95 • C for 15 s and 60 • C for 1 min, followed by a melting curve acquisition step of 95 • C, 1 min; 65 • C, 2 min (optics off); 65 • C to 95 • C at 2 • C/min. Briefly, cDNA synthesis reaction (20 µL per sample) was diluted in 91 µL of RNase-free water, and 102 µL of the aliquot was mixed with 1248 µL of RNase-free water and 1350 µL of 2 × RT 2 SYBR green RT 2 Mastermix. Then, 25 µL of the PCR components mix was added to each well of the RT 2 Profiler PCR Array. Ct values of all samples were exported from the qPCR instrument and uploaded into GeneGlobe Data Analysis Center. Changes in the expression of gene of interest (GOI) normalized to that of the housekeeping gene (HKG) were analyzed using 2 −∆∆C T method.

SYBR Green Real-Time Quantitative Polymerase Chain Reaction
Of the differentially expressed genes (DEGs) analyzed, the top 10 DEGs with the highest discrepancy levels from the controls were selected for further study. GAPDH was considered as a constitutive housekeeping gene. Specific primers and probes targeting these genes were used (Supplementary Table S2). SYBR green qPCR was performed on the StepOnePlus quantitative PCR cycler (Thermo Fisher Scientific, Waltham, MA, USA) using the Thunderbird TM SYBR qPCR Mix (Toyobo, Osaka, Japan), with a thermal cycling condition of 95 • C for 60 s followed by 40 cycles of 95 • C for 15 s and 60 • C for 60 s, followed by a melting curve acquisition step of 95 • C, 1 min; 65 • C, 2 min (optics off); 65 • C to 95 • C at 2 • C/min. Reaction mixture was setup as 10 µL thunderbird SYBR qPCR Mix, 6 pmol forward primer, 6 pmol reverse primer, 2 µL cDNA solution, 0.4 µL of 50X ROX reference dye, and 1.6 µL DDW. Fold changes were analyzed using 2 −∆∆C T method. Briefly, (1) ∆C T for on each array for each gene: ∆C T = C T (a target gene)-C T (GAPDH), (2) average ∆C T for each gene within a group, (3) ∆∆C T for each gene between groups: ∆∆C T = ∆C T (a target sample)-∆C T (a reference sample), (4) fold change: 2 −∆∆C T was determined.

Statistical Analysis
Statistical analysis was performed using GraphPad Prism, version 8.4.3 (GraphPad, San Diego, CA, USA), with a 5% level of significance and two-tailed p values. Values were presented as mean ± standard deviation (SD). Logarithmic transformation of the data was carried out to obtain an approximately normal distribution of the viral titers (PFU) and viral genome copy values, and data were tested for normal distribution using the Shapiro-Wilk test. Log 10 transformed viral titers (PFU) and viral genome copy values were analyzed either in two-group or multiple-group comparisons. Two-group comparisons were analyzed using Student's t-test. Multiple-group comparisons were analyzed by running both parametric (ANOVA) and non-parametric (Kruskal-Wallis) statistical tests with Dunn's and Tukey's post hoc tests. Differences in statistical significance were indicated with asterisks: single asterisk (*) indicates a p value of less than 0.05 (*p < 0.05); double asterisks (**) indicate a p value of less than 0.01 (** p < 0.01); triple asterisks (***) indicate a p value of less than 0.001 (*** p < 0.001); quadruple asterisks (****) indicate a p value of less than 0.0001 (**** p < 0.0001); and ns indicates a p value of over 0.05 (p > 0.05). Number of replicates per experiment is indicated in each figure legend.

Optimizing Multiplicities of Infection (MOI) and Viral Supernatant Harvesting Time Points
To determine the optimal MOI capable of establishing persistent infecting and viral su-

