Chronic hepatitis delta virus infection leads to functional impairment and severe loss of MAIT cells

Patient material and characteristics

The subjects included in this study were seen at the outpatient clinic of the Department of Gastroenterology, Hepatology and Endocrinology at Hannover Medical School in Germany. For selected experiments, peripheral blood was collected from healthy individuals recruited at the Blood Transfusion Clinic at the Karolinska University Hospital Huddinge. Written informed consent was obtained from all subjects for the investigation of immunological parameters as part of the protocols approved by the Ethics Committee of the Hannover Medical School in Germany (5258, 5292) and by the Regional Ethics Review Board in Stockholm. Forty-one patients with chronic HDV infection were included in this study. Furthermore, 38 patients with chronic HBV mono-infection and 57 healthy controls were included for direct comparison with the HDV-infected patients. Patient characteristics are presented in Table 1. Peripheral blood mononuclear cells (PBMCs) were isolated by standard density-gradient separation and cryopreserved until further use. Liver biopsies for immunohistochemistry (IHC) and immunofluorescence (IF) analysis were acquired from three HDV patients and, as controls, from seven viral hepatitis-negative patients undergoing liver biopsy because of suspicion of liver disease but where the liver histology was normal (Supplementary Table S1).

Flow cytometry

PBMCs were thawed and washed once in RPMI-1640 medium supplemented with 25 mM HEPES, 2 mM L-glutamine (all from Thermo Fisher Scientific), 10% FBS (Sigma-Aldrich), 50 μg/mL gentamicin (Life Technologies), and 100 μg/mL normocin (InvivoGen). PBMCs were then used for phenotypic stainings or functional assays. Surface and intracellular stainings for cytokines and transcription factors were performed as previously described.³² The antibodies used are listed in Supplementary Table S2 as well as in Supplementary CTAT Table. Samples were acquired on an LSRFortessa flow cytometer (BD Biosciences) equipped with 355, 405, 488, 561, and 639 nm lasers. Single-stained polystyrene beads (BD Biosciences) and the compensation platform in the FlowJo software v. 9.9 or v. 10 (TreeStar) were used for data compensation. Unless otherwise indicated, MAIT cells were defined as CD161^hiVα7.2⁺ cells being CD8⁺CD4⁻ or CD8⁻CD4⁻. High dimensional analysis of flow cytometry data was performed through Barnes-Hut stochastic neighbor embedding (SNE) analysis using R (version 3.2.1, The R Foundation for Statistical Computing) with an in-house built script as previously described.¹²

In vitro MAIT cell functional and viability assays

MAIT cells in the PBMC samples were stimulated for 24 h with Escherichia coli strain D21 (ratio E. coli CFU: PBMC of 3:1) or with a combination of IL-12p70 (10 ng/mL, Peprotech) and IL-18 (100 ng/mL, Medical & Biological Laboratories). E. coli was mildly fixed in formaldehyde prior to addition to PBMCs as previously described.³³ To assess degranulation, anti-CD107a mAb (BD Biosciences) was added at the beginning of the assay. In the E. coli-mediated stimulation assays, anti-MR1 mAb (clone 26.5, Biolegend) or IgG2a isotype control (clone MOPC-173, Biolegend) was added at the beginning of the assay at the final concentration of 20 μg/mL. Monensin (Golgi Stop, BD Biosciences) and brefeldin A (Golgi Plug, BD Biosciences) were added for the last 6 h of the experiment. In selected experiments, viability of MAIT cells in PBMC samples was assessed by annexin V (AnnV) and dead cell marker (DCM) staining ex vivo or in vitro following either five-day incubation of PBMCs with IL-12 p70 (10 ng/mL) and IL-18 (100 ng/mL) or 24 h incubation with supernatants (pure or at different dilutions) from HDV-infected or uninfected HepG2^hNTCP cells.

