If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Institute of Molecular Medicine and Experimental Immunology, University Hospital Bonn, Germanycurrent address: Peter Medawar Building for Pathogen Research, Nuffield Department of Medicine, University of Oxford, UKthese authors contribute equally
Institute of Molecular Immunology and Experimental Oncology, Technical University of Munich, Germanythese authors contribute equallyGerman Center for Infection Research, Munich site
Antigen-specific T-cells during chronic liver injury feature hallmarks of T-cell exhaustion.
•
IFN-I/IL-10 signaling axis is a determinant of impaired systemic T-cell responses in chronic liver injury and cirrhosis.
•
IFN-I-induced by translocated gut microbiota hampers systemic T-cell immunity via IL-10 release by myeloid cells.
•
IL-10 suppresses T-cell immunity by directly acting on antigen-specific T-cells.
•
IL-10R blockade promotes the reconstitution of T-cell functions in virus infected mice and vaccinated patients.
Abstract
Background & Aims
Patients with chronic liver disease (CLD) such as cirrhosis are at increased risk of intractable viral infections and are hyporesponsive to vaccination. Hallmarks of CLD and cirrhosis include microbial translocation and elevated levels of type I Interferon (IFN-I). We aimed to investigate the relevance of microbiota-induced IFN-I for the impaired adaptive immune responses in CLD.
Methods
We combined BDL and CCl4 models of liver injury with vaccination or LCMV-infection in transgenic mice lacking IFN-I in myeloid cells (LysM-Cre IFNARflox/flox), IFNAR-induced IL-10 (MX1-Cre IL10flox/flox) or IL-10R in T-cells (CD4-DN IL-10R). Key pathways were blocked in vivo with specific antibodies (anti-IFNAR and anti-IL10R). We assessed T-cell responses and antibody titers after HBV and COVID vaccinations in CLD patients and healthy individuals in a proof-of-concept clinical study.
Results
We demonstrate that BDL- and CCL4-induced prolonged liver injury leads to impaired T-cell responses to vaccination in mice and cirrhosis patients as well as defective T-cell responses to viral infection in mice leading to persistent infection. Innate sensing of translocated gut microbiome induced IFN-I-signalling in hepatic myeloid cells that triggered excessive IL-10 production upon viral infection. IL-10R signalling in antigen-specific T-cells rendered them dysfunctional. Antibiotic treatment, IFNAR and IL-10Rα inhibition restored anti-viral immunity without detectable immune pathology in mice. Notably, IL-10Rα blockade restored the functional phenotype of T-cells from vaccinated cirrhosis patients.
Conclusion
Loss of systemic T-cell immunity during prolonged liver injury is driven by IFN-/IL-10 expression induced by innate sensing of translocated microbiota.
Impact and implications
Chronic liver injury and cirrhosis are associated with enhanced susceptibility to viral infections and vaccine hypo-responsiveness. Using different pre-clinical animal models and patient samples, we identified that impaired T-cell immunity in BDL- and CCL4-induced prolonged liver injury is driven by sequential events involving microbial translocation, IFN signalling leading to myeloid cell-induced IL-10 expression, and IL-10-signaling in antigen-specific T-cells. Given the absence of immuno-pathology after interference with IL-10R, our study highlights a novel potential target to reconstitute T-cell immunity in CLD patients that can be explored in future clinical studies.
acquisition of data (CPH, JS, LMA, HH, SVS, RS, SK, TB, JS, BMK, ZA, BS, NK); study concept and design (CPH, CK, JT, NG, BMK, AR, SVS, ZA, PAK); analysis and interpretation of data (CPH, JS, SVS, ZA, KS, NG, AR, EL, JT, PAK); drafting the manuscript (ZA, PAK, CPH, AR, SVS).
Funding
ZA, PAK, SVS were supported by the German Research Foundation (DFG) projects: Excellence Cluster EXC 2151 (ID: 390873048). ZA, PAK were supported by SFBTR57 ((2)-P32 and -P11) (ID: 36842431). ZA, was supported by the GRK 2168 (ID: 272482170), SFB1454 (P10) (ID: 432325352), SFBTR237 (B15) (ID: 369799452), and by the German Federal Ministry of Education and Research (BMBF) for the COVIMMUNE project. CPH was supported by the DFG (project number 403193363) and the Wellcome Trust (WT109965MA). PAK was supported by the European Union’s Horizon 2020 research and innovation programm TherVacB, by the SFB-TRR179 and the German Center for Infection Research, Munich site. The DFG funded equipment used for flow cytometry (project numbers 21637201, 471514137, 387333827, and 216372545).
Conflict of interest statement
The authors declare that they have no conflict of interest
Chronic liver diseases (CLD) such as liver fibrosis and cirrhosis are associated with barrier dysfunction and enhanced microbial translocation into the liver leading to tonic and chronic type I interferon (IFN-I) signalling, release of pro-inflammatory cytokines and chronic immune cell activation
. As the disease progresses, systemic inflammation and structural distortion of liver tissue are believed to determine progressive loss of immune surveillance known as cirrhosis associated immune dysfunction (CAID)
and poor response to vaccination against influenza, hepatitis A and hepatitis B constitute major factors contributing to patient´s morbidity and mortality
. We have previously shown that IFN-I acting on hepatic myeloid cells in the context of bacterial infection during liver damage is responsible for impaired innate immunity in mouse and man
, rendering the immune system unable to contain infection.
Here, we report that tonic IFN-I signalling during prolonged liver injury (hereafter referred to as pLI) and cirrhosis determines the loss of systemic CD8 and CD4 T-cell-mediated immunity through induction of the immune-regulatory cytokine IL-10. Using complementary experimental approaches in preclinical models of BDL- and CCL4-induced pLI and cirrhosis patients, we identified IL-10 as the key mediator of T-cell dysfunction during chronic liver injury and provide evidence that blocking of IL-10 rescues T-cell function and allows for control of viral infection without organ immune-pathology in mice. Given the importance of CD8 and CD4 T-cell responses for immune surveillance our findings contribute to the understanding of failing immune responses in cirrhotic patients and identify the IL-10Rα signalling pathway as a potential molecular target to improve immune control and vaccination efficacy in these patients.
Materials and Methods
Data availability
All data from this study are provided in the source data. RNA-seq expression data are available at NCBI GEO under the accession number: GSE158261. Primary data from flow cytometry or immunohistochemistry are available upon reasonable request.