Antiviral Efficacy and Cytotoxicity in Vero 9013 Cells
Both HD and AC manifested low cytotoxicity to Vero 9013 with CC 50 values of 33.66 ± 1.57 and 30.32 ± 1.74 mg/mL (mean ± SD). Both viabilities increased with decreasing concentration of herbs in a dose-dependent manner ( Figure 1). Balancing between toxicity and viral replication inhibition, a range of concentrations less than or equal to 10 mg/mL of the herbs demonstrated low toxicity, with cells that were chosen for the subsequent antiviral experiments.
The antiviral activity of HD and AC was confirmed by measuring their treatment against viral infection at a range of non-cytotoxic concentrations (0.1~10 mg/mL) in Vero 9013 cells. After each virus adsorption at an MOI of 0.01, Vero 9013 cell monolayers were incubated in the absence or presence of increasing doses of HD or AC as indicated. Supernatants were collected to determine infectious viral titer via plaque assay at 72 h p.i. for JEV OH0566 and ZIKV PRVABC59 and 96 h p.i. for DENV-2 TL-30, respectively.
The antiviral activity of HD and AC was confirmed by measuring their treatment against viral infection at a range of non-cytotoxic concentrations (0.1 ~ 10 mg/mL) in Vero 9013 cells. After each virus adsorption at an MOI of 0.01, Vero 9013 cell monolayers were incubated in the absence or presence of increasing doses of HD or AC as indicated. Supernatants were collected to determine infectious viral titer via plaque assay at 72 h p.i. for JEV OH0566 and ZIKV PRVABC59 and 96 h p.i. for DENV-2 TL-30, respectively.
The antiviral potency and selectivity of HD and AC were evaluated by using plaque assays. The EC 50 , CC 50 , and selectivity index (SI) values of HD and AC are summarized in Table 1. Analysis of EC 50 and SI values showed that HD and AC had high efficacy and safety margins against the three flavivirus strains. For HD treatment, the EC 50 values of viral replication inhibition ranged from 3.4 ± 0.03 mg/mL for DENV to 7.0 ± 0.3 mg/mL for JEV OH0566. For AC treatment, they ranged from 1.78 ± 0.68 mg/mL for DENV-2 TL-30 to 6.5 ± 0.3 mg/mL for JEV OH0566, and were dependent on the species of flavivirus tested. The results revealed that HD and AC had the highest SI against DENV-2 TL-30, with values of 9.8 and 18.9, respectively. These results suggested that HD and AC inhibited JEV OH0566, ZIKV PRVABC59, and DENV-2 TL-30 plaque formation at concentrations with limited cytotoxicity. As HD and AC inhibited plaque formation, the results suggest that the herbs could inhibit virus propagation.

Quantification of Viral RNA Genome Copy by RT-qPCR
To determine the antiviral activity of HD and AC viral genomic RNA in the culture supernatant of herb-treated and untreated cells, infected Vero cells were measured by RT-qPCR. HD and AC significantly inhibited all the tested virus strains with lower viral RNA genome copies by ten-fold in the supernatants at concentrations of 5 and 10 mg/mL ( Table 2.), as compared to the untreated viral-infected control. Significant (1.3 ± 1.5 (p < 0.05) and 2.8 ± 1.1 (p < 0.0001) log 10 viral RNA copies/mL (mean ± SD)) reductions in viral RNA titer were observed in HD-treated DENV-2 TL-30 infected cells at concentrations of 1 and 0.1 mg/mL. However, HD did not significantly suppress the viral RNA titer of JEV OH0566 and ZIKV PRVABC59 at concentrations of 1 and 0.1 mg/mL (Figure 3.).
AC treatment (1 mg/mL) inhibited growth of DENV-2 TL-30 and JEV OH0566, with a log 10 viral RNA copies/mL difference of 1.4 ± 1.4 (p < 0.05) and 0.9 ± 0.5 (p < 0.05), respectively. However, AC (1 mg/mL) did not reduce the ZIKV PRVABC59 viral RNA level. At a concentration of 0.1 mg/mL, AC treatment did not induce a significant reduction in viral RNA level against all the tested virus strains. While there was no statistical significance, viral RNA was reduced by 10-fold genomic copies. As compared to the untreated viral-infected control, viral RNA titers had an inverse association with HD and AC concentrations in a dose-dependent manner. This finding indicates that HD and AC effectively inhibit JEV OH0566, ZIKV PRVABC59, and DENV-2 TL-30 replication in vitro. HD: Hedyotis diffusa; AC: Artemisia capillaris.

