Novel Infectious cDNA Clones of Hepatitis C Virus Genotype 3a (Strain S52) and 4a (Strain ED43): Genetic Analyses and In Vivo Pathogenesis Studies

Source of HCV strains S52 and ED43.Genotype 3a strain S52 and genotype 4a strain ED43 were derived from challenge plasma pools from chimpanzees (5, 16) which had been experimentally infected with sera from chronically infected patients (8, 11).

Amplification, cloning, and sequence analysis.RNA was extracted from 200 μl of the S52 or ED43 plasma pool using high pure viral nucleic acid kit (Roche) or TRIzol reagent (Invitrogen). cDNA was synthesized with SuperScript II or III (Invitrogen) and random hexamers or specific reverse primers (TAG Copenhagen). After treatment of cDNA with RNase H (Invitrogen) and RNase T1 (Ambion), PCR was carried out with BD Advantage 2 polymerase mix (Clontech) (23); PCR of 3′UTR fragments was carried out with Amplitaq Gold DNA polymerase (Applied Biosystems) (83). Gel-purified amplicons were A tailed with Taq DNA polymerase (Invitrogen), cloned in pCR2.1-TOPO or pCR-XL-TOPO (Invitrogen), and transformed in Top10 chemically competent bacteria (Escherichia coli) (Invitrogen). 3′UTR amplicons were subcloned after restriction digestion as previously described (83). Sequence analysis and determination of consensus sequence were done using Sequencher (Gene Codes Corporation) and BioEdit (freeware). Polyprotein alignments and phylogenetic analysis were done using MEGA4.1 (freeware). HCV sequences used for alignments were from the European and American HCV databases (euHCVdb [http://euhcvdb.ibcp.fr/euHCVdb/ ] and LANL [http://hcv.lanl.gov/ ]). Standard molecular techniques, such as restriction digestion-based cloning and fusion PCR, were used for assembly of consensus sequences; fusion PCR was done with Pfu DNA polymerase (Stratagene).

Sequences of strain S52 were obtained by analysis of four amplicons: amplicon i (nucleotides [nt] 24 to 3396), amplicon ii (nt 3359 to 5186), amplicon iii (nt 5065 to 7596), and amplicon iv (nt 7530 to 9401). These amplicons covered the following amino acids: amplicon i, amino acids (aa) 1 to 1019; amplicon ii, aa 1008 to 1615; amplicon iii, aa 1576 to 2419; and amplicon iv, aa 2398 to 3020 on the polyprotein (all nucleotide and amino acid numbers refer to positions on pS52 with nt 1 being the 1st nucleotide of the 5′UTR and aa 1 being the 1st amino acid of the polyprotein; primer sequences were excluded). Another amplicon (amplicon v) contained the C-terminal NS5B sequence (starting from nt 9339) as well as the 3′UTR variable region, poly(U/UC) region, and the first 16 nucleotides of the conserved X region, and was obtained as previously described (83); this amplicon covered aa 3001 to 3021 of the polyprotein sequence. After subcloning, 5 clones of amplicons i, ii, and iv, 6 clones of amplicon iii, and 15 clones of amplicon v were sequenced to determine the consensus sequence. pS52 was constructed using clones derived from amplicons i to iv, a synthesized 3′UTR sequence (Genscript), and pGEM-9Zf-MOD. pGEM-9Zf-MOD was generated by replacement of the NotI/EcoRI fragment containing the HCV H77 sequence in pCV-H77C (83) by a convenient multiple cloning site (5′-GCG GCC GCA ACT TGT TTA AAC GCG CCT TAA TTA AGG ATT GGC GCG CCA ACC GGA ATT C −3′). In pS52, the NotI site is located immediately upstream of the T7 promoter sequence and the C-terminal XbaI site is located immediately upstream of the AscI site.

For strain ED43, 5′UTR and ORF sequences were obtained by two amplicons, amplicon i (nt 28 to 5631) and amplicon ii (nt 5476 to 9376), which covered aa 1 to 1763 and aa 1713 to 3008, respectively (all nucleotide and amino acid numbers refer to positions on pED43). Another amplicon (amplicon iii), spanning the C-terminal NS5B sequence (starting from nt 9301), the 3′UTR variable region, the poly(U/UC) region, and the first 16 nucleotides of the conserved X region, was obtained as previously described (83); this amplicon covered aa 2988 to 3008. After subcloning, 4 clones of amplicon i, 5 clones of amplicon ii, and 10 clones of amplicon iii were sequenced to determine the consensus sequence. pED43 was constructed by using clones derived from amplicons i to iii inserted into pCV-H77C (83) using NotI and NheI sites, thereby retaining the 3′-terminal sequence of pCV-H77C (83). Endotoxin-free maxipreparations (Qiagen) were prepared, and the HCV sequence was confirmed for pS52 and pED43.

