Hepatitis C virus persisting after clinically apparent sustained virological response to antiviral therapy retains infectivity in vitro

Authors

  • Sonya A. MacParland,

    1. Molecular Virology and Hepatology Research Group, Division of Biomedical Sciences, Faculty of Medicine, Health Sciences Centre, Memorial University, St. John's, Newfoundland, Canada
    Search for more papers by this author
  • Tram N. Q. Pham,

    1. Molecular Virology and Hepatology Research Group, Division of Biomedical Sciences, Faculty of Medicine, Health Sciences Centre, Memorial University, St. John's, Newfoundland, Canada
    Search for more papers by this author
  • Clifford S. Guy,

    1. Molecular Virology and Hepatology Research Group, Division of Biomedical Sciences, Faculty of Medicine, Health Sciences Centre, Memorial University, St. John's, Newfoundland, Canada
    Search for more papers by this author
  • Tomasz I. Michalak

    Corresponding author
    1. Molecular Virology and Hepatology Research Group, Division of Biomedical Sciences, Faculty of Medicine, Health Sciences Centre, Memorial University, St. John's, Newfoundland, Canada
    2. Discipline of Laboratory Medicine, Faculty of Medicine, Health Sciences Centre, Memorial University, St. John's, Newfoundland, Canada
    • Molecular Virology and Hepatology Research Group, Faculty of Medicine, Health Sciences Centre, Memorial University, St. John's, Newfoundland, Canada A1B 3V6
    Search for more papers by this author
    • fax: 709-777-8279.


  • Potential conflict of interest: Nothing to report.

Abstract

Hepatitis C virus (HCV) can persist in the liver, lymphoid cells, and serum of individuals with apparently complete spontaneous or therapy-induced resolution of hepatitis C and can replicate in vivo and in vitro in human T cells. The current study was aimed at assessing the infectivity of HCV persisting at very low levels using the previously established HCV infection system in human T cells. Naive lymphoid cells were exposed to plasma and/or supernatants from cultured peripheral blood mononuclear cells from nine individuals with apparent sustained virological response after completion of antiviral therapy. Exposed cells were analyzed for HCV RNA–positive and HCV RNA–negative strands and, in selected cases, for HCV nonstructural protein 5a (NS5a), the appearance of HCV variants, and the release of virions by immunoelectron microscopy (IEM). The results showed that 11 of the 12 established cultures became HCV RNA–positive strand–reactive, whereas 4 also expressed the virus replicative strand. NS5a protein was detected in the de novo infected cells, and clonal sequencing revealed HCV variants not found in inocula. IEM demonstrated enveloped HCV particles in plasma used as inocula and in culture supernatant from T cells exposed to that plasma. Overall, HCV carried in three of the nine individuals studied elicited productive infection in vitro. Conclusion: HCV persisting at very low levels long after therapy-induced resolution of chronic hepatitis C can remain infectious. The retained biological competence of the virus might have implications with respect to the mechanisms of its persistence and the epidemiology of HCV infection. (HEPATOLOGY 2009.)

Hepatitis C virus (HCV) is a single-stranded RNA virus that chronically infects approximately 170 million people worldwide. Up to 85% of the infected individuals may develop chronic hepatitis C (CHC). HCV is infectious even in trace amounts, with approximately 20 virus copies capable of transmitting infection in chimpanzees.1 Recently, the introduction of nucleic acid amplification assays detecting HCV genomes with enhanced sensitivity, which has reached in our laboratory <10 virus genomes or virus genome equivalents (vge)/mL or <2 IU/mL, has revealed that HCV can persist at low levels in individuals with apparently complete resolution of hepatitis C occurring either spontaneously or because of antiviral therapy.2–7 In general, occult HCV infection is considered when small quantities of HCV RNA are identifiable in the serum (usually below 100 vge/mL), peripheral blood mononuclear cells (PBMCs), and/or liver of individuals who are repeatedly serum HCV RNA–nonreactive by clinical laboratory tests with sensitivities ranging between 52 and 1000 vge or 10 and 615 IU/mL and have no clinical or biochemical evidence of liver disease.2, 3, 5–7 In this silent form of HCV infection, the detection of HCV RNA replicative (negative) strand is not uncommon, particularly when ex vivo activated PBMCs are tested by sensitive HCV RNA–negative strand–specific nested reverse-transcription polymerase chain reaction (RT-PCR) combined with nucleic acid hybridization (NAH) analysis of the resulting amplicons.2, 8

