Bisacodyl Limits Chikungunya Virus Replication In Vitro and Is Broadly Antiviral

ABSTRACT

Identifying novel antivirals requires significant time and resource investment, and the continuous threat of viruses to human health necessitates commitment to antiviral identification and development. Developing antivirals requires years of research and validation, and recent outbreaks have highlighted the need for preparedness in counteracting pandemics. One way to facilitate development is to repurpose molecules already used clinically. By screening such compounds, we can accelerate antiviral development. Here, we screened compounds from the National Institutes of Health’s Developmental Therapeutic Program for activity against chikungunya virus, an alphavirus that is responsible for a significant outbreak in the Americas in 2013. Using this library, we identified several compounds with known antiviral activity, as well as several novel antivirals. Given its favorable in vitro activity and well-described in vivo activity, as well as its broad availability, we focused on bisacodyl, a laxative used for the treatment of constipation, for follow-up studies. We find that bisacodyl inhibits chikungunya virus infection in a variety of cell types, over a range of concentrations, and over several rounds of replication. We find that bisacodyl does not disrupt chikungunya virus particles or interfere with their ability to attach to cells, but, instead, bisacodyl inhibits virus replication. Finally, we find that bisacodyl is broadly antiviral against a variety of RNA viruses, including enteroviruses, flaviviruses, bunyaviruses, and alphaviruses; however, it exhibited no activity against the DNA virus vaccinia virus. Together, these data highlight the power of compound screening to identify novel antivirals and suggest that bisacodyl may hold promise as a broad-spectrum antiviral.

INTRODUCTION

Developing antivirals requires significant research effort to identify potential targets or to identify lead compounds with promising antiviral activity. Thus, developing antivirals requires years of work prior to consideration of clinical trials and the potential for approval. A strategy to overcome this significant limitation is to rapidly screen compounds that are approved for human use for novel antiviral activity (1–4). Such screens have shown promise for diverse viruses, including Ebola virus (5), Zika virus (6), Japanese encephalitis virus (7), and coronaviruses (8), including SARS-CoV-2 (9, 10). Outbreaks of viruses lead to a surge in research activity on these pathogens and have enhanced the development of specific antivirals. Recent progress with SARS-CoV-2 antivirals demonstrated the importance of antiviral development, as well as the limitations of drug screens (11, 12). Importantly, the development of drugs such as remdesivir (13) required years of prior research. Thus, antiviral development requires sustained research input to not only address current infections but also has the potential to combat future outbreaks.

Alphaviruses are frequent human pathogens and have caused significant outbreaks. Alphaviruses make up the only genus of the Togaviridae family, comprised of small enveloped viruses containing single-stranded positive-sense RNA genomes. Alphavirus infections are spread through the aid of arthropod vectors such as Aedes aegypti and Aedes albopictus, and through this arthropod vector, alphaviruses rapidly spread. This was best demonstrated by the rapid emergence of chikungunya virus (CHIKV) in the Americas in 2013 (14). Upon introduction and spread of CHIKV, the pathogen established itself, and its geographical reach is unlikely to subside. CHIKV transmission remains prominent in warm and temperate regions such as northern Africa, southeastern Asia, and South America (15). The spread of the mosquito vector compounds virus control. CHIKV infection can be characterized as a musculoskeletal disease (16). Initially, CHIKV symptoms are acute and include fever, headache, and joint pain. Postacute infection can develop into chronic disease involving persistent muscle pain and joint swelling, giving rise to persistent arthralgia (17).

There is no vaccination or drug therapy available for CHIKV infection, and thus, antiviral development is necessary to target CHIKV and quell future outbreaks. Rapid drug screens provide an efficient way to identify unknown compounds with potential antiviral activity and have been used successfully in the identification of several lead compounds, including for the alphavirus Venezuelan equine encephalitis virus (18). The development of direct-acting antivirals (DAA), or molecules specifically targeting the virus, has shown promise in anti-hepatitis C virus (HCV) and anti-HIV drug development but is complicated by the rapid emergence of resistant viruses. In contrast, targeting the host cell to inhibit cellular pathways essential to virus infection shows significant promise (19, 20) in terms of preventing antiviral resistance and has the potential for broad-spectrum antiviral activity. However, few such compounds have been identified and successfully implemented as an antiviral therapy. To facilitate drug development (antivirals and otherwise), NIH has established a repository of chemical compounds, which we and others have used to screen for antiviral activity.

We screened compounds of known structure for antiviral activity against CHIKV. Bisacodyl, a well-studied diphenylmethane derivative, was identified as one of the top hits. Bisacodyl is a stimulant and oral laxative used to treat patients with chronic constipation (21). The molecule targets the colon, and it hydrolyzes to produce bis-(p-hydroxyphenyl)-pyridyl-2-methane (BHPM) (22), which we find also has antiviral activity. Overall, we found that bisacodyl and BHPM exhibit significant antiviral activity in vitro in several cell types. These molecules inhibit viral genome replication to reduce viral titers.

