A summary of the clinical characteristics is provided in Table 1. Serum anti-HCV antibody titers of all patients were determined using the third-generation ELISA technique (4.0 Murex-Abbot), and detectable levels of viral RNA were measured using qualitative PCR (AMPLICOR^® Hepatitis C Virus Test 2.0). Only peripheral blood samples from patients who had not been treated were used to analyze the expression of markers, functional capacity, and correlations with factors predictive of evolution to fibrosis. In contrast, samples from treated patients were included in the genomic study. Liver samples from individuals with CHCV infection were obtained during the surgical procedure for the liver transplant. Data regarding sex, body mass index (BMI), the time when patients became infected (when this information was available), viral load, liver transaminase levels, inflammatory activity, and the degree of fibrosis were obtained to examine factors that predicted clinical progression. For various reasons related to the origin of the samples, we were not able to include all these data from the whole cohort of CHCV patients. Viral load was measured in 96 patients using the qRT-PCR technique with the COBAS^® TaqMan^® HCV Test v2.0 kit (Roche). Individuals were separated into three groups according to the obtained viral load: 12 individuals with a low viral load (<100,000 IU/ml), 25 individuals with a moderate viral load (between 100,000 and 500,000 IU/ml) and 59 individuals with a high viral load (≥500,000 IU/ml).

The COBAS^® HCV GT kit (Roche) was used to analyze the different viral genotypes; this kit distinguishes between genotypes 1 and 6 using a highly sensitive real-time PCR assay. The analysis of 115 CHCV patients showed the following distribution of viral genotypes: genotype 1:37 cases, genotype 2:14 cases, genotype 3:16 cases, and genotype 4:3 cases. BMI was obtained from 47 patients, 22 of whom had normal weight (BMI ≤ 25) and 25 were overweight (BMI > 25). The transaminase activities of SGOT or aspartate aminotransferase (AST) and SGPT or alanine aminotransferase (ALT) were measured in 291 HCV patients, 90 of whom had normal or slightly increased activity levels and 201 had values elevated by more than twofold relative to the upper limit of normal.

To evaluate the degree of inflammation and fibrosis, 140 liver biopsies were performed using a percutaneous or trans-jugular method. Fifty-five cases were evaluated using the METAVIR system. This system classifies fibrosis in stages ranging from F0 to F4, and in the present study, higher stages (F3 and F4) were considered to indicate cirrhosis. The METAVIR system also classifies necroinflammatory activity into stages A0–A3; in our study, stages A2 and A3 were considered to indicate high necroinflammatory activity and were associated with a more rapid progression to cirrhosis (21). Eighty-five individuals were evaluated using the Ishak score, which classifies fibrosis in stages ranging between E1 and E6. For this study, we considered the stages E5 and E6 to indicate cirrhosis. In 77 individuals, the degree of fibrosis was obtained using the non-invasive method of transient elastography (Fibroscan^®), which classifies fibrosis in stages ranging from F0 to F4, and stages F3 and F4 were considered to indicate cirrhosis. Finally, in 134 patients, the presence or absence of cirrhosis was simply determined by the combination of the results of clinical studies and ultrasonography.

Samples were collected from donor livers and recipients using methods previously described by Kelly et al. (22). During orthotopic liver transplantation, 23 samples from donor livers (Table 2) and 6 samples from CHCV-infected recipients were collected at the Austral University Hospital. All samples were identified by a transplant procedure number provided by INCUCAI that did not include the name of the donor and were collected from the Austral University Hospital, Buenos Aires, Argentina. After brain death, the procuration team working in the intensive care unit required the next of kin to sign the donor consent form for transplantation and participation in the research project.

During retrieval (at the time of exsanguination), the donor aorta and superior mesenteric vein were flushed with University of Wisconsin (UW) solution (Bristol-Myers Squibb, Uxbridge, UK) or with HTK (histidine–tryptophan–ketoglutarate) solution. After excising the organ, the liver was again flushed with UW solution until all blood was removed and the perfused solution appeared clear. At implantation and after completing the upper inferior cava anastomosis, livers were flushed with Ringer’s lactate solution through the portal vein at 4°C to wash out the UW before reperfusion. The perfused hepatic fluid was collected from the inferior cava vein (600–1,200 ml).