The consensus sequence of the almost complete ORF of S52 or ED43 genomes recovered from sera of infected chimpanzees was determined by direct sequence analysis of PCR amplicons obtained in a reverse transcription (RT) nested PCR procedure (23). Primers used for RT, 1st- and 2nd-round PCR are given in Tables S1 and S2 in the supplemental material. To determine the 5′-terminal sequences of S52 and ED43 from challenge chimpanzee pools and transfected chimpanzees, we used the 5′RACE system for rapid amplification of cDNA ends (version 2.0; Invitrogen) with dC or dA tailing technology, according to the manufacturer's instruction, using S52- and ED43-specific reverse primers. The second PCR products were directly sequenced and/or TA cloned into pCR2.1-TOPO (Invitrogen) for subsequent sequencing.

Generation of RNA transcripts and transfections.Plasmid DNA was linearized with XbaI (New England BioLabs), purified (Wizard SV gel and PCR clean-up system; Promega), and in vitro transcribed with T7 RNA polymerase (Promega) (23). Before generation of RNA transcripts for in vitro transfection, XbaI-digested pED43 with and without adaptive mutations was treated with mung bean nuclease (New England BioLabs) (65). The amount of RNA transcripts was estimated by standard agarose gel electrophoresis.

For in vitro transfections, Huh7.5 cells (gift from Charles M. Rice) were plated at 4 × 10⁵ cells per well of a 6-well plate and maintained as previously described (23). After 12 to 24 h, lipofection complexes (2.5 μg of RNA transcripts and 5 μl of Lipofectamine 2000 [Invitrogen] in 500 μl of Opti-MEM [Invitrogen]) were added to cells for approximately 16 h.

For in vivo transfections, chimpanzees were housed in compliance with relevant guidelines and requirements (52) in a facility fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. Chimpanzees CH5276 and CH5300 were inoculated intrahepatically by a percutaneous procedure (85) with RNA transcribed as described above from a total of 20 μg XbaI-digested and purified pS52 or pED43, respectively.

Monitoring of HCV infection in Huh7.5 cells.Huh7.5 cells were immunostained for HCV Core antigen using the primary mouse anti-HCV Core protein monoclonal antibody (B2) (Anogen, Yes Biotech Laboratories) at 1:200 in phosphate-buffered saline (PBS) with 5% bovine serum albumin and the secondary antibody Alexa Fluor 594-labeled goat anti-mouse IgG (Invitrogen) at 1:500 in PBS containing Tween; cell nuclei were counterstained with Hoechst 33342 (Invitrogen). The presence of HCV-positive cells was evaluated by fluorescence microscopy. The primary antibody was previously shown to recognize S52 and ED43 Core proteins (23, 65).

Monitoring of HCV infection in chimpanzees.Preinfection sera were obtained at weeks −39, −5, −1, and 0 for chimpanzee CH5276 and at weeks −16, −5, −1, and 0 for chimpanzee CH5300; preinfection liver biopsy samples were obtained at week −5 and −1 for both animals. For CH5276, serum and liver biopsy samples were collected weekly during weeks 1 to 32. For CH5300, serum and liver biopsy samples were taken weekly during week 1 to 18 and every 2 weeks during weeks 20 to 32. Thereafter, both animals were monitored monthly until week 54 to determine the final outcome of infection. Serum samples were monitored for HCV RNA (in-house TaqMan quantitative PCR [qPCR] [16] and Monitor test 2.0; Roche Diagnostics), HCV antibodies (enzyme-linked immunosorbent assay ELISA 2.0; Abbott), and alanine aminotransferase (ALT). Liver biopsy samples were examined for necroinflammatory changes (6) and steatosis.

Investigation of autologous HCV neutralizing serum antibodies in S52- and ED43-infected chimpanzees.Neutralization assays were described previously (30, 65). Briefly, heat-inactivated sera from chimpanzee CH5276 were preincubated with ∼25 focus-forming units (FFU) of S52/JFH1_{T27186,A4550C} (24) and heat-inactivated sera from chimpanzee CH5300 were preincubated with ∼115 FFU ED43/JFH1_{A28196,A3269T} (24) for 1 h at 37°C, followed by 3-h incubation on 6 × 10³ Huh7.5 cells plated out the day before. After 48-h incubation, cultures were immunostained for HCV NS5A (23). The number of FFU was determined on an ImmunoSpot series 5 UV analyzer (CTL Europe GmbH) with customized software kindly provided to the Copenhagen Hepatitis C Program by Alexey Karulin, Wenji Zhang, and Paul Lehmann. The mean FFU count of 24 negative-control wells (∼5 FFU) was subtracted from FFU counts in experimental wells. Experimental counts obtained by this method were comparable to manual counts, and counts of up to 200 FFU/well were in the linear range of test dilution series. After subtraction of the mean FFU counts in negative-control wells, for chimpanzee CH5276, FFU counts ranged from 26 to 73 FFU/well; for chimpanzee CH5300, the counts ranged from 59 to 146 FFU/well. Percentages of neutralization were obtained by comparison with mean FFU counts from wells, in which the respective virus had been preincubated with serum samples from week −1 or from week 0. For a positive control for neutralization and to establish automated FFU counting, we carried out neutralization of genotype 6a (HK6a/JFH1_{T1389C,A1590C}) virus with serum H06 from a persistently infected patient, patient H (2006, year 29 after infection, genotype 1a) (24) (see Fig. S1 in the supplemental material).

