Potential conflict of interest: Nothing to report.
This work was supported by research funds of the Italian Ministry of Health (Ricerca Corrente, Fondazione IRCCS Policlinico San Matteo), by a grant from the Italian Ministry of Education, University and Research MiUR (Fondi di Investimento per la Ricerca di Base, Protocollo: RBAP10TPXK), and by COPEV Associazione per la Prevenzione e Cura dell'Epatite Virale Beatrice Vitiello ONLUS.
Hepatitis C virus (HCV) persistence in the host results from inefficiencies of innate and adaptive immune responses. Most studies addressing the role of innate immunity concentrated on peripheral blood (PB) natural killer (NK) cells, whereas only limited information is available on intrahepatic (IH) NK cells. We therefore examined phenotypic and functional features of IH and PB NK cells in paired liver biopsy and venous blood samples from 70 patients with chronic HCV infection and 26 control persons subjected to cholecystectomy for gallstones as controls. Ex vivo isolated IH NK cells from HCV-infected patients displayed unique phenotypic features, including increased expression of NKp46-activating receptor in the face of reduced tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) and cluster of differentiation (CD) 107a expression, which resulted in impaired degranulation compared with controls. To gain insights into the effect of HCV on NK cells, we exposed peripheral blood mononuclear cells (PBMCs) from patients and healthy donors to cell-culture–derived HCV (HCVcc) and measured NK cell degranulation, TRAIL, and phosphorylated extracellular signal-regulated kinase 1/2 (pERK1/2) expression. Exposure of PBMCs to HCVcc significantly boosted NK degranulation, pERK1/2, and TRAIL expression in healthy donors, but not in patients with chronic HCV infection, a defect that was completely reversed by interferon-alpha. Purified NK cells showed a minimal, though significant, increase in degranulation and TRAIL expression, both in patients and controls, after exposure to HCVcc. Conclusions: These findings indicate dysfunctional IH NK cell cytotoxicity associated with TRAIL down-regulation in chronic HCV infection, which may contribute to virus persistence. PB NK cell impairment upon exposure to HCVcc suggests the existence of an accessory cell-dependent NK cell lytic defect in chronic HCV infection predominantly involving the TRAIL pathway. (HEPATOLOGY 2012;56:841–849)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Innate immunity controls adaptive immune responses through direct interaction and through the exchange of signals between immune cells belonging to both compartments, as reviewed previously.1, 2 Natural killer (NK) cells represent the principal effector cell population involved in innate immune responses to viral infections and are particularly enriched in the liver, where they may account for up to 50% of the intrahepatic (IH) mononuclear cell (MC) pool.3, 4 However, during chronic hepatitis C virus (HCV) infection, NK cells are reduced within the IH compartment5, 6 and their frequency appear to decrease with disease progression.6, 7 Preliminary evidence in patients with chronic HCV infection reported NK cell phenotypic differences between the IH and peripheral blood (PB) compartments, although the vast majority of studies omitted to address IH NK cell function.6, 8, 9 In general, a larger proportion of IH NK cells express activation molecules, such as natural killer group 2 member D (NKG2D),8 NKp46, and tumor necrosis factor–related apoptosis-inducing ligand (TRAIL),9 compared to the PB compartment. Whether such phenotypic differences reflect a physiological setting or are a consequence of chronic HCV infection is presently unclear. Indeed, evaluation of IH NK cells in healthy donors is limited by obvious ethical reasons, and the only available data were obtained using cells isolated from seemingly unaffected areas in subjects with hepatic malignancies or with non-HCV-related chronic liver conditions. In those studies, a higher proportion of IH NK cells expressing the NKG2A inhibitory receptor and a lower percentage of NKp46- and NKp30-activating receptors was observed in patients with chronic HCV infection.10, 11 However, the limited number of cells that can be retrieved from a small fragment of a percutaneous liver biopsy severely limits the number of experiments that can be performed in each patient. As a consequence, only a scarcity of data are available on IH NK cell function, with only one study describing reduced cytokine production and cytotoxic activity in HCV-positive cirrhotic patients, compared to patients with non-HCV-related hepatic malignancies.7
The lack of relevant and reproducible information in the field prompted us to examine phenotypic and functional features of paired peripheral blood mononuclear cells (PBMCs) and intrahepatic mononuclear cells (IHMCs) from patients with chronic HCV infection and control subjects undergoing elective surgery for gallstones. To gain additional insights into the effect of HCV on NK cells, we also analyzed PB NK cytolytic potential upon exposure to cell-culture–derived HCV (HCVcc) in vitro.
