Article Text

Download PDFPDF
Occult infection of peripheral B cells by hepatitis C variants which have low translational efficiency in cultured hepatocytes
  1. Tony Durand1,
  2. Gaëtana Di Liberto1,2,
  3. Hélène Colman1,
  4. Anne Cammas3,
  5. Sebastien Boni1,
  6. Patrick Marcellin2,
  7. Annie Cahour4,
  8. Stephan Vagner3,
  9. Cyrille Féray
  1. 1INSERM U948 and EA 4271, Nantes University, France
  2. 2Centre de Recherche Biomédicale Bichat-Beaujon INSERM 3, France
  3. 3INSERM U563, Toulouse, France
  4. 4UPRES EA 2387, Groupe Hospitalier Pitié-Salpêtrière, Paris, France
  1. Correspondence to Dr Cyrille Féray, INSERM U948, Hôtel-Dieu, 9 quai Moncousu, Nantes 44, France; cyrille.feray{at}


Background Plasma hepatitis C virus (HCV) originates from hepatocytes. However, in certain subjects, B cells may harbour both plasma strains and occult HCV strains tha t are not detected in the plasma. The internal ribosome entry site (IRES) of these latter strains is mutated, suggesting that the efficiency of viral translation could drive the cellular tropism of HCV.

Aims To determine if the translational efficiency of IRES variants in cultured hepatocytes or B cells is correlated with their cellular tropism in vivo.

Methods The efficiency of IRES of 10 B cell-specific variants and nine plasma variants, isolated from six patients with compartmentalised variants in B cells, was estimated by bicistronic dual luciferase expression in hepatocyte cell types (Huh7), in primary cultured human hepatocytes (PCHs) and in two B cell lines (Raji and Daudi).

Results For each of the six subjects, the plasma IRESes were significantly and repeatedly more efficient than B cell IRESes in Huh7 (1.7±0.3 vs 0.7±0.2; p<0.01) and PCH cells. In B cell lines, B cell and plasma IRES had similar low efficiencies (0.8±0.1 vs 0.9±0.1; NS). For three subjects, two IRES variants from the same compartment could be analysed, and had the same efficiency in each cell type. Silencing the lupus antigen, a known IRES trans-acting factor, inhibited plasma IRES variants to a greater extent than B cell-specific IRESes.

Conclusions B cells can harbour occult variants that have a poor translational efficiency in hepatocytes, strongly suggesting their extra-hepatic origin and raising the hypothesis that competition between HCV variants with different IRESes is driven at a translational level in hepatic, as well as in extra-hepatic, sites.

  • HCV
  • hepatitis

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Significance of this study

What is already known about this subject?

  • The 5′ part of the HCV RNA is the internal ribosomal entry site (IRES).

  • Infection by hepatitis C virus (HCV) of cell types other than hepatocyte remains elusive.

  • HCV RNA is frequently detected in the B cell compartment.

  • Mutations of the IRES can be observed when B cells and plasma variants are compared.

What are the new findings?

  • Translational activity of B cell specific IRESes is poor in hepatocytes but not in B cell lines.

  • Translational activity of plasma IRESes depend more on lupus antigen than those of B cells IRESes.

  • Translational activity of mutants found in the same compartment is similar.

How might it impact on clinical practice in the foreseeable future?

  • HCV variants found in B cells are not adapted to hepatocytes, strongly suggesting their extra-hepatic origin.

  • Extra-hepatic sites of HCV replication could be sources for mutants which could escape the immune response or antiviral therapies.


Hepatitis C virus (HCV) can infect B lymphocytes.1–10 However, although detection in these cells of negatively stranded HCV RNA (the replicative intermediate) and viral proteins has been reported9 11–13 there is, as yet, no proof for a complete, productive replication of HCV in B cells or in other extra-hepatic sites.

