Antifungal Triazole Posaconazole Targets an Early Stage of the Parechovirus A3 Life Cycle

Viruses in species Parechovirus A (Picornaviridae) are associated with a wide variety of clinical manifestations. Parechovirus A3 (PeV-A3) is known to cause sepsis-like illness, meningitis, and encephalitis in infants and young children. To date, no specific therapies are available to treat PeV-A3-infected children. We had previously identified two FDA-cleared antifungal drugs, itraconazole (ITC) and posaconazole (POS), with potent and specific antiviral activity against PeV-A3.

ABSTRACT

Viruses in species Parechovirus A (Picornaviridae) are associated with a wide variety of clinical manifestations. Parechovirus A3 (PeV-A3) is known to cause sepsis-like illness, meningitis, and encephalitis in infants and young children. To date, no specific therapies are available to treat PeV-A3-infected children. We had previously identified two FDA-cleared antifungal drugs, itraconazole (ITC) and posaconazole (POS), with potent and specific antiviral activity against PeV-A3. Time-of-addition and synchronized infection assays revealed that POS targets an early stage of the PeV-A3 life cycle. POS exerts an antiviral effect, evidenced by a reduction in viral titer following the addition of POS to Vero-P cells before infection, coaddition of POS and PeV-A3 to Vero-P cells, incubation of POS and PeV-A3 prior to Vero-P infection, and at attachment. POS exerts less of an effect on virus entry. A PeV-A3 enzyme-linked immunosorbent assay inhibition experiment, using an anti-PeV-A3 monoclonal antibody, suggested that POS binds directly to the PeV-A3 capsid. POS-resistant PeV-A3 strains developed by serial passage in the presence of POS acquired substitutions in multiple regions of the genome, including the capsid. Reverse genetics confirmed substitutions in capsid proteins VP0, VP3, and VP1 and nonstructural proteins 2A and 3A. Single mutants VP0_K66R, VP0_A124T, VP3_N88S, VP1_Y224C, 2A_S78L, and 3A_T1I were 4-, 9-, 12-, 34-, 51-, and 119-fold more resistant to POS, respectively, than the susceptible prototype strain. Our studies demonstrate that POS may be a valuable tool in developing an antiviral therapy for PeV-A3.

INTRODUCTION

Parechoviruses are positive-sense, single-stranded RNA viruses of the family Picornaviridae. There are four recognized species within genus Parechovirus. Parechovirus A contains 19 types detected in humans. Parechovirus B contains five types isolated from rodents and birds. The two new parechovirus species, Parechovirus C and D, each comprise a single type, and both have been found only in rodents (1). The parechovirus genome is approximately 7,300 nucleotides and encodes a single polyprotein, which is cleaved by viral protease 3C into structural capsid proteins (VP0, VP1, and VP3) and nonstructural proteins (2A, 2B, 2C, 3A, 3B, 3C, and 3D) (2, 3). The VP1 capsid protein shows a high degree of amino acid sequence divergence among different Parechovirus A types. The cellular receptor for parechovirus A1 (PeV-A1) is a host cell surface integrin bound via an arginine-glycine-aspartic acid (RGD) motif at the carboxyl terminus of VP1 (4, 5). PeV-A2, -A4, -A5, and -A6 contain an RGD amino acid motif at the carboxyl terminus of VP1 that may serve a similar function (6, 7). PeV-A3 and PeV-A7 through PeV-A17 lack the RGD motif (8). The cellular receptor for PeV-A3 is unknown, and the lack of the RGD motif implies that PeV-A3 uses a different receptor than PeV-A1, -A2, -A4, -A5, or -A6 (8). To date, parechovirus types A7 to A19 have not been successfully propagated in cell culture.

Parechoviruses are associated with a wide variety of clinical manifestations. PeV-A3 can cause serious disease in infants and young children, including sepsis-like illness, meningitis, and encephalitis (9–12). PeV-A3 has been the most common parechovirus detected from cerebrospinal fluid (CSF) since its discovery in 1999, emphasizing its importance as an emerging agent of serious infections in young children (4, 6). To date, no specific therapies are available to treat PeV-A3-infected children. FDA-cleared antifungal drugs itraconazole (ITC) and posaconazole (POS) were previously identified as potent broad-spectrum inhibitors of enterovirus (13–15), dengue virus (16), and PeV-A3 (17) at concentrations clinically attainable in pediatric patients. This study explores the possible mechanism(s) of action for in vitro antiviral activity of POS against PeV-A3.

RESULTS

POS is an early-stage inhibitor of PeV-A3 replication.

To determine at which stage in the PeV-A3 life cycle POS and ITC exert their antiviral effects, a time course experiment was performed with compounds added at −1, 0, +1, +2, +4, and +6 h relative to virus infection (Fig. 1). POS at 0.3 μM and ITC at 1 μM or 3 μM were more effective inhibitors at earlier time points than the no-drug control.

FIG 1.

