Synergistic anti-SARS-CoV-2 activity of repurposed anti-parasitic drug combinations

COVID-19 pandemic has claimed millions of lives and devastated the health service system, livelihood, and economy in many countries worldwide. Despite the vaccination programs in many countries, the spread of the pandemic continues, and effective treatment is still urgently needed. Although some antiviral drugs have been shown to be effective, they are not widely available. Repurposing of anti-parasitic drugs with in vitro anti-SARS-CoV-2 activity is a promising approach being tested in many clinical trials. Combination of these drugs is a plausible way to enhance their effectiveness. The in vitro anti-SARS-CoV-2 activity of combinations of niclosamide, ivermectin and chloroquine were evaluated in Vero E6 and lung epithelial cells, Calu-3. All the two-drug combinations showed higher potency resulting in up to 4-fold reduction in the half maximal inhibitory concentration (IC50) values compared to individual drugs. Among these combinations, niclosamide-ivermectin achieved the highest inhibitory level of over 99%. Combination synergy analysis showed niclosamide-ivermectin combination to have the best synergy score with a mean Loewe synergy score of 4.28 and a peak synergy score of 24.6 in Vero E6 cells and a mean Loewe synergy score of 3.82 and a peak synergy score of 10.86 in Calu-3 cells. The present study demonstrated the benefit of drug combinations on anti-SARS-CoV-2 activity. Niclosamide and ivermectin showed the best synergistic profile and should be further tested in clinical trials.


Background
The spread of SARS-CoV-2 and the COVID-19 pandemic has swept through countries and continents causing catastrophic loss of lives, public health, livelihood, and economy. Up to March 2021, more than hundred million cases have been reported with over two million deaths [1]. The hope to get through the pandemic and resume normal life relies heavily on vaccine deployment, which will still take months or years in most less-developed countries [2]. One of the reasons for the heavy loss of lives, hospital overload, and public panic is the lack of effective treatment. Remdesivir is now the only antiviral drug with emergency use authorization by US FDA [3]. The drug is, however, not yet widely available. Other FDA-approved drugs are anti-inflammatory targeting host inflammatory responses [4]. More drugs capable of inhibiting SARS-CoV-2 replication are urgently needed not only for treatment but also for reducing viral load and transmission [5]. Many repurposed anti-parasitic drugs have been shown to possess in vitro activity against SARS-CoV-2.
In vitro screenings of FDA-approved drugs have identified a number of anti-parasitic drugs with anti-SARS-CoV-2 activity and potential for drug repurposing for Open Access *Correspondence: prasert.aue@mahidol.ac.th treatment of COVID-19 patients [6,7]. The early hope to get an effective treatment using these drugs was let down by the failure to show clinical benefit of chloroquine in clinical trials [8]. On the other hand, ivermectin has shown promising results in some clinical trials [9][10][11][12][13]. Ivermectin has been shown to cause up to 5000-fold reduction in SARS-CoV-2 replication in vitro [14][15][16]. The drug has been widely used to treat various parasitic diseases in humans and animals for four decades with little safety concern. It was also used in the mass treatment campaign against river blindness (Onchocerciasis) with good safety record [17]. It is therefore, an attractive option for drug repurposing for COVID-19 treatment. Another anti-parasitic drug, niclosamide, showed a good anti-SARS-CoV-2 activity with a high selective index [7,18]. The drug has been shown to exhibit broad antiviral activity against a wide range of viruses [19]. These antiparasitic drugs with potent in vitro anti-SARS-CoV-2 activity are widely available, inexpensive, and considered relatively safe for short-term usage.
The world urgently needs repurposed drug regimens with higher antiviral activity against SARS-CoV-2 to cope with the pandemic. One of the approaches to enhance drug potency is through drug combination. To find a drug combination with good therapeutic potential, we tested combinations of these common drugs for in vitro synergistic activity against SARS-CoV-2.

