Interferon regulatory factor-3 activation, hepatic interferon-stimulated gene expression, and immune cell infiltration in hepatitis C virus patients


  • Daryl T.-Y. Lau,

    Corresponding author
    1. Liver Center, Department of Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA
    2. Hepatitis Research Center, Institute for Human Infections & Immunity, The University of Texas Medical Branch, Galveston, TX
    • Liver Center, Division of Gastroenterology, Department of Medicine, 110 Francis Street, Beth Israel Deaconess Medical Center, Harvard Medical School, Suite 4A, Boston, MA 02215
    Search for more papers by this author
    • fax: 617-632-1093

  • Penny Mar Fish,

    1. Department of Microbiology, The University of Texas Southwestern Medical Center, Dallas, TX
    Search for more papers by this author
  • Mala Sinha,

    1. Bioinformatics Program, The University of Texas Medical Branch, Galveston, TX
    Search for more papers by this author
  • David M. Owen,

    1. Department of Microbiology, The University of Texas Southwestern Medical Center, Dallas, TX
    2. Department of Immunology, School of Medicine, University of Washington, Seattle, WA
    Search for more papers by this author
    • David M. Owen and Michael Gale Jr are currently affiliated with the Department of Immunology, School of Medicine, University of Washington, Seattle, WA.

  • Stanley M. Lemon,

    1. Hepatitis Research Center, Institute for Human Infections & Immunity, The University of Texas Medical Branch, Galveston, TX
    Search for more papers by this author
  • Michael Gale Jr

    1. Department of Microbiology, The University of Texas Southwestern Medical Center, Dallas, TX
    2. Department of Immunology, School of Medicine, University of Washington, Seattle, WA
    Search for more papers by this author
    • David M. Owen and Michael Gale Jr are currently affiliated with the Department of Immunology, School of Medicine, University of Washington, Seattle, WA.

  • Potential conflict of interest: Nothing to report.


Interferon regulatory factor-3 (IRF-3) activation directs α/β interferon production and interferon-stimulated gene (ISG) expression, which limits virus infection. Here, we examined the distribution of hepatitis C virus (HCV) nonstructural 3 protein, the status of IRF-3 activation, and expression of IRF-3 target genes and ISGs during asynchronous HCV infection in vitro and in liver biopsies from patients with chronic HCV infection, using confocal microscopy and functional genomics approaches. In general, asynchronous infection with HCV stimulated a low-frequency and transient IRF-3 activation within responsive cells in vitro that was associated with cell-to-cell virus spread. Similarly, a subset of HCV patients exhibited the nuclear, active form of IRF-3 in hepatocytes and an associated increase in IRF-3 target gene expression in hepatic tissue. Moreover, ISG expression profiles formed disease-specific clusters for HCV and control nonalcoholic fatty liver disease patients, with increased ISG expression among the HCV patients. We identified the presence of T cell and plasmacytoid dendritic cell infiltrates within all biopsy specimens, suggesting they could be a source of hepatic interferon in the setting of hepatitis C and chronic inflammatory condition. Conclusion: These results indicate that HCV can transiently trigger IRF-3 activation during virus spread and that in chronic HCV, IRF-3 activation within infected hepatocytes occurs but is limited. (HEPATOLOGY 2007.)

Hepatitis C is a significant public health problem affecting nearly 200 million people globally.1 Exposure to hepatitis C virus (HCV) typically results in chronic infection in which the immune response fails to control virus replication and spread. This leads to progressive liver disease that ranges from mild inflammation to severe fibrosis, as well as cancer.2, 3 The immune response to HCV is likely to be triggered initially by innate immune signaling within the infected hepatocyte. An important part of this process involves the recognition of intracellular viral RNA by the product of retinoic acid inducible gene I (RIG-I), leading to RIG-I signaling of downstream activation of interferon regulatory factor-3 (IRF-3).4, 5 IRF-3 may also be activated through recognition of extracellular viral double-stranded RNA by Toll-like receptor 3 (TLR3), which is also expressed by hepatocytes.6 Active IRF-3 translocates from a cytosolic compartment to a nuclear compartment and promotes the expression of specific target genes, including the α/β interferons (IFNs). IFN receptor engagement by secreted IFNs drives a tissue-wide response inducing the expression of hundreds of interferon-stimulated genes (ISGs) that impart antiviral and immunomodulatory activities.7 HCV can control the actions of IRF-3 through the proteolytic targeting of the interferon beta promoter stimulator-1 (IPS-1; also known as MAVS, VISA, and Cardif) and Toll-IL-1 receptor domain-containing adaptor inducing IFN-β (TRIF) adaptor proteins by the viral nonstructural 3/4A (NS3/4A) protease in infected cells, thereby ablating virus activation of IRF-3 and potentially attenuating the expression of IRF-3 target genes that may otherwise control infection.8–11

