INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Hepatitis A virus (HAV) is a picornavirus with a ~7.5-kb positive-strand RNA genome and is the sole member of the Hepatovirus genus (15). The clinical manifestations of HAV infection in humans can vary greatly, ranging from asymptomatic infection, commonly seen in young children, to fulminant hepatitis, which in some cases can result in death (39). HAV is transmitted primarily by the fecal-oral route, and epidemics are common in regions where sanitation conditions are poor.

The single long open reading frame of the HAV genome encodes a large polyprotein which is cleaved by the viral protease to produce three to four structural proteins (encoded by the P1 region) and seven nonstructural proteins (encoded by the P2 and P3 regions). Certain mutations in the 2B and 2C genes of the P2 region increase replication of the HM-175 strain in cell culture (12, 44). Although the function of either protein has not been definitively demonstrated, the presence of a nucleoside triphosphate (NTP)-binding motif within the 2C protein has led to speculation that it may function as a helicase.

HAVs isolated to date have been divided into seven genotypes based on the criterion of 15 to 20% nucleotide sequence diversity over a 168-base segment at the VP1/2A junction (37). All HAV strains isolated from humans are represented in genotypes I, II, III, and VII. The majority (80%) of the human isolates of HAV are categorized as genotype I, of which HM-175 is a prototype strain. Genotypes IV, V, and VI are each represented by a single isolate of simian origin (CY-145, AGM-27, and JM-55/CY-55, respectively) (37). Despite the genetic variability among genotypes, there is evidence for only a single HAV serotype (23, 14).

There is considerable evidence that human strains of HAV can infect nonhuman primates (9, 10, 22, 28). However, whether there are HAV strains that have monkeys as their primary host has not yet been clearly established. HAV or a hepatitis A-like virus is endemic among Malaysian cynomolgus monkeys (3). The viruses infecting these monkeys most likely represent true simian strains since the animals were captured in remote regions of Malaysia and were unlikely to have been infected through contact with humans. Similarly, baboons captured near a sparsely inhabited village in South Africa were positive for anti-HAV antibodies of the immunoglobulin G (IgG) class when they were first tested 2 weeks after capture, again suggesting that these animals had been infected by HAV in nature (41).

The single example of the AGM-27 strain of HAV was isolated from a naturally infected, clinically ill African green monkey (Cercopithecus aethiops) during the acute phase of infection (2). The nucleotide sequence differs from that of the prototype human strain, HM-175 (19), by 17%. This corresponds to 7% divergence in amino acid sequence, with the nonstructural proteins being more highly divergent (9 and 8% for P2 and P3 proteins, respectively) than the structural proteins (3%). The significant divergence of AGM-27 from all other HAV strains (37) and the failure to isolate viruses of the same genotype from humans suggest that AGM-27 may represent a true simian strain of HAV.

There are distinct differences in the biological properties of HM-175 and AGM-27. Wild-type AGM-27 grows in primate cell cultures significantly better than does wild-type HM-175 but not as well as HAV/7, a cell culture-adapted variant of HM-175 (43). Experimental infections of primates have demonstrated that both wild-type HM-175 and AGM-27 are virulent in tamarins (Saguinus mystax) but differ markedly in virulence for chimpanzees (Pan troglodytes). Wild-type HM-175 is virulent in chimpanzees and causes acute hepatitis, whereas AGM-27 is highly attenuated in chimpanzees and infection does not cause hepatitis but does confer protective immunity to chimpanzees subsequently challenged with virulent HM-175 virus (14).

Identifying the genetic parameters that are responsible for HAV pathogenicity is important for understanding the mechanism(s) of HAV pathogenesis and is a critical step toward development of a live attenuated vaccine for HAV. Several inactivated vaccines which effectively protect individuals against HAV infection are currently available (26). However, an effort to generate a live attenuated vaccine is justified because the inactivated vaccines have the limitation that multiple doses are required for effective immunization and it is not known yet if protection conferred by the inactivated vaccines is comparable to that following natural infection. A live vaccine could have the advantage of inducing lifelong immunity following administration of only a single dose. Such a vaccine may also be less expensive to administer, making it more available for use, especially in developing countries. The existence of only one serotype for HAV greatly enhances the prospect of using a single vaccine to protect individuals against infection with all HAV strains.