ALT, alanine aminotransferase; AST, aspartate aminotransferase; CD, cluster of differentiation; ERK1/2, extracellular signal-regulated kinase 1/2; HAI, histological activity index; HBV, hepatitis B virus; HCV, hepatitis C virus; HCVcc, cell-culture–derived HCV; IFN-α, interferon-alpha; IFN-γ, interferon-gamma; IH, intrahepatic; IHMC, intrahepatic mononuclear cells; IL, interleukin; JFH-1, Japanese fulminant hepatitis type 1; mAbs, monoclonal antibodies; MCs, mononuclear cells; MFI, mean fluorescence intensity; NK, natural killer; NKG2D, natural killer group 2 member D; NS3, nonstructural protein 3; PB, peripheral blood; PBMCs, peripheral blood mononuclear cells; pERK1/2, phosporylated extracellular signal-regulated kinase 1/2; pJFH-1, Japanese fulminant hepatitis type 2a strain JFH-1 clone; STAT, signal transducer and activator of transcription; TNF-α, tumor necrosis factor-alpha; TRAIL, tumor necrosis factor–related apoptosis-inducing ligand.
Patients and Methods
Paired liver biopsy and venous blood samples were obtained from a total of 70 consecutive untreated patients with chronic HCV infection and 26 controls. The latter were patients undergoing elective videolaparoscopic cholecystectomy for uncomplicated gallstones. In this group, blood was taken before anesthesia, and no patient developed complications after liver biopsy. All controls had a histologically healthy liver and normal alanine aminotransferase (ALT) and had negative hepatitis B virus (HBV), HCV, and human immunodeficiency virus serology. Patients with chronic HCV infection were all newly diagnosed and received a liver biopsy before treatment. Patients examined in the different assays were comparable for genotype distribution, ALT values, and histology according to Knodell.12 None were active intravenous drug users. A written informed consent was obtained from each individual. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by our institutional ethical committee. The characteristics of HCV-positive patients and controls are listed in Table 1. HCV-positive patients and control subjects analyzed in the different experiments were homogeneous for age and gender.
MC Preparation, Staining, and Phenotypic and Functional Analysis.
These were performed by standard techniques essentially as described before.8 Further technical details, including IHMC isolation and the monoclonal antibodies (mAbs) used for cytofluorimetric analyses, are provided in Supporting Materials. Details on gating strategy and representative dot plots are provided in Supporting Figs. 1 and 2.
HCV Replication System.
Huh-7.5 cells and Japanese fulminant hepatitis HCV genotype 2a strain JFH-1 clone (pJFH-1) were kindly provided by C.M. Rice (Rockefeller University, New York, NY) and by T. Wakita (National Institute of Infectious Diseases, Tokyo, Japan) and were grown as previously described.13 HCV RNA transfection was achieved by electroporation with In Vitro genomic HCV RNA transcribed in vitro from pJFH-1, as previously described.14 To determine the efficiency of infection with JFH-1, Huh-7.5 cells were seeded at a density of 12.5 × 104 per well in six-well plates, infected overnight with 4.6 × 106 copies/mL of HCVcc, and analyzed by flow cytometry and immunofluorescence for HCV protein expression using human mAbs B12.F8 specific for HCV core15 and CM3.B6 specific for nonstructural protein 3 (NS3) helicase/nucleoside triphosphatase domain.16 HCV RNA in culture medium was determined by real-time quantitative reverse-transcription polymerase chain reaction (rt-PCR) (technical details are reported in the legend to Supporting Fig. 3).
Exposure of NK Cells to HCVcc.