RESULTS

Rapid screening of molecules for activity against CHIKV infection.

We initiated a drug screen using a selection of compounds available from the NIH Developmental Therapeutics Program (DTP). Previously, we screened compounds for activity against La Crosse virus (LACV) using a similar strategy to identify novel antivirals (23). In these assays (summarized in Fig. 1A), we plated human Huh7 hepatocytes to confluence in 96-well plates. Four hours prior to infection, we treated with a 1:100 dilution of the compounds. We subsequently infected the cells at a multiplicity of infection (MOI) of 5 PFU per cell with CHIKV to ensure infection of all cells and allowed infection to progress for 48 h, at which point we stained the cells with crystal violet. As controls, we left wells untreated (CHIKV only) or untreated and uninfected. Additionally, we treated cells with ribavirin, a known antiviral with activity against CHIKV (24–26) as a control. In these assays, any compounds with cytotoxicity or compounds that are ineffective against CHIKV both generate a readout of a cleared well and reduced absorbance. However, potential antivirals that were not cytotoxic would stain with the crystal violet and yield a higher absorbance. Thus, our screen is specific for compounds that are both nontoxic and exhibit anti-CHIKV activity.

FIG 1.

Antiviral screen design and identification of primary screen hits. (A) Schematic of drug screen. Huh7 cells were treated with antivirals 4 h prior to chikungunya virus (CHIKV) infection at an MOI of 5. At 48 hpi, cells were fixed with formalin, stained with crystal violet, and quantified by plate reader. (B) Raw crystal violet stain quantification. Each compound is represented by a single dot, and compounds of the highest interest are labeled, along with our control antiviral, ribavirin. (C) Analyzed antiviral activity results compared to nontreated (NT) controls to analyze the relative antiviral activity. A value of 1 represents a treatment that resulted in crystal violet staining equivalent to untreated cells. Top antiviral hits against CHIKV were used to treat Huh7 cells with increasing doses of strychnine N-oxide (D), penzipiperylon (E), naloxone (F), danazol (G), and bisacodyl (H) and infected with CHIKV at an MOI of 0.1. Viral titers were determined at 24 hpi by plaque assay. Error bars represent one standard error of the mean (n ≥ 3 for panels D to H). *, P < 0.05; **, P < 0.01; ***, P < 0.001 by two-tailed Student’s t test.

We quantified this signal by resuspending stained cells in acetic acid and reading absorbance (Fig. 1B). Readings were pooled for each compound and normalized to control values obtained from uninfected, untreated cells on each plate (Fig. 1C). Values equivalent to uninfected, untreated cells were considered to have antiviral activity and were selected for additional screening (Table 1). Satisfyingly, ribavirin exhibited significant antiviral activity and was the third-highest hit in our screen.

TABLE 1.

Top hits in initial antiviral screen

DTP NSC no.a	Primary screen value (absorbance at 590 nm)	Antiviral name	Description	Prior antiviral activity against:
24951	0.590	Strychnine N-oxide	Strychnine-derived Strychnos nux-vomica metabolite	SARS-CoV-2 (31)
73254	0.403	Penzipiperylon	Anti-inflammatory drug	None known
70413	0.296	Naloxone	Opioid receptor antagonist; reversal of opioid overdose and respiratory depression	None known
270916	0.280	Danazol	Attenuated androgen; treatment of gynecologic and hematologic disorders	HTLV-1,b HCV, HIV (30, 45)
755914	0.265	Bisacodyl	Diphenylmethane derivative; laxative	EBOV (5)

^aNSC no., Cancer Chemotherapy National Service Center number.

^bHTLV-1, human T-cell leukemia virus type 1 (HTLV-1).

With the top five hits in our screen, we performed secondary screening. In these assays, we treated cells with escalating doses of the compounds 4 h prior to infection with CHIKV at an MOI of 0.1 to allow for multiple rounds of viral replication. We measured viral titers 24 h later via plaque assay. We observed that each of the compounds exhibited antiviral activity to various degrees. Treatment with strychnine N-oxide did not result in reduced viral titers (Fig. 1D); however, treatment with penzipiperylon, naloxone, danazol, and bisacodyl produced a significant decrease in viral titer (Fig. 1E to H). Our fifth-highest hit in our screen, bisacodyl was of particular interest because of its strong antiviral activity as well as its physiological properties, including its characterization as a constipation treatment and general tolerability to human consumption (22). We also observed dose-dependent reductions in CHIKV titers with bisacodyl treatment, providing for a wide range of concentrations with antiviral activity. For these reasons, we chose to focus on bisacodyl for further study.

Bisacodyl exhibits antiviral activity with limited cytotoxicity.

