Compartmentalization of Hepatitis C Virus Genotype
http://www.100md.com
病菌学杂志 2005年第10期
Laboratoire de Virologie, H?pital Paul Brousse, 12 avenue Paul Vaillant-Couturier, Villejuif 94804, France
Unité Propre de Recherche Enseignement Supérieur (UPRES 3541), Paris XI, Centre Hépato Biliaire, H?pital Paul Brousse
Institut National de la Santé et de la Recherche Médicale (INSERM) U-481, Faculté de Médecine Xavier Bichat, 16 rue Henri Huchard, Paris 75018, France
ABSTRACT
Differences in hepatitis C virus (HCV) variants of the highly conserved 5' untranslated region (UTR) have been observed between plasma and peripheral blood mononuclear cells (PBMC). The prevalence and the mechanisms of this compartmentalization are unknown. Plasma and PBMC HCV variants were compared by single-strand conformation polymorphism (SSCP) and by cloning or by genotyping with a line probe assay (LiPA) in 116 chronically infected patients, including 44 liver transplant recipients. SSCP patterns differed between compartments in 43/109 analyzable patients (39%). Differences were significantly more frequent in patients with transplants (21/38 [55%] versus 22/71 [31%]; P < 0.01) and in those who acquired HCV through multiple transfusions before 1991 (15/20; 75%) or through drug injection (16/31; 52%) than in those infected through an unknown route (7/29; 24%) or through a single transfusion (5/29; 17%; P < 0.001). Cloning of the 5' UTR, LiPA analysis, and nonstructural region 5B sequencing revealed different genotypes in the two compartments from 10 patients (9%). In nine patients, the genotype detected in PBMC was not detected in plasma and was weak or undetectable in the liver in three cases. This genotypic compartmentalization persisted for years in three patients and after liver transplantation in two. The present study shows that a significant proportion of HCV-infected subjects harbor in their PBMC highly divergent variants which were likely acquired through superinfections.
INTRODUCTION
Hepatitis C virus (HCV) frequently leads to chronic infection contributing to liver cirrhosis and hepatocellular carcinoma. HCV is an enveloped, positive-stranded RNA virus that circulates in vivo as a population of closely related variants collectively referred to as a quasispecies. This quasispecies nature may be involved in viral persistence at the early stages of infection (12), allowing the emergence of immune escape mutants (11, 13). The infection of immune cells could be another viral mechanism of evasion contributing to host failure in eradicating the virus (32). The issue of extrahepatic replication of HCV is still debated. Detection of HCV RNA in extrahepatic compartments can be due to the adsorption of plasma variants (17, 31). However, the minus-stranded HCV RNA, the replicative intermediate, which is not detected in plasma, has been detected in peripheral blood mononuclear cells (PBMC) in many studies (4, 8, 10, 15, 18, 21, 22, 23, 25, 28, 37), but this is not sufficient to demonstrate that a complete and productive replication of HCV really occurs in these cells.
Another point supporting autonomous HCV replication in PBMC is the common finding in these cells of HCV variants differing from those circulating in serum (26, 31). The study of different blood cell subsets also revealed that quasispecies compositions can differ significantly between peripheral B and T lymphocytes, monocytes, and plasma (1). We have recently shown that this compartmentalization is a frequent phenomenon (10).
Compartmentalization of HCV variants has been mainly described for the hypervariable region of E2 envelope protein (HVR1). Unlike HVR, the 5' untranslated region (UTR) is highly conserved. The 5' UTR is the internal ribosomal entry site (IRES) essential for viral protein synthesis and virus replication. A small number of mutations are capable of affecting translation efficiency in vitro. The effects of these mutations depend on the cell type tested (16, 18), suggesting that the HCV IRES may adapt to different cell types. In addition, previous work has shown that variations in this region occurred in PBMC in vivo and may be associated with extrahepatic replication evidenced by the detection of the negative-stranded RNA (2, 17, 19). Assuming that HCV compartmentalization reflects differential cellular tropism, the 5' UTR appears to be the region of choice to investigate clinical and viral factors associated with this phenomenon because the small number of expected viral variants within quasispecies allows reliable intercompartment comparisons. HCV compartmentalization has not been studied over time. Its eventual persistence would support HCV lymphotropism. Furthermore, the conditions associated with the detection of compartmentalized variants have not been investigated in large cohorts of patients. The purpose of the present study was thus to determine the frequency and persistence of 5'-UTR variant compartmentalization and its possible relationships with epidemiological, clinical, and virological variables in a large series of chronically infected patients.
MATERIALS AND METHODS
Patients. One hundred sixteen patients managed at our institution were included in this study, which was approved by the local ethics committee. Inclusion criteria were the following: anti-HCV and plasma HCV RNA positivity, HBsAg and anti-human immunodeficiency virus (HIV) negativity, no previous treatment with interferon or ribavirin, no anticancer chemotherapy, and current alcohol intake <10 g/day. Forty-four patients (39%) had received transplants once for HCV-related cirrhosis a mean of 60 months (3 to 10 years) before inclusion in this study. They had no history of surgical complications or steroid-resistant episodes of rejection on standard initial immunosuppression (tacrolimus and steroids), and HCV infection had recurred on the graft in all. The duration of liver infection by HCV was estimated from the year of the transplantation, the first transfusion, or the first intravenous narcotic injection. The likely route of infection (intravenous narcotic use, transfusion, or unknown) was recorded. Among patients with transfusion contamination, those who received blood transfusions at least twice before 1991 were considered multitransfused. The degrees of hepatic fibrosis and of inflammatory activity were graded according to METAVIR scoring (3). The main characteristics of the patients are described in Table 1.
Sampling and HCV RNA amplification. Both plasma and PBMC were collected on the day of liver biopsy. In some cases, PBMC were positively sorted with antibodies against CD19 (B lymphocytes), CD14 (monocytes), CD4, and CD8, as described elsewhere (1). RNA was extracted from 140 μl of plasma by using the QIAmp Viral RNA kit (QIAGEN GmBH, Germany) and from PBMC or liver by using the RNeasy minikit (QIAGEN). To detect the HCV genome, one-fifth of the plasma or cellular RNA extract was subjected to reverse transcription (RT)-PCR with Ready-To-Go RT-PCR beads (Pharmacia Biotech, Uppsala, Sweden). Three microliters of the first PCR product was subjected to a second round of PCR using Ready-To-Go PCR beads, according to the manufacturer's instructions. A 250-base pair fragment (nucleotides 100 to 350) of the 5' UTR was amplified with a nested RT-PCR protocol with previously published primers (1). A 1,500-base pair fragment (nucleotides 100 to 1629) spanning the 5' UTR, core, E1, and HVR1 was amplified with primers published elsewhere (10).
HCV RNA was quantified in plasma and PBMC by using real-time PCR (Perkin-Elmer Corp.-Applied Biosystems, Foster City, CA). The primers, Taqman probes, and protocols used for HCV PCR have been described elsewhere (10, 40). Serial 10-fold dilutions of titered HCV RNA-containing plasma were run in each experiment and used to construct a standard curve.
Analysis of HCV quasispecies by cloning and sequencing. PCR products were cloned with the pGEM-T Easy Vector system (Promega Corporation, Madison, WI) and transformed in Escherichia coli JM109 competent cells (Promega). After overnight incubation at 37°C, insertion was checked on white colonies by PCR using the 5'-UTR inner primer pair. Sequences were read bidirectionally with inner primers on an ABI377 sequencer using the ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction kit with Amplitaq DNA polymerase (Perkin-Elmer Applied Biosystems Division, Foster City, CA). Ten to 20 clones from plasma and PBMC were sequenced for each patient.
Nucleotide sequences of cloned products were aligned by using the CLUSTAL W program version 1.5. Pairwise nucleotide distances were calculated by using the Kimura two-parameter method with a transition-to-transversion ratio of 2. Phylogenetic trees were constructed with the neighbor-joining algorithm, a cluster analysis method suited to sequences with high similarity scores, such as HCV quasispecies. Statistical evaluation of the obtained topology was performed with 1,000 replications of bootstrap sampling. Mantel's test was used to determine if sequences from a given compartment were genetically closer to each other than to sequences from other compartments (1). This test compares the Kimura two-parameter distance matrix with a compartment distribution matrix (Mc) of the same dimensions, where Mc(i, j) is equal to 0 if sequences i and j are from the same compartment, and Mc(i, j) is equal to 1 in other cases. The Pearson correlation coeficient r2 was computed for all pairs, excluding the diagonals of both matrices (observed r2). The null distribution was constructed by permuting the rows and columns of the Mc matrix 1,000 times. The number of times that the observed r2 was exceeded during the 1,000 permutations gave the P value of the observed correlation. Software written by Philippe Casgrain was used for this purpose (http://www.bio.montreal.ca/casgrain/fr/labo/R/index.html).
