Intra-hepatic Depletion of Mucosal Associated Invariant T cells in Hepatitis C Virus-induced Liver Inflammation

Abstract

Background & Aims

Chronic hepatitis affects phenotypes of innate and adaptive immune cells. Mucosal associated invariant T (MAIT) cells are enriched in the liver as compared to the blood, respond to intra-hepatic cytokines, and (via the semi-invariant T-cell receptor) to bacteria translocated from the gut. Little is known about the role of MAIT cells in livers of patients with chronic hepatitis C virus (HCV) infection and their fate after antiviral therapy.

Methods

We collected blood samples from 42 patients with chronic HCV infection who achieved a sustained virologic response after 12 weeks of treatment with sofosbuvir and velpatasvir. Mononuclear cells were isolated from blood before treatment, at weeks 4 and 12 during treatment, and 24 weeks after the end of treatment. Liver biopsies were collected from 37 of the patients prior to and at week 4 of treatment. Mononuclear cells from 56 blood donors and 10 livers that were not suitable for transplantation were used as controls. Liver samples were assessed histologically for inflammation and fibrosis. Mononuclear cells from liver and blood were studied by flow cytometry and analyzed for responses to cytokine and bacterial stimulation.

Results

The frequency of MAIT cells among T cells was significantly lower in blood and liver samples of patients with HCV infection than of controls (median 1.31% vs 2.32% for blood samples, P=.0048 and median 4.34% vs 13.40% for liver samples, P=.001). There was an inverse correlation between the frequency of MAIT cells in the liver and histologically determined levels of liver inflammation (r=−.5437, P=.0006) and fibrosis (r=−.5829, P=.0002). MAIT cells from the liver had higher levels of activation and cytotoxicity than MAIT cells from blood (P<.0001). Production of interferon gamma (IFNG) by MAIT cells was dependent on monocyte-derived interleukin 18 (IL18), and was reduced in patients with HCV infection in response to T-cell receptor-mediated but not cytokine-mediated stimulation, as compared to controls. Anti-viral therapy rapidly decreased liver inflammation and MAIT cell activation and cytotoxicity, and increased the MAIT cell frequency among intra-hepatic but not blood T cells. The MAIT cell response to T-cell receptor-mediated stimulation did not change during the 12 weeks of antiviral therapy.

Conclusions

In analyses of paired blood and liver samples from patients with chronic HCV infection before, during and after antiviral therapy with sofosbuvir and velpatasvir, we found that intrahepatic MAIT cells are activated by monocyte-derived cytokines and depleted in HCV-induced liver inflammation.

Introduction

Intrahepatic inflammation is a critical component in the progression of viral hepatitis to liver cirrhosis and hepatocellular carcinoma ^{1, 2}. Chronic stimulation of both innate and adaptive immune cells contributes to this process ³. Innate immune cells can be activated by virus-induced cytokines and bacteria and endotoxins from the gut that reach the liver via the portal vein, particularly in patients with advanced liver disease ^{4, 5}.

Mucosal associated invariant T (MAIT) cells are innate-like lymphocytes which are enriched at barrier sites such as the intestine, liver and lung ⁶. MAIT cells play a role in the host defense against bacterial infections ^{7, 8}. Their semi-invariant Vα7.2-Jα33 T cell receptor recognizes vitamin B metabolites from bacteria such as E. coli presented on major histocompatibility complex class I-related molecules (MR1) ^9–12. A recent study demonstrated that MAIT cells localize around bile ducts in the portal tracts of the liver and respond to bacterial antigens ¹³. This landmark study on liver-infiltrating MAIT cells from patients with end-stage liver disease provided evidence of MAIT cells as an important guardian of the gut-liver axis ¹³.

The fact that bacterial infections are a major cause of morbidity and mortality in patients with chronic viral hepatitis and liver cirrhosis ¹⁴ raises the question whether MAIT cell-dependent protection is impaired. MAIT cells can be activated by cytokines such as type I IFN ¹⁵, IL-7 ¹⁶, IL-12 and IL-18 ¹⁷. The levels of many of these cytokines are increased in the blood and the liver of HCV-infected patients ^{18, 19}. Indeed, alterations in MAIT cell frequency and phenotype have been described in the blood of patients with HCV ²⁰ and other viral infections ^21,22. They have been described to remain functionally impaired in response to their cognate antigen in sofosbuvir/ribavirin treated patients ²⁰. However, MAIT cells have not been studied in the liver in chronic hepatitis C.

With the introduction of direct acting antivirals (DAAs), even HCV patients with advanced liver disease can be treated safely and effectively. Thus HCV is a valuable model system to study immune cell alterations before and after viral clearance ²³. Initial studies on the effect of DAA-induced HCV clearance on innate immune responses yielded discordant results. The levels of several pro-inflammatory cytokines declined but did not normalize 8 months after the end of treatment with sofosbuvir/ribavirin ¹⁸. NK cell effector responses were shown to normalize in asunaprevir/daclatasvir-treated patients ²⁴. In contrast, the majority of virus-specific T cells remained impaired in cytokine secretion and only a small subset of antigen-specific T cells regained their proliferative capacity in faldaprevir/deleobuvir treated patients ²⁵. These cells represent a memory-like subpopulation of HCV-specific T cells that survive in chronic HCV infection and are maintained after treatment-induced HCV clearance ²⁶.

