key: cord-259744-r9j5yzfc
authors: McDonagh, Phillip; Sheehy, Paul A; Norris, Jacqueline M
title: Identification and characterisation of small molecule inhibitors of feline coronavirus replication
date: 2014-12-05
journal: Vet Microbiol
DOI: 10.1016/j.vetmic.2014.10.030
sha: 
doc_id: 259744
cord_uid: r9j5yzfc

Feline infectious peritonitis (FIP), a feline coronavirus (FCoV) induced disease, is almost invariably fatal with median life expectancy measured in days. Current treatment options are, at best, palliative. The objectives of this study were to evaluate a panel of nineteen candidate compounds for antiviral activity against FCoV in vitro to determine viable candidates for therapy. A resazurin-based cytopathic effect inhibition assay, which detects viable cells through their reduction of the substrate resazurin to fluorescent resorufin, was developed for screening compounds for antiviral efficacy against FCoV. Plaque reduction and virus yield reduction assays were performed to confirm antiviral effects of candidate compounds identified during screening, and the possible antiviral mechanisms of action of these compounds were investigated using virucidal suspension assays and CPE inhibition and IFA-based time of addition assays. Three compounds, chloroquine, mefloquine, and hexamethylene amiloride demonstrated marked inhibition of virus induced CPE at low micromolar concentrations. Orthogonal assays confirmed inhibition of CPE was associated with significant reductions in viral replication. Selectivity indices calculated based on in vitro cytotoxicity screening and reductions in extracellular viral titre were 217, 24, and 20 for chloroquine, mefloquine, and hexamethylene amiloride respectively. Preliminary experiments performed to inform the antiviral mechanism of the compounds demonstrated all three acted at an early stage of viral replication. These results suggest that these direct acting antiviral compounds, or their derivatives, warrant further investigation for clinical use in cats with FIP.

Feline infectious peritonitis (FIP), a feline coronavirus (FCoV) induced disease, is almost invariably fatal with median life expectancy measured in days. Current treatment options are, at best, palliative. The objectives of this study were to evaluate a panel of nineteen candidate compounds for antiviral activity against FCoV in vitro to determine viable candidates for therapy. A resazurin-based cytopathic effect inhibition assay, which detects viable cells through their reduction of the substrate resazurin to fluorescent resorufin, was developed for screening compounds for antiviral efficacy against FCoV. Plaque reduction and virus yield reduction assays were performed to confirm antiviral effects of candidate compounds identified during screening, and the possible antiviral mechanisms of action of these compounds were investigated using virucidal suspension assays and CPE inhibition and IFA-based time of addition assays. Three compounds, chloroquine, mefloquine, and hexamethylene amiloride demonstrated marked inhibition of virus induced CPE at low micromolar concentrations. Orthogonal assays confirmed inhibition of CPE was associated with significant reductions in viral replication. Selectivity indices calculated based on in vitro cytotoxicity screening and reductions in extracellular viral titre were 217, 24, and 20 for chloroquine, mefloquine, and hexamethylene amiloride respectively. Preliminary experiments performed to inform the antiviral mechanism of the compounds demonstrated all three acted at an early stage of viral replication. These results suggest that these direct acting antiviral compounds, or their derivatives, warrant further investigation for clinical use in cats with FIP.

ß 2014 Elsevier B.V. All rights reserved.

A number of compounds have demonstrated an inhibitory effect on the virus in vitro (Barlough and Shacklett, 1994; Hsieh et al., 2010; Keyaerts et al., 2007) , but there is little or no published data regarding their use in treating FIP. The broad spectrum antiviral ribavirin demonstrated in vitro efficacy but provided limited clinical benefit and produced toxicity in cats (Weiss et al., 1993) . More recently in a small study involving experimentally infected cats treatment with chloroquine, a drug with demonstrated in vitro antiviral efficacy, was associated with mild improvements in clinical signs, however there was no statistically significant difference in survival time compared to untreated cats (Takano et al., 2013) . Efficacious and safe antiviral therapeutics are desperately needed for FIP treatment.