The hepatic response to HCV infection has been studied in vivo with functional genomics and biochemical approaches to evaluate human liver from patients with chronic infection and chimpanzee liver from animals undergoing experimental HCV infection. Functional genomics analyses of cohorts of patients with HCV have shown that infection is associated with a hepatic response profile marked by differential levels of immune cell infiltration and expression of ISGs whose levels vary widely among patients.12 These observations suggest that during the course of typical asynchronous infection, HCV can both trigger and control the hepatic innate immune response in vivo. Moreover, immune effector cells are an important source of α/β IFNs,13–15 and their production of IFN as hepatic infiltrates may also impact ISG expression to possibly modulate HCV infection.16

The relationship between the IRF-3 activation state in hepatocytes and intrahepatic ISG expression in HCV patients has not been examined but could provide important clues to the cellular source of the ISGs known to be expressed in the liver during HCV infection. In the current study, we addressed the hypothesis that NS3/4A protease of HCV, by disrupting the viral activation of IRF-3 and IFN production, should attenuate ISG expression in infected hepatocytes. We compared the IRF-3 activation status, ISG expression, and specific immune cell infiltration among liver biopsy samples from IFN treatment–naïve chronic hepatitis C and nonalcoholic fatty liver disease (NAFLD; nonviral hepatitis control) patients. Correlations were made between the presence of nuclear IRF-3 and ISG expression.


ALT, alanine aminotransferase; BDCA2, blood dendritic cell antigen 2; HCV, hepatitis C virus; IFA, immunofluorescence assay; IFN, interferon; IPS-1, interferon beta promoter stimulator-1; IRF-3, interferon regulatory factor-3; ISG, interferon-stimulated gene; MOI, multiplicity of infection; mRNA, messenger RNA; NAFLD, nonalcoholic fatty liver disease; NS, nonstructural; PBS, phosphate-buffered saline; pDC, plasmacytoid dendritic cell; RIG-I, retinoic acid inducible gene-I; STAT1, signal transducer and activator of transcription 1; TLR3, Toll-like receptor 3; WPGMA, weighted pair-group method with arithmetic mean.

Patients and Methods

Cells and Virus.

Huh7 and Huh7.5 cells have been described.5, 17 The JFH1 HCV 2A infectious clone was a gift from Dr. T. Wakita,18 and virus was prepared from Huh7 producer cells (a gift from Dr. G. Luo) as described by Cai et al.19 Huh7 or Huh7.5 cells were infected with JFH1 virus at a multiplicity of infection (MOI) of 0.3. This MOI allowed us to establish an asynchronous HCV infection in vitro. To monitor the cell-to-cell spread of virus, we assessed viral protein expression by microscopy, using an immunostaining procedure to visualize infected cells.10 HCV proteins were detected with a well-characterized hyperimmune patient serum that detects the HCV nonstructural proteins.10 For analysis of gene expression, cells were harvested, RNA was extracted, and reverse-transcription quantitative polymerase chain reaction was performed to measure specific ISG products with methods and primer sets as described.5

Patients and Tissue Samples.

Patients were recruited from University of Texas Medical Branch outpatient clinics and underwent liver biopsy as part of a standard medical evaluation. Patients enrolled in the study gave informed consent for liver tissue in excess of 2.0 cm to be used for research. No biopsy was obtained for research purposes. We studied tissue samples from 15 patients with treatment-naïve chronic hepatitis C and 13 patients with NAFLD. All the liver samples were obtained prior to initiation of therapy. The clinical features of the patients are shown in Table 1. Histological features of the liver biopsies were interpreted by a liver pathologist who was blinded to other clinical data. Subject informed consent was obtained, and the study was approved by the Institutional Review Board.

Table 1. Clinical Features of the Patients with Chronic Hepatitis C and NAFLD
IDAge (Years)GenderHCV GenotypeHCV RNA Serum (IU/mL)ALT (U/L)Liver Biopsy
Inflammation (Grade)Fibrosis (Stage)Steatosis
  • Abbreviations: ALT, alanine aminotransferase; HCV, hepatitis C virus; NAFLD, nonalcoholic fatty liver disease.

  • *

    Nuclei IRF-3(+) HCV cases.

HCV 1*55F1a7.7 × 106108220%
HCV 246F1a2.5 × 10660100%
HCV 3*45F1b5.8 × 10537010%
HCV 4*61M2b1.6 × 10668010%
HCV 546F3a1.3 × 105481130%
HCV 6*46F1a1.0 × 106171340%
HCV 754M1b3.7 × 1061162140%
HCV 8*47M3a1.1 × 10521100%
HCV 9*67F1a1.9 × 106442330%
HCV 10*40F1b1.8 × 10532105%
HCV 1152M13.9 × 10357100%
HCV 1220M1b2.5 × 10562100%
HCV 1318F3a1.2 × 10456110%
HCV 14*47M1b7.7 × 10659210%
HCV 15*48F12.8 × 10514110%
Mean46.1  1.8 × 10663.51.31.17%
IDAge (Years)GenderALT (U/L)Liver Biopsy
Inflammation (grade)Fibrosis (stage)Steatosis
NAFLD 163.2F34115%
NAFLD 253.6M331110%
NAFLD 347.4F24005%
NAFLD 443.4F380130%
NAFLD 544.1F341120%
NAFLD 673.3M45005%
NAFLD 760.3M291120%
NAFLD 854.8F1321250%
NAFLD 951.0F362170%
NAFLD 1057.8F442140%
NAFLD 1155.6F291240%
NAFLD 1239.3F402280%
NAFLD 1360.2M753260%

Microarray Analysis of ISGs.