The possibility of the extra-hepatic replication of HCV is suggested by the presence of non-random mutations when peripheral blood mononuclear cells (PBMCs) and plasma/liver HCV RNA are compared. This compartmentalisation occurs systematically within the hypervariable region of the envelope protein.14 It is also observed less frequently for the conserved 5′ non-coding region,15 which is the internal ribosome entry site (IRES) comprising the 341 nucleotides of the 5′ untranslated region (5′-UTR) and the first 30 nucleotides coding for the core protein. Differences in plasma and lymphocyte HCV IRES sequences are mainly observed in subjects who have received multiple transfusions or who are former intravenous drug users,15 16 a population known to be exposed to multiple infections.17–19 The sequencing of other regions of the viral genome has shown that strains with compartmentalised IRESes are globally distant strains. Liver RNA was available in these studies, and plasma variants originated from the liver in all cases. Finally, the compartmentalisation was stable during a number of years of patient follow-up.15 In the absence of any extra-hepatic replication of HCV, the only explanation for the compartmentalisation of HCV quasi-species should be genomic divergence between non-replicative HCV RNA persisting in cells with a prolonged lifespan, such as PBMCs, and strains replicating actively in hepatocytes and detected in the plasma. The function of these compartmentalised IRESes should not differ from the function of plasma/liver IRESes.

The efficiency of IRES variants detected in an extra-hepatic site has been measured in three studies, using bicistronic reporter experiments in different cell lines.6 20 21 In the first study, an H77 genotype 1 variant was found in peripheral blood mononuclear cells of a chimpanzee inoculated with H77 and also emerged after inoculation of a B cell line by H77.6 This variant had an efficient IRES activity in lymphoid cell lines.6 In the second study, a similar variant was detected in the dendritic cells of one subject and its efficiency was low in various cell types.20 A third team detected a similar variant in the brain of a patient; this variant was efficient in a hepatocytic cell line (Huh7) but not in a microglial cell line.21 These isolated observations, although discordant, suggested that IRES variability may affect cell tropism.

In the present study, we studied six patients whose peripheral B cells harboured HCV IRES variants distant from plasma variants, in order to determine if the mutations of the IRES variants could influence the cell specificity of HCV RNA translation.

Materials and methods

Study subjects (table 1)

Informed consent and local ethics committee approval was obtained from all six patients for the initial study. Blood samples showed that all patients were plasma HCV RNA-positive and HBsAg- and HIV-negative and none had received anti-HCV therapy. The six subjects of the study were selected from previously published series on the compartmentalisation of HCV.14–16 All six had a history of drug use and chronic hepatitis without cirrhosis at liver biopsy. The viral loads in plasma and B cells are reported in table 1. The negatively stranded HCV RNA had been detected in B cells of two of these six patients.

Table 1

Quantification of HCV RNA in plasma and B cells and detection of negative strands in B cells

Patients harboured HCV variants distant from plasma variants in B cells, as demonstrated by subcloning of the 5′ non-coding region (figure 1 and table 2), the E2 region and, in some cases, NS5A (data not shown). In order to avoid sampling bias (Lui SL, Science 1996), the cloning and sequencing of HCV RNA was performed from the same amount of HCV RNA and on more than 1000 copies. The plasma strains were similar to the strains amplified from liver RNA, as determined by direct sequencing in all six patients and cloning of the IRES region in three of them. Finally, in all these subjects, the same compartmentalisation was observed in blood samples harvested 1 year after the first sample (data not shown).

Figure 1

Phylogenetic trees of hepatitis C virus (HCV) variants infecting the six patients showing 23 clusters of two or more clones: B cell-specific clusters are framed with dotted lines, plasma variants are framed with continuous lines. Internal ribosome entry site (IRES) sequences which could not be tested are indicated by a circle. Note that in all subjects except #M, some B cell variants clustered with plasma variants.