Adaptation of HAV for efficient replication in cell culture has resulted in attenuation of the virus for primates (5, 7, 21) and humans (40). The MRC-5-adapted virus, for example, replicates efficiently in MRC-5 cells but fails to replicate to detectable levels in humans (35a, 40). A viable live vaccine candidate must achieve a balance between efficient replication of the virus in cell culture and a limited infection in humans, one that is sufficient to elicit an immune response without causing illness.

In this study, we have utilized chimeras between HM-175 and AGM-27, two HAV strains with distinct molecular and biological properties, to better understand the genetic basis for HAV replication in cell culture and viral pathogenicity in primates. Our initial studies have focused on the 2C gene since 2C gene mutations have previously been shown to be important for enhanced replication of the HM-175 strain in cell culture (12, 44). To evaluate whether importance of the 2C sequence for replication is unique to the cell culture-adapted variants of HM-175 or whether this is a characteristic of 2C in other HAV strains, chimeras between HM-175 and AGM-27, two of the most divergent HAV strains that have been identified (37), were analyzed for their replication efficiency in cell culture. In addition, these simian-human chimeras were utilized to evaluate the importance of the 2C gene for virulence of HAV. Information gained from these studies may have practical implications for vaccine development.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells. A subclone of the FRhK-4 cell line, 11-1, was used in these studies because replication of HAV in this cloned cell line is more efficient than in the parental cell line (16). Cells were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum, glutamine, nonessential amino acids, 50 µg of gentamicin sulfate per ml, and 2.5 µg of amphotericin B (Fungizone) per ml (10% DMEM).

Reverse transcription, PCR, and DNA sequencing. An AGM-27 virus stock consisting of 10% (wt/vol) African green monkey liver homogenate in phosphate-buffered saline (pH 7.2) (43) was the source of viral RNA for cloning of the cDNA fragments generated by reverse transcription (RT)-PCR. RNA was isolated from 5 to 10 µl of the AGM-27 liver homogenate by either the guanidinium thiocyanate extraction procedure (4) or with Trizol reagent (Gibco-BRL, Bethesda, Md.) in accordance with the manufacturer's instructions. Glycogen (20 µg; Boehringer Mannheim, Indianapolis, Ind.) was added as a carrier prior to precipitation with isopropanol. The RNA was resuspended in 10 µl of sterile water to which 1 µl of a 10 µM stock of reverse primer was added. The solution was heated at 65°C for 3 min and cooled at room temperature for 5 min. The RT reaction was performed in a final volume of 20 µl in a reaction mixture consisting of 10 mM Tris-HCl (pH 8.4), 50 mM KCl, 2.5 mM MgCl₂, 40 U of RNasin (Promega Biotech, Madison, Wis.), 1 mM each deoxynucleoside triphosphate (dNTP), and 8 U of avian myeloblastosis virus reverse transcriptase (Promega Biotech). After synthesis of cDNA at 42°C for 60 min, PCR amplification was performed in a total volume of 100 µl of 10 mM Tris-HCl (pH 8.4), 50 mM KCl, 2.5 mM MgCl₂, 0.2 mM each dNTP, 4 U of Taq polymerase (Perkin-Elmer Corp, Norwalk, Conn.), and 0.5 mM each forward and reverse primer. When necessary, silent mutations were incorporated into the primers to generate restriction enzyme sites required for subsequent cloning steps. The PCR consisted of 35 cycles of 1 min at 94°C, 1 min at 45°C, and 1 to 3 min at 72°C followed by a single cycle at 72°C for 10 min. A second round of PCR with nested primers was performed if necessary. The PCR products were purified from low-melting-point agarose gels by phenol extraction or with a gel purification kit (Qiagen, Chatsworth, Calif.).