To begin, we measured viral titers with a range of bisacodyl doses, up to 175 μM (depicted in Fig. 2A). We seeded Huh7 cells and treated them with bisacodyl 24 h prior to infection. Cells were then infected at an MOI of 0.1 to allow for multiple rounds of replication. At 48 h postinfection (hpi), virus was collected and quantified via plaque assay (Fig. 2B). We found that titers were significantly decreased at concentrations above 50 μM compared to untreated control. From these data, we calculated the 50% inhibitory concentration (IC₅₀) value to be 39.1 μM (Table 2). We noted no obvious effect on cell viability in tissue culture with bisacodyl treatment by cellular morphology under the microscope, so to confirm that our treatments were not negatively impacting cell health, we performed a viability assay (Fig. 2C). We seeded a confluent 96-well plate and treated with increasing doses of bisacodyl, including doses that well surpass its significant antiviral-producing phenotype. We observed cells lived beyond the IC₅₀, as shown by our viability assay readout. Toxicity was observed above doses of 400 μM bisacodyl, and the 50% cytotoxic concentration (CC₅₀) value was calculated to be 475.1 μM for a selectivity index (CC₅₀/IC₅₀) of 12.2, which gives us a window of opportunity to treat cells with bisacodyl without impacting viability.

FIG 2.

Bisacodyl exhibits antiviral activity with limited cytotoxicity. (A) Schematic of bisacodyl treatment experiments. (B) Huh7 cells were treated with increasing doses of bisacodyl 24 h prior to CHIKV infection at an MOI of 0.1. CHIKV viral titers were quantified by plaque assay at 48 hpi. (C) Huh7 cells were treated with increasing doses of bisacodyl 24 h prior to analysis for viability. (D) Vero; (E) HeLa; (F) BHK-21; (G) 293T; (H) HFF; and (J) MEF cells were treated with increasing doses of bisacodyl as in panel A and subsequently infected. CHIKV titers were determined 48 hpi by plaque assay. We treated HFF (I) and MEF (K) cells with increasing doses of bisacodyl, and their viability was measured. Error bars represent one standard error of the mean (n ≥ 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001 by two-tailed Student’s t test.

TABLE 2.

IC₅₀ values derived from bisacodyl treatment of diverse viruses

Virus	Cell type	IC50 value (μM)
CHIKV	Huh7	39.1
	293T	68.4
	Vero	89.2
	HeLa	53.7
	BHK-21	69.9
	MEF	1.2
	HFF	4.2
HRV-2	Huh7	13.2
CVB3	Huh7	68.3
MAYV	Huh7	79.2
SINV	Huh7	119.2
KEYV	Huh7	36.9
LACV	Huh7	293.9
ZIKV	Huh7	23.6
VACV	Huh7	519.2

To confirm this antiviral phenotype of bisacodyl using additional in vitro model systems, we measured viral titers after bisacodyl treatment in several cell types, including Vero, HeLa, BHK-21, 293T, HFF, and MEF (Fig. 2D to H and J). Cells were treated with increasing amounts of bisacodyl 24 h prior to infection at an MOI of 0.1. At 48 hpi, virus was collected and quantified via plaque assay. Similar to our experiments with Huh7 cells, all cell types showed a similar decrease in titer with bisacodyl treatment (see IC₅₀ values in Table 2). As with Huh7 cells, we observed minimal effects on cellular viability in HFF and MEFs (Fig. 2I and K). Bisacodyl showed the best antiviral activity in Huh7 cells, which we continued using for additional experiments. Overall, these data suggest bisacodyl can act as an antiviral in several cell types with favorable antiviral properties.

Bisacodyl is antiviral over several rounds of replication and limits viral RNA and protein accumulation.

To characterize the effect of bisacodyl over multiple rounds of replication, we measured viral titers over a time course. We seeded Huh7 cells and treated them with 175 μM bisacodyl 24 h prior to infection and subsequently infected them at an MOI of 0.1 to measure viral replication over several rounds of replication. Samples were collected at regular time points throughout infection and quantified by plaque assay (Fig. 3A). We found that bisacodyl reduced titers below untreated samples after 16 hpi and that titers never recovered in bisacodyl-treated cells, suggesting that bisacodyl treatment remained antiviral over several rounds of virus replication.

FIG 3.

Bisacodyl limits CHIKV infectious virus production, RNA accumulation, and protein translation. (A) Huh7 cells were treated with 175 μM bisacodyl 24 h prior to infection at an MOI of 0.1. Supernatant was collected at indicated times, and viral titers were quantified by plaque assay. (B) Total RNA was purified from infected cells treated as in panel A at 24 hpi and quantified by qPCR using CHIKV-specific primers. (C) Released viral RNA from the supernatant of infected cells was quantified via qPCR. (D) Cell lysates were probed for CHIKV E2 protein following bisacodyl treatment (0, 50, 100, and 200 μM). Bands were quantified with ImageJ and normalized to untreated conditions (bottom). (E) Huh7 cells were treated with 175 μM bisacodyl or left untreated and subsequently infected at an MOI of 0.1 for 24 h prior to staining for dsRNA. Error bars represent one standard error of the mean (n ≥ 3; Western blotting is representative of two independent experiments). *, P < 0.05; **, P < 0.01; ***, P < 0.001 by two-tailed Student’s t test.