Modern antiviral drug discovery often involves high throughput screening of vast chemical libraries. These large scale unfocused screens are expensive and beyond the reach of companion animal medicine. An alternative approach is to utilise a more focused screening strategy, enriching the screening library with compounds considered likely to have an antiviral effect based on a prior knowledge of their pharmacodynamics and the viral life cycle. Focused screening panels may consist of compounds related to those demonstrated effective against the challenge virus or those demonstrated effective against closely related viruses.

In the current study we screened 19 compounds with previously demonstrated antiviral activity against coronaviruses or other RNA viruses, for antiviral activity against FCoV using an optimised resazurin-based CPE inhibition assay. Cytotoxicity of compounds was determined prior to screening using sequential resazurin-and SRB-based assays to determine the optimal minimally toxic test concentration and to enable calculation of selectivity indices. The antiviral effects of compounds identified during screening were confirmed with plaque reduction and virus yield reduction assays. Virucidal suspension assays and time of addition assays provided initial information on the stage of viral replication targeted and the potential mechanism of action.

Crandell Rees Feline Kidney (CRFK) cell line was propagated in Dulbecco's Modified Eagle's Medium (DMEM; Sigma-Aldrich, Castle Hill, NSW, Australia) supplemented with 10% FBS (Sigma-Aldrich) (DMEM-10) in a humidified incubator at 37 8C in 5% CO 2 in air. Two strains of FCoV, FIPV WSU 79-1146 (FIPV1146) and FECV WSU 79-1683 (FECV1683), acquired from the American Type Culture Collection (Virginia, USA), were used. FCoV FECV1683 was originally isolated from mesenteric lymph nodes and intestinal washes of a 1.5 year old female domestic shorthaired cat that died of acute haemorrhagic gastroenteritis while FCoV FIPV1146 was originally isolated from the liver, spleen, and lungs from a case of neonatal death in a 4-day-old male Persian kitten (McKeirnan et al., 1981) . Pathogenicity studies of these two isolates have shown that FIPV1146 is highly virulent and reliably causes signs of classic FIP following oronasal inoculation, while FECV1683 causes a low grade fever and mild enteritis, but no signs of FIP (Pedersen, 2009) . Despite the dissimilar in vivo biological properties of the two isolates, the two have similar in vitro properties in immortalised cell lines.

Compounds were selected for the test panel based on their reported in vitro antiviral properties against coronaviruses or other RNA viruses (see supplementary material for details). The compounds tested and their screening concentrations are shown in Table 1 . Stock solutions were prepared by dissolving compounds in ultrapure water or DMSO (Sigma-Aldrich). Compounds were sterile filtered with a 0.22 mm regenerated cellulose filter (Corning Inc., Corning, NY, USA), aliquoted into sterile single use microtubes (Sarstedt, Numbrecht, Germany), and stored for a maximum of 6 months at À80 8C until use.

To determine an appropriate screening concentration, cytotoxicity of test compounds was determined using sequential resazurin and sulforhodamine B assays. The resazurin-based assay was performed as for the antiviral screening assay except compounds were added in 50 ml volume and there was no infection step. To perform the SRB assay, cells were immediately fixed post fluorescent data acquisition by decanting culture media by inverting plates and adding 10% trichloroacetic acid for 1 h at 4 8C. SRB staining was as previously described by (Vichai and Kirtikara, 2006) except that 0.2% SRB was used for staining. Following solubilisation of bound dye, OD510 was measured using the FLUOstar Omega microplate reader (BMG Labtech, Mornington, Australia). Viability was compared to untreated controls. Test compound concentrations selected for subsequent antiviral screening were those resulting in cell viability of 80% or greater. Compounds showing marked, moderate, or mild antiviral effects were defined as those showing 75-100%, 50-74%, and 25-49% inhibition of CPE respectively. Compounds demonstrating marked CPE inhibition were classified as candidate compounds and were selected for further characterisation.

