Abstract
Tick-borne encephalitis virus (TBEV) is a leading cause of viral meningoencephalitis in many parts of Europe and eastwards in Asia, with high morbidity and often long-term neurologic sequelae. With no treatment available, studies of the immune response to TBEV are essential for the understanding of the immunopathogenesis of tick-borne encephalitis and for the development of therapeutics. We have previously demonstrated that CD8+ T cell responses in peripheral blood in patients with acute TBEV peak at around 7 d after hospitalization in the neuroinvasive phase of the disease. In this study, we identified six novel TBEV HLA-A2– and HLA-B7–restricted epitopes, all derived from the nonstructural proteins of TBEV. This identification allowed for a comprehensive phenotypic and temporal analysis of the HLA-A2– and HLA-B7–restricted Ag-specific CD8+ T cell response during the acute stages of human TBEV infection. HLA-A2– and HLA-B7–restricted TBEV epitope–specific effector cells predominantly displayed a CD45RA−CCR7−CD27+CD57− phenotype at day 7, which transitioned into separate distinct phenotypes for HLA-A2– and HLA-B7–restricted TBEV-specific CD8+ T cells, respectively. At day 21, the most prevalent phenotype in the HLA-A2–restricted CD8+ T cell populations was CD45RA−CCR7−CD27+CD57+, whereas the HLA-B7–restricted CD8+ T cell population was predominantly CD45RA+CCR7−CD27+CD57+. Almost all TBEV epitope–specific CD8+ T cells expressed α4 and β1 integrins at days 7 and 21, whereas the bulk CD8+ T cells expressed lower integrin levels. Taken together, human TBEV infection elicits broad responses to multiple epitopes, predominantly derived from the nonstructural part of the virus, establishing distinct maturation patterns for HLA-A2– and HLA-B7–restricted TBEV epitope–specific CD8+ T cells.
Introduction
Tick-borne encephalitis (TBE) virus (TBEV), a member of the Flaviviridae, is the causative agent of TBE in humans. There are three main subtypes of TBEV, the European, Siberian, and Far-Eastern subtypes. The European subtype of TBEV (TBEV-Eu), which is the subtype for this study, has in most cases a characteristic biphasic course, with an “influenza-like” first phase, which may be followed by a second neuroinvasive phase with neurologic symptoms ranging from meningitis to severe meningoencephalitis (1). Around one-third of the patients with TBEV-Eu will suffer from long-term sequelae following clinically overt infection, including neuropsychiatric problems, headaches, and a substantial decrease in quality of life postinfection (1, 2). Each year, thousands of cases of TBE are reported throughout Europe and Asia, and the endemic areas are increasing (3). With no existing treatment, the only effective protection is properly timed vaccination prior to exposure to infected ticks. The currently two available TBEV vaccines in Europe are based on two different strains of the inactivated TBEV-Eu whole-virus (Neudoerfl and K23 strain) combined with aluminum hydroxide adjuvant. The vaccines require frequent doses to induce sufficient priming and to maintain long-term levels of protection. Vaccination coverage varies highly between European countries, and an increased attention to vaccine failures has been noticed during recent years (4).
The underlying mechanisms for the development of meningoencephalitis in TBEV infection remain largely unexplored. Both direct viral effects as well as immune-mediated mechanisms have been suggested to cause pathology, as both TBEV and cytotoxic CD8+ T cells can be detected in brain tissue in fatal cases (5, 6). However, fatal TBE cases may not be representative for the immune response in the majority of TBE patients. In murine models, the inflammatory reaction of CD8+ T cells has been suggested to play a role in the immunopathology of TBE because CD8−/− and SCID mice display longer survival compared with immunocompetent mice (7). In addition, clonal CD8+ T cell infiltration occurs in the brains of mice dying from TBEV infection (8). The breakdown of the blood brain barrier (BBB) in TBEV-infected mice may be independent of CD8+ T cells (9). The extent to which the BBB is affected has been associated with robust cytokine and chemokine responses and may also be influenced by genetic factors (10–12).
Ag-specific CD8+ T cell responses during acute human viral infections have previously been studied by adopting the live attenuated vaccine strains of yellow fever virus (YFV) and vaccinia virus as models (13–15). Monitoring of Ag-specific CD8+ T cell populations is essential to dissect distinct characteristics in immune responses to a certain pathogen and to delineate the role of virus-specific CD8+ T cells during acute and persistent infections as well as long-term protection. We have previously demonstrated that CD8+ T cell responses in peripheral blood of TBE patients peak at around 7 d after hospitalization during the acute neuroinvasive phase of the disease (16). In that study, we identified one HLA-A2–restricted TBEV epitope and demonstrated that TBEV-specific CD8+ T cells display a distinctive phenotypic and functional profile, paired with a distinct transcription factor expression pattern during the peak of activation. However, the fact that only one TBEV CD8+ T cell epitope had been characterized (16) clearly limited the possibilities to fully characterize the temporal aspects of the response and the contribution of different HLA class I–types to TBEV-specific CD8+ T cell responses in humans. Furthermore, these previous investigations did not encompass any studies on properties bearing characteristics of a migratory capacity to the CNS.
