Persistence of hepatitis C virus in patients successfully treated for chronic hepatitis C

Authors


  • See Editorial on Page 23

  • Conflict of interest: Nothing to report.

Abstract

It is unclear whether the current antiviral treatment for chronic hepatitis C virus (HCV) infection results in complete elimination of the virus, or whether small quantities of virus persist. Our study group comprised 17 patients with chronic HCV who had sustained virological response (SVR) after interferon/ribavirin treatment. Serum and peripheral blood mononuclear cells were collected 2 to 3 times at 3- to 6-month intervals starting 40 to 109 months (mean, 64.2 ± 18.5 months) after the end of therapy. In addition, lymphocyte and macrophage cultures were established at each point. In 11 patients, frozen liver tissue samples were available from follow-up biopsies performed 41 to 98 months (mean, 63.6 ± 16.7 months) after therapy. Presence of HCV RNA was determined by sensitive reverse-transcriptase polymerase chain reaction, and concentration of positive and negative strands was determined by a novel quantitative real-time reverse transcriptase polymerase chain reaction. Only 2 of 17 patients remained consistently HCV RNA negative in all analyzed compartments. HCV RNA was detected in macrophages from 11 patients (65%) and in lymphocytes from 7 patients (41%). Viral sequences were also detected in 3 of 11 livers and in sera from 4 patients. Viral replicative forms were found in lymphocytes from 2 and in macrophages from 4 patients. In conclusion, our results suggest that in patients with SVR after therapy, small quantities of HCV RNA may persist in liver or macrophages and lymphocytes for up to 9 years. This continuous viral presence could result in persistence of humoral and cellular immunity for many years after therapy and could present a potential risk for infection reactivation. (HEPATOLOGY 2005;41:106–114.)

Hepatitis C virus (HCV) is the major etiologic agent of parenterally transmitted non-A, non-B hepatitis.1, 2 In most infected patients, HCV persists indefinitely, leading to chronic hepatitis, cirrhosis, and hepatocellular carcinoma.3–5 The overall prevalence of anti-HCV in the United States is 1.8%, and approximately 2.7 million Americans carry the virus.6 Seeff et al.7 have recently reported on the long-term outcome in HCV-infected patients in several different transfusion studies. Twenty-five year follow-up of the HCV cases showed viremia with chronic hepatitis in 38%, viremia without chronic hepatitis in 39%, anti-HCV without viremia in 17%, and only 7% of patients had no residual markers of HCV infection. Thirty-five percent of HCV patients who underwent biopsies for biochemically defined chronic hepatitis displayed cirrhosis, representing 17% of all patients.7

Modern antiviral therapy is successful in approximately 50% of infected patients, resulting in clearance of HCV RNA from serum, which is usually accompanied by normalization of liver biochemical tests and improvement of liver histology.8, 9 Currently accepted criteria for sustained virological response (SVR) require the patient to remain HCV RNA negative in serum for 6 months after termination of treatment when tested with an assay with a sensitivity of at least 100 viral copies/mL.8 Long-term virological outcome in patients with SVR has been analyzed only recently. Recurrence of infection, defined as reappearance of HCV RNA in serum, was found to be below 2% at 1 to 4 years after induction of SVR,10, 11 although in one study in which patients were followed for 3.5 to 8.8 years, the relapse rate was as high as 8%.12 Obviously, because the length of follow-up of treated patients is still limited, durability of SVR may turn out to be lower over extended periods. It has also been reported that in 2% of SVR patients viral sequences could be detected in liver, and some of these patients ultimately experienced recurrent infection.13 However, the presence of HCV in other compartments except liver has not been analyzed so far in SVR patients.