Immunohistochemistry

For immunohistochemistry analysis, frozen OCT-embedded liver biopsies from HDV-infected patients and controls were sectioned into 5-μm tissue sections, placed on SuperFrost Ultra Plus slides (Histolab) and stored at –80°C until staining. The sections were fixed with 4% paraformaldehyde (Sigma Aldrich) for 20 min on ice, followed by two blocking steps: one for elimination of endogenous peroxidase activity, using Bloxall (Vector Laboratories), and another one for reduction of general background staining, using Innovex background buster (Innovex Biosciences). Blocking was performed at room temperature (RT), for 10 and 20 min respectively. Thereafter, sections were incubated with purified rabbit anti-human CD3 epsilon (EP449E, Epitomics) or mouse anti-human TCR Vα7.2 (3C10, Biolegend), with isotype-matched antibodies serving as the corresponding negative controls. Subsequently, sections were incubated with ImmPRESS mouse or rabbit reagent (Vector Laboratories) for 30 min at RT, followed by signal detection using ImmPACT DAB as the peroxidase substrate solution (Vector Laboratories). Tissue sections were counterstained with Hematoxylin (Histolab) and mounted with Vectamount (Vector Laboratories). Immunoreactivity was visualized and analyzed by light microscopy (Leica DM 4000B).

Multicolor immunofluorescence

Five mm thick cut frozen liver sections were thawed and fixed in ice cold 4% paraformaldehyde for 20 min on ice followed by one wash with Tris Buffer Saline (TBS). Sections were incubated with Image-iT® FX signal enhancer for 30 min at RT and washed once in TBS, then blocked with Background Buster (Innovex Biosciences) for 15 min at RT. Sections were then incubated with the primary antibodies targeting TCR Vα7.2 (clone 3C10, Biolegend) and IL-18Rα (clone AF840, R&D Systems), or IgG1 (Biolegend) and Goat (R&D Systems) isotype control antibodies for 1 h at RT. The primary antibodies were detected with the corresponding donkey anti-mouse and donkey anti-goat secondary antibodies conjugated to Alexa 555 or 647 (all Invitrogen). DAPI was included at 0.0005% w/v with secondary antibody. Following staining background autofluorescence was quenched using TrueBlack® Lipofuscin Autofluorescence Quencher (Biotium) for 5 min followed by three brief washes in TBS, then washed once for 10 min in TBS. The slides were mounted using Prolong Diamond (Invitrogen). Sections were visualized with a Nikon Ti-E spinning-disk confocal microscope. Images were acquired at RT using a 40× Nikon air objective, and Andor DU-897 EM-CCD camera (512×512 pixels. Pixel size 16 μm). Images were generated using Cytosketch (CytoCode, Auckland, New Zealand).

Plasma protein measurements

Plasma levels of sCD14 and sCD163 were measured by ELISA (R&D Systems, catalog number DC140 and DC1630, respectively). Samples were analyzed in triplicates according to the manufacturer’s instructions. The OD was measured at 450 nm with a wavelength correction at 570 nm. Plasma levels of IL-12 and IL-18 were determined using the LUMINEX-based multiplex bead assay (BioPlex Pro Human Cytokine Panel; Bio-Rad), according to the manufacturer’s instructions and optimized protocols.³⁴ The samples were analyzed in one run and acquired on the LUMINEX instrument, using the BioPlex Manager 6.0 software.

In vitro HDV-infection assay

The production of HDV viral particles and HDV infection of the Huh7.5^hNTCP cell line were performed as previously described.^35,36 To obtain supernatants from HDV-infected HepG2^hNTCP cells, cells were seeded in 24-well plates (1–4×10⁴ cells/well) in complete DMEM containing 10 μg/mL blasticidin. The following day, cells were infected with the virus in the presence of polyethyleneglycol (PEG). After 6 h, the medium was exchanged for fresh complete DMEM supplemented with 3% FCS. After five days the supernatant was collected, filtered, and stored at −80 °C until further use.

Statistical analyses

Statistical analyses were performed using Prism software v.6 (GraphPad). The data sets were first evaluated for normality of the data distribution. Statistically significant differences between samples were determined as appropriate using the Kruskal-Wallis test followed by Dunn’s multiple comparisons test, the ordinary one-way ANOVA followed by Holm-Sidak’s multiple comparisons test or the two-way ANOVA followed by Sidak’s multiple comparisons test for multiple unpaired samples, the Mann-Whitney’s test or unpaired t test for pairwise comparison of unpaired samples, and the Wilcoxon’s signed rank test or paired t test for pairwise comparison of paired samples. Differences between categorical variables were assessed through the Fisher’s exact test. Correlations were calculated using the Spearman’s test. Two-sided p values < 0.05 were considered significant.