cDNA clones. Molecular cloning of cDNAs representing the HAV/7 genome has been reported previously (6). All nucleotide number assignments herein are based on the genomic map of wild-type HM-175 (8). Plasmids encoding chimeric viruses were generated by cloning cDNA fragments from AGM-27 into the infectious pHAV/7 plasmid (7) which had been modified to include a PpuMI site at nucleotides (nt) 3987 to 3993 and an AflII site at nt 4353 to 4358. An AflII site at this position is present in the AGM-27 consensus sequence. HAV/7 has a naturally occurring EcoRI site at nt 4977 to 4982 near the 3' end of the gene. Since this site was not present in the AGM-27 consensus sequence, an EcoRI site was engineered at the analogous position in the AGM-27 gene by primer-directed mutagenesis with PCR. A chimera containing most of the AGM-27 2C gene in the background of the HAV/7 genome was constructed (Fig. 1). This pGR2 chimera contained the AGM-27 consensus sequence from nt 3996 to nt 4981 in an HAV/7 background. The 2C gene of pGR2 had three nucleotide differences from the AGM-27 consensus sequence at positions 4211 (C-to-T transition), 4280 (G-to-A transition), and 4397 (T-to-C transition), but these differences did not change the amino acid sequence. Because the EcoRI site at nt 4977 to 4982 was used for cloning, the pGR2 plasmid contained a glutamic acid residue, which is present in HAV/7 at amino acid position 331 in 2C, instead of a lysine residue, which is present in the AGM-27 consensus sequence.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Genomic structure of full-length cDNA clones of chimeras consisting of HAV/7 (white bars) and AGM-27 (grey bars) sequences. Positions of unique restriction sites used in cloning are indicated.

View larger version (4K): [in this window] [in a new window]   FIG. 2.   Diagram showing the predicted amino acid differences between the cell culture-adapted variant, HAV/7, and AGM-27 in the 2C protein. Bars represent predicted amino acid differences between AGM-27 and HAV/7. Asterisks indicate sites that are critical for growth of HM-175 in cell culture (corresponding to mutations at nt 4087 and 4222) (12). Arrows indicate the two conserved putative NTP-binding motifs.

Establishment of virus stocks. All full-length cDNAs were cloned in pGEM1 (Promega Biotech). In vitro transcription and transfection assays were performed as described previously by Emerson et al. (13) with a few modifications. Briefly, HAV RNA transcribed by Sp6 polymerase (Promega Biotech) from 5 µg of cDNA linearized with HaeII was transfected without purification into 11-1 cells. Transfection of cells was accomplished by the DEAE-dextran method. One week after transfection, half of the cells in each flask were passaged to coverslips and the fraction of cells that contained viral antigen was estimated by immunofluorescence microscopy (13). Transfected cells were harvested by trypsinization when >80% of the cells were infected. Viruses were released from the harvested cells by at least three cycles of freeze-thawing to generate the working virus stocks.

RIFA. A radioimmunofocus assay (RIFA) was used to determine focus size and to quantify virus titers. The RIFA was a modification of that described by Lemon et al. (24) and Anderson et al. (1) and was essentially performed as previously described (16). Cells (11-1) were grown on round 25-mm-diameter Thermolux coverslips fixed to the bottom of each well in six-well plates. Virus was adsorbed to cells for 2 to 4 h at 34.5°C in a CO₂ incubator. Infected cells were overlaid with 5 ml of 0.5% agarose medium and incubated at 34.5°C in a CO₂ incubator for 10 or 12 days. When cells were infected with virus from stool samples, G418 sulfate (Gibco-BRL) was added at a final concentration of 0.5 mg/ml to the sample diluent and agarose overlay solutions. Cells were fixed with acetone and either processed immediately or stored at 20°C. Viral antigen was detected with a primary antibody consisting of chimpanzee hyperimmune serum (11a) and a secondary antibody of ¹²⁵I-labelled sheep anti-human IgG F(ab')₂ fragment (Amersham Corporation, Arlington Heights, Ill.). Foci were visualized by autoradiography.

Growth curves. Growth curve assays were performed to evaluate the relative rates of viral replication in 11-1 cells. Cell monolayers which were >80% confluent in 96-well plates (Falcon) were infected with 0.2 ml of virus diluted in 10% DMEM at a multiplicity of infection of 6 radioimmunofocus-forming units (RFU) per cell. Cells were incubated with virus for 2 to 4 h at 34.5°C in a CO₂ incubator, after which the cells were washed four times with 10% DMEM and overlaid with 0.2 ml of 10% DMEM. At various times, the medium, which contained virus released from the cells, was removed and the volume was adjusted to 0.3 ml with 10% DMEM. The adherent cells in the wells were washed once with 0.2 ml of trypsin solution at room temperature and were subsequently incubated with 0.1 ml of trypsin solution at 34.5°C until the cells rounded up. The trypsinized cells were then harvested by vigorous pipetting and combined with the medium that had been previously removed. The total virus sample (0.4 ml) was stored at 80°C. Cells were lysed by at least three cycles of freeze-thawing, and virus was quantified either by slot blot hybridization or by RIFA. Either HAV/7 or a modified clone of HAV/7 containing a silent mutation was used as a standard for estimation of optimal replication in cell culture. Since a difference was not detected in the replication properties of these two viruses, they were used interchangeably and, for simplicity, are always referred to as HAV/7 in this report.

