Activity of Ribavirin against Hantaan Virus Correlates with Production of Ribavirin-5′-Triphosphate, Not with Inhibition of IMP Dehydrogenase

Abstract

Ribavirin (RBV) is a broad-spectrum antiviral agent that inhibits the production of infectious Hantaan virus (HTNV). Although the mechanism of action of RBV against HTNV is not understood, RBV is metabolized in human cells to both RBV-5′-monophosphate, which inhibits IMP dehydrogenase, resulting in a decrease in intracellular GTP levels, and RBV-5′-triphosphate (RBV-TP), which could selectively interact with the viral RNA polymerase. To elucidate which activity of RBV was most important to its anti-HTNV activity, the mechanism of action of RBV was studied in Vero E6 cells. Incubation with 10 to 40 μg/ml RBV resulted in a small decrease in GTP levels that was not dose dependent. Increasing the RBV concentration from 10 to 40 μg/ml resulted in a decrease in viral RNA (vRNA) levels and an increase in RBV-TP formation. Mycophenolic acid (MPA), an inhibitor of IMP dehydrogenase, also resulted in a decrease in vRNA levels; however, treatment with MPA resulted in a much greater decrease in GTP levels than that seen with RBV. Treatment with both MPA and RBV resulted in increased reduction of vRNA levels but did not result in enhanced depression of GTP levels. Although guanosine prevented the depression in GTP levels caused by RBV, guanosine only partially prevented the effect of RBV on vRNA levels. These results suggest that the inhibition of IMP dehydrogenase by RBV is of secondary importance to the inhibition of vRNA replication by RBV and that the interaction of RBV-TP with the viral polymerase is the primary action of RBV.

Hanta virus infections represent an important and growing source of disease in both developed and developing countries (11). Although no vaccines or antiviral agents are approved by the FDA to treat the disease, ribavirin (RBV) has been shown to have antiviral activity against hantaviral infections in in vitro assays and in the suckling-mouse model (9, 12). There is also evidence that it is effective in people infected with Hantaan virus (HTNV) (8). RBV is a broad-spectrum antiviral agent with activity against both DNA and RNA viruses (2, 21). Although much is known about the metabolism and biochemical effects of RBV in human cells (17), the mechanism of action of RBV against HTNV has not yet been determined. Once transported into human cells, RBV is rapidly converted to RBV 5′-monophosphate (RBV-MP) by adenosine kinase (1, 25), and successive phosphorylation leads to the formation and accumulation of RBV-5′-triphosphate (RBV-TP) (5). RBV-MP is a potent competitive inhibitor of IMP dehydrogenase with respect to its natural substrate, IMP (15, 23), and the inhibition of this enzyme by RBV-MP is believed to be responsible for the toxicity of RBV to human cells. Other activities of RBV that could result in antiviral activity include its ability to (i) interfere with capping of the 5′ end of the mRNA (6), (ii) inhibit the viral polymerase by RBV-TP (3, 4, 18, 24, 26), and (iii) induce error catastrophe (7). These activities of RBV are all associated with the production of RBV-TP in virus-infected cells. Severson et al. (20) have shown that treatment with RBV results in an increase in the mutation frequency in the HTNV genome, which suggests that the direct incorporation of RBV in viral RNAs (vRNAs) by the viral polymerase is responsible for its antiviral activity against HTNV.

To better understand the actions of RBV that are responsible for its anti-HTNV activity, we explored the metabolism and biochemical actions of RBV in Vero E6 cells. Our results indicated that the production of RBV-TP correlated with the effect of RBV on vRNA replication and suggested that the interaction of RBV-TP with the viral RNA-dependent RNA polymerase was primarily responsible for the antiviral activity of RBV, which is consistent with the increase in mutation frequency that was observed by Severson et al. (20).

MATERIALS AND METHODS

Reagents.

[G-³H]RBV was obtained from Moravek Biochemicals (Brea, CA). ATP was obtained from Amersham Pharmacia Biotech (Piscataway, NJ). RBV was obtained from ICN pharmaceuticals (Costa Mesa, CA). RBV-TP was obtained from Jena Bioscience (Jena, Germany). Mycophenolic acid (MPA) and guanosine were obtained from Sigma Chemical Company (St. Louis, MO). Ten micrograms per milliliter of mycophenolic acid, guanosine, and RBV are equivalent to 31, 35, and 41 μM of each compound, respectively.

Determination of the effect of drug treatment on hantaviral replication.