Analysis of HCV quasispecies by single-strand conformation polymorphism. HCV quasispecies of the 5' UTR were compared by the single-strand conformation polymorphism (SSCP) assay. This widely used mutation-scanning technique is based on the relation between the electrophoretic mobility of single-stranded DNA and its nucleotide sequence (33). Five microliters (one-fifth) of the amplicons was mixed with 10 μl of denaturing solution (950 ml/liter formamide and 400 mg/liter bromophenol blue), heated for 10 min at 95°C, and chilled on ice. Electrophoretic analysis was carried out with 7 μl of this mixture, using precast polyacrylamide gels (GeneGel Excel 12.5/24; Amersham-Pharmacia Biotech) on a thermostated device (Multiphor II system; Amersham-Pharmacia) at 5°C, 600 V, 25 mA, and 15 W. The gels were silver stained according to the manufacturer's instructions. A similar assay based on 5'-UTR polymorphism has been shown to enable the detection of minor variants representing 3% of the whole population (19). The performance of our 5'-UTR SSCP assay was assessed on 160 cloned and sequenced variants derived from the 250-bp 5'-UTR amplicons. We found that one, two, and three mutations yielded different SSCP patterns in 10%, 44%, and 69% of cases, respectively. SSCP patterns were always different when four or more mutations were present and were never different when the sequences were identical. In all cases, at least two independent PCR-SSCP assays were conducted. Plasma and PBMC SSCP patterns were considered distinct if they differed by at least one band in the two experiments. SSCP-defined compartmentalization was assessed blinded to clinical status.
5'-UTR genotyping. RNA was isolated from plasma and PBMC as described above. 5'-UTR genotyping was performed with the INNO-LiPA HCV II kit (Innogenetics, Ghent, Belgium) according to the manufacturer's instructions. Briefly, the 5' UTR is amplified with biotinylated primers. Biotin-labeled PCR products are reverse hybridized to specific probes attached to nitrocellulose strips. Development results in a purple precipitate that forms a positive line on the strip. The HCV type is deduced on the basis of the patterns of hybridizing bands by using the line probe assay (LiPA) interpretation chart (39).
Genotyping in NS5B. Amplification of the nonstructural region 5B (NS5B) and direct sequencing were performed as previously described (29) in some cases in order to confirm the 5'-UTR genotyping. NS5B sequences were genotyped by using referenced and annotated HCV sequences (http://hepatitis.ibcp.fr/).
Statistical analysis. We sought correlations between SSCP-defined compartmentalization and the following variables: liver transplantation, route of infection (single transfusion, multiple transfusion, intravenous drug use, or unknown), age, gender, genotype (genotype 1 versus others), quantitation of HCV RNA in plasma and lymphocytes, existence of hepatocellular carcinoma, and histological scores. Univariate analysis (chi-square test, t test, and correlation) and logistic regression were run on Statview software (Abacus Concepts, Berkeley, CA).
RESULTS
Detection of HCV RNA in plasma and PBMC. The 5' UTR was amplified by nested PCR in both PBMC and plasma from 109/116 patients (94%): 38 of the 44 transplanted patients and 71 out of the 72 immunocompetent individuals. Plasma and PBMC HCV RNAs from 90 subjects were available for quantification by real-time PCR. Mean viremia was 8.13 x 105 IU/ml (±1.3 x 106; range, 2 x 104 to 6 x 108), and the mean HCV RNA level in PBMC was 1.1 x 104 IU/μg of total RNA (±3.6 x 104; 2 x 102 to 1.6 x 105). PBMC HCV RNA levels did not correlate with HCV viremia. No significant relationship was found between plasma HCV RNA or PBMC HCV RNA levels and clinical or histological variables or the predominant genotype in plasma.
SSCP analysis. 5'-UTR SSCP patterns differed between PBMC and plasma in 43/109 patients (39%) (Fig. 1 and Tables 2 and 3). In 38/43 (88%) cases, at least one specific band was observed in PBMC and not in plasma; in the 5 other cases, all SSCP bands observed for PBMC were observed for plasma, but at least one additional and clear band was observed in plasma. The existence of an SSCP-defined compartmentalization was not related to HCV viremia or HCV RNA levels in PBMC. In univariate analysis, SSCP-defined compartmentalization was significantly more frequent in transplant patients than in nontransplant patients (21/38 [55%] versus 22/71 [31%]; P < 0.01). It was significantly related to the transmission mode (exact Fisher's test; P < 0.001), being more frequent in patients who had received multiple transfusions (15/20; 75%) or in those who had been infected through intravenous drug use (16/31; 52%) than in patients contaminated by an unknown route (7/29; 24%) or by a single transfusion (5/29; 17%) (Table 2). All these variables—transplantation, multiple transfusions, and intravenous drug use—appear significantly and independently correlated to the existence of compartmentalized 5'-UTR variants in multivariate analysis (Table 3).
Phylogenetic and LiPA analyses. Among the 43 patients for whom plasma and PBMC SSCP patterns were different, RNA was available in 28 cases for further analysis (Table 4). Ten of these 28 patients were analyzed through cloning and sequencing. These samples were quantified by real-time RT-PCR, and HCV RNA levels were at least 10,000 copies/ml in plasma and 2,000 copies/μg in PBMC. A quasispecies composition was always observed in plasma and PBMC. In 5 of these 10 patients (no. 1, 2, 3, 18, and 19), phylogenetic analysis showed that plasma and PBMC variants were separated by large genetic distances, exceeding 0.1 substitution/nucleotide in 4 of them. Phylogenetic analysis, including 5'-UTR reference sequences of different genotypes and subtypes, confirmed the presence of two genotypes or subtypes in these five patients. For patient 3, a mixed infection by subtypes 1a and 1b was further confirmed by cloning and sequencing of a larger fragment comprising one part of the IRES, the entire core and E1 regions, and the HVR (Fig. 2). In four out of these five mixed infections, one of the genotypes or subtypes found in PBMC was not detected in the plasma (up to 21 plasma clones were studied). These results were confirmed by repeated cloning of the remaining extracted RNAs. In these 5 patients infected by two different genotypes and in 1/5 others, Mantel's test was highly significant, supporting the nonrandom distributions of plasma and PBMC variants in these six cases. In addition, a 1,500-bp fragment encompassing the 5' UTR to the HVR was amplified in plasma and PBMC in three patients with similar 5'-UTR SSCP patterns in plasma and PBMC. In these three subjects, <4 mutations were observed between compartments in the 5' UTR (nucleotides 100 to 350), while the nonrandom distribution of plasma and PBMC HVR variants was evidenced through Mantel's test (Fig. 2).
In order to detect additional cases of genotypic compartmentalization, we performed 5'-UTR genotyping with the line probe assay in plasma and PBMC samples of all 28 SSCP-compartmentalized cases with available RNA. LiPA results were consistent with the cloning results in all 10 cases. SSCP, phylogenetic trees, and LiPA for four patients with genotypic compartmentalization are shown in Fig. 3. Four additional patients (no. 4, 16, 17, and 20) were found through line probe assay to have a genotype in PBMC not detectable in plasma. In two cases (no. 16 and 17), NS5B direct sequencing of plasma and PBMC HCV RNAs confirmed the results of 5'-UTR genotyping. Finally, in patient 15, 5'-UTR genotyping failed in PBMC but could be done through NS5B direct sequencing, showing HCV genotype 1b in PBMC and confirming genotype 3a in plasma.
Altogether, SSCP followed by cloning and LiPA showed that 10 patients had different genotypes in plasma and PBMC (genotypic compartmentalization) and that 9 of these patients had in their PBMC a strain assigned to a genotype different from the genotype detected in plasma.
To assess the possible sampling bias associated with starting low viral loads in the RT-PCR assay, we performed the following control. Two plasmas, one containing genotype 1 and one genotype 3, were quantified by our quantitative RT-PCR assay. Viral loads in both samples were adjusted by dilution. Serial dilutions of a 1/1 mixture of these samples were tested by 5'-UTR RT-PCR, SSCP, and cloning experiments. The SSCP experiment shows that the pattern was not altered from 2 x 105 to 2 x 103 copies/ml (Fig. 4), the latter sample corresponding to 103 copies/assay of each genotype. Cloning experiments on samples containing 2 x 105 and 2 x 103 copies/ml evidenced the same repartition of type 1 and 3 genotypes. This control showed that both genotypes could be detected with small amounts of starting material.
Longitudinal follow-up of genotypic compartmentalization. Longitudinal analysis (Fig. 5) of PBMC and plasma HCV RNA was possible for four of the seven patients with genotypic compartmentalization. In the two patients studied after liver transplantation (no. 19 and 20), plasma and PBMC harvested before transplantation (1 and 6 years before the index sample) showed the same genotypic compartmentalization. This demonstrates that the plasma genotype, but not the PBMC genotype, had infected the liver graft and that the compartmentalization had been stable for years. In the two nontransplant patients, samples were available 1 to 4 years after the index sample. One of the two plasma genotypes (genotype 3) disappeared from a patient (no. 2) who had received interferon during the interval, while genotype 1 remained detectable in both plasma and PBMC. In the last patient (no. 4), who was not treated, the same genotypic compartmentalization was found in the two samples.
Origin of compartmentalized genotypes. In order to identify the cell type harboring a specific genotype, we conducted a detailed analysis of CD19+, CD14+, CD4+, and CD8+ peripheral blood cell subsets in two immunocompetent patients with occult mixed infections (Fig. 6). Patient 3 had subtype 1a in serum and subtypes 1a and 1b in PBMC, but only B cells harbored the compartmentalized 1b subtype while monocytes and CD8 T cells harbored the circulating subtype 1a; this was evidenced by cloning and sequencing (Fig. 2). As evidenced by LiPA analysis, patient 4 had genotype 4 in plasma and genotypes 4 and 1 in PBMC. SSCP analysis showed that the compartmentalized type 1 was present in both B and CD8 T cells, while monocytes harbored the circulating genotype 4.
Frozen liver specimens were available for six patients. In three cases of genotypic compartmentalization (no. 4, 19, and 20), liver HCV genotypes were analyzed by line probe assay. The plasma genotype was found in the liver in all three cases, and a faint signal corresponding to the PBMC-specific genotype was detected in two cases (no. 4 and 20). The PBMC genotype was absent from the liver of patient 19. For this patient, liver RNA was also available from 6 years before the index case, and the same genotypic compartmentalization was observed. In patient 20, a liver-transplant patient contaminated through drug use, an additional genotype, absent from plasma and PBMC, was found in the liver.