Chemicals
All drugs were prepared to 10 mM stock solutions in 100% DMSO (Sigma) for niclosamide (N3510, Sigma) and ivermectin (I8898, Sigma), or water for chloroquine (HY-17589, MCE) and stored at − 80 °C. The drugs were diluted to the working concentrations in 2%FBS-MEM or 2%FBS-DMEM/F12 for the experiments in Vero E6 or Calu-3 cells, respectively. The final concentration of DMSO was 0.5% in all experiments.
SARS-CoV-2 (SARS-CoV-2/01/human/Jan2020/Thailand) representing the original Wuhan strain was isolated from nasopharyngeal swabs of a COVID-19 patient in Thailand in the previous study [20]. The protein sequence of surface glycoprotein of SARS-CoV-2/01/human/ Jan2020/Thailand is available at GenBank: QYZ85362.1. Vero E6 cells was used for viral propagation. The supernatants containing virus were harvested when 50% of the infected cells display cytopathic effect (CPE). Centrifugation was performed to remove cell debris. The virus supernatants were aliquoted and stored at -80 °C as a virus stock.

Viral quantifications Plaque assay for SARS-CoV-2
Vero E6 cells were plated in 24-well plates at a density of 1.3 × 10 5 cells per well, which allowed 100% confluence to be reached within 24 hours. The culture medium was removed and 300 μl of serum free-MEM was added. After that, the cells were incubated with 10-fold serial dilution of virus supernatants for 1 hour at 37 °C with 5%CO 2 . Subsequently, the virus inoculums were removed, and the cells were immediately overlaid with 1 ml of 1.56% microcrystalline cellulose (Avicel, RC-591) in 2%FBS-MEM. The infected cells were further incubated in the standard condition for 3 days. At 3 days after infection, the overlaid media were removed. The infected cells were fixed with 10% (v/v) formalin in phosphate-buffered saline for 2 hours. After that, fixed cells were washed in tap water and stained with 1% (w/v) crystal violet in 20% (v/v) ethanol for 5 min. The excess dyes were removed by washing in tap water. The titers of virus were calculated in plaque forming units per ml (pfu/ml).

50% tissue culture infectious dose (TCID 50 ) endpoint dilution assay coupled with ELISA
Before the day of infection, Calu-3 cells were seeded at a density of 2.0 × 10 4 cells/well in 96-well plates. Next day, the culture medium was removed, and the cells were inoculated with 100 μl of half-log10 serial dilution of the virus supernatants for 2 days in the standard condition. Subsequently, the supernatants were discarded, and the infected cells were fixed with the mixture of absolute methanol and acetone in 1:1 ratio for 30 min at 4 °C. The infected cells were detected based on ELISA assay using the antibody against the SARS-CoV-2 nucleocapsid protein (NP) (40143-R001, Sino Biological) and anti-rabbit IgG-conjugated HRP. The TCID 50 titers were calculated following the Reed and Muench method [21].

One-step quantitative reverse-transcription PCR (qRT-PCR)
In this study, one-step qRT-PCR was used as a screening assay to detect the RNA of SARS-CoV-2 directly from the virus supernatants. The procedure was previously described elsewhere [22]. For the sample preparation, the virus supernatants were subjected to disinfection by heat inactivation at 70 °C for 20 min. Then the heat inactivated virus supernatants were diluted for a 1:10 ratio in DNase/RNase free distilled water. Subsequently, one-step qRT-PCR was performed using the Power SYBR one-step kit (Applied Biosystems) in the LightCycler 480 (Roche, LC480) following the kit's instructions for a 10 μl reaction volume. The forward and reverse primers used in this study were N-Fw: 5′-GGG GAA CTT CTC CTG CTA GAAT-3'and N-Rv: 5′-CAG ACA TTT TGC TCT CAA GCTG-3′, respectively. TRIzol-LS (Invitrogen) purified RNA of SAR-CoV-2 stock virus was used as a positive control. The nuclease-free water and mock-infected supernatants were used as no-template control. The thermocycler was run following the instructions of Power SYBR one-step kit. The reverse transcription and the activation of polymerase were performed at 48 °C for 30 min and 95 °C for 10 min, respectively. The amplification step was performed for 45 cycles at 95 °C for 15 s, 60 °C for 1 min and the melting curve step at 95 °C for 30s, 60 °C for 30s. The Abs Quant/2nd derivative method was used to calculate threshold cycle (Ct) values. The melting temperature (T m ) of the PCR products were analyzed and compared with the product amplified from positive control to exclude the reactions with non-specific amplification. The percent inhibition was calculated relative to the cells treated with 0.5% DMSO.