Confluent Vero E6 cells (ATCC CRL 1586) in six-well cell culture plates (confluent 3-day-old cultures) were infected with HTNV (strain 76-118) at a multiplicity infection of 0.1 as described previously (20). After infection for 1 h at 37°C, the medium was removed and replaced with 2 ml of Dulbecco's modified Eagle medium containing 10% fetal bovine serum and various compounds. There was no toxicity to the Vero cells at the concentrations of drugs used in this study. After incubation for 3 days at 37°C, the medium was discarded and the cells were washed with phosphate-buffered saline once. PFU were measured by agarose overlay as described previously (20). To evaluate vRNA replication, total intracellular RNA was isolated from each well with Trizol reagent (Gibco-BRL) as described in the manufacturer's protocol. The purified RNA was suspended in The RNA Storage Solution (Ambion) and stored at −80°C until it was used. The amount of vRNA was measured with a quantitative real-time RT-PCR assay employing the comparative cycle threshold method (Applied Biosystems). The cDNA was synthesized from 1 μg of total cellular RNA by SuperScript III Reverse Transcriptase (Invitrogen, CA) with 10 pmol of vhRT primer (5′-TAGTAGTAGACTCCCTAAAGAGCT-3′) for the S segment of the vRNA and 10 pmol of 18SRT primer (5′-TCTTCTCAGCGCTCCGCCA-3′) for the endogenous control of 18S rRNA. The forward primer and reverse primer for real-time PCR were 5′-GCAGATAGGATTGCAACTGGGAAAA-3′ and 5′-GCATTGACCTCTCTTTCAGATGGT-3′, respectively. The HTNV replicon was measured using the TaqMan probe, which was labeled with either a reporter dye (6-carboxylfluorescein) at the 5′ end or the minor-groove binder at the 3′ end. Quantification of the endogenous control, 18S RNA, was done with a predeveloped 18S RNA endogenous control reagent kit from Applied Biosystems. All quantitative real-time PCRs were done in triplicate and were prepared with TaqMan universal PCR master mix (Applied Biosystems).

Measurement of intracellular nucleotide levels and RBV metabolites.

Confluent Vero E6 cells prepared as described above for the antiviral studies were incubated at 37°C with various agents. At the end of the incubation periods, the cells were washed twice with sterile phosphate-buffered saline and then treated with 0.4 ml of 0.5 M ice-cold perchloric acid as described previously (16). The perchloric extracts were centrifuged at 12,000 × g for 10 min, and the supernatant fluid was removed, neutralized with 4 M of KOH, and buffered with 1 M of potassium phosphate (pH 7.4). The KClO₄ precipitate was removed by centrifugation, and a portion of the supernatant fluid was injected onto a Partisil-10 strong-anion-exchange (SAX) high-performance liquid chromatography (HPLC) column (10 μm; 250 by 4.6 mm; Keystone Scientific Inc., Bellfonte, PA). Elution of the nucleotides was accomplished with a 50-min linear gradient from 5 mM of NH₄H₂PO₄ (pH 2.8) to 750 mM of NH₄H₂PO₄ (pH 3.7) buffer with a flow rate of 2 ml/min. The natural nucleotides (ATP and GTP) were detected by measurement of the UV absorbance at 260 nm, and radioactive metabolites of RBV were detected by counting 1-min fractions. Because of the difficulty in obtaining reliable cell counts from the Vero E6 cell monolayers, we used intracellular ATP levels as a measure of the number of cells that were treated. None of the treatments shown in the current work resulted in changes to ATP levels. Similar results were obtained if the data were not normalized, except that the variability was increased.

RESULTS

Effect of RBV and/or MPA on vRNA replication.

Treatment of HTNV-infected Vero E6 cells with RBV inhibited both the amount of vRNA that was produced and the level of infectious virus (Fig. 1A). Although, MPA was more potent than RBV in its ability to decrease vRNA levels (Fig. 1B), its effect on vRNA levels was limited (all concentrations above 2 μg/ml resulted in approximately 90% inhibition) and the maximal effect was not as great as that seen with RBV (40 μg/ml RBV inhibited the production of vRNA by 99.9%). Treatment with both RBV and MPA resulted in greater decreases in vRNA levels (Fig. 1C). For example, incubation with 0.5 μg/ml MPA plus 10 μg/ml RBV resulted in a 98% decrease in vRNA levels, whereas incubation with either compound alone resulted in only a 70% to 80% decrease in vRNA levels.

FIG. 1.