Patients and blood samples
Blood and serum samples as well as questionnaire-based assessment of donor characteristics and disease stage from cirrhosis patients and healthy volunteers before vaccination with TWINRIX (HAV and HBs Antigen of the HBV) were collected at the University Hospital Bonn and general practices in Bonn (healthy individuals; n = 16 and cirrhotic patients n = 16). All enrolled subjects were never exposed to HBV or been vaccinated against HBV or HAV, proven by the absence of anti-HBs antibodies by serological testing before vaccination and patient information. Informed consent was obtained from all patients and healthy individuals enrolled in the trial in accordance with the Declaration of Helsinki protocol. The study was approved by and performed according to the guidelines of the local ethics committees of the University of Bonn (310/16). All enrolled subjects received 4 doses of the vaccine on d1, d7, d21 and after 6 months. Sample collection was performed 4 to 5 weeks after the last injection. PBMCs were isolated by Ficoll-Hypaque (PromoCell, Germany) density gradient centrifugation and stored at −80 °C until further use. Serum was separated by centrifugation for 10 min and supernatant was stored at −80 °C. Detailed characteristics of healthy donor and patients with CLD and the study design are provided in Tables S1 and S2 and Fig.S.1,e,f.
HBS-specific T-cell activation
43 (80-90% purity) 15-mer synthetic peptides overlapping by ten amino acids overlap (Xaia Custom Peptides, Sweden) covering the sequence of the HBV surface antigen (HBs Antigen) protein according Galibert sequence (Table S3). Peptides were pooled and used for detection of HBs-specific T-cells and their function. After isolation from peripheral blood, untouched CD3 T-cells were labelled with cell trace violet (CTV) (0.5 μM; Thermo Fisher Scientific) and cultured in round-bottom 96-well plate at 5 x 104/well, in the presence of autologous irradiated (4,000 rads) PBMC (105/well), HBs-peptides and recombinant human IL-2 (50 U/ml; PeproTech). After 5 days of incubation, proliferation and cytokine production in HBS-specific T-cells were assessed by flow cytometry.
Anti-RBD (S) SARS-CoV-2 antibodies assay
Serum samples were analyzed for anti-S-RBD IgG titers using the SARS-CoV-2 Plate 7 V-PLEX Serology Kit from MSD (Meso Scale Diagnostics, LLC). Antibody concentrations were quantified via the MESO SECTOR S 600. Raw data was analyzed with the MSD Discovery Workbench tool (V 4.0.13), that quantifies anti-RBD IgG. All assays were performed by trained laboratory technicians according to the manufacturer’s standard procedures.
SARS-CoV-2 neutralization assay
Heat-inactivated sera were serially two-fold diluted starting with 1:5 dilutions. 120 μl of each serum dilution was mixed with 120 μl of OptiPROTM SFM cell culture media (Gibco) containing 80 plaque-forming units (PFU) of SARS-CoV-2 (isolate B.3). After 1 h at 37°C, 200 μl of each mixture was added to wells of a 24 well plate seeded the day before with 1.5 x 10e5 Vero E6 cells/well. After incubation at 37°C for 1 h, the inoculum was removed and cells were overlaid with a 1:1 mixture of 1.5% carboxymethylcellulose (Sigma) in 2x MEM (Biochrom) with 4% FBS (Gibco). After a 3-day incubation at 37°C, cells were fixed with 6% formaldehyde and stained with 1% crystal violet in 20% ethanol.
In vitro activation and proliferation assay of human T-cells
PBMCs were labelled with CTV (0.5 μM; Thermo Fisher Scientific) and cultured in RPMI supplemented with 10% FCS at 1 x 106 cells/well in 96-well plates (Corning) coated with anti-CD3 (clone HIT3a, BioLegend) and anti-CD28 (clone L293, BD Biosciences) (each at 3 μl/ml) and IL-2 (10 IU/ml; PeproTech). T-cell proliferation and intracellular cytokine expression as well mitochondrial assays were assessed by flow cytometry.
In vitro IL-10Rα blockade on human T-cells
PBMCs were cultured in RPMI supplemented with 10% FCS at 1 x 106 cells/well in 96-well plates (Corning) coated with anti-CD3 (clone HIT3a, BioLegend) and anti-CD28 (clone L293, BD Biosciences) (each at 3 μl/ml) and IL-2 (10 IU/ml; PeproTech). 5μg/ml of the blocking anti IL-10Rα antibody (3F9; BioLegend) was added 12 hours later.
ELISA for detection of cytokine expression
Human IFNγ, TNF and IL-2 in the supernatant of in vitro activated cells and murine IL-10 in the serum were determined using ELISA MAX Deluxe Set (BioLegend) according to manufacturer instructions.
Quantification of mitochondrial membrane potential, mitochondrial mass and mitochondrial superoxide production
For mitochondrial studies in human T-cells were determined after anti-CD3 stimulation. Cells were incubated with 50 nM MitoTracker Green (MTG) and/or 25 nM MitoTracker DeepRed (MTDR) for 30 min at 37C beforecell surface staining. Mitochondrial superoxide levels in human T-cells were determined, after cell surface staining, by incubation (15 min at 37 °C) of overnight anti-CD3 stimulated and unstimulated cells in the presence of MitoSOX Red (5μM; Molecular Probes).
Mice
6–9-week-old mice C57BL/6J (B6) were purchased from Janvier (Le Genest-Saint-Isle, France. LysMCre-IFNARfl/fl (previously described18), OT-I, OT-II, CD45.1-B6 (B6.SJL-Ptprca Pepcb/BoyJ), 6C2.36 and P14 mice were originally purchased from The Jackson Laboratory and maintained in the House of Experimental Therapy, University Clinic Bonn. Mx1Cre-Il10fl/fl x Mx1Cre ((Il10tm1Roer) x (C.Cg-Tg(Mx1-Cre)1Cgn/J)) were kindly provided by Axel Roers. IL10RaDN mice43 were kindly provided by Samuel Huber. All mice were maintained under specific pathogen-free conditions specific pathogen-free (SPF) conditions and were handled according to the guidelines of the institutional animal guidelines of the animal facilities of the University of Bonn. Experimental procedures were approved by the Animal Ethics Committee of the state of North Rhine-Westphalia, Germany. For antibiotic treatment, mice were given a combination of vancomycin (1 g/l), ampicillin (1 g/l), kanamycin (1 g/l), and metronidazole (1 g/l) in drinking water72. All antibiotics were obtained from Sigma Aldrich. Colonization of Germ Free (GF) mice with caecal microbiota of SPF mice was perform as previously described73. Briefly GF mice received 3 times a suspension of the cecum content of sham or BDL operated wildtype per oral gavage on 3 consecutive days and were BDL operated one week after the last transfer.