Time of Drug Addition
To further evaluate the time point of infection by which the herbs exhibited antiviral activity against flavivirus infection, the time of drug addition study was conducted. In the pre-treatment study, at a non-cytotoxic concentration (5 mg/mL) of HD and AC, no infectious virus particles in the cell culture supernatant were detected for ZIKV and DENV-2 strains. In comparison with the untreated viral-infected control, pre-treatment of HD inhibited JE viral growth (89.5% log reduction value (LRV) (6.1 ± 1.7 log 10 PFU/mL, p < 0.001)). Similarly, pre-treatment of AC also demonstrated an inhibitory activity of 94.8% LRV (6.5 ± 0.8 log 10 PFU/mL, p < 0.001) on JEV (Figure 4). A similar significant reduction pattern of infection for HD and AC against all the three virus strains was observed in coand post-treatment processes (p < 0.0001). Post-treatment of the herbs also significantly suppressed all the viral replication (p < 0.0001). Compared with viral plaque in the cotreatment group, HD pre-treatment significantly reduced viral titers, with a log 10 PFU/mL difference of 5.1 ± 0.2 (p < 0.001) for JEV. Similarly, the viral titers of AC pre-treatment were significantly lower that of co-treatment, with a log 10 PFU/mL difference of 5.4 ± 0.9 (p < 0.001) for JEV. Co-treatment of HD or AC exhibited anti-ZIKV and DENV-2 activities; by comparison, pre-treatment showed significantly higher (p < 0.001) and absolute inhibition of viral plaque formation. Although co-and post-treatment of the herbs showed strong inhibition of the tested viral replication, pre-treatment demonstrated comparatively higher antiviral efficacy (Table 3).     Inhibition of HD and AC at 5 mg/mL concentration against JEV OH0566, ZIKV PRVABC59, and DENV-2 TL-30 at an MOI of 0.01 in Vero cells was quantified by plaque assay in pre-co-, and post-treatment processes. Log 10 transformed viral titer values of herb-treated viral-infected cell culture supernatants were compared with that of the untreated viral-infected control. The inhibition was calculated as follows: inhibition (%) = (viral titer (log 10 PFU/mL) of treated viral-infected cell culture supernatants/mean viral titer (log 10 PFU/mL) of untreated viral-infected cell culture supernatants) ×100%.
In this study, pre-treatment of the herbs significantly reduced viral titers in the cell culture supernatant among all the methods tested. Additionally, both herbs strongly inhibited infectious progeny replication in all the tested virus strains, regardless of the time of drug addition.
In the K562, T98G, and HEK293T cell lines, significant reduction in viral titers following the herb treatments were demonstrated; however, the inhibition level of viral replication was higher in T98G cell lines in comparison to the other two human cells. Therefore, this experiment confirmed the antiviral potency of HD and AC in human cells against JEV, ZIKV, and DENV-2 infection.

Screening of Differentially Expressed Genes (DEGs)
Total RNA isolated from T98G cells at 6 h post treatment of the herbs against DENV-2 TL-30 strain (MOI = 1) infection were analyzed by using RT 2 Profiler PCR array to evaluate the expression of 84 genes associated with human inflammatory cytokines and receptors (Supplementary Table S4). Comparative analysis of the gene expression profile demonstrated the broadest dysregulation of genes in the number of differentially expressed genes (DEGs), as shown in Supplementary Figure S2. The top ten DEGs were then selected for further studies.

Herbs Altered Gene Expression Profile Response to Viral Infection
In the next series of experiments, the selected DEGs were validated by SYBR Green real-time PCR (Table 4). Viral-infected or uninfected T98G cells were treated or untreated with herbs, and their proliferation potentials were analyzed at 6 min (m p.t.) and 6 h (h p.t.) post-treatment. The experimental data were gathered from the following five pairwise comparisons: (1) (Figure S4.). However, the majority of the tested genes reversed the expression profiles displayed by uninfected cells by a statistically significant degree following herb treatment in the viral-infected cells. Uninfected cells were exposed to herbs and analyzed in parallel to exclude gene expression change induced by the direct effect of herb treatment. As compared to untreated cells, HD treatment repressed CCL26 (3.7-fold, p = 0.0006) and CCL5 (1.3-fold, p = 0.015) at 6 m p.t and 6 h p.t., respectively ( Figure S5). The expression of major genes tested were elevated (range 15.6-68.2-fold) upon AC treatment at 6 m p.t., except CCL5 and CXCL13; however, uninfected cells showed no significant effect of herbs on the gene expression at 6 h p.t.