Antifungal azoles POS and ITC exert antiviral activity against PeV-A3 in the early stages of infection. The cytopathic effect of POS (0.3 and 1 μM) and ITC (1 and 3 μM) on PeV-A3 strain US-WI-09 was evaluated in a time course assay. Results represent means ± standard deviations (SD) from five independent experiments.

To further characterize the mechanism of action of POS, we used a time-of-addition assay to evaluate the antiviral effect on virus replication of drug pretreatments, coaddition, and addition postinfection (Fig. 2A). The virus titer was decreased by ∼100-fold (2 log₁₀) when cells were pretreated with POS prior to infection (Fig. 2B, pretreatment-1) and decreased by ∼31-fold (1.49 log₁₀) when PeV-A3 and POS were added simultaneously (Fig. 2B, coaddition). When POS was added after virus infection, titers decreased by only ∼3-fold (0.49 log₁₀) (Fig. 2B, postinfection), which suggests POS is more effective earlier in the PeV-A3 infection cycle. To assess whether POS targets the PeV-A3 particle directly and not the host cell, Vero-P cells were pretreated with POS and any unbound drug was washed away prior to virus infection. A reduction of antiviral activity resulted when unbound POS was removed prior to addition of virus; only a modest decrease of viral titer (0.44 log₁₀; ∼2.75-fold) was observed (Fig. 2B, pretreatment-2).

FIG 2.

POS is more effective early in PeV-A3 life cycle. (A) Schematics of the treatments for the time-of-drug-addition assays, utilizing Vero-P cell monolayers. Shaded areas indicate the presence of 0.5 μM POS. One hundred CCID₅₀ PeV-A3 (strain US-WI-09) or PeV-A1 (Harris strain) was used for all experiments. Titration results were obtained using the virus titration assay previously described (45). (B) The pretreatment-1 experiment showed the effect of POS on Vero-P cells before PeV infection. Monolayers were incubated with 0.5 μM POS, followed by the addition of PeV. Residual POS and unbound PeVs were removed by washing. The pretreatment-2 experiment showed the effect of POS on Vero-P cells with residual drug removed by washing before inoculation. PeV was added and incubated for 2 h, followed by a second wash to remove unbound virus. For the coaddition test, Vero-P monolayers were inoculated simultaneously with POS and PeV and incubated for 2 h, followed by washing to remove unbound POS and PeV. In the postinfection configuration, Vero cells were infected with PeV, followed by removal of unbound virus by washing. POS was added and remained present throughout incubation. Data shown represent mean values from three independent experiments.

POS blocks early steps in viral entry by binding to virus particles and inhibiting viral attachment.

The time course and time-of-drug-addition assays suggested that POS exerts its antiviral effect early in the PeV-A3 infection cycle, independent of interaction with the Vero-P cells. To further demonstrate it directly targets the virus, we performed in vitro synchronized infection assays (18) to evaluate the antiviral effect of POS when combined with PeV-A3 prior to inoculation of cells (free virus particles), on PeV-A3 attachment, and on PeV-A3 entry (Fig. 3A and B). One hundred microliters of concentrated PeV-A3 (500 CCID₅₀ [50% cell culture infectious dose]) was incubated with 100 μl (0.5 μM) POS for 2 h at 35°C and then diluted 20-fold in growth medium. This dilution reduces the drug concentration (0.025 μM) to approximately 10-fold below the 50% effective concentration (EC₅₀) (0.24 μM). Vero-P cells were inoculated with the diluted drug-virus mixture and titers determined as described previously. The PeV-A3 titer was reduced by ∼27.5-fold (1.44 log₁₀) compared to that of the virus-only control (Fig. 3B, free virus particles), suggesting that POS bound directly to PeV-A3 prior to inoculation of cells. That is, the drug exerted an antiviral effect when incubated directly with virus before inoculation of cells, even though the drug concentration in the final assay was well below the EC₅₀ for POS.

FIG 3.

POS blocks early viral entry by binding to PeV-A3 particles and inhibiting viral attachment. (A) Schematics of the treatments for the synchronized infection assays, utilizing Vero-P cell monolayers. Shaded areas indicate the presence of 0.5 μM POS. One hundred CCID₅₀ PeV-A3 (strain US-WI-09) or PeV-A1 (Harris strain) was used for all experiments. Titration results were obtained using the antiviral activity assay previously described (45). Data represent mean values from three independent experiments. (B) For the free virus particle experiment, POS and PeV were mixed in a sterile tube and incubated for 2 h at 35°C, followed by a 20-fold dilution to reduce the POS concentration to approximately 10-fold below the EC₅₀. Subsequently, Vero-P monolayers were inoculated with the diluted POS/PeV solution and incubated for 2 h at 35°C, followed by a wash to remove any unbound POS and PeV. For attachment, Vero-P cells were held at 4°C and then inoculated with PeV and POS simultaneously. After 2 h of incubation at 4°C, the monolayers were washed to remove residual PeV and POS. The temperature was raised to 35°C for the duration of the experiment. For the entry test, Vero-P monolayers were held at 4°C, inoculated with PeV, and incubated for 2 h, followed by a wash to remove unbound PeV. The temperature was shifted to 35°C and POS was added for 2 h of incubation, followed by washing to remove unbound POS.