MAIT cell levels are severely reduced in peripheral blood during chronic HDV infection

MAIT cells can be defined in healthy adult humans as T cells carrying the TCR Vα7.2 segment in combination with high expression levels of CD161.³⁷ We initially determined the frequency of CD161^hiVα7.2⁺ MAIT cells in the peripheral blood of patients with chronic HDV infection, chronic HBV mono-infection, and healthy controls (Fig. 1A). Levels of total MAIT cells in HDV-infected patients were significantly lower than in HBV mono-infected patients, whereas HBV mono-infected patients exhibited only a mild decrease in MAIT cells compared with healthy controls (Fig. 1B). In contrast, and as a reference, no difference was detected in the levels of CD161⁻Vα7.2⁺ (non-MAIT) T cells between the groups (Fig. 1B). In HDV-infected patients, loss of MAIT cells occurred preferentially in the CD8⁺ compartment, as the levels of CD8⁺CD4⁻MAIT cells out of total MAIT cells (Fig. 1C) and out of T cells (Supplementary Fig. S1A) were significantly lower than in healthy controls and HBV mono-infected patients.

With respect to clinical characteristics, HDV-infected patients had significantly higher levels of the enzymes aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyl transpeptidase (gGT), and significantly lower levels of prothrombin compared to HBV mono-infected patients (Table 1), indicative of a more extensive liver damage. The level of hepatic fibrosis was higher in the HDV-infected patients as compared with HBV mono-infected patients (Table 1). However, only when considering the HBV- and HDV-infected patients together, there was a significant correlation between the loss of MAIT cells and the levels of fibrosis (Supplementary Fig. 1B). This indicates that the decrease in circulating MAIT cell levels was not directly associated to the level of fibrosis in the HDV-infected patients. With respect to virological characteristics, the vast majority of HBV mono-infected patients and 57% of HDV-infected patients were HBeAg⁻ (Fig. 1D and Table 1). No significant differences in MAIT cell levels were detected between HBeAg⁻ and HBeAg⁺ HDV-infected patients (Fig. 1D).

Patients with HBV and HDV received different types of antiviral treatments, and thus had disparate treatment histories. This might possibly have affected the results. However, previous or on-going treatment did not have a major effect on the levels of MAIT cells (Fig. 1E, F). Since only five or three HBV mono-infected patients had received nucleos(t)ide analogue (NUC) treatment or pegylated (peg)-IFNα, respectively, at the time of sampling, only a minor bias, if any, might have been introduced by this treatment (Fig. 1E). HDV-infected patients who had received NUC treatment presented with a minor increase in levels of MAIT cells (Fig. 1F, left), and no significant differences were detected in the levels of MAIT cells between either patients that had, or had not, been treated with peg-IFNα at the time of sampling (Fig. 1F, middle), or HDV-infected patients before and after twelve weeks of treatment (Fig. 1F, right).

Taken together, these data indicate that peripheral blood MAIT cells are profoundly depleted in chronic HDV infection.

Intrahepatic MAIT cells in chronic HDV infection

We next examined whether the loss of MAIT cells in peripheral blood was a consequence of their homing to the liver. The presence of CD3⁺ and Vα7.2⁺ cells was analyzed by IHC in liver biopsies from HDV-infected patients and controls (Fig. 1G). There was no difference in the presence of Vα7.2⁺ cells in the livers of HDV-infected patients as compared with control liver sections, whereas an accumulation of CD3⁺ cells was clearly detected (Fig. 1G). As conventional T cells can also express Vα7.2, we next performed IF staining using Vα7.2 and IL-18Rα co-expression as a strategy to accurately identify MAIT cells, as previously reported in other human tissues.^{15,21,38–40} In contrast to the control samples, very few Vα7.2⁺ cells in the liver biopsies from HDV-infected patients co-expressed IL-18Rα (Fig. 1H). Thus, the Vα7.2⁺ cells detected through both IHC and IF in the livers of HDV-infected patients are likely not MAIT cells, indicating loss of intrahepatic MAIT cells during HDV infection.