Complementary RNA probes were prepared from total RNA extracted from liver biopsy material and subsequently hybridized to HG-U95A version 2 Human GeneChips (Affymetrix Inc., Santa Clara, CA), each of which contains 12,625 probe sets (representing mostly full-length human genes). Details of the methods of probe preparation and analysis of the Affymetrix data have been published previously.20 In brief, scaled hybridization signal intensities for those probe sets representing RNA transcripts were analyzed by one-way analysis of variance to identify genes for which differences in detectable transcript abundance (scaled hybridization signal intensity) were likely to be due to the disease state (NAFLD versus chronic hepatitis C) rather than random variation. The weighted pair-group method with arithmetic mean (WPGMA) hierarchical clustering method was subsequently applied to construct dendrograms and a heat map on the representative ISG probe sets identified by this approach, mostly at a P value ≤0.01.

Immunohistochemical Staining and Confocal Microscopy.

For sections of paraffin-embedded liver biopsies, we used the following procedure involving 4 major steps, including deparaffinization of tissue, antigen retrieval, tissue permeabilization, and immunostaining. Slide-mounted paraffin-embedded biopsy tissue was deparaffinized first by the placement of slides on a 58°C heat block for 5 minutes in order to preheat them to the working temperature. Preheated slides were then placed in a glass Coplin jar containing 1× EZ Dewax solution (Biogenex) and were incubated for 5 minutes at 20°C. Slides were then placed in xylene and incubated for 5 minutes at 20°C followed by sequential 5-minute incubations in Coplin jars, each containing 100% ethanol, 95% ethanol, or distilled ultrapure water. Slides were then gently rinsed in immunofluorescence assay (IFA) buffer [phosphate-buffered saline (PBS) containing 0.05% Tween-20] and were then subjected to antigen retrieval. For antigen retrieval of deparaffinized slide-mounted tissue, slides were placed in a plastic Coplin jar containing 1× AR10 solution (Biogenex), and a vented lid was placed on the jar. The jar was then set inside a microwave oven and incubated with the oven setting on high with constant monitoring through the oven door to define the first point of boiling of the liquid within the jar. The oven power was turned off when the first point of boiling was observed, the liquid was removed from the jar, and the jar was refilled with 1× AR10 solution maintained at 4°C. The jar (containing the slides) was then allowed to cool for 10 minutes at room temperature. This cycle of heating and cooling was repeated, after which the jar was allowed to cool at room temperature for 45 minutes. Slides were then removed from the jar and rinsed sequentially with distilled ultrapure water and PBS. Tissues were permeabilized by the placement of slides in a Coplin jar containing PBS/1.0% TritonX-100 for 10 minutes. Slides were then rinsed with PBS.

For immunostaining, slides were first incubated in PBS containing 10% normal goat serum for 1 hour. After rinsing with PBS, primary antibody was applied, and the slides were incubated overnight at 20°C in a humidified chamber, after which slides were rinsed three times with IFA buffer followed by application of secondary polyclonal antibody coupled to Alexa488 or Alexa 594 fluorophores (Molecular Probes). After a 1-hour incubation at 20°C, the slides were rinsed three times with IFA buffer and allowed to dry, and cover slips were mounted with mounting medium as described.10 Tissues were visualized with a Zeiss Axiovert 200M inverted microscope with the LSM5 Pascal laser confocal module with a 25-mW ultraviolet diode laser emitting light at a wavelength of 405 nm, a 30-mW argon laser emitting light at a wavelength of 458, 488, or 514 nm, and a 1-mW HeNe laser emitting light at a wavelength of 453 nm. Digital images were collected as 0.3 μM optical sections and were processed with Zeiss Image Examiner software. Multiple images were collected for each sample analyzed. For quantification of cell frequency, the images of the triple-stained samples were imported into the analysis software package Metamorph version 6.2 (Molecular Devices, Downingtown, PA). On each field, the total number of cells was estimated by the counting of the nuclei with the “Automated Morphometry Analysis” feature, whereas the number of positive cells per field was determined by an expert observer using the function “Manually Count Objects”. All counts were automatically logged into Microsoft Excel for statistical analysis.