A caveat of these studies is that they focused exclusively on the blood and did not assess immune responses in the liver, the site of HCV infection and inflammation. In addition, these results are based on patients treated with diverse treatment regimens, some of which are not recommended anymore. Here, we studied paired liver biopsies and blood samples prior to and after viral clearance achieved with sofosbuvir/velpatasvir. This regimen is approved in the US, Europe and other regions for the treatment of genotype 1–6 HCV infection ^{27, 28}. We provide insight into the respective roles of cytokines and bacteria in MAIT cell and monocyte activation and function and asked to what extent immune alterations in chronic HCV infection resolve upon antiviral therapy.

Materials and Methods

Study cohort

This study included 42 patients with chronic HCV infection and compensated liver disease who were treated with a 12-week course of a fixed-dose combination of sofosbuvir/velpatasvir (400 mg/100 mg p.o. once daily, Gilead Inc, Foster City, CA). The patient characteristics are shown in supplementary table 1. All patients achieved a sustained virologic response, defined as undetectable HCV RNA 24 weeks after completion of antiviral therapy. Mononuclear cells were isolated from heparin-anticoagulated blood samples prior to, at weeks 4 and 12 of therapy and at week 24 after the end of therapy. Paired liver biopsies from 37 patients with chronic HCV infection were studied prior to and at week 4 of therapy. Mononuclear cells from NIH blood donors (n=56) and livers that were not used for transplantation (n=10, Research Triangle Labs, Research Triangle Park, NC) were studied as non-viral controls. All were seronegative for HBV, HCV and HIV.

Liver biopsies were graded by an expert hepatopathologist at the NIH using the modified Knodell scoring system to grade inflammation and the ISHAK scoring system to stage fibrosis. The FIB-4 index was calculated based on platelet count, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activity and age ^{29, 30}. All patients gave written informed consent for research testing under protocol NCT02468648 (clinicaltrials.gov) approved by the institutional review board of NIDDK/NIAMS. All authors had access to the study data and reviewed and approved the final manuscript.

Serological analysis

Serum HCV RNA was quantified using the Cobas TaqMan real-time PCR (Roche Molecular Diagnostics, Branchburg, NJ) with a lower limit of detection of 10 IU/ml and a lower limit of quantification of 25 IU/ml.

Plasma analyses

IL-18 was quantified in human plasma samples and in supernatants from E. coli-stimulated PBMC using the Human IL-18 Platinum ELISA (eBioscience, San Diego, CA).

Isolation of lymphocytes from liver and blood

Liver biopsies from chronic HCV patients prior to and at week 4 of antiviral therapy were placed in RPMI, mechanically homogenized and washed with PBS.

Non-parenchymal cells from livers not used for transplantation (Triangle Research Labs, Research Triangle Park, NC) were studied for comparison. Cells were shipped overnight in Dulbecco’s Modified Eagle’s Medium (DMEM) with 4.5 g/L glucose. The next day debris and residual parenchymal cells were removed by centrifugation on a 37% Percoll gradient and intrahepatic lymphocytes were isolated by density gradient centrifugation using Ficoll-Histopaque (Mediatech, Manassas, VA).

PBMC were separated from heparin-anticoagulated blood on Ficoll-Histopaque density gradients and washed three times with phosphate-buffered saline (PBS, Mediatech). PBMC were either studied immediately together with cells isolated from paired liver biopsies or cryopreserved in 70% fetal bovine serum (FBS, Serum Source International, Charlotte, NC), 20% RPMI1640 (Mediatech) and 10% DMSO (Sigma Aldrich, St. Louis, MO).

Flow cytometry

(i) Staining of paired blood and liver samples

Mononuclear cells were stained with ethidium monoazide (EMA, Sigma-Aldrich), anti-CD161-BV421, anti-CD16-V500, anti-CD8-BV605, anti-CD19-PECy5, anti-CD107a-PE (BD Biosciences), anti-TCRVα7.2-FITC, anti-CD3-AlexaFluor700 and anti-CD14-PE-Cy7 (Biolegend) on the day they were isolated from blood samples and liver biopsies. Anti-CD69-APC-Cy7 (Biolegend), anti-HLA-DR-BV711 and anti-CD107a-PE (BD Biosciences) were included to assess activation (CD69, HLA-DR) and degranulation (CD107a).