Although originally thought to be strictly hepatotropic, HCV has been shown in numerous studies to also invade and replicate in immune cells.9–11 In our recent work, conclusive evidence of HCV replication in CD4+ and CD8+ T lymphocytes, B cells, and monocytes was presented.7 It has also been shown that the same immune cell subsets can be infected in both CHC and persistent low-level HCV infection continuing after resolution of CHC.7 Furthermore, primary T lymphocytes from healthy individuals have been found to be susceptible to HCV infection in vitro.9 Along this line, mathematical modeling has independently predicted that HCV originating from extrahepatic reservoirs, possibly the immune system, constitutes approximately 3% of the circulating virus pool in CHC.12 However, an analysis of HCV quasispecies occurring in the liver, plasma, PBMCs, and lymphoid tissue of a patient waiting for a liver transplant demonstrated that extrahepatic variants may constitute more than 50% of those occurring in serum.10

The clinical relevance of low-level HCV carriage, including its potential ability to transmit infection, is yet to be determined. Nonetheless, HCV reactivation has been reported in patients with HCV RNA clearance confirmed by standard clinical laboratory assays prior to liver transplantation.13, 14 However, the opposite has also been described, although no details have been given regarding assay sensitivity and the quantity of the template analyzed.15 Also, HCV RNA has been identified in anti-HCV–reactive patients receiving HCV-negative bone marrow16 or an HCV-negative kidney.17 Taken together, these findings suggest the possibility that occult HCV infection could have both pathogenic and epidemiological importance.

At present, sustained virological response (SVR) is defined as serum HCV RNA negativity by clinical laboratory assays for at least 6 months after completion of antiviral therapy. However, given that the identification of low-level (occult) HCV infection is made possible only by the use of research tests with a sensitivity much greater than that of those applied for clinical use, it is not surprising that low levels of HCV RNA are frequently escaping detection, giving conflicting results on the occurrence and infectivity of HCV persisting at trace levels. Implementation of assays detecting HCV RNA with a greater sensitivity (preferably <10 vge/mL) for clinical and population-based testing should meaningfully contribute to the identification of the scope of potential problems associated with low-level HCV infection.18

An HCV cell culture system allowing for authentic propagation of wild-type HCV in primary human T cells has previously been established in this laboratory.9 In the current study, this system was employed to assess the potential infectivity of HCV persisting at trace quantities for years in patients who achieved SVR after completion of interferon alpha (IFNα) therapy with or without ribavirin. Our investigation has focused on randomly selected cases that, although repeatedly serum-negative by the clinical test, were found by RT-PCR/NAH to be positive for HCV RNA in sera and in PBMCs after their ex vivo stimulation.2, 8 We have discovered that the residual virus carried by some of the individuals has the capacity to de novo infect and propagate in T cells, and this strongly suggests that the virus, persisting as an occult infection, can retain its biological competence and thus be potentially infectious.

Abbreviations

5′-UTR, 5′-untranslated region; CHC, chronic hepatitis C; F, female; FACS, fluorescence activated cell sorting; HCV, hepatitis C virus; HCV sRNA, synthetic hepatitis C virus RNA; IEM, immunoelectron microscopy; IFNα, interferon alpha; IL-2, interleukin-2; IVDU, intravenous drug use; M, male; mAb, monoclonal antibody; NAH, nucleic acid hybridization; ND, not detectable; NS5a, nonstructural protein 5a; P-IFN, pegylated interferon alpha; PBMC, peripheral blood mononuclear cell; PCR, polymerase chain reaction; PHA, phytohemagglutinin; R, ribavirin; rHCV UTR-E2, recombinant hepatitis C virus 5′-untranslated region E2 fragment; RT-PCR, reverse-transcription polymerase chain reaction; SVR, sustained virological response; vge, virus genome equivalents.

Patients and Methods

HCV Inocula and Cell Targets.

Nine patients who achieved SVR after completion of IFNα or IFNα/ribavirin therapy, as defined by repeated serum HCV RNA negativity by the Roche Amplicor HCV version 2.0 assay (sensitivity, 500 IU/mL or 1000 vge/mL; Roche Molecular Diagnostics, Pleasanton, CA) and normal liver function tests assessed at 6- to 12-month intervals, were investigated in this study (Table 1). All patients were anti-HCV antibody–positive by enzyme immunoassay (Abbott Diagnostics, Mississauga, Canada). The follow-up period after SVR ranged from 24 to 72 months. All nine individuals were found to carry HCV RNA at the time of this study when total RNA isolated from 500 μL of serum was assayed by highly sensitive RT-PCR/NAH (sensitivity of ≤10 vge/mL or ≤2 IU/mL) that was previously established.2 The estimated HCV RNA loads in the patients' sera ranged from <40 to 400 vge/mL, with the exception of 6/F (59/F), who carried as much as 1.6 × 103 vge/mL (Table 1). Also, although PBMCs collected from the patients were seemingly HCV RNA–nonreactive by the same highly sensitive assay, a 72-hour culture of the PBMCs with phytohemagglutinin (PHA; 5 μg/mL; Sigma, Oakville, Canada) and recombinant interleukin-2 (IL-2; 20 IU/mL; Roche)2 enabled detection of the virus genome in seven individuals at estimated levels between <10 and 300 vge/μg of total RNA (Table 1). Furthermore, HCV RNA–negative strand in PBMCs was detected in five cases (Table 1).