First, 1 × 106/mL of freshly isolated PBMCs or negatively selected (>90% enriched) NK cells (EasySep; Stemcell Technologies, Vancouver, British Columbia, Canada) from patients with chronic HCV infection and healthy donors were incubated for 24 hours with supernatants from uninfected or JFH1-infected Huh-7.5 cells in the absence of cytokines. More than 85% of HCV-infected Huh-7.5 cells used in the experiments expressed HCV core and NS3 proteins (Supporting Fig. 3), and supernatants were used at the concentrations of 2.5 and 15 × 106 HCV RNA copies/mL, comparable to the serum HCV RNA titers usually observed in patients. After incubation, cells were used as effectors in a 3-hour CD107a degranulation assay with K562 target cells at an E:T ratio of 1:1. In some experiments, interferon-alpha (IFN-α) (1,000 U/mL) was added in culture to examine its effect on TRAIL and CD107a expression on PB and enriched NK cells. In addition, extracellular signal-regulated kinase 1/2 (ERK1/2) phosphorylation (pERK1/2) was evaluated by flow cytometry in PB NK cells in time-course experiments (0, 5, 10, and 20 minutes) after incubation at 37°C with an equal number of K562 cells. Cells were fixed with BD Fix solution for 10 minutes at 37°C, washed with phosphate-buffered saline, and permeabilized with Perm III BD solution for 30 minutes in ice. After two washings, cells were stained with CD56 PC5, CD3 phycoerythrin (Beckman Coulter, Inc., Brea, CA) and anti-phospho ERK1/2 Alexa 488 (BD Biosciences, Franklin Lakes, NJ).
Mann-Whitney U test was used to compare data, which were not normally distributed, between different groups. Wilcoxon's signed-rank test was used to compare data within the same group. Phenotypic and functional data were expressed as median values and interquartile range or as “box and whiskers” graphs. Correlations between variables were analyzed using Spearman rank-correlation coefficient. A P value ≤0.05 was deemed statistically significant.
IH and PB NK Cells Are Characterized by Different Receptor Expression.
Comparison between PB and IH compartments showed similar trends in HCV patients and controls for several NK receptors analyzed. Thus, there was a statistically significant enrichment of intrahepatic NKp44-, NKp46-, CXCR4-, NKG2A-, and CD107a-expressing NK cells, whereas NK cells expressing NKp30 and KIR3DL1 were significantly reduced, compared to PB compartment in both patients and controls (Fig. 1). Changes in NKp44, NKp46, and NKp30 frequencies were paralleled by changes of their mean fluorescence intensity (MFI) on NK cells (Supporting Fig. 4). There were no differences in PB NK receptor expression between patients and controls, with the exception of NKp30, which was slightly increased in HCV-infected patients (P = 0.056). Interestingly, in agreement with our previous findings8 and with Sène et al.,17 NKG2D-expressing NK cells were significantly enriched in the IH compartment of HCV patients, compared with their own PBMCs, whereas no such enrichment was observed in control subjects (Fig. 1E). CD16 expression was significantly decreased on IH NK cells, compared to PB NK cells, in both groups (Supporting Fig. 5).
Unique Phenotypic Features of IH NK Cells in Chronic HCV Infection.
In agreement with others,6, 18 we found that the percentage of IH NK cells was significantly decreased in patients with chronic HCV infection, compared to controls (Fig 2A), which correlated inversely with serum ALT and AST levels (Fig. 2C,D) and was significantly decreased in patients with the highest histological grading (histological activity index [HAI] score), according to a modified Knodell score12 (Fig. 2B). No statistically significant correlation was found between fibrosis stage and NK receptor expression.
HCV-infected patients showed significant up-regulation of NKp46 and NKG2D on IH NK cells, compared to control subjects, and, although not statistically significant, there was a trend toward up-regulation of NKp30 (P = 0.08) (Fig. 3). In contrast with this finding, expression of TRAIL and CD107a on IH NK cells was significantly reduced in HCV patients, compared to control subjects (Fig. 3).
Impaired IH NK Cell Degranulation and Normal Cytokine Secretion in Chronic HCV Infection.
In view of these apparently contrasting phenotypic data, which suggested simultaneous up-regulation of activating receptors, together with down-regulation of killing-related molecules, we examined CD107a degranulation on IH NK cells from patients and controls after stimulation with interleukin (IL)-2 and IL-12. Significantly decreased IH NK cell degranulation was observed in HCV-infected patients, compared to control IH NK cells, using K562 and P815 cell line targets (Fig. 4A,B). Redirected cytotoxicity experiments with anti-NKG2D- and anti-NKp30-activating receptors confirmed the reduced cytolytic potential of IH NK cells in these patients (Fig. 4C,D). Surprisingly, NK cell ability to produce interferon-gamma (IFN-γ) and tumor necrosis factor-alpha (TNF-α) through the NKG2D pathway remained conserved in both PB and IH compartments (Supporting Fig. 6).