MAIT cells were identified as CD3⁺CD161⁺TCRVα7.2⁺ cells and studied in 35 paired blood samples and 32 paired liver biopsies prior to and at week 4 of antiviral therapy. Monocytes were identified as SSC^highHLA-DR⁺CD14⁺ cells and studied in 34 paired blood samples and 29 paired liver biopsies prior to and at week 4 of therapy. Liver biopsies with less than 41 events in the MAIT cell gate were excluded from analysis for activation and degranulation markers. CD14⁺⁺CD16⁻, CD14⁺⁺CD16⁺ and CD14⁺CD16⁺⁺ monocyte subsets were identified as described ³¹ and analyzed for activation marker expression if their event counts exceeded the 25^th percentile of the respective subset in all liver biopsies (n=33, 33, 23 and 23 events, respectively).

(ii) Stimulation of MAIT cells with IL-12 and IL-18

Thawed PBMC were incubated with or without IL-12 (50 ng/ml; Miltenyi Biotec) and IL-18 (50 ng/ml; R&D Systems) in RPMI1640 with 10% FBS (Serum Source International), 1% penicillin/streptomycin, 2 mM L-glutamine and 10 mM HEPES (Mediatech) for 13h, followed by addition of brefeldin A (BD Biosciences) for 5h.

(iii) Stimulation of MAIT cells with E. coli

E. coli (K12 strain, Bioline, Taunton, MA) was cultivated in Luria Bertani medium (LB) overnight, fixed in 1% formaldehyde and washed three times in PBS. The concentration of bacteria was determined based on spectrophotometer readings at OD₆₀₀ (OD₆₀₀ of 1.0=8 × 10⁸ cells/ml). Thawed PBMC were incubated with or without formaldehyde-fixed E. coli at a MOI of 5 (i) for 2h in the presence of IL-15 (50 ng/ml; R&D Systems) followed by a 4h-incubation with brefeldin A (BD Biosciences) or (ii) for 13h in the presence of IL-15 (50 ng/ml; R&D Systems) followed by a 5h-incubation with brefeldin A. In some experiments, culture supernatants were harvested prior to addition of brefeldin A and frozen at −80°C for later quantitation of IL-18.

Thereafter, cytokine- or E. coli-stimulated PBMC were washed and stained with ethidium monoazide (EMA, Sigma-Aldrich), anti-TCRVα7.2-FITC, anti-CD3-AlexaFluor700, anti-CD4-APC-Cy7, anti-CD8-BV510 (Biolegend), anti-CD16-PE-Cy5, anti-CD19-PE-Cy5 (BD Bioscience), anti-CD161-APC (Miltenyi Biotec) and anti-CD14-PE-Cy5 (AbD Serotec, Raleigh, NC). After fixation and permeabilization with BD Cytofix/Cytoperm^™, cells were stained with anti-TNFα-BV421, anti-IFNγ-PE (BD Biosciences) and anti-IL-17A (Biolegend).

All samples were immediately acquired on a BD LSR-II flow cytometer using FACSDiva Version 6.1.3 (BD Biosciences, San Jose, CA) and FlowJo Version 9.8.5 software (Tree Star, Ashland, OR). Gates were placed based on fluorescence-minus-one (FMO) controls.

Statistical analysis

D’Agostino & Pearson omnibus normality tests were performed. Dependent on the distribution of the statistical dataset, t-tests, Wilcoxon-signed-rank tests or Mann-Whitney tests were applied. Two-sided p-values <.05 were considered significant. All analyses were performed with GraphPad Prism version 7.0a (GraphPad Software, La Jolla, CA).

Results

MAIT cells are depleted in HCV-induced liver inflammation

To study MAIT cells in blood and liver, we took advantage of paired blood and liver samples of 37 patients with chronic HCV infection prior to and after HCV clearance with sofosbuvir and velpatasvir. MAIT cells were identified by multicolor flow cytometry as CD3⁺CD161⁺TCRVα7.2⁺ cells (Fig. 1A). Consistent with previous reports ^{15, 20, 22}, the frequency of MAIT cells was lower in the blood of patients with chronic HCV infection than of uninfected controls (1.31%, interquartile range [IQR] 0.61%–1.92% vs 2.32%, IQR 0.96%–3.5%; P=.0048, Fig. 1B left panel).

Figure 1. MAIT cells are depleted in HCV-induced liver inflammation.

(A) Representative flow cytometry plots for identification of MAIT cells (time gate and singlet gate are not shown). EMA, ethidium monoazide; SSC-A, side scatter area.

(B) Frequency of MAIT cells in the blood of 37 HCV patients and 57 uninfected controls (left panel). Frequency of MAIT cells in the liver of 37 HCV patients and 10 uninfected controls (right panel).

(C) Linear regression analysis (non-parametric Spearman correlation) of intrahepatic MAIT cell frequency and liver fibrosis (ISHAK) in chronic HCV infection.

(D) Linear regression analysis (non-parametric Spearman correlation) of intrahepatic MAIT cell frequency and liver inflammation (histologic activity index) in chronic HCV infection.

(E) Serum HCV RNA levels (n=37) and liver biopsy HAI score (n=28) of HCV-infected patients prior to and at week 4 of antiviral therapy. Liver biopsies of 11 patients were of insufficient size for staging and grading.