We further analyzed cells for viral genomes to determine if bisacodyl was affecting viral RNA accumulation. We sampled bisacodyl-treated and CHIKV-infected cells and collected cells for RNA purification at 24 hpi. RNA was purified, reverse transcribed, and analyzed by qPCR using primers specific to CHIKV and normalized to cellular actin using the threshold cycle (ΔΔC_T) method (Fig. 3B). These values were then normalized to untreated control conditions, set to 1, to compare untreated to treated conditions. We observed that bisacodyl treatment up to 200 μM significantly reduced intracellular CHIKV genome accumulation in a dose-dependent manner. To complement these data, we also measured viral genomes in cellular supernatant, representing released viral genomes. We collected infected cell supernatant, purified and reverse transcribed RNA, and performed quantitative PCR (qPCR) with CHIKV-specific primers (Fig. 3C). Released viral genome quantity was similarly reduced, fitting with diminished titers with bisacodyl treatment, up to 200 μM. To confirm that this reduction in viral genome accumulation similarly reduced viral protein accumulation, we measured viral E2 protein by Western blotting after treatment with escalating doses of bisacodyl (Fig. 3D). We observed that bisacodyl significantly limited E2 accumulation in infected cells in a dose-dependent manner, similar to reductions in viral titers and genome accumulation. Finally, we imaged infected cells via immunofluorescence, staining for double-stranded RNA (dsRNA) (Fig. 3E). We found that bisacodyl treatment did not change the gross morphology of these replication complexes within infected cells under the conditions tested. Thus, overall bisacodyl appears to disrupt viral RNA and protein accumulation, but not the formation of replication complexes, suggesting that while these replication complexes can form, they are not producing viral genomes or the viral E2 structural protein.

Bisacodyl inhibits CHIKV replication.

In all previous assays, we added bisacodyl to cells prior to infection, and the compound was present throughout infection (see schematic in Fig. 4A). To identify the stage of virus replication sensitive to bisacodyl, we again treated Huh7 cells with 175 μM bisacodyl at distinct times before and after infection at MOI of 5 to infect all cells. Titers were measured by plaque assay at 24 hpi. As expected, treatment of cells prior to infection led to a significant decrease in viral titers; however, viral titers were also reduced when bisacodyl was added up to 8 hpi. However, treatment after 8 hpi resulted in no significant change in viral titers (Fig. 4B), suggesting that early events up to and including viral genome replication were affected by bisacodyl treatment. To determine if this antiviral effect of bisacodyl could be reversed, we treated cells with bisacodyl and, prior to infection, washed cells with phosphate-buffered saline (PBS) before replenishing the cells with complete media lacking bisacodyl. When we replaced drug-free media, we observed a nearly full rescue of viral titers (Fig. 4C), suggesting that bisacodyl’s antiviral activity requires consistent treatment throughout infection.

FIG 4.

Bisacodyl inhibits CHIKV replication. (A) Schematic of treatment of cells with bisacodyl relative to infection. (B) Huh7 cells were treated with 175 μM bisacodyl at distinct times before and after infection with CHIKV at an MOI of 5. Viral titers were determined at 24 hpi by plaque assay. (C) Huh7 cells were treated with 175 μM bisacodyl 24 h before being washed with PBS to remove bisacodyl and subsequently infected with CHIKV at an MOI of 0.1. Viral titers were determined at 24 hpi by plaque assay. (D) Stock CHIKV was treated with bisacodyl and incubated at 37°C at the indicated doses. Titers were determined by plaque assay after 4 h of incubation. (E) Stock CHIKV was incubated with 175 μM bisacodyl at 37°C for the indicated amount of time prior to viral quantification by plaque assay. (F) Scheme of attachment assays. (G, H) CHIKV attachment assay was used to quantify CHIKV attachment with bisacodyl treatment. Cells were treated with increasing doses of bisacodyl 24 h prior to a 5-min time period (G) or 175 μM bisacodyl over a 30-min time period (H) of CHIKV infection on ice. Cells were subsequently washed to remove unbound virus and bisacodyl, and attached virus was allowed to form plaques. Plaques were subsequently quantified 48 h later by crystal violet staining. (I) Huh7 cells harboring a CHIKV replicon encoding Renilla luciferase were treated with increasing doses of bisacodyl for 24 h before analysis for luciferase activity. (J) Replicon-harboring Huh7 cells were treated with 175 μM bisacodyl for the indicated times, and Renilla luciferase activity was measured. Error bars represent one standard error of the mean (n ≥ 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001 by two-tailed Student’s t test.