Table 1. Clinical and Virological Characteristics of Individuals with Occult HCV Infection
CaseAge/SexRoute of InfectionAntiviral Treatment (Weeks)HCV GenotypeFollow-Up Period After SVR (Months)HCV RNA
Serum Load (vge/mL)*PBMCs
HCV RNA–Positive Strand (vge/μg)HCV RNA–Negative Strand
  • Abbreviations: F, female; HCV, hepatitis C virus; IFN, interferon alpha; IL-2, interleukin-2; IVDU, intravenous drug use; M, male; n.a., not applicable; NAH, nucleic acid hybridization; ND, not detectable; P-IFN, pegylated interferon alpha; PBMC, peripheral blood mononuclear cell; PHA, phytohemagglutinin; R, ribavirin; RT-PCR, reverse-transcription polymerase chain reaction; SVR, sustained virological response; vge, virus genome equivalents.

  • *

    The serum HCV RNA load was determined by real-time RT-PCR.

  • †PBMCs were stimulated with 5 μg/mL PHA and 20 IU/mL IL-2, the HCV RNA–positive strand was measured by nested RT-PCR/NAH, and the RNA-negative strand was measured by strand-specific RT-PCR/NAH as described in the Patients and Methods section.

  • Plasma was used as the HCV inoculum for the in vitro infection.

  • §

    The supernatant from PBMCs after a 72-hour culture with PHA and IL-2 was used as the HCV inoculum for the in vitro infection.

1/M48/MUnknownIFN (24)472<40300+
2/M§46/MIVDUIFN/R (24)1a6040050ND
3/M43/MIVDUIFN/R (48)2a6040<50ND
4/F§43/FIVDUIFN/R (48)Unknown6040NDn.a.
5/F§44/FInhalationP-IFN/R (48)1a42100100+
6/F59/FUnknownP-IFN/R (48)1b361.6 × 103NDn.a.
7/F48/FEndemicIFN/R (48)1a2450100+
8/F33/FInhalationP-IFN/R (48)124100<50+
9/F§45/FIVDUP-IFN/R (48)3a18<50<10+

Plasma from eight patients was used for in vitro infection experiments (see Table 1). In four cases (Table 1), supernatant from PBMCs cultured in the presence of PHA and IL-2 for 72 hours served as an inoculum for in vitro infection. Infectivity of both plasma and PBMC supernatants was examined in three cases (Table 1).

Lymphoid cells serving as in vitro HCV targets were isolated from healthy donors who had no clinical history or molecular indication of HCV exposure, as confirmed by RT-PCR/NAH assay2 and the absence of anti-HCV antibody by enzyme immunoassay (Abbott).

HCV Infection.

Monocyte-depleted lymphoid cells from a healthy donor were treated with 5 μg/mL PHA for 48 hours.9 Following stimulation, 7 × 106 lymphoid cells were exposed to 500 μL of test plasma in 6.5 mL of culture medium9 or 7 mL of supernatant from in vivo infected PBMCs, which were cultured as already indicated and described in detail previously.9 In parallel, the same number of target cells was exposed to 250 μL of plasma from a patient with CHC carrying HCV genotype 1b at 7.3 × 105 vge/mL as a positive control and to 500 μL of plasma from a healthy donor as a negative control (mock infection). Inocula were removed after 24 hours, and the cells were washed and cultured under alternating stimulation with PHA and IL-2 (phases A-D) for 14 days, as reported.9 It was previously established that after 14 days in culture, approximately 98% of the cells were T cells.9 Culture supernatants were collected at 1, 4 (phase A), 7 (phase B), 11 (phase C), and 14 (phase D) days post-infection and stored at −80°C, whereas cells recovered at 14 days post-infection (phase D) were cryopreserved for analysis.

Modification of HCV Infectivity by Anti-HCV E2, Anti-CD81, and IFN-α Treatments.

Neutralization of HCV was carried out by the incubation in duplicate of 250 μL of 48/F plasma with an anti-HCV E2 monoclonal antibody (mAb; AP33; provided by Dr. A. Patel, Institute of Virology, University of Glasgow, Glasgow, United Kingdom) for 1 hour at 37°C and then for 1 hour at 4°C prior to the addition of T cell targets. Infection was also inhibited by pre-incubation of T cell targets with anti-CD81 mAb (Pharmingen, San Diego, CA) before exposure to 44/F or 48/F plasma, as described previously.9 Appropriate isotype-matched mAbs were used in control experiments. To further reaffirm that active HCV replication was established in T cells, the cells were treated in duplicate with 1000 U/mL recombinant human IFNα 2b (Research Diagnostics, Flanders, NJ) at the time of HCV inoculation, as reported.9 The cells exposed to the same amount of 44/F or 48/F plasma, but not treated with IFNα, served as positive controls.