(F) Frequency of MAIT cells in the liver of HCV-infected patients prior to and at week 4 of antiviral therapy compared to uninfected controls (n = 10). Frequency of MAIT cells in the blood of HCV-infected patients prior to, at week 4 of antiviral therapy and at week 24 after the end of antiviral therapy (sustained virological response, SVR) compared to uninfected controls (n = 57).

Wk, week. Median and IQR are shown.

MAIT cells were enriched in the liver as compared to the blood in chronic HCV infection (P<.0001) but their intrahepatic frequency was significantly lower than that of uninfected controls (4.34%, IQR 2.77%–8.59% vs 13.40%, IQR 8.36%–18.88%; P=.001, Fig. 1B right panel). The frequency of MAIT cells in the HCV-infected liver correlated inversely with liver fibrosis (ISHAK score; r=−0.5829, P=.0002, Fig. 1C) and liver inflammation (histologic activity index; r=−0.5437, P=.0006, Fig. 1D). To investigate the effect of antiviral therapy on intrahepatic MAIT cells, we studied paired liver biopsies prior to and at week 4 of therapy. Antiviral therapy was associated with a significant decline of HCV RNA levels by 6.4 ± 0.6 log₁₀ IU/ml and HCV RNA was undetectable in the blood of all patients by week 4 of therapy (Fig. 1E left panel). In parallel, liver inflammation decreased significantly (P=.0001, Fig. 1E right panel), but liver fibrosis did not (data not shown).

The reduction in liver inflammation was associated with a significant increase in the frequency of MAIT cells in the liver (4.34%, IQR 2.77%–8.59% prior to treatment compared to 6.99%, IQR 2.86%–13.93% at week 4 of treatment; P=.0012) but it did not reach the level of uninfected controls (P=.0283, Fig. 1F left panel). The frequency of MAIT cells in the blood remained unchanged up to week 24 after the end of antiviral therapy (Fig. 1F, right panel). The inverse correlation between MAIT cell frequency and liver inflammation and the change in MAIT cell frequency during antiviral therapy were not due to activation-induced downregulation of CD161 (Suppl. Fig. 1A–C).

Collectively, these findings demonstrate that intrahepatic MAIT cells are depleted in HCV-induced liver inflammation. Their frequency increases concomitant to a treatment-induced reduction of liver inflammation by week 4 of antiviral therapy, whereas the frequency of peripheral blood MAIT cells does not change for at least 6 months after the end of therapy.

Activation and cytotoxic effector function of intrahepatic MAIT cells decrease within four weeks of antiviral therapy

To study the effect of antiviral therapy on activation and cytotoxic effector function of peripheral and intrahepatic MAIT cells we assessed the ex vivo expression of the short-term and long-term activation markers CD69 and HLA-DR, and the degranulation marker CD107a by flow cytometry. The frequency of CD69⁺ and HLA-DR⁺ MAIT cells and the activation level per cell were higher in the liver than in the blood of patients with chronic HCV infection (Fig. 2A–D, left column). The frequency of CD69⁺ and HLA-DR⁺ MAIT cells and the activation level per cell decreased significantly in blood (Fig. 2A–D, middle column) and liver (Fig. 2A–D, right column) within four weeks of antiviral therapy. Similar results were obtained for CD161⁻TCR-Va7.2⁺ cells (Suppl. Fig. 1D–F). In addition, intrahepatic MAIT cells displayed higher levels of the degranulation marker CD107a on the cell surface than peripheral MAIT cells in chronic HCV infection. CD107a expression decreased on MAIT cells in blood and liver within four weeks of antiviral therapy (Fig. 2E).

Figure 2. MAIT cell activation and cytotoxicity decrease in blood and liver within four weeks of antiviral therapy for HCV infection.

(A–B) Frequency of CD69⁺ MAIT cells (A) and CD69 expression (MFI, mean fluorescence intensity) of MAIT cells (B) in blood and liver of HCV-infected patients (left panel). Effect of antiviral therapy on the frequency of CD69⁺ MAIT cells in blood (middle panel) and liver (right panel). Mean and SD (A, right panel) or median and IQR (all other panels) are shown.

(C–D) Frequency of HLA-DR⁺ MAIT cells (C) and HLA-DR expression (MFI, mean fluorescence intensity) of MAIT cells (D) in blood and liver of HCV-infected patients (left panel). Effect of antiviral therapy on the frequency of HLA-DR⁺ MAIT cells in blood (middle panel) and liver (right panel). Mean and SD (right panels) or median and IQR (left and middle panels) are shown.

(E) Frequency of degranulated (CD107a⁺) MAIT cells in blood and liver of HCV-infected patients (left graph). Effect of antiviral therapy on CD107a MFI of MAIT cells in blood (middle graph) and liver (right panel). Median and IQR (left and middle panels) are shown.

Outliers in panels A, B and D do not affect the statistical significance.

Wk, week.

Collectively, these findings indicate that the increase of MAIT cells in the liver is associated with a decrease in their activation status and cytotoxic effector function by week four of antiviral therapy.