Cell viability assay
Vero E6 or Calu-3 cells were seeded in 96 well-plates at a density of 2.5 × 10 4 or 2.0 × 10 4 cells per well, respectively. Then niclosamide, ivermectin, and chloroquine at various concentrations in 2%FBS-MEM or 2%FBS-DMEM/ F12 were added to Vero E6 or Calu-3 cells, respectively, for 48 hours. Subsequently, the cell viability was assessed using MTT dyes (Invitrogen) in duplicate (Supplementary file 1). The procedure was described elsewhere [23]. The viability of drug-treated cells was expressed as percent cell viability relative to 0.5% DMSO-treated cells.

Antiviral activity against SARS-CoV-2 in vitro Single drug treatments
Vero E6 or Calu-3 cells were seeded in 96 well-plates at a density of 2.5 × 10 4 or 2.0 × 10 4 cells per well, respectively. The cultured medium was removed, and the cells were incubated with twofold serially diluted drugs in 2%FBSmedia for 1 hour at 37 °C with 5% CO 2 . The medium containing 0.5% DMSO was used as no drug control. After that, the Vero E6 or Calu-3 cells were inoculated with SARS-CoV-2 at multiplicity of infection (MOI) 0.01 or 500 TCID 50 , respectively, for 1 hour. Then the inoculum was discarded, and the cells were further maintained in the media containing drugs at various indicated concentrations or 0.5%DMSO. The virus supernatants were collected at 48 hours after infection. The virus titers were determined using both plaque assay and one-step qRT-PCR. The experiments were repeated at least three times (Supplementary file 1).

Two-drug combinations treatments
The experiments were performed according to the single drug treatment protocol. Except the cells were treated for 1 hour with 16 different pairwise combinations of two drugs. Four concentrations of each single drug were used which are at 2×, 1×, 0.5×, and 0.25× of IC 50 values that were evaluated from the single drug treatments. The experiments were repeated at least three times (Supplementary file 1). The cell viability was also assessed using MTT dyes as mentioned earlier.

The combination synergy analysis
The combination synergy of two-drug combinations was analyzed using SynergyFinderPlus (www. syner gyfin derpl us. org) [24]. Four reference models were used in this study, including the Loewe additivity (Loewe) [25], Zero Independence Potency (ZIP), Highest Single Agent (HSA), and Bliss independence models.

Statistical analysis
The independent experiments were performed in triplicate, and data are shown as mean ± SD. The 50% cytotoxic concentration (CC 50 ) and the half-maximal inhibitory concentration (IC 50 ) were calculated from the dose-response curves of drug treatment by non-linear regression analysis using GraphPad Prism 8 (GraphPad Software, Inc., CA).  Table 1 show the anti-SARS-CoV-2 activities and cytotoxicity of the repurposed drugs in Vero E6 cells. The plaque assay was used to determine the viral production and is expressed as the percent inhibition relative to the viral titer of DMSO-treated cells. The one-step qRT-PCR was used to quantitate the viral RNA in virus supernatants and is also expressed as the percent inhibition relative to the DMSO-treated cells. The IC 50 values calculated from the dose-response determined by plaque assay for niclosamide, ivermectin, and chloroquine were 0.049, 1.23, 0.046 and 0.83 μM, respectively. The IC 50 values calculated from the dose-response determined by one-step qRT-PCR for niclosamide, ivermectin, and chloroquine were 0.043, 1.27, and 0.89 μM, respectively. Both methods used for viral quantification resulted in similar IC 50 values. Thus, the viral RNA quantification by the onestep qRT-PCR accurately determined the infectious virus output in these experiments and could be used for the further two-drug combination experiments for the high throughput screening.

Evaluation of two-drug combination treatments against SARS-CoV-2 in Vero E6 cells
Firstly, the antiviral activities of two-drug combinations were assessed in vitro in Vero E6 cells. The cells were treated with 16 different pairwise combinations of two drugs, including, niclosamide-ivermectin, niclosamidechloroquine and ivermectin-chloroquine.