Inhibition of vRNA replication by RBV and MPA. Vero E6 cells infected with HTNV were treated with various concentrations of RBV or MPA for 3 days, and the amounts of HTNV S-segment vRNA or PFU were determined as described in Materials and Methods. The results are presented as percentages of untreated sample. Each data point represents the mean ± standard deviation from three independent experiments.

The effect of RBV and MPA on intracellular GTP levels.

MPA is a potent inhibitor of IMP dehydrogenase, and its antiviral activity is due to the decrease in GTP levels as a consequence of the inhibition of this enzyme (14). As indicated above, inhibition of IMP dehydrogenase by RBV-MP could also be responsible for the antiviral activity of RBV. Therefore, to assess the inhibition of IMP dehydrogenase activity, we determined the effects of RBV and MPA on intracellular GTP pools (Fig. 2A). Incubation with 10 μg/ml RBV resulted in only a small decline in intracellular GTP levels (40 to 50%) that was not sustained over the 72-h incubation period. The maximal depression of intracellular GTP occurred after 4 h of incubation with RBV, and GTP levels had almost returned to control values by 72 h. In contrast to these results, incubation with 2.5 μg/ml MPA resulted in 90% reduction in intracellular GTP levels that was sustained for the 72-h incubation period. Interestingly, 24 h of incubation with MPA was required to cause maximal depression of GTP levels. Treatment with 10 μg/ml RBV had approximately the same effect on vRNA levels as 1 μg/ml of MPA, but RBV had much less effect on intracellular GTP levels.

FIG. 2.

Effect of RBV or MPA on GTP levels. Acid-soluble extracts of cell pellets from cell cultures treated with RBV and/or MPA were analyzed by SAX HPLC to determine intracellular GTP and ATP concentrations. (A) Vero E6 cells were treated with 10 μg/ml of RBV (41 μM), 2.5 μg/ml of mycophenolic acid (7.8 μM), or no drug for 0, 1, 2, 4, 24, 48, or 72 h. (B) Vero E6 cells were treated with RBV for 4 h or MPA for 24 h. (C) Vero E6 cells were treated with RBV, MPA, or RBV plus MPA for 24 h. Each data point represents the mean ± standard deviation from three measurements. These experiments were repeated twice with similar results.

Increasing the concentration of RBV did not result in increased depression of GTP levels (Fig. 2B). Incubation of cells with 40 μg/ml RBV for 4 h caused a depression of GTP levels that was similar to that seen at 10 μg/ml RBV. In addition, the effect of 40 μg/ml RBV on intracellular GTP levels was also not sustained, and GTP levels had returned to near control levels by 72 h (data not shown). There was only a small decline (less than 5%) in the concentration of RBV in the medium surrounding the cells during the 72-h incubation period (data not shown), which indicated that the lack of sustained effect on GTP pools was not due to the disappearance of RBV from the culture medium.

Since incubation with both MPA and RBV resulted in a greater reduction of vRNA levels than was seen with either agent alone (Fig. 1C), it was of interest to determine whether incubation with both MPA and RBV would result in enhanced effects on intracellular GTP. As was seen before, treatment with RBV alone resulted in a decrease in GTP levels of approximately 50% (Fig. 2C), and treatment with only MPA resulted in 80 or 90% decline in GTP levels (Fig. 2C). However, intracellular GTP levels in cells treated with both RBV and MPA were not significantly different than those seen in cells treated with MPA alone (Fig. 2C), even though incubation with both RBV and MPA resulted in greater decreases in vRNA levels than either agent alone (Fig. 1C).

The small effect of RBV on intracellular GTP levels in Vero E6 cells and the lack of correlation between RBV's effect on intracellular vRNA and GTP levels indicated that the antiviral activity of RBV was not due to its ability to decrease intracellular GTP levels.

Effects of guanosine on intracellular vRNA and GTP levels.