DISCUSSION
The present study shows that a significant proportion of HCV-infected subjects harbor in their PBMC highly divergent variants which are not detectable in plasma and which are likely acquired through coinfections or superinfections. In 9 out of 109 analyzable patients, viral sequences detected in PBMC differed enough to be assigned to a genotype different from the predominant strain in plasma. This genotypic compartmentalization was confirmed by analysis of different regions and has persisted for years in three patients, including after liver transplantation in two.
The identification of this population with occult mixed infections is an additional argument supporting the still-debated issue of the infection of PBMC by HCV. Indeed, the main argument for such replication in PBMC is based on the detection of negative-stranded HCV RNAs. The detection of such replicative intermediates does not mean that the replication is complete (e.g., leading to infectious virions). Conversely, the assays for the detection of such intermediates are far from being standardized, and the absence of detection does not formally exclude HCV replication. In the present study, two lines of evidence support the replication of HCV in this compartment: the existence of significant differences between HCV variants found in plasma and PBMC and the persistence of these differences over time.
Given the large phylogenetic distances between different HCV genotypes (35), we can exclude the possibility that a genotype found only in PBMC was derived from the genotype found in plasma. This genotypic compartmentalization has been described in brain and blood monocytes sampled postmortem (37). The persistence of such compartmentalization over years, as shown here, rules out any possible bias due to PCR cross-contamination (9). More importantly, as PBMC have a limited life span, the most plausible explanation for the persistence of compartmentalized quasispecies is replication and propagation of HCV RNA within these cells. However, it remains possible that PBMC-specific strains with an extremely high affinity for PBMC could be produced by hepatocytes. In this view, the small amounts of compartmentalized genotypes which were detected in the livers of two patients could be due either to monocytes and lymphocytes infiltrating the liver or to replication by hepatocytes (6).
By using different assays, we have shown that the existence of compartmentalized 5'-UTR variants is a frequent phenomenon, detected in 31% of immunocompetent subjects and in 55% of liver transplant patients. The existence of compartmentalized 5'-UTR variants correlated independently with the contamination mode, either multiple transfusions or intravenous drug use, two conditions for which coinfections or superinfections are possible. It is obvious that the detection of distinct genotypes, but also of distinct strains belonging to the same genotype, was facilitated by the studied region: the very conserved 5' UTR. Coinfection by different genotypes or strains occurs logically when the patients are repeatedly contaminated from different subjects (by sharing of contaminated objects or multiple transfusions). In those subjects exposed to multiple strains, one strain usually becomes predominant in plasma (21, 41). The observed concomitant infection of the same host by two different genotypes, each replicating in a different compartment, strongly suggests that during coinfection with different strains, the outcome of this strain competition may depend on differences in cellular tropism. It is possible that adaptation of some HCV variants to PBMC allows them to avoid competitive exclusion by the dominant strain, at least in this compartment. However, coinfection is not the mechanism of HCV compartmentalization. Indeed, when established in a far more variable region, such as the HVR, HCV compartmentalization appeared as a constant phenomenon in patients contaminated by a single transfusion, as in others (10).
The frequency of compartmentalized 5'-UTR variants also correlated with liver transplantation. The hypothesis that the replication of HCV in PBMC may be favored by immunosuppression is supported by the common presence of HCV replication (negative-stranded RNA) in PBMC of HIV-infected subjects (17) or liver transplant recipients (36). Moreover HCV replication was demonstrated in human PBMC inoculated into SCID mice (5). The frequent compartmentalization of PBMC strains observed in liver transplant patients could be due, in this setting, to the difference between the durations of infection of liver and PBMC by HCV. Finally, we could not exclude an increased risk of nosocomial transmissions in this grafted population.
An important question is how the different HCV strains got into PBMC and persisted. In the present study, the peripheral blood cell subsets involved in genotypic compartmentalization were studied in two patients. Both had a specific genotype in B or T CD8+ lymphocytes. Although some of these cells could have a prolonged life span, the genotypes they harbored would have disappeared in the absence of autonomous HCV replication. This was not the case in any of the patients in whom genotypic compartmentalization could be monitored long term (up to 6 years). The line probe assay is capable of detecting very minor populations, but PBMC-specific genotypes were not detected in the serum of any of the patients harboring a specific genotype in PBMC. Since cell-specific virions were not detected in the circulation and the life span of the cells involved is limited, the question arises as to how naive PBMC become infected. One hypothesis is that compartmentalized variants are produced in plasma at an undetectable level by PBMC or hepatocytes and have a very strong affinity for PBMC. A second hypothesis is that cell-to-cell transmission occurs among PBMC without the need for cell-free virions. This is a classical viral strategy used to escape immunity (14); it is well documented in the case of HIV but remains to be shown for HCV. Whatever the respective roles of these mechanisms, the persistence of compartmentalized variants despite the limited PBMC life span strongly points to autonomous HCV replication and propagation within this compartment. This replication within PBMC could be associated with specific mutations reflecting an adaptation to PBMC. In this view, the previously described set of "adaptive" IRES mutations (24, 38) was detected in one patient whose PBMC were coinfected by subtypes 1a and 1b. Superinfection by different types could favor the opportunity for such a strain to overcome the previous strains, at least in PBMC. Further studies are required to analyze the translational efficiencies of compartmentalized IRES in hepatocyte or lymphocyte cell lines in the hypothesis of an IRES-driven cellular tropism of HCV, which has been suggested for strain H77 (24). However, plasma and PBMC 5'-UTR variants were apparently identical in 60% of the patients (while plasma and PBMC HVR variants were different in all studied subjects), suggesting that other viral regions than the IRES are involved in the cellular tropism of HCV.
Very few clinical studies have investigated the potential consequences of PBMC infection by HCV. This infection does not seem to play a key role as a viral reservoir (7, 20, 30) but could have important implications in the pathogenesis of HCV infection. The relation between HCV and B lymphoma is well established (42), and lymphotropic variants of HCV might be directly involved in lymphomagenesis. Recently, one study has pointed to the persistence of HCV RNA in PBMC of long-term responders who had no detectable plasma HCV RNA for years (34). The unexpectedly high frequency of mixed infections and the persistent replication/propagation of HCV within PBMC shown here appear particularly frequent in intravenous drug users, who are often exposed to repeated infections and to different HCV strains (27) and who are becoming the major population with acute and chronic hepatitis C. With this changing epidemiology, it may be important to investigate further in a larger group of naive and treated patients the clinical and therapeutic implications of this particular and relatively frequent type of HCV coinfection.
ACKNOWLEDGMENTS
This work was supported in part by the Agence Nationale de la Recherche sur le SIDA et les Hépatites (ANRS, study HEP 012) and the Fondation pour la Recherche Médicale.
Delphine Ducoulombier was supported by a grant from the Ministere de l'Education Nationale, de la Recherche et de la Technologie.
REFERENCES
Afonso, A. M., J. Jiang, F. Penin, C. Tareau, D. Samuel, M. A. Petit, H. Bismuth, E. Dussaix, and C. Feray. 1999. Nonrandom distribution of hepatitis C virus quasispecies in plasma and peripheral blood mononuclear cell subsets. J. Virol. 73:9213-9221.
Bain, C., A. Fatmi, F. Zoulim, J. P. Zarski, C. Trepo, and G. Inchauspe. 2001. Impaired allostimulatory function of dendritic cells in chronic hepatitis C infection. Gastroenterology 120:512-524.
Bedossa, P., T. Poynard, et al. 1996. An algorithm for the grading of activity in chronic hepatitis C. 24:289-293.
Boisvert, J., X. S. He, R. Cheung, E. B. Keeffe, T. Wright, and H. B. Greenberg. 2001. Quantitative analysis of hepatitis C virus in peripheral blood and liver: replication detected only in liver. J. Infect. Dis. 184:827-835.
Bronowicki, J., M. Loriot, V. Thiers, Y. Grignon, A. Zignego, and C. Brechot. 1999. Hepatitis C virus persistence in human hematopoietic cells injected into SCID mice. Hepatology 28:211-218.
Cabot, B., M. Martell, J. I. Esteban, M. Piron, T. Otero, R. Esteban, J. Guardia, and J. Gomez. 2001. Longitudinal evaluation of the structure of replicating and circulating hepatitis C virus quasispecies in nonprogressive chronic hepatitis C patients. J. Virol. 75:12005-12013.
Cabot, B., M. Martell, J. I. Esteban, S. Sauleda, T. Otero, R. Esteban, J. Guardia, and J. Gomez. 2000. Nucleotide and amino acid complexity of hepatitis C virus quasispecies in serum and liver. J. Virol. 74:805-811.
Caussin-Schwemling, C., C. Schmitt, and F. Stoll-Keller. 2001. Study of the infection of human blood derived monocyte/macrophages with hepatitis C virus in vitro. J. Med. Virol. 65:14-22.
Damen, M., H. Cuypers, H. Zaaijer, H. Reesink, W. Schaasberg, W. Gerlich, H. Niesters, and P. Lelie. 1996. International collaborative study on the second EUROHEP HCV-RNA reference panel. J. Virol. Methods 58:175-185.
Ducoulombier, D., A. M. Roque-Afonso, G. Di Liberto, F. Penin, R. Kara, Y. Richard, E. Dussaix, and C. Feray. 2004. Frequent compartmentalization of hepatitis C virus variants in circulating B lymphocytes and monocytes. Hepatology 39:817-825.