Ultracentrifugation.

To concentrate the virus and to recognize its general biophysical properties, plasma (5 mL) or pooled T cell culture supernatants (10 mL) collected after phases C and D, which were preclarified at 400g for 30 minutes in the presence of a protease inhibitor cocktail (1:200; Sigma), were layered onto 1-mL 30% sucrose cushions and centrifuged at 28,000g for 2.5 hours at 4°C in a TH641 rotor with a Sorvall Discovery 100SE ultracentrifuge (Mandel Scientific Co., Inc., Guelph, Canada). On the basis of the findings from preceding experiments (data not shown), 8.6 mL was removed from the top of each tube, and the remaining 2.4 mL was collected in 300-μL fractions (n = 8) for the evaluation of the HCV RNA content and sucrose density. In some instances, two 10-mL samples of pooled T cell culture supernatant collected after phases A to D of the same infection experiment were concentrated as indicated previously, and the resulting equivalent fractions were pooled and used for analysis.

RNA Extraction and RT-PCR/NAH Assays.

Total RNA was extracted with Trizol (Invitrogen Life Technologies, Burlington, Canada) from ∼1 × 107 cells (yielding ∼10 μg of RNA) or from 150 μL of the 300-μL sucrose fractions. RNA was reversely transcribed with Moloney murine leukemia virus reverse transcriptase (Invitrogen). HCV RNA–positive and HCV RNA–negative strands were detected with complementary DNA derived from 1 to 2 or 2 to 4 μg of total RNA, respectively, and primers, amplification conditions, and controls were exactly as reported in our previous studies.2, 7, 8 A water sample and a mock extraction were always included as contamination controls. Complementary DNA prepared from the mock infection served as an additional RT-PCR–negative control. Recombinant hepatitis C virus 5′-untranslated region E2 fragment (rHCV UTR-E2) served as a positive control.2 The specificity of the detection and validity of controls were routinely confirmed by NAH (i.e., Southern blot hybridization) with 32P-labeled rHCV UTR-E2 as a probe.2 The sensitivity of the RT-PCR assay for HCV RNA–positive strand detection was <10 vge/mL (<2 IU/mL) or 5 vge/μg of total RNA, whereas that for HCV RNA–negative strand detection was 25 to 50 vge/μg of total RNA.2 As a rule, HCV RNA–negative strand was tested only in RNA samples from T cells that had been found reactive for HCV RNA–positive strand.

Clonal Sequencing.

The nucleotide sequences of 5′-untranslated region (5′-UTR) HCV amplicons detected in cultured T cells exposed to 44/F or 48/F plasma were compared to those amplified from the respective plasma and from PBMCs isolated from the patients who provided those plasma samples. The amplicons were cloned with the TOPO-TA cloning system (Invitrogen). Ten clones for each polymerase chain reaction (PCR) product were sequenced in both directions with M13 primers and the ABI-Prism 7000 Sequence Detection System (Applied Biosystems, Streetsville, Canada). The resulting sequences were aligned with the help of Sequencher software version 4.7 (Gene Codes Corp., Ann Arbor, MI).7

Confocal Microscopy and Flow Cytometry.

To detect HCV nonstructural protein 5a (NS5a) in in vitro infected T cells and to estimate the number of positive cells, confocal immunofluorescent microscopy and fluorescence activated cell sorting (FACS) were applied. For confocal microscopy, infected cells were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, blocked with 10% normal goat serum, and double-stained with rat anti-tubulin (Chemicon International, Temecula, CA) and with either mouse anti-HCV NS5a mAb (Chemicon) or mouse isotype control.7 Then, cells were incubated with Cy2-labeled donkey anti-mouse and Cy5-labeled donkey anti-rat antibodies (both from Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). Cultured HCV-naïve T cells, Huh7 cells, and Huh7 cells carrying HCV AB12-A2FL replicon (provided by Dr. C. Richardson and Dr. J. Wilson, Ontario Cancer Institute, Toronto, Canada), stained as previously described, were used as controls. Cells were examined in a FluoView FV300 confocal system (Olympus America, Inc., Melville, NY). Approximately 1000 cells per preparation were examined, and NS5a-positive cells were counted. For FACS analysis, cells were fixed with 4% paraformaldehyde, permeabilized with 0.5% saponin, and double-stained with anti-NS5a mAb and anti-tubulin and then with Cy2-labeled and Cy5-labeled secondary antibodies. Cells were examined by flow cytometry with a FACSCalibur cytometer (BD Biosciences Pharmingen, San Jose, CA), and the results were analyzed with CellQuest Pro software (BD Biosciences).

Immunoelectron Microscopy (IEM).

