We have studied three groups that differed with respect to the degree of exposition to HCV: 1) chronic HCV carrier (CI); 2) individuals who remained noninfected despite repeated exposure through sexual contact contact with these carriers (EUI); 3) healthy apparently unnexposed volunteers (UI). Specific cellular immune responses against HCV Core or NS3 were assessed by Elispot or T cell proliferation assays.
The highest proportion of HCV-specific response was observed among chronically-infected subjects (Table 1). The fact that such a response was detected in less than half of the patients is in agreement with previous reports on HCV T-cell responses in chronic HCV infection . Elispot responses were detected the most frequently (in 8/20 chronically-infected volunteers tested for Core, and in 2/6, tested for NS3 responses) whereas only few chronically-infected individuals presented Core or NS3-specific proliferative reactions (3/17, and 2/16, respectively).
The synthetic peptides used to screen cellular responses to HCV represent sequence of HCV genotype 1 since the majority of individuals were infected by a genotype 1 virus. Two of the three chronically-infected subjects who gave positive NS3- responses were infected with a genotype 1 virus. Seven of the eight chronically-infected subjects who were positive for Core by the Elispot assay, were also infected by HCV genotype 1. Meanwhile, chronically-infected patients carried also HCV of three other genotypes: one was infected by a genotype 2 virus, and two by each of genotypes 3 and 4 (Table 4). Very low frequency of NS3-responders amongst patients infected with HCV genotype 2, 3 or 4, may reflect a limited number of non-HCV genotype 1 infected individuals included in this study and also the genotype variation of NS3 sequence. The latter explanation is, however, hardly applicable for the core-specific responses, since HCV core is highly conserved with very few amino acid inter-genotype differences .
The low frequency of proliferative responses compared to Elispot could be attributed to a higher sensitivity of the latter (assay). However, in our view, decisive is the type of the registered response. Elispot assays performed with ex vivo isolated PBMC preferentially detect effector lymphocytes that do not need to expand, while assays using in vitro expanded T lymphocytes rather detect precursors of memory T cells with a proliferative capacity . The low frequency of proliferative responses among chronic HCV carriers may rather reflect a weak HCV memory response (specifically when comparing chronic hepatitis C patients to those resolving HCV infection; for review, 13). Of particular note, relatively few individuals gave a concomitant positive response in both assays. This absence of correlation between Elispot and proliferative responses in chronically-infected individuals suggests that effector and memory T cells are distinct T cell populations, probably recognizing different epitopes. Such phenomenon was described earlier .
Interestingly, HCV Core-specific Elispot responses were observed in a relatively high proportion (30%) of the uninfected partners of chronically-infected individuals (Table 1). This is in agreement with previous reports on the populations of uninfected seronegative individuals exposed to HCV, including healthy relatives of HCV-infected individuals, intravenous drug users, and individuals with occupational exposure [17–21].
The most striking result of the current study was that despite stringent criteria of the positive cellular response, an HCV-specific response was registered in 20% of uninfected subject tested (13/65; Tables 2 and 3). This group was split into two subgroups, depending on the possibility of exposure to HCV. Indeed, eight individuals who displayed a positive result could have been exposed to HCV (professionally), although there was no clear history of contamination (Table 3). No such risk was, however, identified to explain positive results in the remaining five individuals (Table 2).
The detection of HCV specific cellular responses in uninfected volunteers reflects the difficulty to precisely identify all (possible) risks of exposure to HCV. Furthermore, it may also reflect a past inapparent HCV infection. Clearance of HCV viremia associated with cellular immunity in the absence of seroconversion has been reported in populations at risk for HCV exposure [5, 22, 23].
Other causes for detecting HCV-response in healthy risk-free individuals cannot be categorically ruled out. Two uninfected volunteers had positive proliferative response for Core: one with no risk of exposure to HCV (EFS 20) gave a response that involved both CD4 and CD8 populations, whereas the other that only implicated a CD4 population response (EFS 11) was retrospectively shown to have been exposed to HCV. For EFS 20, we could map the reactive sequence to Core amino acid residues 173-190 (not shown). An extensive sequence search using the BLAST tool  revealed a eight amino acid homology between HCV Core 174-185 (FSIFLLALLSCL) and HBs antigen 41-52 (FIIFLFXLLXCL). While it remains possible that the observed reactivity corresponds to a cross-reactive immunization [8–10], it is noteworthy that EFS 20 was neither infected nor immunized with HBV.