Using the resazurin-based CPE inhibition assay a concentration-response experiment was conducted with serial dilutions of identified candidate compounds (nine concentrations per compound). To enable calculation of the selectivity index, a repeat cytotoxicity screen was performed concurrently. Each treatment was performed in triplicate and repeated in three independent experiments. Data were exported to Microsoft Excel for calculation of cell viability and CPE inhibition according to the formulae described above. Data analysis were conducted in GraphPad Prism, with the 50% inhibitory concentration (IC50) and 50% cytotoxic concentration (CC50) values calculated using the inbuilt non-linear curve fitting functions following log 10 transformation of compound concentrations. The selectivity index (SI) for each compound was calculated according to the following formula:

2.5. Confirmatory assays Plaque reduction and virus yield reduction assays were performed to confirm antiviral effects of candidate compounds identified using the CPE inhibition assay. Virus yield reduction assays were performed in 24-well plates (Sarstedt). Wells were seeded with 4.0 Â 10 4 cells well À1 in 400 ml DMEM-10. Plates were kept at room temperature for 30 min and then at 37 8C in 5% CO 2 in air for 5 h prior to the addition of test compounds. Compounds were diluted in DMEM to the required concentrations with 75 ml added to each well.

Cells were incubated at 37 8C in 5% CO 2 in air for an additional 1 h prior to infection with FCoV FIPV1146 at MOI 0.1 in 25 ml DMEM. Cells were incubated for a further 48 h at 37 8C in 5% CO 2 . At 24 and 48 h post-infection (hpi) cell monolayers were visually assessed for CPE using an Olympus CKX41 inverted phase-contrast microscope (Olympus, Melville, NY, USA) and culture media was collected and stored at À80 8C for virus titration. Untreated infected cells, untreated uninfected cells, and treated uninfected cells were included as controls. This latter control was included to allow assessment of morphological changes to cells due to compound treatment. Titration of extracellular virus harvested at 24 and 48 hpi was performed using the TCID50 method as described by McDonagh et al. (2011) . Each treatment and time point was performed in triplicate and repeated in two independent experiments, with results representing Mean AE SE. Plaque reduction assays were performed in 12-well plates (Corning). Cells seeded at a density of 6 Â 10 4 cells well À1 in 1 ml DMEM-10 were held at room temperature for 30 min prior to incubation at 37 8C in 5% CO 2 in air for 60 h, by which time monolayers were approximately 90% confluent. Culture media was discarded and replaced with 400 ml DMEM supplemented with 2% FBS plus 75 ml of various concentrations of test compounds in DMEM (or 75 ml DMEM only for control wells) using five or six concentrations per compound. After exposure to the compound for 1 h, cells were infected with 30 pfu well À1 FCoV FIPV1146 in 25 ml DMEM. Virus was allowed to adsorb for 90 min with plates rocked every 15 min to ensure an even distribution of inoculum. Culture media was discarded after 90 min and cells overlaid with 1 ml 0.9% carboxymethylcellulose, 2% FBS in DMEM containing the same concentration of compound as present prior to and during infection. Cells were fixed and stained with 0.1% (w/v) crystal violet 48 hpi prior to manual plaque counting. The relative plaque number was calculated for each treatment, with the value of untreated control defined as 100%. Each treatment was performed in duplicate, and repeated in three independent experiments, with data representing Mean AE SE.

A virucidal suspension assay was performed to assess virucidal effects of test compounds. The assay was performed as above with the exception that virus was mixed and incubated with test compounds prior to infection. Stock FCoV FIPV1146, diluted in DMEM to 2 Â 10 6 pfu ml À1 , was mixed with an equal volume of test compound diluted in DMEM to 2Â the test concentration used during screening. The control virus suspension was mixed with DMEM containing an equal concentration of DMSO as the test samples. Virus suspensions were incubated for 1 h at room temperature before serial dilution in DMEM to infect cells with 25 pfu well À1 in 100 ml. Following serial dilution of the virus, cells were exposed to test compounds at concentrations greater than 4 log 10 lower than concentrations previously shown to have no antiviral effect. The experiment was performed in triplicate and repeated in two independent experiments. Data represent Mean AE SE.

A modification of the resazurin-based CPE inhibition assay was performed to assess the effect of time of compound addition on the antiviral efficacy of identified compounds. The CPE inhibition assay was performed as previously described with the exception that test compounds were added at various time points before and after infection. The selected time points were 1 h prior to infection, concurrent with infection, and 1, 3, or 6 h postinfection. Treatments were performed in triplicate and repeated in three independent experiments. Data represent Mean AE SE.