The primary antibodies used were mouse monoclonal antibody to HCV NS3 (catalog no. NCL-HCV-NS3, Novo Castra Laboratories), rabbit polyclonal antibody to human IRF-3 (a gift from Dr. Michael David, University of California at San Diego), blood dendritic cell antigen 2 (BDCA2) mouse monoclonal antibody (Miltenyi Biotech), mouse monoclonal antihuman CD3–fluorescein isothiocyanate (Chemicon), and rabbit polyclonal anti-ISG15 (a gift from Dr. A Haas). Nuclei were visualized through the staining of cells or tissues with 4′,6-diamidino-2-phenylindole. For examination of IRF-3 and HCV proteins in Huh7 or Huh7.5 cells, the cells were cultured on chamber slides and were fixed and processed for immunostaining exactly as described.10

Statistical Methods.

Unless otherwise indicated, comparisons between two groups were carried out with the Student t test and between multiple groups by one-way analysis of variance. Where indicated, the chi-square test was used. A P value of less than 0.05 was considered statistically significant unless otherwise noted.


A previous study demonstrated that HCV infection can trigger a transient activation of IRF-3 in vitro. However, these studies were limited to relatively high-MOI, synchronous infection of cells.10 By comparison, HCV infection in vivo represents a long-term process of asynchronous infection and cell-to-cell virus spread. We therefore examined IRF-3 activation in Huh7 cells undergoing asynchronous HCV infection initiated under conditions of low-MOI virus challenge. In mock-infected control cells, IRF-3 was present in its resting cytoplasmic state. At 36 hours after virus challenge, we detected the active, nuclear isoform of IRF-3 in a small proportion of cells (Fig. 1A). In particular, activated IRF-3 was detectable only in the nuclei of cells harboring low levels of HCV proteins and was not detected in the nucleus of cells that stained brightest for viral protein expression, as determined by immunofluorescence microscopy. This pattern of nuclear IRF-3 abundance and relationship with viral protein levels suggest that IRF-3 activation can occur during a low, asynchronous infection. In this system, the low-frequency activation of IRF-3 was associated with a detectable level of messenger RNA expression of ISG56 and ISG15, target genes of IRF-3 stimulation (Fig. 1B).21 Thus, during low-MOI and asynchronous infection, HCV can impart ISG expression.

Figure 1.

Low-level IRF-3 activation and ISG expression during HCV infection in vitro. Huh7 cells were cultured alone or were infected with HCV 2A at a multiplicity of infection of 0.3. At the time points shown in hours, the cells were immunostained for detection of HCV proteins and IRF-3 or were harvested for analysis of ISG expression. (A) Fluorescence micrographs (20× magnification) of mock-infected cells (upper) or HCV-infected cells (lower) costained with patient antiserum specific to the HCV nonstructural proteins (green) and rabbit serum specific to IRF-3 (red). Nuclei were stained with 4′,6-diamidino-2-phenylindole (blue). Arrows mark the cells exhibiting nuclear IRF-3. (B) ISG56 and ISG15 expression was measured by reverse-transcription quantitative polymerase chain reaction analysis of RNA recovered from cells that were mock-infected/mock-treated (white bars), treated with 10 units/mL IFN-α for 18 hours (gray bars), or infected with HCV for 36 hours (black bars). HCV, hepatitis C virus; IFN, interferon; IRF-3, interferon regulatory factor-3; ISG, interferon-stimulated gene; and mRNA, messenger RNA.

To determine if IRF-3 activation is a characteristic of HCV infection that results in measurable ISG expression, we examined the hepatic IRF-3 activation status and gene expression profiles of patients with chronic HCV infection, comparing these results with similar studies of patients with NAFLD. Chronic hepatitis C and NAFLD are both important causes of chronic liver disease that can lead to hepatic inflammation and progressive fibrosis. The characteristics of the patients that we studied are summarized in Table 1. The nonviral etiology of NAFLD provided a useful control for analysis of IRF-3 activation and gene expression in HCV-infected liver.

IRF-3 Distribution in Chronic HCV and NAFLD Patients.

Applying immunohistochemical staining and confocal microscopy approaches to sections of liver biopsies from HCV-infected patients, we were able to identify the activated isoform of IRF-3 in the nuclei of some hepatocytes. We found that IRF-3 is widely expressed in liver tissue and that hepatocytes in both the NAFLD and HCV tissues had a high abundance of cytoplasmic IRF-3. However, although none of the NAFLD samples demonstrated the active, nuclear isoform of IRF-3 within hepatocytes, 9 of 15 (60%) of the HCV tissues demonstrated positive nuclear staining for IRF-3. The proportion of nuclei containing the activated IRF-3 isoform varied from 2.2% to 18% of the hepatocytes analyzed in each image field.