Exposure to HCVcc Results in Impaired PB NK Cell Degranulation, Failure to Up-regulate TRAIL, and Reduced pERK1/2 in HCV Patients.
To examine whether HCVcc could directly induce modification in NK cell cytolytic potential, we tested the degranulation capacity and TRAIL expression of ex vivo freshly isolated PB and negatively selected NK cells from patients with chronic HCV infection and healthy controls after incubation with HCVcc-positive or -negative supernatants from Huh-7.5 cells. Exposure of PB NK cells to HCVcc at titers that are normally found in patients with chronic HCV infection resulted in a HCV-RNA dose-dependent increase in NK cell degranulation (Fig. 5A,B) and a concomitant increment of TRAIL expression (Fig. 5C,D) in healthy donors, with the latter being predominantly found on CD56bright NK cells (Supporting Fig. 7). In contrast, exposure to HCVcc failed to induce a significant increase in degranulation on patients' PB NK cells (Fig. 5A,B), and no concomitant significant up-regulation of TRAIL was observed (Fig. 5C,D). There was no statistically significant difference in PB NK degranulation or TRAIL expression upon exposure to uninfected Huh-7.5 cell supernatant between the two groups analyzed. Purified NK cells showed reduced CD107a and TRAIL expression, compared to unsorted PB NK cells, and the addition of HCVcc induced a weak, but significant, up-regulation of these molecules, both in controls and HCV patients (Fig. 5).
Because IFN-α is known to determine the up-regulation of TRAIL and to increase NK cell degranulation,9, 19 we analyzed virus-induced modulation of CD107a and TRAIL with or without IFN-α. There was a significant increase in CD107a (Fig. 6A,B) and TRAIL (Fig. 6C,D) expression on NK cells treated with IFNα, which was already evident in the absence of virus, both in patients and controls. Exposure to HCVcc did not further increase or reduce expression of these molecules in both groups, indicating that HCVcc does not influence IFN-α activation of NK cells.
To understand whether the observed decrement in NK cell degranulation upon exposure to HCV could be ascribed to inhibition of the ERK-signaling pathway, we examined pERK1/2 in PBMC treated with supernatants containing or not containing HCVcc and exposed to K562 cells in a time-course experiment. NK cells from healthy donors showed increased pERK1/2 in the presence of HCVcc, compared to NK cells treated with control Huh-7.5 media, which seemed to be predominant in the CD56dim population (Fig.7a). In contrast, NK cells from HCV-infected patients failed to up-regulate pERK1/2 after incubation with HCVcc-positive medium (Fig. 7B).
In this study, in agreement with others,6, 7 we found a reduced proportion of IH NK cells, which was inversely associated with the HAI and serum aminotransferase values, suggesting that NK cells would be diluted in the context of a large inflammatory cell infiltrate in severe liver disease, which would be constituted predominantly by T cells. This interpretation is supported by our findings of a proportional enrichment of IH T cells in this setting (not shown).
Direct comparison between control and HCV-infected livers disclosed novel, major differences in NK-cell–activating receptor expression. Thus, NKp46 and NKG2D were increased in HCV patients, suggesting a prevalent activating phenotype. Our data are at variance with those of Nattermann et al.,11 who found decreased proportions of NKp30- and NKp46-positive IH NK cells in patients with chronic HCV infection. These apparent discrepancies could be explained by the smaller number of subjects analyzed in that study, which also used patients with liver disease caused by different etiologies as a control group, whereas we employed control subjects without parenchymal hepatic diseases. Interestingly, this prevalent activating phenotype was not associated with enhanced cytolytic potential, because IH NK cell degranulation was significantly reduced in HCV-infected patients, compared to controls. This impaired cytolytic function was confirmed in redirected degranulation experiments, in which NKG2D- and NKp30-mediated cytotoxicity was impaired, despite NKG2D up-regulation on IH NK cells in HCV subjects. This finding may have several explanations. Indeed, it is known that HCV is able to inhibit NK cells by interaction between the E2 protein and CD81,20, 21 and that the HCV core protein induces up-regulation of major histocompatibility complex class I on hepatocytes22 and stabilizes the expression of human leukocyte antigen E on liver cells inhibiting NK-mediated cytolysis.10 Moreover, it has been shown that IH levels of IL-10 determine an immunosuppressive environment, both in mice23 and humans,24 and, in agreement with the aforementioned, it has been reported that IH, HCV-specific IL-10-producing regulatory CD8+ T cells may prevent liver damage during chronic infection.24, 25 This, coupled to exhaustion induced by continuous receptor engagement, would eventually lead to defective cytolytic function.