Niclosamide-ivermectin combination
The presence of ivermectin induced a shift in the doseresponse curve of niclosamide, with approximately 2-fold reduction of niclosamide IC 50 value in the presence of 0.6 and 0.3 μM ivermectin ( Fig. 2A, Table 2). In a similar way, the presence of 0.0225 μM and 0.01125 μM niclosamide resulted in 4.06 and 1.92-fold reduction of ivermectin IC 50 value, respectively (Fig. 2B, Table 2). The dose-response matrix of niclosamide and ivermectin combination showed the obvious increasing inhibitory effects with the maximal inhibitory activity of over 99% at the concentrations of 2-fold of the individual drug IC 50 (Fig. 2C). From Fig. 2D, a synergy score plot shows positive Loewe synergy scores in the combinations with 1.2 μM and 2.4 μM ivermectin. Moreover, the combination of 0.0225 μM niclosamide and 1.2 μM ivermectin shows a peak Loewe synergy of 24.6, 95% confidence intervals (CI) [19.22, 30.9], which indicated a strong synergistic effect. The scores were slightly negative in the other part of the plot with lower ivermectin concentration indicating only additive effect at these lower concentrations. The mean Loewe synergy score is 4.28, which accounted for the additive effects between niclosamide and ivermectin in Vero E6 cells. Similar synergy scores of 3.97 and 4.26 were obtained using ZIP and Bliss independence reference models, respectively. The synergy scores calculated using HSA model was 16.02, which indicated a synergistic effect between niclosamide and ivermectin. No significant cytotoxicity in all 16 pairwise combinations ( Fig. 2A, B).

Niclosamide-chloroquine combination
It was found that the presence of chloroquine induced a shift in niclosamide dose-response curve, with 3.308 and 1.483-fold reduction of niclosamide IC 50 value in the presence of 0.425, and 0.2125 μM chloroquine, respectively (Fig. 3A, Table 2). A similar trend was   3B, Table 2). The dose-response matrix shows increasing inhibitory effect of the combination with higher concentrations of niclosamide and chloroquine (Fig. 3C). From Fig. 3D, most parts of a synergy score plot show negative to low positive synergy scores with a mean Loewe synergy score of 0.68, indicating an additive effect between niclosamide and chloroquine.
Except for the combination of 0.045 μM niclosamide and 0.85 μM chloroquine that shows a peak synergy score of 20.11, 95% CI [12.15, 23.29], indicating a synergistic effect at these concentrations. Additionally, the synergy scores calculated using ZIP and Bliss independence reference models gave the values of − 2.72 and − 2.8, respectively, which similarly indicated the additive effect. The HSA model resulted in the synergy score of 10.23, which accounted for the small level in synergistic effect. No significant cytotoxicity in all 16 pairwise combinations (Fig. 3A, B).

Ivermectin-chloroquine combination
The results showed that the presence of chloroquine induced a shift in ivermectin dose-response curve, with 2.47, 1.55-fold reduction of ivermectin IC 50 value in the presence of 0.425, and 0.2125 μM chloroquine, respectively (Fig. 4A, Table 2). Similarly, the presence of ivermectin also induced a shift in chloroquine dose-response curve, with 2.83 and 1.73-fold reduction of chloroquine IC 50 value in the presence of 0.6 and 0.3 μM ivermectin, respectively (Fig. 4B, Table 2). The dose-response matrix shows increasing inhibitory effect with higher concentrations of ivermectin and chloroquine (Fig. 4C). Most parts of a synergy score plot show negative synergy scores except for a peak positive score of 6.85, 95% CI [− 4.92, 18.37] in the combination of the 0.6 μM ivermectin and 0.85 μM chloroquine (Fig. 4D). The peak negative synergy score is − 7.5. As all of the different combinations had Loewe synergy scores between − 10 and 10 with a mean score of − 3.08, it suggests an additive effect between ivermectin and chloroquine. Moreover, ZIP, Bliss independence and HSA reference models showed the synergy scores of − 7.61, − 7.66 and 6.66, respectively, which indicated the additive effect. No significant cytotoxicity in all 16 pairwise combinations (Fig. 4A, B).