Murine liver injury models
Liver injury was induced in male, 8–9-week-old mice via bile duct ligation (BDL) or treatment with carbon tetrachloride (CCl4) following established protocols75. In brief, to induce BDL, the animals were treated with painkillers and anaesthetized before peritoneal cavity was opened along the linea alba. Two ligatures were placed around the common bile duct in order to obstruct it, the incisions in the peritoneum and the skin were then closed and the mice were allowed to recover. During the first 5 days after operation all animals received additional injections of painkillers and liver injury was allowed to develop for 10 days before experiments were performed. Alternatively, mice received 0.5 μl CCl4/g body weight for 12 weeks intraperitoneally. CCl4 was dissolved 1:7 in olive oil and was administered in 3-day intervals. After the final injection, mice were allowed to recover for 10 days before further experiments were performed. Control mice were sham-operated (no ligation of the bile duct) or received olive oil (i.p.) injections respectively. During all experiments, animals were monitored closely on a daily basis. To inhibit of IFNAR or IL10Rα signaling in-vivo, mice were treated intraperitoneally with 250μg/mouse of blocking antibodies targeting TGFBR-II (clone: 1D11.16.8) IFNAR1 (clone: MAR1-5A3) or IL10Rα (clone: 1B1.3A) from BioXcell intravenously. Control animals received injections containing the HPRN and MOPC-21 antibodies respectively.
LCMV infection
Mice were infected with 2x104 plaque-forming units (PFU) of LCMV strain WE or Armstrong diluted in sterile PBS intravenously.
Irradiation and adoptive cell transfers
Mice were subjected to sublethal irradiation (6 Gy) one day before transfer. T-cells were isolated from the spleens and lymph nodes of donor mice and 4x106 CD3+ T-cells (1:1 of WT and TG cells) per recipients were transferred intravenously. Before any subsequent experiment, recipient mice were allowed to recover for 12 days in order to ensure proper engraftment. For all other infection or vaccination studies, Naïve P14, OT-I and OT-II cells were isolated from the spleens of mice and 2.5x105 cells were transferred one day before vaccination.
Murine in vitro T-cell proliferation assay
48-well plates were coated with antibodies directed against murine CD3 (clone: 500A.2, BD Biosciences) and CD28 (clone: 37.51, Biolegend) at a concentration of 0.5μg/ml and 10μg/ml respectively, incubated at 37°C for 2 h and washed. T-cells isolated from the spleens of Sham-or BDL operated mice were stained with Cell Trace Violet (Invitrogen) and added at a concentration of 2x106 cells/ml. T-cell proliferation was measured after three days by flow cytometry.
Flow cytometry and antibodies
Single cell suspensions were acquired on a FACSCanto II or LSRII Fortessa (DFG ID 216372401) and analyzed with FlowJo (version 10.0.7, Tree star). In order to determine the expression of surface molecules, cells were stained on ice for 20 min. The LIVE/DEAD fixable Near-IR Dead Cell Stain kit (Life Technologies) was used in all staining to detect dead cells, also, an in-house made antibody (clone: 2.4G2) directed against the epitopes shared by Fc-gamma receptors was added. Single cell suspension from spleen or liver were stimulated with 100ng/ml PMA, 200ng/ml in the presence of Brefeldin A and monensin for 3 hours, before collecting and intracellular staining for flow cytometry. To block unspecific staining. Cytokines were stained after fixation with 4% PFA and Permeabilization with 1x Permeabilization buffer (FoxP3/Transcription Factor Staining Buffer Set, eBioScience) and transcription factors with the FoxP3/Transcription Factor Staining Buffer Set according to the manufacturer’s instructions. Antibodies directed against the following targets in mice were purchased from Biolegend, eBioScience or Miltentyi: anti-CD3e (145-2C11), anti-CD4 (GK1.5 or RM4-5), anti-CD8a (53-6.7), anti-CD11b (M1/70), anti-CD11c (N418), anti-CD44 (1M7), anti-CD45.1(A20), anti-CD45.2(104), anti-CD146 (ME-9F1), anti-CD210a (1B1.3a), anti-CTLA-4 (UC10-4B9), anti-Eomes (Dan11mag), anti-F4/80 (BM8), anti-GzmB (GB11), anti-IFNγ (XMG1.2), anti-IL-2 (JES6-5H4), anti-IL-21 (FFA2), anti-LAG3 (C9B7W), anti- PD-1 (29F.1A12), anti-T-bet (4B10), anti-IRF4 (IRF4.3E4), anti-TOX (TXRX10), anti-TCF-1 (S33-966), anti-TCRβ (H57.597), anti-TIM3 (RMT3-23) and anti-TNF (MP6-XT22), anti-phospho-STAT3 (Stat3Y705-B12) and anti-phospho-SMAD2 (Ser250 (SD207-1)). CD8 and CD4 T-cell tetramers specific for the LCMV epitopes gp33-41 and gp66-77 respectively, were purchased from Immudex (gp33-dextramer) or provided by the NIH Tetramer Core Facility (Emory University). The following antibodies were used to detect protein expression on human cells: anti-CD3 (SK3), anti-CD62L (DREG-56), anti-CD8 (PRA-T8), anti-CD45RA (HI100), anti-CD45 (HI30) from ThermoFisher and anti-CD4 (PRA-T4), anti-PD1 (EH12.2H7), anti-CTLA4 (BNI3), anti-CD3 (OKT3), anti-IFNγ (4S.B3), anti-IL-21 (3A3-N2) anti anti-IL-10Rα (3F9) from BioLegend.
Immunofluorescence microscopy
Liver tissue samples were fixed in a 0.05 M phosphate buffer containing, 0.1 M L-lysine, 2 mg/ml NaIO4, and 10 mg/ml paraformaldehyde (PLP) at pH7.4 overnight. Subsequently, samples were washed in phosphate buffer and dehydrated in 30% Sucrose overnight. Finally, samples were place int Tissue TEK (Sakura Finetek), snap-frozen and stored at -80 °C. 20μm sections were acquired on a CM3050S cryostat, rehydrated and stained with antibodies in buffer containing 1% normal mouse serum. Pictures were acquired on an LSM 710 confocal Microscope (Zeiss). A BV421-conjugated antibody directed against F4/80 (BM8) was purchased from Biolegend, for detection of the LCMV nucleoprotein an unconjugated rat-anti-LCMV antibody (VL4) was purchased from BioXCell; expression was detected via an AF647-conjugated goat-anti-rat antibody (Invitrogen).
RNA extraction, cDNA-synthesis and RT-PCR
Small samples of liver tissue were homogenized in 1ml Quiazol and total RNA was isolated using the RNeasy Lipid Tissue Kit (Quiagen, 74804). Afterwards, 2-5μg of RNA was reverse transcribed into cDNA at 37°C for 2 hours using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, 4368814). cDNA was stored at -20°C and real-time PCR was performed using Taqman primers and probes for Il10, Ifnb1, Mx1, Hprt and Gapdh in murine samples and PDCD1, CTLA4, CD244, EOMES, BATF, GAPDH and HPRT in human samples. The relative mRNA expression was calculated with the ΔΔCt-method.
