Cytokines induced during chronic hepatitis B virus infection promote a pathway for NK cell-mediated liver damage

Abstract

Hepatitis B virus (HBV) causes chronic infection in more than 350 million people worldwide. It replicates in hepatocytes but is non-cytopathic; liver damage is thought to be immune mediated. Here, we investigated the role of innate immune responses in mediating liver damage in patients with chronic HBV infection. Longitudinal analysis revealed a temporal correlation between flares of liver inflammation and fluctuations in interleukin (IL)-8, interferon (IFN)-alpha, and natural killer (NK) cell expression of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) directly ex vivo. A cross-sectional study confirmed these findings in patients with HBV-related liver inflammation compared with healthy carriers. Activated, TRAIL-expressing NK cells were further enriched in the liver of patients with chronic HBV infection, while their hepatocytes expressed increased levels of a TRAIL death-inducing receptor. IFN-alpha concentrations found in patients were capable of activating NK cells to induce TRAIL-mediated hepatocyte apoptosis in vitro. The pathogenic potential of this pathway could be further enhanced by the ability of the IFN-alpha/IL-8 combination to dysregulate the balance of death-inducing and regulatory TRAIL receptors expressed on hepatocytes. We conclude that NK cells may contribute to liver inflammation by TRAIL-mediated death of hepatocytes and demonstrate that this non-antigen-specific mechanism can be switched on by cytokines produced during active HBV infection.

Figure 1.

IL-8 and IFN-α concentrations are elevated in the serum of CHB patients with liver inflammation. (a) Circulating concentrations of multiple cytokines detected in longitudinal serum samples taken from a representative patient (patient 1) assayed by CBA (IL-8, IL-1β, IL-6, IL-10, TNF, and IL-12p70) and sandwich ELISA (IFN-α). The concentrations of inflammatory cytokines were determined by CBA software or Prism. (b) Temporal relationship between serum IL-8 and IFN-α concentrations and liver inflammation (ALT) and viral load (HBV-DNA) in 4 representative patients of 14 patients assayed. Cross-sectional comparison of IL-8 (c) and IFN-α (d) levels quantitated by sandwich ELISA in healthy donors, HBV patients with low ALT (ALT < 60 IU/l for the last year), and HBV patients with raised ALT (ALT > 60 IU/L at time of sampling). Significance testing was performed using the Mann-Whitney U test.

Figure 2.

Direct ex vivo correlation between NK cell TRAIL expression and liver inflammation in CHB patients. (a) Representative flow cytometry dot plot from a CHB patient stained with mAb to CD3, CD56, and TRAIL, and gated on CD3⁻ cells. The percentages denote the proportion of freshly isolated CD3⁻CD56⁺ NK cells staining with TRAIL. (b) Top: PBMCs from patients with eAg-CHB were stained ex vivo, and the percentage of NK (CD3⁻CD56⁺) -expressing TRAIL upon flow cytometry was correlated with ALT. CD69⁺ NK cells are presented as a percent of total lymphocytes. Bottom: The percent of CD56^bright NK cells out of total CD3⁻CD56⁺ NK cells and the percent of those CD56^bright NK cells that were TRAIL⁺ was plotted against ALT. (c) Cross-sectional analysis of ex vivo surface TRAIL expression on CD3⁻CD56⁺ NK cells from healthy donors, HBV patients with low ALT (ALT < 60 IU/l for the last year), and HBV patients with raised ALT (ALT > 60 IU/L at time of sampling). Significance testing was performed using the Mann-Whitney U test.

Figure 3.

Enrichment of NK cell numbers, TRAIL expression, and activation in the liver compared with periphery. (a) Mononuclear cells from the periphery and liver of a representative CHB patient were stained with antibodies to CD3 and CD56, and the proportion of CD3⁺ T cells, CD3⁺CD56⁺ NKT cells, and CD3⁻CD56⁺ NK cells (highlighted in box) was determined by flow cytometry. NK cells (CD3⁻CD56⁺) from liver-infiltrating (IHL) and circulating (PBL) lymphocytes from five chronically infected HBV patients were assessed ex vivo for CD69 expression (b) and TRAIL expression (d). p-values were determined by the Mann-Whitney U test. (c) Flow cytometry dot plot analysis of a representative CHB patient comparing intrahepatic NK cell activation in the CD56^bright and CD56^dim NK cell subsets. (e) A histogram comparing TRAIL expression on the CD56^bright and CD56^dim NK cell subsets and CD3⁺ T cells isolated from the liver. (f) Paraffin-embedded liver sections taken from seven HBV patients were stained with an anti-TRAIL mAb. The boxed area on the left indicates the field of view on the right panel. TRAIL⁺ cells are stained brown and are highlighted with black arrows. Bar, 40 μm.