Investigation of cellular immune responses in S52- and ED43-infected chimpanzees.For in vitro stimulation of CD4⁺/CD8⁺ T cells, synthetic peptides 20 aa in length, overlapping by 10 residues, and spanning the entire HCV polyprotein were used. Genotype 3a (strain K3a/650; GenBank accession number D28917) peptides were obtained from the NIH Biodefense and Emerging Infections Research Resources Repository, NIAID, NIH. Information regarding these peptides is available at http://www.beiresources.org . The amino acid sequence of the peptide set differed at 177 amino acid positions (5.9%) compared to the sequence encoded by pS52. Nine peptide pools (41 to 72 peptides per pool) were assembled, roughly corresponding to HCV proteins Core/E1 (aa 1 to 383), E2 (aa 384 to 752), p7/NS2 (aa 753 to 1032), NS3-1 (aa 1033 to 1365), NS3-2 (aa 1349 to 1663), NS4A/B (aa 1664 to 1978), NS5A (aa 1979 to 2426), NS5B-1 (aa 2427 to 2752), and NS5B-2 (aa 2735 to 3021). Synthetic peptides spanning the entire polyprotein of HCV genotype 4a (strain ED43; GenBank accession number Y11604) were purchased from Genemed Synthesis. The patient sequence used to design this peptide set differed from the sequence encoded by pED43 by 67 aa (2.2%). Peptides were organized into 9 pools (30 to 40 peptides per pool) that represented Core/E1 (aa 1 to 400), E2 (aa 381 to 760), p7/NS2 (aa 741 to 1040), NS3-1 (aa 1021 to 1360), NS3-2 (aa 1341 to 1670), NS4A/B (aa 1651 to 1990), NS5A (aa 1971 to 2430), NS5B-1 (aa 2411 to 2720), and NS5B-2 (aa 2701 to 3008).

Direct ex vivo gamma interferon (IFN-γ) enzyme-linked immunospot (ELISPOT) assays were used to quantify HCV-specific IFN-γ-secreting cells in peripheral blood mononuclear cells (PBMC) according to the manufacturer's instructions (U-Cytech). PBMC (2 × 10⁵) isolated on Ficoll gradients were plated in triplicate wells of an IFN-γ ELISPOT plate with peptide pools at a final concentration of 1 μg/ml of each peptide in Aim-V lymphocyte medium (Invitrogen) supplemented with 2% human AB serum (Gemini-BioScience) (AimV-2% HS). The plates were incubated at 37°C for 36 h and then developed according to the manufacturer's protocol. The cutoff for positive responses was set at 10 spots above the number of spots in negative-control wells or three times the number of spots in negative-control wells, whichever was higher. At no time did the number of IFN-γ spots in the negative-control wells exceed 20.

For evaluation of intrahepatic immune responses, liver biopsy samples were gently homogenized in PBS with 2% fetal calf serum (FCS). CD8⁺ T cells were positively isolated using Dynal magnetic beads (Invitrogen). Positive isolation of CD4⁺ T cells was performed from the remaining CD8⁺-depleted fraction by the same procedure. Enriched CD8⁺ and CD4⁺ T cells were separately stimulated with 6 × 10⁶ freshly isolated autologous irradiated PBMC pulsed with the whole HCV peptide set (1 μg/ml final concentration of each peptide) in medium (RPMI 1640 supplemented with 10% FCS, 50 μg/ml penicillin-streptomycin, and 50 IU/ml of recombinant human interleukin 2). The cells were cultured at 37°C under 7% CO₂, and half of the culture medium was replaced every 3 or 4 days with fresh medium. After 3 weeks, the cells were polyclonally stimulated using soluble anti-CD3 antibodies at a final concentration of 0.05 μg/ml and 3 × 10⁶ irradiated human PBMC as feeder cells. Four to five weeks later, CD4⁺ or CD8⁺ T cells were plated in duplicate at 5 × 10⁴ cells per well together with 1 × 10⁵ autologous irradiated PBMC pulsed with peptide pools (1 μg/ml final concentration) as antigen-presenting cells in a 36-h IFN-γ ELISPOT assay. The cutoff for positive responses was set as described above for blood responses.

Nucleotide sequence accession numbers.The nucleotide sequences of the S52 consensus sequence, pS52 (infectious clone of S52), ED43 consensus sequence, and pED43 (infectious clone of ED43) have been deposited in the GenBank database with the accession numbers GU814263, GU814264, GU814265, and GU814266, respectively.