To further elucidate the stage of viral replication affected by each compound the effect of time of addition on viral antigen expression was examined. Cells were seeded at a density of 5.0 Â 10 3 cells well À1 in 100 ml DMEM-10 in 96-well plates (mClear 1, Greiner Bio-One). After seeding plates were kept at room temperature for 30 min and then incubated at 37 8C in 5% CO 2 in air for 5 h prior to the first time-point of compound addition.

Compounds were added in 30 ml to duplicate wells at different time points prior to, concurrent with, or postinfection. Cells were infected with FCoV FIPV1146 at MOI 0.5 in 20 ml or mock infected with 20 ml DMEM for an infection period of 1 h. An infection period of 12 h was selected based on the reported one step growth curve of FCoV (Rottier et al., 2005) . At 12 hpi (measured from the end of the infection period) cells were fixed in 20% formaldehyde in PBS and permeabilised in ice cold methanol. Viral antigen was detected with a biotinylated anti-FCoV antibody (CCV2-2; Custom Monoclonals International, Sacramento, CA, USA) and visualised with streptavidin-conjugated Alexafluor 555 (Life Technologies, Mulgrave, VIC, Australia). To enable accurate segmentation, cells were stained with the whole cell stain HCS Cell Mask Blue (Life Technologies) and DAPI (Life Technologies) to enhance nuclear visualisation. Fluorescent imaging was performed using the BD Pathway 855 Bioimager (BD Bioscience, Franklin Lakes, NJ, USA). Images of wells were acquired using a 10Â objective (NA 0.4) using a 3 Â 3 montage with laser autofocus performed for each montage frame. HCS Cell Mask Blue/DAPI, images were acquired with Ex 380/10 BP and Em 435 LP filters, and Alexa Fluor 555 images acquired with Ex 548/20 BP and Em 570 LP filters. Image analysis was performed using the free opensource image analysis software CellProfiler (R11710, www.cellprofiler.org) with data exported to FCS Express Image Cytometry (version 4.07.0005, De Novo Software, Los Angeles, CA, USA) for analysis. Each treatment was performed in duplicate, and data represents Mean AE SD.

To assess efficacy against different FCoV strains, identified candidate compounds were tested against FCoV FECV1683 using the resazurin-based CPE inhibition assay. The assay was performed as described, except that cells were infected with either FCoV FIPV1146 or FECV1683 at MOI 0.01. Each treatment was performed in triplicate and repeated in three independent experiments, with data representing Mean AE SE.

Three of nineteen tested compounds showed marked inhibition of virus induced CPE (Fig. 1) and were selected for further characterisation. Pre-treatment with chloroquine at 25 mM, mefloquine at 10 mM, and hexamethylene amiloride at 10 mM resulted in 93.3%, 89.8%, and 77.6% inhibition of CPE respectively. A further two compounds, glycyrrhizic acid at 25 mM and cinanserin at 20 mM displayed a mild antiviral effect with a 26.7% and 34.0% reduction in CPE respectively. All other compounds demonstrated limited or no inhibitory effect on CPE. Included among these ineffective compounds was ribavirin, a broad spectrum antiviral compound that had previously shown in vitro (Barlough and Scott, 1990; Weiss and Oostrom-Ram, 1989) , and to a limited extent in vivo efficacy against FCoV (Weiss et al., 1993) , as well as rFeIFN-v which had previously shown in vitro efficacy against FCoV (Mochizuki et al., 1994; Truyen et al., 2002) .

A concentration-response study was conducted with chloroquine, mefloquine, and hexamethylene amiloride. A repeat cytotoxicity screen was concurrently performed for these compounds to allow calculation of selectivity indices. All compounds demonstrated a clear concentration-response effect over the tested range (Fig. 2) . Calculated IC50, CC50, and SI values for the compounds are shown in Table 2 .