RNA sequencing and data analysis
Transcriptomic differences in isolated T-cells from BDL-treated or sham control mice were determined by QuantSeq 3’mRNA sequencing (Lexogen). FACS (Aria Fusion DFG ID 387333827) sorted cells were lysed in 700μl Trizol and stored until RNA extraction was performed with the RNeasy micro kit (Qiagen). Library production for 3’-mRNA sequencing was performed with up to 160 ng purified RNA according to the manufacturers’ protocol and sequenced on a HiSeq2500 (Illumina) with a sequencing depth of 15 Mio reads per sample (NGS Core Facility, University Hospital, Bonn, Germany). The alignment was performed with STAR (v2.5.3a) against the murine reference genome mm10. Transcripts were quantified with the Partek E/M algorithm and further processed for normalization in R (v3.5.0) with the DEseq2 algorithm (v1.20.0). The data set was further optimized by flooring transcripts with minimal gene counts at least to ≤1 and the exclusion of transcripts with a mean expression ≤10 in every test condition. Differentially expressed genes were identified in the Partek Genomics Suite (v7.18.0402) for T-cells in BDL versus sham control cells by a one-way-ANOVA (fold-change |1.5|, FDR-adjusted p-Value ≤0.05). Data visualization and biological interpretation were performed with the Partek Genomics Suite, ClueGo plugin (v2.5.2) for Cytoscape (v3.7.2) and R packages ggplot2 (v3.2.1), Enhanced Volcano (v1.6) and tidyr (v1.0.2). Heatmaps of two groups were created using means across the two groups, with expression being centered around 0 and visualized with Mayda (v2.14).
16S qPCR for quantification of bacterial DNA
DNA was extracted from samples using MoBio PowerSoil kit (Qiagen). DNA concentration was calculated using a standard curve of known DNA concentrations from E. coli K12. 16S qPCR using primers identifying different regions of the V6 16S gene was performed using Kappa SYBR fast mix. Absolute number of bacteria in the samples was then approximated as DNA amount in a sample/DNA molecule mass of bacteria. Liver tissue of germ-free mice was used as an internal control.
Statistical analysis
To determine statistical differences, a two-tailed unpaired or paired Student’s t-test was used when two groups were compared; a repeated-measurements one-way ANOVA was used when three groups were compared. For non-parametric data, the Mann-Whitney test was used when comparing two groups and the Kruskal-Wallis test when comparing three or more groups, respectively. Analysis was done with Prism 8. Statistical significance was set at p < 0.05.
Results
Failure of T-cell immunity to control viral infection in mice with chronic liver disease.
To explore the impact of chronic liver diseases (CLD) on T-cell mediated antiviral immune responses, we combined two different mouse models of prolonged liver injury (hereafter referred to as pLI) (i.e., bile duct ligation (BDL) and CCl4-treatment) with lymphocytic choriomeningitis virus (LCMV-WE and LCMV-Armstrong strain) infection, that is rapidly cleared by the T-cell response in healthy mice. While healthy mice successfully controlled LCVM replication latest on day 12 (d12) (Fig.1a-c, Fig. 1a-d, data not shown), BDL- and CCl4-treated mice failed to clear LCMV-WE from blood, liver and spleen and virus replication persisted for at least 30 days post infection (p.i.), indicating a systemic loss of antiviral immune surveillance, rather than a local attenuation of anti-viral immunity selectively in the liver. At the peak of the immune response, i.e. d8 post-infection, pLI-mice showed significantly lower numbers of LCMV-specific CD8 T-cells, recognizing the gp33 epitope of LCMV, compared to healthy mice (Fig.1d,e and Figs. S1e and f). Moreover, LCMV-specific CD8 T-cells from pLI-mice had lower frequencies and numbers of IFNγ/TNF-producing T-cells (Fig.1f,g, Fig. 1g), indicating loss of anti-viral immune surveillance during pLI independent of its aetiology. Of note, we observed reduced numbers of LCMV-specific CD4 T-cells in BDL-mice (Fig.1h,i), that had diminished effector cytokine production with only few cells co-expressing IFNγ, TNF and IL-2 (Figs. S1h and i). To study the defect in virus-specific CD8 T-cells in more detail, we transferred TCR-transgenic P14 CD8 T-cells that express a T-cell receptor specific for the LCMV-gp33 epitope, into BDL-mice one day before LCMV infection. BDL-mice showed reduced frequencies and numbers of IFNγ/TNF-producing T-cells compared to sham-mice (Fig. S1j-m), confirming the results from the endogenous T-cells. Together, these experiments demonstrated a broad and severe dysfunction of virus-specific T-cell immunity in mice with pLD that was associated with failure to control LCMV infection.
Figure 1Impaired antiviral T-cell responses in mice with pLI
BDL- or sham-operated mice were infected with LCMV (2×104 PFU) on d9 post-operation. a-c, LCMV-titers (PFU) in blood, liver and spleen. d,e, total and gp33-specific CD8 T-cells from spleen and liver on d8. f,g, IFNγ/TNF-production by LCMV-specific CD8 T-cells. h,i, frequencies of total and LCMV-specific CD4 T-cells on d8. a-c, data representative of ≥3independent experiments; e,g,i, pooled data from ≥3 independent experiments; statistical analysis by unpaired t-test.
Defective T-cell responses to vaccination in patients and mice with liver injury
To assess whether the observations made in the mouse models of BDL- and CCL4-induced prolonged liver injury recapitulate the phenotype of T-cells from patients, we stimulated PBMC-derived T-cells from healthy individuals and cirrhosis patients (Table S1) with anti-CD3/CD28 antibodies to induce antigen-mediated T-cell activation. T-cells from CLD patients produced significantly lower levels of IL-2, IFNγ and TNF (Fig.2a) which were associated with enhanced apoptosis (AnnexinVpositive), lower mitochondrial membrane potential, higher numbers of depolarized mitochondria, and higher levels of reactive-oxygen species that failed to increase upon stimulation (Figs. S2a–d). To explore the mechanisms of CLD-associated T-cell dysfunction, we studied the responses of cirrhotic patients and healthy individuals (Table S1) to hepatitis B virus (HBV) vaccination in a proof-of-concept clinical study (Fig. S2e). In line with previous studies
94% of the healthy vaccinees had Anti-HBs antibody titer >100 IU/L (strong responders) after vaccination while only 38% of cirrhosis patients had a titer above 100U/L and 62% between 10-100IU/L (week responders) (Fig.2b). Furthermore, HBs-specific blood-derived CD8 and CD4 T-cells from cirrhosis patients failed to proliferate and to produce IFNγ and IL-21 after stimulation with HBs-specific peptides (Fig.2c-f). Additionally, we assessed the response to COVID-19 vaccination in naïve cirrhotic patients and healthy individuals (Table S2) who received two doses of COVID-19 mRNA (BNT162b2) vaccine (Fig. S2f). Blood samples were obtained before and 7-10 days after the second vaccination. RBD IgG binding to wildtype as well as variants of concern such as such as B1.351 (beta), B.1.1.7(alpha) and P.1(gamma) were significantly lower in cirrhotic patients compared to healthy individuals and neutralization capacity (Fig.2g, Fig. S2gFig.2h,). Likewise, CD4 and CD8 T-cells from cirrhotic patients produced significantly lower levels of IFNγ upon in vitro stimulation with spike protein peptides (Fig.2i). Together, these results indicate impaired antigen-specific B and T-cell immunity after vaccination in cirrhotic patients.