viral-infected cells vs. cells, (2) herb-treated viral-infected cell vs. cells, (3) herb-treated cells vs. cells, (4) herb-treated viral-infected cells vs. herb-treated cells, and (5) herb-treated viral-infected cells vs. viral-infected cells.
To control for non-virus-specific immunomodulatory effects of herbs, the changes in gene expression were analyzed between herb-treated viral-infected and treated uninfected cells. Compared to the herb response in uninfected cells, several genes following HD treatment showed small changes at 6 m p.  Figure S6). Upon AC treatment, the majority of gene expressions were largely decreased (range: 16.4-to 86.6-fold) at 6 m p.t., and increased little (IL21: 1.5-fold, p = 0.002; CCL15: 1.7-fold, p = 0.02) at 6 h p.t. Relatively few gene expressions were similar in viral-infected and uninfected cells following herb treatment. Therefore, the gene expression differences were detected after herb treatment in viral-infected cells, and their uninfected counterparts were associated with the human response to viral infection.
To assess the effect of antiviral treatment, pairwise comparison analyses of herbtreated and untreated viral-infected cells were undertaken. Compared to untreated viralinfected cells, during the early stages of HD treatment (6 m p.t) in viral-infected cells, CCR1 (1.7-fold, p = 0.01), CCL26 (1.6-fold, p = 0.004), and CCL15 (1.6-fold, p = 0.02) were repressed ( Figure 6, Table 4  Multiple comparison analysis indicated that up/downregulated genes by viruses in infected cells relative to their uninfected counterparts demonstrated statistically significant reversal in viral-infected cells following herb treatment ( Figure 6, and Figure S3). Elevated cytokines and receptors including CCL26, CCL15, and CCR1 were responsive to statistically significant reversal following exposure of infected cells to HD for 6 min. CCL15 was also significantly reversed upon AC treatment for 6 min.

Discussion
Medicinal plants have historically been valuable sources of potential therapeutic products and continue to be an attractive source of discovery of novel antiviral compounds [51]. While Hedyotis diffusa (HD) and Artemisia capillaris (AC) have been identified in TCM to possess "heat-clearing" and detoxifying effects, there is limited scientific evidence of the antiviral potential of HD and AC against Flaviviridae. Since Vero cells are highly susceptible to a variety of viruses and are incapable of producing any type I interferons, they have been proven to be the first-choice cell model for various types of lifethreating emerging viral pathogens [52,53]. In this study, HD and AC were used to determine the utility in viral replication inhibition by using an in vitro model of infection.
A selectivity index of at least four has been recommended as an index with good antiviral selectivity [54,55]. In this context, HD and AC demonstrated SI values of >4, indicating that the herbs are potentially useful antivirals for JEV, DENV, and ZIKV. Time of drug addition assays with the herbs against various virus strains were performed to max- Multiple comparison analysis indicated that up/downregulated genes by viruses in infected cells relative to their uninfected counterparts demonstrated statistically significant reversal in viral-infected cells following herb treatment (Figures 6 and S3). Elevated cytokines and receptors including CCL26, CCL15, and CCR1 were responsive to statistically significant reversal following exposure of infected cells to HD for 6 min. CCL15 was also significantly reversed upon AC treatment for 6 min.