Given the effect of bisacodyl on virus titers and genome accumulation, as well as its chemical structure, containing multiple phenol rings, we hypothesized that bisacodyl may be directly inactivating virus. To test this, we directly incubated CHIKV stock with increasing doses of bisacodyl for 4 h, at which point we directly determined the titer of the samples (Fig. 4D). We observed no significant difference in viral titers, suggesting that bisacodyl was not directly impacting virus infectivity. To confirm, we incubated CHIKV stocks with bisacodyl over a time course up to 24 h. In untreated samples, we saw a decay in viral titers, but this decay was indistinguishable when incubated with bisacodyl (Fig. 4E), again suggesting that bisacodyl did not directly inactivate viral particles.

Observing no change in CHIKV infectivity when incubated with bisacodyl, we next tested whether bisacodyl could inhibit viral attachment. To this end, we performed a viral attachment assay (27) in which we incubate virus on treated or untreated cells for 5 min prior to washing and removing unbound virus and allowing a plaque to form (see schematic in Fig. 4F). Importantly, in these assays, bisacodyl is present only prior to and during viral attachment before it is washed away to allow for plaque formation. We observed that bisacodyl treatment only modestly reduced viral attachment (Fig. 4G), and the differences were not statistically significant. When this attachment assay was repeated with a 175-μM dose of bisacodyl over a 30-min period, we observed a significant reduction in bound virus (Fig. 4H). These data suggest that bisacodyl inhibits viral binding by approximately 2-fold, which suggests that additional stages in infection may be impacted by bisacodyl.

We next considered whether bisacodyl affected viral replication. Using Huh7 cells harboring a CHIKV replicon, encoding Renilla luciferase as a readout (27, 28), we treated cells with increasing doses of bisacodyl and measured luciferase activity 24 h later. We observed a dose-dependent decrease in luciferase activity (Fig. 4I), suggesting that bisacodyl was inhibiting CHIKV replication. We confirmed this phenotype by treating with bisacodyl over a 24-h period, observing a steady and significant decrease in luciferase activity with increasing treatment time (Fig. 4J). Together, these data suggest that at least one of bisacodyl’s antiviral effects is on CHIKV replication.

The bisacodyl derivative BHPM is antiviral.

Bisacodyl must be initially converted to the metabolite bis-(p-hydroxyphenyl)-pyridyl-2-methane (BHPM; Fig. 5A), the active metabolite, also known as desacetyl bisacodyl, to relieve constipation (29). We aimed to investigate if this compound is also responsible for inhibiting CHIKV replication. We treated cells with increasing doses of BHPM 24 h prior to infection, as with bisacodyl treatment. The cells were infected with CHIKV, and virus was quantified 24 hpi via plaque assay (Fig. 5B). We found a similar response to bisacodyl treatment with BHPM treatment, with significantly reduced viral titers above 50 μM and an IC₅₀ of 38.6 μM. To ensure the compound treatment was responsible for this phenotype, we again, as described earlier, performed a viability assay on Huh7 cells using increasing doses of BHPM (Fig. 5C). We discovered that BHPM, like bisacodyl, is not toxic to cells until far beyond the effective dosage, with a CC₅₀ value of 464.9 μM and a stimulation index (SI) of 12.0. These data suggest that BHPM inhibits CHIKV infection similarly to bisacodyl, and to confirm that BHPM similarly impacted CHIKV replication, we treated Huh7 cells harboring the luciferase-containing CHIKV replicon and measured Renilla luciferase activity 24 h after treatment (Fig. 5D). As with bisacodyl, we observed a significant reduction in CHIKV replication above 50 μM, again suggesting that bisacodyl and its active derivative BHPM inhibit CHIKV replication.

FIG 5.

Bisacodyl’s active metabolite, BHPM, is antiviral. (A) Structures of bisacodyl and desacetyl bisacodyl, or BHPM. (B) Huh7 cells were treated with increasing doses of BHPM 24 h prior to infection with CHIKV at an MOI of 0.1 Viral titers were determined 24 hpi by plaque assay. (C) Huh7 cell viability with BHPM treatment was measured after 24 h treatment. (D) Huh7 cells harboring a CHIKV replicon encoding Renilla luciferase were treated with increasing doses of BHPM. Luciferase activity was subsequently measured 24 h later. Error bars represent one standard error of the mean (n ≥ 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001 by two-tailed Student’s t test.

Bisacodyl exhibits broad-spectrum activity against RNA viruses.