Genetic analysis of genotype 3a strain S52.The S52 HCV source was an acute-phase plasma pool from a chimpanzee, experimentally infected with serum from a chronically infected Italian patient (8). In this pool, the HCV RNA titer was 10^4.3 IU/ml, and the infectivity titer was 10³ chimpanzee infectious doses (CID)/ml (5, 16).

The S52 consensus sequence was determined by clonal sequence analysis of five overlapping RT-PCR amplicons, spanning the complete ORF and partial UTRs as described in Materials and Methods. At each nucleotide position, 5 to 15 clones were analyzed. The S52 ORF consisted of 9,063 nucleotides (nt 340 to 9402; all nucleotide positions refer to pS52), encoding a 3,021-aa polyprotein, followed by a single stop codon. Genetic heterogeneity, with at least one of the analyzed clones being different from the S52 consensus sequence, was found at 199 nt (2.2%) and 67 aa (2.2%) positions (Table 1). Compared to the entire polyprotein, a high percentage of amino acid positions with genetic heterogeneity was found in E2, p7, and NS2 (Table 1). The amino acid sequence of E2 hypervariable region 1 (HVR1) was identical among the clones. As seen in Table S3 in the supplemental material, which shows amino acid positions with genetic heterogeneity, there was evidence for 2 different S52 quasispecies populations. For each sequenced clone, differences to the consensus sequence were found on average at 0.5% of positions at the nucleotide level (range, 0.1 to 1.1%), as well as on the amino acid level (range, 0.1 to 1.5%). At nucleotide position 5358, no distinct consensus could be determined, since 3/6 clones had T, and the other 3 clones had C, with T and C encoding the same amino acid.

In the 5′UTR sequence initially determined for S52 (nt 24 to 339), genetic heterogeneity among the analyzed clones was found at three nucleotide positions (Table 1). In the S52 3′UTR, the first 23 nt of the variable region (nt 9403 to 9425) were identical in all 15 clones (see Fig. S2 in the supplemental material). This sequence was followed by a TTC motif (nt 9426 to 9428) that was present in 13/15 clones. Assuming a variable region of 26 nt (nt 9403 to 9428), the length of the poly(U/UC) region, which could be determined in 3/15 clones, was 108, 111, and 123 nt. The first 16 nucleotides of the 3′UTR X region were identical in all analyzed clones.

Partial sequences were previously determined for the S52 source patient; the partial 5′UTR and complete Core/E1 sequence (nt 58 to 1488) determined from HCV recovered from this patient (7-9) differed from the S52 consensus sequence obtained for the first passage chimpanzee pool at only one nucleotide/amino acid position (in E1). In contrast, the consensus S52 ORF sequence determined in the present study differed from 3 published genotype 3a isolates (55, 64, 81) by 4.8 to 6.5% and 3.6 to 5.8% at the nucleotide and amino acid level, respectively (Table 2). A phylogenetic analysis of the polyprotein of published HCV cDNA clones and representative HCV isolates showed that pS52 clustered with other genotype 3a isolates (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 1.

Phylogenetic tree of pS52, pED43, in vivo infectious cDNA clones, and representative isolates of HCV genotypes 1 to 4. GenBank accession numbers are shown (the names of the isolates/clones are shown in parentheses). In vivo infectious cDNA clones are indicated by an asterisk. Multiple polyprotein sequence alignment and neighbor-joining tree analysis using amino acid p-distance model were performed using MEGA4.1. The scale shows the genetic distance in amino acid substitutions per site.

Compared to 2 other genotype 3a 5′UTR sequences, GenBank accession numbers D17763 and D28917 (64, 81), the obtained S52 consensus sequence showed differences at one and three nucleotide positions, respectively. A higher degree of variability was seen in the short 3′UTR variable region; a comparison of this region to that of other genotype 3a isolates, D28917 (81), AF009075 (36), D17763 (64), and D85024 and D85025 (82), is shown in Fig. S2 in the supplemental material. A consensus ACACTCC motif (nt 9418 to 9424) was highly conserved. The TG dinucleotide preceding the beginning of the poly(U/UC region) for certain isolates (36, 78) was found in only 1 of 15 S52 clones analyzed. The lengths of the 3′UTR poly(U/UC) tracts were 84, 86, and 110 nt for D85024 (82), D85025 (82), and AF009075 (36), respectively. The S52 consensus sequence of the first 16 nt of the 3′UTR X region was identical to genotype 3a isolates AF009075, D85024, and D85025 (36, 82), as well as to the genotype 1a cDNA clone pCV-H77C (AF011751) (83).