After each operation, the liver from the CHCV patient was placed in a container with a physiological solution, and 500 ml of lactated Ringer’s solution was infused through the portal vein; after traversing the organ, this solution was recovered through the suprahepatic veins. The liver remained in this solution during the transplant process. Subsequently, this solution was discarded, and the infusion of 500 ml of lactated Ringer’s solution through the portal vein was repeated; the solution from the second infusion was then collected.

Peripheral blood mononuclear cells (PBMCs) from adult healthy controls and liver mononuclear cells (LMCs) from healthy cadaveric donors and recipients were obtained using Ficoll-Hypaque density gradient centrifugation (GE Healthcare Biosciences, Uppsala, Sweden).

Peripheral blood mononuclear cells and LMCs were stained with unconjugated antibodies against KIR 3DL1/3DL2 (clone 5.133, courtesy of Dr. M. Colonna), KIR 2DL2/2DS2/2DL3 (clone CHL, courtesy of Dr. S. Ferrini), KIR 2DL1/2DS1 (clone HPMA4), KIR 2DL1/2DS1/2DS3 (clone HP3E4), CD94 (clone 3D9), and NKG2A (clone Z199, courtesy of Dr. M. Bottet) or unconjugated IgG1, IgG2a, IgG2b, and IgM isotype control antibodies (Becton Dickinson, San Jose, CA, USA). In addition, mouse monoclonal antibodies against 3DL1-FITC, 2DL3-PE (Becton Dickinson, San Jose, CA, USA), 2DS4-PE, NKG2C-PE (R&D Systems, MN, USA), NKp44-PE, NKp46-PE, 2B4-FITC, CD16-FITC, CD57-APC, CD8-PE/APC, CD4-FITC, CD45RA-PE-Cy7, CD28-PE, CD27-PE-CF594, CCR7-FITC, CD11b-FITC, CD161-APC/PE, HLA-DR-FITC, and labeled IgG1 and IgG2a isotype control antibodies (Biolegend, San Diego, CA, USA) were used. Cells were incubated with the above mentioned antibodies at room temperature (RT) for 30 min, then were washed twice with 1 ml of 1× PBS and centrifuged at 500 g for 5 min. For unconjugated antibodies, a subsequent incubation with 3 µl of FITC- or PE-conjugated antimouse immunoglobulin antibody (Dako, Glostrup, Denmark) was performed for 30 min at RT. After washing, 3 µl of normal mouse serum was added. Cells were incubated with 3 µl of anti-CD3 (Becton Dickinson, San Jose, CA, USA) conjugated to PerCP/FITC and 5 µl of anti-CD56 conjugated to PE/APC (Becton Dickinson, San Jose, CA, USA) or to BV421 (Biolegend, San Diego, CA, USA) for 15 min at RT to analyze labeling corresponding to the NK, NKT (CD3⁺CD56⁺), and T cells. A suitable fluorochrome combination was performed for each labeling scheme. Liver samples were also incubated with the anti-CD45 pan-leukocyte marker (Becton Dickinson) conjugated to APC-H7. After washes with 1× PBS, cells were fixed with 2% paraformaldehyde and subsequently analyzed using a FACSAria II flow cytometer (Becton Dickinson). The results were analyzed using FlowJo 7.6.2 software (Tree Star, Inc., Ashland, OR, USA). All analyses of a particular population of interest were based on the gating of at least 100,000 events. The surface expression of three KIR receptors was indirectly inferred as follows: the antibody 3DL1/3DL2 showed two populations with different mean fluorescence intensities, the lower population corresponded to 3DL2 and the upper population to 3DL1. In all cases, the expression of the upper population was compared with the monoclonal antibody against 3DL1 and a significant difference was not detected. The expression of 2DS3 was identified by subtracting the expression detected using the 2DL1/2DS1/2DS3 antibodies from 2DL1/2DS1 expression. The expression of 2DS1 was deduced by subtracting the expression of 2DL1 from the expression detected using the 2DL1/2DS1 antibody. The expression of 2DL2/2DS2 was inferred after subtracting the expression detected by the anti-2DL3 antibody to the expression detected by the anti-2DL2/2DS2/2DL3 antibody. In all cases, the presence of the KIR gene was verified. When using polyclonal antibodies, the correct method to deduce the expression of a single KIR protein is to use polyclonal and monoclonal antibodies together in the same tube (23); however, because of the limited number of fluorochrome combinations, we performed the reactions in different tubes and inferred the expression of single KIR proteins (see Figure S1 in Supplementary Material).