Table 2

Alignment of the 19 analysed internal ribosome entry site (IRES) sequences showing mutations between plasma (P) and peripheral B cells (Ly). The numbering of mutated positions is deduced from H77 IRES

Amplification of IRES

Mononuclear cells were isolated from 14 ml EDTA-treated blood by centrifugation through a Ficoll density gradient. Recovered cells were washed three times and counted. Peripheral B lymphocytes were positively sorted with anti-CD19 magnetic beads with more than 95% purity. Cellular RNA (from 106 cells) and serum RNA (from 150 μl serum) were extracted as described previously.14

HCV IRES RNA was reverse-transcribed and amplified from both plasma and B cells. Complementary DNAs were synthesised from plasma (150 μl) and cellular RNAs (1 μg) using 100 mmol/l random hexanucleotides (Invitrogen, Cergy-Pontoise, France) and Moloney murine leukaemia virus reverse transcriptase (Invitrogen) for 1 h at 37°C. They were further amplified by nested PCR with high-fidelity Expand taq DNA polymerase (Roche, Meylan, France) using external primers S9 (5′-CCTGATGGGGGCGACACT-CC-3′; nt 9–28); 94 (5′-CTTGACGTCCTGTGGGCG-3′; nt 393–410) and internal primers S40BamH1 (5′-AGCGGGATCCATCACTCCCCTGTGAGG-3′; nt 26–52) and 54Pst1 (5′-GCGCCCTGCAGTTTTCTTTGAGGTTTAGG-3′; nt 354–383).

Plasmid construct

Bicistronic vectors were derived from the vector pRF as described previously.22 This plasmid named pRF contains the CMV and T7 promoters for transcription, the coding sequence for the firefly luciferase (Fluc), followed by a polylinker containing BamH1 and Pst1 and the Renilla luciferase gene (Rluc) in the pcDNA3.1Zeo(+) vector (Invitrogen) (figure 2). The tested IRESes were inserted in the previously described vector using Bamh1 and Pst1 restriction sites. We called pIRF, the inter-experimental construction containing a reference IRES from a genotype 1 strain with an A at positions 204 and 243.22 Both firefly and Renilla luciferases were measured for their respective luminometric activities in cell lysates with the Dual Luciferase Assay (Promega, Charbonnières, France) 24 h after transfection. IRES relative translational activity was represented by the Rluc/Fluc ratio and was normalised to the R/F ratio of pIRF measured in each experiment. Northern blot was tested on RNA extracted from Huh7 transfected with pIRF and other constructions as previously described.22

Figure 2

Transfection of bicistronic RNA transcribed in vitro through the T7 polymerase promoter which was partially capped, confirming the stronger efficiency of plasma IRES in Huh7 and Raji cell lines, the results obtained with the corresponding plasmid vectors. RNA translated from the pRF (empty vector) was transfected and firefly (F) luciferase but no Renilla (R) luciferase activity was observed.

In vitro transcription and translation

DNA templates for in vitro transcription were amplified using the Expand High Fidelity PLUS PCR system (Roche) on the bicistronic vectors, allowing amplification of the gene encoding F luciferase, followed by the tested IRES and the R luciferase coding region. The 5′ and 3′ primers used for this amplification were 5′-TAATACGACTCACTATAGG-3′, and 5′-T(35)TAGAAGGCACAGTCGAGG-3′, respectively. The 3′ primer contained a 35-nt polyT overhang, allowing transcription of an mRNA with a 35-nt polyA tail. The templates were purified using the QIAquick PCR purification kit (Qiagen, Courtaboeuf, France).

Capped polyadenylated mRNAs were synthesised using the MEGAscript T7 kit (Ambion, Austin, TX, USA) according to the manufacturer's instructions, with 5 ml of PCR product template. 6 mmol/l cap analogue M7G(5′)ppp(5′)G was added to cap mRNAs, and the GTP solution was used at a 1:5 dilution. The mix was incubated for 4 h at 37°C and DNA was removed by digestion for 15 min with TURBODNase. All RNAs were purified using the MEGAclear kit (Ambion, Austin, Texas, USA) to remove unincorporated nucleotides and cap analogue. All RNA transcripts were quantified by absorbance at 260 nm and their integrity was checked by northern blot as previously described.23 In vitro transcribed RNAs were also translated in rabbit reticulocyte lysates (Promega) as described previously.24