Treatment with guanosine can prevent the activity of an IMP dehydrogenase inhibitor by replenishing the GTP pool. Guanine, which is generated from guanosine by purine nucleoside phosphorylase, is converted to GMP by hypoxanthine/guanine phosphoribosyltransferase and thus bypasses the block of IMP dehydrogenase by RBV-MP or MPA (17). Therefore, to determine the role of IMP dehydrogenase inhibition in the activity of RBV, we evaluated the effect of guanosine on vRNA levels in cells treated with RBV and MPA. The addition of 10 μg/ml guanosine totally prevented the effect of MPA on vRNA levels but only partially prevented the effect of RBV on vRNA levels (Fig. 3). Furthermore, the addition of 10 μg/ml of guanosine to cells treated with either 40 μg/ml RBV or 1 μg/ml MPA totally prevented the depression of intracellular GTP levels that was caused by these two agents (Fig. 3). Treatment of Vero E6 cells with 10 μg/ml guanosine alone resulted in only a small increase in intracellular GTP levels (145% of control). However, in combination with 40 μg/ml RBV, the GTP levels were approximately three times those seen in untreated cells. Even though GTP levels were three times normal, treatment with RBV still resulted in a decrease in vRNA levels of 80% (Fig. 3). These results supported the above-mentioned conclusion that an action other than inhibition of IMP dehydrogenase activity was responsible for the antiviral activity of RBV.

FIG. 3.

Effects of guanosine on intracellular vRNA and GTP levels. Vero E6 cells were treated with 10 μg/ml guanosine (35 μM), 40 μg/ml RBV (164 μM), 40 μg/ml RBV plus 10 μg/ml guanosine, 1 μg/ml MPA (3.1 μM), or 1 μg/ml MPA plus 10 μg/ml guanosine. After 24 h of treatment, acid-soluble extracts of the cell pellets were analyzed by SAX HPLC to determine the amounts of GTP and ATP in each sample. The effect of each treatment on HTNV S-segment vRNA was determined after 3 days of treatment. Each number represents the mean plus standard deviation from three measurements.

Metabolism of RBV in Vero E6 cells.

Because the conversion of RBV to phosphorylated metabolites is an important aspect of its mechanism of action, we evaluated the amounts of RBV-MP and RBV-TP that were formed in cells treated with RBV and correlated its metabolism with its effect on vRNA levels. The primary intracellular metabolite of RBV in Vero E6 cells was RBV-TP (Fig. 4). Its concentration quickly rose to approximately 1% of the ATP concentration in the cell. Since the intracellular concentration of ATP is about 3 mM (10), these results indicated that RBV-TP accumulated to approximately 30 μM in cells that were treated with 10 μg/ml RBV. The RBV-TP concentration peaked after 4 h of incubation and then declined by about 50% over the next 68 h. The concentration of RBV-MP was approximately 10% of that of RBV-TP and remained steady over the 72-h incubation period. Increasing the extracellular concentration of RBV to 40 μg/ml resulted in approximately fourfold-greater intracellular concentrations of both RBV-MP and RBV-TP (Table 1). Since increasing the concentration of RBV from 10 to 40 μg/ml resulted in a greater reduction of vRNA levels (Fig. 1B), these results indicated that the metabolism of RBV was positively correlated with the effect of RBV on vRNA synthesis. Neither MPA nor guanosine affected the metabolism of RBV (Table 1), which indicated that the effects of these agents on vRNA levels of cells treated with RBV (Fig. 1B and Fig. 3) was not due to either the enhancement or inhibition (by MPA or guanosine, respectively) of RBV metabolism.

FIG. 4.

Metabolism of RBV in Vero cells. Vero E6 cells were treated with 10 μg/ml [³H]RBV (41 μM) for 0, 1, 2, 4, 24, 48, or 72 h. Acid-soluble extracts of the cell pellets were analyzed by SAX HPLC, and the intracellular metabolites were determined. The experiment was repeated two times with similar results. The error bars represent standard deviations.

TABLE 1.

Effects of guanosine and MPA on the metabolism of RBV^a

Expt	RBV (μg/ml)b	MPA (μg/ml)	Guanosine (μg/ml)	RBV-MP (pmol/nmol ATP)	RBV-TP (pmol/nmol ATP)
1	10	0	0	1.1 ± 0.1	6.5 ± 0.6
	10	1	0	1.3 ± 0.1	7.8 ± 0.5
	10	2	0	1.3 ± 0.2	6.8 ± 0.2
	10	5	0	1.2 ± 0.0	7.0 ± 0.5
	40	0	0	3.2 ± 0.3	22 ± 1.6
	40	1	0	4.2 ± 0.4	26 ± 1.5
	40	2	0	3.8 ± 0.4	27 ± 1.1
	40	5	0	3.9 ± 0.3	26 ± 0.7
2	10	0	0	1.3 ± 0.1	11 ± 1.3
	10	0	10	1.9 ± 0.2	14 ± 3.3
	40	0	0	4.3 ± 0.2	35 ± 6.9
	40	0	10	8.6 ± 2.8	65 ± 20

^aIn experiment 1, Vero E6 cells were treated with RBV with and without MPA, and in experiment 2, Vero E6 cells were treated with RBV with and without guanosine. After 24 h of incubation, acid-soluble extracts of the cell pellets were analyzed by SAX HPLC, and the intracellular metabolites (RBV-MP and RBV-TP) were determined. Each number represents the mean ± standard deviation from three measurements. The experiment was repeated two times with similar results.