Erickson, A. L., Y. Kimura, S. Igarashi, J. Eichelberger, M. Houghton, J. Sidney, D. McKinney, A. Sette, A. L. Hughes, and C. M. Walker. 2001. The outcome of hepatitis C virus infection is predicted by escape mutations in epitopes targeted by cytotoxic T lymphocytes. Immunity 15:883-895.
Farci, P., A. Shimoda, A. Coiana, G. Diaz, G. Peddis, J. Melpolder, A. Strazzera, D. Chien, S. Munoz, A. Balestrieri, R. Purcell, and H. Alter. 2000. The outcome of acute hepatitis C predicted by the evolution of the viral quasispecies. Science 288:339-344.
Farci, P., A. Shimoda, D. Wong, T. Cabezon, D. De Gioannis, A. Strazzera, Y. Shimizu, M. Shapiro, H. J. Alter, and R. H. Purcell. 1996. Prevention of hepatitis C virus infection in chimpanzees by hyperimmune serum against the hypervariable region 1 of the envelope 2 protein. Proc. Natl. Acad. Sci. USA 93:15394-15399.
Johnson, D., and M. Huber. 2002. Directed egress of animal viruses promotes cell-to-cell spread. J. Virol. 76:1-8.
Lanford, R. E., D. Chavez, F. V. Chisari, and C. Sureau. 1995. Lack of detection of negative-strand hepatitis C virus RNA in peripheral blood mononuclear cells and other extrahepatic tissues by the highly strand-specific rTth reverse transcriptase PCR. J. Virol. 69:8079-8083.
Laporte, J., I. Malet, T. Andrieu, V. Thibault, J. J. Toulme, C. Wychowski, J. M. Pawlotsky, J. M. Huraux, H. Agut, and A. Cahour. 2000. Comparative analysis of translation efficiencies of hepatitis C virus 5' untranslated regions among intraindividual quasispecies present in chronic infection: opposite behaviors depending on cell type. J. Virol. 74:10827-10833.
Laskus, T., M. Radkowski, A. Piasek, M. Nowicki, A. Horban, J. Cianciara, and J. Rakela. 2000. Hepatitis C virus in lymphoid cells of patients coinfected with human immunodeficiency virus type 1: evidence of active replication in monocytes/macrophages and lymphocytes. J. Infect. Dis. 181:442-448.
Laskus, T., M. Radkowski, L. F. Wang, M. Nowicki, and J. Rakela. 2000. Uneven distribution of hepatitis C virus quasispecies in tissues from subjects with end-stage liver disease: confounding effect of viral adsorption and mounting evidence for the presence of low-level extrahepatic replication. J. Virol. 74:1014-1017.
Laskus, T., M. Radkowski, L. F. Wang, H. Vargas, and J. Rakela. 1998. The presence of active hepatitis C virus replication in lymphoid tissue in patients coinfected with human immunodeficiency virus type 1. J. Infect. Dis. 178:1189-1192.
Laskus, T., M. Radkowski, J. Wilkinson, H. Vargas, and J. Rakela. 2002. The origin of hepatitis C virus reinfecting transplanted livers: serum-derived versus peripheral blood mononuclear cell-derived virus. J. Infect. Dis. 185:417-421.
Laskus, T., L. Wang, M. Radkowski, H. Vargas, M. Nowicki, J. Wilkinson, and J. Rakela. 2001. Exposure of hepatitis C virus (HCV) RNA-positive recipients to HCV RNA-positive blood donors results in rapid predominance of a single donor strain and exclusion and/or suppression of the recipient strain. J. Virol. 75:2059-2066.
Lerat, H., F. Berby, M. A. Trabaud, O. Vidalin, M. Major, C. Trepo, and G. Inchauspe. 1996. Specific detection of hepatitis C virus minus strand RNA in hematopoietic cells. J. Clin. Investig. 97:845-851.
Lerat, H., S. Rumin, F. Habersetzer, F. Berby, M. A. Trabaud, C. Trepo, and G. Inchauspe. 1998. In vivo tropism of hepatitis C virus genomic sequences in hematopoietic cells: influence of viral load, viral genotype, and cell phenotype. Blood 91:3841-3849.
Lerat, H., Y. K. Shimizu, and S. M. Lemon. 2000. Cell type-specific enhancement of hepatitis C virus internal ribosome entry site-directed translation due to 5' nontranslated region substitutions selected during passage of virus in lymphoblastoid cells. J. Virol. 74:7024-7031.
Lin, L., J. Fevery, and S. Hiem Yap. 2002. A novel strand-specific RT-PCR for detection of hepatitis C virus negative-strand RNA (replicative intermediate): evidence of absence or very low level of HCV replication in peripheral blood mononuclear cells. J. Virol. Methods 100:97-105.
Maggi, F., C. Fornai, A. Morrica, M. L. Vatteroni, M. Giorgi, S. Marchi, P. Ciccorossi, M. Bendinelli, and M. Pistello. 1999. Divergent evolution of hepatitis C virus in liver and peripheral blood mononuclear cells of infected patients. J. Med. Virol. 57:57-63.
Mehta, S., A. Cox, D. Hoover, X. Wang, Q. Mao, S. Ray, S. Strathdee, D. Vlahov, and D. Thomas. 2002. Protection against persistence of hepatitis C. Lancet 359:1478-1483.
Mellor, J., G. Haydon, C. Blair, W. Livingstone, and P. Simmonds. 1998. Low level or absent in vivo replication of hepatitis C virus and hepatitis G virus/GB virus C in peripheral blood mononuclear cells. J. Gen. Virol. 79:705-714.
Morice, Y., D. Roulot, V. Grando, J. Stirnemann, E. Gault, V. Jeantils, M. Bentata, B. Jarrousse, O. Lortholary, C. Pallier, and P. Deny. 2001. Phylogenetic analyses confirm the high prevalence of hepatitis C virus (HCV) type 4 in the Seine-Saint-Denis district (France) and indicate seven different HCV-4 subtypes linked to two different epidemiological patterns. J. Gen. Virol. 82:1001-1012.
Muller, V., A. Maree, and R. De Boer. 2001. Release of virus from lymphoid tissue affects human immunodeficiency virus type 1 and hepatitis C virus kinetics in the blood. J. Virol. 75:2597-2603.
Okuda, M., K. Hino, M. Korenaga, Y. Yamaguchi, Y. Katoh, and K. Okita. 1999. Differences in hypervariable region 1 quasispecies of hepatitis C virus in human serum, peripheral blood mononuclear cells, and liver. Hepatology 29:217-222.
Oldstone, M. B. 1997. How viruses escape from cytotoxic T lymphocytes: molecular parameters and players. Virology 234:179-185.
Orita, M., H. Inahana, H. Kanazawa, K. Hayashi, and T. Sekiya. 1989. Detection of polymorphisms of human DNA by gel electrophoresis as single-strand conformation polymorphisms. Proc. Natl. Acad. Sci. USA 86:2766-2770.
Pham, T. N., S. A. MacParland, P. M. Mulrooney, H. Cooksley, N. V. Naoumov, and T. I. Michalak. 2004. Hepatitis C virus persistence after spontaneous or treatment-induced resolution of hepatitis C. J. Virol. 78:5867-5874.
Pybus, O., M. Charleston, S. Gupta, A. Rambaut, E. Holmes, and P. Harvey. 2001. The epidemic behavior of the hepatitis C virus. Science 292:2323-2325.
Radkowski, M., L. F. Wang, H. E. Vargas, J. Rakela, and T. Laskus. 1998. Detection of hepatitis C virus replication in peripheral blood mononuclear cells after orthotopic liver transplantation. Transplantation 66:664-666.
Radkowski, M., J. Wilkinson, M. Nowicki, D. Adair, H. Vargas, C. Ingui, J. Rakela, and T. Laskus. 2002. Search for hepatitis C virus negative-strand RNA sequences and analysis of viral sequences in the central nervous system: evidence of replication. J. Virol. 76:600-608.
Shimizu, Y. K., H. Igarashi, T. Kanematu, K. Fujiwara, D. C. Wong, R. H. Purcell, and H. Yoshikura. 1997. Sequence analysis of the hepatitis C virus genome recovered from serum, liver, and peripheral blood mononuclear cells of infected chimpanzees. J. Virol. 71:5769-5773.
Stuyver, L., A. Wyseur, W. van Arnhem, F. Hernandez, and G. Maertens. 1996. Second generation line-probe assay for hepatitis C virus genotyping. J. Clin. Microbiol. 34:2259-2266.
Takeuchi, T., A. Katsume, T. Tanaka, A. Abe, K. Inoue, K. Tsukiyama-Kohara, R. Kawaguchi, S. Tanaka, and M. Kohara. 1999. Real-time detection system for quantification of hepatitis C virus genome. Gastroenterology 116:636-642.
Vargas, H. E., T. Laskus, L. F. Wang, R. Lee, M. Radkowski, F. Dodson, J. J. Fung, and J. Rakela. 1999. Outcome of liver transplantation in hepatitis C virus-infected patients who received hepatitis C virus-infected grafts. Gastroenterology 117:149-153.