HCV is not a strictly hepatotropic virus, and there is evidence that it can also replicate in peripheral blood mononuclear cells (PBMCs). The infected cells were reported to contain HCV RNA–negative strand, which is a viral replicative intermediate, and viral genomic sequences were often found to be distinct from those found in serum and liver.14–17 Furthermore, it was also reported that human T- and B-cell lines are capable of supporting HCV infection in vitro,18, 19 and some viral strains were found to be lymphotropic both in vitro and in vivo in infected chimpanzees.20 Within the population of PBMCs, the cells harboring replicating virus have been identified as belonging to T-cell and B-cell lineage and monocytes/macrophages.21–23 Although the presence of extrahepatic replication of HCV is becoming well recognized, whether it is affected by current antiviral treatment is unclear. Persistence of extrahepatic sites of HCV replication could potentially play a role in late recurrence after treatment.

In the current study, we provide evidence that in most patients with SVR, low-level HCV RNA can be detected in lymphocytes and monocytes/macrophages, and occasionally in liver and serum for up to 9 years after the end of therapy.

Abbreviations

HCV, hepatitis C virus; SVR, sustained virological response; PBMC, peripheral blood mononuclear cell; IFN, interferon; PBS, phosphate-buffered saline; RT, reverse transcription; PCR, polymerase chain reaction.

Patients and Methods

Biological Samples.

The study group comprised 17 randomly chosen patients with chronic HCV who responded to antiviral therapy and fulfilled the criteria for SVR by being negative for HCV RNA in serum at the end of therapy and 6 months afterwards. The other inclusion criteria were willingness to participate in the study and at least 3 years' follow-up after the end of treatment. Moreover, patients had to be persistently HCV RNA negative in serum by routine testing performed at least once a year after the end of therapy. During the 4- to 9-year follow-up, HCV RNA status was checked in all patients first by in-house assay (sensitivity limit, 500 viral copies/mL) and from 1997 on by Amplicor HCV version 2.0 (sensitivity limit, 135 viral copies/mL). Liver function tests, which were repeated every 6 months during the follow up, were consistently normal. Before treatment, the diagnosis of chronic hepatitis was based on the presence of HCV RNA and anti-HCV in serum, increased activity of serum aminotransferases, and on liver biopsy findings, which was performed in 15 patients. Five patients treated between 1993 and 1997 received interferon (IFN) α2b (Intron-A) monotherapy 3 to 5 million units for 24 to 72 weeks 3 times per week, whereas 12 patients received combination treatment consisting of intron-A 5 million units 3 times per week and ribavirin 1,200 mg daily (Table 1). The likely sources of infection were surgery with or without blood transfusions (10 cases), occupational exposure (3 cases), and past intravenous drug abuse (2 cases). In 2 patients, no risk factors could be identified. The study protocol conformed to institutional review board guidelines at participating institutions.

Table 1. Clinical and Histopathological Data on 17 Patients With Sustained Virological Response After Antiviral Treatment
Pt #Age at Treatment/SexAntiviral Treatment/Duration (wk)Liver Histopathology Before TreatmentLiver Histopathology After Treatment
Necroinflammatory Changes* (0–8)Fibrosis (0–6)Time After Treatment (mo)Necroinflammatory Changes (0–8)Fibrosis (0–6)
  • *

    Necroinflammatory score included periportal hepatitis (0–4), confluent necrosis (0–6), spotty necrosis (0–4), portal infiltrate (0–4) for a possible total score of 18.

148/FIFN 3MU/72736113
239/MIFN 5MU/R 48735410
346/MIFN 5MU/R 48215500
447/MIFN 3MU/24219811
543/MIFN 5MU/R 48825751
643/MIFN 5MU/R 4831NDNDND
760/FIFN 5MU/R 24736621
831/FIFN 5MU/R 48535011
941/MIFN 5MU/R 4861NDNDND
1037/MIFN 3MU/24328901
1140/MIFN 5MU/R 72636832
1241/MIFN 5MU/R 48744111
1350/MIFN 5MU/4894NDNDND
1448/MIFN 5MU/R 4874NDNDND
1547/FIFN 5MU/R 48NDNDNDNDND
1647/MIFN 5MU/R 24NDNDNDNDND
1746/FIFN 5MU/72726010

For the current study, serum and PBMC samples were collected 2 to 3 times at 3- to 6-month intervals from each patient starting 40 to 109 months (mean, 64.2 ± 18.5 months) after the end of therapy. In addition, frozen tissue samples kept at −80°C were available from 11 control liver biopsies performed 41 to 98 months (mean, 63.6 ± 16.7 months) after therapy. Liver histopathology was read in a blinded fashion as to clinical data by a trained pathologist (T.V.C.), using Hepatitis Activity Index scoring system proposed by Knodell et al.24

Control sera and PBMC samples were collected from 15 healthy anti-HCV–negative subjects. In 13 of 15 controls, sera and PBMCs were collected again after 1 month.