Generation of consensus clone pS52.The consensus full-length cDNA clone pS52 was constructed as described in Materials and Methods. The S52 sequence contained the following three structural elements: (i) 5′UTR, (ii) ORF, and (iii) 3′UTR. The first structural element, the 5′UTR, consisted of 339 nt, in which nt 24 to 339 were the S52 consensus sequence, while nt 1 to 23 were deduced from published genotype 3a 5′UTR sequences (D17763 and D28917) (64, 81). For nt 1, at which G (D28917) and A (D17763) occurred, G was chosen to optimize in vitro transcription. The second structural element, the ORF, consisted of 9,063 nt (nt 340 to 9402) with two coding changes, G1037A and G1913A, in comparison to the S52 consensus sequence. However, at both positions, A encoded by pS52 was present in 2/5 clones analyzed. In addition, in clonal analysis previously performed for a S52 Core-E2 amplicon (50), A was found as consensus at both positions. Noncoding changes compared to the S52 consensus sequence were A639G, A915T, C1488T, G1575A, C1707T, C2655T, C2805T, C3069T, G3792A, T5187C, T7755A, T8469C, and G8745A. The A915T and T7755A changes were introduced to remove consensus XbaI sites at both positions, which were present in all 5 clones analyzed. All other noncoding changes in pS52 were found in at least 1 of the clones covering the respective position. The third structural element, the 3′UTR, consisted of 235 nt (nt 9403 to 9637) and contained three regions, a variable region of 26 nt (nt 9403 to 9428), which was identical to the S52 nucleotide consensus sequence, a poly(U/UC) region of 111 nt (nt 9429 to 9539), chosen from a representative clone, and a conserved X region of 98 nt (nt 9540 to 9637), determined by the sequence of pCV-H77C (83). The X region of 2 genotype 3a isolates (D85024 and D85025 [82]) was identical to that of pCV-H77C (83), whereas genotype 3a isolate AF009075 (36) differed at positions 9594 and 9635. An XbaI site was inserted immediately downstream of the HCV 3′UTR for generation of the HCV 3′ end (83).

Genetic analysis of genotype 4a strain ED43.The ED43 HCV source was an acute-phase challenge plasma pool from a chimpanzee, experimentally infected with serum from a chronically infected Egyptian patient. This plasma pool had an HCV RNA titer of 10^5.6 IU/ml and an infectivity titer of 10⁵ CID/ml (5, 16). Previously, the complete ORF of the source patient's virus was sequenced by Chamberlain et al. (11) (GenBank accession number Y11604), and the complete 3′ UTR of the patient virus was sequenced by Kolykhalov et al. (43E; Peter Simmonds, personal communication) (36).

In the present study, the consensus sequence of ED43 from the chimpanzee plasma pool was determined by clonal sequence analysis of three overlapping RT-PCR amplicons spanning the complete ORF and partial UTRs as described in Materials and Methods. In agreement with the patient virus sequence (11), the ED43 ORF was found to consist of 9,024 nt (nt 341 to 9364), coding for 3,008 aa, and terminated by two stop codons. Genetic heterogeneity, with at least one of the analyzed clones deviating from the ED43 consensus sequence, was found at 145 nucleotide (1.6%) and 64 amino acid (2.1%) positions (Table 3). Compared to the average for the entire polyprotein, heterogeneity was relatively high in Core, E1, p7, NS2, and NS4A. The nucleotide and amino acid sequences of HVR1 were identical among the clones. Amino acid positions, at which individual clones differed from the ED43 consensus sequence, are shown in Table S4 in the supplemental material. For each sequenced clone, differences to the consensus sequence were found on average at 0.3% (range, 0.1 to 0.5%) at the nucleotide level and 0.4% (range, 0.1 to 0.8%) at the amino acid level. No distinct consensus could be determined at the noncoding positions 1966 (G/A), 1999 (C/T), 3751 (A/G), and 3871 (C/T) where 2/4 clones represented each sequence.

In the determined ED43 5′UTR sequences (nt 28 to 340), genetic heterogeneity among 4 clones was found at 6 positions (Table 3). In the ED43 3′UTR, the 36-nt variable region (nt 9365 to 9400) was identical in the 10 clones analyzed (see Fig. S3 in the supplemental material). The exact length of the poly(U/UC) region could be determined in all 10 clones and ranged from 72 to 86 nt. The first 16 nt of the 3′UTR X region were identical in all clones analyzed.

For ED43 derived from the source patient (GenBank accession number Y11604) (11), nucleotides 62 to 340 of the 5′UTR were determined; this sequence differed from the ED43 consensus sequence derived from the chimpanzee plasma pool at 2 positions. The ED43 consensus ORF sequence, determined in the present study, differed at 125 nucleotide (1.4%) and 67 amino acid (2.2%) positions from the patient sequence (Tables 4 and 5). Differences of at least 2.2% at the amino acid level were detected in NS2, NS4B, NS5A, and NS5B. Differences of less than 1% were detected in E1, E2, and NS5A, and notably, the HVR1 sequence of both isolates was identical at the nucleotide and amino acid levels. At aa 2011 (position 39 of NS5A) of the ED43 polyprotein, C was found; C at this position was previously shown to be critical for replication (71). However, in the source patient, W was reported to be present at this position (11). The 3′UTR variable region of the ED43 consensus sequence determined in the present study was identical to the equivalent sequence of the source patient determined by Kolykhalov et al. (GenBank accession number AF009077) (36). In this sequence, the poly(U/UC) region of 46 nt was shorter than we observed for ED43. The consensus sequence of the first 16 nt of the ED43 X region (nt 9482 to 9497) was identical to the equivalent sequence of AF009077.