Evaluation of single drug treatment against SARS-CoV-2 in Calu-3 cells
The best antiviral activity and calculated synergy scores demonstrated in the treatment with niclosamide-ivermectin combination in Vero E6 cells. Therefore, this twodrug combination was selected for the further evaluation in the human lung cancer cell line, Calu-3. The antiviral activities of single niclosamide and ivermectin treatments were assessed in Calu-3 cells (Fig. 5). The IC 50

Evaluation of Niclosamide-ivermectin combination treatment against SARS-CoV-2 in Calu-3 cells
The strong shifts were observed in the dose-response curves of niclosamide combined with 0.4 and 0.2 μM ivermectin (Fig. 6A). The presence of 0.1 and 0.05 μM ivermectin also induced a shift in the dose-response curve of niclosamide, with 2.38 and 2.33-fold reduction of niclosamide IC 50 values, respectively (Fig. 6A, Table 3). In a  Table 3).
The dose-response matrix shows the increasing antiviral activity compared to the single drug treatments (Fig. 6C).
The combination synergy analysis showed the mean Loewe synergy score of 3.82, which accounted for the additive effect between niclosamide and ivermectin in Calu-3 cells (Fig. 6D). Additionally, a peak Loewe synergy score of 10.86 showed in the combination of niclosamide and ivermectin at the highest concentrations (0.4 μM for both drugs). The synergy score obtained from ZIP, Bliss independence and HSA reference models were − 7.09, − 7.34 and 8.12, respectively, which also accounted for the additive effect between niclosamide and ivermectin. All 16 pairwise combinations showed no significant cytotoxicity (Fig. 6A, B).