To examine the effect on the viral attachment stage, cells were incubated with PeV-A3 and POS at 4°C for 2 h (Fig. 3A and B). At this temperature, the viral particles bind to the cell surface but entry is prevented (18). After washing to remove unbound virus and drug, the cells were shifted to 35°C to allow the infection to progress. POS had a moderate antiviral effect on the attachment phase, reducing the PeV-A3 titer by ∼9-fold (0.95 log₁₀) (Fig. 3B, attachment). We next examined the effect of POS on PeV-A3 cell entry. Vero-P monolayers and virus were incubated at 4°C before removal of free PeV-A3 by washing. POS was added to the medium and the temperature shifted to 35°C (Fig. 3A and B, entry). POS addition had only a minimal effect on virus titer (∼1.8-fold; 0.25 log₁₀, entry) compared to the effects of POS on virus binding and attachment (Fig. 3B, free virus particles and attachment).

To further investigate the interaction of POS with the PeV-A3 capsid, we developed a PeV-A3 enzyme-linked immunosorbent assay (ELISA). Incubation of PeV-A3 with POS inhibited subsequent binding by the PeV-A3 monoclonal antibody (MAb) compared to that of the no-drug control (Fig. 4A and C). Voriconazole (VRC), which is inactive against PeV-A3, had little effect on PeV-A3 antibody binding (Fig. 4B and D).

FIG 4.

POS binds directly to PeV-A3. Microplates were coated with 0.5 μg/well rabbit anti-PeV-A3 pAb (16-1071) and incubated overnight at 4°C. The next day, 100 CCID₅₀ PeV-A3 (US-WI-09) was incubated with 10 μM drug (A and B), drug titration (C and D), or without drug (A to D) for 2 h at room temperature. The contents were transferred to pAb-coated microplates and incubated for 1 h. Unbound drug and PeV-A3 were removed by washing, and bound components were incubated with mouse anti-PeV-A3 MAb (clone AB6-BA9-AH7) for 1 h. The effects of drug, PeV-A3, and MAb were detected using a goat anti-mouse IgG (H+L) peroxidase-labeled antibody and TMB substrate. (A and C) The effect of POS binding to PeV-A3. (B and D) The effect of VRC binding to PeV-A3. Data represent mean absorbance at 620 nm from three independent experiments.

POS and ITC have an additive antiviral effect when added in combination.

To examine the effect of these drugs when added in combination, we used the MacSynergy II program (19) to calculate a theoretical additive value for each drug combination. POS and ITC, when tested in combination, generated a synergy volume of 8.5 μM²%, which represents an additive antiviral effect (Fig. 5). This suggests that the structurally similar drugs POS and ITC have a similar mechanism(s) of action against PeV-A3.

FIG 5.

In vitro combination effect of antifungal azoles POS and ITC. POS (termed PSZ here) exhibits an additive antiviral effect when tested in combination with ITC (ITZ). Differential surface plots were derived from data generated using MacSynergy II. Volumes of synergy or antagonism that deviate significantly from the expected additive drug interactions are shown. The shading indicates the level of synergy or antagonism. The volumes are calculated at the 95% confidence interval (CI). Five parallel repeats were tested in each measurement.

Generation and characterization of POS- and ITC-resistant PeV-A3 variants.

To obtain PeV-A3 variants that are resistant to POS and ITC, we passaged PeV-A3 strain US-WI-09 in the presence of gradually increasing concentrations of POS or ITC (Fig. 6A). Vero-P passages 1 through 7 had lower titers than the parental virus for both POS and ITC, showing a consistent antiviral effect (Fig. 6B and C). The drug-treated PeV-A3 strains exhibited measurable titer increases starting at passage 8. By passage 13 the titers for the drug-treated strains approached those of the parental strain, suggesting the emergence of drug-resistant phenotypes (Fig. 6B and C).

FIG 6.

Selection of POS- and ITC-resistant PeV-A3. (A) Schematic of drug-resistant virus selection. POS- and ITC-resistant PeV-A3 was selected by culturing the PeV-A3 strain US-WI-09, either POS treated (B) or ITC treated (C), for 14 passages on confluent Vero-P cells in the presence of gradually increasing concentrations of POS or ITC. Viral titers were determined after each passage as described previously (45). Results represent averages from three experiments.

The complete genomes of the parental PeV-A3 strains (P0) and the drug-resistant variant strains US-WI-09(ITC) and US-WI-09(POS) revealed amino acid substitutions in the VP0, VP3, VP1, 2A, and 3A regions (Fig. 7). Amino acid substitutions in VP1 were observed in all of the resistant variants. The 3A T1I substitution for US-WI-09(POS) and US-WI-09(ITC) is notable, as it changes the 2C-3A cleavage site from Q/T to Q/I. The POS-resistant variant was 142-fold more resistant to POS than parental strain US-WI-09. Various amounts of cross-resistance were observed between the POS- and ITC-resistant variants (Fig. 7).