Virus yield reduction assays confirmed the CPE inhibition identified during screening was associated with a marked reduction in extracellular viral titre. Determination of extracellular virus titre was performed at 24 and 48 hpi with results shown in Fig. 3 . For chloroquine and mefloquine there was a considerable difference in the resulting concentration-response curves at 24 and 48 hpi, while for hexamethylene amiloride the shape of the curve was similar at both time points. Differences in concentration-response curves between the two time points is reflected in the IC50 values, with increased IC50 values for chloroquine and mefloquine at 48 hpi compared to 24 hpi, while for hexamethylene amiloride IC50 values were similar at both time points (Table 3) . Plaque reduction assays confirmed the findings of the CPE inhibition and virus yield reduction assays. Pre-treatment with chloroquine, mefloquine, or hexamethylene amiloride resulted in a concentration-dependent decrease in plaque number, with high concentrations completely inhibiting macroscopic plaque formation. For all compounds plaque morphology was similar between treated and untreated wells however plaque size was smaller in treated versus untreated wells.

During the virus yield reduction assay cells were monitored for the development of CPE using phase contrast microscopy. It was noted that infected and uninfected cells treated with chloroquine, mefloquine, or hexamethylene amiloride displayed characteristic morphological changes. These changes consisted of a large number of variably sized cytoplasmic (predominantly perinuclear) inclusions in addition to the presence, in some cells, of an increased number of cytoplasmic vacuoles. To investigate the nature of these inclusions, separate wells were stained with 33 mg ml À1 neutral red in DMEM for 2 h.

These inclusions appeared to accumulate the vital dye neutral red following suggesting they were likely dilated endosomes/lysosomes (Fig. 4) .

Using a virucidal suspension assay no virucidal effects were seen for chloroquine, mefloquine, or hexamethylene amiloride, with the infectivity of virus suspensions exposed to the compounds not significantly different from virus incubated with media alone.

The effect of time of addition on the antiviral activity of selected compounds was assessed using a modification of the resazurin-based CPE inhibition assay and through IFA of viral protein expression. Based on the CPE inhibition assay maximum antiviral effect was seen when compounds were added prior to or concurrent with infection, following which there was a time-dependent reduction in CPE inhibition (Fig. 5) . For all tested compounds CPE inhibition remained greater than 50% when compounds were added at the latest tested time point of 6 h postinfection.

The CPE inhibition assay encompasses multiple rounds of viral replication. To further elucidate the stage of viral replication affected by test compounds a single replication cycle IFA-based assay was conducted which confirmed that, based on viral antigen expression, all three compounds possess antiviral properties when added prior to, or at the time of infection. Furthermore all compounds displayed a time of addition dependent reduction in antiviral effect; however the extent and timing of this reduction varied. The inhibitory effect of chloroquine was reduced, based on an increase in the percentage of FCoV antigen positive cells, when added at any time postinfection (Fig. 6) . A similar result was seen for hexamethylene amiloride, although in this case a significant increase in the number of infected cells was not seen until compound addition was delayed until 1 hpi. In contrast, mefloquine remained effective when added up to 5 hpi suggesting it may act at a later stage of viral replication than chloroquine and hexamethylene amiloride.

The efficacy of the three identified candidate compounds was tested against FCoV FECV1683, a serotype II enteric biotype FCoV. Comparison of the virus control (no treatment) wells showed FCoV FIPV1146 infection resulted in more pronounced CPE over the 72 h infection period compared to FCoV FECV1683. Pre-treatment with chloroquine, mefloquine, or hexamethylene amiloride provided a degree of protection against strain FCoV FECV1683. Pretreatment with hexamethylene amiloride provided protection against virus induced CPE that was similar for the two strains, with a reduction in CPE of 89.5% and 86.0% for FCoV FIPV1146 and FECV1683 respectively. Both chloroquine and mefloquine however were more effective against FCoV FIPV1146 than FECV1683, with CPE inhibition for chloroquine of 76.9% for versus 63.8%, and for mefloquine 79.0% versus 67.5% for strains FIPV1146 and FECV1683 respectively. 

This study identifies three compounds (chloroquine, mefloquine, and hexamethylene amiloride) demonstrating a marked inhibitory effect on FCoV replication in vitro by significant reductions in virus induced CPE and viral titres at low micromolar concentrations when present during the early stages of viral replication. An antiviral effect of chloroquine had previously been demonstrated against FCoV, and hexamethylene amiloride had previously demonstrated efficacy against other coronaviruses, however this is the first demonstration of antiviral efficacy of mefloquine against a coronavirus.