MAIT cells are activated during chronic HDV infection

Next, we investigated the expression of immune activation markers on MAIT cells (Fig. 2A). In healthy controls and HBV mono-infected patients, peripheral blood MAIT cells expressed low levels of CD69, CD38, and HLA-DR (Fig. 2A,B). In contrast, both CD38 and HLA-DR were significantly upregulated on MAIT cells during HDV infection (Fig. 2B). Furthermore, the expression of PD-1 was significantly higher in HDV-infected patients than in healthy controls (Fig. 2B). We also assessed the expression of the co-stimulatory molecule CD28 as well as CD127 on MAIT cells, and found these receptors to be significantly lower in HDV-infected patients than in healthy controls and HBV mono-infected patients (Fig. 2C,D). In contrast, there was no difference in the expression of Ki-67 between the groups. Only few alterations were noted in HBV mono-infected patients as compared to healthy controls (Fig. 2A–D). As an additional control, we also assessed the phenotype of CD161⁻Vα7.2⁺ T cells (non-MAIT) and found no differences in the expression of the aforementioned markers between healthy controls, HBV mono-infected, and HDVinfected patients (Supplementary Fig. S1C).

MAIT cells express the innate-like T cell transcription factor promyelocytic leukemia zinc finger protein (PLZF), as well as eomesodermin (Eomes), T box transcription factor 21 (TBX21 or T-bet), and Helios (IKZF2).^37,41,42 To explore whether the phenotypic signs of activation (Fig. 2A–D) was paralleled by altered expression of master transcription factors, we examined the expression of these transcription factors in MAIT cells (Fig. 2E). Helios expression was decreased in HDV-infected patients compared with healthy controls, whereas the expression of PLZF, Eomes, and T-bet was unchanged (Fig. 2F).

In summary, these data show that MAIT cells exhibit an abnormal phenotype in chronic HDV infection with upregulation of activation markers, downregulation of co-stimulatory molecules, and altered expression of the transcription factor Helios.

High-dimensional analysis reveals a compound MAIT cell-phenotype driven by chronic HDV infection

In order to investigate the data set for complex phenotypic patterns that can easily be overlooked by conventional flow cytometry analysis, and to investigate the interrelatedness of the identified phenotypic alterations, we undertook an unsupervised high-dimensional SNE analysis to determine the overall impact of chronic HDV infection on the MAIT cell population. This type of analysis projects high-dimensional data into a low dimensional space at the single-cell level without perturbing the high-dimensional data relationships. To this end, a set of thirteen surface receptors, intracellular molecules and transcription factors, including the ones used in the analysis described above, were simultaneously assessed in MAIT cells at the single-cell level from ten healthy controls, ten HBV mono-infected patients, and seven HDV-infected patients in different pairwise comparisons (Fig. 2G–I and Supplementary Fig. S1D,E). Electronic barcodes were individually assigned to the data from each subject, data were then merged, analyzed using Barnes-Hut SNE, and finally displayed in SNE maps for patients and controls. For each comparison, we also obtained the residual plot for the differences between the groups, which shows the relative position in the SNE map of the clusters of cells that are more abundant in each of the groups being compared. This analysis confirmed substantial differences in the phenotype of MAIT cells from HDV-infected patients compared with healthy controls (Fig. 2G), as well as compared to HBV mono-infected patients (Fig. 2H and Supplementary Fig. 1D). In contrast, and in agreement with the conventional flow cytometry analysis (Fig. 2A–F), differences in the MAIT cell phenotype between HBV mono-infected patients and healthy controls were minor (Fig. 2I and Supplementary Fig. S1E).

