INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The antiviral drug standard for the treatment of herpes simplex virus (HSV) infections including herpes labialis and genital herpes for almost 2 decades has been acyclovir (ACV) [9-(2-hydroxyethoxymethyl)guanine]. However, with the more recent introduction of penciclovir (PCV) (BRL 39123) [9-(4-hydroxy-3-hydroxymethylbut-1-yl)guanine] and its oral prodrug, famciclovir, the usage of antivirals alternative to ACV for the management of herpesvirus infections has also increased. Identical activation pathways and similar modes of action suggest that the mechanisms of HSV resistance to PCV and ACV are likely to be analogous (2, 40). An assumption that the frequency with which resistance in HSV arises is identical for PCV and ACV can be based on the biochemical similarities of the two compounds and the cross-resistance of thymidine kinase (TK)-negative mutants; however, direct genetic evidence is not available.

A low level of replication errors is typically associated with DNA synthesis (10, 33). Resistance to PCV or ACV can arise by a single base mutation in the DNA encoding the HSV TK protein which activates the antiviral agent (6, 23, 29). These spontaneous mutations occur during DNA replication and are independent of the presence of antiviral drug (16). These errors, or random mutations, provide genetic diversity to facilitate the adaptation and evolution of an organism (15). Data from a study of the molecular evolution of HSV type 1 (HSV-1) show that its evolution is slow; the mutation rate was estimated to be 3.5 × 10⁸ substitutions per site per year (36). Mispaired deoxyribonucleoside triphosphates are often removed by the HSV polymerase (Pol) through its associated 3'-5' exonuclease activity (24-26), a common property of DNA Pols (10, 14, 33). Nonetheless, the HSV DNA Pol is not completely error proof, and mutations occur constantly throughout the viral DNA replication cycle. However, most changes remain unnoticed because they are lethal or silent or affect a gene product without a discernible phenotype.

DNA Pols with 3'-5' exonuclease deficiencies exhibit elevated spontaneous mutation frequencies, thereby verifying that these synthesis and repair activities can regulate the production of mutations in vivo (10, 13, 27, 28, 38). When such errors occur within a gene whose activity is easily detected, the frequency of the error itself can be measured (17), and such errors are commonly found within the open reading frame of nonessential viral proteins, such as the HSV TK (31). Resistance to an antiviral agent such as PCV or ACV, for example, can serve as a mutation frequency marker since resistant variants can readily be detected in a mixture of sensitive and resistant viruses. For example, a modified drug susceptibility assay was used to screen for HSV variants able to propagate in the presence of antiviral agent and resulted in the identification of six antimutator variants of HSV (17).

HSV mutants resistant to PCV or ACV carry mutations within either the viral TK or DNA Pol open reading frames. The TK gene is not essential for virus replication in cell culture (8), although in vivo analyses implicate involvement in HSV virulence, pathogenicity, and reactivation from latency (4, 11, 21, 39). Even with an in vivo role for TK, approximately 95% of clinical HSV isolates resistant to ACV are TK mutants rather than Pol mutants (3, 31), although the prevalence of resistant isolates due to double mutants cannot be assessed. Hence, mutations in the viral DNA Pol gene, which encodes a polypeptide essential for virus viability, appear to be less favored than mutations in the TK coding sequence.

This natural phenomenon of spontaneous mutation, which occurs in the absence of drug selection, results in the accumulation of approximately six to eight TK-deficient variants per 10⁴ plaque-forming viruses in virus populations that have never been exposed to selective pressure (7, 17, 19). Although this error frequency was determined by genetic analyses with laboratory-adapted HSV-1 strains, it has been extrapolated to explain the natural heterogeneous occurrence of both ACV-resistant and ACV-sensitive viruses within all clinical HSV isolates (34). Parris and Harrington confirmed that HSV variants resistant to relatively high ACV concentrations were present in populations of uncloned, low-passage clinical isolates (30).