Figure 2Poor immune responses to vaccination in patients and mice with liver cirrhosis
a, IL-2, IFNγ/TNF-release from anti-CD3/CD28 stimulated T-cells from healthy donors or cirrhotic patients. b, anti-HBs IgG titers in sera of healthy donors (n=16) and cirrhotic patients (n=16) vaccinated against HBsAg. c-f, frequencies of proliferating and IFNγ-producing CD8 T-cells or IL-21-producing CD4 T-cells after ex-vivo stimulation with HBsAg peptides. g, Serum IgG binding to RBD of SARS-CoV-2 at d7-10 after the second COVID-19 mRNA vaccination (BNT162b2). h, Neutralization of wildtype SARS-CoV-2 by the sera of healthy (n=7) or cirrhosis (n=9) vaccinees. i, frequencies of IFNγ-producing CD8 and CD4 T-cells after in vitro stimulation with SARS-CoV-2 spike protein. j-l, Oil or CCl4-treated (12 weeks) mice were vaccinated with HBs antigen/ polyI:C. j, Anti-HBs IgG titer on d7. k, percentages of splenic CD45.1+ HBs-specific (6C2.36) CD8 T-cells. l, frequency of IFNγ/ TNF-producing CD8 T-cells. data representative of three (a) or two (h, j, k, l) independent experiments. a-f, i-l, statistical analysis by unpaired t-test, g, paired Wilcoxon.
To characterize impaired T-cell immunity after vaccination in-vivo, we transferred naïve HBs-specific (6C2.36) or ovalbumin-specific CD45.1+CD8 (OT-I) and CD4 (OT-II) transgenic T-cells in CCl4-treated and BDL-operated mice, respectively one day before vaccination with HBs antigen or ovalbumin with polyI:C as adjuvant. In line with the results from CLD patients, we detected significantly lower titers of anti-HBs antibodies (Fig.2j) as well as lower frequencies of 6C2.36 (Fig.2k,l), and lower frequencies of OT-I and OT-II cells and reduced frequencies of IFNγ, TNF and IL-2 producing specific T-cells (Figs. S2h–k) in the spleen of pLI-mice, which suggest reduced T-cell expansion and effector function. Together, these findings suggest that systemic antigen-specific T-cell responses to vaccination were impaired during chronic liver disease and cirrhosis.
To fine-map the transcriptional profile of antigen-specific CD8 T-cells during pLI, we performed bulk RNA-seq analysis of sorted P14 cells from BDL- or sham-operated mice at d8 post LCMV infection identifying 2153 differentially expressed genes (Fig. S3a). Gene ontology enrichment analysis (GOEA) and network visualization revealed differences in processes associated with lymphocyte activation, immune effector functions, metabolic and signal transduction processes (Fig.3a). Genes encoding inhibitory receptors (Havrc2, Pdcd1, Ctla4, and Lag3) and inflammation-associated cytokines like Tnf and Il10 were up-regulated in T-cells from BDL-mice, whereas transcription factors associated with effector T-cell functions namely, Tbet, Tcf7, Eomes and Bcl6
were down-regulated (Fig.3b,c). P14 T-cells from BDL-mice showed increased expression of transcription factors associated with T-cell exhaustion such as Tox, Batf, Irf4 and Id3
, showed enrichment for these genes in P14 T-cells from pLI-mice (Fig.3d,e). Flow cytometric analysis confirmed expression of these genes at protein level in both transferred P14 (Fig.3f) and endogenous LCMV-specific CD8 T-cells in BDL-mice (Figs. S3b and c). We further detected increased expression of PD1, TIM3 and LAG3 in virus-specific CD8 and CD4 T-cells from liver and spleen of both BDL- and CCl4-treated mice (Figs. S3d–f). To assess whether T-cell dysfunction in pLI coincided with expression of TOX, BATF and IRF4, we determined the expression of IFNγ/TNF and these transcription factors in T-cells on d5 p.i., when precursors of exhausted T-cells are known to emerge
. Although P14 T-cell already lost their effector function on d5 p.i. in BDL-mice, no increased expression of TOX or BATF were detected at this time point besides a marginal IRF4 expression (Fig.3g). These results indicated that T-cell dysfunction during pLI was distinct from TOX-dependent exhausted T-cells.
Figure 3T-cell exhaustion signature in antigen-specific T-cells during pLI
a,b,c, Gene-enrichment Network (a), volcano plot (b) heatmap of centered mean expression (c) of DE genes in P14 T-cells from on d8. d, GSEA for exhausted T-cells-associated genes. e, heatmap for DE genes in P14 T-cells. f,g expression of inhibitory receptors and transcription factors in P14 T-cells. (g) IFNγ/TNF-production, inhibitory receptors and transcription factor expression on d5 or d8. f,g, data representative of two independent experiments.
, which may curtail T-cell mediated immunity. GSEA revealed enrichment of 148 genes associated with TGFβ-signalling in P14 T-cells from BDL-mice (Fig.4a and Fig. S4a). Consistently, we detected increased levels of TGFβ expression in the liver (Fig.4b) and enhanced levels of phospho-SMAD2 (Fig.4c) a key downstream molecular of TGFβ−receptor signalling in LCMV-specific CD8 T-cells from BDL-mice. To test the relevance of TGFβ-receptor signalling for T-cell dysfunction during pLI, we treated BDL-mice with TGFβRII-blocking antibodies during LCMV infection. However, 60% of BDL-mice succumbed after infection when TGFβ-receptor signalling was blocked (Fig.4d), without evidence for increased viral clearance (Fig.4e), suggesting a non-redundant and specific function of TGFβ in tissue-protection but not immune surveillance during liver injury.
Figure 4| microbiota-induced IFN-I drives T-cell dysfunction during pLI.
a, GSEA of TGFβ-associated genes in P14 T-cells d8 post infectoin. b,expression of TGFβ-mRNA in the liver of mice in (a). c,phospho-Smad2 in LCMV-specific CD8 T-cells. d,Kaplan-Meier survival curves for LCMV-infected BDL-mice receiving anti-TGFβR-II or isotype control. e, Liver and spleen LCMV-titers in mice in (d). f,IFNb or Mx1 mRNA expression in the liver. g,h, GSEA (g) and volcano plot (h) of IFNAR-associated genes in P14 T-cells. i-k, Numbers and frequency of IFNγ/TNF-producing P14 T-cells from in antibiotics (Abx)-treated BDL-mice. a,One-way ANOVA, b-f,i-k,data from ≥2 independent experiment. b,e,f,k, statistics were assessed by unpaired t-test, i,k,one-way ANOVA with Dunnett’s multiple comparisons test.