Zignego, A., C. Ferri, F. Giannelli, C. Giannini, P. Caini, M. Monti, M. Marrocchi, E. Di Pietro, G. La Villa, G. Laffi, and P. Gentilini. 2002. Prevalence of bcl-2 rearrangement in patients with hepatitis C virus-related mixed cryoglobulinemia with or without B-cell lymphomas. Ann. Intern. Med. 137:571-580.(Anne-Marie Roque-Afonso, )
Unité Propre de Recherche Enseignement Supérieur (UPRES 3541), Paris XI, Centre Hépato Biliaire, H?pital Paul Brousse
Institut National de la Santé et de la Recherche Médicale (INSERM) U-481, Faculté de Médecine Xavier Bichat, 16 rue Henri Huchard, Paris 75018, France
ABSTRACT
Differences in hepatitis C virus (HCV) variants of the highly conserved 5' untranslated region (UTR) have been observed between plasma and peripheral blood mononuclear cells (PBMC). The prevalence and the mechanisms of this compartmentalization are unknown. Plasma and PBMC HCV variants were compared by single-strand conformation polymorphism (SSCP) and by cloning or by genotyping with a line probe assay (LiPA) in 116 chronically infected patients, including 44 liver transplant recipients. SSCP patterns differed between compartments in 43/109 analyzable patients (39%). Differences were significantly more frequent in patients with transplants (21/38 [55%] versus 22/71 [31%]; P < 0.01) and in those who acquired HCV through multiple transfusions before 1991 (15/20; 75%) or through drug injection (16/31; 52%) than in those infected through an unknown route (7/29; 24%) or through a single transfusion (5/29; 17%; P < 0.001). Cloning of the 5' UTR, LiPA analysis, and nonstructural region 5B sequencing revealed different genotypes in the two compartments from 10 patients (9%). In nine patients, the genotype detected in PBMC was not detected in plasma and was weak or undetectable in the liver in three cases. This genotypic compartmentalization persisted for years in three patients and after liver transplantation in two. The present study shows that a significant proportion of HCV-infected subjects harbor in their PBMC highly divergent variants which were likely acquired through superinfections.
INTRODUCTION
Hepatitis C virus (HCV) frequently leads to chronic infection contributing to liver cirrhosis and hepatocellular carcinoma. HCV is an enveloped, positive-stranded RNA virus that circulates in vivo as a population of closely related variants collectively referred to as a quasispecies. This quasispecies nature may be involved in viral persistence at the early stages of infection (12), allowing the emergence of immune escape mutants (11, 13). The infection of immune cells could be another viral mechanism of evasion contributing to host failure in eradicating the virus (32). The issue of extrahepatic replication of HCV is still debated. Detection of HCV RNA in extrahepatic compartments can be due to the adsorption of plasma variants (17, 31). However, the minus-stranded HCV RNA, the replicative intermediate, which is not detected in plasma, has been detected in peripheral blood mononuclear cells (PBMC) in many studies (4, 8, 10, 15, 18, 21, 22, 23, 25, 28, 37), but this is not sufficient to demonstrate that a complete and productive replication of HCV really occurs in these cells.
Another point supporting autonomous HCV replication in PBMC is the common finding in these cells of HCV variants differing from those circulating in serum (26, 31). The study of different blood cell subsets also revealed that quasispecies compositions can differ significantly between peripheral B and T lymphocytes, monocytes, and plasma (1). We have recently shown that this compartmentalization is a frequent phenomenon (10).
Compartmentalization of HCV variants has been mainly described for the hypervariable region of E2 envelope protein (HVR1). Unlike HVR, the 5' untranslated region (UTR) is highly conserved. The 5' UTR is the internal ribosomal entry site (IRES) essential for viral protein synthesis and virus replication. A small number of mutations are capable of affecting translation efficiency in vitro. The effects of these mutations depend on the cell type tested (16, 18), suggesting that the HCV IRES may adapt to different cell types. In addition, previous work has shown that variations in this region occurred in PBMC in vivo and may be associated with extrahepatic replication evidenced by the detection of the negative-stranded RNA (2, 17, 19). Assuming that HCV compartmentalization reflects differential cellular tropism, the 5' UTR appears to be the region of choice to investigate clinical and viral factors associated with this phenomenon because the small number of expected viral variants within quasispecies allows reliable intercompartment comparisons. HCV compartmentalization has not been studied over time. Its eventual persistence would support HCV lymphotropism. Furthermore, the conditions associated with the detection of compartmentalized variants have not been investigated in large cohorts of patients. The purpose of the present study was thus to determine the frequency and persistence of 5'-UTR variant compartmentalization and its possible relationships with epidemiological, clinical, and virological variables in a large series of chronically infected patients.
MATERIALS AND METHODS
Patients. One hundred sixteen patients managed at our institution were included in this study, which was approved by the local ethics committee. Inclusion criteria were the following: anti-HCV and plasma HCV RNA positivity, HBsAg and anti-human immunodeficiency virus (HIV) negativity, no previous treatment with interferon or ribavirin, no anticancer chemotherapy, and current alcohol intake <10 g/day. Forty-four patients (39%) had received transplants once for HCV-related cirrhosis a mean of 60 months (3 to 10 years) before inclusion in this study. They had no history of surgical complications or steroid-resistant episodes of rejection on standard initial immunosuppression (tacrolimus and steroids), and HCV infection had recurred on the graft in all. The duration of liver infection by HCV was estimated from the year of the transplantation, the first transfusion, or the first intravenous narcotic injection. The likely route of infection (intravenous narcotic use, transfusion, or unknown) was recorded. Among patients with transfusion contamination, those who received blood transfusions at least twice before 1991 were considered multitransfused. The degrees of hepatic fibrosis and of inflammatory activity were graded according to METAVIR scoring (3). The main characteristics of the patients are described in Table 1.
Sampling and HCV RNA amplification. Both plasma and PBMC were collected on the day of liver biopsy. In some cases, PBMC were positively sorted with antibodies against CD19 (B lymphocytes), CD14 (monocytes), CD4, and CD8, as described elsewhere (1). RNA was extracted from 140 μl of plasma by using the QIAmp Viral RNA kit (QIAGEN GmBH, Germany) and from PBMC or liver by using the RNeasy minikit (QIAGEN). To detect the HCV genome, one-fifth of the plasma or cellular RNA extract was subjected to reverse transcription (RT)-PCR with Ready-To-Go RT-PCR beads (Pharmacia Biotech, Uppsala, Sweden). Three microliters of the first PCR product was subjected to a second round of PCR using Ready-To-Go PCR beads, according to the manufacturer's instructions. A 250-base pair fragment (nucleotides 100 to 350) of the 5' UTR was amplified with a nested RT-PCR protocol with previously published primers (1). A 1,500-base pair fragment (nucleotides 100 to 1629) spanning the 5' UTR, core, E1, and HVR1 was amplified with primers published elsewhere (10).
HCV RNA was quantified in plasma and PBMC by using real-time PCR (Perkin-Elmer Corp.-Applied Biosystems, Foster City, CA). The primers, Taqman probes, and protocols used for HCV PCR have been described elsewhere (10, 40). Serial 10-fold dilutions of titered HCV RNA-containing plasma were run in each experiment and used to construct a standard curve.
Analysis of HCV quasispecies by cloning and sequencing. PCR products were cloned with the pGEM-T Easy Vector system (Promega Corporation, Madison, WI) and transformed in Escherichia coli JM109 competent cells (Promega). After overnight incubation at 37°C, insertion was checked on white colonies by PCR using the 5'-UTR inner primer pair. Sequences were read bidirectionally with inner primers on an ABI377 sequencer using the ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction kit with Amplitaq DNA polymerase (Perkin-Elmer Applied Biosystems Division, Foster City, CA). Ten to 20 clones from plasma and PBMC were sequenced for each patient.
Nucleotide sequences of cloned products were aligned by using the CLUSTAL W program version 1.5. Pairwise nucleotide distances were calculated by using the Kimura two-parameter method with a transition-to-transversion ratio of 2. Phylogenetic trees were constructed with the neighbor-joining algorithm, a cluster analysis method suited to sequences with high similarity scores, such as HCV quasispecies. Statistical evaluation of the obtained topology was performed with 1,000 replications of bootstrap sampling. Mantel's test was used to determine if sequences from a given compartment were genetically closer to each other than to sequences from other compartments (1). This test compares the Kimura two-parameter distance matrix with a compartment distribution matrix (Mc) of the same dimensions, where Mc(i, j) is equal to 0 if sequences i and j are from the same compartment, and Mc(i, j) is equal to 1 in other cases. The Pearson correlation coeficient r2 was computed for all pairs, excluding the diagonals of both matrices (observed r2). The null distribution was constructed by permuting the rows and columns of the Mc matrix 1,000 times. The number of times that the observed r2 was exceeded during the 1,000 permutations gave the P value of the observed correlation. Software written by Philippe Casgrain was used for this purpose (http://www.bio.montreal.ca/casgrain/fr/labo/R/index.html).
Analysis of HCV quasispecies by single-strand conformation polymorphism. HCV quasispecies of the 5' UTR were compared by the single-strand conformation polymorphism (SSCP) assay. This widely used mutation-scanning technique is based on the relation between the electrophoretic mobility of single-stranded DNA and its nucleotide sequence (33). Five microliters (one-fifth) of the amplicons was mixed with 10 μl of denaturing solution (950 ml/liter formamide and 400 mg/liter bromophenol blue), heated for 10 min at 95°C, and chilled on ice. Electrophoretic analysis was carried out with 7 μl of this mixture, using precast polyacrylamide gels (GeneGel Excel 12.5/24; Amersham-Pharmacia Biotech) on a thermostated device (Multiphor II system; Amersham-Pharmacia) at 5°C, 600 V, 25 mA, and 15 W. The gels were silver stained according to the manufacturer's instructions. A similar assay based on 5'-UTR polymorphism has been shown to enable the detection of minor variants representing 3% of the whole population (19). The performance of our 5'-UTR SSCP assay was assessed on 160 cloned and sequenced variants derived from the 250-bp 5'-UTR amplicons. We found that one, two, and three mutations yielded different SSCP patterns in 10%, 44%, and 69% of cases, respectively. SSCP patterns were always different when four or more mutations were present and were never different when the sequences were identical. In all cases, at least two independent PCR-SSCP assays were conducted. Plasma and PBMC SSCP patterns were considered distinct if they differed by at least one band in the two experiments. SSCP-defined compartmentalization was assessed blinded to clinical status.