NS3-specific proliferative responses were observed in four uninfected volunteers (COC 13, and EFS 14, 21 and 24). The PBMC of these individuals were also reactive to the CEF peptides including 12 influenza epitopes. As immunization against Influenza virus neuraminidase was reported to generate immune responses crossreactive with HCV NS3 , we cannot formally exclude that T-cell proliferation in response to NS3 resulted from cross-reactivity.
In all three groups, a much higher proportion of individuals tested positive for CEF-specific response registered by Elispot and proliferation tests. The proportion of responders (number of positive/total number tested) varied between groups, but was within the limits of stochastic variations: between 56% (9/16) for the chronically-infected subjects and 29% (5/17) for their uninfected partners (in proliferation). These figures matched the range of proportions seen in CEF-positive Elispots: 60% (18/30) for at risk uninfected volunteers and 46% (13/28) for uninfected volunteers with no known risk of HCV-exposure (Table 1). This was somewhat lower that the data reported by Currier et al. , but similar to that reported by Horton et al.  possibly reflecting the heterogeneity of the HLA alleles in the studied groups. All CEF-specific proliferative responses involved the CD8 subpopulation, and in 1 of 95 individuals, both the CD8 and CD4 compartments. This is not surprising since most of the CEF peptides were 8 to 9 mers representing CD8 class I-restricted epitopes, although CD4-specific cytotoxic responses have also been reported in human viral infections [26–28].
As there were no statistical difference between the groups in the frequencies of proliferative or Elispot responses to the control (CEF) antigens (Table 1), exposure to or infection by HCV did not seem to have any major impact on the frequency of cellular responses to unrelated viruses. Hence, it is unlikely that the number of positive cellular responses to HCV antigens could be explained by antigen stimulation(s) specific to other viral antigens. In addition, pair-wise comparisons revealed no difference in the occurrence of cellular immune response against HCV core and/or NS3 among CEF-negative versus CEF-positive individuals in any of the groups (UI with known risk, UI at risk, EUI, or CI; all p values > 0.3). Thus, there is no evidence demonstrating that anti-CEF cellular reactivity interfere with the detection of anti-HCV cellular responses.
Alternatively, atypical HCV-specific immune responses may be generated by the occult HCV infections of the liver . Such infections have been described for patients with abnormal liver function of unknown origin, who present negative HCV PCR and Elisa results in the serum but where HCV RNA is detected in the liver . However, in our study, all uninfected CIC volunteers had normal liver biology. For the twenty COC individuals, liver function was investigated using the Fibrotest , and all gave a normal value (not shown). Thus, it is likely that, in this study, the detection of a positive HCV-specific cellular response did not reflect an occult HCV infection.
The polymorphism of the IL28B gene has been recently associated with both spontaneous resolution of HCV infection and sustained virologic response in pegylated interferon/ribavirin treated patients [32–34]; we can speculate that such a polymorphism may explain partially our results but this study was initiated before the first report and we are unauthorized to make a retrospective genetic study.
In summary, the detection of HCV-specific immune responses in uninfected volunteers may reflect an under-estimated prevalence of inapparent and resolving acute HCV infections. This changes our understanding of the epidemiology and the physiopathology of HCV infection. An alternative, not mutually exclusive, hypothesis is the existence of cross-reactivity between HCV antigens and other viral or common antigens present in the general population, as previously suggested by other researchers.