To interpret the differences between the identified clusters in the residual plots, these were overlaid onto the individual SNE maps showing the expression intensity of each parameter in the analysis (Fig. 2J and Supplementary Fig. S1F). This analysis revealed an HDV-specific cluster of MAIT cells expressing higher levels of CD38 and PD-1, and lower levels of CD28, CD127, and of the transcription factors PLZF, Eomes, and Helios, in comparison with healthy controls (Fig. 2J). Thus, this high-dimensional analysis revealed the appearance of a compound phenotype within MAIT cells of HDV-infected patients not uncovered by the conventional single-parameter flow cytometry analysis (Fig. 2A–F). On the other hand, other differences were found in both types of analysis, such as the higher levels of CD8⁻CD4⁻ MAIT cells in HDV-infected patients when compared with healthy controls (Fig. 1C and 2J). When comparing HDV with HBV mono-infected patients, a similar cluster of CD38^hiPD-1^hiCD28^loCD127^loPLZF^loEomes^loHelios^lo MAIT cells was enriched in the HDV-infected patients (Supplementary Fig. S1F).

In conclusion, SNE analysis of this dataset revealed that higher expression of CD38 as well as PD-1 and lower expression of CD28 on MAIT cells from HDV-infected patients, as initially detected by single parameter flow cytometry analysis, occurs in the same MAIT cell subpopulation. Also, this cluster of MAIT cells simultaneously expressed lower levels of CD127 as well as lower levels of the transcription factors PLZF, Eomes, and Helios.

MR1-dependent MAIT cell responses are defective in chronic HDV infection

We next assessed the functional capacity of MAIT cells upon specific bacterial and innate cytokine-stimulation in chronic HDV infection. PBMCs were stimulated for 24 h with either mildly fixed E. coli to assess MR1-dependent responses (Supplementary Fig. S2A) or with a combination of IL-12 and IL-18. The expression of activation markers on MAIT cells (Fig. 3A), as well as their capacity to produce cytokines and degranulate (Fig. 3C), was then assessed by flow cytometry. Upon stimulation with E. coli, MAIT cells from HDV-infected patients failed to upregulate CD69 and CD25 to the same extent as healthy control MAIT cells (Fig. 3A,B). They also produced significantly lower levels of IFNγ, and degranulated less than healthy control and HBV mono-infected patient MAIT cells (Fig. 3C,D). In contrast, MAIT cells from HDV-infected patients responded to IL-12 and IL-18 stimulation equally well as those from healthy controls, with only slightly lower levels of CD25 (Fig. 3A–D). Importantly, no differences in activation readouts were observed between MAIT cells from HBV mono-infected patients and healthy controls for any of the stimulations (Fig. 3A–D), indicating that the MAIT cell functionality is not compromised by chronic HBV infection alone.

Unsupervised SNE analysis, integrating data on eleven activation markers as well as cytokines and cytotoxic molecules after E. coli stimulation, revealed extensive differences in the MAIT cell functionality between HDV-infected patients and healthy controls, and smaller differences for the HBV versus HDV, and the HBV versus healthy comparisons (Fig. 3E–G and Supplementary Fig. S2C,D). Compared to healthy controls and HBV mono-infected patients, MAIT cells in HDV infection concomitantly failed to upregulate CD69 and CD25. Furthermore, MAIT cells from patients with chronic HDV infection degranulated to a lower extent with reduced expression of CD107a and GrzB (Fig. 3H and Supplementary Fig. S2E). Taken together, these data indicate that MAIT cells in chronic HDV infection exhibit severely impaired responses against MR1-restricted bacterial antigen.

Finally, we aimed to investigate the functional relevance of the compound phenotype identified in MAIT cells from HDV infected patients. Upon IL-12 and IL-18 stimulation, PD-1⁺ MAIT cells from HDV-infected patients, but not their PD-1⁻ counterparts, failed to upregulate CD25 and to produce TNF and perforin to the same extent as PD-1⁺ MAIT cells from healthy controls and HBV mono-infected patients (Supplementary Fig. S2F). However, the functional deficient was not directly linked to increased PD-1 expression in HDV as no restoration of function was noted upon PD-1/PD-L1 blockade (data not shown). Thus, the MAIT cell compound phenotype of activation and exhaustion observed during HDV infection (Fig. 2J and Supplementary Fig. S1F) is also linked to MAIT cell dysfunction.