5'-UTR genotyping. RNA was isolated from plasma and PBMC as described above. 5'-UTR genotyping was performed with the INNO-LiPA HCV II kit (Innogenetics, Ghent, Belgium) according to the manufacturer's instructions. Briefly, the 5' UTR is amplified with biotinylated primers. Biotin-labeled PCR products are reverse hybridized to specific probes attached to nitrocellulose strips. Development results in a purple precipitate that forms a positive line on the strip. The HCV type is deduced on the basis of the patterns of hybridizing bands by using the line probe assay (LiPA) interpretation chart (39).
Genotyping in NS5B. Amplification of the nonstructural region 5B (NS5B) and direct sequencing were performed as previously described (29) in some cases in order to confirm the 5'-UTR genotyping. NS5B sequences were genotyped by using referenced and annotated HCV sequences (http://hepatitis.ibcp.fr/).
Statistical analysis. We sought correlations between SSCP-defined compartmentalization and the following variables: liver transplantation, route of infection (single transfusion, multiple transfusion, intravenous drug use, or unknown), age, gender, genotype (genotype 1 versus others), quantitation of HCV RNA in plasma and lymphocytes, existence of hepatocellular carcinoma, and histological scores. Univariate analysis (chi-square test, t test, and correlation) and logistic regression were run on Statview software (Abacus Concepts, Berkeley, CA).
RESULTS
Detection of HCV RNA in plasma and PBMC. The 5' UTR was amplified by nested PCR in both PBMC and plasma from 109/116 patients (94%): 38 of the 44 transplanted patients and 71 out of the 72 immunocompetent individuals. Plasma and PBMC HCV RNAs from 90 subjects were available for quantification by real-time PCR. Mean viremia was 8.13 x 105 IU/ml (±1.3 x 106; range, 2 x 104 to 6 x 108), and the mean HCV RNA level in PBMC was 1.1 x 104 IU/μg of total RNA (±3.6 x 104; 2 x 102 to 1.6 x 105). PBMC HCV RNA levels did not correlate with HCV viremia. No significant relationship was found between plasma HCV RNA or PBMC HCV RNA levels and clinical or histological variables or the predominant genotype in plasma.
SSCP analysis. 5'-UTR SSCP patterns differed between PBMC and plasma in 43/109 patients (39%) (Fig. 1 and Tables 2 and 3). In 38/43 (88%) cases, at least one specific band was observed in PBMC and not in plasma; in the 5 other cases, all SSCP bands observed for PBMC were observed for plasma, but at least one additional and clear band was observed in plasma. The existence of an SSCP-defined compartmentalization was not related to HCV viremia or HCV RNA levels in PBMC. In univariate analysis, SSCP-defined compartmentalization was significantly more frequent in transplant patients than in nontransplant patients (21/38 [55%] versus 22/71 [31%]; P < 0.01). It was significantly related to the transmission mode (exact Fisher's test; P < 0.001), being more frequent in patients who had received multiple transfusions (15/20; 75%) or in those who had been infected through intravenous drug use (16/31; 52%) than in patients contaminated by an unknown route (7/29; 24%) or by a single transfusion (5/29; 17%) (Table 2). All these variables—transplantation, multiple transfusions, and intravenous drug use—appear significantly and independently correlated to the existence of compartmentalized 5'-UTR variants in multivariate analysis (Table 3).
Phylogenetic and LiPA analyses. Among the 43 patients for whom plasma and PBMC SSCP patterns were different, RNA was available in 28 cases for further analysis (Table 4). Ten of these 28 patients were analyzed through cloning and sequencing. These samples were quantified by real-time RT-PCR, and HCV RNA levels were at least 10,000 copies/ml in plasma and 2,000 copies/μg in PBMC. A quasispecies composition was always observed in plasma and PBMC. In 5 of these 10 patients (no. 1, 2, 3, 18, and 19), phylogenetic analysis showed that plasma and PBMC variants were separated by large genetic distances, exceeding 0.1 substitution/nucleotide in 4 of them. Phylogenetic analysis, including 5'-UTR reference sequences of different genotypes and subtypes, confirmed the presence of two genotypes or subtypes in these five patients. For patient 3, a mixed infection by subtypes 1a and 1b was further confirmed by cloning and sequencing of a larger fragment comprising one part of the IRES, the entire core and E1 regions, and the HVR (Fig. 2). In four out of these five mixed infections, one of the genotypes or subtypes found in PBMC was not detected in the plasma (up to 21 plasma clones were studied). These results were confirmed by repeated cloning of the remaining extracted RNAs. In these 5 patients infected by two different genotypes and in 1/5 others, Mantel's test was highly significant, supporting the nonrandom distributions of plasma and PBMC variants in these six cases. In addition, a 1,500-bp fragment encompassing the 5' UTR to the HVR was amplified in plasma and PBMC in three patients with similar 5'-UTR SSCP patterns in plasma and PBMC. In these three subjects, <4 mutations were observed between compartments in the 5' UTR (nucleotides 100 to 350), while the nonrandom distribution of plasma and PBMC HVR variants was evidenced through Mantel's test (Fig. 2).
In order to detect additional cases of genotypic compartmentalization, we performed 5'-UTR genotyping with the line probe assay in plasma and PBMC samples of all 28 SSCP-compartmentalized cases with available RNA. LiPA results were consistent with the cloning results in all 10 cases. SSCP, phylogenetic trees, and LiPA for four patients with genotypic compartmentalization are shown in Fig. 3. Four additional patients (no. 4, 16, 17, and 20) were found through line probe assay to have a genotype in PBMC not detectable in plasma. In two cases (no. 16 and 17), NS5B direct sequencing of plasma and PBMC HCV RNAs confirmed the results of 5'-UTR genotyping. Finally, in patient 15, 5'-UTR genotyping failed in PBMC but could be done through NS5B direct sequencing, showing HCV genotype 1b in PBMC and confirming genotype 3a in plasma.
Altogether, SSCP followed by cloning and LiPA showed that 10 patients had different genotypes in plasma and PBMC (genotypic compartmentalization) and that 9 of these patients had in their PBMC a strain assigned to a genotype different from the genotype detected in plasma.
To assess the possible sampling bias associated with starting low viral loads in the RT-PCR assay, we performed the following control. Two plasmas, one containing genotype 1 and one genotype 3, were quantified by our quantitative RT-PCR assay. Viral loads in both samples were adjusted by dilution. Serial dilutions of a 1/1 mixture of these samples were tested by 5'-UTR RT-PCR, SSCP, and cloning experiments. The SSCP experiment shows that the pattern was not altered from 2 x 105 to 2 x 103 copies/ml (Fig. 4), the latter sample corresponding to 103 copies/assay of each genotype. Cloning experiments on samples containing 2 x 105 and 2 x 103 copies/ml evidenced the same repartition of type 1 and 3 genotypes. This control showed that both genotypes could be detected with small amounts of starting material.
Longitudinal follow-up of genotypic compartmentalization. Longitudinal analysis (Fig. 5) of PBMC and plasma HCV RNA was possible for four of the seven patients with genotypic compartmentalization. In the two patients studied after liver transplantation (no. 19 and 20), plasma and PBMC harvested before transplantation (1 and 6 years before the index sample) showed the same genotypic compartmentalization. This demonstrates that the plasma genotype, but not the PBMC genotype, had infected the liver graft and that the compartmentalization had been stable for years. In the two nontransplant patients, samples were available 1 to 4 years after the index sample. One of the two plasma genotypes (genotype 3) disappeared from a patient (no. 2) who had received interferon during the interval, while genotype 1 remained detectable in both plasma and PBMC. In the last patient (no. 4), who was not treated, the same genotypic compartmentalization was found in the two samples.
Origin of compartmentalized genotypes. In order to identify the cell type harboring a specific genotype, we conducted a detailed analysis of CD19+, CD14+, CD4+, and CD8+ peripheral blood cell subsets in two immunocompetent patients with occult mixed infections (Fig. 6). Patient 3 had subtype 1a in serum and subtypes 1a and 1b in PBMC, but only B cells harbored the compartmentalized 1b subtype while monocytes and CD8 T cells harbored the circulating subtype 1a; this was evidenced by cloning and sequencing (Fig. 2). As evidenced by LiPA analysis, patient 4 had genotype 4 in plasma and genotypes 4 and 1 in PBMC. SSCP analysis showed that the compartmentalized type 1 was present in both B and CD8 T cells, while monocytes harbored the circulating genotype 4.
Frozen liver specimens were available for six patients. In three cases of genotypic compartmentalization (no. 4, 19, and 20), liver HCV genotypes were analyzed by line probe assay. The plasma genotype was found in the liver in all three cases, and a faint signal corresponding to the PBMC-specific genotype was detected in two cases (no. 4 and 20). The PBMC genotype was absent from the liver of patient 19. For this patient, liver RNA was also available from 6 years before the index case, and the same genotypic compartmentalization was observed. In patient 20, a liver-transplant patient contaminated through drug use, an additional genotype, absent from plasma and PBMC, was found in the liver.