In this report, we examined the spontaneous DNA replication-associated error rate of both HSV-1 and HSV-2 strains in the human cell line MRC-5. Our work provides the first genetic evidence that the frequencies with which resistance to PCV and that to ACV arise in HSV are identical. Additionally, the mutation frequency for the HSV-1 laboratory strain SC16 in these studies is consistent with that previously reported for other type 1 laboratory strains. Surprisingly, we discovered that the naturally occurring error frequency associated with laboratory HSV-2 strains is greater than that for HSV-1 SC16 by 9- to 16-fold. Since the spontaneous mutation frequency may play a significant role in the process of drug resistance in vivo, we also examined the mutation frequency of clinical isolates of HSV-1 and HSV-2 and confirmed a higher frequency in HSV-2 isolates, averaging 30-fold that of HSV-1 isolates. It remains to be determined whether this difference results in the in vivo selection of resistant HSV-2 more readily than resistant HSV-1 during drug therapy.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell lines and virus strains. The MRC-5 diploid human embryonic lung cell line was obtained from the American Type Culture Collection (no. 177 CCL) and grown in Dulbecco's modified Eagle's medium supplemented with 10% calf serum and incubated at 37°C with 5% CO₂. TK-negative (TK) human osteosarcoma (143) cells were propagated in the medium described above.

Compounds. PCV (BRL 39123) was synthesized at SmithKline Beecham Pharmaceuticals. ACV used in these studies was from Sigma Chemical Company. For cell culture assays, 10-mg/ml stock solutions were prepared in dimethyl sulfoxide and stored at 20°C. Working dilutions were prepared in assay medium immediately before use as described below.

Preparation of virus stocks. Virus stocks were prepared by inoculating MRC-5 cells with progeny from a single plaque at a multiplicity of infection of 0.01 PFU per MRC-5 cell. Virus stocks were harvested in culture medium, sonicated, clarified by centrifugation, and stored at 80°C. MRC-5 cells (3.5 × 10⁵ cells/well) were plated into 12-well dishes and grown overnight. Duplicate plates of cells were inoculated with 5 to 15 PFU of HSV-1/well or 5 to 18 PFU of HSV-2/well in 500 µl of serum-free medium at 37°C for 1 h. For each virus sample, five or six replicate wells were evaluated in each of two independent experiments, except for HSV-2 SB5 and HSV-2 6757, for which a total of 34 and 20 replicate wells, respectively, were tested from four independent experiments. Following adsorption, the inoculum was removed from the wells, and one set of plates for each virus received 2 ml (per well) of an overlay containing Dulbecco's modified Eagle's medium, 5% calf serum, and 0.4% SeaPlaque agarose (FMC BioProducts). The remaining dishes were supplemented with 1 ml of liquid medium per well and were incubated for 2 to 3 days at 37°C until full cytopathic effect appeared. The dishes with the agarose overlay were fixed by overlaying them with 10% formaldehyde for 1 h at room temperature, and the agarose plug was then removed. The monolayers were stained with crystal violet (0.5% [wt/vol] in 70% methanol), and plaques were counted to ensure that a low initial inoculum was used. Virus was harvested from the dishes containing the liquid overlay by scraping the infected cells into the medium and freezing the cell suspension at 80°C. Virus stocks were thawed and sonicated, and cell-free virus supernatants were titrated. Between 10 and 34 replicates were prepared for each virus specimen.

PRA. PRAs were performed with MRC-5 cells to determine 50% inhibitory concentrations (IC₅₀s) of PCV or ACV for the virus preparations. Cell monolayers (12-well dishes) were infected with approximately 100 PFU, and following adsorption for 1 h, the inoculum was removed and 2 ml of an agarose overlay containing 5% (vol/vol) heat-inactivated fetal calf serum and the appropriate concentration of antiviral was added to each well. The compounds were tested over a fourfold dilution range to give concentrations from 0.02 to 25 µg/ml. The cells were incubated for approximately 40 h and then fixed and stained as described above. Plaques were counted, and IC₅₀s were calculated using the Kärber method (22).

Plating efficiency. The plating efficiency of the virus preparations was determined according to the method developed by Hall et al. with minor modifications (17). Briefly, six serial 10-fold dilutions of virus were inoculated onto MRC-5 cells in the absence of antiviral drug or in the presence of 3 µg of either PCV or ACV per ml for type 1 strains or 8 µg of either drug per ml for type 2 strains. These concentrations of PCV and ACV were chosen because they are 10 times higher than the average IC₅₀ for HSV-1 and HSV-2, respectively. The mutation frequency was calculated as follows: mutation frequency = (virus titer in the presence of drug)/(virus titer in the absence of drug). The frequency is often expressed as a plating efficiency, or percentage of viruses that are resistant, which is calculated as percent resistant = (mutation frequency) × 100.