Discussion
Medicinal plants have historically been valuable sources of potential therapeutic products and continue to be an attractive source of discovery of novel antiviral compounds [51]. While Hedyotis diffusa (HD) and Artemisia capillaris (AC) have been identified in TCM to possess "heat-clearing" and detoxifying effects, there is limited scientific evidence of the antiviral potential of HD and AC against Flaviviridae. Since Vero cells are highly susceptible to a variety of viruses and are incapable of producing any type I interferons, they have been proven to be the first-choice cell model for various types of life-threating emerging viral pathogens [52,53]. In this study, HD and AC were used to determine the utility in viral replication inhibition by using an in vitro model of infection.
A selectivity index of at least four has been recommended as an index with good antiviral selectivity [54,55]. In this context, HD and AC demonstrated SI values of >4, indicating that the herbs are potentially useful antivirals for JEV, DENV, and ZIKV. Time of drug addition assays with the herbs against various virus strains were performed to maximize the possibility of antiviral efficacy. Compared with the co-treatment (0 h p.i.), the addition of the herbs markedly inhibited infectious progeny production following pre-treatment (−1 h p.i.) against all the tested virus strains. The greatest inhibition of virus attachment was inhibited with the addition of the herbs at the 1 h earlier time points (−1 h p.i.) than when they were added at 0 h p.i. (Figure 4). The finding indicated that the herbs may suppress the virus uptake and inhibit the viral attachment to cells in the initial step in the infection process. HD and AC inhibition of viruses by the co-treatment process indicated the virucidal potential. On the other hand, HD and AC significantly reduced the viral titers by post-treatment process, indicating that both herbs may suppress the infectious progeny replication. This data suggests that the antiviral efficacy of HD and AC by acts by inhibiting viral attachment, entry, and replication, with the potential of involvement of an array of immunological cascades in each phase of the viral infection cycle.
Apart from Vero cells, which is a cell line that is commonly used in flavivirus propagation assays, various human cell lines (K562, HEK293T, and T98G) were used to investigate the antiviral activity of HD and AC. Viral growth inhibition was observed in the three cell lines infected with all the tested virus strains in the presence of herb treatment. Notably, the addition of HD and AC herbs resulted in 100% inhibition of viral plaque formation against all the tested virus strains in T98G cells ( Figure 5). After entering the CNS, neurotropic flaviviruses can infect neurovascular unit cells such as astrocytes, and lead to general neuroinflammation and blood-brain barrier (BBB) impairment [56,57]. Astrocytes are mediators of neuroinflammation, and these infections may further disrupt neuronal activity [58,59]. Numerous neurotropic flaviviruses reportedly target astrocytes, such as DENV [60], ZIKV [61,62], and JEV [63]. T98G cell lines have been proven useful as a human astrocyte model, which have similar morphological and functional properties in comparison to other human astrocytes [64]. In this context, T98G was chosen as a model cell line for DENV-2 infection. These results suggest that HD and AC may represent a novel therapeutic agent to mitigate CNS manifestations of viral infection.
Acute flavivirus infections are characterized by extensive inflammation and chemokine expression, which is associated with outcomes including enhanced viral dissemination, tissue damage, and viral burden [65,66]. Downregulated expression during HD treatment in viral-infected cells as compared to their untreated viral-infected counterparts includes CCR1, CCL26, and CCL15 at 6 m p.t, and CCL5, IL21, and IL17C at 6 h p.t. Upon AC treatment, compared to untreated viral-infected samples, CCL5, CCl26, IL21, and CCL15 levels were decreased at 6 m p.t. In contrast, CCR1, CCL26, and AIMP1 levels were elevated following AC treatment in viral-infected cells. Of the genes demonstrating downregulated expression level in HD-treated viral-infected cells, chemokine CCL5 ( Figure 6) expression was significantly lower. C-C motif chemokine ligand 5 (CCL5) has been identified as a chemokine that is localized in white matter tracts undergoing demyelination following viral infection. The chemokine has been hypothesized to participate in viral pathogenesis by attracting activated leukocytes and activated macrophages into the CNS, and in turn, leading to neurological impairment. Selective neutralization of CCL5 resulted in diminished leukocyte infiltration into the CNS and reduced neurological disability in a viral model of multiple sclerosis [67]. The cytokine CCR1 had a role in the development of detrimental pulmonary responses during respiratory syncytial virus (RSV), and the absence of CCR1 may lead to lead to cellular damage and/or altered immune activation [68]. In contrast, CCL15 was significantly associated with the expression of Hepatitis B virus X protein and negatively correlated with clinical outcome for HBV-positive hepatocellular carcinoma (HCC) patients [69,70]. Overall, together with inhibiting virus propagation, the cytokine signature suggests that the HD and AC could potentially alter the expression of proinflammatory cytokines during viral infection, which in turn results in alleviating pathogenicity in the target tissue.

Conclusions
Hedyotis diffusa (HD) and Artemisia capillaris (AC) inhibited flaviviral replication and potentially limit the inflammatory response generated by DENV infection. The results suggest that HD and AC should be viable candidates for further in vivo study to demonstrate their efficacy against viral infection and pathogenesis. Our results indicate that the herbs have efficacies as potential agents for prophylaxis or treatment of flavivirus infection Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/plants11192589/s1, Table S1: Oligonucleotide primers and fluorogenic probes used in the JEV, ZIKV, and DENV-2 real-time RT-qPCR assays, Table S2: Oligonucleotide primers of the selected inflammatory genes used in SYBR green qPCR, Table S3: Antiviral efficacy of herbs in Vero 9013, Figure S1: Time-dependent viral titer of Vero 9013 cells inoculated with JEV, ZIKV, and DENV-2 at different virus concentrations, Figure S2: Relative expression comparison of 84 genes involved in human inflammatory cytokines and receptors, Figure S3: The fold changes of the selected inflammatory genes in DENV-2 TL-30 infected T98G cells relative to T98G cells, Figure S4: The fold changes of the selected inflammatory genes in herb-treated DENV-2 TL-30 infected T98G cells relative to T98G cells, Figure S5: The fold changes of the selected inflammatory genes in herb-treated T98G cells relative to T98G cells, Figure S6: The fold changes of the selected inflammatory genes in herb-treated DENV-2 TL-30 infected T98G cells relative to herb-treated uninfected cells.

Data Availability Statement:
The dataset is available upon reasonable request to the corresponding author.