Patients and methods
Patients and volunteers
Sixty-five presumably unexposed and uninfected volunteers (Uninfected individuals, UI) were studied. All volunteers were negative in HCV PCR assay (ABBOTT Real Time HCV, Abbott, Rungis, France, threshold < 12 I.U/ml) and had a negative HCV-specific humoral response according to a commercial Elisa assay (MONOLISA anti-HCV Plus V2, Biorad, Marnes-la-Coquette, France). This enzyme immunoassay contains HCV recombinant proteins expressed in E coli including sequences from NS3 and NS4 and from the structural core protein. All volunteers were not infected by HBV or HIV. These volunteers were categorized according to the putative risk of exposure to HCV [no known risk (n = 33), Table 2; at risk (n = 32), Table 3]. Exclusion factors for exposure to HCV  were: professional exposure, drug abuse, blood transfusion or injection of blood products, sexually transmitted diseases, incarceration, alcoholism, dialysis, endoscopy, acupuncture, mesotherapy, invasive cosmetic treatment, piercing, tattooing, sexual exposure, familial exposure, and hospitalization or outpatient treatment in a developing country. The 65 volunteers were recruited in three distinct centers located in the Paris area. Initialy enrolled was a group of 20 uninfected volunteers [mean age: 46 year; range: 27-65; sex ratio: 1] (Necker Clinical Investigation Center, CIC volunteers). However, it was retrospectively reported that fifteen individuals from this group might have been exposed to HCV due to their occupational status. A second group of 25 volunteers was recruited at a french blood center in Paris (Etablissement Français du Sang (EFS), Paris; EFS 01 to 25). This group comprised 8 volunteers without any known risk for exposure to HCV [mean age: 27.8 year; range: 18-40; sex ratio: 0.14] and 17 volunteers at risk [mean age: 43.1 year; range: 21-64; sex ratio: 0.13]. The third group of 20 volunteers with no known risk for exposure to HCV was recruted at the Center for clinical investigation of the Cochin Hospital, Paris [COC 01 to 20; mean age: 27.4 year; range: 18-41; sex ratio: 1.2].
Twenty chronic HCV infected carriers and their exposed uninfected sexual partners were included as positive controls for HCV infection and potential exposure, respectively (Table 4). Infected patients [mean age: 46 year; range: 24-66; sex ratio: 1] were all HCV seropositive and viraemic. All viruses were genotyped except for one volunteer; the HCV genotypes were: 1b (n = 10), 1a (n = 4)], 2a/c (n = 1), 3 (n = 2), and 4 (n = 2). The mode of contamination was established for fifteen individuals; ten were infected by blood transfusion, one after surgery, one following a tattooing procedure, and three were intravenous drug users. The 20 exposed uninfected partners [mean age: 44 year; range: 26-63; sex ratio: 1] were active sexual contacts (> 2 years) of these infected HCV carriers. All exposed uninfected individuals were HCV seronegative and HCV-RNA negative by PCR.
None of the volunteers was infected by HIV, and all had a normal blood cell count the day of harvesting PBMCs. Biomedical research was approved by the local ethics committee (RBM 01-24), and was carried out in accordance with the Helsinki Declaration.
Preparation of PBMC
PBMCs were isolated from heparinized blood as described . The PBMCs were frozen at -80°C in 90% fetal calf serum (D. Dutscher, Strasbourg, France) containing 10% DMSO (Pierce, ThermoFisher, Brebières, France), and stored in liquid nitrogen until used.
The consensus sequence of the Core protein (genotype 1a) was covered by thirty-seven 15 mer peptides that overlapped by 10 residues, as described . NS3 [consensus 1b, aa 1027-1657] was represented by sixty-eight overlapping 15 mer peptides corresponding to regions encoding the CD4 and CD8 epitopes were used. These clusters of T4 and T8 epitopes corresponded to the following regions: aa 1072-1111 (TCVN... LVGW); 1167-1191(GPLL... GVAK); 1199-1355(SMET... TDAL); 1461-1475 (TVDF... IETT); 1531-55(TPAE... QDHL); 1576-1652(TQKA... ACMS), according to the Los Alamos databases [36, 37]. A pool of unrelated 12-to 15-mer peptides derived from Gag and Nef of simian immunodeficiency virus (SIVmac239) were used as a negative control. Core and SIV peptides were purchased from NeoMPS (Strasbourg, France), and NS3 ones from Proimmune (Oxford, UK). Each peptide was certified to be > 80% pure, by RP-HPLC. Positive control was a pool of 32 peptides (CEF) corresponding to well-characterized CD8 class I restricted epitopes of human cytomegalovirus (CMV), Epstein-Barr virus (EBV) and Influenza virus . CEF pool was obtained through the NIH AIDS Research and reference reagent program, or Anaspec Inc., San Jose, CA, USA. The peptides were dissolved in DMSO at 1 mg/ml, and were stored at -80°C until used.