TK assay. Viral TK activity was determined by a modification of the method previously described (4). Human 143 TK-negative cells seeded in duplicate 100-mm-diameter dishes were infected with a multiplicity of infection of 5 PFU/cell in 4.0 ml of serum-free medium. Parallel cell monolayers were mock infected. At 1 h postinfection, monolayers were rinsed with phosphate-buffered saline and fresh medium was added for 8 h. Infected cells were then rinsed with phosphate-buffered saline, scraped, and centrifuged for 10 min at 1,000 × g (4°C), and cell pellets were frozen at 80°C. Thawed pellets were resuspended in 300 µl of 10 mM sodium phosphate buffer (pH 6.0)-5 mM 2-mercaptoethanol-10% glycerol-50 µM thymidine. Extracts were sonicated on ice and centrifuged to remove cellular debris. This extract (9 µl) was added to a mixture to yield final concentrations of 100 mM sodium phosphate (pH 6.0), 10 mM ATP, 10 mM magnesium acetate, 6 µCi of [³H]thymidine (11 Ci/mmol; NEN Research Products), 50 µM TTP, 25 mM NaI, 0.67 mM dithiothreitol, and 10 µg of bovine serum albumin per ml in a final volume of 30 µl. Reaction mixtures were incubated at 30°C. At various times ranging from 0 to 180 min after addition of the cell extract, 5-µl aliquots were removed, added to 20 µl of 1 mM thymidine, and boiled for 2 min. Samples were mixed briefly and spotted onto Whatman DE81 circle filters. After drying, the filters were washed three times with 4 mM ammonium formate-10 µM thymidine, once with distilled water, and twice with ethanol. Dry filters were placed into scintillation vials with Betafluor and counted. Values from duplicate samples were averaged. Radioactivity from the mock-infected control was processed in parallel and used to subtract background. Data points from the linear portion of thymidine phosphorylation kinetics were used. Activities of the test panel of viruses were normalized to their counterpart wild-type virus, either HSV-1 SC16 or HSV-2 SB5, which was set at 100%. The limit of detection was estimated to be 0.3%, consistent with previous reports (4).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of parental viruses. To study spontaneous mutation in the HSV genome, we examined the occurrence of mutations which inactivate the viral TK gene. The production of PCV-resistant or ACV-resistant mutants was utilized as a measure of TK mutation, since both PCV and ACV require viral TK function to initiate phosphorylation to nucleoside triphosphates in order to generate the triphosphate forms which inhibit viral DNA synthesis (1). Three laboratory HSV strains (HSV-1, SC16; HSV-2, SB5 and 333) as well as eight first-passage clinical isolates (HSV-1, 484, 1123, 1761, and 424; HSV-2, 6757, 6653, 64C, and 83D) were used in this study. All parental samples were susceptible to PCV, ACV, and cidofovir (HPMPC), a viral DNA Pol inhibitor (18), which, unlike PCV and ACV, is independent of HSV TK activity for its function (data not shown).

Characterization of progeny virus stocks. Progeny virus stocks, prepared by low-multiplicity-of-infection with the drug-sensitive parental viruses, were evaluated by the PRA. All virus progeny stocks remained sensitive to PCV and ACV (Table 1) as well as HPMPC (see partial data in Table 3). It was important to demonstrate that these working stocks were susceptible to PCV and to ACV in order to verify that a chance resistant mutant from the virus preparation was not inadvertently used to initiate the infection and subsequently propagated. Notably, differential susceptibilities of HSV strains to PCV and ACV, as determined by PRA with MRC-5 cells, have been reported previously and are not unexpected (1). Furthermore, despite similar mechanisms of action for PCV and ACV, large host cell line-dependent variations in relative antiviral potency for PCV and ACV measured in vitro by PRA have been reported (2). Therefore, the diploid, limited-passage human fibroblast line, MRC-5, which provides an acceptable degree of similarity between PCV and ACV IC₅₀s and thereby allows the comparison of PCV and ACV spontaneous mutation frequencies at similar drug concentrations, was chosen for this study.