To determine whether complete HCV virions can circulate in individuals with clinically apparent SVR and be secreted by de novo infected T cell cultures, 500 μL of unfractionated 48/F plasma and culture supernatant from T cells exposed to that plasma was incubated with anti-E2 AP33 mAb as reported.9 In addition, HCV RNA–positive fractions 4 and 7 (shown later in Fig. 5A), obtained after centrifugation of 44/F plasma over sucrose, were similarly incubated with anti-E2 mAb. Reacting particles were detected with anti-mouse IgG conjugated with 12-nm gold particles (Jackson ImmunoResearch) and counterstained with 1% phosphotungstic acid. Examinations were carried out in a JEM 1200 EX microscope (JEOL, Ltd., Tokyo, Japan).

Figure 5.

Sedimentation velocity of HCV RNA–reactive particles in sucrose. Plasma samples from individuals 44/F and 48/F with clinically apparent SVR and supernatants from cultured T cells exposed to this plasma were separately layered onto 1-mL sucrose cushions and ultracentrifuged. Eight 300-μL fractions collected from the bottom of each tube were assayed for HCV RNA–positive strands by reverse-transcription polymerase chain reaction/nucleic acid hybridization and for sucrose density. (A) 44/F plasma collected 3.5 years after SVR was achieved, (B) culture supernatant from T cells infected with HCV contained in 44/F plasma, (C) 48/F plasma collected 2 years after clinical SVR was achieved, and (D) culture supernatant from T cells infected with HCV contained in 48/F plasma. On the left side of each panel, the HCV RNA level is expressed in relative density units given by hybridization signals shown on the blot under each panel. On the blots, contamination and specificity controls are marked as in the legend to Fig. 1. Abbreviations: HCV, hepatitis C virus; rHCV UTR-E2, recombinant hepatitis C virus 5′-untranslated region E2 fragment; SVR, sustained virological response.

Figure 1.

Detection of HCV RNA–positive and HCV RNA–negative strands in T cell cultures after exposure to plasma from patients followed for up to 60 months after apparently complete clinical clearance of HCV. (A) HCV RNA–positive strand and (B) HCV RNA–negative (replicative) strand. HCV sRNA–positive and HCV sRNA–negative strands were used to confirm the specificity of the detections. Water instead of complementary DNA amplified in direct (D/W) and nested (N/W) reactions and a mock extraction (M) treated as test RNA were included as contamination controls. Positive samples showed the expected 244-bp amplicons. *A hybridization signal for 48/F with appropriate controls was overexposed to visualize its presence. Abbreviations: HCV, hepatitis C virus; HCV sRNA, synthetic hepatitis C virus RNA; rHCV UTR-E2, recombinant hepatitis C virus 5′-untranslated region E2 fragment.

Results

HCV Genome Expression in T Cells Exposed to Plasma from Individuals with Clinically Apparent SVR.

With the previously established system allowing for de novo infection and propagation of wild-type HCV in vitro,9 the infectivity of residual HCV occurring in the plasma of individuals with clinically apparent SVR was tested. Prestimulated lymphoid cells exposed to plasma from eight individuals and then cultured under alternate stimulation became reactive for HCV RNA–positive strand in seven of the cases (Fig. 1A). HCV RNA–negative strand, indicative of active virus replication, was evident in three of the seven (42.8%) cell cultures that were positive for the virus-positive strand. The HCV loads were estimated to be between 1 × 103 and 5 × 104 vge/107 cells.

HCV NS5a Protein in In Vitro Infected T Cells.

To determine whether expression of HCV RNA in in vitro infected T cells was accompanied by synthesis of viral protein, cells exposed to 44/F and 48/F plasma were examined for HCV NS5a protein by confocal microscopy. As illustrated in Fig. 2A, HCV NS5a occurred predominantly as granular intracytoplasmic deposits and at the plasma membrane of the positive cells. The percentages of NS5a-reactive cells enumerated under a confocal microscope were between 0.78% and 1.35%. A flow cytometric analysis gave comparable results of 1.05% to 1.52% of HCV NS5a protein–positive cells (Fig. 2B).

Figure 2.

Expression of HCV NS5a protein in cultured T cells exposed to plasma from individuals with SVR. (A) Cultured T cells after exposure to (A,B) 43/M plasma, (C-E) 48/F plasma, and (F) plasma from a patient with chronic hepatitis C (positive control) were double-stained with anti-HCV NS5a and anti-tubulin mAbs and analyzed by confocal microscopy. (G) Cultured T cells exposed to HCV and stained with an isotype control antibody served as a negative control. (H) Huh7 cells transfected with HCV AB12-A2FL full-length HCV replicon served as an additional positive control. Images were captured at 60× magnification. (B) Flow cytometric quantification of HCV NS5a–positive T cells exposed to plasma from individuals with therapy-induced SVR. Prestimulated cells were incubated with 44/F or 48/F plasma and cultured as described in the Patients and Methods section. T cells were stained with anti-NS5a mAb or isotype control antibody. Gates were set on the basis of isotype controls. Stimulated, uninfected T cells similarly stained with anti-NS5a mAb served as a negative control. Huh7 cells transfected with HCV AB12-A2FL replicon were used as a positive control. Percentages indicate positive cells. Abbreviations: HCV, hepatitis C virus; mAb, monoclonal antibody; NS5A, nonstructural protein 5a; SVR, sustained virological response.