It is interesting to note that in our study, TRAIL was down-regulated on IH NK cells in HCV-infected patients, adding further evidence in support of a role of this molecule in IH NK cell-killing activity.26 Others9 reported increased TRAIL MFI in IH, compared with PB NK, cells of HCV-infected patients; however, the absence of a control group makes it difficult to interpret this finding. The importance of the role of TRAIL in chronic HCV infection is further emphasized by evidence that TRAIL is up-regulated at the gene level in patients who have successfully responded to IFN-α treatment19 and by data showing up-regulation of this molecule on NK cells from healthy donors after IFN-α exposure in vitro.9 Moreover, our in vitro experiments clearly demonstrated a failure to up-regulate TRAIL on PB NK cells upon exposure to HCVcc in HCV patients, compared to control subjects. Similarly, there was no significant, dose-dependent increment of NK cell degranulation in HCV patients in the presence of HCVcc. Negatively selected NK cells showed instead a minimal, though significant, increase in degranulation and TRAIL expression, both in patients and controls, after exposure to HCV, suggesting that a functional impairment of accessory cells is responsible for the lack of NK activation in chronic HCV infection. IL-12 is produced primarily by monocytes, macrophages, and dendritic cells and is known to affect NK cell cytolytic function by up-regulating TRAIL and inducing ERK1/2, signal transducer and activator of transcription (STAT)1, and STAT4 phosphorylation.27 Interestingly, our preliminary data show that the proportion of IL-12-producing monocytes from healthy donors, but not from HCV patients, treated with HCVcc positively correlates with that of TRAIL-positive NK cells, particularly with the CD56bright subset (Supporting Fig. 8). Studies are under way in our laboratory to investigate the role of other cells and cytokines in HCVcc-induced NK cell activation. On the other hand, failure to up-regulate TRAIL or CD107a in HCV+ patients was completely rescued by IFN-α treatment, independently of the presence of HCV, indicating that this functional defect is reversible and not pervasive.
Contrasting results about the effect of HCVcc on NK cell function have been published by others, with one group showing no effect of HCVcc on NK cell activity28 and another one showing reduced IFN-γ secretion upon exposure of NK cells to plastic-coated, but not to soluble, HCVcc.29 The use of freshly isolated PBMCs before and after NK enrichment, which were exposed to HCV titers that are normally present in this clinical setting, as well as different patient characteristics, may, at least in part, account for discrepancies with previously published data.
IH NK cells can also be inhibited by continuous engagement of activating receptors, as recently reported.30 Because during HCV infection hepatocytes up-regulate NKG2D ligands,17 and because we found up-regulation of NKG2D on IH NK cells of HCV patients, we speculate that incessant triggering of NKG2D receptor by ligands expressed on infected or stressed hepatocytes could induce NK cell exhaustion.
Interestingly, cross-tolerization to many unrelated NK receptors can be caused by sustained engagement of NKG2D31 and this would result in a more general impairment of NK cell function. In relation to that, several articles demonstrated the requirement of ERK mitogen-activated protein kinases for NK cell cytotoxicity,32-34 and many NK receptors, including NKG2D, NKp30, and NKp46, merge into a common cascade that ends in ERK activation independently of the proximal signaling used.35 Unfortunately, it was not possible to analyze ERK phosphorylation in IH NK cells because of the limited number of cells retrieved from liver biopsies, and, consequently, we tested whether ERK phosphorylation in PB NK cells was influenced by HCVcc-conditioned media. In line with the data showing reduced TRAIL and CD107a expression, ERK phosphorylation was impaired in NK cells exposed to HCVcc-conditioned media in chronic HCV infection, suggesting that alteration of this signaling pathway was responsible for this defect, in agreement with previous studies reporting reduced ERK phosphorylation in NK cells after engagement with the HCV E2 protein.21, 36
In conclusion, our study reveals novel, hitherto unrecognized findings in the IH compartment of patients with chronic HCV infection showing impaired NK cytolytic function associated with TRAIL down-regulation, which may in part explain HCV persistence in the host liver. An alternative, though not mutually exclusive, explanation could be that dysfunctional NK cells may contribute to maintain low-level liver inflammation while preventing excessive liver damage, as inferred for IH regulatory T cells.24, 37