Mutation frequency in laboratory isolates. The proportions of virus resistant to PCV or ACV, which arose during amplification of stocks in MRC-5 cells in the absence of drug selection, were measured by the method described by Hall et al. (17). The method relies upon plaquing virus in the presence of high concentrations of an antiviral. The concentrations used in this study were sufficiently high to ensure that only preexisting resistant mutants formed plaques, since 3 and 8 µg/ml are in the linear, plateau portion of the PCV dose-response curve for HSV-1 and HSV-2, respectively (data not shown). TK activity was measured on a random sampling of HSV-1 SC16 and HSV-2 SB5 viruses grown in the presence of 3 or 8 µg/ml, respectively, to verify that the mutant resistant phenotype was expressed upon amplification in these antiviral concentrations (data not shown).

Mutation frequency in clinical isolates. Four HSV-1 clinical isolates generated either a similar (HSV-1 1761 and HSV-1 424) or a 10-fold-lower proportion of mutant viruses (HSV-1 484 and HSV-1 1123) compared with the laboratory strain HSV-1 SC16 (Table 2). Overall, the percentage of resistant virus detected in the plating efficiency assay using 3 µg of PCV or ACV per ml ranged from 0.003 to 0.05% PCV resistant or 0.003 to 0.07% ACV resistant. Moreover, all four clinical HSV-2 strains consistently yielded higher mutation frequencies compared with the clinical or laboratory HSV-1 strains (Table 2) even though higher concentrations of PCV and ACV were used for HSV-2 strains (8 µg/ml). Two HSV-2 isolates (HSV-2 6652 and HSV-2 83D) contained between 0.6 and 0.9% PCV-resistant virus or 0.5 and 0.8% ACV-resistant virus, values comparable to the laboratory isolates HSV-2 SB5 and HSV-2 333. One clinical isolate, HSV-2 64C, generated twofold-more resistant virus after amplification in vitro, relative to the HSV-2 laboratory strains.

Non-TK-dependent mutation frequency. Mutation frequencies were also assessed using a non-TK-dependent inhibitor of the viral DNA Pol, HPMPC, in order to discount the possibility that the 30-fold differential in mutation frequencies identified between HSV-1 and HSV-2 and the apparent high error rate associated with virus HSV-2 6757 were both due to a TK-dependent phenotype. Differences in spontaneous mutation frequency between HSV-1 SC16 and HSV-2 strains were observed with HPMPC in the plating efficiency assay (Table 3), similar to those for PCV and ACV. In this instance, the proportions of HPMPC-resistant viruses detected for HSV-2 SB5 and HSV-2 333 laboratory strains relative to those for HSV-1 SC16 were 24- and 62-fold higher, respectively (Table 3). Indeed, this difference was confirmed with four clinical HSV-2 isolates, ranging from 32- to 85-fold higher than HSV-1 in proportion of resistant virus. Thus, the relative infidelity of HSV-2 replication compared with that of HSV-1 was confirmed by scoring for mutations in a gene other than TK, through resistance to HPMPC.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This work provides the first genetic evidence that the frequencies of naturally occurring, or spontaneous, HSV resistance to PCV and of that to ACV are identical. Among strains of one virus type, the percentages of resistant virus generated after amplification in human MRC-5 cells are remarkably similar (excluding HSV-2 6757). This is not altogether surprising, since both PCV and ACV interact with the viral TK and DNA Pol in order to inhibit virus replication, although subtle differences in the affinity between the PCV- and ACV-TK or Pol interactions exist (2, 9). These results are consistent with data indicating that ACV-resistant clinical HSV isolates, and laboratory-selected resistant variants, are predominantly TK deficient (31) and therefore are usually cross-resistant to PCV (2).

The plating efficiency assay used here to monitor the spontaneous mutation frequency for the HSV-1 laboratory strain SC16 yielded data consistent with previous reports for two other HSV-1 laboratory strains (7, 17). The frequency of PCV- and ACV-resistant variants of HSV-1 SC16 in natural populations (virus stocks grown in cell culture following infection at a low multiplicity) is between five and seven in 10⁴ PFU in MRC-5 cells, equivalent to mutation error rates of HSV-1 laboratory strains (KOS and Cl101) ascertained in Vero cells. Thus, the DNA replication-associated error rates are similar in both Vero and MRC-5 cells for all three wild-type HSV-1 laboratory strains. The choice of MRC-5 diploid, limited-passage fibroblast cells provided an acceptable degree of similarity between PCV and ACV IC₅₀s and therefore was ideal for this study.