The frequency of pro-inflammatory monocytes correlates with MAIT cell activation

MAIT cells can be activated in a TCR-independent manner by cytokines and in a TCR-dependent manner by riboflavin-synthesizing bacteria ¹⁷. Monocytes play a critical role in MAIT cell activation because of their ability to release cytokines in response to HCV ^{32, 33} and their ability to present vitamin B metabolites from bacteria ¹⁰. Thus, we studied the activation of monocytes in blood and liver prior to and at week four of antiviral therapy. We distinguished three subsets of monocytes based on their expression of CD14 and CD16 (Fig. 3A).

Figure 3. The frequency of intermediate/pro-inflammatory monocytes correlates with the activation of MAIT cells.

(A) Representative flow cytometry plots for identification of monocytes and their subsets: a, CD14⁺⁺CD16⁻ ‘classical’ monocytes; b, CD14⁺⁺CD16⁺ ‘intermediate/pro-inflammatory’ monocytes; c, CD14⁺CD16⁺⁺ ‘non-classical’ monocytes. EMA, ethidium monoazide; SSC-A, side scatter area.

(B) HLA-DR expression on blood and liver monocytes of HCV-infected patients. Median and IQR are shown.

(C) Linear regression analysis (non-parametric Spearman correlation) of the frequency of CD14⁺⁺CD16⁺ intermediate/pro-inflammatory monocytes and MAIT cell activation in the blood of HCV-infected patients.

(D) Plasma IL-18 levels of HCV-infected patients prior to and at week 4 of antiviral therapy compared to uninfected controls. Mean and SD are shown.

(E–F) HLA-DR expression of total monocytes and monocyte subsets in blood (E) and liver (F) of HCV-infected patients prior to and at week 4 of antiviral therapy. Median and IQR (E) and Mean and SD (F) are shown.

Wk, week.

Monocytes were more activated in the liver than in the blood in chronic HCV infection (P<.0001, Fig. 3B), and the frequency of CD14⁺⁺CD16⁺ ‘intermediate’ monocytes correlated with the percentage of activated MAIT cells (R=.4832, P=.0028, Fig. 3C). Since we had previously described IL-18 as one of the main cytokines monocytes release in a co-culture with hepatoma cells that are transfected with subgenomic HCV RNA replicons ³³, we quantified plasma levels of IL-18 in HCV patients prior to and at week four of antiviral therapy. Antiviral therapy resulted in a significant decrease in plasma levels of IL-18 (P=.0002) but they did not reach levels of uninfected controls yet (P=.001, Fig. 3D). Consistent with the reduction in plasma IL-18 levels, the expression of the activation marker HLA-DR decreased significantly on total monocytes and their classical (CD14⁺⁺CD16⁻) and intermediate (CD14⁺⁺CD16⁻) subsets in blood (Fig. 3E) and liver (Fig. 3F) by week four of antiviral therapy.

Combined, the rapid decrease in the activation of MAIT cells, classical monocytes and intermediate monocytes along with a decrease in plasma levels of IL-18 during antiviral therapy indicate the contribution of virus-induced cytokines to MAIT cell activation in chronic HCV infection.

Changes in MAIT cell and monocyte phenotype during antiviral therapy are greater in the liver than in the blood

Although statistically significant in both compartments, the decrease in MAIT cell activation (Fig. 4A, B) and cytotoxicity (Fig. 4C) was more pronounced in the liver than in the blood. Similar results were obtained for the activation of total monocytes (Fig. 4D) and their classical (Fig. 4E) and intermediate (Fig. 4F) subsets. These findings show that MAIT cell and monocyte activation primarily take place in the liver at the site of HCV infection and inflammation.

Figure 4. Changes in MAIT cell and monocyte phenotype during antiviral therapy are greater in the liver than in the blood.

(A–C) Change in the frequency of HLA-DR⁺ MAIT cells (A) and the HLA-DR MFI (B) and CD107a MFI (C) of MAIT cells by week 4 of antiviral therapy in blood and liver.

(D–F) Change in HLA-DR MFI of total monocytes (E) and their CD14⁺⁺CD16⁻ ‘classical’ (E) and CD14⁺⁺CD16⁺ ‘intermediate/pro-inflammatory’ subsets (F). Median and IQR are shown.

CD8⁺ rather than CD4⁺ MAIT cells respond to in vitro stimulation

To assess the function of MAIT cells, we stimulated PBMC with either recombinant IL-12/IL-18 or with formaldehyde-fixed E. coli. Whereas stimulation with recombinant IL-12/IL-18 bypasses the requirement for monocytes in MAIT cell activation, stimulation with E. coli requires monocytes to take up, process and present vitamin B metabolites ¹². IL-15 was added to enhance the in vitro response of MAIT cells to E. coli, but is not sufficient to induce a response on its own ³⁴. For functional analysis we distinguished between CD8⁺ and CD4⁺ MAIT cells by flow cytometry (Fig. 5A). A majority of peripheral MAIT cells was CD8⁺, whereas minor fractions were either CD4⁺ or double negative (Fig. 5B). The percentage of CD8⁺ MAIT cells was significantly lower in patients with chronic HCV infection than in uninfected controls (75.1%, IQR 68.1%–82.4% vs 82.1, IQR 75.2%–87.6%; P=.0168, Fig. 5B) whereas the percentage of CD4⁺ MAIT cells was higher (9.6%, IQR 4.3%–16.3% vs 4.6%, IQR 2.2%–6.8%; P=.0072, Fig. 5B).