Isolation of PBMC and Lymphocyte and Macrophage Cultures.

PBMCs were isolated from blood by centrifugation over density gradient (Ficoll-Hypaque, Pharmacia, Kalamazoo, MI). Cells were washed 3 times with Mg2+ and Ca2+ free phosphate-buffered saline (PBS) (pH 7.4) and analyzed by microscopy and flow cytometry to make sure they were free of granulocytes and platelets. Isolated PBMCs were split into two parts: one part was stored unfractioned, and the remaining cells were used for establishing lymphocyte and macrophage cultures. To separate macrophages from other cells, PBMCs were resuspended in 10 mL RPMI 1640 medium (Gibco/BRL, Grand Island, NY) containing 10% fetal bovine serum, and incubated in plastic 25-cm2 culture flasks (Costar, Corning, NY). After incubation at 37°C for 4 hours, nonadhering cells were removed for subsequent mitogen stimulation whereas adhering cells (approximately 3–6 × 106 cells/flask) were washed 3 times with PBS and were maintained in RPMI 1640 with 10% fetal bovine serum. Half of culture medium was changed every 2 to 3 days, and the cultures were terminated after 2 weeks. Nonadhering cells were cultured with polyclonal activator by use of standard procedures.25 Briefly, approximately 107 cells were resuspended at 106/mL RPMI in the presence of phytohemagglutinin (10 μg/mL; Wellcome Diagnostics, Research Triangle Park, NC). After 72 hours of incubation, the cells were pelleted, washed 3 times with PBS, and stored at −80°C until final analysis.

RNA was extracted from cells and serum by means of a modified guanidinium thiocyanate-phenol/chloroform technique by using a commercially available kit (TRIZOL LS, Gibco/BRL). Liver tissue samples were homogenized before being extracted with TRIZOL and 1 μg, as determined by spectrophotometry, was used for the reverse transcriptase (RT) step. To maximize the sensitivity of HCV RNA detection, all RNA extracted from 1 mL serum was put into a single reaction.

RT-Polymerase Chain Reaction for the Detection of 5′UTR Sequences.

The assay has been previously described.26 Briefly, extracted RNA was incubated for 20 minutes at 42°C in a 30-μL reaction containing 50 pmol of the antisense primer (5′ TGA/GTGCACGGTCTACGAGACCTC 3′; nt 342-320), 1 × RT buffer (Gibco, Gaithersburg, MD), 5 mmol/L dithiothreitol, 5 mmol/L MgCl2, 1 mmol/L dNTP, and 20 units of Moloney murine leukemia virus RT (Gibco). After heating to 99°C for 10 minutes, 50 pmol of the sense primer (5′ A/GAC/TCACTCCCCTGTGAGGAAC; nt 35–55), 7 μL 10 × polymerase chain reaction (PCR) buffer II (Perkin Elmer, Norwalk, CT), and 2.5 units of Taq DNA polymerase (Perkin Elmer) were added, and the volume was adjusted to 100 μL. Amplification was run in a DNA Thermal Cycler 480 (Perkin Elmer) as follows: initial denaturing at 94°C for 4 minutes, followed by 50 cycles of 94°C for 1 minute, 58°C for 1 minute, and a final extension at 72°C for 7 minutes. Twenty microliters of the final product was analyzed by agarose gel electrophoresis and Southern hybridization with a 32P-labeled internal oligoprobe (5′ ACTGTCTTCACGCAGAAAGCGTC 3′; nt 57–79). As reported before, this assay was capable of detecting 10 genomic equivalent (Eq) molecules of the correct template but was not strand specific.27 The sensitivity of the assay was not affected by addition of total RNA extracted from 1 mL serum. Appropriate measures, described elsewhere,17, 27 were employed to prevent and detect contamination. Negative controls included PBMCs, lymphocyte, and macrophage cultures from uninfected subjects and normal sera.