Initial compound screening was performed using a CPE inhibition assay, with subsequent virus yield reduction assays and plaque reduction assays used for confirmatory testing. For the effective compounds the IC50 values, and corresponding selectivity index, varied with the assay method utilised. This is not unexpected given the assays measure different endpoints, and has been reported for other antiviral drugs such as the retroviral protease inhibitor saquinavir where the reported IC50 calculated based on production of viral p24 antigen is approximately 30-fold lower than that based on production of mature virions (Buss and Cammack, 2001) . Similarly variation in assay conditions may result in the calculation of significantly different IC50 values. The concentration-response curve of chloroquine against SARS-CoV determined using a PCR based virus yield reduction assay was shown to shift considerably to the right when viral genome copies were assayed 3 days post-infection compared to 1 day postinfection (Keyaerts et al., 2004) . A similar finding was noted in the current study for both chloroquine and mefloquine, with differences in potency reported with the TCID50 based virus yield reduction assay performed at 24 and 48 hpi, however this was not seen for hexamethylene amiloride.

Two compounds, ribavirin and rFeIFN-v, which had previously demonstrated in vitro efficacy against FCoV, failed to demonstrate significant inhibition of CPE during screening. For both compounds these discordant results are likely attributable to testing at concentrations below their useful therapeutic range and variations in assay conditions and sensitivity compared with previous work. The screening concentration of compounds used in this study was determined based on cytotoxicity testing to achieve cell viability greater than 80%. Previous studies with ribavirin demonstrated IC50 values of 41.7 mg ml À1 (170 mM) (Barlough and Scott, 1990 ) based on a visual assessment of protection from cytopathic effect and 2.5 mg ml À1 (10.2 mM), based on the reduction of extracellular viral titre (Weiss and Oostrom-Ram, 1989) . The concentration used for screening (2.5 mM) was therefore more than 60 times lower than the IC50 previously calculated based on a similar assay endpoint. From the results of the current study, virus yield reduction assays appear to provide a more sensitive assessment of antiviral efficacy than CPE inhibition assays, with the IC50 values calculated based on viral titre reduction significantly lower than those calculated based on CPE inhibition for all compounds. A small antiviral effect of ribavirin cannot therefore be ruled out based on the current findings, as although the tested concentrations did not provide protection against virus induced CPE, it may have been associated with a reduction in extracellular viral titre. The practical relevance of such a small antiviral effect is questionable, particularly given the known toxicity profile of this compound in cats.

For rFeIFN-v reductions in viral titres of 0.2-1.2 logs have been reported when CRFK cells were treated with 50,000 U ml À1 1 h post-infection (Truyen et al., 2002) and 0.5-0.6 logs when FCWF cells were pre-treated with 100-100,000 U rFeINF-v (Mochizuki et al., 1994) . Protection from CPE was not seen in the current study when cells were pre-treated with rFeINF-v at 100 U ml À1 , a concentration significantly lower than that previously shown to be effective using the same virus strain and cell line (Truyen et al., 2002) . The tested concentration was however similar to that used by Mochizuki et al. (1994) . This apparent lack of efficacy in this case may reflect differences in the drug exposure and infection conditions, the viral isolate tested, or an intrinsic enhanced susceptibility to the antiviral effects of interferon in FCWF cells compared to CRFK cells as used in this study (Weiss and Toivio-Kinnucan, 1988) . Alternatively it may be that, as suggested for ribavirin, virus yield reduction assays provide a more sensitive assessment of antiviral effects than CPE inhibition assays, and that the screening method utilised failed to identify mild antiviral effects.

A number of different mechanisms of action have been suggested to account for the antiviral properties of the compounds identified in this study against other viruses. For chloroquine antiviral effects have been ascribed to inhibition of glycosylation of viral proteins (Savarino et al., 2004) or cellular receptors for viral attachment (Vincent et al., 2005) , inhibition of glycoprotein expression (Dille and Johnson, 1982) , or inhibition of endosome mediated viral entry (Savarino et al., 2003) . The antiviral effect of mefloquine against JC virus has been postulated to be due to its action as an adenosine mimetic (Brickelmaier et al., 2009) , while for hexamethylene amiloride it has been suggested antiviral properties against different viruses may arise through competitive inhibition of viral RNA polymerase (Gazina et al., 2011) , an indirect mutagenic effect (Levi et al., 2010) , or inhibition of viroporins (Wilson et al., 2006) .