FIG 7.

Antiviral phenotypes and the location of the nonsynonymous amino acid substitutions in the capsid and the nonstructural regions. Superscript letters in the table: a, results determined in the antiviral activity assay as previously described from three independent experiments; b, relative resistance (EC₅₀ of resistant isolate/EC₅₀ of parental); c, the substitution Y224C is at the C terminus of VP1 and is not included in the X-ray crystallography, which ends with GSRAA, whereas the predicted sequence is GSRAAALYDE (the green annotation marks the amino acid A at three positions preceding Y224C); d, the T1I substitutions at the 3A region are detected at the 2C-3A cleavage site, changing it from Q/T to Q/I.

We used reverse genetics to generate a prototype infectious PeV-A3 clone to confirm the resistant residues identified in the variant screen. Six single mutants, VP0_K66R, VP0_A124T, VP3_N88S, VP1_Y224C, 2A_S78L, and 3A_T1I, were 4-, 9-, 12-, 34-, 51-, and 119-fold, respectively, more resistant to POS than its susceptible prototype strain. Itraconazole exhibited cross-resistance with the POS-resistant mutants at 6-, 4-, 4-, 10-, 15-, and 11-fold, respectively (Table 1).

TABLE 1.

Antiviral activity of POS and ITC against reverse-engineered PeV-A3 mutants

PeV-A3 (US-WI-09)	Clone	POS/EC50a ± SD (μM)	RRb	ITC/EC50a ± SD (μM)	RRb
PeV-A3-R1	Prototype	0.27 ± 0.06		1.61 ± 0.29
PeV-A3-R2	VP0_K66R	0.88 ± 0.22	4	10.2 ± 1.37	6
PeV-A3-R3	VP0_A124T	2.57 ± 0.63	9	6.51 ± 0.75	4
PeV-A3-R4	VP3_N88S	3.41 ± 0.54	12	6.81 ± 1.05	4
PeV-A3-R5	VP1_Y224C	9.56 ± 1.15	34	15.5 ± 2.43	10
PeV-A3-R6	2A_S78L	14.3 ± 2.84	51	24.5 ± 2.55	15
PeV-A3-R7	3A_T1I	33.2 ± 3.75	119	17.3 ± 2.04	11

^aShown are results determined in the antiviral activity assay, as previously described from five independent experiments.

^bRR, relative resistance (EC₅₀ of resistant clone/EC₅₀ of prototype).

The growth kinetics and endpoint titers for the POS-resistant mutants were not significantly different from those of the parental (US-WI-09) virus (unpublished observation). We tested the thermostability of POS on the parental PeV-A3 strain (US-WI-09) and the POS-resistant mutant (VP1_Y224C). The presence of 0.25 μM POS (continuous throughout the experiment) had a greater effect on thermostability for the parental PeV-A3 strain (Fig. 8A) than the VP1_Y224C mutant (Fig. 8B). These results suggest that POS stabilizes the viral capsid at elevated temperatures.

FIG 8.

Thermostability assay. Effect of POS on heat inactivation of PeV-A3 strain US-WI-09. Five hundred CCID₅₀ of virus was incubated with 0.25 μM POS or an equivalent volume of medium at temperatures ranging from 37°C to 55°C for 5 min and then rapidly cooled to 4°C. Viral infectivity was quantified in Vero-P cells using the virus titration assay. (A) PeV-A3 parental strain US-WI-09. (B) PeV-A3 drug-resistant strain VP1_Y224C. Results represent means from three experiments, and error bars are standard deviations.

DISCUSSION

PeV-A3 infections are an important cause of sepsis-like illness in infants and young children (9–11). There are no antiviral drugs available to treat diseases caused by PeV infection. The antifungal triazole drug ITC has been shown to inhibit enterovirus replication in cell culture (13, 14), and we recently demonstrated that ITC and POS are also active against PeV-A3 (17). In clinical studies for its antifungal indication, POS was safe and well tolerated, without severe side effects in pediatric patients (20–22), and efficacious mean plasma concentrations were achieved for treatment of invasive fungal disease in pediatric patients (23). Despite an incomplete understanding of POS’s pharmacokinetic characteristics, another recent study concluded that POS may be a valuable antifungal agent in children based on the correlation between oral POS dose and trough plasma levels (24). A retrospective review of intravenous POS dosing demonstrated attainment of therapeutic trough plasma levels in 9 of 10 pediatric patients (25).