Slot blot hybridization. RNA was isolated from cell cultures with Trizol reagent (Gibco-BRL) in accordance with the manufacturer's instructions with the following modifications. A 130-µl aliquot of sample was extracted with an equal volume of Trizol. After the 15-min centrifugation step, 100 µl of the aqueous phase was removed and added to 320 µl of a 1:1 solution of 10× SSC (1.5 M NaCl and 0.15 M sodium citrate [pH 7.0]) and formaldehyde. Viral RNA was quantified by slot blot analysis by using a negative-strand ³²P-labelled RNA probe spanning the complete HAV genome. Hybridization was at 50°C for at least 16 h. Blots were washed three times for 30 min each with 2× SSPE (1× SSPE is 10 mM sodium phosphate, 0.15 M NaCl, 1 mM EDTA [pH 7.4]) with 0.1% sodium dodecyl sulfate at room temperature and once with 0.1× SSPE-0.1% sodium dodecyl sulfate at 64°C for 1 h. Autoradiography was performed, and the viral RNA from each time point in the growth curve was analyzed with a Deskscan II scanner (Hewlett-Packard) and a Macintosh computer, using the public domain NIH Image program (developed at the U.S. National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih-image/). At least two or more sister clones were assayed for each virus construct.

Virulence studies with tamarins and chimpanzees. Each of two tamarins was inoculated intravenously with approximately 10^3.8 tissue culture infectious doses (TCID) of GR2 virus or 10^4.8 TCID of the GR3 or GR4 virus in a 0.5-ml volume of inoculum. Four chimpanzees were inoculated (two each) with 10⁵ TCID of either GR2 or GR4 virus in a 0.5-ml volume of inoculum. The number of TCID in the inoculum was determined by RIFA. Blood samples were collected and percutaneous biopsies of the liver were performed weekly on each animal for at least 2 weeks before and 16 weeks after inoculation with virus. The blood samples were analyzed for seroconversion to anti-HAV positivity with a commercial assay (Abbott Laboratories, North Chicago, Ill.) and for serum alanine aminotransferase (ALT) and isocitrate dehydrogenase (ICD) levels by standard techniques (Metpath, Rockville, Md.). Serum liver enzyme values were considered to be above background levels if the value was greater than two times the geometric mean preinoculation value. Histopathology was determined under code and scored on a scale of 1 (mild hepatitis) to 4 (severe hepatitis).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Chimeras between AGM-27 and HAV/7. Transcripts from chimeras containing sequences from the simian virus 2C gene in the HAV/7 background (Fig. 1) were infectious when transfected into 11-1 cells (data not shown). Of the 31 amino acid differences between the 2C proteins of AGM-27 and HAV/7, 30 are present in GR2. Since an EcoRI site near the 3' end of the 2C gene was utilized in constructing the chimeras, they lacked the last AGM-27-specific residue at position 331 of the 2C protein because this residue is encoded just downstream of the restriction site. The GR3 and GR4 chimeras differ from HAV/7 by 17 and 13 amino acid residues, which are clustered near the amino and carboxy termini of 2C, respectively.

RIFA and growth curve analyses. RIFA can be used to evaluate the relative abilities of HAV mutants to replicate in vitro. The cell culture-adapted mutant of HM-175 virus replicated well in cell culture and formed large foci (Fig. 3, HAV/7). AGM-27, which is a wild-type virus, replicated in cell culture but had a small-focus phenotype (Fig. 3). Foci formed by wild-type HM-175 were too small to visualize by using this assay (11a). Replacement of the 2C gene of HAV/7 with AGM-27 sequences (chimera GR2) drastically reduced the ability of the virus to replicate (Fig. 3), confirming the importance of 2C sequence for HAV/7 replication in cell culture and suggesting that the 2C sequence may be important for replication of most if not all HAV strains. The intragenic 2C chimeras GR3 and GR4 formed foci of intermediate size (Fig. 3).

View larger version (45K): [in this window] [in a new window]   FIG. 3.   Comparison of the sizes of the foci formed by HAV/7, AGM-27, and chimeras that contain AGM-27 2C sequences in the HAV/7 background. GR2, GR3, and GR4 chimeras encompass the AGM-27 sequence from nt 3996 to 4981, 3996 to 4357, and 4354 to 4981, respectively.