We coupled this analysis with concomitant immunostaining of liver sections with antibody against the HCV NS3 protein, specific ISGs, or immune cell markers. Among HCV-infected patients, the NS3 protein was detected in 9% to 58% of the hepatocytes in a cytoplasmic/perinuclear context (Table 2). We consistently observed two general patterns of staining in which NS3 either was not detected (Fig. 2A) or was present within focal clusters of hepatocytes (Fig. 2C,D) in a given field. When present, the active nuclear isoform of IRF-3 was focally distributed in hepatocytes in specimens from the HCV-infected patient cohort, possibly reflecting focal regions of infection. It is notable that when present in the hepatocyte nucleus of patient liver tissue, IRF-3 exhibited a partial nuclear accumulation with residual IRF-3 in the cytoplasm, whereas cultured Huh7 cells exhibited patterns of nuclear IRF-3 ranging from partial to compete nuclear accumulation in HCV-infected cells. Cells exhibiting nuclear IRF-3 were identified in either case adjacent to cells that stained strongly positive for the presence of HCV proteins (compare Figs. 1A and 2A). As shown in the example in Fig. 2E, we found that IRF-3 was consistently excluded from the nucleus in NS3-positive hepatocytes within patient tissue. This pattern most likely reflects the blockade of IRF-3 activation imposed by the NS3 protein that has accumulated within infected cells.10 In contrast, NS3 and the active nuclear isoform of IRF-3 were not detected within tissue specimens from any of the NAFLD patient cohort (data not shown).

Table 2. Comparison of the Features of HCV Nuclei IRF-3(+), HCV Nuclei IRF-3(−), and NAFLD Cases
 HCVP Value for HCV: IRF-3(+) Versus IRF-3(−)*NAFLD (n = 13)
IRF-3(+) Nuclei (n = 9)IRF-3(−) Nuclei (n = 6)
  • Abbreviations: ALT, alanine aminotransferase; HCV, hepatitis C virus; IRF-3, interferon regulatory factor-3; NAFLD, nonalcoholic fatty liver disease; NS3, nonstructural 3.

  • *

    The P value reflects the comparison between HCV nuclei IRF-3(+) and HCV nuclei IRF-3(−) cases.

Age [years (range)]51 (40-67)39 (18-54)0.254 (39-73)
HCV genotype 17/94/60.6NA
HCV RNA [MIU/mL (range)]2.3 × 106 (6 × 105 to 7.7 × 106)1.1 × 106 (3.9 × 103 to 3.7 × 106)0.3NA
ALT [IU/L (range)]62 (14-171)66 (48-116)0.846 (29-132)
Inflammation [grade (range)]1.3 (0-3)1.2 (1-2)0.71.2 (0-3)
Fibrosis [stage (range)]1.4 (0-4)0.5 (0-1)0.11.1 (0-2)
NS3-positive [cells (range)]23% (9%-58%)18% (12%-31%)0.20%
Figure 2.

NS3 and IRF-3 staining of HCV-infected human liver specimens. Panels show confocal micrographs of liver biopsy specimens from HCV-infected patients. Tissues were costained with antibodies against HCV NS3 (green) and IRF-3 (red); nuclei were stained with 4′,6-diamidino-2-phenylindole (blue). Images show 0.3 μM confocal optical sections. (A-D) These images were collected at 20× magnification and show complete fields of four representative image sets ranging from (A) no detectable NS3 staining to (B-D) focal NS3 staining. (E) This image set presents the same specimen shown in panel D but was collected at 40× magnification, showing IRF-3 staining and a merged composite image of a smaller field. Arrows mark the cells expressing nuclear IRF-3. The bar indicates 10 μM. HCV, hepatitis C virus; IRF-3, interferon regulatory factor-3; NS3, nonstructural 3.

Innate Immune Response Gene Expression and IRF-3 Activation Profile in Patient Cohorts.

To assess the intrahepatic expression of ISGs, we conducted functional genomics analysis of total RNA isolated from frozen liver biopsy specimens. Hierarchical clustering identified a bioset of innate response genes that included IRF-3 target genes and ISGs. This gene set suggests that ISGs are differentially expressed between the two disease states, namely, NAFLD and HCV at P < 0.01 (Fig. 3A). Interestingly, the profiles of 2 patients with chronic HCV infection segregated into cluster A, which included predominantly NAFLD specimens; likewise, the gene expression of 2 NAFLD samples segregated into cluster B with predominantly chronic HCV. The 2 HCV patients that were grouped into cluster A had the lowest level of viremia, at 3925 and 12,261 IU/mL, respectively, in the cohort. There was no distinguishing characteristic of the 2 NAFLD samples that were grouped into cluster B. HCV samples generally showed increased expression of ISGs and IRF-3 target genes compared to NAFLD cases, consistent with the lack of expression of nuclear IRF-3 in the latter (see Fig. 2). A number of genes such as GBP2 guanylate binding protein-2 (GBP2), IFRD1 interferon-related developmental regulator-1 (IFRD1), TRIF, protein kinase RNA activated (PRKRA), interferon alpha-10 (IFNA10), interferon omega-1 (IFN W1), and interferon-related developmental regulator-2 (IFRD2) formed a separate cluster and were not significantly different between NAFLD and HCV samples, and this suggests that their expression may be a more general characteristic of hepatic inflammation.