We distinguished between apoptotic and viable cells based on differences in forward and side scatter, and the results showed a good correspondence with the results obtained using FITC-annexin staining (24, 25). We previously assessed the quality of the results through dead cell staining, showing that dead cell discrimination by forward/side scatter displayed an excellent correlation with the results obtained using the LIVE/DEAD^® Fixable Aqua Dead Cell Stain Kit, which specifically stains lymphocytes.

The conditions used to identify the presence or absence of each KIR gene have been documented previously (20). Briefly, two PCR amplifications were performed: PCR-1 amplified the combined D1 and D2 domains and PCR-2 amplified the transmembrane and cytoplasmic regions. Nineteen 5′-digoxigenin-labeled probes were used in the SSOP approach, with 13 for PCR-1 and 6 for PCR-2. The KIR gene content in each individual was inferred after analyzing the combinations of all probes. The KIR2DS4 gene was amplified by PCR with previously described primers to analyze the cell membrane-anchored receptor (designated KIR2DS4-FL or FL) and a truncated soluble protein (designated KIR1D), which is generated when exon 5 contains a 22-bp deletion. HLA-A, B, and C genotyping was performed for sequences from exons 2 and 3. The primers and conditions for PCR amplification were the same as those described by Cereb et al. (26). HLA-C SSOP typing was also conducted using two probes (5′-digoxigenin label); one of the probes (5′-TGACCGAGTGAACCTGC-3′) was specific for the HLA-C alleles that belong to the C1 group (Asn 80), and the other probe (sequence: 5′-ACCGAGTGAGCCTGCG-3′) anneals with the HLA-C alleles that belong to the C2 group (Lys 80).

Mononuclear cells obtained from the liver or peripheral blood were placed in 24-well plates at a density of 1 × 10⁶ lymphocytes per well and cultured in 1 ml of complete RPMI medium. Three wells were used for each sample according to the following scheme: (1) cells stimulated with the target antibody to study functionality, (2) cells stimulated with the isotype control, and (3) unstimulated cells that served as a negative control. Twenty microliters of the anti-CD107a antibody (Becton Dickinson) conjugated to FITC were added to wells 1 and 3 and 20 µl of the FITC-conjugated IgG1 (Becton Dickinson) isotype control was placed in well 2 as a control. PMA (phorbol 12-myristate 13-acetate; Sigma) was added to wells 1 and 2 at a final concentration of 25 ng/ml, and ionomycin (calcium salt from Streptomyces conglobatus; Sigma) was added at a final concentration of 0.5 µg/ml. The culture plate was incubated for 1 h at 37°C in a 5% CO₂ atmosphere. One microliter of monensin (BD-Golgi Stop Protein Transport Inhibitor, BD Biosciences) was then added to all wells, and the cells were incubated for an additional 5 h. The contents of each well were transferred to FACS tubes; cells were washed with 1 ml of 1× PBS/A and centrifuged at 500 g for 5 min. In all samples, cell surface labeling was performed with the following monoclonal antibodies: 3 µl of anti-CD3 conjugated to PerCP (Becton Dickinson) and 5 µl of anti-CD56 conjugated to APC (Becton Dickinson). Pan-leukocyte anti-CD45 conjugated to APC-H7 (Becton Dickinson) was added to the liver samples. After a 20-min incubation at RT, cells were washed twice with 1× PBS/A. During intracytoplasmic labeling, cells were treated with 100 µl of 1× fixation and permeation solution (Cytofix/Cytoperm, BD Biosciences) for 20 min at 4°C, washed twice with 1 ml of 1× Perm/Wash buffer (Becton Dickinson), and centrifuged at 500 g for 5 min. Twenty microliters of IgG1 isotype PE control (Becton Dickinson) were added to well 2, and 20 µl of anti-IFNγ conjugated to PE (Becton Dickinson) were placed in wells 1 and 3; then, the samples were incubated for 30 min at 4°C. Cells were washed with 1× Perm/Wash buffer, the pellet was suspended in 200 µl of 1× PBS/A, and the cells were fixed with 2% paraformaldehyde. Finally, the cells were analyzed using a FACSAria II flow cytometer (Becton Dickinson).