The loss of MAIT cells during HDV infection cannot be explained by altered MR1 expression or exposure to HDV-infected cells

We next investigated possible mechanisms behind the loss of MAIT cells in HDV-infected patients. We first assessed if HDV increases the expression of MR1 upon infection of hepatoma cells. To this end, Huh7.5^hNTCP cells were productively infected with HDV viral particles (Supplementary Fig. S3A), and the expression of MR1 was evaluated by flow cytometry (Supplementary Fig. S3B). Both surface and intracellular expression of MR1 remained unchanged in HDV-infected cells (Supplementary Fig. S3B). In addition, no difference in the expression of MR1 was observed by IHC in liver biopsies of HDV-infected patients when compared to healthy controls (Supplementary Fig. S3C). This suggests that alterations in MR1 expression are not responsible for driving the loss of circulating MAIT cells in HDV infection. Next, we assessed MAIT cell viability after exposure to supernatants from HDV-infected cells. For these studies, HepG2^hNTCP cells were infected with HDV and MAIT cells were then exposed for 24 h with different dilutions of supernatants from the infected cultures. Interestingly, and in contrast to our hypothesis, we observed that at lower supernatant dilutions, the frequency of alive (AnnV⁻DCM⁻) MAIT cells was slightly, although significantly, higher in the presence of infected supernatants as compared to control supernatants (Supplementary Fig. 3D). Thus, the loss of MAIT cells during HDV infection cannot be explained by exposure to HDV-infected cells. Instead, MAIT cells exposed to high dilutions of viral supernatant showed signs of activation with increased expression of CD107a and perforin (Supplementary Fig. S3E).

Increased levels of IL-12 and IL-18 may contribute to loss of MAIT cells in chronic HDV infection

We next determined the levels of sCD14 and sCD163, markers of monocyte/macrophage activation, in the plasma of healthy controls and HBV- and HDV-infected patients. The plasma levels of sCD14 were significantly increased in HDV-infected patients when compared with healthy controls. A significant stepwise increase in the levels of sCD163 was detected from healthy controls through HBV-infected patients to HDV-infected patients (Fig. 4A). The loss of MAIT cells significantly correlated with the levels of sCD163 (and not with the levels of sCD14) when both HBV and HDV patients were considered in the analysis (Supplementary Fig. 3F and G). However, the significant correlation was lost when HDV-infected patients were analyzed independently (Supplementary Fig. 3G). Thus, although plasma levels of inflammatory markers are increased in HDV-infected patients they cannot alone predict the loss of MAIT cells during HDV infection.

We next assessed the levels of pro-inflammatory cytokines IL-12 and IL-18 in the plasma of healthy controls and HDV-infected patients. While subunit IL-12p40 could only be detected in 2 out of 35 healthy controls, it was more abundant in HDV-infected patients where it was found in detectable levels in 25 out of 115 investigated patients (Fig. 4B). Moreover, the plasma levels of IL-18 were significantly higher in HDV-infected patients as compared to healthy controls (Fig. 4B). The continuous exposure to high levels of these pro-inflammatory cytokines could potentially drive MAIT cell loss in these patients. To test this hypothesis more directly, we compared the viability of MAIT cells between healthy controls, HBV mono-infected, and HDV-infected patients either ex vivo (Fig. 4C,D), or after in vitro incubation with a combination of IL-12 and IL-18 (Fig. 4E,F). The ex vivo apoptosis status of MAIT cells was similar between the three groups, and in general few apoptotic cells were present (Fig. 4C,D). Instead, long-term in vitro exposure to IL-12 and IL-18 significantly increased the frequency of non-viable MAIT cells in all three groups (Fig. 4E,F). Importantly, stimulation of PBMCs from healthy controls with IL-12 and IL-18 induced an abnormal MAIT cell phenotype similar to what we observed ex vivo in HDV-infected patients (Fig. 2A–D), with increased expression of CD38 and PD-1 as well as decreased expression of CD28 and CD127 (Fig. 4G). Altogether, these data suggest that dysregulated plasma levels of IL-12 and IL-18 may contribute to MAIT cell activation and loss in HDV-infected patients.