In the current study, we describe the identification of six novel HLA class I–presented TBEV epitopes recognized by CD8+ T cells. Using these epitopes, we carry out a comprehensive characterization of the TBEV-specific CD8+ T cell response during the course of acute infection in the neuroinvasive phase during TBE. Detectable frequencies of TBEV-specific CD8+ T cells were found at days 7 and 21 following hospitalization. At day 7, the predominant maturation profile shared between A2- and B7-restricted CD8+ T cells was CD45RA−CCR7−CD27+CD57−, phenotypically defining them as effector memory cells. However, the dominant phenotype of TBEV-specific CD8+ T cells differed between A2- and B7-restricted responses at day 21. The most prevalent phenotype in the HLA-A2–restricted cells was CD45RA−CCR7−CD27+CD57+, whereas the most prevalent phenotype from HLA-B7–restricted cells was CD45RA+CCR7−CD27+CD57+. Almost all of the TBEV-specific CD8+ T cells expressed both α4 and β1 integrin at days 7 and 21. The high level of integrin expression on these cells may indicate capabilities toward migration into the CNS: the major site of pathological reactions and disease-associated symptoms.
Materials and Methods
Ethics statement
All included patients and healthy volunteers gave written informed consent to participate in the study. The Kaunas Regional Research Ethics Committee, Lithuania, and the Regional Ethical Review Board in Stockholm, Sweden, approved the study.
Study design and subjects
Peripheral blood was obtained from 10 confirmed TBEV-infected patients from the Department of Infectious Disease at the Lithuanian University of Health Sciences in Kaunas, Lithuania. The donors in this study consisted of a separate cohort from previously published work (16, 17). TBEV infection was confirmed by specific IgM- and/or IgG-TBEV analysis in serum according to standard clinical diagnostic criteria together with symptoms associated with meningitis or meningoencephalitis. Blood samples were collected in CPT tubes (BD Biosciences, San Jose, CA) within 3 d after hospital admission (day 0) and subsequently on day 7, 21, and 90. PBMC were isolated and cryopreserved for later analysis in 90% FCS and 10% DMSO. Uninfected TBEV IgM and IgG seronegative community-matched individuals not previously vaccinated against TBEV were used as the reference group. EDTA tubes were used to collect whole blood for DNA isolation.
HLA class I genotyping
Genomic DNA was isolated from whole blood using a DNeasy kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions. HLA class I typing was performed using a multiplexed reverse sequence–specific oligonucleotide probe method (LABType SSO; One Λ, Canoga Park, CA) according to the manufacturer’s instructions. The haplotypes of the respective donors are presented in Supplemental Table I.
Generation of HLA class I monomers and UV exchange
HLA-A2 and HLA-B7 monomers were generated as previously described (18). In brief, HLA class I H chains were expressed in bacteria, grown on Luria-Bertani media, and isopropyl β-d-thiogalactoside was added to induce protein expression. Bacteria were lysed, and the protein was isolated. Refolding of HLA class I molecules was done using the conditional ligands KILGFVGJV (for HLA-A2, in which the J indicates the UV-sensitive amino acid) and AARGJTLAM (for HLA-B7) (18). Aliquots were stored at −80°C until further use.
UV exchange and generation of MHC tetramers was performed as previously described (18, 19). In brief, 200 μg/ml monomer was mixed with 400 μM of peptide diluted in PBS in a 96-well plate. The mixture was then exposed for 1 h to long-wave UV light, after which it was centrifuged for 5 min at 3000 × g. The supernatant was transferred to a fresh 96-well plate and incubated for 30 min with streptavidin-conjugated fluorescent Abs at 4°C. A mixture of 25 μM of d-biotin and 0.02% Sodium Azide (final concentration) was added to prevent exchange of streptavidin-conjugated Abs during tetramer staining. Tetramers were stored up to 1 wk at 4°C.
To identify new epitopes, each tetramer/epitope complex was conjugated with two different streptavidin fluorescent Abs as described previously (19). The following streptavidin Abs were used: streptavidin APC, streptavidin PE, streptavidin Brilliant Violet 421, streptavidin Brilliant Violet 605, streptavidin Brilliant Violet 650 (all from BioLegend, San Diego, CA), and streptavidin Brilliant Violet 711 (BD Biosciences). Brilliant Violet streptavidin Abs were mixed in a 4:1 ratio with APC and PE to maintain equal fluorescence. 3 × 106 cells per well were then stained with 2 μl of each tetramer/epitope mixture and incubated for 15 min at 37°C before further staining.
Flow cytometry Abs
The following Abs were used in this study: from BD Biosciences: CD14 FITC, CD19 FITC, CD14 V500, CD19 V500, granzyme B PE-CF594, CXCR3 Alexa Fluor 488, and CD103 Brilliant Violet 711; from Invitrogen (Waltham, MA): CD4 Q605, CD8 Q605, CD8 Q705, Green Dead Cell Marker Kit, and Aqua Dead Cell Marker Kit; from AbD Serotec (Kidlington, U.K.): CD40 FITC; from BioLegend: CD8 Alexa Fluor 700, HLA-ABC Alexa Fluor 647, CD29 (Integrin β-1) APC, CCR5 Alexa Fluor 700, CD14 APC-Cy7, CD19 APC-Cy7, CD57 Pacific Blue, CD49d (integrin α-4) Brilliant Violet 510, CD27 Brilliant Violet 570, CCR6 Brilliant Violet 650, CCR7 Brilliant Violet 785, CD4 PE-Cy5, and CD45RA PE-Cy7; and from Beckman Coulter (Brea, CA): CD3 ECD and CD3 PE-Cy5.