Interestingly, it was noted that the spontaneous mutations in HSV-2 laboratory strains (SB5 and 333) accumulate at approximately a 9- to 16-fold-greater frequency than do errors in HSV-1 SC16. This higher frequency of mutations in HSV-2 strains was confirmed and was even more apparent for a set of three clinical isolates, averaging 30-fold.

The difference in error rates between HSV-1 and HSV-2 reported here would indicate the potential for a more rapid evolution of drug resistance in HSV-2 than in HSV-1 under selection and may help to explain why clinically significant resistance is commonly associated with HSV-2 infection in severely immunocompromised patients (35). A higher level of spontaneous mutations with HSV-2 may contribute a selective advantage to the virus during therapy by providing a greater genetic diversity allowing for efficient propagation of the virus at particular sites of infection. Since mutations within the viral DNA Pol gene have been previously shown to affect spontaneous viral mutation rates (16), the type-specific difference may be due to an inherent virus type-specific property of the DNA Pols. Alternatively, other viral proteins such as the TK, dUTPase, and uracil-DNA glycosylase may directly or indirectly contribute to modulate the inherent mutation frequency of HSV (32).

Progeny virus from HSV-2 6757 contained high levels of resistant virus ranging from 21 to 35% resistant variants within the mixture. However, the drug susceptibility of this mixture was verified by the PRA, which yielded low IC₅₀s for PCV (0.82 µg/ml) and ACV (0.49 µg/ml). Modified nucleotide selection during polymerization or impaired 3'-5' exonuclease activity of the viral DNA Pol as described by Hwang et al. (20) may account for the extremely high mutation frequency of HSV-2 6757. Surprisingly, amino acid changes within the three highly conserved exonuclease motifs (exonucleases I, II, and III) of Pol were not present in HSV-2 6757, although mutations within this coding region were identified (data not shown). It is unclear whether the changes identified within the Pol or other alterations within the HSV-2 6757 genome account for the high error rate, and experiments to clarify the mechanism are in progress.

Plating efficiency assays were also performed with HPMPC, a nucleotide analog inhibitor of the viral DNA Pol which does not require the viral TK for activation and therefore is distinct from PCV and ACV. Selection of mutants with 8 µg of HPMPC per ml (10 times above the wild-type control strain IC₅₀ for SC16 and SB5) revealed virus type-specific differences in mutation frequencies, as with PCV and ACV. Five of the six HSV-2 strains exhibited approximately a 20- to 80-fold-higher spontaneous mutation rate to HPMPC compared with HSV-1 SC16. Therefore, detection of spontaneous mutations in HSV is not unique to the antiviral agents PCV and ACV, which require activation by the viral TK, but has now been extended to an HSV DNA Pol inhibitor, HPMPC. Moreover, the atypical HSV-2 clinical isolate, 6757, consistently possessed a high error rate (300-, 700-, or 1,250-fold above that of SC16) regardless of the antiviral agent (PCV, ACV, or HPMPC) used in the screen.

The PRA measures the overall sensitivity of a virus population and equates that determination to an IC₅₀, whereas the plating efficiency assay measures a distinct parameter, the percentage of resistant virus within a mixed population. Although plating efficiency analysis is sufficiently powerful to detect the presence of low levels of resistant HSV in a mixed virus population, to date there has been no reported evaluation of the correlation between clinical outcome, or treatment resistance in vivo, and the percent resistant virus. In contrast, such a correlation has been established for the PRA IC₅₀ (35). However, the plating efficiency assay may serve as a useful adjunct to the PRA. Interestingly, although plating efficiencies for HSV-2 strains were 9- to 700-fold higher than those for HSV-1 SC16, there was at most only a threefold-higher IC₅₀ by PRA for HSV-2 viruses. Previously reported data on the heterogeneity of HSV clinical strains are consistent with our results (30).

The prevalence of ACV-resistant HSV isolates from immunocompetent patients has remained relatively unchanged over many years of antiviral use regardless of the introduction of long-term suppressive therapy for patients with recurrent genital herpes (12, 37). These experiences, together with the lower mutation frequency associated with HSV-1 strains than with HSV-2 strains, suggest that the prevalence of antiviral resistance in infections caused by HSV-1 may be less than that suggested by HSV resistance prevalence studies predominantly based on diseases typical of HSV-2 infection (3, 5).