Besides TGFβ, we also detected enhanced expression of IFN-I and the interferon-stimulated gene Mx1 in livers of BDL-mice that further increased after LCMV infection (Fig.4f). GSEA indicated significant enrichment of 230 interferon alpha receptor (IFNAR)-signalling associated genes
in P14 T-cells from BDL-mice (Fig.4g, Fig. S4b). Among the top 20 upregulated genes, we found Irf4, Nr4a2, Mt2, Egr2, Lclat1 and Frmd4a (Fig.4h), that have been associated with T-cell dysfunction in cancer and chronic viral infections
, we found that translocation of gut microbiota in mice with BDL- and CCL4-induced pLI- and CLD patients (Figs. S4c and d) induced tonic IFNAR-signalling in hepatic myeloid cells (Figs. S4e and f). Interestingly, colonization of BDL-operated germ-free mice with the microbiome of sham- or BDL-operated mice (Fig. S4g) led to comparable levels of IFN-I (Figs. S4h and i) suggesting that chronic IFNAR-signalling during pLI was triggered by sensing translocated gut microbiota independent of its composition. More importantly, reduction of intestinal microbial burden by antibiotic treatment led to reduction of IFN-I expression in liver and intestine of BDL-operated mice, improved expansion and effector cytokine production by LCMV-specific CD8 T-cells and consequently enhanced viral clearance (Fig. S4j and Fig.4i-k). These results demonstrated a critical role of microbial translocation in IFN-I production and T-cell dysfunction during chronic liver injury.
High IFN-I expression during pLI drives loss of T-cell immunity
Next, we investigated the relevance of IFN-I-signalling for loss of T-cell immunity during liver injury. Antibody mediated blockade of IFNAR in LCMV-infected pLI-mice led to significant reduction of IFNAR-induced genes (Fig. S5a). Strikingly, inhibition of IFNAR-signalling increased the numbers of total (Fig. S5b) and LCMV-specific CD8 and CD4 T-cells in BDL- as well as CCl4-treated mice (Fig.5a,c, Figs. S5b–h). Blockade of IFNAR-signalling led to enhanced expression of IFNγ/TNF in T-cells and pronounced reduction of the viral load in pLI-mice (Fig.5a-e, Figs. S5d–h). Further we observed a reduction in PD1, TIM3 and LAG3 and TOX, IRF4 and BATF expression (Fig.5f-i) in LCMV-specific CD8 T-cells. Of note, IFNAR-blockade induced higher numbers of TCF1+P14 T-cells and TCF1+TIM3- progenitor-exhausted cells (Fig.5g,i). Thus, tonic IFNAR-signalling in pLI-mice determined T-cell dysfunction and virus persistence in a similar manner as in chronic LCMV infection with clone 13 and HIV infection
pLI was by BDL operation or CCl4 treatment before LCMV infection. Mice were treated with IFNAR1-blocking or isotype antibodies every second day starting at d5 p.i.. a-c, numbers of virus-specific and cytokine producing LCMV-specific CD8 T-cells on d8 (a), d15 (b). d, liver immune-fluorescence on d15. e, quantification of LCMV-titers in mice from (e). f, expression of inhibitory receptors on LCMV-specific CD8+ T-cells. g-i, co-expression of the indicated marker in LCMV-specific CD8 T-cells d15. c,e,i, statistics assessed by unpaired t-test.
To explore the relevance of T-cell-specific IFNAR-signalling for their dysfunction during pLD, we generated CD4-CreERT2x IFNARfl/fl mice, in which tamoxifen injection deleted IFNAR selectively in ∼60% of T-cells (Fig. S6a). Tamoxifen application from day 3 after LCMV infection, neither restored T-cell function (Figs. S6b and c) nor improved viral clearance (Fig. S6d). Although this does not rule out a direct effect of IFN-I signalling on LCMV-specific CD8 T-cells, it rather suggests a role of IFNAR-signalling on other immune cell populations that then cause T-cell dysfunction. We therefore ablated IFNAR-signalling in myeloid cells (i.e., macrophages, monocytes and neutrophils), using LysM-Cre x IFNARfl/fl mice. Strikingly, BDL-operated LysM-Cre x IFNARfl/fl mice showed a better control of LCVM infection compared to IFNARfl/fl littermates (Crenegative), had higher numbers of total and LCMV-specific CD8 T-cells that produced more IFNγ/TNF (Fig.6a-g) and expressed lower levels of PD1, LAG3 and TIM3 (Figs. S6e and f). This suggests that IFNAR-signalling in myeloid immune cells was a key event for T-cell dysfunction during pLI. Furthermore, we detected elevated IL-10 levels in the blood of pLI-mice after LCMV infection that were reduced by anti-IFNAR or antibiotics treatment (Fig.6h). Likewise, we observed higher levels of phosphorylated-STAT3, the canonical downstream signalling molecule of IL-10-Receptor, upon anti-IFNAR treatment (Fig.6i). GSEA showed enrichment of IL10R-stimulated genes in P14 T-cells from BDL-mice (Fig.6j), indicating a role for IFNAR-induced IL-10 in T-cells dysfunction during pLI. To prove this hypothesis, we generated MX1-Cre x IL-10fl/fl mice, in which Il10 gene is deleted upon expression of the IFNAR-stimulated gene MX1
. LCMV-specific CD8 and CD4 T-cell numbers were higher in BDL-operated MX1-Cre x IL-10fl/fl mice (Fig.6k,l and Figs. S6g and h), with higher levels of IFNγ/TNF production (Fig.6m, Figs. S6i and j) and a reduced viral load (Fig.6n,o) compared to littermate control mice. Overall, these data strongly indicate that IFNAR-induced IL-10 expression in myeloid cells contributed to T-cell dysfunction and failure to control viral infection during BDL- and CCL4-induced pLI.
Figure 6Abrogation of IFNAR/IL-10 axis restores T-cell immunity in pLI.
a-g, IFNARflox/flox and IFNARflox/floxxLysm-Cre mice were infected with LCMV after BDL-operation and analysed on d8. a, liver immune-fluorescence for LCMV nucleoprotein and macrophages, b, Liver LCMV-titers. c,d, quantification of virus-specific CD8 T-cells and e-g cytokine-producing T-cells. h, IL-10 levels in blood of BDL-mice treated with anti-IFNAR antibodies or antibiotic (Abx). i, phospho-STAT3 in LCMV-specific CD8 T-cells at d8 p.i. and after IFNAR-blockade. j, GSEA of IL-10R-associated genes in LCMV-specific P14 T-cells. k-o, IL10f/f Mx1-Cre+ or IL10f/f (Cre-) littermates were infected with LCMV after BDL-operation. k-m, frequencies and numbers of hepatic CD8+ T-cells, LCMV-specific CD8 T-cells and IFNγ/TNF-producing LCMV-specific CD8 T-cells. m, liver immune-fluorescence of LCMV nucleoprotein and macrophages. o, Liver LCMV-titers. Data obtained from three (a-g), two (k-o) independent experiments. b,d,f,g,j,n,o, statistics assessed by unpaired t-test, h, one-way ANOVA with Dunnett’s multiple comparisons test.