Extra- and intracellular flow cytometry staining
1–3 × 106 cells per well were incubated for 30 min in 4°C in the dark with surface Abs, followed by washing with PBS containing 0.1% BSA and 2 mM of EDTA. When no intracellular staining was needed, cells were fixed with 4% PFA for 30 min at 4°C. For intracellular stainings, cells were fixed and permeabilized in the dark at 4°C for 30 min with fix/perm buffer (eBioscience, San Jose, CA). Cells were then washed and stained with intracellular mAbs in perm/wash buffer (eBioscience). Samples were measured on a BD LSRFortessa instrument (BD Biosciences) and analyzed using FlowJo software version 9.7.5 (Tree Star, Ashland, OR), GraphPad Prism 6 (GraphPad Software, La Jolla, CA), and SPICE 5.32 software provided by National Institute of Allergy and Infectious Disease (20).
CD8+ T cells were identified as CD8+ cells among single, live, CD14−, CD19−, and CD3+ lymphocytes (Supplemental Fig. 1A). Gating strategy for tetramer-positive CD8+ T cells can be viewed in Supplemental Fig. 1A. Supplemental Fig. 1B demonstrates the gating strategy for the cell surface stabilization assay. Gating strategy in large-scale screening can be found in Supplemental Fig. 2B.
Epitope prediction and generation of synthetic peptides
Candidate epitopes presented by HLA-A2 or HLA-B7 were predicted using the netCTL search engine (version 1.2) (21). The HLA class I epitope predictions were performed using the polyprotein of the Neudoerfl TBEV strain, representative of TBEV-Eu (access code UL27495.1). For epitope selection, standard settings were used as suggested by netCTL, and the 44 best scoring peptides for both HLA-A2 and -B7 supertypes were selected for synthesis. Peptides were synthesized by GenScript (GenScript USA, Piscataway, NJ) and dissolved in 10% DMSO and 90% PBS to a concentration of 5 mg/ml. The peptides were stored at −20°C until further use.
Generation of RMA-S cells expressing HLA-I
Tumor cell line RMA-S was used to generate HLA-A2– and HLA-B7–expressing cells for cell surface stabilization assay. Coding sequences of HLA-A*0201 and HLA-B*0702 were amplified and His-tagged by PCR using primers 5′-cac cAT GGC CGT CAT GGC GCC-3′ and 5′-CTC GAG TCA GTG ATG GTG ATG GTG ATG CAC TTT ACA AGC TGT GAG-3′ (for A2) and 5′-cac cAT GCT GGT CAT GGC GCC-3′ and 5′-CTC GAG TCA ATG GTG ATG GTG ATG AGC TGT GAG AGA CAC ATC AGA G-3′ (for B7). The amplicons were cloned into pENTR from ThermoFisher Scientific (Waltham, MA) (plasmids no. 108213 and no. 104849, respectively; Addgene) and further recombined into the retroviral expression plasmid pMP71 (plasmids no. 108214 and no. 104850, respectively; Addgene). Retroviral particles were prepared and used to transduce RMA-S cells, which were sorted to purity using antiHLAABC (W6/32) Ab (our collection).
Cell surface stabilization assay
RMA-S HLA-A2 β2m and RMA-S HLA-B7 β2m were cultured in complete RPMI 1640 medium (Life Technologies, Waltham, MA) containing 10% FCS, 2 mM l-glutamine, 1% penicillin, streptomycin (Invitrogen, Carlsbad, CA), and 250 μg/ml G418 (Life Technologies) at 37°C. Prior to the assay, cells were cultured overnight in fresh RPMI medium lacking FCS (to deprive the medium of β2m) at 26°C to accumulate HLA class I molecules. Concentration series started at 20 μg/ml (HLA-A2) or 100 μg/ml (HLA-A2 and HLA-B7) and were diluted six times by a factor of 2 in complete RPMI medium containing no FCS. Influenza A2 (GILGFVFTL), CMV-A2 (NLVPMVATV), and CMV-B7 (TPRVTGGGAM) were used as positive and negative controls in the experiment. A total of 1 × 105 cells per well were added and incubated for 4 h at 37°C. Wells containing only RMA-S cells were included as a baseline. Afterward, cells were stained with HLA-ABC Alexa Fluor 647 (BioLegend), and the mean fluorescence intensity was measured by flow cytometry. The binding index was calculated as follows: binding index = mean fluorescence intensity (sample)/mean fluorescence intensity (RMA-S cells alone).
Statistical analysis
Analyses were performed using GraphPad Prism software 5.0 for Mac OS X (GraphPad Software). Data were analyzed by nonparametric repeated measures Mann–Whitney U test or Wilcoxon signed-rank test; p values <0.05 were considered statistically significant.
Results
Identification of novel HLA-A2– and HLA-B7–restricted TBEV epitopes
To characterize CD8+ T cell responses during acute TBEV infection, we set out to first identify a broader set of TBEV-specific CD8+ T cell epitopes than previously described, using a combinatorial tetramer screening method (19). Fluorescently labeled MHC class I multimers enable the identification of specific T cell epitopes from a limited amount of patient samples, especially when combinatorial staining strategies are used (19, 22). This technology has been successfully used for the identification of MHC class I–restricted bacterial and viral epitopes as well as novel minor histocompatibility Ags (23–25). Patients selected for the epitope screen fulfilled the criteria for acute TBEV infection and were carriers of either HLA-A2 (n = 5) or HLA-B7 (n = 5). Patients were included in the study within 3 d of hospitalization (day 0), and PBMCs were isolated at days 0, 7, 21, and 90.