Figure 5. Stimulation with IL-12/IL-18 or E. coli activates CD8⁺ rather than CD4⁺ MAIT cells and results in a loss of CD8⁺ MAIT cells.

(A) Representative flow cytometry plot for identification of CD8⁺ and CD4⁺ MAIT cells.

(B) Frequency of CD8⁺, CD4⁺, double positive (DP) and double negative (DN) MAIT cells in the peripheral blood of 20 chronic HCV patients and 23 uninfected controls. Median and IQR are shown. * P ≤ 0.05 and ** P < 0.01.

(C–D) Frequency of IFN-γ⁺ cells and IFN-γ MFI within the CD8⁺ and CD4⁺ MAIT cell populations after 18h-stimulation of PBMC from chronic HCV patients with either IL-12/IL-18 (C) or E. coli/IL-15 (D). Mean and SD (left panel in C, D) or median and IQR (right panel in C) are shown.

(E) Change in the frequency of CD8⁺ and CD4⁺ MAIT cells after IL-12/IL-18 (left panel) or E. coli/IL-15 (right panel) stimulation of PBMC from chronic HCV patients.

(F) Comparison of the ex vivo frequency of CD8⁺ (left panel) and CD4⁺ MAIT cells (right panel) of 20 HCV patients prior to (Pre) and at week 24 after the end of antiviral therapy (Wk 24 post). Median and IQR are shown.

Wk, week.

As shown in the left panels of figure 5C and 5D the IFN-γ response of MAIT cells to both stimulation conditions was almost exclusively confined to the CD8⁺ MAIT cell subset (33.02% ± 16.05% of CD8⁺ MAIT cells responding to an 18h IL-12/IL-18 stimulation; 52.00% ± 12.51% of CD8⁺ MAIT cells responding to an 18h E. coli stimulation), whereas CD4⁺ MAIT cells responded poorly (4.98% ± 4.40% and 6.51% ± 4.01%, respectively). Similar results were obtained for the amount of IFN-γ produced per cell (Fig. 5C, D right panels). In vitro stimulation was associated with a selective decrease in the percentage of CD8⁺ MAIT cells (P<.0001) and a relative increase in the fraction of CD4⁺ MAIT cells (P<.0001, Fig. 5E). Conversely, antiviral therapy resulted in an increase of the ex vivo CD8⁺ MAIT cell frequency (P<.0001) and in a decrease of the ex vivo CD4⁺ MAIT cell frequency (P<.0001, Fig. 5F).

Together, these findings demonstrate both in vivo and in vitro that MAIT cell activation results in a selective loss of CD8⁺ MAIT cells and a relative increase in CD4⁺ MAIT cells. CD8⁺ rather than CD4⁺ MAIT cells have the functional capacity to influence immune responses and inflammation in chronic HCV infection.

TCR-dependent and -independent MAIT cell stimulation induce different cytokine profiles

Next, we characterized the cytokine profile of MAIT cells from patients with chronic HCV infection. Stimulation of PBMC in a strictly TCR-independent manner with recombinant IL-12 and IL-18 induced predominantly IFN-γ-producing CD8⁺ MAIT cells, a small fraction of TNF-α-producing MAIT cells and almost no IL-17-producing MAIT cells (Fig. 6A). In contrast, a 6h TCR-dependent stimulation with E. coli resulted in a higher frequency of TNF-α- than IFN-γ-producing CD8⁺ MAIT cells (Fig. 6A left and middle panel). Again, the frequency of IL-17-producing CD8⁺ MAIT cells was negligible (Fig. 6A right panel). Stimulation of PBMC with E. coli for 18h resulted in the strongest response with equal frequencies of IFN-γ and TNF-α-producing CD8⁺ MAIT cells (52.64% ± 14.85% and 47.50% ±12.69% respectively) and about 6% IL-17-producing CD8⁺ MAIT cells (Fig. 6A). As shown in figure 6B, the 18h, but not the 6h-stimulation of PBMC with E. coli induced IL-18. This suggests that monocytes exposed to E. coli for 18h deliver both antigen-specific TCR-dependent and cytokine-mediated TCR-independent signals to MAIT cells. Consistent with this notion, neutralization of IL-18 during the 18h-stimulation period with E. coli resulted in a significant decrease in the frequency of IFN-γ and TNF-α producing MAIT cells (Fig. 6C). Interestingly, monocytes, activated in vivo in HCV infection, displayed a reduced IL-18 response to in vitro stimulation with E. coli compared to monocytes of uninfected controls (Fig. 6D).

Figure 6. Stimulation with IL-12/IL-18 or E. coli induces different cytokine profiles of MAIT cells.