Virus-specific circulating effector T lymphocyte responses were studied using two distinct functional assays:
HCV-specific T cell responses of freshly isolated or frozen PBMC were studied by ex vivo ELISPOT assays , using the panels of Core or NS3 peptides described above. Peptides were used at a final concentration of 1 μg/ml. Negative controls consisted of cells incubated in medium. Phorbol myristate acetate and ionomycin (25 and 100 ng/ml, respectively; Sigma-Aldrich Chimie, Saint-Quentin Fallavier, France) were used as positive controls. The frequencies of IFN-γ producing cells were expressed as the number of spot-forming cells (SFC) per 106 cells. Frequencies lower than 50 spots/l million PBMC were considered unspecific. An assay was considered positive if: 1/. The number of spots generated in response to stimulation with specific peptides exceeded the mean of the number of spots obtained with culture medium plus 2 SD; and 2/. Its ratio to the number of spots with culture medium was > or = 4.
PBMC (2 ×106/ml) were labelled with 10 mM carboxyfluorescein diacetate succinimidyl ester (CFDA-SE; Invitrogen, ref C1157) in serum-free medium for 30 min at 37°C . Labeled PBMC were washed with complete medium (D-MEM +1% non essential aminoacids, 1 mM L-glutamine, Invitrogen, Cergy, France) supplemented with 10% heat inactivated human AB serum (SAB, Biowest, France), and incubated in complete D-MEM culture medium at 37°C under 5% CO2. The following antigen stimulations were performed: 1/. HCV-specific with pools of Core or NS3-specific peptides each at a final concentration of 1 μg/ml; 2/. Common antigen-specific CEF peptides as positive control (final concentration 0.5 μg/ml); 3/. Mitogen (superantigen) Staphylococcus Enterotoxin B (Ref S4881, Sigma, St Louis, MI) at 500 ng/mL as positive control for PBMC viability; 4/. An irrelevant SIV-peptide pool, and complete medium plus 0.05% DMSO (peptide diluent) as negative controls.
After 6 days incubation, cells were washed in PBS and incubated for 30 min at 25°C with anti CD3 phycoerythrin-Texas Red (ECD)-, anti-CD8β phycoerythrin-cyanin 5 (PCy5)-, and anti CD4 phycoerythrin-cyanin 7 (PCy7)-conjugated monoclonal antibodies (refs A07748, 6607101, and 737660 from Beckman-Coulter respectively). At the end of the incubation period, cells were washed twice in PBS and fixed with 200 μL of 2% formaldehyde solution in PBS for 15 min at 25°C. Cell division accompanied by CFSE dilution  was analyzed by flow cytometry. For each sample, at least 105 events were acquired using a FC500 cytometer (Beckman Coulter). Data were analysed with FlowJo (TreeStar). Lymphocytes were gated based on their forward and side scattering dot plot. T lymphocytes were defined based on their expression of CD3 and CD4 or CD8. The following criteria for antigen-specific proliferation were set: 1/. Background of proliferation without antigen (DMSO) < 4%; 2/. Antigen proliferation ratio (Antigen/SIV) > or = 4.; 3/. Absolute number of proliferating cells (i.e. CFSE negative) > 100; 4/. Threshold value > mean of difference between control antigen (SIV) + 2 SD.
Frequencies of HCV-specific proliferative and IFN-γ ELISPOT responses between groups were compared between the groups pairwisely using two-sided t-test for independant samples assessing difference in proportions. Tests were done using Quick Calcs, Graph Pad Software.