Cell cultures

We tested one hepatocytic cell line, HUH7, and two human B cell lymphoma cell lines, Raji and Daudi. HUH7 cells were maintained in Dulbecco's modified Eagle medium supplemented with 10% fetal calf serum, 1% penicillin–streptomycin, 1% l-glutamine and, for HepG2, 1% non-essential amino acids. Primary cultured human hepatocytes (PCHs) (monolayers) were supplied by Biopredic (Rennes, France) and grown according to a previously described protocol.25 Non-adherent cell-lines were maintained with RPMI 1640 medium with 10% fetal calf serum, 1% l-glutamine, 1% penicillin–streptomycin, 1% non-essential amino acids and 1% sodium pyruvate. Raji and Daudi cell lines were purchased from the American Type Culture Collection and the Huh7 cell line was a generous gift from Christine Neuveut (Pasteur Institute, Paris, France). All medium components were purchased from Sigma Aldrich (Fallavier, France). Cells were passaged twice weekly at the appropriate dilution for exponential growth and incubated at 37°C in a 5% CO2 atmosphere. Transfections were carried out in triplicate if viability was >90%. Adherent cells were transfected at 70% confluency. One day before transfection, 2×105 adherent cells (Huh7, HepG2) were seeded in 24-well plates. For DNA transfection, 2×106 Raji or Daudi cells were submitted to electroporation (Biorad, 260 V, 950 μF). Lipofectamine 2000 (Invitrogen) was used for the transfection of PCH and Huh7 cells. For RNA transfection, 2 μg mRNA was electroporated with nucleofector kit V (Amaxa, Koln, Germany) for Raji cells and with lipofectamine 2000 for Huh7 and PCH cells. In each experiment, pIRF (genotype 1 with a C204A; G243A) was included to normalise the different experiments.

Silencing F luciferase and the lupus antigen

To search for RNA transcribed from a cryptic promoter located within the tested IRES, an siRNA specifically targeting F luciferase, which was the upstream reporter gene (pGL2 luciferase control duplex firefly; Eurogentec, Angers, France) was cotransfected with the constructs. The siRNA directed against lupus antigen (LA) (ON-Targetplus SMARTpool SSB) and control siRNA (SiControl Non Targeting #4) were purchased from Dharmacon and all experiments were carried out as described previously.26 Western blots were performed using standard protocols with a polyclonal antibody against La/SSB (D-20; Santa Cruz Biotechnology, Santa Cruz, California, USA) and normalisation by actin (Santa Cruz Biotechnology). The siRNAs were transfected using lipofectamine.

Sequence analysis

Nucleic acid sequences were aligned using MEGA software version 4 (Tamura, Dudley, Nei, Kumar, 2007). The IRES sequences were compared to consensus sequences of genotypes 1–6 and also to that of H77. Analysis of quasi-species clustering in relation to their original compartment (plasma or B cells) was based on Mantel's test, as described previously.5

Statistical analysis

IRES efficiency was determined in triplicate and in two to three independent experiments. Each measure was normalised to the efficiency of the reference pIRF, also performed in triplicate for each transfection experiment. Statview software (SAS institute) was used to analyse the repeated measures of the efficiency of each IRES variant, according to the cell type or experimental conditions, with the Mann–Whitney U test.


Selection of IRES sequences in B cell and plasma compartments

PCR products were sub-cloned and a mean of 12 clones were derived from each compartment. At least one cluster of variants derived from B cell HCV RNA (figure 1) was detected in all patients. In five of six cases, B cells also harboured minor strains similar to those detected in plasma, as reported previously,15 while full compartmentalisation between plasma and B cells was found in patient #M. In one of the six patients (#D), genotypes 1 and 2 were found in both plasma and B cells. In one patient (#G), genotypes 2 and 3 were both detected in plasma but only genotype 3 was detected in B cells. Table 2 shows the alignment of the main selected variants.

Figure 1 shows the 23 main clusters of IRES sequences in these six patients, each represented by at least two clones. One clone representing each of these 23 clusters was selected. The clones corresponding to four minor clusters of IRES sequences could not be re-amplified and, thus, a total of 19 cloned IRES sequences (10 from B cells and nine from plasma) were then tested in the luciferase assay with hepatocyte and B cell lines. It is worth noting that, in three patients (#G, #D, #M), we could analyse two distinct HCV major variants derived from the plasma and the B cell compartments.