View larger version (14K): [in this window] [in a new window]   FIG. 4.   Growth curves of simian-human chimeras GR2, GR3, and GR4 and the cell culture-adapted human virus, HAV/7, as measured by hybridization. Viral RNA was quantified by slot blot hybridization followed by autoradiography and densitometry analysis. The 2C gene is demarcated in the bar diagram identifying the virus genotype and is not drawn to scale. AGM-27 sequences are shaded in grey. O.D., optical density.

Virulence studies with tamarins. AGM-27 is virulent in tamarins (14), while HAV/7 is attenuated (5). The GR2 chimeric virus was inoculated intravenously into two tamarins, and levels of liver enzymes in serum, antibody titers, liver pathology, and levels of virus excretion in the feces were evaluated. The GR2 chimera was virulent in tamarins, causing seroconversion to anti-HAV positivity by week 5 after inoculation and significant increases in serum liver enzyme levels (Fig. 5A and B). The pattern of ALT levels (data not shown) during the course of infection was similar to that of ICD. Liver histology was indicative of mild (1+) to moderate (2+) hepatitis 6 to 9 weeks after inoculation in one animal (Fig. 5A) and mild (1+) to moderately severe (3+) hepatitis 5 to 8 weeks after inoculation in the second animal (Fig. 5B). In both animals, the most severe liver pathology was observed coincident with the peak in serum liver enzyme levels. The peak titers of anti-HAV antibodies elicited in response to infection with the GR2 chimera were 1:16,000 and 1:32,000 in S. mystax animals 782 and 783, respectively (Table 1). Virus-specific RT-PCR amplification of fecal samples showed that the highest level of virus was excreted just prior to and concurrent with the peak serum liver enzyme level and seroconversion for HAV in S. mystax animal 782 (Fig. 5A). HAV was detected in the feces of S. mystax animal 783 for 14 of the 16 weeks studied and at significant levels for 10 of these weeks (Fig. 5B). Stool samples that were positive for HAV after one round of PCR were pooled, and the virus titer was determined in duplicate by RIFA. The mean peak levels of virus in the pooled stool samples were 1.7 × 10⁶ and 1.4 × 10⁶ RFU/g of feces for S. mystax animals 782 and 783, respectively. Partial sequence analyses of viral genomes, amplified from stool samples, from regions of the 2B and 2C genes were performed to confirm that the excreted virus contained 2B sequence derived from HAV/7 and 2C sequence derived from AGM-27, as was expected for the GR2 chimera. These results show that the 2C gene of AGM-27 alone can confer the phenotype of virulence for tamarins to an otherwise attenuated virus (HAV/7).

View larger version (26K): [in this window] [in a new window]   FIG. 5.   Biochemical (ICD), serological (anti-HAV), and histopathological (0 to 4+) responses and virus excretion in feces (PCR) of tamarins inoculated with simian-human chimeras GR2 (A and B), GR3 (C and D), and GR4 (E and F). PCR symbols: , stool samples positive for HAV after one round of PCR; , stool samples positive for HAV only after two rounds of PCR; , stool samples negative for HAV even after two rounds of PCR. Histopathology scores: 1+, mild hepatitis; 2+, mild to moderate hepatitis; 3+, moderately severe hepatitis; 4+, severe hepatitis. The preinoculation ICD value is the geometric mean of three values prior to inoculation. NA, liver sample not available. Mystax, S. mystax animal.

Virulence of the GR2 and GR4 chimeras in chimpanzees. Two chimpanzees each, inoculated with either the GR2 or the GR4 chimera, did not display biochemical or histological changes associated with hepatitis (Fig. 6A to D). Animals inoculated with the GR2 chimera seroconverted at week 8 (chimp 1545) and week 17 (chimp 1547) after inoculation, and the peak anti-HAV antibody titers elicited in the two animals were 1:200 and 1:40, respectively (Table 2). Excreted virus was detected in the feces of both chimpanzees infected with the GR2 chimera although the levels of virus present in the stool samples were low (<400 RFU/g of feces) and the duration of excretion was short (Fig. 6A and B; Table 2). Infection of chimpanzees with GR4 resulted in seroconversion at week 8 (chimp 1564) and week 10 (chimp 1558) postinoculation with peak anti-HAV antibody titers of 1:40 and 1:800, respectively. RT-PCR amplification of fecal samples failed to detect virus in stool samples from chimp 1564, indicating that only very low levels of virus, if any, were being excreted. Virus that is present in levels lower than 10 RFU/g of feces is below the limit of the sensitivity of the RT-PCR analysis. Virus was detected in the feces of chimp 1558 just prior to and at the time of seroconversion (Fig. 6D). For comparison, chimpanzees inoculated with at least 80-fold more wild-type AGM-27 virus seroconverted 6 weeks after infection and had peak anti-HAV titers of <1:20 and 1:200 (14). Like the parent strains HAV/7 and AGM-27, both GR2 and GR4 were attenuated for chimpanzees.