(A) Frequency of IFN-γ⁺ (left graph), TNF-α⁺ (middle graph) or IL-17⁺ (right graph) CD8⁺ MAIT cells after stimulation of PBMC from chronic HCV patients with either IL-12/IL-18 (18h), E. coli/IL-15 (6h) or E. coli/IL-15 (18h).

(B) IL-18 concentration in supernatants of PBMC from chronic HCV patients after stimulation with E. coli/IL-15 for 6h or 18h.

(C) Frequency of IFN-γ⁺ (left graph) and IFN-γ⁺ TNF-α⁺ MAIT cells (right graph) and IFN-γ MFI of MAIT cells (middle graph) after 18h stimulation of PBMC with E. coli/IL-15 in the presence or absence of IL-18 neutralizing antibodies.

(D) IL-18 concentration in supernatants after 18h stimulation of PBMC from chronic HCV patients and uninfected controls with E. coli/IL-15.

W/o, without. Median and IQR are shown.

These results show that monocyte-derived IL-18 plays a central role in the activation and effector function of MAIT cells in chronic HCV infection. Other factors, such as viral load, HCV genotype or previous treatment with PEG-IFN-α did not affect the MAIT cell response (data not shown).

IFN-γ production of MAIT cells in response to antigen but not cytokine stimulation is impaired in chronic HCV infection and does not change with antiviral therapy

To explore whether MAIT cell function is altered by chronic HCV infection and whether it recovers during antiviral therapy, we studied a representative subgroup of patients prior to and at the end of treatment (week 12). All patients were end-of-treatment responders and achieved a sustained virologic response at week 24 post therapy. PBMC were stimulated either with E. coli for 6h or with recombinant IL-12/IL-18 for 18h to reflect either strictly TCR-dependent or strictly TCR-independent stimulation conditions. When compared to MAIT cells of uninfected controls (clinical characteristics shown in Suppl. Table 2), CD8⁺ MAIT cells of HCV patients were impaired in their IFN-γ (P=.026 for %IFN-γ⁺ CD8⁺ MAIT cells and P=.037 for MFI IFN-γ) and TNF-α (P=.058 for %TNF-α⁺ CD8⁺ MAIT cells and P=.0406 for MFI TNF-α) response to E. coli stimulation but not to IL-12/IL-18 stimulation (Fig. 7A–C). The IL-17 response was not altered in chronic HCV infection (data not shown). Of note, the antigen-dependent function of MAIT cells did not change with successful antiviral therapy (Fig. 7A–C). Interestingly, after strictly TCR-dependent E. coli stimulation, the frequency of IFN-γ and TNF-α positive MAIT cells correlated inversely with the FIB-4 index (r=−0.5115, P=.0212 for IFN-γ and r=−0.4827, P=.0311 for TNF-α, Fig. 7D), which is used to non-invasively predict liver fibrosis in patients with chronic hepatitis C ^{29, 30}. The number of patients with cirrhosis (ISHAK score 5–6) was too small to compare MAIT cell function in patients with and without cirrhosis.

Figure 7. IFN-γ production of MAIT cells in response to antigen but not cytokine stimulation is impaired in chronic HCV infection and does not change with antiviral therapy.

(A) Frequency of IFN-γ⁺ and TNF-α⁺ CD8⁺ MAIT cells (left graph) of PBMC from HCV patients prior to (Pre) and at the end (Wk 12) of antiviral therapy and PBMC from uninfected controls after stimulation with IL-12/IL-18 (18h).

(B) Frequency of IFN-γ⁺ and TNF-α⁺ CD8⁺ MAIT cells (left graph) of PBMC from HCV patients prior to (Pre) and at the end (Wk 12) of antiviral therapy and PBMC from uninfected controls after stimulation with E. coli/IL-15 (6h).

(C) IFN-γ and TNF-α MFI of MAIT cells in PBMC from HCV patients and PBMC from uninfected controls after stimulation with E. coli/IL-15 (6h).

(D) Linear regression analysis (non-parametric Spearman correlation) of FIB-4 index and the percentage of IFN-γ⁺ and TNF-α⁺ MAIT cells in response to stimulation with E. coli/IL-15 (6h).

Collectively, these findings demonstrate an impaired MAIT cell function in response to TCR- dependent E. coli stimulation in chronic HCV infection.

Discussion

In this study we performed a comprehensive analysis of MAIT cells in both blood and liver in chronic HCV infection. Whereas all previous studies focused exclusively on the blood compartment, the current study provides insight into MAIT cells in the liver, the site of HCV infection and inflammation.