Figure 3.

Gene expression profiles in human liver tissues. (A) Heat map produced by hierarchical clustering of intrahepatic ISG expression profiles derived from Affymetrix GeneChip assays of biopsies. Tissue samples clustered into 2 main groups (cluster A and cluster B) defined by gene expression profile and disease pathogenesis. Red color indicates an increased relative abundance of the transcript in the indicated tissue sample, whereas green indicates a decreased relative abundance according to the scale shown. Dendrograms at the top and to the left show the relatedness of the expression patterns for individual tissue samples and individual probe sets (genes), respectively. (B) Hybridization signal intensities of representative IFN-induced transcripts between HCV samples with and without positive IRF-3 nuclei stain. Results are shown as the mean hybridization signal intensities obtained with tissue samples grouped by IRF-3 nuclei stain patterns; error bars represent the standard deviation. HCV, chronic hepatitis C virus infection; GIP2, interferon, alpha-inducible protein; IFIT, interferon-induced protein with tetratricopeptide repeats; IFITM, interferon inducible transmembrane; IFN, interferon; IRF-3, interferon regulatory factor-3; ISG, interferon-stimulated gene; MX1, myxovirus resistance-1; NAFLD, nonalcoholic fatty liver disease; PRKRA, protein kinase, interferon-inducible double stranded RNA dependent inhibitor; STAT1, signal transducer and activator of transcription 1; TLR3, Toll-like receptor 3.

Subsequent analyses focused on the 15 liver biopsy samples from patients with HCV infection (Table 1). Each of these patients had HCV RNA detectable in serum, with the majority infected with genotype 1 HCV. We compared the signal intensities of representative ISGs in the liver samples with and without nuclei containing IRF-3. The liver tissues from the 9 patients staining positive for nuclear IRF-3 segregated into cluster B and were associated with significantly higher expression levels of a number of innate immune response genes that included IRF-3 target genes such as ISG56, ISG44, and ISG54 and ISGs such as ISG6-16 and signal transducer and activator of transcription 1 (Fig. 3B). There was no significant difference in the proportion of NS3-positive hepatocytes in specimens that did or did not exhibit nuclear IRF-3 (Table 2). Furthermore, those with and without nuclear IRF-3 expression had similar serum HCV RNA and alanine aminotransferase levels, hepatic inflammation, and fibrosis (Table 2). The insignificant P values are likely a reflection of the dynamic nature of hepatocyte turnover and HCV infectivity. The P values compared the mean values between the 2 groups but did not take into account the different durations of infection of the hepatocytes within the same liver tissue. Similarly, the hepatic inflammation and fibrosis represent the cumulative results of chronic infection over variable periods of time.

To verify the expression patterns of ISG56 and ISG15, we conducted immunostaining and confocal microscopy analysis of patient liver tissues. Figure 4 shows representative images from NAFLD and chronic HCV patients whose gene expression profile segregated into cluster A or cluster B. The staining patterns of ISG56 and ISG15 expression revealed two main patterns of low (cluster A, Fig. 4A) or high (cluster B, Fig. 4B) abundance in these patients. In cluster B patients, ISG15 and ISG56 were typically expressed in focal areas in hepatocytes that could reflect localized IFN production and response. This expression pattern was not observed in cluster A patients, who overall demonstrated weaker staining patterns with fewer cells exhibiting ISG15 or ISG56 expression, thus directly validating the microarray data (see Fig. 3A). Quantification of ISG15-positive cells in tissues from individual patients further confirmed the two response clusters as exhibiting a low or high ISG response (Fig. 4C). Moreover, a higher frequency of ISG15-positive and ISG56-positive cells was observed within specimens of HCV patients that expressed the nuclear IRF-3 isoform (Fig. 4D).

Figure 4.

Expression and distribution of ISG15 and ISG56 in human liver tissues. Tissues were stained with antibodies to ISG15 or ISG56 (red), and nuclei were stained with 4′,6-diamidino-2-phenylindole (blue) and were analyzed as 0.3 μM optical sections collected by confocal microscopy. Panel sets show 10× (left) and 20× magnifications (right) of merged color images from the indicated chronic HCV and NAFLD specimens that segregated into (A) cluster A and (B) cluster B as defined in Fig. 3. (C) The frequency of ISG15+ cells within multiple 20× fields was quantified by cell counting and is expressed as a mean from cluster A and cluster B specimens (as indicated in Fig. 3). (D) The frequency of ISG15+ cells within specimens expressing (IRF-3+) or lacking nuclear IRF-3 (IRF-3−) was quantified within multiple 20× fields by cell counting. Bars represent the mean frequency of ISG15+ cells. HCV, hepatitis C virus; IRF-3, interferon regulatory factor-3; ISG, interferon-stimulated gene; NAFLD, nonalcoholic fatty liver disease.

Immune Cell Infiltration.