View larger version (25K): [in this window] [in a new window]   FIG. 6.   Biochemical (ALT), serological (anti-HAV), and histopathological (0 to 4+) responses and virus excretion in feces (PCR) of chimpanzees inoculated with simian-human chimeras GR2 (A and B) and GR4 (C and D). See legend to Fig. 5 for symbols and details.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

There is considerable evidence that adaptation of HAV to replicate efficiently in cell culture is associated with increased attenuation for several species of primates (5, 21, 33, 35) and for humans (27, 34, 40). This suggests that the two biological properties replication in cell culture and virulence in primates share at least some important genetic determinants. We have utilized chimeras between HAV/7 and AGM-27, two HAV strains with distinctly different molecular and biological properties, to identify genetic factors that contribute to HAV replication in vitro and to HAV virulence in vivo. Information gained from these studies will be applied toward development of a live vaccine for HAV.

Replacement of the HAV/7 2C gene with that of AGM-27 dramatically reduced the ability of the virus to replicate in cell culture, as shown by reduced focus size (Fig. 3) and by lower yields of both viral RNA (Fig. 4) and infectious virus (data not shown) in growth curve analyses. These results confirm and expand the previous studies demonstrating the significant contribution of 2C gene mutations to replication of the HM-175 strain of HAV in cell culture (12, 36, 44) and suggest that the 2C protein may have a central role in replication of most if not all HAV strains in cell culture.

The function of the 2C protein has not yet been definitely demonstrated for HAV. The amino acid differences in 2C between HM-175 and AGM-27 are clustered at the amino and carboxy-terminal ends, while the central region of the protein is highly conserved (only 1 amino acid difference between residues 105 and 261, inclusive, of AGM-27 and HAV/7 [Fig. 2] and no differences in this region between AGM-27 and wild-type HM-175). These comparisons suggested that sequences within the central region may encode important structures necessary for 2C function. Alternatively, it seemed possible that sequences throughout the protein may be important for function and that the sequences at the ends may have coevolved in each virus so that 2C function was maintained even though the sequences had diverged. The two clusters of amino acid sequence differences between the two strains were separated in the intragenic chimeras, GR3 and GR4, resulting in viruses that were still viable (Fig. 3 and 4). The intragenic chimeras replicated at levels intermediate between HAV/7 and the chimera containing the entire 2C gene of AGM-27 (Fig. 3 and 4). These observations are consistent with the idea that the conserved central region of 2C was responsible for the critical function of this protein but that the sequences at the amino- and carboxy-terminal ends contributed to the efficiency of this function. The central region contains an NTP-binding motif (18), a structure that is present in numerous positive-strand RNA viruses (17), including poliovirus, a distantly related picornavirus. The poliovirus 2C protein has been shown to have GTPase and ATPase activities (30, 38) that likely provide energy for one or more processes required for virus replication. Mutation of conserved residues in the NTP-binding motif of the poliovirus 2C protein severely inhibits virus replication, thus greatly compromising virus viability (29, 42). The absolute conservation of the NTP-binding motif between two very different HAV strains (HM-175 and AGM-27) suggests that such an activity may also be functionally important for the role of 2C in HAV replication.

It is difficult at this time to ascertain whether the reduced replication ability of the GR2 chimera in cell culture is a direct result of diminished 2C function. It is plausible that replacement of a large number of amino acid residues at the amino and carboxy termini of HAV/7 2C with simian HAV sequences may have sufficiently altered the secondary or tertiary structure of the polyprotein, thereby interfering with the efficiency of a necessary viral function such as polyprotein processing. Alternatively, the structural properties of the RNA genome itself may have been altered, which might result in decreased stability or less efficient packaging of the RNA genome. Such effects may be responsible for, or at least contribute to, the diminished GR2 chimera replication that was observed in cell culture (Fig. 4). This interpretation seems unlikely, however, since viral functions were not significantly compromised in vivo as the GR2 chimera replicated sufficiently well in tamarins to cause disease (Fig. 5A and B).