For the selection of TBEV epitope candidates, the Neudoerfl TBEV strain sequence was analyzed using the netCTL search engine (21). The top 44 scoring nonamer peptides for HLA-A2 and HLA-B7 were synthesized (Supplemental Fig. 2A, Supplemental Table II). MHC class I/peptide tetramers were subsequently generated for each candidate peptide using a UV exchange technique in which the MHC class I molecule is initially refolded with a high affinity epitope that breaks up following exposure to UV light and can be exchanged for other peptides (22). The influenza-derived HLA-A2 epitope influenza M1 (GILGFVFTL) and the CMV-derived HLA-B7 epitope CMV pp65 (TPRVTGGGAM) were used as internal positive controls. A combinatorial staining approach based on two fluorophores per epitope was applied to increase staining sensitivity and the number of epitopes that can be tested in a single sample (19, 25). Using this approach and six fluorophores, we were able to screen CD8+ T cell responses toward 15 peptides per 3 × 106 PBMCs from TBEV-infected patients.
A total of 10 patients were screened at days 0, 7, 21, and 90 following hospitalization. Identification of tetramer-positive cells was performed in accordance with previously described gating strategies (Supplemental Fig. 2B, 2C) (19). Cumulative analysis of the results revealed significant CD8+ T cell populations specific for the four TBEV-derived HLA-A2–restricted epitopes A2-1 (MLLQAVFEL), A2-2 (ILLDNITTL), A2-19 (SLINGVVKL), and A2-30 (NIWGAVEKV) (“major epitopes”), including the epitope previously identified by us (Fig. 1A, Supplemental Fig. 2A) (16). Significant CD8+ T cell responses were also observed against the HLA-B7–restricted epitopes B7-3 (RPVWKDARM), B7-4 (RVRFHSPAV), and B7-6 (LPLGHRLWL) (major epitopes) (Fig. 1B, Supplemental Fig. 2A). Small but still detectable populations specific for the TBEV-derived HLA-A2–restricted peptides A2-4 (RMMGILWHA), A2-40 (RTAWFVPSI), and A2-43 (SALDVFYTL) were also observed in a few donors (“minor epitopes”) (Fig. 1A).
Screening of HLA-A2– and HLA-B7–binding TBEV peptides.
(A) Frequencies of HLA-A2 tetramer–positive CD8+ T cells at days 0, 7, 21, and 90 (n = 5). (B) Frequencies of HLA-B7 tetramer–positive CD8+ T cells at days 0, 7, 21, and 90 (n = 5). Data are shown as mean ± SEM of five donors for each HLA type, each donor was run as an independent experiment.
All the identified major HLA-A2– and HLA-B7–restricted CD8+ T cell epitopes were located within the nonstructural proteins of TBEV (Table I, Supplemental Fig. 2A). One of the identified minor epitopes (A2-4) was located in the C protein and has previously, as part of a 15-mer, been shown to give rise to a CD4 T cell response in previously TBEV-infected and vaccinated healthy volunteers (26). All the major epitopes were highly conserved among different TBEV subtype sequences, including the Western European, Siberian, and Far-Eastern subtypes. However, the A2-1 and B7-3 epitopes comprised a limited amount of amino acid differences between these viral strains (A2-1: M/T at peptide position 1, B7-3: K/R at position 5) (Table I).
The TBEV HLA-A2–restricted epitope A2-2 has been previously identified by assessing degranulation as well as cytokine and chemokine expression in CD8+ T cells following in vitro stimulation of PMBC with a pool of potential TBEV-associated peptides predicted by the netCTL algorithm (16). The positive control tetramers with epitopes (GILGFVFTL) and (TPRVTGGGAM) stained positive throughout the course of infection, indicating previous exposure to these pathogens in all patients. However, it should be noted that no expansion of either CMV nor influenza epitope–specific CD8+ T cells was observed (Fig. 1A, 1B), supporting the notion that very little bystander activation of these CD8+ T cell populations occurs during the course of TBEV infection (16).
Magnitude and kinetics of the CD8+ T cell–specific responses to HLA-A2 and HLA-B7 TBEV epitopes
All tested patients stained positive for HLA-A2 tetramers presenting the TBEV peptide epitopes A2-1, A2-2, and A2-30 (Fig. 2A). Furthermore, 2–3% of the CD8+ T cells stained positive at day 21 in one patient for the TBEV A2-19 tetramer complex, whereas this epitope was not recognized in the other four patients (Fig. 2A). All tested patients also stained positive for at least one of the HLA-B7 TBEV–specific epitope tetramers (Fig. 2B). The HLA-B7–restricted B7-6 peptide tetramer detected the highest percentage of specific CD8+ T cells (range 0.12–1.16%) and was recognized by CD8+ T cells derived from four out of five patients. CD8+ T cells from four out of five patients also recognized efficiently the TBEV HLA-B7–restricted epitopes B7-3 and B7-4.
TBEV epitope–specific CD8+ T cells vary among patients.
