Inhibition of in vitro Ebola infection by anti-parasitic quinoline derivatives

Introduction

Since its discovery in 1976, Ebola virus has been responsible for numerous outbreaks, with case fatalities varying from 20% to 90%^1,2. From 2014 to 2016, West Africa was devastated with the largest Ebola outbreak in recorded history with six countries affected: Nigeria, Senegal, Guinea, Liberia, Mali and Sierra Leone. With no effective therapies or licensed vaccines available, little could be done to treat patients and prevent the spread of the virus³. Over a period of two years, Ebola virus was responsible for 28,646 reported infections, with 11,323 deaths^1,3.

Since the outbreak, a vaccine for Ebola virus (rVSV-EBOV) has been developed by the National Microbiological Laboratory in Winnipeg, Manitoba, Canada, and has been shown to be highly protective⁴. Undoubtedly, the rVSV-EBOV vaccine is to be a powerful defense against future Ebola outbreaks. However, it is unlikely to be sufficient to stop future outbreaks. As the natural host of Ebola virus remains unknown, eradication of Ebola reservoir populations is impossible⁵. Because of this, future spillover events are an inevitability⁶. Additionally, the unstable political environment of many affected African countries further complicates the effective administration of vaccines; as seen in the Democratic Republic of Congo during the current Ebola outbreak^2,7. These factors emphasize the need for acute infection therapies.

Previous publications have focused on potential drugs predicted to have efficacy against Ebola virus disease (EVD) and included drugs already approved and widely used that could perhaps be repurposed as treatments for EVD^8,9. Through a virtual screening process, Veljkovic et al.⁸ identified 267 US Food and Drug Administration (FDA)-approved drugs with possible inhibitory effects against Ebola virus. Among these were the antimalarial quinoline drugs amodiaquine and chloroquine. Chloroquine has been shown to be protective against Ebola virus disease in an in vivo mouse model⁹. Additionally, amodiaquine has been identified as having inhibitory effects in a pseudo-type entry assay of Ebola virus⁹. Furthermore, retrospective analysis of relative risks for patients from the 2014 to 2016 West Africa outbreak identified amodiaquine in combination with artesunate as having therapeutic effects. Data have suggested that patients infected with Ebola virus that were prescribed artesunate-amodiaquine had a significantly reduced risk of death compared to those that were prescribed artemether-lumefantrine, the latter treatment not being statistically significantly different from no treatment¹⁰. This suggests that the quinoline derivative drug class may have therapeutic use in the treatment of EVD.

In this article, the efficacy of 36 novel quinolines (Table 1) derivatives and previously approved quinoline compounds (amodiaquine and chloroquine) were examined for their ability to inhibit Ebola virus replication. Our results support previous reports that suggested that amodiaquine and chloroquine could be potential treatments for EVD. In addition, we identified additional, novel, quinolines that could be candidates for further study of their potential for inhibition of Ebola virus infection.

Table 1. Antiparasitic quinolines examined, chemical structures, inhibitory activity and toxicity testing.

Drug	Structure	Inhibition at 10µM	Viability at 10µM
Amodiaquine		97.67± 10.62%	87.71± 1.00%
Chloroquine		92.16± 6.23%	106.26± 1.24%
2NH2Q		89.74± 6.53%	81.10± 2.09%
2PentQ		-6.39± 9.15%	103.48± 0.99%
2OHQ		89.69± 4.09%	31.15± 1.27%
2COOHQ		30.58± 6.15%	88.54± 1.54%
2CNQ		3.07± 8.34%	87.95± 3.63
2PQSel		10.85± 32.36%	97.61± 2.06%
2QPOH		-113.56± 92.5%	88.33± 3.26%
2Q16OH		-123.88± 30.63%	74.05± 0.75%
2Qi15		93.27± 4.94%	109.52± 0.25%
2QQ		74.33± 13.45%	57.31± 0.29%
XF906		0.21± 20.27%	104.49± 1.65%
BS460		4.04± 44.15%	107.53± 0.59%
DD1		90.49± 8.32%	59.69± 0.03%
DD2		80.77± 1.38%	77.05± 4.78%
DD3		-156.04± 97.28%	105.49± 2.54%
DD4		49.7± 1.17%	90.09± 1.64%
DD5		-66.45± 17.78%	51.44± 1.31%
MBN111		-48.86± 87.76%	42.50± 4.97%
MBN132		-10.40± 57.69%	104.57± 1.39%
MBN91		81.81± 8.80%	43.52± 0.73%
KHD291		-68.13± 58.12%	71.17± 0.12%
KHD288		-8.41± 60.44%	62.68± 0.72%
MBN115		-400.96± 223.28%	77.08± 0.93
MD823		76.25± 3.16%?	58.86± 1.59%
FZ49		21.83± 28.39%	72.72± 1.22%
DD6		-8.49± 18.58%	96.73± 2.01%
DD7		-31.56± 80.91%	90.32± 1.64%
DD8		-74.03± 100.51%	91.54± 0.41%
FZ142		63.08± 2.64%	101.60± 1.39%
TOF411		71.04± 0.67%	104.87± 0.60%
TOF401		-44.62± 11.87%	106.21± 1.08%
MD20		-124.41± 93.37%	108.35± 0.56%
MBN87		-17.91± 51.80%	70.50± 0.78%
MBN140		-68.13± 58.12%	67.58± 0.73%
FS48		-8.41± 60.44%	81.47± 0.97%
AS65		-10.17± 65.40%	77.51± 0.40%