Infection of chimpanzees with AGM-27 does not cause overt hepatitis but does protect these animals against heterotypic cross-challenge with a wild-type human virus, HM-175 (14). These observations suggest that the AGM-27 virus functions as a live attenuated vaccine for HAV in chimpanzees, the primate model which most closely resembles humans in its response to HAV. However, the infectivity and/or replication of AGM-27 in chimpanzees is relatively inefficient, and this may be a practical limitation for development of the simian virus itself as a vaccine candidate. It is interesting to note that HAV strains that have been isolated from Old World monkeys (AGM-27, CY-145, and JM-55/CY-55) have amino acid substitutions in VP3 and VP1 that have been found in neutralization escape mutants of human HAV strains (25, 31, 32, 43). This has led to the speculation that the antigenic differences in the capsid proteins (20, 31, 43) may have evolved to facilitate interaction of the phylogenetically distinct viruses with cellular receptors in their natural hosts. This is one possible explanation for the apparent difficulty that is observed in infecting Old World monkeys with human HAV strains and the poor susceptibility of chimpanzees to infection with AGM-27. A chimera between AGM-27 and HAV/7, which has the human HAV-encoded capsid proteins, could be a more promising candidate than the simian virus itself for development of a live attenuated vaccine. If the efficiency of virus infectivity is indeed a function of the virus capsid, infectivity for humans would be greater if the chimeric viral genome was encapsidated by structural proteins from a human strain of virus. This would also be advantageous because such a chimera should induce serum neutralizing antibodies against the major antigenic determinants in VP3 and VP1, which are present in all human strains of HAV.

Tamarins and chimpanzees have been valuable animal models for predicting HAV pathogenicity for humans. In general, partially attenuated HAV strains normally show a greater level of attenuation in chimpanzees and humans than in tamarins (21, 34). This was also true for the GR2 and GR4 chimeras, which appeared to be more attenuated for chimpanzees than for tamarins (Fig. 5 and 6).

Several cell culture-adapted strains of HM-175 have previously been considered as candidate live vaccines. These include the HM-175 P32 virus (21), which was passaged 32 times in primary AGMK cells, HAV/7 (5, 7), which was derived from a molecular clone of a strain that was passaged 35 times in AGMK cells, and the MRC-5 virus (40), which had been adapted to replicate in MRC-5 cells, a substrate that is licensed for vaccine production. The level of attenuation of the HM-175 P32 and HAV/7 strains is comparable to that of the GR4 chimera for tamarins (5, 7, 21) and chimpanzees (21). The MRC-5 virus is significantly more attenuated for tamarins and chimpanzees (35a) than is either its parent strain, HM-175 P32 (21), or the GR4 chimera (this study) and appears to be overattenuated for humans (40).

In previous studies with a large number of chimeras between wild-type HM-175 and the cell culture-adapted variant, HAV/7, combinations of mutations throughout the genome resulted in similar efficiencies of replication of the virus in cell culture (11a, 12). To date, no single mutation that resulted in attenuation of the virus for primates has been identified (11, 11a). Among the simian-human chimeras discussed in this study, the GR4 chimera appears to have a constellation of changes that provides the necessary balance between replication of the virus in cell culture and the appropriate level of attenuation of the virus for both tamarins and chimpanzees that is required in a potential live vaccine candidate. Further studies with simian-human chimeras that contain only a subset of the differences in 2C that are present in GR4 would be useful in determining whether multiple changes contribute to the attenuation of GR4. However, with 13 amino acid differences between HAV/7 and GR4, the number of possible combinations of mutations that can potentially be tested is considerable and is a practical limitation.

Generation of chimeras between two distinctly different strains of HAV, both of which are attenuated for chimpanzees, offers an alternative and novel approach to developing a live vaccine for HAV. With the availability of a new AGMK cell line that is acceptable for vaccine production (32a), the number of HAV strains that can be considered for vaccine development has been greatly expanded. These include both the AGM-27 and HAV/7 parent strains as well as the simian-human chimeras discussed in this study. Our initial observations suggest that the GR4 chimera may be a promising candidate for a live HAV vaccine.