Given the need for PeV antivirals and existing drugs with activity at physiologically achievable concentrations, we sought to better understand the mechanism of POS’s PeV-A3-specific antiviral activity. Itraconazole inhibits enterovirus replication through interaction with the cellular oxysterol binding protein (OSBP), a protein involved in intracellular lipid transfer (14). Despite the importance of the triazole moiety for the antifungal effect, encephalomyocarditis virus (EMCV; another picornavirus, in the genus Cardiovirus) was inhibited by stereoisomers of ITC, suggesting the triazole structure is not important for antiviral activity (26). POS and ITC are structurally similar to one another, and when tested in combination, an additive antiviral effect was demonstrated, which is consistent with a similar mechanism of action. Previous studies suggested that an extended chemical side chain is responsible for enterovirus antiviral activity (13). Further exploration into the structure-activity relationship of POS may lead to development of POS-like antiviral compounds with more favorable drug-like properties (27). Recent studies have identified the oxysterol-binding protein (OSBP) and viral protein 3A as important targets responsible for the broad-spectrum enterovirus antiviral activity of ITC and POS (13, 14, 26). Our drug resistance experiments confirm the importance of the 3A viral protein as one of several targets responsible for PeV-A3 antiviral activity (Table 1). We recently identified ITC and POS as specific inhibitors of PeV-A3 activity, but the two drugs had no detectable antiviral activity against other cultivable parechoviruses (PeV-A1, -A2, and -A4 to -A6), suggesting a different mechanism for inhibition of PeV-A3 by these antifungal azoles (17). A recent study, using different assay methods, confirmed that ITC had specific antiviral activity against PeV-A3 and not PeV-A1 (28).

Members of a class of antienteroviral compounds known as capsid inhibitors exert their antiviral effect by inserting into the hydrophobic pocket in the virus canyon and replacing a lipid pocket factor, thereby preventing virus uncoating (29–31). Recently completed clinical trials of capsid inhibitors pleconaril and pocapavir have demonstrated safety and efficacy (32, 33), and studies using pocapavir as a treatment for enterovirus infection in neonates have shown some success (34–36). Recently, a novel early-stage capsid inhibitor, 4-dimethylaminobenzoic acid (4EDMAB), was shown to inhibit replication of coxsackievirus B3 (CVB3) with a mechanism targeting the capsid outside the typical binding pocket (37). Collectively, these results suggest that the viral capsid is an important target for development of antivirals and, for 4EDMAB, that a noncanonical mechanism is possible. 4EDMAB also protected both parental and resistant variants from heat inactivation. Recently, nucleoside viral polymerase inhibitors 7-deaza-2′-C-methyladenosine and 2′-C-methylcytidine were identified as in vitro replication inhibitors of PeV-A1 and -A3 (28). Preliminary time course studies demonstrated that POS and ITC exert their PeV-A3 antiviral effect at an early stage in the viral life cycle (Fig. 1). Both POS (0.3 μM) and ITC (1 μM, 3 μM) exhibited antiviral activity when Vero-P monolayers were treated with drug prior to, simultaneously with, or after the addition of PeV-A3. At these concentrations, the antiviral activity continued to decrease as either drug was introduced later in the infection cycle, except for the postinfection 1 μM POS treatments. Low PeV-A3 titers were measured for all the postinfection 1 μM POS addition treatments. This is likely due to the inhibition of PeV-A3 progeny by the high level of POS in the medium and perhaps provides indirect evidence that POS acts on the capsid.

Time-of-addition assays (Fig. 2) demonstrated that POS exerts an antiviral effect on the PeV-A3 capsid and found little measurable effect on PeV-A3 titers due to POS interaction with the Vero-P cells prior to monolayer infection. A decrease in antiviral activity was observed when POS was added after PeV-A3 infection (Fig. 2B, postinfection). Antiviral activity was also decreased when POS was added to Vero-P cells and then removed by washing prior to PeV-A3 infection (Fig. 2B, pretreatment-2). The latter observation suggests that POS exerts its antiviral effect by binding the PeV-A3 capsid directly versus inhibiting any step in the PeV-A3 intracellular replication process.

To confirm that POS binds directly to PeV-A3, we developed an ELISA that utilized a polyclonal antibody (pAb) to capture PeV-A3 on the plate and a monoclonal antibody as a detector. When dilutions of PeV-A3 (US-WI-09) were incubated with POS prior to PeV-A3 MAb addition, a reduced absorbance signal was produced (Fig. 4A and C), indicating that POS binds to the PeV-A3 capsid and prevents binding by the PeV-A3 MAb. For comparison, we tested the antifungal azole voriconazole, which is inactive against PeV-A3, under similar conditions. PeV-A3 plus VRC incubation, followed by treatment with PeV-A3 MAb, resulted in a small decrease in absorbance relative to that of the no-VRC control, suggesting only weak VRC binding to PeV-A3 (Fig. 4B and D).