MAIT cells were depleted in HCV-induced liver inflammation and their intrahepatic frequency correlated inversely with liver inflammation and fibrosis. We suggest that MAIT cells are recruited from the blood to the liver in chronic HCV infection where they are activated by inflammatory cytokines and undergo activation-induced cell death. Activation-induced death of MAIT cells was shown by Cosgrove et al. after in vitro stimulation of MAIT cells with E. coli ²¹. Moreover, a pro-apoptotic phenotype of MAIT cells with increased expression of caspase 3 and 7 has been reported ³⁵. This is consistent with our observation of a selective loss of CD8⁺ MAIT cells upon in vitro stimulation with either IL-12/IL-18 or E. coli/IL-15. It is also consistent with the increase in the frequency of MAIT cells in the liver along with the resolution of liver inflammation by week four of antiviral therapy. In contrast, and consistent with previous reports ^{15, 20, 22}, antiviral therapy did not improve the reduced MAIT cell frequency in the peripheral blood even during a longer follow up to week 24 post therapy. Since the liver is equipped with a high density of immune cells, an increase in MAIT cell frequency in the blood may lag significantly behind any changes the liver.

MAIT cells were more activated in the liver than in the blood in chronic HCV infection. During antiviral therapy, the decrease in the expression of the long-term activation marker HLA-DR exceeded the decrease in the expression of the short-term activation marker CD69. Indeed, about 60% of intrahepatic MAIT cells remained CD69⁺ at week 4 of antiviral therapy, a time point at which HCV RNA was not detectable anymore in the blood. A similar percentage of CD69⁺ MAIT cells has been observed in “healthy” liver tissue ¹⁶ and CD69 has recently been recognized as a marker for tissue resident cells ³⁶. Thus, we propose that the intrahepatic MAIT cell population comprises a large subset of CD69⁺ tissue-resident cells.

The decrease in MAIT cell activation parallels the decrease in monocyte activation during antiviral therapy. Among the three monocyte subsets, ‘intermediate’ CD14⁺⁺CD16⁺ monocytes expressed the highest levels of the activation marker HLA-DR and their frequency correlated with the percentage of activated MAIT cells. CD14⁺⁺CD16⁺ monocytes are recognized for their ability to produce high levels of pro-inflammatory cytokines, such as TNF-α, IL-12 and inflammasome-derived cytokines, such as IL-1β and IL-18 ³⁷. Since we had previously shown inflammasome activation and IL-18 production of monocytes in response to HCV-replicating hepatoma cells ³³, we quantified IL-18 as a representative cytokine. IL-18 is expressed at high levels in liver ³⁸ and blood ¹⁹ in HCV infection and has been shown to play a role in MAIT cell activation and IFN-γ production ²². The activation of MAIT cells and intermediate monocytes decreased within four weeks of antiviral therapy along with plasma levels of IL-18. Combined, our data suggest that MAIT cell activation is primarily driven by virus-induced cytokines in patients with compensated chronic HCV-related liver disease. This is further supported by the fact that we did not detect any bacterial DNA in the plasma of these patients with a TLR-9 reporter cell line (HEK-Blue^™ TLR9 cells, InvivoGen) (data not shown).

We found an impaired MAIT cell function in response to TCR-dependent E. coli stimulation but not in response to TCR-independent IL-12/IL-18 stimulation in the peripheral blood of patients with chronic HCV infection. This is consistent with Hengst et al. who reported an impaired MR1-dependent MAIT cell function ²⁰. In contrast to Hengst et al. we stimulated for 6h to reflect strictly TCR-dependent MAIT cell function (longer stimulation results in the induction of IL-12 and IL-18) and to avoid a selective loss of functional CD8⁺ MAIT cells. The high variability in the MAIT cell response upon TCR stimulation required the analysis of a large cohort of uninfected controls. This may be related to comorbidities such as obesity and diabetes that have been reported to affect MAIT cell frequency and function ³⁹.

It has recently been shown that MR1-mediated MAIT cell activation is tightly regulated and dependent on the integration of innate signals by antigen presenting cells ⁴⁰. These innate signals may be altered in HCV infected patients as compared to uninfected controls, and may result in the differential MAIT cell response to E. coli and IL-12/IL-18 stimulation. The fact that MAIT cell function was not impaired in response to IL-12/IL-18 indicates that MAIT cells may contribute to antiviral immune responses by production of IFN-γ in chronic HCV infection.

A limitation of this study is that functional assays were not performed with intrahepatic MAIT cells due to the small size of the percutaneous liver biopsies and the limited number of isolated intrahepatic lymphocytes. Whether the effector profile of MAIT cells vary in peripheral blood and liver remains to be elucidated.

In conclusion, intrahepatic MAIT cells are depleted depending on the severity of liver inflammation in chronic HCV infection. They are activated by monocyte-derived cytokines and can produce the antiviral cytokine IFN-γ Furthermore, they exert cytotoxicity as indicated by increased expression of the degranulation marker CD107a. In contrast, MAIT cell function in response to E. coli stimulation is impaired in patients with chronic HCV infection which may impact MAIT-cell dependent protection against bacterial infections in patients with advanced liver disease.

Abbreviations

DAA

direct acting antivirals

HCV

hepatitis C virus

IFN-γ

interferon gamma

IQR

interquartile range

MAIT

mucosal associated invariant T cells

PBMC

peripheral blood mononuclear cells

SD

standard deviation

TCR

T cell receptor

TNF-α

tumor necrosis factor alpha