To investigate how other viruses respond to bisacodyl treatment, we treated Huh7 cells with bisacodyl 24 h prior to infection with the enteroviruses human rhinovirus 2 (HRV-2) and coxsackievirus B3 (CVB3; Fig. 6A), the alphaviruses Mayaro virus (MAYV) and Sindbis virus (SINV; Fig. 6B), the bunyaviruses keystone virus (KEYV) and La Crosse virus (LACV; Fig. 6C), the flavivirus Zika virus (ZIKV; Fig. 6D), and the DNA virus vaccinia virus (VACV; Fig. 6E). At 24 hpi, virus was quantified via plaque assay. Identical doses of bisacodyl were used during treatments. We observed significant antiviral activity with bisacodyl treatment for all RNA viruses (see IC₅₀ values in Table 2). KEYV was particularly susceptible, and titers could not be detected above 50 μM. Interestingly, the DNA virus vaccinia virus was not susceptible to bisacodyl treatment. These data suggest that bisacodyl is broadly antiviral, though its activity is limited to a subset of viruses.

FIG 6.

Bisacodyl exhibits broad activity against RNA viruses. Huh7 cells were treated with increasing doses of bisacodyl prior to infection with the enteroviruses human rhinovirus 2 (HRV2) and coxsackievirus B3 (CVB3) (A), alphaviruses Mayaro virus (MAYV) and Sindbis virus (SINV) (B), bunyaviruses keystone virus (KEYV) and La Crosse virus (LACV) (C), flavivirus Zika virus (ZIKV) (D), and the poxvirus vaccinia virus (VACV) (E). Viral titers were quantified by plaque assay 24 hpi in all cases. Error bars represent one standard error of the mean (n ≥ 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001 by two-tailed Student’s t test.

DISCUSSION

Antiviral development requires significant investment in identification of lead compounds, and testing approved molecules can provide a “jumpstart” to antiviral development. While a wide variety of antiviral screens have identified FDA-approved molecules, a limited set of these molecules has been used successfully as antivirals (12). However, testing these approved molecules has several benefits in terms of well-studied pharmacokinetic data in human patients that could be a significant hurdle for antiviral development de novo. While certainly not the only approach for antiviral development, screening approved compounds for alternative use as an antiviral holds significant promise. Here, we identified several interesting compounds (Table 1) with antiviral activity. Many of these molecules had previously shown antiviral activity (30, 31), validating our screening method. Our molecule of choice for follow-up study, bisacodyl, is used primarily as a treatment for constipation and has been used both clinically and over the counter. No prior work has suggested that treatment with bisacodyl reduces virus infection to our knowledge, though it was identified in a screen for antivirals to treat Ebola virus infection (5). Additional hits within our screen included naloxone, commonly used as a treatment for opioid overdose. Again, no studies have shown antiviral activity for this molecule. Danazol is used in the treatment of endometriosis (32), among other indications. These compounds, along with bisacodyl, are orally bioavailable (33), which is also promising for potential antiviral activity. While our screen identified a plethora of compounds with potential antiviral activity, follow-up study will be necessary to confirm these results.

Bisacodyl has several properties that make it a potential antiviral. First, it exhibits broad-spectrum activity in vitro, suggesting that it could be used in the treatment of diverse viruses. Additionally, the drug itself is relatively nontoxic, both in vitro in our studies and in humans when it is taken to treat constipation. Toxic doses can be reached in humans when consistently exceeding 30 mg for several weeks (34), which may limit its potential antiviral activity. The drug is orally bioavailable and can also be taken as a suppository. The oral route and action on the gut have important implications for antiviral development, especially for fecal-oral pathogens, like several enteroviruses (including coxsackievirus B3, studied here). The impact on the intestines and its effect as a laxative may require additional consideration for in vivo treatment, as laxative treatment can have multiple systemic effects (35, 36), and the effect of this amalgamation on virus infection would require significantly more study.

The precise mechanism by which bisacodyl impacts virus infection is not clear, though we believe that the molecule interferes at the stage of viral genome replication, as evidenced by the effect of bisacodyl on CHIKV replicons. However, bisacodyl may impact several viral proteins or their expression or localization. We observe a significant reduction in the accumulation of viral RNA and E2 structural protein, though we observe no obvious change in replication compartment formation. Thus, bisacodyl may specifically impact the function of these replication compartments rather than their formation itself. Further, we observe a 4-fold reduction in released viral genomes despite reductions in titer of greater than 100-fold. Thus, bisacodyl may impact the formation of infectious virions. Importantly, bisacodyl may impact a cellular process that, in turn, impacts virus replication. Bisacodyl’s molecular mechanisms are not fully understood. At an organismal level, bisacodyl induces peristalsis (37) by stimulating enteric neurons. Bisacodyl’s impact on chloride (38) and calcium (39) ions may play a role in this stimulation, though it has not been formally shown. Within the context of our antiviral screens, bisacodyl’s antiviral activity could be a result of perturbation of ion signaling, though the impact of chloride and calcium signaling on a broad array of viruses is incompletely understood. Thus, further pharmacological characterization of bisacodyl’s molecular mechanism will be necessary to fully understand its antiviral activity.