(A) Frequency of HLA-A2–positive CD8+ T cells at days 0, 7, 21, and 90 after hospitalization for A2-1, A2-2, A2-19, and A2-30, respectively (n = 5). (B) Frequency of HLA-B7–positive CD8+ T cells at days 0, 7, 21, and 90 after hospitalization for B7-3, B7-4, and B7-6, respectively (n = 5). Data for (A) and (B) are shown as individual data points, with one data point for each donor (n = 5), each donor was run as an independent experiment. (C) Flow plots show representative stainings of CD8+ T cells with single-stained PE-labeled tetramers in TBE patients (top row) and community-matched healthy controls (bottom row) for HLA-A2 (top panels) and HLA-B7 (bottom panels), respectively. Number of donors, n = 3 for A2-1, A2-2, and B7-6; for others n = 2. Negative control stainings: n = 3 for A2-1, A2-2, and B7-6; n = 2 for all other epitopes.
In general, the levels of epitope-specific CD8+ T cells were very low or undetectable at day 0. Detectable frequencies of epitope-specific cells were found at days 7 and 21 following hospitalization, with the highest percentages of TBEV HLA-A2 and HLA-B7 tetramer–positive CD8+ T cells at day 21. These numbers decreased by day 90 (Fig. 2A, 2B). Overall, HLA-B7–restricted CD8+ T cell populations were smaller than those specific for HLA-A2–restricted epitopes. To confirm the specificity of the TBEV-specific epitopes, additional comparative single tetramer stains were performed on PBMC from patients and healthy controls (Fig. 2C). Although epitope-specific CD8+ T cells were detected in PBMC from TBEV-infected patients, all healthy controls were consistently negative.
TBEV-derived peptides form stable HLA-I/peptide cell surface complexes
To study if binding to the respective restriction elements (HLA-A2 and -B7, respectively) correlated with the observed patterns of immunodominance, binding of the peptides was evaluated for each identified peptide using cell surface stabilization assays. All the tested HLA-A2–restricted peptides were able to form stable HLA-A2 complexes at a peptide concentration of 20 μg/ml (Fig. 3A). The HLA-A2–restricted peptides A2-1 and A2-30 displayed relatively low stabilization capacity, and therefore additional assays with up to 100 μg/ml peptide were performed. Both peptides displayed low stabilization capacity (IC50 >100 μg/ml), although both were able to bind to HLA-A2 (data not shown). In contrast, peptides A2-2 and A2-19 displayed high stabilization capacity (IC50 <25 μg/ml) for HLA-A2. The identified HLA-B7–restricted peptides all displayed high stabilization capacity (IC50 <25 μg/ml) (Fig. 3B). Notably, TBEV peptide A2-2, with the highest stabilization capacity, was also on average the most strongly recognized epitope, supporting the expected association between its dominant recognition and HLA class I stabilization capacity. However, the TBEV epitope A2-1, which displayed the lowest stabilization index, was still recognized in all patients tested but to a lower magnitude compared with A2-2. The TBEV peptide B7-3 displayed the highest stabilization capacity in the B7 cell surface stabilization assay, followed by peptide B7-6. Both these peptides correspond to the most strongly recognized epitopes for HLA-B7 in patients. Based on the stabilization capacity and the magnitude of recognition in patient material, A2-2 and B7-3/B7-6 displayed the characteristics of immunodominant epitopes.
Stability of TBEV peptide binding to HLA-A2 and HLA-B7.
(A) Stability index of HLA-A2 epitopes in stability assay using RMA-S A2 β2m cells. (B) Stability index of HLA-B7 epitopes in stability assay using RMA-S B7 β2m cells. Data are shown as stability index and were calculated as follows: stability index = mean fluorescence intensity (sample)/mean fluorescence intensity (RMA-S cells alone). The stability of each peptide was measured in two independent experiments. A representative of one experiment is shown.
CD8+ T cell effector and memory profiles toward HLA-A2 and HLA-B7 in complex with TBEV-associated epitopes
With access to both HLA-A2– and -B7–restricted TBEV-specific epitopes, we subsequently studied the temporal aspects of the CD8+ T cell response based upon the HLA restriction element. With the highest percentage of TBEV-specific CD8+ T cells found at days 7 and 21, we used samples from these time points to analyze the epitope-specific CD8+ T cell populations. Based on the higher percentages of epitope-specific CD8+ T cells for the HLA-A2–restricted peptides A2-1 and A2-2, as well as the HLA-B7–restricted peptides B7-3 and B7-6, we selected these epitopes to further study the phenotype of TBEV-specific CD8+ T cells in more detail using multiple HLA tetramers in parallel.
Using the surface expression of CD45RA and CCR7, we defined naive (CD45RA+CCR7+), central memory (CD45RA−CCR7+), effector memory (CD45RA−CCR7−), and effector memory RA (CD45RA+CCR7−) CD8+ T cell populations (27, 28). TBEV epitope–specific (tetramer-positive) CD8+ T cells were additionally characterized for their expression of CD45RA, CCR7, and CD27 as well as the senescence marker CD57 (Fig. 4A). First, we analyzed these individual markers on tetramer-positive CD8+ T cells, either using the immunodominant HLA-A2–restricted epitope A2-2 in HLA-A2+ donors (n = 5) or the immunodominant HLA-B7 epitope B7-6 in HLA-B7+ donors (n = 4). Expression level of CD27, as measured by median fluorescence intensity (MFI), was significantly higher among tetramer-positive cells at days 7 and day 21, as compared with tetramer-negative cells (Fig. 4B). There were significantly lower percentages of CD45RA+ and CD57+ tetramer–positive cells at day 7, as compared with tetramer-negative CD8+ T cells in infected donors and healthy controls (Fig. 4B). At day 21, the percentage of CD45RA+ tetramer–positive cells were higher than at day 7 (Fig. 4B). However, the frequency of this population within tetramer-positive cells was still lower than in tetramer-negative population in infected donors and healthy controls (Fig. 4B).