DISCUSSION
The present study shows that a significant proportion of HCV-infected subjects harbor in their PBMC highly divergent variants which are not detectable in plasma and which are likely acquired through coinfections or superinfections. In 9 out of 109 analyzable patients, viral sequences detected in PBMC differed enough to be assigned to a genotype different from the predominant strain in plasma. This genotypic compartmentalization was confirmed by analysis of different regions and has persisted for years in three patients, including after liver transplantation in two.
The identification of this population with occult mixed infections is an additional argument supporting the still-debated issue of the infection of PBMC by HCV. Indeed, the main argument for such replication in PBMC is based on the detection of negative-stranded HCV RNAs. The detection of such replicative intermediates does not mean that the replication is complete (e.g., leading to infectious virions). Conversely, the assays for the detection of such intermediates are far from being standardized, and the absence of detection does not formally exclude HCV replication. In the present study, two lines of evidence support the replication of HCV in this compartment: the existence of significant differences between HCV variants found in plasma and PBMC and the persistence of these differences over time.
Given the large phylogenetic distances between different HCV genotypes (35), we can exclude the possibility that a genotype found only in PBMC was derived from the genotype found in plasma. This genotypic compartmentalization has been described in brain and blood monocytes sampled postmortem (37). The persistence of such compartmentalization over years, as shown here, rules out any possible bias due to PCR cross-contamination (9). More importantly, as PBMC have a limited life span, the most plausible explanation for the persistence of compartmentalized quasispecies is replication and propagation of HCV RNA within these cells. However, it remains possible that PBMC-specific strains with an extremely high affinity for PBMC could be produced by hepatocytes. In this view, the small amounts of compartmentalized genotypes which were detected in the livers of two patients could be due either to monocytes and lymphocytes infiltrating the liver or to replication by hepatocytes (6).
By using different assays, we have shown that the existence of compartmentalized 5'-UTR variants is a frequent phenomenon, detected in 31% of immunocompetent subjects and in 55% of liver transplant patients. The existence of compartmentalized 5'-UTR variants correlated independently with the contamination mode, either multiple transfusions or intravenous drug use, two conditions for which coinfections or superinfections are possible. It is obvious that the detection of distinct genotypes, but also of distinct strains belonging to the same genotype, was facilitated by the studied region: the very conserved 5' UTR. Coinfection by different genotypes or strains occurs logically when the patients are repeatedly contaminated from different subjects (by sharing of contaminated objects or multiple transfusions). In those subjects exposed to multiple strains, one strain usually becomes predominant in plasma (21, 41). The observed concomitant infection of the same host by two different genotypes, each replicating in a different compartment, strongly suggests that during coinfection with different strains, the outcome of this strain competition may depend on differences in cellular tropism. It is possible that adaptation of some HCV variants to PBMC allows them to avoid competitive exclusion by the dominant strain, at least in this compartment. However, coinfection is not the mechanism of HCV compartmentalization. Indeed, when established in a far more variable region, such as the HVR, HCV compartmentalization appeared as a constant phenomenon in patients contaminated by a single transfusion, as in others (10).
The frequency of compartmentalized 5'-UTR variants also correlated with liver transplantation. The hypothesis that the replication of HCV in PBMC may be favored by immunosuppression is supported by the common presence of HCV replication (negative-stranded RNA) in PBMC of HIV-infected subjects (17) or liver transplant recipients (36). Moreover HCV replication was demonstrated in human PBMC inoculated into SCID mice (5). The frequent compartmentalization of PBMC strains observed in liver transplant patients could be due, in this setting, to the difference between the durations of infection of liver and PBMC by HCV. Finally, we could not exclude an increased risk of nosocomial transmissions in this grafted population.
An important question is how the different HCV strains got into PBMC and persisted. In the present study, the peripheral blood cell subsets involved in genotypic compartmentalization were studied in two patients. Both had a specific genotype in B or T CD8+ lymphocytes. Although some of these cells could have a prolonged life span, the genotypes they harbored would have disappeared in the absence of autonomous HCV replication. This was not the case in any of the patients in whom genotypic compartmentalization could be monitored long term (up to 6 years). The line probe assay is capable of detecting very minor populations, but PBMC-specific genotypes were not detected in the serum of any of the patients harboring a specific genotype in PBMC. Since cell-specific virions were not detected in the circulation and the life span of the cells involved is limited, the question arises as to how naive PBMC become infected. One hypothesis is that compartmentalized variants are produced in plasma at an undetectable level by PBMC or hepatocytes and have a very strong affinity for PBMC. A second hypothesis is that cell-to-cell transmission occurs among PBMC without the need for cell-free virions. This is a classical viral strategy used to escape immunity (14); it is well documented in the case of HIV but remains to be shown for HCV. Whatever the respective roles of these mechanisms, the persistence of compartmentalized variants despite the limited PBMC life span strongly points to autonomous HCV replication and propagation within this compartment. This replication within PBMC could be associated with specific mutations reflecting an adaptation to PBMC. In this view, the previously described set of "adaptive" IRES mutations (24, 38) was detected in one patient whose PBMC were coinfected by subtypes 1a and 1b. Superinfection by different types could favor the opportunity for such a strain to overcome the previous strains, at least in PBMC. Further studies are required to analyze the translational efficiencies of compartmentalized IRES in hepatocyte or lymphocyte cell lines in the hypothesis of an IRES-driven cellular tropism of HCV, which has been suggested for strain H77 (24). However, plasma and PBMC 5'-UTR variants were apparently identical in 60% of the patients (while plasma and PBMC HVR variants were different in all studied subjects), suggesting that other viral regions than the IRES are involved in the cellular tropism of HCV.
Very few clinical studies have investigated the potential consequences of PBMC infection by HCV. This infection does not seem to play a key role as a viral reservoir (7, 20, 30) but could have important implications in the pathogenesis of HCV infection. The relation between HCV and B lymphoma is well established (42), and lymphotropic variants of HCV might be directly involved in lymphomagenesis. Recently, one study has pointed to the persistence of HCV RNA in PBMC of long-term responders who had no detectable plasma HCV RNA for years (34). The unexpectedly high frequency of mixed infections and the persistent replication/propagation of HCV within PBMC shown here appear particularly frequent in intravenous drug users, who are often exposed to repeated infections and to different HCV strains (27) and who are becoming the major population with acute and chronic hepatitis C. With this changing epidemiology, it may be important to investigate further in a larger group of naive and treated patients the clinical and therapeutic implications of this particular and relatively frequent type of HCV coinfection.
ACKNOWLEDGMENTS
This work was supported in part by the Agence Nationale de la Recherche sur le SIDA et les Hépatites (ANRS, study HEP 012) and the Fondation pour la Recherche Médicale.
Delphine Ducoulombier was supported by a grant from the Ministere de l'Education Nationale, de la Recherche et de la Technologie.
REFERENCES
Afonso, A. M., J. Jiang, F. Penin, C. Tareau, D. Samuel, M. A. Petit, H. Bismuth, E. Dussaix, and C. Feray. 1999. Nonrandom distribution of hepatitis C virus quasispecies in plasma and peripheral blood mononuclear cell subsets. J. Virol. 73:9213-9221.
Bain, C., A. Fatmi, F. Zoulim, J. P. Zarski, C. Trepo, and G. Inchauspe. 2001. Impaired allostimulatory function of dendritic cells in chronic hepatitis C infection. Gastroenterology 120:512-524.
Bedossa, P., T. Poynard, et al. 1996. An algorithm for the grading of activity in chronic hepatitis C. 24:289-293.
Boisvert, J., X. S. He, R. Cheung, E. B. Keeffe, T. Wright, and H. B. Greenberg. 2001. Quantitative analysis of hepatitis C virus in peripheral blood and liver: replication detected only in liver. J. Infect. Dis. 184:827-835.
Bronowicki, J., M. Loriot, V. Thiers, Y. Grignon, A. Zignego, and C. Brechot. 1999. Hepatitis C virus persistence in human hematopoietic cells injected into SCID mice. Hepatology 28:211-218.
Cabot, B., M. Martell, J. I. Esteban, M. Piron, T. Otero, R. Esteban, J. Guardia, and J. Gomez. 2001. Longitudinal evaluation of the structure of replicating and circulating hepatitis C virus quasispecies in nonprogressive chronic hepatitis C patients. J. Virol. 75:12005-12013.
Cabot, B., M. Martell, J. I. Esteban, S. Sauleda, T. Otero, R. Esteban, J. Guardia, and J. Gomez. 2000. Nucleotide and amino acid complexity of hepatitis C virus quasispecies in serum and liver. J. Virol. 74:805-811.
Caussin-Schwemling, C., C. Schmitt, and F. Stoll-Keller. 2001. Study of the infection of human blood derived monocyte/macrophages with hepatitis C virus in vitro. J. Med. Virol. 65:14-22.
Damen, M., H. Cuypers, H. Zaaijer, H. Reesink, W. Schaasberg, W. Gerlich, H. Niesters, and P. Lelie. 1996. International collaborative study on the second EUROHEP HCV-RNA reference panel. J. Virol. Methods 58:175-185.
Ducoulombier, D., A. M. Roque-Afonso, G. Di Liberto, F. Penin, R. Kara, Y. Richard, E. Dussaix, and C. Feray. 2004. Frequent compartmentalization of hepatitis C virus variants in circulating B lymphocytes and monocytes. Hepatology 39:817-825.