The column on the left indicates the percent inhibition of trVLP replication using 10 µM, and the column on the right indicates the drug toxicity as assessed using MTT viability.

Inhibition: HEK 293T cells were either transfected with replication machinery plasmids VP30, VP35, NP and Tim-1 (-L) or transfected with all replication machinery plasmids and Tim-1, allowing trVLP entry and replication (+L). Cells were treated with a 10µM final dose of each quinoline derivative, two hours pre-infection. Cells were infected with 100µL of viral stock diluted in 200µL of DMEM with 5% FBS. Cells were lysed, and luciferase activity was assessed 72 hours post-infection. Data represents three biological replicates. Data background corrected and displayed as a percentage of the positive control. Error shown is standard error of the mean. Viability: Drugs were added at a final dose of 10µM, 0.1% DMS. 26 hours later, media was replaced with 100µL of 0.5 mg/mL MTT in media without phenol red. After incubating for two hours at 37°C, 100 μL 10% SDS was added. MTT was fully dissolved at 37°C, then absorbance was read at 570nm. Wells containing no cells served as blanks. Cell viability was determined as percent absorbance of treatment to 0.1% DMSO control. Data shown represents four biological replicates. Data background corrected and displayed as a percentage of the no treatment control. Error shown is the standard error of the mean.

Methods

Results

To assess the ability of quinolines to inhibit Ebola virus infection, we employed an established mini-genome model of Ebola replication, trVLP, allowing us to work under biosafety containment level 2 (CL 2) conditions^27–31.

In our first series of tests, the ability of amodiaquine and chloroquine to inhibit in vitro infection of Ebola virus replication was tested (Figure 1). Following transfection with the required replication machinery and attachment receptor expression plasmids, HEK-293T cells were treated with the drugs at a concentration of 10 µM two hours pre-infection. Luciferase activity was measured 72 hours post-infection²². Both amodiaquine and chloroquine demonstrated significant reductions in luciferase activity, indicating inhibited viral transcription and replication³². Following these findings, dose-response experiments for amodiaquine and chloroquine were conducted. Inhibition activity was assessed at concentrations ranging from 10 µM to 0.31 µM (Figure 2A and 2B). Amodiaquine exhibited a half maximal inhibitory concentration (IC₅₀) of 1.45 µM, while chloroquine exhibited an IC₅₀ of 3.95 µM. Cellular toxicity for both amodiaquine and chloroquine was measured using an MTT viability assay (Figure 2A and 2B) to quantify drug toxicity along the concentration response²².

Figure 1. Amodiaquine and chloroquine inhibit trVLP replication.

293T cells were either transfected with replication machinery plasmids VP30, VP35, NP and Tim-1 (-L) or transfected with all replication machinery plasmids and Tim-1, allowing trVLP entry and replication (+L). Cells were treated with either 10µM amodiaquine or chloroquine two hours pre-infection or received no treatment. Cells were infected with 100µL of viral stock diluted in 200µL of DMEM with 5% FBS. Cells were lysed, and luciferase activity assessed 72 hours post-infection. Data shown represents four biological replicates. Data background corrected and displayed as a percentage of the positive control. Error bars shown are the standard error of the mean. Results from drug treatments were statistically compared with the (+) control group. * Denotes a p-value <0.05.

Figure 2. Chloroquine, amodiaquine and 2NH2Q inhibit trVLP in a dose dependent manor.

Circular markers, solid line: 293T cells were either transfected with replication machinery plasmids VP30, VP35, NP and Tim-1 (-L) or transfected with all replication machinery plasmids and Tim-1, allowing trVLP entry and replication (+L). Cells were treated with either amodiaquine, chloroquine or 2NH2Q at concentrations of 10, 5, 2.5, 1.25, 0.625, or 0.31µM, two hours pre-infection. Cells were infected with 100µL of viral stock diluted in 200µL of DMEM with 5% FBS. Cells were lysed, and luciferase activity assessed 72 hours post-infection. Data shown represents four biological replicates. Data background corrected and displayed as a percentage of the positive control. Error bars shown are the standard error of the mean. Square markers, dashed line: 293T cells were seeded to mimic day two (post-transfection) stage (30% confluency). Drugs were added at a final dose of 0.47, 0.94, 1.88, 3.75, 7.5, 15, or 30 µM, 0.1% DMS. 26 hours later, media was replaced with 100µL of 0.5 mg/mL MTT in media without phenol red. After incubating for two hours at 37°C, 100 μL 10% SDS was added. MTT was fully dissolved at 37°C then absorbance was read at 570nm. Wells containing no cells served as blanks. Cell viability was determined as a percentage of the absorbance of treatment to 0.1% DMSO control. Data shown represents three biological replicates. Data background corrected and displayed as a percentage of the no treatment control. Error bars shown are the standard error of the mean.