Inhibition of HCV Infection by Anti-E2, Anti-CD81, and IFNα.

Pre-incubation of HCV present in 48/F plasma with anti-HCV E2 mAb, but not with an isotype control, neutralized the virus infectivity, as evidenced by the absence of HCV RNA–negative strand detection in the cells exposed to the treated inoculum (Fig. 3A). Similarly, pre-incubation of T cells with anti-CD81 mAb, but not with an isotype control mAb, blocked HCV replication in experiments in which 44/F or 48/F plasma was used as inoculum (Fig. 3A). Furthermore, treatment with recombinant IFNα prevented establishment of HCV replication in T cells, as shown in Fig. 3B.

Figure 3.

Effect of HCV neutralization with anti-E2 mAb, pre-incubation of target cells with anti-CD81 mAb, or their treatment with IFNα on the replication of HCV in cultured T cells. (A) HCV contained in 48/F plasma was pretreated with anti-HCV E2 mAb in duplicate (ex 1 and ex 2) or an isotype immunoglobulin control mAb and then incubated with T cells, or T cell targets were first incubated with anti-CD81 mAb or an appropriate isotype control and then exposed to 44/F plasma (ex 1) or 48/F plasma in duplicate (ex 1 and ex 2). (B) T cell targets were exposed to 44/F or 48/F plasma in the presence or absence (UT) of 1000 U/mL recombinant human IFNα 2b in duplicate (ex 1 and ex 2). RNA was analyzed for HCV RNA–positive and HCV RNA–negative strands after 14 days of culture. Specificity and contamination controls were as outlined in the legend to Fig. 1. Positive samples showed the expected 244-bp amplicons. Abbreviations: HCV, hepatitis C virus; HCV sRNA, synthetic hepatitis C virus RNA; IFNα, interferon alpha; mAb, monoclonal antibody; rHCV UTR-E2, recombinant hepatitis C virus 5′-untranslated region E2 fragment.

Unique HCV Variants in De Novo Infected T Cells.

To assess whether HCV replication in de novo infected T cells led to the appearance of variants distinct from those present in the plasma used for their inoculation, as observed in our previous study,9 5′-UTR amplicons from 44/F and 48/F plasma, PBMCs, and cultured T cells exposed to this plasma were cloned, bidirectionally sequenced, and compared. As shown in Fig. 4, clonal sequence analysis of the 147-bp fragment revealed a deletion at position 120 in all clones from PBMCs and in eight clones from de novo infected T cells and a C to T change at position 249 in all clones from PBMCs and in seven clones from T cells in comparison with the sequence amplified from 44/F plasma used as an inoculum; this indicated that unique HCV variants were present in the PBMCs and, importantly, that the same variants emerged in the T cells exposed to the plasma. On the other hand, sequencing of the cloned 5′-UTR amplicons derived from 48/F plasma, PBMCs, and cultured naïve T cells exposed to 48/F plasma revealed random single nucleotide polymorphisms, but no lymphoid cell–specific variants were identified.

Figure 4.

Nucleotide sequence alignment of clones derived from HCV 5′-UTR fragments amplified from 44/F and 48/F plasma, peripheral blood mononuclear cells, and cultured T cells exposed to that plasma. The 5′-UTR amplicons derived from each sample were cloned, and 10 randomly selected clones were sequenced bidirectionally. As a reference, the HCV genotype 1a sequence (GenBank accession number M67463) is shown on the top line. Nucleotides in the sequences identical to those of the reference are shown as dots, deletions are shown as dashes, and differences are shown as letters. Numbering of the nucleotides is according to the M67463 HCV genotype 1a sequence. Abbreviations: 5′-UTR, 5′-untranslated region; HCV, hepatitis C virus.

Physically Distinct HCV RNA–Reactive Particles Occur in Plasma and Culture Supernatants from T Cells Exposed to That Plasma.