^bTen micrograms per milliliter of mycophenolic acid, guanosine, and RBV are equivalent to 31, 35, and 41 μM of each compound, respectively.

DISCUSSION

In the current work, we evaluated the effect of RBV on intracellular GTP levels and the metabolism of RBV to RBV-TP in an effort to understand which actions of RBV are important to its antiviral effect against HTNV. We show in many examples that the effect of RBV on intracellular GTP did not correlate with its ability to decrease vRNA levels, which indicated that the inhibition of IMP dehydrogenase by RBV-MP was not the primary mechanism by which RBV inhibited the replication of HTNV in Vero E6 cells. In contrast to RBV, there was good correlation between the effect of MPA on vRNA levels and depression of GTP levels. Treatment with MPA depressed intracellular GTP levels by 90% and also resulted in approximately a 90% decrease in viral RNA. The results with MPA are consistent with the understanding of MPA's mechanism of action, i.e., its only known action is the inhibition of IMP dehydrogenase (14).

The antiviral activity of RBV against some viruses is prevented by treatment with guanosine (13, 19, 22, 27). However, with HTNV, the addition of guanosine to the viral cultures did not completely reverse the antiviral activity of RBV, even though addition of guanosine to the cell cultures resulted in intracellular GTP levels that were 300% of control values. Although RBV was still active in the presence of guanosine, its effect on vRNA levels was much less than that seen in cultures treated with RBV alone, which is evidence that intracellular GTP levels do play some role in the antiviral activity of RBV. Based on the lack of correlation observed between GTP levels and the effect on vRNA levels described above, it is likely that the effect of guanosine on RBV activity was due to the ability of the increased intracellular levels of GTP to compete with RBV-TP as a substrate for the viral RNA polymerase, thus reducing the effect of RBV on vRNA replication.

The results with MPA indicated that decreasing the intracellular GTP level was sufficient to inhibit vRNA levels. Although, our results with RBV indicated that inhibition of IMP dehydrogenase was not the primary mechanism of RBV activity, they did indicate that GTP levels could modulate the antiviral activity of RBV-TP. Therefore, since treatment with RBV results in only a small decrease in intracellular GTP pools, it would only self-potentiate its antiviral activity by a small amount. Since (i) MPA reduced GTP levels much more than RBV, (ii) the toxicity of RBV to human cells is primarily due to its inhibition of IMP dehydrogenase, and (iii) the combination of MPA plus RBV did not result in greater decline in intracellular GTP levels than that seen with MPA alone, it is possible that a combination of MPA with RBV in patients infected with HTNV would result in greater antiviral activity without enhanced toxicity.

In contrast to the effect of RBV on GTP levels, the decline in vRNA levels caused by treatment with RBV correlated with the production of RBV-TP in Vero E6 cells, which is good evidence in support of the hypothesis that the interaction of RBV-TP with the viral RNA polymerase was primarily responsible for the antiviral activity of RBV. A nucleotide analog can interact with a polymerase in one of three ways: (i) it could compete with the natural nucleotides without being a substrate for the enzyme, (ii) it could be used as an alternative substrate for the polymerase and cause chain termination, or (iii) it could be used as an alternative substrate for the polymerase without causing chain termination. Recently, our laboratory has shown that RBV induced error-prone replication in HTNV, resulting in a reduction in viral titer (20). Because RNA-dependent RNA polymerases of RNA viruses do not have proofreading exonucleases, RNA viruses may exist as a quasispecies (7). Hence, the virus functions on the edge of mutation crisis, and even a slight increase in the mutation frequency could cause a dramatic reduction in virus viability. Our data suggested that ribavirin is incorporated into the viral RNA in place of GTP and then causes mutations by base pairing with both UTP and CTP during subsequent rounds of replication. It is also possible that RBV could compete with ATP for incorporation into the viral RNA, but at present we have no evidence for this possibility. Further studies are needed to fully characterize the interaction of RBV-TP with the viral RNA polymerase. An understanding of the mechanism of action of RBV against HTNV is important to optimize its use in the treatment of people infected with HTNV and to provide strategies to aid the design and development of better antiviral drugs against this virus.