Quantitative Strand-Specific Real-Time RT-PCR.

Strand specificity of our RT-PCR for the detection of 5′UTR HCV RNA negative and positive strands was ascertained by conducting cDNA synthesis at high temperature with the thermostable enzyme Tth. A detailed description of the Tth-based assay was published previously.17 In brief, the cDNA was generated in 20 μL of reaction mixture containing 50 pmol/L of sense primer, 1× RT buffer (Perkin Elmer), 1 mmol/L MnCl2, 200 μmol/L (each) dNTP, and 5 units of Tth (Perkin Elmer). After 20 minutes at 70°C, Mn2+ were chelated with 8 μL 10× ethyleneglycoltetra-acetic acid chelating buffer (Perkin Elmer), 50 pmol/L antisense primer was added, the volume was adjusted to 100 μL, and the MgCl2 concentration was adjusted to 2.2 mmol/L. For the detection of positive strands, the primers were added in reverse order. The amplification was performed in Perkin Elmer GenAmp PCR System 9600 thermocycler as follows: initial denaturing for 1 minute at 94°C, followed by 20 cycles of 94°C for 15 seconds, 58°C for 30 seconds, and 72°C for 30 seconds. Two microliters of the final product was directly added into the second round real-time PCR.

Real time PCR employed LightCycler FastStart DNA Master SYBR Green I (Roche, Indianapolis, IN) and nested primer set as recently described.28 Each amplification was followed by melting curve analysis to ensure that a single-size product was amplified and no significant primers–dimers were present. In addition, amplification products were run on agarose gel to confirm the correct product size. Amplification was run in LightCycler (Roche) as follows: initial denaturation and activation of enzyme for 10 minutes at 95°C, followed by 35 cycles of 95°C for 30 seconds, 55°C for 5 seconds, and 72°C for 30 seconds. This strand-specific assay was capable of detecting approximately 100 genomic Eq molecules of the correct strand while unspecifically detecting ≥ 107 to 108 genomic Eq of the incorrect strand. Figure 1A-D shows the performance of the assay on synthetic template: it detected 102 to 108 synthetic HCV RNA genomic molecules while maintaining linearity and remaining strand specific. The sensitivity and dynamics of this assay were not affected when total RNA extracted from 1 mL HCV RNA–negative serum was added into reaction (Fig. 1B). Figure 2 shows performance of our assay on RNA extracted from an autopsy liver of an HCV-infected patient. As can been seen, both positive and negative HCV RNA strands were detected, and the assay remained linear over the range of five 10-fold dilutions (from 10−1 to 10−5 μg liver RNA).

Figure 1.

Quantitative detection of HCV RNA negative and positive strands by real-time strand-specific assay. Reverse transcription was performed at 70°C using thermostable enzyme Tth, and the first round of PCR amplification was limited to 20 cycles. The second nested round used LightCycler FastStart DNA Master SYBR Green I (Roche). (A) Detection of positive-strand HCV RNA synthetic template. (B) Detection of positive-strand HCV RNA synthetic template in the presence of RNA extracted from 1 mL uninfected serum. (C) Detection of negative-strand HCV RNA synthetic template. (D) Specificity of the assay for the detection of negative-strand HCV RNA. Reverse transcription used positive-sense primer and serial dilution of positive-strand HCV RNA served as template.

Figure 2.

Quantitative detection of HCV RNA negative and positive strands by real-time strand-specific assay in liver tissue. Total RNA was extracted from an infected liver and serially diluted in water; 1 μg RNA extracted from uninfected liver was added to each dilution to keep the quantity of RNA constant. (A) Detection of HCV RNA–positive strand. Negative-antisense primer used in the reverse transcription step. (B) Detection of HCV RNA–negative strand: sense primer used in the reverse transcription step.