Interestingly all three compounds showing marked antiviral efficacy against FCoV in this study resulted in similar morphological changes in cells exposed to sub-toxic concentrations. Increased numbers of variably size cytoplasmic inclusions that accumulate the viral dye neutral red suggests these compounds result in a perturbation of the normal endocytic pathway in CRFK cells. Alterations in the endocytic pathway have previously been reported for chloroquine (Dean et al., 1984) , mefloquine (Labro and Babin-Chevaye, 1988) , and for amiloride and some of its derivatives (Dutta and Donaldson, 2012 ). This suggests a common physiological effect on treated cells for all three candidate antivirals and possibly a shared mechanism of action. Viruses are known to usurp a variety of host endocytic pathways for cell entry and intracellular movement and inhibition of these pathways may be a useful therapeutic approach. Although targeting a cellular pathway may be associated with an increased risk of toxicity, if that pathway is critical for viral replication this approach may slow or limit the development of resistance.

Time of addition studies demonstrated all compounds were most effective when added prior to infection, suggesting a mechanism of action involving early stages of viral replication. The CPE inhibition based time of addition assay involved infection at low MOI with a 72 h infection period, allowing for multiple rounds of viral replication. As a result of this, even with the delayed addition of compounds, cells uninfected by the original inoculum are effectively pre-treated prior to challenge with progeny virions produced during the primary replication cycle. Using an IFA-based time of addition study involving a single replication cycle we were able to further clarify of the effect of time of addition, and refine the possible stage of the viral life cycle targeted by each compound. Based on the IFA results chloroquine was effective only if present at the time of infection, supporting the hypothesis that chloroquine acts during cell entry for FCoV FIPV1146, possibly through inhibition of endosomal pH (Takano et al., 2008) . Hexamethylene amiloride and mefloquine provided significant antiviral effects when compound addition was delayed for up to 1 and 5 hpi respectively, suggesting that if the antiviral effects of these compounds arise through perturbation of endosomal function, the effects occur at different stages of the viral life cycle. Alternatively distinct mechanisms of action may account for the observed effects of these compounds, as suggested for other viruses.

There is limited published pharmacokinetic or safety data to inform the potential therapeutic application of the identified compounds in cats and given the relatively low SI of all three compounds consideration must be given to their in vivo safety in this species. The human approved pharmaceuticals chloroquine and mefloquine are generally considered well-tolerated drugs, albeit with a narrow therapeutic index, while the clinical use of hexamethylene amiloride has not been reported. Pharmacokinetic data available for chloroquine and mefloquine in humans would suggest that effective plasma concentrations could be achieved at standard therapeutic doses (Pussard and Verdier, 1994; Simpson et al., 1999) . Chloroquine has been shown to accumulate in leukocytes, where the concentration may be two orders of magnitude greater than that of plasma (Mackenzie, 1983) , with the highest concentration reported in monocytes (French et al., 1987) . Thus therapeutic concentrations may be attained in the target cells of virulent biotype FCoVs at relatively low plasma concentrations, minimising the risk of dose-dependent adverse effects. Mefloquine is known to accumulate within brain parenchyma at concentrations approximately 10-30 times higher than found in serum, with tissue concentrations of up to 50 mM reported (Nevin, 2009; Pham et al., 1999) . Mefloquine may therefore be useful in the treatment of dry (non-effusive) FIP, where CNS lesions are common (Pedersen, 2009 ) although the potential for neurotoxicity must be considered. Although the concentration of mefloquine achieved in the CNS is greater than the CC50 of this compound in immortalised feline kidney cells, in humans this tissue concentration is achievable at therapeutic doses, despite in vitro data in human cells showing a CC50 approximately equal to that determined in the current study (Brickelmaier et al., 2009) . It may be therefore that the more static cell population of the CNS is more refractory to the toxic effects of mefloquine than mitotically active immortalised cells.