Erickson, A. L., Y. Kimura, S. Igarashi, J. Eichelberger, M. Houghton, J. Sidney, D. McKinney, A. Sette, A. L. Hughes, and C. M. Walker. 2001. The outcome of hepatitis C virus infection is predicted by escape mutations in epitopes targeted by cytotoxic T lymphocytes. Immunity 15:883-895.
Farci, P., A. Shimoda, A. Coiana, G. Diaz, G. Peddis, J. Melpolder, A. Strazzera, D. Chien, S. Munoz, A. Balestrieri, R. Purcell, and H. Alter. 2000. The outcome of acute hepatitis C predicted by the evolution of the viral quasispecies. Science 288:339-344.
Farci, P., A. Shimoda, D. Wong, T. Cabezon, D. De Gioannis, A. Strazzera, Y. Shimizu, M. Shapiro, H. J. Alter, and R. H. Purcell. 1996. Prevention of hepatitis C virus infection in chimpanzees by hyperimmune serum against the hypervariable region 1 of the envelope 2 protein. Proc. Natl. Acad. Sci. USA 93:15394-15399.
Johnson, D., and M. Huber. 2002. Directed egress of animal viruses promotes cell-to-cell spread. J. Virol. 76:1-8.
Lanford, R. E., D. Chavez, F. V. Chisari, and C. Sureau. 1995. Lack of detection of negative-strand hepatitis C virus RNA in peripheral blood mononuclear cells and other extrahepatic tissues by the highly strand-specific rTth reverse transcriptase PCR. J. Virol. 69:8079-8083.
Laporte, J., I. Malet, T. Andrieu, V. Thibault, J. J. Toulme, C. Wychowski, J. M. Pawlotsky, J. M. Huraux, H. Agut, and A. Cahour. 2000. Comparative analysis of translation efficiencies of hepatitis C virus 5' untranslated regions among intraindividual quasispecies present in chronic infection: opposite behaviors depending on cell type. J. Virol. 74:10827-10833.
Laskus, T., M. Radkowski, A. Piasek, M. Nowicki, A. Horban, J. Cianciara, and J. Rakela. 2000. Hepatitis C virus in lymphoid cells of patients coinfected with human immunodeficiency virus type 1: evidence of active replication in monocytes/macrophages and lymphocytes. J. Infect. Dis. 181:442-448.
Laskus, T., M. Radkowski, L. F. Wang, M. Nowicki, and J. Rakela. 2000. Uneven distribution of hepatitis C virus quasispecies in tissues from subjects with end-stage liver disease: confounding effect of viral adsorption and mounting evidence for the presence of low-level extrahepatic replication. J. Virol. 74:1014-1017.
Laskus, T., M. Radkowski, L. F. Wang, H. Vargas, and J. Rakela. 1998. The presence of active hepatitis C virus replication in lymphoid tissue in patients coinfected with human immunodeficiency virus type 1. J. Infect. Dis. 178:1189-1192.
Laskus, T., M. Radkowski, J. Wilkinson, H. Vargas, and J. Rakela. 2002. The origin of hepatitis C virus reinfecting transplanted livers: serum-derived versus peripheral blood mononuclear cell-derived virus. J. Infect. Dis. 185:417-421.
Laskus, T., L. Wang, M. Radkowski, H. Vargas, M. Nowicki, J. Wilkinson, and J. Rakela. 2001. Exposure of hepatitis C virus (HCV) RNA-positive recipients to HCV RNA-positive blood donors results in rapid predominance of a single donor strain and exclusion and/or suppression of the recipient strain. J. Virol. 75:2059-2066.
Lerat, H., F. Berby, M. A. Trabaud, O. Vidalin, M. Major, C. Trepo, and G. Inchauspe. 1996. Specific detection of hepatitis C virus minus strand RNA in hematopoietic cells. J. Clin. Investig. 97:845-851.
Lerat, H., S. Rumin, F. Habersetzer, F. Berby, M. A. Trabaud, C. Trepo, and G. Inchauspe. 1998. In vivo tropism of hepatitis C virus genomic sequences in hematopoietic cells: influence of viral load, viral genotype, and cell phenotype. Blood 91:3841-3849.
Lerat, H., Y. K. Shimizu, and S. M. Lemon. 2000. Cell type-specific enhancement of hepatitis C virus internal ribosome entry site-directed translation due to 5' nontranslated region substitutions selected during passage of virus in lymphoblastoid cells. J. Virol. 74:7024-7031.
Lin, L., J. Fevery, and S. Hiem Yap. 2002. A novel strand-specific RT-PCR for detection of hepatitis C virus negative-strand RNA (replicative intermediate): evidence of absence or very low level of HCV replication in peripheral blood mononuclear cells. J. Virol. Methods 100:97-105.
Maggi, F., C. Fornai, A. Morrica, M. L. Vatteroni, M. Giorgi, S. Marchi, P. Ciccorossi, M. Bendinelli, and M. Pistello. 1999. Divergent evolution of hepatitis C virus in liver and peripheral blood mononuclear cells of infected patients. J. Med. Virol. 57:57-63.
Mehta, S., A. Cox, D. Hoover, X. Wang, Q. Mao, S. Ray, S. Strathdee, D. Vlahov, and D. Thomas. 2002. Protection against persistence of hepatitis C. Lancet 359:1478-1483.
Mellor, J., G. Haydon, C. Blair, W. Livingstone, and P. Simmonds. 1998. Low level or absent in vivo replication of hepatitis C virus and hepatitis G virus/GB virus C in peripheral blood mononuclear cells. J. Gen. Virol. 79:705-714.
Morice, Y., D. Roulot, V. Grando, J. Stirnemann, E. Gault, V. Jeantils, M. Bentata, B. Jarrousse, O. Lortholary, C. Pallier, and P. Deny. 2001. Phylogenetic analyses confirm the high prevalence of hepatitis C virus (HCV) type 4 in the Seine-Saint-Denis district (France) and indicate seven different HCV-4 subtypes linked to two different epidemiological patterns. J. Gen. Virol. 82:1001-1012.
Muller, V., A. Maree, and R. De Boer. 2001. Release of virus from lymphoid tissue affects human immunodeficiency virus type 1 and hepatitis C virus kinetics in the blood. J. Virol. 75:2597-2603.
Okuda, M., K. Hino, M. Korenaga, Y. Yamaguchi, Y. Katoh, and K. Okita. 1999. Differences in hypervariable region 1 quasispecies of hepatitis C virus in human serum, peripheral blood mononuclear cells, and liver. Hepatology 29:217-222.
Oldstone, M. B. 1997. How viruses escape from cytotoxic T lymphocytes: molecular parameters and players. Virology 234:179-185.
Orita, M., H. Inahana, H. Kanazawa, K. Hayashi, and T. Sekiya. 1989. Detection of polymorphisms of human DNA by gel electrophoresis as single-strand conformation polymorphisms. Proc. Natl. Acad. Sci. USA 86:2766-2770.
Pham, T. N., S. A. MacParland, P. M. Mulrooney, H. Cooksley, N. V. Naoumov, and T. I. Michalak. 2004. Hepatitis C virus persistence after spontaneous or treatment-induced resolution of hepatitis C. J. Virol. 78:5867-5874.
Pybus, O., M. Charleston, S. Gupta, A. Rambaut, E. Holmes, and P. Harvey. 2001. The epidemic behavior of the hepatitis C virus. Science 292:2323-2325.
Radkowski, M., L. F. Wang, H. E. Vargas, J. Rakela, and T. Laskus. 1998. Detection of hepatitis C virus replication in peripheral blood mononuclear cells after orthotopic liver transplantation. Transplantation 66:664-666.
Radkowski, M., J. Wilkinson, M. Nowicki, D. Adair, H. Vargas, C. Ingui, J. Rakela, and T. Laskus. 2002. Search for hepatitis C virus negative-strand RNA sequences and analysis of viral sequences in the central nervous system: evidence of replication. J. Virol. 76:600-608.
Shimizu, Y. K., H. Igarashi, T. Kanematu, K. Fujiwara, D. C. Wong, R. H. Purcell, and H. Yoshikura. 1997. Sequence analysis of the hepatitis C virus genome recovered from serum, liver, and peripheral blood mononuclear cells of infected chimpanzees. J. Virol. 71:5769-5773.
Stuyver, L., A. Wyseur, W. van Arnhem, F. Hernandez, and G. Maertens. 1996. Second generation line-probe assay for hepatitis C virus genotyping. J. Clin. Microbiol. 34:2259-2266.
Takeuchi, T., A. Katsume, T. Tanaka, A. Abe, K. Inoue, K. Tsukiyama-Kohara, R. Kawaguchi, S. Tanaka, and M. Kohara. 1999. Real-time detection system for quantification of hepatitis C virus genome. Gastroenterology 116:636-642.
Vargas, H. E., T. Laskus, L. F. Wang, R. Lee, M. Radkowski, F. Dodson, J. J. Fung, and J. Rakela. 1999. Outcome of liver transplantation in hepatitis C virus-infected patients who received hepatitis C virus-infected grafts. Gastroenterology 117:149-153.
Zignego, A., C. Ferri, F. Giannelli, C. Giannini, P. Caini, M. Monti, M. Marrocchi, E. Di Pietro, G. La Villa, G. Laffi, and P. Gentilini. 2002. Prevalence of bcl-2 rearrangement in patients with hepatitis C virus-related mixed cryoglobulinemia with or without B-cell lymphomas. Ann. Intern. Med. 137:571-580.(Anne-Marie Roque-Afonso, )