Viral genotypes were determined by sequencing of 5′UTR amplified by nested RT-PCR.17 Statistical analysis was performed using the SPSS version 9.0 package (SPSS, Inc., Chicago, IL). Means were compared by the nonparametric Mann-Whitney U test, and paired data were compared by Wilcoxon signed rank sum test.

Results

HCV RNA in Liver Tissue.

Follow-up liver biopsies were performed 41 to 98 months (mean, 63.6 ± 16.7 months) after the end of treatment in 11 of 17 patients. Importantly, all patients in whom follow-up biopsies were done showed improvement in necroinflammatory changes (mean, 5.5 ± 2.2 vs. 1.5 ± 1.4; P = .001 by Wilcoxon matched pairs test), and 9 patients also had improvement in liver fibrosis. Mean fibrosis score in all 11 patients with follow-up biopsy was significantly lower after treatment when compared with pretreatment values (2.5 ± 0.9 vs. 1.0 ± 0.9; P = .039). HCV RNA sequences were detected in liver samples from three patients (Fig. 3). Interestingly, the only 2 cases without improvement in fibrosis were also HCV RNA positive (cases 1 and 4 in Table 1). This result raises the possibility that the detection of viral sequences in the liver despite ostensible SVR could have prognostic value with relation to fibrosis.

Figure 3.

Detection of HCV RNA in liver tissue samples from 11 patients with SVR after treatment. The presence of viral sequences was determined by RT-PCR. Twenty microliters (20%) of the reaction mixture was fractionated on agarose, transferred to a nylon membrane by Southern blotting, and subsequently hybridized to a 32P-labeled probe. The amount of RNA loaded into each RT-PCR was 1 μg. Positive control (lane +) consisted of 3 × 102 genomic Eq of the synthetic strand mixed with 1 μg RNA extracted from liver tissue from an HCV-negative subject. Negative control (lane −) consisted of 1 μg RNA extracted from uninfected liver.

HCV RNA–positive livers were examined for viral RNA load by our real-time strain-specific assay. All three samples had low but quantifiable amounts of virus ranging from 180 genomic Eq/μg RNA in patients 10 to 245 and 466 genomic Eq in patients 4 and 1, respectively. However, HCV RNA–negative strand was not detected in any of the 3 liver samples. This result is not unexpected, because the viral-negative strand is typically 1- to –2-log lower than the positive strand,28, 29 and thus it was likely to have been below the sensitivity of the assay. In all 3 cases, the HCV genotype found in the liver was identical to that present in serum before initiation of treatment (Table 2).

Table 2. Presence of HCV RNA in Liver and Sequential Samples of Serum, PBMC and Lymphocyte and Macrophage Cultures in 17 Patients With Sustained Virological Response After Antiviral Treatment
IDHCV RNA in LiverTime After Treatment (mo)HCV RNA in SerumHCV RNA in PBMCsHCV RNA in LymphocytesHCV RNA in MacrophagesGenotype
1st*2nd3rd1st2nd3rd1st2nd3rd1st2nd3rdPretreatment§Posttreatment
  • Abbreviations: HCV, hepatitis C virus; ND, not done.

  • *

    1st, 2nd, and 3rd represent sequential samples collected at 3- to 6-month intervals.

  • Quantifiable by real-time RT-PCR.

  • HCV RNA–negative strand present.

  • §

    Determined in serum only.

  • Genotype could be determined.

1+61+1b1b
252+NDND++ND++ND1b1b
351+++1a1a
4+98+++++1a1a
555+1aND
6ND50ND+NDND++ND3a3a
764++++1a1a
851+++3a1a
9ND62NDNDNDND1bND
10+901b1b
1164NDNDNDND1bND
1240+1a1a
13ND109+++1b1b
14ND51++2a2a
15ND61NDNDND++ND1b1b
16ND73+++++++1b1b
1760ND+ND++NDND1a1a

HCV RNA in Sera.