From 7 other genotype 4a isolates with reported ORF consensus sequence (19, 76), the ED43 consensus sequence differed at 8.8 to 9.5% and 5.5 to 6.7% at the nucleotide and amino acid levels, respectively (Tables 4 and 5). Phylogenetic analysis showed that the ED43 consensus sequence determined in this study clustered with other genotype 4a isolate sequences (Fig. 1). The obtained ED43 5′UTR consensus sequence differed from a published genotype 4a 5′UTR sequence (GenBank accession number D45193) (20) at only 1 position. High homology was found between the 3′UTR variable region of ED43 and that of several other genotype 4a sequences (20, 77) (see Fig. S3 in the supplemental material). The ED43 variable region was terminated by a TG dinucleotide as reported for other isolates (36, 78). A similar length of the poly(U/UC) region of 82 nt was reported for another genotype 4a isolate (GenBank accession number AF009076). In addition, the first 16 nt of the ED43 X region were identical to the equivalent sequence of this genotype 4a isolate (36), as well as to pCV-H77C (83).

Generation of consensus clone pED43.The consensus full-length cDNA clone pED43 was constructed with the following three structural elements: (i) 5′UTR, (ii) ORF, and (iii) 3′UTR. The first structural element, the 5′UTR, consisted of 340 nt, with nt 28 to 340 having the ED43 consensus sequence, while nt 1 to 27 were derived from D45193 (20); a G was inserted immediately upstream of the 5′UTR to facilitate in vitro transcription. The second structural element, the ORF, consisted of 9024 nt (nt 341 to 9364), encoding the ED43 amino acid consensus sequence. Compared to the ED43 consensus sequence, noncoding changes were A2458G, A2593G, C3988T, A4459C, C4915T, and T5428C; each of these changes was present in 1/4 clones analyzed. For determination of the pED43 sequence at nt 1966 and 1999, at which no distinct consensus was obtained, we used a previously determined consensus sequence (50). The third structural element, the 3′UTR, consisted of 215 nt (nt 9365 to 9579) with a variable region of 36 nt (nt 9365 to 9400) identical to the ED43 consensus sequence, with a representative poly(U/UC) region of 81 nt (nt 9401 to 9481) and with a conserved X region of 98 nt (nt 9482 to 9579) determined by the sequence of pCV-H77C (83). The sequence of the X region in pED43 is identical to that of another genotype 4a isolate (AF009076) (36). An XbaI site was introduced immediately downstream of the HCV 3′ UTR (83).

RNA transcripts of pS52 and pED43 were nonviable in Huh7.5 hepatoma cells.Previously, Huh7.5 cells were shown to be permissive to infection with strain JFH1 (79, 89) and JFH1-based intra- and intergenotypic recombinants, including recombinants with Core-NS2 of S52 and ED43 (23, 24, 30, 42, 55, 65, 86). Therefore, we tested whether full-length S52 and ED43 replicated or led to productive infection of Huh7.5 cultures. Thus, replicate cultures were transfected with RNA transcripts from pS52, pED43, and positive-control pJ6/JFH1 (42). For J6/JFH1, HCV-Core antigen-positive cells were detected after transfection, and viral spread to almost the complete Huh7.5 culture occurred in 4 to 10 days. In contrast, we detected no HCV-Core-positive cells in S52 and ED43 cultures, stained two or three times per week, and monitored for 4 weeks. In total, four independent transfections with RNA transcripts from pS52 and two transfections with pED43 transcripts were analyzed.

We next tested whether selected adaptive mutations, leading to efficient growth of intergenotypic recombinants S52/JFH1 (23) and H77/JFH1 (65, 86) and/or JFH1 (33) in hepatoma cell lines could confer replication capability to the full-length S52. Therefore, we constructed pS52 with single-nucleotide exchanges in different proteins as follows: in p7, T2717G (identified in S52/JFH1 [23]); in NS3, A4097T (identified in H77/JFH1 [65, 86]) or A4549C (identified in S52/JFH1 [23]); and in NS5A, G7171C (identified in S52/JFH1 [23]) or G7621C (identified in JFH1 [33]). Similarly, we introduced two NS2 mutations in pED43 (A2819G and A3269T) that have been shown to confer cell culture viability to ED43/JFH1 (65). However, after transfection of Huh7.5 cells with the respective RNA transcripts, no HCV-Core-positive cells were observed; the ED43_{A2819G,A3269T} culture was monitored for 1 week, and all other cultures were monitored for 4 weeks. Thus, cDNA clones pS52 and pED43, with or without putative adaptive mutations, were apparently not replication competent in Huh7.5 cells, and long-term cultures did not lead to adaptation that yielded infectious particles.