To further demonstrate that POS directly targets PeV-A3, we performed synchronized infection assays that evaluated POS plus PeV-A3 pretreatment, PeV-A3 attachment, and PeV-A3 entry as previously used to study a hepatitis C virus entry inhibitor (18). POS reduced PeV-A3 activity when incubated with PeV-A3 prior to Vero-P infection (free virus particles condition) and also inhibited PeV-A3 attachment to Vero-P cell monolayers. A minimal antiviral effect on preventing PeV-A3 entry into host cells was observed. These results suggest that POS specifically inhibits PeV-A3 by directly binding the PeV-A3 capsid and preventing attachment. A thermostability assay demonstrated that POS protects PeV-A3 from heat inactivation, which also suggests that POS directly interacts with the viral capsid.

We hypothesize that the structurally similar compounds POS and ITC have the same capsid target. Our drug synergy results suggest a similar viral target. The anti-PeV-A3 MAb may target a distinct site or may interact at the POS site. It remains possible that POS acts by multiple mechanisms, as evidenced by drug-resistant PeV-A3 strains acquiring substitutions in multiple regions of the genome (Fig. 7 and Table 1).

We generated drug-resistant PeV-A3 variants that were less sensitive to POS than the PeV-A3 parental strains by over 100-fold. POS exhibited strong cross-resistance (72-fold) to an ITC-resistant variant, supporting the idea that the two structurally similar drugs target a common binding site on the PeV-A3 capsid. Whole-genome sequencing revealed amino acid substitutions across the viral genome. While this list provides potential candidate substitutions underlying the phenotypic changes, we cannot rule out that some are the result of PeV-A3 adaptive changes to cell culture passage or simply neutral amino acid substitutions. We generated an infectious PeV-A3 clone to experimentally confirm the exact amino acid substitutions responsible for the drug-resistant phenotype. The results confirmed that the amino acid substitutions identified in the variant screen were responsible for the resistant phenotype (Fig. 7 and Table 1).

Our previous study indicated POS had antiviral activity against PeV-A3 clinical isolates at concentrations clinically attainable in pediatric patients (17). Other published studies have shown that POS dosing can achieve these clinically relevant concentrations in the central nervous system (38–41).

This study explored the in vitro antiviral mechanism of the FDA-cleared drug posaconazole against PeV-A3 and demonstrated that POS inhibits PeV-A3 infection by directly targeting the capsid and interfering with virus-cell interactions. Future molecular modeling studies may be important to help further identify the precise molecular interactions between PeV-A3 and POS. Little is known about compounds with PeV-A3 antiviral activity or the mechanism of inhibition; thus, POS may become a valuable tool in developing an antiviral therapy for PeV-A3.

MATERIALS AND METHODS

Cells and compounds.

Vero-P cells (African green monkey kidney) were supplied by the Division of Scientific Resources, CDC. Posaconazole (POS) and the related compounds, itraconazole (ITC) and voriconazole (VRC), were purchased from Sigma-Aldrich (St. Louis, MO). Compounds were stored in 500-μl aliquots at −70°C as 10 mM solutions in high-performance liquid chromatography-grade dimethyl sulfoxide (DMSO; Sigma-Aldrich).

Viruses.

PeV-A3 virus isolate US-WI-09 was previously described (42), and the PeV-A1 prototype strain (Harris) was obtained as an NIH Reference Reagent. Drug-resistant variants US-WI-09(POS) and US-WI-09(ITC) were generated from parental strains as described below. All viruses were cultivated in Vero-P cells and titrated in triplicate to standardize the virus concentration in the assays.

Multicycle antiviral activity assay.

Antiviral activity of compounds was assessed in a homogeneous cell-based assay that measured inhibition of viral cytopathic effect (CPE) in Vero-P cells. For the CPE inhibition assay, 50 μl of half-log₁₀ dilutions of drug compound (70 μM to 0.032 μM) was combined with 50 μl of 100 CCID₅₀ (50% cell culture infectious dose) of virus diluted in growth medium (minimal essential medium [MEM] with 2% fetal bovine serum [FBS]). The virus-drug mixtures were incubated for 1 h at 35°C before addition of 100 μl cells (12,500 cells per well) in 96-well, white, flat-bottom microplates (Corning, Inc., Kennebunk, ME). Plates were incubated at 35°C with 5% CO₂ for 12 days, and cell viability was assessed using ATPLite (Perkin Elmer, Waltham, MA) by adding 50 μl of cell lysis buffer and then 50 μl of substrate solution by following the manufacturer’s recommendations. Luminescence was read in a Victor⁵ plate reader (Perkin Elmer), and the 50% effective concentration (EC₅₀) of each compound was calculated by 4-parameter curve fitting with GraphPad Prism (version 5.0.3; GraphPad Software, La Jolla, CA).

Virus titration assay.

The assay for virus titer was a modification of the WHO Global Polio Laboratory Network cell sensitivity assay (21). The dilutions of virus that resulted in at least 80% destruction of the cell monolayer (signal-to-noise ratio, ≥5) were used to calculate the virus titer in the sample (as CCID₅₀), using the Kärber formula (43). Titer values were expressed as log(CCID₅₀).