Sera were collected starting 40 to 109 months after the end of antiviral treatment. All 17 patients were repeatedly nonreactive for HCV RNA by commercial tests (Amplicor 2.0, Roche). Whereas all initial sera were HCV RNA negative with our current in-house assay (sensitivity limit ≥ 10 genomic Eq/mL), 4 sera collected from 4 different patients at later times were positive (Table 2). All sera, which were positive by the RT-PCR/hybridization assay, were tested for RNA viral load by the positive-strand real-time PCR (sensitivity ≥ 102 genomic Eq/mL). Of 4 samples, only 1 contained quantifiable amounts of HCV RNA (208 genomic Eq/mL). Because the remaining 3 sera were negative when tested by the real-time RT-PCR, the quantity of HCV RNA must have been below the sensitivity limit of this assay and was therefore likely to be between 10 and 100 viral copies/mL.

HCV RNA in PBMCs and Lymphocyte and Macrophage Cultures.

Two of 17 studied patients were HCV RNA positive in unfractioned PBMCs on first testing, and this number increased to 9 when additional samples collected at 3- to 6-month intervals were analyzed (Table 2). The presence of viral sequences was confirmed in 4 cases by positive-strand real-time RT-PCR, and in one of these samples HCV RNA–negative strand was detected as well. The quantity of viral load varied from 166 to 560 genomic Eq per 106 cells, and the proportion of positive to negative HCV RNA strand in the single case where both were detected was 6.3.

We have previously found that mitogen stimulation of lymphocytes may enhance HCV replication,21 and similar observations have been made by other authors with respect to woodchuck hepatitis virus.30 We have also found that HCV positive and negative strands can be detected in cultured macrophages from infected patients21 and that cultured native human macrophages are susceptible to HCV infection in vitro.31 For these reasons, we analyzed the presence of viral RNA in phytohemagglutinin mitogen (PHA) stimulated lymphocytes and in cultured macrophages. As seen in Table 2, HCV RNA was detectable either in stimulated lymphocytes or cultured macrophages in 9 patients (53%) on the first testing, and 14 patients (82%) were positive in at least one culture when the analysis was repeated 1 to 2 times using cells collected at later times. In 16 of 29 positive lymphocyte and macrophage samples, there was enough viral RNA to allow for quantification of HCV RNA–positive strand, which ranged from 1.43 × 102 to 4.26 × 103 per 106 cells (mean, 9.51 × 102 genomic Eq/106 cells). HCV RNA–negative strand, which is a viral replicative form, was detected in 6 (21%) of 29 HCV RNA–positive cultures (Table 3). The ratio of positive to negative strand in the samples ranged from 2.1 to 11.3 (mean, 6.6), and is lower than the ratio reported for liver but similar to the ratio determined previously in HCV-infected macrophage cell cultures.28 Reactions detecting viral-negative strand were unlikely to represent false-positive results because nonspecific detection of the incorrect strand might be expected when the latter is present at a concentration of at least 107 to 108 genomic Eq/reaction. However, the concentration of HCV RNA in samples containing viral-negative strand was below 104 genomic Eq/reaction (Table 3).

Table 3. The Quantity and Ratio of Positive (+) and Negative (−) HCV RNA Strands in 2 Lymphocyte and 4 Macrophage Cell Cultures as Determined by Quantitative Real-Time RT-PCR
ID(+) Strand/106 Cells(−) Strand/106 Cells(+)/(−) Ratio
  1. Abbreviations: HCV, hepatitis C virus; PCR, polymerase chain reaction.

Lymphocyte culture   
 Pt 72.41 × 1032.60 × 1029.3
 Pt 164.26 × 1039.91 × 1024.2
Macrophage culture   
 Pt 23.09 × 1035.02 × 1026.2
 Pt 64.40 × 1022.15 × 1022.1
 Pt 137.13 × 1021.11 × 1026.4
 Pt 163.85 × 1033.42 × 10211.3