This study has identified three compounds demonstrating marked in vitro inhibition of FCoV in an immortalised cell line at low micromolar concentrations, including the first demonstration of antiviral effects of mefloquine against a coronavirus. Although the low SI of the three compounds may limit their therapeutic utility, these preliminary studies open the way for further investigation and potential optimisation of these compounds as antiviral agents.

Effectiveness of three antiviral agents against FIP virus in vitro

Antiviral studies of feline infectious peritonitis virus in vitro

Identification and characterization of mefloquine efficacy against JC virus in vitro. Antimicrob

Measuring the effectiveness of antiretroviral agents

Effects of exogenous amines on mammalian cells, with particular reference to membrane flow

Inhibition of vesicular stomatitis virus glycoprotein expression by chloroquine

Search for inhibitors of endocytosis: intended specificity and unintended consequences

Randomized, placebo controlled study of the effect of propentofylline on survival time and quality of life of cats with feline infectious peritonitis

Uptake of chloroquine and hydroxychloroquine by human blood leucocytes in vitro: relation to cellular concentrations during antirheumatic therapy

Amiloride is a competitive inhibitor of coxsackievirus B3 RNA polymerase

Treatment of cats with feline infectious peritonitis

Synergistic antiviral effect of Galanthus nivalis agglutinin and nelfinavir against feline coronavirus

In vitro inhibition of severe acute respiratory syndrome coronavirus by chloroquine

Plant lectins are potent inhibitors of coronaviruses by interfering with two targets in the viral replication cycle

Effects of amodiaquine, chloroquine, and mefloquine on human polymorphonuclear neutrophil function in vitro

Fidelity variants of RNA dependent RNA polymerases uncover an indirect, mutagenic activity of amiloride compounds

Pharmacologic actions of 4-aminoquinoline compounds

In vitro inhibition of feline coronavirus replication by small interfering RNAs

Isolation of feline coronaviruses from two cats with diverse disease manifestations

Inhibitory effects of recombinant feline interferon on the replication of feline enteropathogenic viruses in vitro

Epileptogenic potential of mefloquine chemoprophylaxis: a pathogenic hypothesis

A review of feline infectious peritonitis virus infection: 1963-2008

Cerebral uptake of mefloquine enantiomers in fatal cerebral malaria

Antimalarial 4-aminoquinolines: mode of action and pharmacokinetics

Effect of feline interferon-omega on the survival time and quality of life of cats with feline infectious peritonitis

Acquisition of macrophage tropism during the pathogenesis of feline infectious peritonitis is determined by mutations in the feline coronavirus spike protein

Effects of chloroquine on viral infections: an old drug against today's diseases

Anti-HIV effects of chloroquine: inhibition of viral particle glycosylation and synergism with protease inhibitors

Population pharmacokinetics of mefloquine in patients with acute falciparum malaria&ast

Analysis of the mechanism of antibody-dependent enhancement of feline infectious peritonitis virus infection: aminopeptidase N is not important and a process of acidification of the endosome is necessary

Effect of chloroquine on feline infectious peritonitis virus infection in vitro and in vivo

Antiviral potency of interferon-omega (IFN-omega) against selected canine and feline viruses

Sulforhodamine B colorimetric assay for cytotoxicity screening

Chloroquine is a potent inhibitor of SARS coronavirus infection and spread

Evaluation of free or liposomeencapsulated ribavirin for antiviral therapy of experimentally induced feline infectious peritonitis

Inhibitory effects of ribavirin alone or combined with human alpha interferon on feline infectious peritonitis virus replication in vitro

Inhibition of feline infectious peritonitis virus replication by recombinant human leukocyte (alpha) interferon and feline fibroblastic (beta) interferon

Hexamethylene amiloride blocks E protein ion channels and inhibits coronavirus replication

This study was supported by donations from the Rex Cat Club (especially Sharon Barton and Tracey Gleeson), participants of the National Annual Feline Health Seminars, Christine Atkins, Ruth Thurling, Cat Fanciers Association of NSW, and the Cat Protection Society of NSW.

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.vetmic.2014.10.030.