The use of a laxative such as bisacodyl in the treatment of virus infection has several implications, both in the treatment of disease and in the potential molecular mechanisms involved in infection. As mentioned, bisacodyl has several impacts on the body, both at the molecular and organismal levels. Because bisacodyl’s primary effects are seen within the alimentary canal, it may function well in the context of a virus infecting these tissues, such as enteroviruses. However, the systemic effect of bisacodyl may well impact the replication of nonenteric viruses, like CHIKV. Significant additional study, including animal studies, would illuminate the possibility of using bisacodyl to treat CHIKV infection. For instance, it is unclear if oral bisacodyl would reduce viral replication in a systemic or arthritis model of CHIKV infection. Further, treating virus-infected patients with a laxative like bisacodyl may exacerbate their condition. Regardless, understanding the effects of bisacodyl on CHIKV infection, as well as other virus infection, has the potential to highlight additional cellular pathways or physiological processes that are critical to virus infection.

Our data suggest that bisacodyl exhibits broad-spectrum activity against a variety of RNA viruses, from alphaviruses to bunyaviruses to enteroviruses. This could suggest a conserved mechanism of action whereby bisacodyl affects a common target. However, it is also possible that bisacodyl is targeting a host pathway that these RNA viruses rely on. Bisacodyl, as a laxative, has a variety of functions within the cell and the body, including releasing intracellular calcium and chloride ions, manipulating aquaporins (40, 41), and altering cAMP levels within the cell (42), though its molecular target is not fully understood. Thus, our studies are limited by the current understanding of bisacodyl’s mechanism of action within the cell. Interestingly, we find that vaccinia virus (VACV) is insensitive to bisacodyl. VACV, a DNA virus that replicates within the cytoplasm, has several characteristics that distinguish it from the other viruses in our study, most prominently that it is a DNA virus. However, VACV is vastly different in its replication strategies, and untangling the precise mechanism whereby VACV is insensitive to bisacodyl is unclear at this point. However, these data suggest that despite broad antiviral activity, bisacodyl is not a global antiviral.

MATERIALS AND METHODS

Cell culture.

Cells were incubated at 37°C containing 5% CO₂ in Dulbecco’s modified medium (DMEM; Life Technologies) containing bovine serum and penicillin-streptomycin. Vero and HeLa cells (BEI Resources) were supplemented with 10% newborn calf serum (NBCS; Thermo Fisher). Huh7, MEF, HFF, 293T, and BHK-21 cells were supplemented with 10% fetal bovine serum (FBS; Thermo Fisher).

Drug treatment.

Standard treatment experiments were as follows. Huh7 cells were infected at a multiplicity of infection (MOI) of 0.1 PFU/cell, unless indicated differently, with CHIKV, HRV-2, CVB3, MAYV, SINV, KEYV, LACV, ZIKV, and VACV and then were simultaneously treated with bisacodyl (Cayman Chemical), strychnine N-oxide (NIH DTP compound), penzipiperylon (NIH DTP compound), ribavirin (VWR), naloxone (NIH DTP compound), and danazol (NIH DTP compound) dissolved in dimethyl sulfoxide (DMSO) or water. Cells were incubated at 37°C in 5% CO₂ for 24 to 48 h. Chemical structures were recreated using Adobe Illustrator (Adobe).

Rapid screening of antiviral compounds.

Huh7 cells were seeded on 96-well plates and treated with a 1:100 dilution of each compound from the NIH DTP compound plates 4 h prior to infection. Cells were infected at an MOI of 5 PFU/mL and incubated for 48 h at 37°C in 5% CO₂. Medium was removed, cells were fixed with 4% formalin, and living cells were stained using crystal violet solution (Sigma-Aldrich). Excess stain was removed in a mild bleach solution, and the cells were allowed to dry for 24 h. Crystal violet stain was resuspended in 10% acetic acid. The absorbance at 590 nm was detected using a BioTek Synergy H1 plate reader.

Viability assay.

Huh7 cells, unless otherwise noted, were seeded in a 96-well plate. Twenty-four hours later, cells were treated with increasing doses of bisacodyl or BHPM. After 24 h of treatment, 30 μL of CytoTox-Fluor cytotoxicity assay reagent (Promega) was added to each well and incubated in the dark at room temperature (RT) for 45 min. Fluorescence of the plate was taken using a SpectraMax iD3 fluorometer (485 nm excitation/520 nm emission).

Infection and enumeration of viral titers.