Overall, taking into account the results of all testing, including cell cultures as well as detection of HCV RNA in liver and serum, evidence of infection was found in 15 (88%) of 17 patients. In one patient, the only evidence of infection was HCV RNA presence in the liver tissue, whereas in the remaining patients viral sequences were detected in serum or cultured macrophages and lymphocytes. Thus, HCV RNA was detected in macrophages from 11 patients (65%) and in lymphocytes from 7 patients (41%); in 4 patients (24%), both cell types tested positive. Viral replicative forms were found in lymphocytes from 2 (12%) and in macrophages from 4 (24%) patients. Only 2 of 17 patients in the study remained consistently HCV RNA negative in all analyzed compartments; however, follow-up liver biopsy sample was unavailable for analysis in one of these cases. Because of limitations in clinical material available, repeated independent confirmation of positive results was limited to liver samples and 16 cell cultures. All 3 liver samples and 15 cultures (94%) were positive on repeated testing with our RT-PCR. However, 24 (53%) of 45 originally positive samples were confirmed by the less sensitive real-time quantitative RT-PCR (Table 2).

Comparison of 5′UTR sequences amplified from different compartments was done in 14 patients. For this purpose, respective RT-PCR products were cloned (TA cloning system, Invitrogen) and 2 to 7 individual clones were sequenced directly. In all 3 liver-positive cases, viral sequences amplified after therapy from liver tissue were identical to those found in pretreatment serum. Altogether, in 6 patients, all analyzed sequences were identical in the same patient. In 7 cases, there were small sequence differences, suggesting a possibility of viral evolution, and in one patient (patient 8), sequence differences were significant enough to qualify pretreatment and posttreatment viral variants as belonging to various genotypes. Figure 4 shows comparison of sequences amplified from different compartment in patients 2, 8, and 13. In case 2, the pretreatment serum-derived virus differed from all posttreatment sequences by two nucleotide substitutions (C to A at position 204 and G to A at position 243) and one nucleotide insertion (C between positions 126 and 127). Interestingly, all of these changes were previously identified by us in sequences amplified from extrahepatic sites,16, 32 and substitutions at positions 204 and 243 were also described by others.18, 33 The C to A change at position 204 and G to A change at position 243 were also found in patient 16 (not shown). In patient 13, pretreatment and posttreatment sequences differed by 3 to 4 substitutions (Fig. 4).

Figure 4.

Nucleotide sequence alignment of the 5′ UTR fragments of HCV recovered from pretreatment sera (S pre) and various posttreatment samples: sera (S), peripheral blood mononuclear cells (PBMC), lymphocytes cultured with PHA mitogen (PHA), and cultured macrophages (Macr) from patients 2, 8, and 13. (.), sequence identity; (-), gap introduced to preserve sequence integrity.

Importantly, the genotypes after treatment were compatible with the genotypes found in pretreatment serum, the only exception being patient 8. However, this patient had a history of intravenous drug use, and thus the genotype discordance may have been the result of fresh superinfection. Alternatively, there could have been different genotypes in PBMCs and serum compartments before treatment. However, no PBMCs from the pretreatment period were available for analysis.

All sera and PBMC samples as well as all lymphocyte and macrophage cultures from 15 control subjects were HCV RNA negative.

Discussion

Currently available antiviral therapy for chronic hepatitis C leads to SVR in approximately 50% of patients.8, 9 However, whether successful treatment results in sterilization, or whether low viral replication persists and is perhaps kept in check by cellular and humoral immune responses, is unclear. Thus, the observation that specific antibodies and cellular immune response may persist for 2 decades after spontaneous resolution of acute hepatitis C could imply continuous antigen stimulation.7, 34 The critical role of virus-specific CD4+ T cells in long-term control of the virus is suggested by the observation that loss of this specific T-cell response is followed by HCV recurrence in patients recovering from acute hepatitis C.35 Similarly, in the chimpanzee hepatitis C model, depletion of CD4+ T cells results in incomplete control of viremia associated with emergence of viral escape mutations.36