B cell specific IRES had a lower efficiency in hepatocytes than plasma/liver IRES

In each experiment, bicistronic constructs were tested in triplicate and we repeated all experiments at least three times. Because repeated experiments were performed, we normalised the efficiency of tested IRES sequences by dividing the experimental Rluc/Fluc ratio by the Rluc/Fluc ratio of the reference IRES sequence (pIRF) measured in each experiment. The efficiency of the nine plasma IRES sequences was significantly higher in Huh7 cells than that of the 10 corresponding B cell-specific IRES variants (1.7±0.3 vs 0.7±0.2; p<0.01; figure 3). Similarly, the least efficient of the nine tested plasma IRESes was 50% more efficient than the most efficient of the 10 tested B cell IRESes. We repeated tests of pairs of plasma and B cell-specific IRES variants from four patients, using primary cultured human hepatocytes from three different donors (figure 4). The differences between plasma and B cell-specific variants were similar to those observed with Huh7 cells in three cases (#C, #M, #D), while this difference was smaller in patient #G. In these six patients co-infected by distinct strains, the strains which predominated in the liver, and thus in the plasma, were those with a more-efficient IRES in hepatocyte cells.

Figure 3

R/F ratio of 19 internal ribosome entry sites (IRESs) in Huh7 and in Raji. In all cases and in Huh7, the plasma variants had a significantly higher translational efficiency than B cell variants (p<0.01 indicated by *). In Raji, the efficiency of plasma IRESs was lower than in Huh7 and similar to the efficiency of B cell IRES variants. The R/F value of pRF (the empty vector) was very weak.

Figure 4

R/F ratio of plasma and B cell internal ribosome entry sites (IRESs) from four patients transfected in primary cultured human hepatocytes, confirming the low efficiency of B cell IRESs in hepatocyte cells.

In Raji cells, for five of six patients, the Rluc/Fluc ratios were similar with B cell and plasma IRES variants and were not significantly different to those measured for B cell IRESes in Huh7 (figure 3). Similar results were observed with Daudi cells (data not shown). These results support the hypothesis of competition between IRES variants mediated by the translation efficiencies of co-infecting strains, since all five patients harboured both specific B cell variants and variants whose IRESes were identical to those of plasma variants in their peripheral cells. Conversely, the only patient (#M) whose B cell-specific IRESes were slighty, but significantly, more efficient than paired plasma IRESes in Raji cells was the only patient who harboured specific variants only in his B cells, with no plasma strains (complete compartmentalisation). In the same view, two distinct IRESes could be isolated and analysed in plasma and in B cells in three patients. In each of these subjects, the efficiencies of the two IRESes detected in the same compartment were similar in all cell lines tested. This suggests that the coexistence of two IRES variants within a compartment is related to the same efficiency of IRES in the cell type.

Major IRES variants were also tested in vitro in a rabbit reticulocyte translation system. We observed a relatively wide range of efficiencies among the different IRESes (1.3–5.1) with a small, but significant, difference between paired plasma and B cell IRESes (data not shown). No correlation was found between the R/F ratios determined in reticulocyte lysates and those in hepatocytes or B cell cultures, suggesting a role of hepatocyte or B cell-specific factors in explaining our observations.