LACV (product no. NR-540; BEI Resources) and KEYV (strain B64-5587.05; product no. NR537; BEI Resources) were derived from Huh7 cells. ZIKV (strain MR766), VACV (WR strain; provided by Tom Gallagher), CHIKV (strain 181/25 from BEI Resources), SINV, and MAYV (43) were derived from Vero cells. CVB3 (Nancy strain) and human rhinovirus 2 (HRV2) were derived from HeLa cells. Various drug amounts were maintained throughout the course of infection as noted. For infection unless otherwise noted, virus was diluted in serum-free DMEM. The viral inoculum was overlaid on cells for 30 min, and the cells were washed with PBS before medium replenishment. Supernatant was collected at the times noted. For plaque assays, dilutions of cell supernatant were prepared in serum-free DMEM and used to inoculate a confluent monolayer of Vero cells for 10 min at 37°C. Cells were overlaid with 0.1% agarose in DMEM containing 2% NBCS. CHIKV, MAYV, SINV, CVB3, and HRV2 samples were incubated for 2 days, LACV samples were incubated for 4 days, VACV samples were incubated for 24 h, and KEYV and ZIKV samples were incubated for 5 days at 37°C. Following appropriate incubation, cells were fixed with 4% formalin and stained with crystal violet solution (10% crystal violet; Sigma-Aldrich). Plaques were enumerated and used to back-calculate the number of PFU per milliliter.

Direct incubation.

Stock CHIKV was generated on Vero cells. Stock was combined with 10, 100, 300, and 600 μM bisacodyl and incubated at 37°C for 4 h. Virus was then directly quantified via plaque assay for both untreated and treated samples. In addition, this experiment was repeated with a single drug dose of bisacodyl over the course of 24 h. Virus was quantified from both treated and untreated samples via plaque assay every 8 h.

Wash-away assay.

Huh7 cells were left untreated or treated with bisacodyl overnight. Bisacodyl-treated cells were subsequently washed with PBS and replenished with fresh media not containing drug. The plate was infected with CHIKV at an MOI of 0.1. At 24 hpi, virus was quantified via plaque assay.

Immunofluorescence imaging.

Huh7 cells were seeded on coverslips and treated with 175 μM bisacodyl or untreated 24 h prior to CHIKV infection. Cells were infected with CHIKV at an MOI of 0.1 for 24 h and then fixed with 4% formalin. Cells were blocked using a 0.2% Triton X-100 and 2% BSA in PBS at RT. Cells were washed using PBS and incubated with primary anti-dsRNA monoclonal antibody J2 (1:500; Sigma-Aldrich) for 1 h at RT. Cells were washed with PBS and incubated with secondary goat anti-mouse antibody (1:1,000) for 1 h at RT. Coverslips were mounted onto microscope slides using mounting media containing DAPI (4′,6-diamidino-2-phenylindole; Biotium) for nuclei visualization. As control, uninfected samples were imaged. Samples were imaged with a Zeiss Axio Observer 7 with Lumencor Spectra X LED light system and a Hamamatsu Flash 4 camera with appropriate filters using Zen Blue software with a 40× objective.

Western blot analysis.

Samples were collected using SDS buffer and run on 15% polyacrylamide gels. Gels were transferred in the Trans-Blot Turbo transfer system (Bio-Rad). Membranes were blocked using 5% BSA in 1× Tris-buffered saline with 0.2% Tween (TBST) and probed with primary CHIKV E2 (1:1,000; BEI Resources) and actin (1:1,000) antibody (ProteinTech). Membranes were washed in 1× TBST and placed in secondary anti-mouse IgG antibody (1:15,000; Sigma-Aldrich) and incubated at room temperature for 1 h. Again, membranes were washed in 1× TBST and SuperSignal West Pico Plus chemiluminescent substrate (Thermo Fisher Scientific) was applied to membranes and developed on a molecular imager, Bio-Rad GelDoc XR+ imaging system (Bio-Rad).

RNA extraction and qPCR.

Cells were collected using TRIzol reagent (Zymo Research). RNA was purified, and DNase treated (Zymo Research) and used for cDNA synthesis with 5× All-In-One RT-PCR mastermix (BioBasic). cDNA was analyzed by qPCR using CHIKV-specific primers (44), control actin primers, and SYBR green mastermix (Thermo Fisher Scientific) by using a QuantStudio real-time PCR system (Thermo Fisher) and the ΔΔC_T method. These values were then normalized to untreated controls to allow for direct comparison between samples.

Replicon analysis.

Huh7 CHIKV replicon cells (27) (provided by Maryam Ehteshami and Raymond Schinazi) were seeded in a 96-well plate. After 24 h, cells were treated with either bisacodyl or BHPM. Twenty-four hours posttreatment, cells were lysed for 20 min using passive lysis buffer. Well contents were transferred to a solid white plate, and luminescence was measured using a microplate reader with the addition of Renilla substrate. This experiment was repeated to incorporate multiple time points up until 48 h after treatment with 175 μM bisacodyl. Luminescence was measured at the indicated times.

Statistical analysis.

Prism 9 (GraphPad) was utilized to generate graphs and perform statistical analysis. For all analyses, analysis of variance (ANOVA) and two-tailed Student's t test was used to compare groups unless otherwise noted with α values of 0.05.