To gain insight into the general biophysical properties of HCV RNA–reactive particles occurring in the plasma of individuals with SVR and those released into culture medium by T cells exposed to this plasma, samples of plasma and supernatants from the infected T cells were ultracentrifuged over sucrose, and the eight bottom fractions were collected. The analysis of 44/F plasma showed that HCV RNA–reactive particles occurred in fractions 1, 4, and 7, which corresponded to sucrose densities of 1.092, 1.064, and 1.027 g/mL, respectively, with apparent HCV RNA peak reactivity in fraction 4 (Fig. 5A). In the culture supernatant of T cells exposed to 44/F plasma, HCV RNA–positive particles were found in fractions 5 to 8 at densities between 1.024 and 1.011 g/mL with the peak RNA positivity in fraction 5 at a density of 1.024 g/mL (Fig. 5B). HCV RNA–reactive particles after centrifugation of 48/F plasma banded in fraction 4 at a sucrose density of 1.047 and in fractions 6 and 7 at densities of 1.018 to 1.013 g/mL (Fig. 5C), whereas those occurring in the culture supernatant of T cells exposed to this plasma were found in fractions 5 to 8, having a sucrose density of 1.027 to 1.019 g/mL, with the HCV RNA peak reactivity in fraction 5 at a density of 1.027 g/mL (Fig. 5D).

Infectivity of HCV Released by PBMCs After Clinically Apparent SVR.

To assess the infectivity of HCV found in in vivo infected PBMCs of individuals with clinically apparent SVR, PBMCs from four such patients followed for up to 5 years (Table 1) were stimulated with PHA and IL-2 for 72 hours, and the resulting supernatants were used as inocula to infect T cells. The data revealed that all cell cultures exposed to the PBMC supernatants acquired HCV RNA–positive strand reactivity, whereas HCV RNA–negative strand was detected in one of the cultures (Fig. 6).

Figure 6.

Detection of HCV RNA–positive and HCV RNA–negative strands in cultured T cells exposed to culture supernatants derived from in vivo HCV-infected, ex vivo stimulated peripheral blood mononuclear cells obtained from individuals with follow-up of up to 60 months after clinical sustained virological response was achieved. Contamination and specificity controls are marked as outlined in the legend to Fig. 1. Positive samples show the expected 244-bp amplicons. Abbreviations: HCV, hepatitis C virus; HCV sRNA, synthetic hepatitis C virus RNA; rHCV UTR-E2, recombinant hepatitis C virus 5′-untranslated region E2 fragment.

Ultrastructural Identification of HCV Particles in SVR Plasma and Culture Supernatants of De Novo Infected T Cells.

HCV particles were visualized with anti-E2 mAb by IEM (Fig. 7). Figure 7A-C depicts HCV virions detected in unfractionated (total) plasma obtained from patient 48/F 24 months after SVR was achieved. Figure 7D-F shows HCV particles found in HCV RNA–reactive fractions 4 (panel D) and 7 (panels E and F) after fractionation of 44/F plasma collected 42 months after SVR. As shown in Fig. 7G-K, HCV virion particles, either singly or as aggregates, were also detected in culture supernatants collected from T cells in vitro infected with virus carried in 44/F plasma. Particle sizes ranged from 50 to 75 nm in diameter.

Figure 7.

Ultrastructural features of HCV RNA–reactive particles in the plasma of individuals with clinical sustained virological response and in the culture supernatant obtained from T cells exposed to one of the plasma samples as visualized by immunogold staining with anti-E2 mAb. (A-C) HCV virion particles in unfractionated plasma of patient 48/F. HCV virions in (D) fraction 4 (sucrose density, 1.064 g/mL) and (E,F) fraction 7 (sucrose density, 1.027 g/mL) of 44/F plasma. (G-K) HCV particles in the supernatant of cultured T cells infected with HCV contained in 44/F plasma. (L) The same culture supernatant pool shown in panels G to K exposed to the isotype control instead of anti-E2 mAb. Preparations were counterstained with 1% phosphotungstic acid. Bars indicate 100 nm. Abbreviations: HCV, hepatitis C virus; mAb, monoclonal antibody.

Discussion

These findings provide in vitro evidence that trace quantities of HCV persisting in the circulation for a long time after therapeutically induced resolution of CHC can remain infectious. The transmission of HCV infection was exemplified by the detection of HCV RNA–negative strand and NS5a protein and the emergence of unique HCV variants in cultured T cells exposed to plasma from individuals with long-term follow-up after SVR. Furthermore, HCV replication in T cells was prevented following the neutralization of virus with anti-E2 mAb, blocking with anti-CD81 mAb, and treatment of the cells with recombinant human IFNα 2b. In addition, HCV residing in PBMCs after clinical resolution of infection was also found to be infectious. HCV virion particles specifically recognized by anti-E2 mAb were uncovered in plasma of these individuals and in the supernatant derived from de novo infected T cell cultures exposed to the plasma.