Phenotype of HLA-A2 and HLA-B7 TBEV tetramer–defined CD8+ T cell populations.
(A) Representative stainings of CD45RA, CD27, CD57, CCR7, and A2-2 tetramer at day 7 (upper panel) and day 21 (lower panel) after hospitalization. Gated on total CD8+ T cells. (B) Bar plots show the expression of CD27 in terms of MFI, and the percentages of cells expressing CD45RA and CD57 in tetramer-positive CD8+ T cells (A2-2 or B7-6) and tetramer-negative CD8+ T cells at days 7 and 21 after hospitalization, respectively, and healthy control CD8+ T cells. n = 5 (HLA-A2), 5 (HLA-B7), and 6 (HC), data are from three independent experiments, including either two patients and one healthy individual (two experiments) or one patient and one healthy individual (one experiment). (C) Bar chart represents the subset distribution of CCR7, CD27, CD45RA, and CD57 in tetramer-positive CD8+ T cells at day 7 and day 21 for A2-1, A2-2, B7-3, and B7-6, respectively (n = 5). Statistical analysis was performed using Mann–Whitney U test or Wilcoxon signed-rank test. *p < 0.05, **p < 0.005, ***p < 0.0005.
Next, we used the HLA-A2–restricted peptides A2-1 and A2-2 and the HLA-B7–restricted peptides B7-3 and B7-6 to further study the maturation profile of TBEV-specific (tetramer-positive) CD8+ T cells. The most prevalent phenotype in A2-1 and A2-2 as well as B7-3 and B7-6 tetramer–positive CD8+ T cells at day 7 was CD45RA−CCR7−CD27+CD57− cells (∼40–50% of the cells) (Fig. 4C). This phenotype decreased over time, and at day 21 the A2-1– and A2-2–specific CD8+ T cell populations were predominantly CD45RA−CCR7−CD27+CD57+ (∼25%), whereas the most prevalent phenotype in the B7-3– and B7-6–specific CD8+ T cell population was CD45RA+CCR7−CD27+CD57+ (∼40% of the cells) (Fig. 4C). Altogether, these results indicate that different CD8+ T cell responses may behave differently, segregated by HLA allele.
Chemokine receptor expression on TBEV-specific CD8+ T cells
TBEV infection causes often-severe pathology affecting the CNS (1). This raises the possibility of pathogenesis being related to the migration of activated T cells to this site. We therefore set out to study the integrin and chemokine receptor expression on the TBEV-specific CD8+ T cells to assess if they expressed homing receptors associated with a potential to migrate to the CNS following activation. CD8+ T cells were stained for the markers α4 integrin (CD49d), β1 integrin (CD29), and CCR6, all associated with T cell migration through the BBB (29). Furthermore, CD8+ T cells were also stained for the E-cadherin receptor CD103, previously suggested to be expressed on T cells in the CNS during infection with other flaviviruses (30), and for the chemokine receptors CXCR3 and CCR5, which have been suggested in previous studies to be involved in TBE, potentially by facilitating recruitment of T cells into the cerebrospinal fluid (31–33). Patient samples were stained with these markers, together with peptide A2-1, A2-2, B7-3, and B7-6 tetramers. Although detected in tetramer-negative CD8+ T cell population, CCR5, CCR6, and CD103 were not detected on tetramer-positive cells, from neither HLA-A2– nor HLA-B7–positive donors at days 7 or 21 (Fig. 5A). In contrast, almost 100% of the HLA-A2 and HLA-B7 tetramer–positive cells expressed both α4 and β1 integrins, whereas these two markers were expressed in around 80% of tetramer-negative CD8+ T cells from the patients and healthy controls (Fig. 5A, 5B). It should be noted that a general decrease in MFI levels was consistently observed for both α4 and β1 in TBEV-specific CD8+ T cells over time (Fig. 5B).
Chemokine profile of HLA-A2 and HLA-B7 tetramer–defined CD8+ T cell populations.
(A) Representative staining of CXCR3, integrin α4, integrin β1, CCR5, CCR6, CD103, and HLA-A2-2 TBEV epitope tetramer at day 7 (upper panel) and day 21 (lower panel) after hospitalization. Gated on total CD8+ T cells. (B) Line graphs show the expression of CXCR3, α4, and β1 in terms of MFI in HLA-A2–, A2-1–, and A2-2–specific cells and HLA-B7–, B7-3–, and B7-6–specific cells, respectively (upper panel). Bar plots show the expression of CXCR3, α4, and β1 in tetramer-positive CD8+ T cells (A2-2 or B7-6) and tetramer-negative CD8+ T cells at days 7 and 21 after hospitalization and healthy control CD8+ T cells (lower panel). n = 5 (HLA-A2), 5 (HLA-B7), 6 (HC), data are from three independent experiments, including either two patients and one healthy individual (two experiments) or one patient and one healthy individual (one experiment). (C) Bar chart represents the subset distribution of α4, β1, and CXCR3 in tetramer-positive cells at day 7 and day 21 for A2-1, A2-2, B7-3, and B7-6, respectively. Statistical analysis was performed using Mann–Whitney U test or Wilcoxon signed-rank test. *p < 0.05, **p < 0.005, ***p < 0.0005.
CXCR3 expression followed different patterns of expression on tetramer-positive CD8+ T cells in the individual patients. In six out of 10 patients, CXCR3 expression was higher at day 7 but decreased at day 21, whereas the opposite pattern was observed in three patients, and in one donor the levels remained stable over time (Fig. 5B). The most prevalent phenotype for the tetramer-positive CD8+ T cells, with specificity for either HLA-A2 or HLA-B7/peptide complexes, was CXCR3−α4+β1+ (∼50–70%), whereas the remainder displayed a CXCR3+α4+β1+ phenotype at day 7 (Fig. 5C). Taken together, with almost all tetramer-positive CD8+ T cells expressing α4 and β1 integrin, these results may indicate that TBEV-specific CD8+ T cells in peripheral blood have properties indicative of a capacity to enter tissues, including the CNS.
Discussion
Surprisingly little is known about the role of CD8+ T cells in TBEV infection, and human studies on CD8+ T cell responses in TBEV pathogenesis are still scarce (6, 16). Further studies of these cells are thus necessary to gain further insights into the mechanisms of immune pathogenesis to help the design of new treatments and for the development of new, more effective vaccines.
In the current study, we identified three novel HLA-A2–restricted and three novel HLA-B7–restricted TBEV CD8+ T cell epitopes. Epitope-specific responses against these epitopes were observed at days 7 and 21 following hospitalization. At day 7, ∼40–50% of the TBEV-specific CD8+ T cells were CD45RA−CCR7−CD27+CD57−, phenotypically defining them as effector cells. The HLA-A2– and HLA-B7–restricted TBEV-specific CD8+ T cells then transitioned into separate distinct phenotypes at day 21. Almost all of the HLA-I/peptide tetramer–positive cells expressed α4 and β1 integrins at days 7 and 21, whereas the bulk CD8+ T cells expressed lower levels of integrins, findings that might be indicative of phenotypical changes that relate to the capacity of TBEV-specific CD8+ T cells to migrate into the CNS.
CD8+ T cells have been suggested to be involved in the control of other flaviviral infections. Previous studies during acute dengue virus (DENV) infection have indicated an association between HLA class I genotypes and susceptibility to DENV disease. Certain HLA class I alleles were clearly associated with a significantly reduced epitope-specific CD8+ T cell response and thus increased susceptibility to DENV disease (34). Although a number of CD8+ T cell epitopes specific for other flaviviruses such as DENV (34), West Nile virus (35), and YFV (15) have previously been identified, to date, only one TBEV-specific viral CD8+ T cell epitope is known (16). This has significantly hindered our capacity to characterize TBEV-specific CD8+ T cell responses in more detail.
We used the previously described UV exchange technology (36), along with the generation of multicolor codes (19), to identify new T cell epitopes. This approach allows for the screening of a relatively large number of epitope candidates using only a small amount of patient sample, which is usually the limiting factor for studies based on clinical material during natural causes of infections. This strategy enabled us to identify and characterize four HLA-A2– and three HLA-B7–restricted TBEV epitopes, including the previously identified HLA-A2–restricted epitope (16). All the major TBEV epitopes formed stable complexes with their cognate HLA class I at the cell surface. Notably, the highest stabilization capacity for HLA-A2 was observed for peptide A2-2. The use of this peptide in tetramer constructs resulted in the highest percentage of tetramer-positive cells in most patients, confirming its immunodominant features.
The identification of a new set of epitopes allowed us to study if the temporal aspects of the CD8+ T cell response depend on epitope specificity. In the phenotypic studies of tetramer-defined CD8+ T cells specific for the HLA-A2–restricted peptides A2-1 and A2-2 and the HLA-B7–restricted epitopes B7-3 and B7-6, we observed that the dominant phenotype is CD45RA−CCR7− at day 7 after hospitalization. This was followed by memory formation on day 21, where epitope-specific CD8+ subsets gained expression of CD45RA and CD57. Interestingly, the major phenotype of TBEV-specific CD8+ T cells differed in HLA-A2 and HLA-B7 donors at day 21. The most prevalent phenotype in the TBEV-specific CD8+ T cell populations from HLA-A2 donors was CD45RA−CCR7−CD27+CD57+, whereas the most prevalent phenotype from HLA-B7 donors was CD45RA+CCR7−CD27+CD57+, suggesting different kinetics in effector memory RA CD8+ T cell development. The HLA-A2– or HLA-B7–restricted epitope-specific CD8+ T cells were either absent or present at very low levels at the day of hospitalization and appeared 1 wk later, consistent with our previous findings (16). These findings further support our hypothesis that primary T cell responses to TBEV infection occur in the meningoencephalitic phase and likely not in the first phase of infection.
The mechanisms underlying TBE immunopathogenesis in humans are still largely unknown. It is believed that the CNS is infected during the meningoencephalitic phase. To clear the virus, Ag-specific T cells would need to enter the CNS. Virus-specific effector T cells are recruited to the CNS during infections by chemokines and integrins (37). The migration of CD8+ T lymphocytes across the BBB mainly depends on α4 integrin, as demonstrated both in vitro and in vivo in animal models for experimental autoimmune encephalomyelitis (EAE) and coronavirus-infected (mouse hepatitis virus-A59) mice (38, 39). However, it should be noted that conflicting data exist on the importance of α4 integrin for CD8+ T cell entry into the CNS upon coronavirus-induced encephalitis in mice (40). Most previous studies on CNS recruitment of T cells have been performed in neuroinflammatory diseases, such as multiple sclerosis (MS) and in EAE animal models for MS as well as in mouse models for viral infections. However, to date, not so much is known on CNS recruitment of T cells during human acute viral CNS infections, and it remains unknown how TBEV-specific CD8+ T cells enter the CNS. In this paper, we established that almost all the epitope-specific CD8+ T cells expressed α4 integrin and β1 integrin at day 7 and day 21, properties associated with an ability to infiltrate tissues and cross the BBB. CXCR3-expressing T cells have been suggested to play an important part in TBEV infection, as increased CXCL10 expression has been found in the CNS of TBEV-infected patients (31, 32). However, our present results did not establish higher percentages of epitope-specific CD8+ T cells compared with bulk CD8+ T cells expressing CXCR3, even if we noticed a trend for higher levels of expression of CXCR3 at day 7 compared with day 21 in epitope-specific CD8+ T cells.
CD8+ T cells localized in the brain of mice infected with West Nile virus display a significantly increased expression of the E-cadherin receptor CD103, suggesting a role for CD103 expression in allowing memory CD8+ T cells to remain in the brain postinfection (30). Other studies indicated that CCR5 may play a protective role in TBEV pathogenesis, as patients with a 32-bp deletion in the CCR5 gene displayed an increased risk of developing clinical TBE infection (33). Further, it has been previously established that CCR6+ α4β1+ T cells may act as “brain-homing” T cells in patients with MS, with similar observations in EAE animal models (29). However, in the current study we did not observe expression of the markers CCR5, CCR6, or CD103 in TBEV-specific CD8+ T cells in the peripheral blood, although these markers were expressed to some extent in the bulk CD8+ T cell population. The lack of expression on virus-specific CD8+ T cells could possibly be explained by the high proportion of virus-specific CD8+ T cells that may already reside in the CNS during this time of infection or by the possibility that these receptors are not crucial for CD8+ T cell migration across the BBB in TBEV infection. To elucidate this, further investigations of CD8+ T cells in the cerebrospinal fluid will be necessary.
All the major TBEV epitopes identified in the current study are located in nonstructural proteins of the virus, whereas T cell epitope studies in TBEV infection and vaccination so far have focused only on virion proteins (26, 41). The majority of identified CD8+ T cell viral epitopes in YFV-17D or DENV are also derived from nonstructural proteins in these viruses (13, 14, 16, 34). Studies of DENV infection have demonstrated that CD4+ T cell responses are skewed toward recognition of the envelope, capsid, and NS1, whereas CD8+ T cell responses are mainly focused on viral nonstructural proteins (42). In Japanese encephalitis virus infection, T cell responses were mostly CD8+ T cell–mediated, targeting nonstructural proteins in asymptomatic Japanese encephalitis virus infection, whereas the response in symptomatic infection were mostly CD4+ T cell–mediated, targeting both structural proteins and the secreted protein NS1 (43).
In conclusion, we describe in this paper the identification of six novel TBEV-specific CD8+ T cell epitopes recognized during acute severe TBEV infection in humans. Our study extensively delineates the phenotype and maturation status of virus Ag–specific CD8+ T cells, including properties associated with a capacity to migrate to tissues, including the CNS. With no antiviral treatment to TBEV infection available, new strategies for treatment and vaccination are needed. Access to a larger number of TBEV-specific epitopes provides an opportunity for the design of novel vaccines. Likewise, insights into mechanisms of T cell entry to the CNS may pave the way for new treatment strategies.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank the patients and healthy donors at the Department of Infectious Diseases, Lithuanian University of Health Sciences, Kaunas, Lithuania. We also thank Dr. Carsten Linnemann for technical advice on the combinatorial tetramer screening method, Dr. M. Schaffer for HLA-typing of the patients, and I. Pilitauskiene and L. Lesnikauskiene for help with sample preparation.
Footnotes
This work was supported by the Wenner-Gren Foundations, the Marianne and Marcus Wallenberg Foundation, the Jeansson Foundations, the Åke Wiberg Foundation, the Siri Lindström Foundation, the Magnus Bergvall Foundation, and the Swedish Society of Medicine (all to S.G.-R.). Parts of this study were also supported by the Swedish Research Council (to A.A., H.-G.L., J.K.S.), the Swedish Cancer Society (to A.A.), and the Swedish Foundation for Strategic Research (to H.-G.L.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Abbreviations used in this article:
- BBB
- blood brain barrier
- DENV
- dengue virus
- EAE
- experimental autoimmune encephalomyelitis
- MFI
- median fluorescence intensity
- MS
- multiple sclerosis
- TBE
- tick-borne encephalitis
- TBEV
- TBE virus
- TBEV-Eu
- European subtype of TBEV
- YFV
- yellow fever virus.
The online version of this article contains supplemental material.
- Received April 27, 2018.
- Accepted June 11, 2018.
- Copyright © 2018 The Authors
This article is distributed under the terms of the CC BY-NC-ND 4.0 Unported license.