IL-10 acts directly on antigen-specific T-cells during liver injury
To identify cell types on which IL-10 exerts its inhibitory effect, we determined IL-10Rα expression in different immune cell populations. While antigen-experienced (CD44+) CD8, CD4 and LCMV-specific T-cells in BDL mice expressed higher IL-10Rα levels (Fig.7a,b, Fig. S7a), no changes in IL-10Rα expression were observed in myeloid immune cells or CD44- T-cells (Figs. S7b–d). Importantly, we also detected elevated surface expression levels of IL-10Rα on CD45RAneg T-cells in cirrhotic patients compared to healthy individuals (Fig.7c) suggesting a higher sensitivity to IL-10 signalling in activated T-cells during pLI. To assess the relevance of IFNAR-signalling for elevated IL-10Rα expression on activated T-cells, we treated murine T-cells with IFN-I for 48hrs before T-cell receptor stimulation in vitro or induced IFN-I production in-vivo by polyI:C administration for 1-3 days. Pre-exposure of T-cells to IFN-I upregulated IL-10Rα expression in a time and dose-dependent manner (Fig.7d,e).
a,b, expression level and percentage of IL-10Rα-expressing T-cells from BDL- or sham-operated mice at d8. c, IL10Rα expression on peripheral blood CD3+CD45RAneg T-cells of healthy donors or cirrhotic patients. d, IL10Rα expression of murine T-cells stimulated ex-vivo with αCD3/CD28 in presence of increasing IFNβ concentrations. e, IL10Rα expression after ex-vivo stimulation with αCD3/CD28 in murine T-cells isolated d3 after pI:C application. f-k, sub-lethally irradiated B6 mice (CD45.1) recieved 1:1 mixture of naive wildtype (WT,CD45.2+) and transgenic (TG,CD45.1/2+) T-cells. 4 weeks later mice underwent sham or BDL operation and infected with LCMV. f-h abundance of transgenic and wildtype CD8 T-cells (f), LCMV-specific CD8 T-cells or LCMV-specific CD4 T-cells (g) IFNγ/TNF production in LCMV-specific T-cells (h). i, Liver and spleen LCMV-titer in BDL-mice recieved either WT or TG T-cells. j, expression of inhibitory receptors in LCMV-specific CD8 T-cells. data from three independent experiments (a-j) and (c) with individual donors. b,i, statistics assessed by unpaired t-tests, f-j two-tailed paired Student’s t-test.
To investigate whether IL-10 was acting directly on T-cells, we co-transferred equal numbers of wild type (WT, CD45.2+) or CD4–DN IL-10Rα transgenic (TG, CD45.1/2+) T-cells, which overexpressed a dominant-negative (DN) IL-10Rα and therefore had impaired IL-10 signalling
, into irradiated wild type mice. After induction of liver injury, chimeric mice (Fig. S7e) were infected with LCMV and T-cells were analyzed on d8 p.i. Strikingly, total numbers of activated TG CD8 T-cells in the liver were significantly higher compared to WT-cells (Fig.7f). More importantly, CD4 and CD8 T-cells with impaired IL-10Rα signalling were more prevalent among LCMV-specific T-cells and had superior effector cytokine production compared to WT-cells (Fig.7g,h). Consistently, mice which received IL-10Rα-impaired T-cells showed significantly lower viral load compared to their counterparts that received WT T-cells (Fig.7i). Thus, T-cells with low responsiveness to IL-10 showed an improved proliferative capacity and effector cytokine production during pLI. Of note, transgenic LCMV-specific T-cells had lower expression levels of PD-1, LAG-3 or TIM3 but no differences in expression of IRF4, TOX or BATF (Fig.7j, Fig. S7f). Collectively, these results suggested that elevated levels of IL-10 during liver injury were responsible for T-cell dysfunction and impaired viral clearance by acting directly on T-cells.
Therapeutic targeting and inhibition of IL-10 signalling restores T-cell responses during liver injury.
Next, we investigated whether therapeutic interference with IL-10R downstream signalling restores antiviral T-cell immune surveillance during liver injury. Efficient IL-10Rα-blockade, as indicated by reduced pSTAT3 expression (Fig. S8a), in LCMV-infected BDL- and CCL4-treated mice led to increased numbers of LCMV-specific CD8 T-cells with higher IFNγ/ TNF production and reduced PD-1, LAG-3 and TIM3 expression (Fig.8a-d, Figs. S8b–e). Consistently, IL-10Rα blockade reconstituted clearance of viral infection in BDL- and CCl4-treated mice similar to healthy mice by d15 p.i. (Fig.8e-g). Moreover, IL10Rα-blockade enhanced the frequencies of TCF1+TIM3neg T-cells (Fig. S8f) and reduced TOX, IRF4 and BATF expression (Fig.8h,i). Likewise, we observed significant increase in the frequency of HBs-specific CD8 T-cells as well IFNγ/ TNF CD8 T-cells in HBs-vaccinated pLI-mice upon treatment with anti-IL-10R antibodies (Figs. S8g and h). Next, we wondered whether interference with IL-10Rα-signalling would also restore dysfunctional T-cell from patients with CLD. Antibody-mediated blockade of the IL-10Rα in αCD3/CD28 stimulated T-cells from cirrhosis patients reduced PD-1 and CTLA-4 expression (Fig.8j and Fig. S8i). More importantly, IL-10Rα-blockade enhanced the proliferation as well as IL-21 and IFNγ production (Fig.8k,l, Figs. S8j and k) by T-cells from vaccinated cirrhotic patients upon stimulation with HBs peptides. Together, these data revealed that therapeutic blockade of IL-10Rα-signalling restored T-cell immune surveillance during BDL- and CCL4-induced pLI and in patients with cirrhosis.
Figure 8IL-10Rα blockade restores T-cell immunity during pLI
BDL- or CCl4-treated mice received anti-IL-10Rα antibody every second day during LCMV infection. a-d frequencies and numbers of LCMV-specific (a,c), IFNγ/TNF−producing (b,c) expression of inhibitory receptors by LCMV-specific CD8 T-cells (d). e-g, liver immune-fluorescence (e,f), Liver LCMV-titers (e-g). h,i, expression of indicated marker in LCMV-specific CD8 T-cells. j, co-expression of PD1 and CTLA4 in CD8 T-cells from PBMC of healthy controls or cirrhotic patients after stimulation with anti-CD3/CD28 in presence of anti-IL10Rα or isotype control. k,l, T-cells from HBsAg-vaccinated healthy controls (n=4) and cirrhotic patients (n=4) stimulation with HBsAg peptides in the presence of anti-IL10Rα or isotype control. k, frequencies of proliferating l, cytokine producing CD4 and CD8 T-cells. data from three independent experiments (a-g) and one (j-l) experiment with individual donors. c,g,i, Statistics assessed by unpaired t-tests. j-l two-tailed paired Student’s t-test
ns not significant, *P≤0.05, **P≤0.01, ***P≤0.001, ****P≤0.0001.