Figure 4.

Concentrations of IFN-α observed in patient sera induce increased surface TRAIL expression and activation of NK cells isolated from CHB patients. PBMCs from healthy donors (white bars) and CHB patients (black bars) were incubated for 24 h in vitro with 1,000 U/ml IFN-α, 5 ng/ml IL-8, or IFN-α and IL-8. The effect of this incubation on TRAIL expression (a) and NK cell activation (b) was assessed by flow cytometry analysis with NK cells identified as CD3⁻CD56⁺. Graphs were plotted by subtracting baseline levels of CD69 or TRAIL observed in the untreated controls from those observed after cytokine treatment.

Figure 5.

TRAIL receptor expression on hepatocytes in HBV infection. (a) Paraffin-embedded sections from HBV-infected (left) and healthy control (right) livers were stained with an anti–TRAIL-R2 mAb. Membrane-localized (arrows) and cytoplasmic (*) TRAIL-R2 expression is indicated by the brown chromogen reactivity. Bars: top, 40 μm; bottom, 16 μm. (b) Mean fluorescence intensity of HepG2 TRAIL-R2 levels after IL-8 incubation (10 ng/ml for 24 h) compared with untreated and isotype controls. (c) Mean fluorescence intensity of HepG2 TRAIL-R4 levels after 1,000 U/ml IFN-α for 24 h compared with untreated and isotype controls. These are representative of five separate experiments.

Figure 6.

IFN-α–activated NK cells from CHB patients can mediate TRAIL-induced hepatocyte apoptosis. (a) HepG2 cells were incubated for 24 h with or without 10 ng/ml IL-8. Simultaneously, PBMCs were incubated for 24 h with or without 1,000 U/ml IFN-α. Top: PBMCs were then added to HepG2 at an E/T ratio of 10:1 for 4 h before visualization of caspase activation with the fluorescein-labeled Z-VAD-fmk and detection by flow cytometry, expressed as mean fluorescence intensity (MFI). Bottom: Experimental procedure as above except for the addition of 10 ng/ml of a TRAIL blocking antibody. (b) Representative results of PBMCs from healthy donors, CHB patients with low ALT, and CHB patients with high ALT incubated with 1,000 U/ml IFN-α for 24 h and then assessed for caspase activation of IL-8–treated HepG2 as above. (c) Representative HepG2 caspase induction by ex vivo PBMCs from high ALT HBV patient and reduction upon the addition of TRAIL blocking mAb. (d) Representative HepG2 caspase induction upon the addition of PBMCs taken directly ex vivo from a healthy donor, CHB patient with low ALT, and CHB patient with high ALT. (e) Summary level of HepG2 caspase induction using PBMCs directly ex vivo from HBV patients with liver injury (ALT high patients, n = 6) compared with PBMCs from controls (HBV patients without raised ALT, n = 3; healthy controls, n = 3; P = 0.03, Mann- Whitney U test).

Figure 7.

NK cells from CHB patients can mediate TRAIL-induced apoptosis in primary human hepatocytes. Primary human hepatocytes were cultured for 48 h with the addition of 10 ng/ml IL-8 and 1,000 U/ml IFN-α for the last 24 h. Simultaneously, PBMCs from three healthy donors, three CHB patients with low ALT, and four CHB patients with high ALT were incubated with or without IFN-α for 24 h at 37°C. The hepatocytes and PBMCs were incubated together for 18 h at an E/T ratio of 10:1, with or without a TRAIL blocking antibody in the IFN-treated wells. The degree of apoptosis was determined by in situ DNA end labeling (ISEL) for the detection of DNA fragmentation. (a) A representative image of control hepatocytes incubated without PBMCs. Bar, 40 μm. (b) A representative image of hepatocytes after an 18-h incubation with PBMCs from a CHB patient with high ALT. The arrows represent ISEL⁺ hepatocytes. (c) Summary data of percentage of ISEL⁺ hepatocytes using PBMCs from high ALT patients (n = 4) versus controls (low ALT patients, n = 3; healthy donors, n = 3) without (white bars) or with (black bars) IFN-α treatment and with TRAIL blocking of IFN-α–treated wells (hatched bars). Results are presented after subtraction of the mean baseline level of hepatocyte apoptosis of 14% seen without the addition of PBMCs, and significance was tested with the Mann-Whitney U test.