Time course, time-of-addition, and synchronized infection assays.

In the time course assay, drug was added to the cell monolayer in 96-well plates at 0, +1, +2, +4, or +6 h postinfection at 35°C. PeV-A3 (US-WI-09 strain; 100 CCID₅₀) was added at 0 h, and both virus and drug remained present in the assay at 35°C with 5% CO₂ for 12 days. The time-of-addition assay evaluated the antiviral effect of POS pretreatment with PeV-A3, POS, and PeV-A3 coaddition and POS added postinfection. These experiments used 100 CCID₅₀ of PeV-A3 (US-WI-09 strain) or PeV-A1 (Harris strain) and POS at 0.5 μM. In pretreatment-1, monolayers were incubated with 0.5 μM POS for 2 h, followed by the addition of virus to the cell and drug mixture for 2 h. Residual POS and unbound virus were removed by washing 4× with 300 μl of phosphate-buffered saline (PBS). In the pretreatment-2 experiment, Vero-P cells were pretreated with POS for 2 h, and residual drug was removed by washing 4× with 300 μl PBS before addition of virus. PeV was added and incubated for 2 h, followed by a second wash step (4× with 300 μl PBS) to remove unbound virus. For the coaddition test, Vero-P monolayers were inoculated simultaneously with POS and PeV, incubated for 2 h, and then washed 4× with 300 μl PBS to remove unbound POS and PeV. The postinfection configuration had Vero-P cells infected with PeV for 2 h, followed by removal of unbound virus by washing 4× with 300 μl PBS. POS was added and remained present throughout the incubation.

The synchronized infection assay evaluated the effect of POS on virus prior to inoculation of cells (pretreatment), during virus attachment, and during virus entry (18). For the pretreatment experiment, PeV-A3 (500 CCID₅₀) was incubated with 0.5 μM POS for 2 h at 35°C, and then the drug-virus mixture was diluted 20-fold in growth medium. This dilution produces a viable virus concentration but an ineffective drug concentration. Vero-P monolayers were inoculated with diluted drug-virus mixture. To examine the effect on the viral attachment stage, cells were inoculated with PeV-A3 at 4°C for 2 h. At this temperature, the viral particles are allowed to bind to the cell surface but entry is prevented; thus, inoculation and incubation at 4°C allows for the specific examination of the attachment phase. After washing 4× with 300 μl PBS to remove unbound virus, the cells were incubated at 35°C for 12 days to allow for the subsequent infection to progress. In addition, we examined the effect of POS on virus entry. Cells and virus were allowed to prebind at 4°C for 2 h before removal by washing 4× with 300 μl PBS. Drug was added to the virus-bound cells, temperature was shifted to 35°C, and the mixture was incubated for 12 days to maximize virus entry. The titers for these experiments were determined as previously described.

PeV-A3 ELISA.

High-binding microplates (Immulon 2HB, 96-well polystyrene, flat bottom; Thermo-Fisher) were coated with 0.5 μg/well rabbit anti-PeV-A3 pAb (CDC no. 16-1071) and incubated overnight at 4°C. The following day, the contents were removed and the wells were blocked with 200 μl of blocking buffer (1% bovine serum albumin in PBS, pH 7.2). POS and VRC at 10 μM or serially diluted (0.077 to 10 μM) were incubated individually or combined with 100 CCID₅₀ of PeV-A3 (US-WI-09) in a total volume of 100 μl 0.01 M PBS, pH 7.2, for 2 h at room temperature. The contents were transferred to pAb-coated microplates and incubated for 1 h to allow the virus alone or virus-drug complex to bind to the plate. Unbound drug and virus were removed by washing 4× with 300 μl 0.01 M PBS, pH 7.2, with 0.05% Tween 20 (Sigma-Aldrich) and then incubated for 1 h at room temperature with 200 μl of blocking buffer. The wells were washed 4× with 300 μl PBS, followed by the addition of 50 μl (4.8 μg/well) anti-PeV-A3 mouse MAb (CDC clone AB6-BA9-AH7) diluted in blocking buffer. After incubating for 1 h at room temperature, the wells were washed 4× with 300 μl PBS, followed by the addition of 100 μl (0.032 μg/well) of horseradish peroxidase-conjugated goat anti-mouse IgG(H+L) (KPL, Inc., Gaithersburg, MD) and then by incubation for 1 h at room temperature. The microplates were washed 4× with 300 μl PBS, and anti-PeV-A3 MAb binding was detected by adding 100 μl of 3-3′-5′ tetramethylbenzidine (TMB) SureBlue Reserve peroxidase substrate (KPL, Inc.). The color was allowed to develop for 15 min before stopping the reaction with 100 μl TMB BlueStop solution (KPL, Inc.). Absorbance was determined at 620 nm using a Victor⁵ microplate reader (Perkin Elmer).

Drug synergy assay.