Cirrhosis-associated immune dysfunction (CAID) is linked to poor responses to vaccination and occurrence of infections that can cause loss-of-function of remaining liver tissue and thereby trigger life-threatening liver failure [
], against which no specific therapeutic intervention exists. While intractable bacterial infections pose the most prominent threat to cirrhotic patients [
]. Here, we identified the IFN-I/IL-10 signaling axis as a determinant of CAID, through which liver damage is linked to suppression of systemic T-cell immunity to acute LCMV infection in preclinical models of prolonged liver injury (pLI) and to vaccination in pLD-mice and cirrhosis patients.
The outcome of immunity to viral infection is determined by virus-intrinsic properties and by host factors that both influence anti-viral immunity. Virus-intrinsic properties and their contribution to development of persistent viral infection have been extensively studied, such as lymphocytic choriomeningitis virus clone 13 infection overcoming anti-viral immunity by inducing T-cell exhaustion
. In COVID-19 patients with cardiovascular, lung or metabolic diseases, overshooting immune responses causing organ pathology are more frequently observed
. During CLD and cirrhosis however patients suffer from rather an attenuated and dysfunctional immune response largely driven by loss of tissue architecture and systemic chronic inflammation
. Our results demonstrated that enhanced microbial-translocation after liver injury caused increased tonic IFNAR-signalling in liver myeloid cells and that abrogation of IFNAR-signalling in these cells prevented loss of systemic T-cell immunity. It is well established that intestinal microbiota play a key role in the pathology and progression of chronic liver diseases
. Interestingly, reconstitution of germfree mice with gut microbiota from either healthy mice or mice with liver injury led to similar levels of IFN-I and similar loss of systemic T-cell immunity. This suggests that IFNAR-induced loss of systemic T-cell immunity during pLI is rather driven by the innate sensing of translocated gut bacteria, and raised the question how IFN-signalling in myeloid cells might regulate T-cell function. None of the current mouse models faithfully reflect all etiological and pathophysiological characteristics of chronic liver disease and cirrhosis in humans. The BDL and CCl4 models used in this study mainly induce prolonged liver injury, which replicate several relevant aspects and complications observed in patients with cirrhosis, namely microbial translocation, chronic IFNAR signalling and immune dysfunction. It therefore remains unclear whether acute liver injury or changes in the liver architecture occurring in cirrhosis patients further contributes to the mechanisms described here, such as enhanced gut bacterial translocation, or whether cirrhosis-associated ultrastructural changes themselves may influence hepatic or systemic immunity.
Genome-wide transcriptional and protein profiling of LCMV-specific T-cells from pLI-mice revealed gene signatures characteristic of exhausted T-cells found in cancer or chronic infection. However, loss of T-cell effector function in pLI-mice preceded the upregulation of exhaustion-associated transcription factors; TOX, BATF and IRF4, corroborating our assumption that loss of T-cell immunity during liver injury is not a cell-intrinsic process but is driven in a paracrine fashion. Of note, we detected gene signatures for TGFβ-signalling, IFN-signalling and IL-10-signalling among others. It was previously shown that blockade of TGFβ-signalling does not lead to control of persistent viral infection under healthy conditions
. Our experiments confirmed these findings in the context of pLI and revealed a non-redundant role of TGFβ in organ-protection during pLI as its blockade led to severe liver immune pathology. Importantly, after LCMV-infection of mice with liver injury we detected increased IL-10 expression in monocytes/macrophages that depended on IFN-I-signalling, which is unexpected because IFN-I is considered to be involved in self-perpetuating inflammation
. In contrast to the immune pathology-promoting effect of blocking TGFβ, inhibition of IL-10 by antibodies or T-cell-specific ablation of IL-10R signalling rescued effector function of LCMV-specific T-cells in pLI-mice and promoted viral clearance in the absence of notable immune pathology. Likewise, elimination of intestinal microbiota and reducing gut microbial-translocation to the liver and blockade of IFN-I-signalling in myeloid cells all prevented the excessive production of IL-10 and rescued antiviral T-cell function. Translating these results from preclinical models of liver injury to patients with CLD, our data suggest that inhibition of IL-10R signalling rescued vaccination-induced antigen-specific T-cells from their dysfunction.
Although our data provides in vitro, in vivo and human evidences on this particular mechanism, the study may still have some limitations. First, none of the current mouse models faithfully reflect all etiological and pathophysiological characteristics of chronic liver disease and cirrhosis in humans. The BDL and CCl4 models used in this study mainly induce prolonged liver injury, which indeed recapitulate several relevant aspects and complications observed in patients with cirrhosis, namely microbial translocation, chronic IFNAR signalling and immune dysfunction. It remains unclear whether acute liver injury or changes in the liver architecture occurring in cirrhosis patients further contributes to the mechanisms described here, such as enhanced gut bacterial translocation, or whether cirrhosis-associated ultrastructural changes themselves may influence hepatic or systemic immunity. Second, IL-10 signalling is a key regulator of immune effector functions and tissue integrity
, and its blockade may therefore carry the risk to induce immunopathology or autoimmunity. Therefore, future studies aiming at overcoming the limitations of IL-10 signalling in cirrhosis-associated immune dysfunction will need to fine-tune the duration, timing and levels before implementing IL-10R blockade to increase vaccination efficacy in CLD patients.
Taken together, our work provides a mechanistic understanding of the loss of systemic T-cell immunity during chronic liver injury, discriminates organ-protective TGFβ-signalling from T-cell suppressing IL-10 signalling and identifies IFN-I and IL-10 as molecular targets for immune interventions that reconstitute T-cell immunity.
Acknowledgments
We thank Alisa Ismaili, Maike Kreutzenbeck, Janett Wieseler, Sven Kröcker and Rebecca Balduf for technical assistance. We thank Dr. Michael Brinkmann and Lisa Brinkmann for the excellent support of the vaccination studies. We would like to acknowledge the support by the Flow Cytometry and Next Generation Sequencing Core Facilities at the Medical Faculty at the University of Bonn. The DFG funded equipment used for flow cytometry (project numbers 21637201, 471514137, 387333827, and 216372545).
Appendix A. Supplementary data
The following is/are the supplementary data to this article:
00158-7&id=ga1.jpg){kind=link}
00158-7&id=gr1.jpg){kind=link}
00158-7&id=gr2.jpg){kind=link}
00158-7&id=gr3.jpg){kind=link}
00158-7&id=gr4.jpg){kind=link}
00158-7&id=gr5.jpg){kind=link}
00158-7&id=gr6.jpg){kind=link}
00158-7&id=gr7.jpg){kind=link}
00158-7&id=gr8.jpg){kind=link}