Finally, for comparison, we also showed that previously developed in vivo infectious cDNA clones of genotype 1a (pCV-H77C) (83), 1b (pCV-J4L6S) (85), and 2a (pJ6CF) (84) were nonviable in Huh7.5 cells. After transfection with the respective RNA transcripts, these cultures were monitored for 2 weeks by immunostaining for HCV-Core.

RNA transcripts of pS52 were infectious in vivo.After intrahepatic transfection of chimpanzee CH5276 with pS52 RNA transcripts, the HCV RNA titers increased to peak levels of 10⁵ to 10^5.5 IU/ml during weeks 6 to 14 (Fig. 2A). The animal became persistently infected with relatively high virus titers at the end of follow-up (week 54). Furthermore, CH5276 developed acute hepatitis with elevated liver enzyme levels. ALT levels of ∼100 U/liter coincided with significant necroinflammatory liver changes, detected during weeks 19 to 32 (Fig. 2A). Analysis of liver biopsy samples showed no evidence of steatosis, which has been reported to be correlated with genotype 3a infection (61).

• Open in new tab
• Download powerpoint

FIG. 2.

Course of HCV infection and host immune responses in chimpanzee CH5276 following intrahepatic transfection with RNA transcripts of pS52 (genotype 3a). (A) Course of infection. Serum HCV RNA was monitored by in-house TaqMan assay (detection limit of 10 IU/ml) and/or by Roche Monitor test 2.0 (detection limit of 600 IU/ml). At the top of the panel, HCV RNA test results, anti-HCV antibody results, and liver histology scores are shown. Serum samples that were positive by TaqMan and/or by Monitor test 2.0 (solid black rectangles) and serum samples negative by TaqMan (white rectangle) are indicated in the top row. Anti-HCV antibodies were detected by 2nd generation ELISA: +, positive; −, negative. Necroinflammatory liver changes were scored as follows: 0, healthy or normal; 1, mild; 2, mild-moderate; 3, moderate-severe; 4, severe. HCV Monitor titers are indicated by the solid black circles in the graph; samples below the detection limit are shown as not detected (ND). The gray shaded area in the graph shows serum ALT levels (U/liter). The time points of direct sequence analysis of recovered viral genomes are indicated by the two white arrows. (B) Serum neutralizing antibodies. Percent neutralization of JFH1-based intergenotypic recombinants expressing S52 envelope proteins. The serum samples were diluted 1:20; >50% neutralization was considered significant. Values represent the means of three neutralizations; the standard error of the mean (SEM) range was 3 to 13%. Negative values are shown as 0%. For 1:80 serum dilutions, neutralization was <20% at all time points. (C and D) Peripheral and intrahepatic CD4⁺/CD8⁺ T-cell responses. The number of IFN-γ-secreting cells after stimulation with a panel of overlapping peptides specific for genotype 3a (strain K3a/650) and spanning the entire HCV polyprotein was determined in ELISPOT assays. PBMC were used directly. Intrahepatic CD4⁺ and CD8⁺ T cells were expanded from liver biopsy samples as described in Materials and Methods. The bars represent the total numbers of IFN-γ-secreting CD4⁺ and CD8⁺ T cells following stimulation with the different pools, after background subtraction. Cutoff points (dotted lines) were determined for individual experiments as described in Materials and Methods. Results below the cutoff are indicated by bars up to the dotted line. ND, not detectable.

The consensus sequence (nt 296 to 9387) of viral genomes recovered from serum at peak HCV RNA titers (weeks 7 and 10) was identical to pS52. Using the 5′RACE procedure, the 5′-terminal sequence (nt 1 to 350) was determined for week 10 viruses; the consensus sequence was identical to pS52. However, at nt 1, we observed a quasispecies G/A (pS52 had G). Subsequent analysis of the 5′-terminal sequence of S52 from the acute-phase source plasma pool, not previously analyzed, showed that S52 had consensus A at position 1. Thus, it is possible that the heterogeneity at nucleotide position 1 in the transfected animal represented a reversion to the original sequence of strain S52.

The chimpanzee became anti-HCV positive in a commercial enzyme-linked immunosorbent assay (ELISA) test at week 19 postinfection (Fig. 2A). However, chimpanzee CH5276 did not develop significant levels of autologous neutralizing antibodies during the acute infection, since preincubation of S52/JFH1 viral particles (24) with 1:20 and 1:80 dilutions of week 2 to 32 sera did not lead to >50% neutralization of S52/JFH1 infectivity in Huh7.5 cells compared to preincubation with preinfection sera (Fig. 2B).