The HLA allele frequencies and expression of KIR genes in the patients and controls were compared using Fisher’s exact test to evaluate significant differences, and the p-value was corrected with the Bonferroni method when appropriate. Additionally, 3 × 2 contingency tables were analyzed using the Chi-square test for independent samples or the maximum likelihood method when appropriate (viral load factor). The Mann–Whitney U test was used to compare independent groups. Spearman’s test was used to establish correlations. For all statistical analyses, we used GraphPad Prism 5 (GraphPad Software). Data are presented as medians and ranges. Reported p-values are two-tailed, and p < 0.05 was considered significant.

No significant differences in the typing of HLA-A, HLA-B, and HLA-C genes were observed between HCV patients (n = 351) and healthy individuals (n = 273) (Table 3).

The KIR genotyping results did not show any differences from the data reported in a previous study (11) (Table 4). However, the typing of different KIR2DS4 alleles (27) revealed an increased frequency of the functional KIR2DS4 alleles (KIR2DS4-FL) in patients with CHCV infections (0.52 vs. 0.42 in healthy controls, p = 0.02) (Table 5). Similarly, the deleted forms (KIR1D) showed a decreased frequency of the homozygous KIR1D genotype and a clear tendency toward an increase in the frequency of the homozygous KIR 2DS4FL genotype in patients with CHCV infections.

Killer cell immunoglobulin-likereceptors must recognize their ligands to perform their functions. Among the inhibitory KIRs, KIR2DL1 recognizes HLA-C2 allotypes containing a lysine at position 80 of the HLA-Cα₁ domain and KIR2DL2/3 recognize HLA-C1 alleles containing Ser77/Asp80 (28, 29). However, KIR2DL2 and, to some extent, KIR2DL3 interact with some C2 alleles with weaker affinity than KIR2DL1. The genotype frequency of individuals with a known ligand is shown in Table 6. After analyzing the whole population of CHCV patients, the comparison between healthy individuals and CHCV patients showed no significant differences.

In contrast to the genotype frequency, the expression of KIR genes in PBMCs from patients with CHCV infections showed differences compared with PBMCs from healthy controls. NK (CD56⁺CD3⁻), NKT (CD3⁺CD56⁺), and T (CD3⁺) cells were analyzed. CD56^dim cells from CHCV patients carrying the KIR gene showed higher expression of KIR2DS1 (10.1 vs. 5.3% in CD56^dim cells from healthy controls, p = 0.003) (Figure 1C) and KIR3DL2 (22.5 vs. 15.1% in healthy controls, p = 0.009) (Figure 1C) and lower expression of KIR3DL1 (11.4 vs. 16.4% in healthy controls, p = 0.04) (Figure 1C). CD3⁺CD56⁺ cells also showed increased expression of KIR3DL2 (12.3 vs. 7.1% in healthy controls, p = 0.02) and KIR2DS4 (13.3 vs. 4.0% in healthy controls, p = 0.01) (Figure 1D). No differences were found in the expression of KIRs within CD56^bright NK cells (Figure 1B) and T cells (Figure 1E). Importantly, PBMCs collected from CHCV patients, and healthy LMCs were collected from untreated individuals. LMCs from CHCV patients were collected before liver transplantation on previously treated patients. However, LMCs from CHCV patients showed levels of KIR expression that were similar to those in LMCs from healthy controls (Figures 1F–I). Expression was measured only in individuals carrying the KIR gene, which partially explains the small number of CHCV-infected liver samples examined.