Similarly to CHC,7 circulating lymphomononuclear cells have been found to be the sites of active HCV replication in low-level infection continuing after resolution of hepatitis C, although ex vivo activation of the cells is usually required to uncover the virus presence.2, 8 Among circulating immune cells, CD4+ and CD8+ T lymphocytes, B cells, and monocytes have been identified to be infected to varying degrees with HCV, but with overall viral loads greater in CHC than in occult infection.7 As with PBMCs from occult infection, ex vivo stimulation of T cells affinity-purified from patients with low-level infection also significantly augments HCV replication, allowing for more ready detection of the residing virus, as a recent study showed.7

It was also previously uncovered that mitogen activation of normal human T lymphocytes predisposes the cells to infection by wild-type HCV of different genotypes (MacParland et al., unpublished observations, 2008).9 Productive replication of HCV in such treated cells, after exposure to either plasma from patients with CHC or culture supernatants from serial passage of wild-type HCV in T cell–enriched cultures, was shown by methods comparable to those used in the current study, that is, by detection of HCV RNA–negative strand, virus proteins (NS5a and E2), and HCV variants distinct from those occurring in respective inocula and by identification of secreted complete virions, as evidenced by isopycnic banding and ultrastructural examinations.9 In the present work, with the same HCV replication system and similar evaluation criteria, it became evident that the virus, occurring at low levels in three of eight plasma samples collected 2 to 5 years after SVR (cases 43/F, 44/F, and 48/F; Table 1), was able to establish active HCV replication in vitro. In addition, HCV derived from in vivo infected lymphoid cells obtained from one (case 44/F) of four patients induced de novo infection in the culture system. Taken together, these findings indicate that HCV carried by three of the nine individuals investigated in this study established infection in vitro, and this was confirmed by at least one criterion of active HCV replication, that is, the appearance of HCV RNA–negative strand. However, the infection initiated by the virus originating from two (cases 44/F and 48/F) of these three convalescent individuals was also confirmed by detection of viral protein, secretion of HCV RNA–reactive particles physically distinct from those occurring in inocula, and identification of complete virions by IEM.

HCV RNA–reactive particles occurring in 44/F and 48/F plasma and those secreted by T cells exposed to that plasma displayed different sedimentation profiles after ultracentrifugation over sucrose, and this implied distinct biophysical properties. This further supported the conclusion that the released virus originated from the de novo infection process. In general, although the plasma virions, as confirmed by IEM, predominantly banded at higher sucrose densities (1.047–1.064 g/mL), those released from in vitro infected T cells tended to sediment at densities not exceeding 1.027 g/mL. Because (1) HCV virions in the plasma of patients with CHC have been shown to have heterogeneous densities, particularly when associated with immunoglobulins and lipids,19–21 (2) the majority of plasma virions should or do originate from infected hepatocytes, and (3) the viral particles found in culture supernatants in the current study had low densities and were exclusively produced by T cells, a possibility exists that virions assembled in hepatocytes and lymphoid cells could be biophysically distinct because of association with different host proteins and/or lipids, giving, as a result, different sedimentation profiles. Studies have yet to compare the biochemical properties of plasma virus in CHC and in persistent low-level HCV infection and of virions produced by lymphoid cells in CHC and occult infection.

The infectivity of HCV traces persisting during a naturally acquired occult HCV infection has not yet been investigated. The present study, to our knowledge, is the first attempt in this regard. However, in early studies in chimpanzees, diluted plasma from a patient with acute posttransfusion hepatitis containing approximately 10 virions was capable of inducing infection, which was characterized by elevated serum alanine aminotransferase and liver inflammation.22 More recently, as few as 20 copies of HCV RNA prepared by the dilution of serum obtained during the pre-acute phase of hepatitis C of an infected chimpanzee has been demonstrated to cause HCV RNA–positive infection in the absence of alanine aminotransferase elevation.1 However, because the sensitivity of the PCR assay used for the detection of serum HCV RNA in the latter study appeared to be between 100 and 250 copies,1 the possibility remains that lower doses of HCV may also transmit infection in this model. Our present findings reveal that HCV circulating in some individuals with resolved hepatitis C is capable of inducing productive infection in vitro at doses of 20 to 50 copies. This can be interpreted as a strong indication of potential virus infectivity in vivo. In future studies, it would be of interest to determine the molecular mechanisms explaining why HCV circulating in some individuals, but not in others, is infectious to T cells despite comparable levels of virus being present.

In summary, the current study provides the first experimental evidence that HCV RNA detectable at low quantities for years after apparently complete resolution of CHC reflects the existence of traces of biologically competent virus that in some situations can retain infectivity.

Acknowledgements

The authors thank Dr. S. B. Reddy and D. King, a hepatology nurse specialist, from the Gastroenterology Clinic and Dr. J. S. McGrath from the General Hospital of the Health Science Centre (St. John's, Canada) for providing clinical samples. They also thank Dr. D. Richardson and J. Wilson from the Ontario Cancer Institute (Toronto, Canada) for supplying the HCV AB12-A2FL replicon and Dr. A. Patel from the Institute of Virology of the University of Glasgow (Glasgow, United Kingdom) for AP33 mAb against the HCV E2 protein.

Ancillary