To further examine the immunopathogenesis of S52 infection, we monitored the occurrence of genotype 3a-specific IFN-γ-secreting CD4⁺/CD8⁺ T cells in peripheral blood (Fig. 2C) and liver biopsy (Fig. 2D) samples. Throughout follow-up, PBMC collected from chimpanzee CH5276 did not have any detectable IFN-γ secretion above the background level in ELISPOT assays, after stimulation with 9 different genotype 3a peptide pools (Fig. 2C and Table 6). Intrahepatic IFN-γ-secreting CD4⁺/CD8⁺ T cells were studied similarly after in vitro expansion and were detected from week 9 (Fig. 2D and Table 6). An increase in the percentage of IFN-γ-secreting intrahepatic T cells was detected during weeks 11 to 32, several weeks before the occurrence of peak ALT levels; it also preceded the most pronounced necroinflammatory liver changes, which were observed during weeks 19 to 32. Intrahepatic T-cell responses were detected at 14/23 time points tested, with the strongest response at week 22 (Fig. 2D and Table 6). Intrahepatic T-cell responses were characterized by limited breadth; the highest diversity was recorded at week 18, with responses to 5/9 peptide pools (Core/E1, p7/NS2, NS3-2, NS4A/B, and NS5B-1). Stimulation with the NS3-2 peptide pool resulted in the highest response rate (14/23 time points tested) and the highest number of IFN-γ-secreting cells (Table 6).

RNA transcripts from pED43 were infectious in vivo.After intrahepatic transfection of chimpanzee CH5300 with pED43 transcripts, HCV RNA titers increased to peak levels of 10^4.5 to 10^5.5 IU/ml during weeks 1 to 8 posttransfection (Fig. 3A). During weeks 9 to 12, there was a 100-fold decrease in HCV RNA titers, and the virus titers remained at relatively low levels (10^2.5 to 10⁴ IU/ml) throughout 54 weeks of follow-up. The course of infection was characterized by a fast onset of acute hepatitis with peak serum ALT levels of 100 to 200 U/liter during weeks 5 to 10 (Fig. 3A). Peak ALT levels coincided with significant necroinflammatory liver changes during weeks 5 to 13. Following week 10, ALT levels decreased to 20 to 70 U/liter, and subsequently, we observed a decrease in the score of liver necroinflammatory changes. There was no evidence of steatosis in liver biopsy samples.

• Open in new tab
• Download powerpoint

FIG. 3.

Course of HCV infection and host immune responses in chimpanzee CH5300 following intrahepatic transfection with RNA transcripts of pED43 (genotype 4a). (A) See the legend to Fig. 2A. (B) JFH1-based intergenotypic recombinants expressing ED43 envelope proteins were used as described in the legend to Fig. 2B; the SEM range was 1 to 16%. For 1:80 serum dilutions, neutralization was <20% at all time points. (C and D) Peptides specific for genotype 4a (strain ED43) were used as described in the legend to Fig. 2C and D.

The consensus sequence (nt 297 to 9359) of viral genomes recovered from week 1 and 6 serum of chimpanzee CH5300 did not show any changes compared to the pED43 sequence. Using the 5′RACE procedure, the 5′-terminal sequence (nt 1 to 351) was determined for week 6 viruses from transfected chimpanzee CH5300 and from the original chimpanzee plasma pool source sample; both consensus sequences were identical to pED43 with A at position 1. However, the G added before the 5′-terminal A in pED43 to enhance in vitro transcription was not present in viruses recovered from CH5300.

In chimpanzee CH5300, the ELISA test became positive from week 6 posttransfection (Fig. 3A). However, CH5300 did not develop significant levels of autologous neutralizing antibodies during the acute infection (Fig. 3B), since preincubation of ED43/JFH1 particles (65) with 1:20 and 1:80 dilutions of week 2 to 32 sera did not lead to significant neutralization of ED43/JFH1 infectivity in Huh7.5 cells (Fig. 3B).

Intrahepatic CD4⁺/CD8⁺ T cells, secreting IFN-γ upon stimulation with HCV genotype 4a peptide pools, were detected at week 4 posttransfection (Fig. 3D), as ALT levels increased and the first evidence of necroinflammatory changes was seen in the liver (Fig. 3A). The strongest intrahepatic T-cell response was observed during weeks 7 to 11, coinciding with significant necroinflammatory liver changes. However, intrahepatic T cells reactive to genotype 4a peptides were detected throughout follow-up. CH5300 mounted a multispecific T-cell response. At 12/21 time points with a T-cell response above the threshold level, more than 4/9 peptide pools induced IFN-γ secretion by intrahepatic T cells (Table 7). The greatest diversity was observed at weeks 4 and 7 (response to 8/9 pools) and at weeks 8, 11, and 15 (response to 7/9 pools). The highest stimulation rates were achieved with the NS5B-2, Core-E1, NS3-2, and NS5A peptide pools (response at 20, 16, 16, and 15 of 25 time points tested, respectively) (Table 7).

In chimpanzee CH5300, detection of HCV-specific intrahepatic T cells was followed by occurrence of HCV genotype 4a-reactive PBMC at several time points during week 7 to 15 (Fig. 3C and D). At 7 of 24 time points tested, IFN-γ secretion was observed upon stimulation with NS3-2 peptide pools. In addition, reactivity to the NS3-1 pool was seen at weeks 7 and 8 (Table 7).