The drug synergy experiments were performed as previously described (44). Briefly, a region of synergy or antagonism was indicated by a peak or depression composed of a cluster of adjacent cells significantly above or below zero. A synergy volume was calculated by adding all of the positive values for each drug combination. Similarly, all of the negative values are added to give an antagonistic volume. These volumes are then statistically evaluated using the 95% confidence level and are expressed in percent square micromolars, used to categorize the degree of synergy. The MacSynergy II program (version 1.0; Ann Arbor, MI) was used to calculate a theoretical additive value for each drug combination based on the values generated by the drugs alone using the Bliss Independence model in a Microsoft Excel-based interface (19).

Selection of POS- and ITC-resistant PeV-A3 variants.

Drug-resistant PeV-A3 variants were isolated by culturing PeV-A3 strain US-WI-09 on confluent Vero-P cells in the presence of gradually increasing concentrations of POS or ITC. Vero-P cells (250,000 cells per well) in 24-well plates were incubated overnight in MEM with 10% FBS at 35°C in a humidified 5% CO₂ incubator. The following day, the cell medium was removed and cells were infected with 100 CCID₅₀ PeV-A3 alone or with drug (0.2, 0.4, 0.6, 1, 2, or 10 μM). A dose escalation was performed with drug at 0.2 μM (passages 1 to 2), 0.4 μM (passages 3 to 4), 0.6 μM (passages 5 to 7), 1 μM (passages 7 to 10), 2 μM (passages 11 to 13), and 10 μM (passage 14). The plates were incubated for 12 days, and the supernatant was collected from wells that exhibited full CPE. Virus was filtered (0.45 μm) and diluted 10-fold prior to infection of subsequent passages. PeV-A3 titers were determined after each passage as described previously for poliovirus (45).

Whole-genome sequencing and protein model of PeV-A3 parental virus and POS- and ITC-resistant variants.

To identify genetic changes that correlate with drug resistance, we sequenced strains US-WI-09(POS) and US-WI-09(ITC). Viruses from Vero-P passages 0 through 14 were sequenced using a Nextera XT next-generation sequencing (NGS) library kit and a 300-cycle (2× 150-bp paired-end) MiSeq reagent kit, v2 (Illumina, San Diego, CA), as described previously (46). At least 500,000 reads were generated for each sample. Reads were trimmed for quality and to remove adaptor sequences using an in-house bioinformatics pipeline (VPipe) (46). To generate full-genome consensus sequences, trimmed reads were mapped to parechovirus A3 reference sequences using Geneious (version 10; Biomatters, Auckland, New Zealand). All PeV-A3 genomes were supported by at least 10× coverage at every base across the entire coding region.

The whole-genome alignment was performed using MAFFT (47) and was inspected for nonsynonymous substitutions using Geneious. Nonsynonymous substitutions between the initial and final passage were visualized using a published X-ray crystallography-based capsid model of parechovirus A3 (PDB accession no. 5APM) (48) with the software PyMOL (Schrödinger, Inc., New York, NY).

Generation of PeV-A3 POS-resistant mutants.

The PeV-A3-R1 prototype clone was manufactured by GenScript (Piscataway, NJ) in the pCC1-4K vector. The viral insert was cloned into pUC by standard methods to conform with our normal laboratory work flow. The clone was designed with 19 bases of the T7 promoter immediately upstream of the 5′ end of the viral sequence and a 30-base poly(A) tail after the 3′ end, followed by an SbfI restriction endonuclease site used for linearizing the plasmid. Single-base substitutions were made by using a QuikChange Lightning site-directed mutagenesis kit (Agilent, Santa Clara, CA). In vitro transcripts of viral RNA made using a MEGAscript kit (Life Technologies, Carlsbad, CA) were transfected into semiconfluent Vero cell monolayers in 12-well plates using a TransIT-mRNA transfection kit (Mirus, Madison, WI). Cytopathic effect was observed for viruses after incubation at 35°C for 10 days when over half the cells in each well were affected. Plates were frozen and thawed two times, and 750 μl from each transfection was transferred to a confluent Vero cell monolayer in a 75-cm² flask containing complete minimal essential medium with 2% FBS. In this passage, >75% cytopathic effect was observed after 6 days of incubation at 35°C when the flasks were frozen and thawed two times. The complete sequences of all viruses were verified by in vitro amplification of two overlapping PCR products for each viral clonal stock and capillary sequence analysis.

Thermostability assay.

To assess the effect of POS on the thermostability of parental and drug-resistant virus, aliquots of concentrated (500 CCID₅₀) PeV-A3 (US-WI-09) and VP1_Y224C mutant, each containing a fixed amount of POS (0.25 μM) or an equivalent volume of medium, were incubated at temperatures ranging from 37°C to 55°C for 5 min and then rapidly cooled to 4°C. Viral infectivity was quantified in Vero-P cells using the virus titration assay as described previously (21).

Data availability.

The genome sequences of the parental strains, resistant variants, and infectious clones were deposited in GenBank under accession numbers MN781627 to MN781636.