We next examined the patient liver tissues for the presence of infiltrating T cells and plasmacytoid dendritic cells (pDCs). In order to assess possible pDC association with hepatic ISG expression, tissues were costained with the BDCA2 antibody, which identifies pDCs, and with anti-ISG15. To define hepatic T cells, liver specimens were costained with anti-CD3 and anti–IRF-3. Because IRF-3 is highly abundant in hepatocytes, it served as a general cell stain for the hepatic tissue. Specimens were analyzed by confocal microscopy. Figure 5 shows representative confocal images of the staining patterns observed in the patient specimens. We identified pDC and T cell infiltrates within hepatic tissues from both NAFLD and chronic HCV patients. In general, patients whose gene expression profile segregated to cluster A exhibited a lower frequency and staining intensity of CD3 or BDCA2+ cells in their liver than patients whose gene expression profile segregated to cluster B (Fig. 5A,B). In some cases, we observed an overlap in the distribution of pDCs and ISG15-positive cells in NAFLD samples (Fig. 5B). Overall, however, there were no significant differences in the frequency of hepatic pDCs or T cells between NAFLD and HCV-infected patient tissues (Fig. 5C).

Figure 5.

Distribution of pDCs and T cells in chronic HCV and NAFLD liver tissues. Tissues from the indicated HCV and NAFLD patients of (A) cluster A or (B) cluster B (as defined in Fig. 3) were costained in sets. Tissues were analyzed by confocal microscopy. Image sets show a 20× field collected from a 0.3 μM optical section for each analysis, except for the BDCA2/ISG15 image set of NAFLD-8 shown in panel B, which was collected at 40× magnification in order to demonstrate the overlapping distribution of each stain. Stains are shown alone and as merged composites as indicated. For each patient, the left image set shows tissue stained with the BDCA2 monoclonal antibody (green) and an antibody against ISG15 (red). The right image set shows tissue stained with anti-CD3 monoclonal antibody (green) and anti–IRF-3 (red). Nuclei were stained with 4′,6-diamidino-2-phenylindole (blue). (C) The frequency of BDCA2+ cells or CD3+ cells within multiple 20× fields within liver tissue that segregated into cluster A or cluster B (defined in Fig. 3) was quantified by cell counting. Bars show the mean frequency for each. BDCA2, blood dendritic cell antigen 2; HCV, hepatitis C virus; IRF-3, interferon regulatory factor-3; ISG, interferon-stimulated gene; NAFLD, nonalcoholic fatty liver disease; and pDC, plasmacytoid dendritic cell.


Our results provide evidence for interactions of HCV with the innate immune response in vitro and in vivo and lend support to a model of dynamic regulation of this response during infection. HCV RNA is a potent trigger of RIG-I signaling events that effect IRF-3 activation,4, 5 and this occurs early during a discrete window period prior to viral inactivation of IPS-1 in cells undergoing synchronous HCV infection in vitro.10 Recognition of double-stranded RNA by TLR3 can also lead to IRF-3 activation and ISG expression, but its role in HCV infection is not clear, and TLR3 is not expressed at levels sufficient for virus activation of IRF-3 in the Huh7 cells used to model HCV infection in cell culture.6 Hence, it remains uncertain whether HCV infection can activate IRF-3 through this pathway. Studies of HCV-infected chimpanzees show that infection is asynchronous in vivo and that viral load is reduced with the onset of innate immune defenses and hepatic ISG expression.22 Our analysis of asynchronous HCV infection in vitro identified a low frequency of cells expressing the nuclear, active form of IRF-3, suggesting that IRF-3 is activated, at least transiently, during early infection and subsequent cell-to-cell spread of HCV. The presence of nuclear IRF-3 in HCV-infected cultures was a characteristic of cells exhibiting low or undetectable levels of viral proteins, and IRF-3 was excluded from the nucleus in cells exhibiting the brightest staining for HCV proteins (see Fig. 1). These observations are consistent with a recent study and a model suggested by us previously involving dynamic control of the innate immune response by HCV in which IRF-3 activation is triggered through viralpathogen-associated molecular pattern engagement by RIG-I during initial infection.10 The triggering of IRF-3 activation promotes a brief period of time when the products of expressed ISGs may suppress HCV levels. However, as HCV proteins accumulate beyond a threshold level in a given cell, the NS3/4A protease cleaves and inactivates IPS-1 to ablate cell-specific RIG-I signaling of IRF-3 activation and ISG expression.10, 23 We previously found evidence for the variable cleavage of IPS-1 in vivo during chronic HCV infection,10 thus supporting this dynamic model of innate immune control. After fully developing and evaluating techniques for immunohistochemical staining and confocal microscopy of paraffin-embedded clinical liver specimens, we examined the patterns of HCV NS3 protein and IRF-3 distribution with respect to ISG expression and immune cell infiltration in liver samples from patients with chronic hepatitis C and NAFLD. Our results demonstrate that a low frequency of hepatocytes express the nuclear, active isoform of IRF-3 during chronic HCV infection, suggesting that IRF-3 is at least transiently activated in some hepatocytes during infection in vivo.