Discussion
Our study shows that the repurposed anti-parasitic drugs, niclosamide, ivermectin and chloroquine possess high in vitro activity against SARS-CoV-2 as the IC 50 values are in the low micromolar range. These results of single drugs treatments are in agreement with the previous studies [7,14,18,26]. Niclosamide showed broad-spectrum antiviral activity against a wide range of viruses such as SARS-CoV [19,27,28], MERS-CoV [29], Zika virus [30], hepatitis C virus [31], Ebola virus [32] and Human immunodeficiency virus type 1 (HIV-1) [23]. The evidence found in other viruses suggested the plausible mechanisms of niclosamide in SARS-CoV-2 inhibition by blocking of viral entry via altering endosomal pH and the prevention of autophagy that led to the inhibition of virus replication [29,33,34]. Although niclosamide was originally thought to act on parasitic worms in the gut lumen and is barely absorbed to the blood stream, it was tested for various systemic repurposed treatments, and a maximal plasma concentration ranged from 35.7 to 182 ng ml − 1 (corresponding to 0.11-0.56 μM) was observed in a pharmacokinetic study [35][36][37][38]. This level exceeds the in vitro niclosamide IC 50 against SARS-CoV-2, especially when used in the tested combinations. However, there have been little clinical data on niclosamide in COVID-19 treatment.
Chloroquine inhibits a broad range of viruses by blocking viral entry via inhibition of endosomal acidification [39]. It was recently shown that chloroquine could not inhibit SARS-CoV-2 in human lung cells because of the expression of TMPRSS2 [40]. This may at least partially explain the lack of clinical efficacy of this drug. Despite these in vitro anti-SARS-CoV-2 activities, clinical application of this drug to COVID-19 treatment has not yet been successful [8].
Previous in vitro studies suggested that ivermectin inhibits host importin alpha/beta-1 nuclear transport proteins, thus preventing the viruses from suppressing the host antiviral response [41]. Several studies reported antiviral activity of ivermectin on other viruses such as Zika virus [42], dengue virus [43][44][45], HIV-1 [46] and influenza A viruses [47]. Recently, it was found that ivermectin may interfere with the attachment of SARS-CoV-2 spike protein to the ACE2 receptor on human cell membrane [48]. Various possible mechanisms of action of ivermectin against SARS-CoV-2 had been proposed in both direct action on SARS-CoV-2 and host cellular targets [49]. However, uncertain clinical trial results varying from effective to no significant benefits were found in COVID-19 treatment using ivermectin [9][10][11][12][13][50][51][52][53]. Oral administration of ivermectin (200 μg/kg) in humans resulted in a maximum plasma concentration at 0.049 ± 0.02 μM (mean ± SD) [54], which was lower than the in vitro IC 50 values for anti-SARS-CoV-2 activity (IC 50 Vero E6 = 1.23 μM, Calu-3 = 0.2 μM). However, higher levels of ivermectin were found in other tissues including fat, skin, and nodular tissues [55]. Different lung tissue concentrations of ivermectin were reported. The predicted maximum lung concentration calculated based on lung: plasma ratio in cattle was around twofold of the maximum plasma concentration (0.0873 μM) [56,57]. However, a previous study in animals reported that the concentration of ivermectin in lung tissue may be 20 times higher than the plasma concentration [58], which  50 . Then the cells were maintained in the medium containing drugs or 0.5%DMSO for two days. Virus titers were determined using the plaque assay. The dose-response curves were expressed as the percent inhibition in relative to DMSO-treated cell control. The cell viability was determined using MTT assay and is expressed in relative to the DMSO-treated cell control. The experiments were repeated at least three times, and data are shown as mean ± SD  [59,60]. This suggested that ivermectin tissue levels may be much higher than in plasma and may reach therapeutic antiviral level. The lack of obvious clinical efficacy suggests that either these in vitro activities could not take effect in vivo, or the activities may not be sufficiently potent. An obvious strategy to enhance the potency is drug combination. While combining direct acting antivirals with different targets almost always results in additive or synergistic effect, combining drugs that act on host machineries does not always cause a synergistic effect and can even result in an antagonistic effect [26,61]. Selecting proper drug combinations with synergistic effect is therefore crucial for development of efficacious regimens. In this study, two-drug combinations improve anti-SARS-CoV-2 activity as the greater viral inhibitory effects were observed with lower IC 50 values compared to individual drugs. Although all combinations resulted in comparable levels of IC 50 reduction indicating additive/synergy effect at levels lower than IC 50 of individual drugs, only niclosamide-ivermectin resulted in enhanced activity at most of higher concentrations resulting in almost complete After that, the cells were infected with SARS-CoV-2 at 500 TCID 50 for 1 hour. The virus inoculum was discarded, and the cells were further maintained in the medium containing drugs for 2 days. The viral titers were determined using the plaque assay. The dose-response curves of two-drug combination treatments against SARS-CoV-2 are shown; (A) serial dilutions of niclosamide in the presence of different fixed concentrations of ivermectin, (B) serial dilutions of ivermectin in the presence of different fixed concentrations of niclosamide. The synergy scores of two-drug combinations were calculated using SynergyFinderPlus. The dose-response matrix (C) and the Loewe synergy score map of two-drug combination treatment (D) are shown. The synergy scores less than − 10 accounted for the antagonistic effect; from − 10 to 10 accounted for the additive effect; and larger than 10 accounted for the synergistic effect between two drugs. The experiments were repeated at least three times, and data are shown as mean ± SD in A, B and C or mean [95% confidence intervals (CI)] in D inhibition (> 99%) at the concentrations of about 2 times of IC 50 of single drugs. At these concentrations, the single drugs could achieve only 70-80% inhibition. The enhanced combine-activity at both lower and higher concentrations resulted in the higher mean synergy score as compared to other combinations. Our data may be useful in guiding the design of clinical trials that may generate a badly needed efficacious regimen for COVID-19 treatment and prevention. Although only the original strain was tested in this study, we do not expect the inhibitory effect to be much different among strains as the drugs target either the more conserved part of the virus or host machineries. Nevertheless, before attempting to use these drugs in clinical trials, the sensitivity of circulating viral strains should be confirmed.

Conclusions
In conclusion, our study demonstrated the benefit of combining ivermectin, niclosamide and chloroquine on their anti-SAR-CoV-2 activities. Among the combinations, ivermectin and niclosamide showed the best synergistic profile. This combination should be further tested in clinical trials.