Compared with those from healthy controls, PBMCs from CHCV patients showed decreased frequencies of CD56^dim NK cells (6.5 vs. 8.3%, p = 0.002) and CD3⁺CD56⁺ cells (2.0 vs. 3.8%, p = 0.001) (Figure 2A). Similarly, decreased frequencies of CD56^dim NK cells (5.0 vs. 12.2% in CD56^dim cells from healthy controls, p < 0.001) and CD56^bright cells (5.4 vs. 15.0% in healthy controls, p < 0.001) (Figure 2B) were detected in the liver of CHCV patients, but the CD56^dim/CD56^bright ratio was maintained (50% of each subset). Livers of CHCV patients showed significant infiltration of T lymphocytes, suggesting that the decrease in the frequency of NK cells occurred at the expense of an increase in the frequency of T lymphocytes (79.1 vs. 54.5% in peripheral blood, p < 0.001). The lack of changes in the frequency of liver CD3⁺CD56⁺ cells indicates that the increase in the frequency of T cells observed in CHCV patients involved T cell subsets aside from CD3⁺CD56⁺ cells (Figure 2B).

Compared with those from healthy controls, PBMCs-CD56^bright NK cells from CHCV patients showed no differences in the expression of CD94, NKG2A, NKG2C or CD16 (Figure 3A). In contrast, PBMCs-CD56^dim NK cells from CHCV patients showed a decreased frequency of CD16 expression (69.5 vs. 92.0%, p = 0.002; Figure 3B). Similarly, liver CD56^dim cells maintained the same low expression of CD16 in liver healthy controls (Figure 3D), which implies a potentially lower capacity to exert ADCC. We detected more frequent expression of the activating receptor NKG2C in the liver CD56^bright cells from patients with CHCV infections (13.2 vs. 4.2% in healthy controls, p = 0.04; Figure 3C).

We next evaluated the correlations of altered expression of NK receptors with clinical features of CHCV patients.

We postulated that degranulation, as measured by CD107a staining, is a representative marker of cytotoxicity. Compared with healthy livers, liver CD56^bright, and CD56^dim cells from CHCV patients showed an increased cytotoxic capacity (71.1 vs. 42.5%, p < 0.05, and 47.0 vs. 18.8%, p < 0.05, respectively; Figure 6). Similarly, liver CD3⁺CD56⁺ and T cells from CHCV patients also showed a higher cytotoxic capacity (79.3 vs. 35.0, p = 0.03, and 53.5 vs. 26.7%, p = 0.03, respectively). In contrast, patients and healthy controls displayed a similar capacity to secrete IFNγ (Figure 6).

Peripheral blood CD56^bright and CD56^dim cells from CHCV patients showed decreased expression of CD45RA (89.1 vs. 95.4% in healthy controls, p = 0.02, and 98.0 vs. 100.0 in healthy controls, p = 0.01, respectively; Figures 7A,B). This profile was even more evident in the liver CD56^bright and CD56^dim cells (69.6 vs. 96.2% in healthy livers, p < 0.0001, and 82.5 vs. 97.0% in healthy livers, p = 0.009, respectively; Figures 7E,F). Loss of this marker suggests that the cell population included more activated NK cells. On the other hand, CD161 expression was markedly increased in peripheral blood CD56^bright cells from CHCV patients (68.0 vs. 40.3% in healthy controls, p = 0.0008). Peripheral blood T lymphocytes from patients with CHCV infections displayed decreased expression of the markers CD45RA, CD27, and CD28 (33.8 vs. 54.2% in healthy controls, p = 0.004, 83.0 vs. 95.2% in healthy controls, p = 0.009, and 77.6 vs. 90.1% in healthy controls, p = 0.002, respectively; Figure 7D). A similar trend was observed for the expression of CD45RA, CD28, and CD27 in CD3+CD56⁺ PBMCs (Figure 7C) and in hepatic T and CD3⁺CD56⁺ lymphocytes (45.7 vs. 69.0% in healthy livers, p = NS, 38.3 vs. 58.3% in healthy livers, p = NS, and 26.0 vs. 65.1%, p = 0.03, 23.7 vs. 41.1%, p < 0.05, respectively; Figures 7G,H). Thus, the increased frequency of T cells in the livers of CHCV patients is associated with a more differentiated stage.