A cryptic promoter did not play a role in bicistronic plasmid constructs

In most studies, HCV IRES activity has been measured in the same bicistronic dual luciferase expression system as that used here. One important potential source of bias is that reporter gene transcription may be initiated by a cryptic promoter located in the plasmid constructs. This has been demonstrated with cellular IRESes27 as well as HCV IRESes.28 We applied siRNAs directed against upstream firefly luciferase in Huh7 cells and in primary cultured human hepatocytes transfected with different bicistronic plasmids. SiRNAs specifically directed to firefly luciferase (ineffective on R luciferase) were tested with nine clones and with pIRF. In each case a drastic reduction in firefly and Renilla luciferase activity was observed (figure 5). However, silencing Renilla luciferase activity was slightly less efficient than silencing firefly luciferase, confirming the existence of an internal transcription promoter located in DNA encoding the HCV IRES.28 The contribution of transcription of this promoter was weak in all tested clones and, therefore, did not markedly affect our results. The efficiency of luciferase silencing was poor in non-adherent cells (Raji and Daudi cells). The empty vector pRF without IRES was also tested in Huh7: the F luciferase activity was comparable to those obtained with plasmid with IRES such as pIRF but as expected in the absence of IRES, the R luciferase activity was extremely low (figure 3). Finally, northern blot on RNAs extracted from Huh7 transfected with pIRF and four other constructs showed the expected 3 kb long bicistronic transcript (supplementary figure 1).

Figure 5

Effect of firefly luciferase silencing on the activity of firefly (F) and Renilla (R) luciferase on different internal ribosome entry sites (IRESs). The residual Renilla expression was observed, confirming the existence of a low cryptic promoter within the plasmid sequence coding IRES. Note the logarithmic expression of IRES activity, magnifying weak residual Renilla expression.

Transfection of bicistronic RNA constructs confirmed results obtained with plasmid constructs (figure 2)

The other important control was the transfection of capped bicistronic RNA containing B cell and plasma IRES variants. In this case, artefacts due to cryptic promoters are impossible. Although the efficiency of RNA transfection was poor, reproducible results were obtained with Huh7 and Raji cells for two pairs of plasma and B cell IRESes. In the former cell type, plasma IRES variants were more efficient than B cell variants in both tested subjects. In subject #M, plasma IRES was less efficient than B cell IRES in Raji cells, as demonstrated previously with plasmid constructs. In subject #D, activity was similar in plasma and B cell variants, as observed using corresponding plasmid constructs. It should be noted that the efficiency of IRES measured with bicistronic RNA was higher than that obtained with the plasmid constructs. Indeed, the Rluc/Fluc ratio was between 2 and 9, while it was below 2 when measured with the plasmid constructs. This was because all RNA molecules cannot be capped, reducing the translation of firefly luciferase. Finally, we translated in vitro the RNA from the pRF (empty vector) and transfected it in Huh7 cells. We did not observe any R luciferase activity after RNA transfection while the measured F luciferase was comparable to the activity observed with RNA translated from pIRF. Altogether, these controls showed that the marked correlation between cell tropism and IRES efficiency was not due to reproducible artefacts.

The function of plasma IRES depends more on lupus antigen than B cell specific IRES

Because LA has been shown to be a potent cellular IRES transacting factor of HCV IRES, we silenced this gene in PCH and Huh7 cells. Although LA silencing was relatively modest (50%), a strong reduction in the Rluc/Fluc ratio was observed compared with scrambled siRNA, confirming that LA was a positive regulator of HCV IRES. In Huh7 cells, the reduction in IRES efficiency by LA silencing was more marked for plasma IRES (mean 2.7±0.7) than for B cell IRES (1.7±0.2, p<0.01) (figure 6). Similar results were observed with PCH cells. Despite repeated attempts, LA silencing could not be obtained in Raji or Daudi cells.

Figure 6

Effect of lupus antigen silencing on the efficiency of plasma and B cell-specific internal ribosome entry sites (IRESs) expressed in Huh7 and primary cultured hepatocytes.


The main finding of this study was the correlation between the cellular tropism of HCV variants for liver or B lymphocytes in vivo, and the translational efficiency of their IRESes, tested by transfection in cultured hepatocytes and B cells. The IRESes of the B cell-specific strains had a common activity profile and were consistently less efficient than paired plasma IRESes in hepatocyte cells (Huh7 and primary cultured hepatocytes) but as efficient as these latter in non-hepatocytic cells (Raji, Daudi). These findings establish a new argument suggesting the existence of an extra-hepatic replication of HCV. If we hypothesise the absence of an extra-hepatic replication of HCV, the variants found in B cells are obligatorily of hepatocyte origin and the nucleotide sequence divergences between compartments are only due to mutations occurring in infected hepatocytes. Following this view, the fact that occult IRESes were systematically less efficient than plasma IRESes strongly suggests that B cell-specific strains originated from cells other than hepatocytes.