The current study provides evidence that HCV RNA can indeed persist for years in patients fulfilling the accepted criteria for SVR after treatment of chronic hepatitis C.8 Whereas HCV RNA was detectable in sera from only 4 of 17 patients, 14 patients harbored viral sequences in circulating lymphocytes or macrophages. This situation may be analogous to infection with hepatitis B virus, which also demonstrates lymphotropism.37, 38 Thus, hepatitis B virus DNA sequences were found in PBMCs over 5 years after complete clinical and serological recovery from acute viral hepatitis.39 Small quantities of hepatitis B virus DNA have been shown to persist in serum for decades after recovery from acute hepatitis, and these traces of virus maintain specific cytotoxic T-lymphocyte response.40 Also, in woodchuck hepatitis virus infection model, a lifelong persistence of scanty amounts of replicating virus occurs in both the liver and the lymphatic system after spontaneous resolution of experimental hepatitis.41

HCV RNA was detected in unfractioned PBMCs in 9 patients, and viral-negative strand was detected only in one. However, culture of lymphoid cells with PHA as well as culture of macrophages increased the number of patients harboring HCV sequences to 14, and in 6 of these patients, viral replicative forms were detected as well. This finding suggests that HCV replication may be more efficient in activated cells. PHA is nonspecific inducer of T-cell proliferation stimulating CD8+ and CD4+ cells via lectin binding to glycosylated T-cell receptor complex proteins, whereas culturing of macrophages on polystyrene delivers a powerful activation signal for the latter cells.42 Interestingly, it was shown that unspecific mitogen stimulation of lymphocytes facilitates detection of woodchuck hepatitis virus, which is becoming recognized as being an inherently lymphotropic virus.41 Another important factor aiding HCV RNA detection was repeated testing. Whereas 9 patients were positive in serum or lymphocyte and macrophage culture on testing of initial samples, this number increased to 14 when samples collected at 2 to 3 different points were analyzed. Because viral load was typically low and close to the detection limit of used assays, this intermittent detection could be explained by sampling differences related to stochastic phenomena.

Infection of monocytes/macrophages by HCV is not unexpected, because these cells are known to be permissive to a wide range of viruses, including some other flaviviruses,43 and many RNA and DNA viruses are lymphotropic.44 We have previously reported the presence of viral replicative forms in monocytes/macrophages from HCV-infected patients,21 and in subsequent studies we showed that native human macrophages are susceptible to HCV infection in vitro and that this infection may be facilitated by concomitant human immunodeficiency virus (HIV) infection.28 Whereas the mere presence of HCV RNA in phagocytic cells could come from virions entrapped inside these cells or adsorbed on their surface, presence of viral-negative strand in macrophage and lymphocyte cultures argues for the presence of genuine viral replication. HCV RNA–negative strand was detected only in a minority of lymphocyte and macrophage cultures, but it is likely that the strand-specific assays are not sensitive enough to detect low-level extrahepatic replication. Indeed, in several studies they were found to be at least 1 log less sensitive than standard RT-PCR.27, 29 Moreover, in cells supporting HCV replication, negative RNA strands are generally detected at levels lower than the levels of positive strands,27, 29 and the titer of positive strand HCV RNA in cultured cells was low to begin with. Interestingly, using a novel real-time quantitative PCR, we found the ratio of positive to negative strands in lymphocytes and macrophages may be lower than that found in the liver,27, 29 which suggests that replication dynamics in extrahepatic and hepatic sites may be different. However, it is also possible that the high ratio in the liver tissues is caused by contamination by circulating virions containing positive-strand HCV RNA.

Our findings of the continuing presence of HCV RNA years after ostensible successful treatment are compatible with the very recent findings by Pham et al.,45 who were able to amplify viral sequence from follow-up sera or PBMC in 11 of 11 SVR patients for up to 5 years after therapy. Most monocyte-derived dendritic cell cultures and mitogen-stimulated PBMCs contained HCV RNA–negative strand as well. All of those patients were HCV RNA negative in serum by commercial assays. However, liver tissue was not analyzed in this report, and pretreatment samples were not available for genotype comparison.

In summary, our results suggest that in patients with SVR after IFN or IFN/ribavirin therapy, small quantities of HCV RNA may persist in liver or PBMCs for up to 9 years. This continuous presence of HCV RNA could explain the phenomenon of relatively common persistence of humoral and cellular immunity for many years after supposed viral clearance and could present a potential risk for transmission or infection reactivation.

Ancillary