To rapidly assess the inhibitory ability of the 36 novel quinoline derivatives (Table 1), drugs were screened at a concentration of 10 µM as described above, with drug treatments occurring two hours pre-infection. From these tests, 2OHQ¹², 2NH2Q¹³, 2QQ¹⁷, 2Qi15¹⁸, MD823, TOF411, DD2, DD1, and MBN91 were identified as having inhibitory activity in our in vitro assay (Table 1). However, due to drug toxicity or availability of synthesized drug, only 2NH2Q was pursued for further study. A dose-response for 2NH2Q was conducted, with inhibition being assessed between 10 µM and 0.31 µM (Figure 2C). Results indicated the IC₅₀ of 2NH2Q to be 4.66 µM, slightly higher than either amodiaquine or chloroquine.

Discussion

As outbreaks of EVD continue to occur, treatment and preventative strategies are urgently needed.In August 2018, the Ministry of Health of the Democratic Republic of the Congo declared a new outbreak of Ebola virus disease in North Kivu Province. As of January 19, 2019, there have been 636 confirmed cases of EVD and 370 confirmed deaths^2,7. Previously, it was suggested that anti-malarial drugs may be effective for the treatment of EVD^8–10. In the report, we have provided supporting evidence that antiparasitic drugs can inhibit Ebola virus replication in vitro, which could be a breakthrough in fighting EVD if proven by additional studies or clinical trials.

Quinoline is a heterocyclic aromatic organic compound and is the base structure for many small molecule drugs. Quinoline-based drugs are widely used for the treatment of malarial infections. As their production is cheap, they are easy to synthesize with high yields, and are chemically stable while being readily accessible, quinolines offer many advantages for the treatment of infectious disease in tropical climates, such as the climates found in Africa^8,27.

Our study aimed to test anti-malarial compounds against an Ebola replication model using the previously published trVLP in vitro assay, which has been shown to accurately reflect CL4 studies using intact infectious Ebola virus²⁹. We find amodiaquine to be effective at inhibiting Ebola viral replication. Amodiaquine is a useful anti-malarial agent in regions where Ebola virus outbreaks are possible, and this drug has been suggested to help with EVD outbreaks¹⁰. We also show using our assay that other antiparasitic malarial drugs can inhibit Ebola virus replication with low cellular toxicity. These include chloroquine, 3/35 novel synthetic quinolines, one termed 2NH2Q, with high efficacy in inhibiting Ebola virus replication and low toxicity, and 4/35 showing moderate efficacy in inhibiting Ebola virus replication while being relatively non-toxic (Table 1).

Quinolines, such as amodiaquine or chloroquine, have been used in regions at risk for EVD to prevent malaria and some have stated that use of these compounds is correlated with reduced death from EVD, suggesting that quinolines could be useful as both an anti-malarial and EVD treatment/prevention drug.

In summary, 38 compounds selected for their antiparasitic activity were evaluated against Ebola virus and their cytotoxicity was also determined. Some of them had antileishmanial activity, such as 2PQsel¹³ and others exhibited antiplasmodial activity, such as chloroquine, amodiaquine and BS460³³. The most interesting compounds emerging from viral inhibition and cytotoxicity assays were chloroquine and amodiaquine, with the advantage of chloroquine having no toxicity at 10 µM. From the quinoline series, the most promising compounds were 2Qi15 and 2NH2Q, both of them being 2-substituted quinolines. When studying structure-activity relationships, the presence of an amino group at the extremity of the propylene chain (compound 2NH2Q) allowed a better cell viability (81%) than a hydroxyl group (compound 2OHQ, 31%), despite their similar capacity for viral inhibition (89%). Replacement by a carboxylic group (compound 2COOHQ) strongly reduced the antiviral inhibition but maintained good cell viability. The other substitutions performed on the quinoline scaffold were not really satisfactory in regard to their biological activity.

Our results support the hypothesis that antiparasitic quinolines active against malaria and leishmaniasis could be efficacious as a treatment for EVD. Further studies are necessary to confirm the utility of quinolines as anti- Ebola virus therapeutics, including consideration of these drugs when designing future Ebola drug trials in Africa.

Data availability