Although we detected negative-stranded HCV RNA in B cells of two of six cases studied, our findings do not demonstrate that replication obligatorily occurs in peripheral B cells. Peripheral B cells could harbour strains coming from non-liver tissues, such as lymph nodes.9 A recent study showed clearly that B cell-associated HCV readily infected hepatoma cells and had an enhanced infectivity compared with extracellular virus.23 In this view, our results largely exclude a positive adaptation of the translation of HCV RNA to B cells. In only one of six cases, the efficiency of B cell IRESes was modestly higher in B cell lines. However, we cannot exclude that these extra-hepatic strains are produced by tissues or cells in which the IRESes could be more efficient. Finally, these results suggest that these IRESes are not defective since their functions were similar to the plasma IRESes in Raji, as well as in Daudi, cells.

The group of patients studied is particular. They were selected according to the existence of an occult, persistent infection of their B cells by HCV strains not found in plasma/liver. They were clearly infected by at least two distinct strains with different IRESes in a setting of past and repeated exposures. We used this clinical setting to test the hypothesis of a relation between IRES function and extra-hepatic replication of HCV. The fact that plasma IRESes were more competent than B cell IRESes in hepatocytes suggested a selective pressure on extra-hepatic strains and, therefore, an extra-hepatic replication. Besides the issue of cellular tropism of different HCV variants, these findings suggest that strains co-infecting an individual, and having divergent IRESes, may compete through the efficiency of their IRES. Indeed, we observed in all available cases that the IRES variants co-infecting the same compartment in vivo had a similar translational efficiency in the corresponding cell system. Conversely, IRES variants which were dominant in a compartment had a stronger translational efficacy in the corresponding cell system. Since all other viral regions of plasma and occult variants are also divergent, we have to postulate that efficacy of translation is dominant over other viral functions. The fact that the efficiency of IRES in different cell types was related to its distribution in plasma and B cells suggests that the initiation of viral translation is dominant over other viral functions. Although the differences in IRES efficiency were highly significant and reproducible, they could seem relatively modest. However, for viral replication, any differences among competing strains rapidly favour those that initiate translation more efficiently. Finally, the analysis of more patients or more quasi-species might show that HCV variants with less-efficient translation, but with another strong selective advantage, exist.

This study, like others on the HCV IRES, was based mainly on the use of the classical bicistronic expression vector, which is more reproducible than monocistronic constructs but which may also be affected by artefacts. Silencing of reporter genes, RNA transfection, and northern blot rules out any significant role of cryptic promoters or RNA splicing in our observations. Another point is that the 3′-untranslated region (UTR) of HCV is also important for translation, although the isolated IRES is sufficient to initiate viral translation. We did not test the role of the 3′-UTR, although it likely enhances the translational efficiency of the IRES.29 The main reason is that the complete amplification of the whole genome in a single PCR is quite challenging and, thus, the isolation of the 3′-UTR corresponding to a given IRES is still impossible.

Although it is non-coding, the IRES is a well-preserved region because of the secondary and tertiary structures necessary to bind ribosomal proteins, eiF3, and probably because of trans-acting factors such as the HCV core protein,24 liver-specific microRNA mi12230 or lupus antigen.26 31 Few mutations have been described in this region which could regulate translational efficiency. The reproducible differences in the sensitivity of compartmentalised IRES to silencing of the LA suggest that this trans-acting factor plays in our observations. Other trans-acting factors could also be involved.32

The most important issue raised by the hypothesis of IRES-driven competition between HCV strains is to determine the viral and host conditions leading to optimal translation of HCV RNA in hepatocytes, as this could lead to the discovery of novel therapeutic targets.


The authors thank Kate Vassaux for reviewing the manuscript prior to submission.



  • Linked articles 210054.

  • Funding This work was supported by the French National agency for AIDS and hepatitis research (ANRS).

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the CHU Nantes.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles