Identification of a First Human Norovirus CD8+ T Cell Epitope Restricted to HLA-A*0201 Allele

Norovirus (NoV) causes a substantial global burden of acute gastroenteritis in all age groups and the development of NoV vaccine is a high priority. There are still gaps in understanding of protective NoV-specific immunity. Antibody mediated immune responses have been widely studied, but in contrast, the research on NoV-specific human T cell-mediated immunity is very limited. We have recently reported NoV capsid VP1-specific 18-mer peptide (134SPSQVTMFPHIIVDVRQL151) to induce strong CD8+ T cell immune responses in healthy adult donors. This work extends to identify the precise NoV T cell epitope and the restricting human leucocyte antigen (HLA). Pentamer technology was used to detect HLA-A*0201-restricted T cell-mediated responses to 10-mer peptide 139TMFPHIIVDV148 of four healthy adult blood donors. Immunogenicity of the 10-mer epitope was confirmed by ELISPOT IFN-γ and intracellular cytokine staining (ICS) on flow cytometry. A population of CD3+CD8+ T lymphocytes binding to HLA-A*0201/TMFPHIIVDV pentamers was identified in two HLA-A*0201-positive donors. Recognition of the 10-mer epitope by T cells resulted in a strong IFN-γ secretion as shown by ELISPOT assay. In addition, ICS confirmed that high proportion (31 and 59%) of the TMFPHIIVDV epitope-responsive CD3+CD8+ T cells in the two donors had multifunctional phenotype, simultaneously producing IFN-γ, IL-2 and TNF-α cytokines. In the present study novel human NoV HLA-A*0201-restricted minimal 10-mer epitope 139TMFPHIIVDV148 in the capsid VP1 was identified. The HLA-peptide pentamer staining of T cells from healthy donor PBMCs and cytokine responses in ex-vivo ELISPOT and ICS assays suggest that this epitope is recognized during NoV infection and activates memory phenotype of the epitope-specific multifunctional CD8+ T cells. The importance of this epitope in protection from NoV infection remains to be determined.


INTRODUCTION
Noroviruses (NoV) are group of very contagious viruses that cause >90% of non-bacterial and approximately half of all-cause epidemic gastroenteritis worldwide. NoV gastroenteritis is a significant public health problem with high clinical and economic costs, that is estimated to cause over 200,000 deaths annually, mainly in resource-limited countries (1). Despite scientific efforts and increasing awareness of NoV burden, there is no vaccine available yet (2,3). NoV exhibits a great genetic diversity with six currently recognized genogroups (GI-GVI), human NoVs belonging mainly to GI and GII, that include over 30 genotypes and numerous intra-genotype variants (4). GII.4 genotype has been predominant since the mid-90s, including pandemic variants US 1995/96, Farmington Hills 2002, Hunter 2004, New Orleans (NO) 2009, and most recently, Sydney 2012 (5). GII.4 variants share >95% of their VP1 capsid amino acid (aa) sequence, whereas <85% of the capsid aa are identical between different genotypes (6). The antigenic diversity of NoVs is likely to have major impact on evading host immune responses and complicates designing of NoV vaccines. NoV research is further limited due to the lack of small animal model and efficient NoV propagation in cell culture (7,8), but spontaneously formed NoV VP1 virus-like particles (VLPs) are successfully utilized as viral surrogates for assaying immune responses and as vaccine candidates (9,10). NoV infections cause acute but self-limited illness already at early age and several consecutive infections occur frequently (11)(12)(13). Protective immunity and clearance of NoV infections are not well-characterized (1,2). Even though there is some indication of pre-existing antibodies conferring protection (14) the level of serum NoV-specific antibodies is not directly linked to protection from infection (15). The best correlate of protection from NoV infection identified so far are blocking (neutralizing) antibodies, that are able to block the binding of NoV VLPs to the putative receptors, histo-blood group antigens (HBGA's) (15). NoV-specific antibodies are cross-reactive to some extent within genogroups however, blocking antibodies are mostly genotype-specific with low protective capacity against more distinct strains (1). Although NoV-specific antibodies are extensively studied very little is published on human T cell immunity to NoV and its' role in protection (16)(17)(18)(19). Studies using a murine norovirus (MNV)-mouse model have suggested T cell-mediated immunity to be important in clearance of MNV infection (20,21). To this end, cytotoxic CD8 + T cellmediated immune responses may provide substantial protection against serologically distinct viruses via recognition of crossreactive, conserved epitopes, such as observed in influenza virus infection (22).
Using matrix peptide approach we have recently identified a NoV-specific 18-mer peptide ( 134 SPSQVTMFPHIIVDVRQL 151 ) containing CD8 + T cell epitope likely restricted to human leukocyte antigen (HLA)-A * 0201 allele (17). Furthermore, peptide binding prediction tools have predicted 10 amino acid (aa) stretch as a minimal epitope ( 139 TMFPHIIVDV 148 ). In here, experiments were undertaken to experimentally confirm the recognition of the 10-mer epitope and its presentation by the HLA-A * 0201 allele.

Blood Donors and Cell Isolation
Four healthy adults (age 35-45, laboratory personnel) volunteering in a study recently conducted by our laboratory were selected based on their CD8 + T cell responses to NoV GII.4 VP1-specific 18-mer peptide 134 SPSQVTMFPHIIVDVRQL 151 (originally named 99-20) recently described (17). The two peptide responders were HLA typed as HLA-A * 02:01 carriers (17). The cells of the two donors identified as nonresponders to the 18-mer peptide were used as negative controls. Peripheral blood mononuclear cells (PBMCs) of the heparin blood sample were obtained by Ficoll-Pague PLUS (GE Healthcare, Little Chalfont, United Kingdom) density gradient centrifugation. PBMCs were frozen in 10% DMSO in fetal bovine serum (FBS) using freezing container (Mr. Frosty TM , Thermo Scientific,Waltham, MA, United States) with controlled rate of cooling at −80 • C and transferred to liquid nitrogen. Prior to analysis, PBMCs were thawed, washed, and resuspended in culture medium (CM) containing RPMI 1640 with Glutamax R and HEPES (Gibco TM by Thermo Fisher Scientific) supplemented with 10 µg/ml Gentamicin (Gibco TM ) and 10% fetal bovine serum (FBS, Sigma-Aldrich, St. Louis, MO, United States). The samples tested here are collected at a single bleed. Each sample aliquot was tested simultaneously by pentamer staining, enzyme-linked immunosorbent spot interferon-gamma (ELISPOT IFN-γ) assay and intracellular cytokine staining (ICS) assays, at least two times. Written informed consent was obtained from each volunteer prior to the sample collection in accordance with the Declaration of Helsinki. No approval by an ethics committee was required as per the local legislation.

Synthetic Peptides and Pentamers
To predict the optimal HLA-A * 02:01 allele binding epitope within the 18-mer NoV-specific peptide 134 SPSQVTMFPHIIVDVRQL 151 , artificial neural networks (ANN) (23) implemented at Immune Epitope Database and Analysis Resource (IEDB) was employed (24)(25)(26). A 10-mer high binding affinity sequence 139 TMFPHIIVDV 148 was identified (17) and the peptide was synthetized with purity >75% (Synpeptide Co. Ltd, Shanghai, China). SYFPEITHI prediction algorithm (http://www.syfpeithi.de) (27) was used for scoring the peptide affinity for HLA-A * 0201 allele, as a binding score of >21 ensures synthesis of a custom pentamer by Proimmune Ltd. (Oxford, United Kingdom). HLA-A * 0201/TMFPHIIVDV pentamer (Pro5 R MHC Class I Pentamer) labeled with Rphycoerythrin (R-PE) was synthetized by Proimmune Ltd. In addition, HLA-A * 0201 negative control pentamer conjugated to R-PE was synthetized as well. The degree of conservation of the 10-mer epitope among different NoV genotypes and genogroups was investigated using Basic Local Alignment Search Tool (BLAST) for sequence identity. Evolutionary analyses of the major capsid VP1 aa sequence of the aligned NoV genotypes were conducted in MEGA X (28). The evolutionary distances were computed using the Poisson correction method (29).

Pentamer Staining
PBMCs were treated for 10 min with Human BD Fc Block to prevent non-specific staining. To discriminate viable from non-viable cells, Horizon Fixable Viability stain 780 was used according to manufacturer's instructions. The cells (1 × 10 6 ) were incubated for 30 min on ice with either HLA-A * 0201/TMFPHIIVDV pentamer or negative control pentamer, using 0.25 or 0.5 µg of the pentamers per condition. Cells were further stained with monoclonal antibodies against human CD3 (clone UCHT1, fluorescein isothiocyanate (FITC) conjugate) and CD8a (clone RPA-T8, PerCP Cy5.5 conjugate) for 30 min on ice. All reagents used for staining were purchased from BD Pharmingen (San Jose, CA, United States). After washing the cells were resuspended in 1% FBS, 2.5% formaldehyde in PBS for flow cytometry acquisition. At least 400 000 events were acquired for analysis on a 2-laser FACS CantoII flow cytometer (BD) with FACSDiva Software V 6.1.3 (Becton Dickinson, Heidelberg, Germany). Data were analyzed by FlowJo software version 10.1 (Tree Star, San Carlos, CA, United States). Lymphocytes were gated according to forward/sideward scatter (FSC/SSC) and dead cells were excluded by gating on the population negative for the viability dye. CD3 + CD8 + population was further plotted on SSC-A vs. pentamer population.

Intracellular Cytokine Staining (ICS)
An ICS assay was employed to quantify IFN-γ, TNF-α, and IL-2 producing CD3 + CD8 + T cells. PBMCs were stimulated according to the previously published protocol (17) with 4 µg/ml of 10-mer peptide 139 TMFPHIIVDV 148 , 4 µg/ml irrelevant 9-mer peptide (negative control) or 1 µg/ml Staphylococcal enterotoxin B (SEB, Sigma) in the presence of 1 µg/ml CD28 and 1 µg/ml CD49d costimulatory antibodies (BD Biosciences, San Jose, CA, United States). PBMCs incubated in CM supplemented with the costimulatory antibodies only were used as additional control. The protein transport inhibitor brefeldin A (GolgiPlug, BD Biosciences, San Jose, CA, United States) at a concentration of 10 µg/ml was added after 2 h and the incubation was continued for 16 h at 37 • C. After stimulation, the cells were treated with EDTA for 15 min and washed with FACS Stain buffer. Prior to fixation and permeabilization for ICS, PBMCs were blocked for non-specific staining, stained for viable/non-viable discrimination and surface markers CD3 and CD8a as described above for the pentamer staining. BD Fixation/Permeabilization solution was used according to the manufacturer's instructions and cells were intracellularly stained with the mixture of IFN-γ (clone 4S.B3) PE-Cy7 conjugate, IL-2 (clone MQ1-17H12) PEconjugate, and TNF-α (clone MAb11) allophycocyanin (APC)conjugate in 50 µl Perm/Wash buffer for 30 min on ice in the dark. Cells were resuspended in FACS Staining Buffer for acquisition and analysis using FACS CantoII flow cytometer and FACSDiva Software V 6.1.3. All reagents used for ICS were purchased from BD Pharmingen (San Jose, CA, United States). The data analysis was performed using FlowJo software version 10.1.

The 10-mer Minimal Epitope Prediction and Conservation
At the outset of our study, an 18-mer NoV VP1-specific CD8 + T cell epitope (99-20, 134 SPSQVTMFPHIIVDVRQL 151 ) had been identified in two subjects, both having HLA-A * 0201 allele. When analyzed by IEDB database ANN method, a 10 aa sequence 139 TMFPHIIVDV 148 within the 18-mer showed the highest binding affinity of IC 50 21.5 nM ( Table 1), whereas all other predicted sequence lengths indicated affinity >80 nM (data not shown). Peptides with IC 50 < 50 nM are considered high affinity and <500 nM intermediate affinity (30). Congruently, the same 10-mer epitope was identified to have the highest binding affinity by SMM method [data not shown (23)]. SYFPEITHI prediction algorithm confirmed a score of 27 for the peptide affinity for HLA-A * 0201 allele ( Table 1). When comparing the aa sequence of different NoV genotypes ( Table 1) the 10-mer epitope was found to be highly conserved among different GII.4 variants and also more distant NoV genotypes belonging to GII and GI. The phylogenetic distances of the NoV genotypes compared in the Table 1 are shown in Figure 1. In addition, the 10-mer epitope sequence in all genotypes belonging to GI and GII NoV listed in the Table 1 also had good affinity binding scores for HLA-A * 0201 allele regardless of up to three aa substitutions compared to the GII.4-1999 10-mer sequence. On the contrary, the 10-mer sequence in the human GIV 2010 NoV had a unique substitution 140 M to 140 Q which seemed to abolish its binding affinity resulting in IC 50 552.6 nM and syfpeithi score 17 ( Table 1). In addition, a change of the aa 148 V to 148 D at the c-terminus of the empirical 10-mer peptide abolished the predicted binding completely. This is not surprising as the aa at positions 140 and 148 are reported anchoring positions for HLA-A * 0201 binding (31)(32)(33). In addition, deletions at the each terminus of the peptide affected the predicted binding capability of the peptide.

DISCUSSION
Despite significant efforts to define correlates of protection to NoV infection (1,2), the role of T lymphocytes in protection and clearance of NoV infection is still largely unexplored area, with only few papers published so far (16)(17)(18)(19)34). We have recently tested ten healthy volunteer donors for NoV GII.4 capsid VP1-specific T cell responses using matrix peptide pools and found two CD8 + T cell responders to 18-mer NoVspecific peptide ( 134 SPSQVTMFPHIIVDVRQL 151 , 99-20) (17). In here, we defined a 10-mer 139 TMFPHIIVDV 148 minimal epitope within this sequence and its restriction to HLA-A * 0201 allele. While the 10-mer 139 TMFPHIIVDV 148 epitope is derived from GII. 4-1999 strain (17), the alignment with several other GII.4 variants showed epitope sequence to remain invariable until 2012 (Table 1). Interestingly, alignment with other quite distant genotypes (Figure 1) even belonging to GI viruses, indicated high degree of conservation of the epitope. Importantly, up to three aa substitutions as in the GI.1 West Chester 2001 NoV do not necessarily reduce the binding affinity to HLA-A * 0201, depending on the aa position and characteristics. Aa at the position 6 and 8 of the 10-mer showed most variability among the aligned genotypes, however, these changes did not lead to considerable change in binding affinity predictions. However, a change at aa position 2 from a hydrophobic (M) to a polar aa (Q) in a human GIV 2010 NoV and at aa position 9 from V to D in an empirical sequence seemed to drastically abolish binding affinity. This is not surprising as HLA-A * 0201 binding epitopes have a restricted size of 9/10 aa and hydrophobic anchor residues at positions 2 and 9/10 likewise the epitope identified in this study. The NoV 10-mer epitope contains methionine (M) at position 2 and valine (V) at position 10 which is described as a typical C-terminal anchor residue (31)(32)(33).
Direct staining of PBMCs with HLA-A * 0201 pentamer loaded with 139 TMFPHIIVDV 148 peptide attested relatively high frequency of circulating NoV capsid-specific memory CD8 + T cells in two HLA-A * 0201 positive healthy donors (0.51 and 1.52% of CD3 + CD8 + gated lymphocytes, respectively). HLA-A2 is the most prevalent MHC allele family in human population and the gene frequency of HLA-A * 02:01 is especially high in Caucasians (>20%), and other western ethnic groups (African Americans 12%, Hispanics 23%, North American natives 21%) (35)(36)(37), further emphasizing the importance of the identified epitope. However, it has been reported that only a small portion of the peptides with predicted and experimentally confirmed high affinity binding, are actually capable to induce T-cell responses (38). Therefore, the utilization of functional assays is essential for determining the immunogenicity of an epitope. To that end, the functionality of T cell immune responses to 10-mer epitope were determined by two functional assays, ELISPOT IFN-γ and ICS. The results of ELISPOT IFN-γ assay showed robust NoV VP1 10-mer epitope-specific CD8 + T cell immune responses in HLA-A * 0201-positive individuals, whereas control donors were negative. Titration of 10-mer peptide down to 0.05 µg/ml revealed high avidity of CD8 + T cells to 10-mer peptide. These results were expanded using ICS assay to detect CD8 + T cells secreting cytokines IFN-γ, IL-2, and TNF-α. By this approach, the presence of circulating monofunctional and polyfunctional memory CD8 + T lymphocytes specific for 10mer NoV epitope was detected in HLA-A * 0201-positive healthy donors. Polyfunctional CD8 + T cells that produce more than one immune mediator, are associated with protection from viral infections, such as HIV-1 and human herpes virus (39)(40)(41). In the future we aim to utilize the pentamer technology to validate CD8 + T cell responses in a larger number of serologically positive NoV-infected HLA-A * 02:01 positive subjects to strengthen the clinical significance of the finding in this study.
Pentamer staining, ELISPOT IFN-γ and ICS assays, performed simultaneously using the same sample of each donor, resulted in highly congruent data supporting high immunogenicity of the novel NoV 139 TMFPHIIVDV 148 CD8 + T cell epitope. Many viral infections are not contained by antibody responses alone and it is likely that T cell responses play a role in clearance of NoV infection and may have a role in protective immunity as well. Virus-specific CTL responses play a crucial role in a clearance of many other viral infections such as human immunodeficiency virus-1 (HIV-1) (40,42,43), influenza (22,44) and human papilloma virus (45). Patients with chronic hepatitis B (HBV) infection typically lack effective HBV-specific T cells, whereas fully recovered patients display strong CD8 + T-cell responses (46).
To conclude, the present work describes the first human NoV CD8 + T cell epitope 139 TMFPHIIVDV 148 restricted to HLA-A * 02:01 allele. Regarding the high conservation of the identified epitope and high frequency of HLA-A * 02:01 allele in human population, it can be speculated that ∼20% of the individuals exposed to divergent NoV genotypes, will develop strong CD8 + T cell responses to this 10-mer epitope. While NoV blocking antibodies are largely genotype specific, a role of T cells targeted to broadly conserved epitopes, as described in here, may be of large significance. However, the role and importance of this epitope as well as overall T cell responses in protection from NoV infection needs to be further investigated.

AUTHOR CONTRIBUTIONS
MM sample acquisition, processing, and laboratory analysis (Pentamer staining, ELISPOT, and ICS by flow cytometry). Data acquisition, analysis, and interpretation, and writing the manuscript. TV the Head of Vaccine Research Center, revision of the manuscript text. VB the Head of the Laboratory, conception and designing the study, data interpretation, drafting, and writing the paper, critical revision of article for important intellectual content.

FUNDING
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.