Importantly, the NS3 protein and nuclear isoform of IRF-3 were detectable only within specimens from the HCV patient cohort. Because the activation of IRF-3 is a viral-mediated process, the consistent absence of the nuclear isoform of IRF-3 in the hepatocytes of the NAFLD patient samples confirmed the specificity of our assays. The heterogeneous patterns of NS3 protein expression among the chronic hepatitis C liver samples are consistent with the constant turnover of the hepatocytes resulting in asynchronous HCV infection, as previously modeled.24 We found that 60% of the HCV patient samples examined showed a low frequency of cells containing detectable IRF-3 in the nucleus and that IRF-3 was excluded from the nucleus of hepatocytes that stained positive for a high abundance of the viral NS3 protein. These observations indicate that HCV can both trigger and control cellular pathways of IRF-3 activation in a dynamic and likely continual process during infection in vivo.

The microarray profiling revealed distinct patterns of hepatic ISG expression between HCV and NAFLD liver samples. Overall, the HCV cohort exhibited generally increased levels of ISGs, although these were variable among patients. Importantly, the HCV liver tissues positive for nuclear IRF-3 were associated with significantly higher levels of RNA transcripts of IRF-3 target genes. However, HCV patients with an absence of detectable nuclear IRF-3 and NAFLD patients also expressed hepatic ISGs, although to lower levels. This suggests that direct IRF-3 activation and triggering of endogenous cellular pathways of the innate immune signaling within hepatocytes may not be the sole mechanism of ISG induction in the liver. NAFLD is a metabolic liver disease and, similar to HCV, is associated with variable degrees of hepatic inflammation and fibrosis. Immune cells that infiltrate the liver could potentially serve as sources of α/β IFN production and could drive an ISG response within hepatocytes through a paracrine mechanism.5 We therefore evaluated the relationship between infiltrating T cells and pDCs in the liver and ISG expression. Hepatic T cells play a critical role in directly controlling HCV infection,25 and cytokines secreted by infiltrating T cells may further suppress virus replication. pDCs specifically represent an important and mobile source of α/β IFN production in response to viruses, and their systemic production of α/β IFN is essential for control of virus infection.13, 14 Both HCV and NAFLD patients had T cells and pDCs distributed in their liver tissues. In the context of HCV, inflammation driven by virus infection might serve to stimulate immune cell infiltration, where a higher viral load could be associated with increased levels of cell infiltrates whose actions might possibly impact hepatic ISG expression levels. In support of this, we found an overall trend toward increased hepatic pDCs and T cells in HCV patients of cluster B compared to cluster A (see Table 1). We found that the patterns of the pDCs and ISG15 distribution were different between the HCV and NAFLD liver samples. There was a significant overlap of pDCs and ISG15 distribution in the NAFLD samples, suggesting that the expression of ISG15 might be directly stimulated by α/β IFNs secreted by locally infiltrating pDCs. In contrast, only a minority of the HCV samples showed ISG15 expression in hepatocytes within tissues staining positive for pDCs.

There are reports that the pDCs isolated from peripheral blood mononuclear cells of patients with chronic HCV infection have reduced frequency and IFN-production capacity compared to healthy controls and patients with NAFLD.26–28 Dolganiuc and coworkers26 also provided evidence that these pDC defects are secondary to the HCV core protein–induced interleukin-10 and tumor necrosis factor-α secretion by monocytes. In the same study, however, they found that there was an increase in hepatic pDC frequency among patients with chronic HCV compared to normal controls. This raises the question of compartmentalization of pDC location and function. In contrast, Longman et al.29 observed that peripheral blood pDCs from HCV-infected patients generated similar levels of IFN-α compared to those generated from healthy control subjects when analyzed on a per-cell basis. Thus, differential IFN production by pDCs that have infiltrated the liver may impact overall hepatic ISG expression during HCV infection. It is important to note, however, that our immunohistochemical staining method could not distinguish between activated and naïve immune cells or between their contributions to IFN production and hepatic ISG expression. Further studies are required to ascertain the role of lymphocytes and pDCs in the hepatic ISG response during HCV infection.

In conclusion, our study provides evidence that HCV can transiently trigger hepatocellular IRF-3 activation and ISG production in the setting of chronic HCV infection. However, the IRF-3 activation is limited, supporting a role for HCV control of innate immune function. We suggest that immune cells infiltrating the liver, especially pDCs, could be an important source of the α/β IFN expressed within the hepatic environment. However, the resulting ISG response is insufficient for viral clearance. An understanding of the mechanisms by which HCV interferes with the host innate immunity will have significant implications on therapeutic intervention of chronic infection.


The authors are grateful for the expert technical assistance on quantification of the confocal images by Dr. Leoncio A. Vergara of the University of Texas Medical Branch Optical Imaging Laboratory.