Novel mumps virus epitopes reveal robust cytotoxic T cell responses after natural infection but not after vaccination

Mumps is nowadays re-emerging despite vaccination. The contribution of T cell immunity to protection against mumps has not been clearly defined. Previously, we described a set of 41 peptides that were eluted from human leukocyte antigen (HLA) class I molecules of mumps virus (MuV)-infected cells. Here, we confirmed immunogenicity of five novel HLA-B*07:02- and HLA-A*01:01-restricted MuV T cell epitopes from this set of peptides. High frequencies of T cells against these five MuV epitopes could be detected ex vivo in all tested mumps patients. Moreover, these epitope-specific T cells derived from mumps patients displayed strong cytotoxic activity. In contrast, only marginal T cell responses against these novel MuV epitopes could be detected in recently vaccinated persons, corroborating earlier findings. Identifying which MuV epitopes are dominantly targeted in the mumps-specific CD8+ T- response is an important step towards better understanding in the discrepancies between natural infection or vaccination-induced cell-mediated immune protection.


Results
In the present study, peptides from a previously characterized set of 41 naturally processed MuV epitopes 16 , predicted to bind best to HLA-B*07:02 (n = 12) and HLA-A*01:01 molecules (n = 10) were selected for immunogenicity assessment. Selected epitopes were tested for their capacity to recall a specific T cell response, by using expanded CD8 + T-effector cells isolated from three HLA-B*07:02 positive and three HLA-A*01:01 positive mumps patients (Fig. 1).
Using p:HLA dextramers, high frequencies of MuV epitope-specific T cells were observed in all three B*07:02and three HLA-*01:01-positive patients, even exceeding 2% of the total CD8 + T cell pool ( Fig. 2A). The frequencies of the epitope-specific CD8 + T cells and the dominance of the response per epitope varied between patients. Further analyses showed that the majority of peptide-specific T cells were within the T-effector cell subset (CD3 + CD8 + CD45RO -CCR7 − ). Functional characterization revealed that, upon stimulation with peptides, the epitope-specific T cells expressed the activation marker CD137 and displayed a cytotoxic phenotype by expression of CD107a. In addition to pro-inflammatory cytokines IFN-γ, TNF and IL-17A, epitope-specific T cells produced granzyme-A/B, granulysin, and to a lesser extent sFas(L), and perforin that are involved in cell-mediated cytotoxicity (Fig. 2B). Finally, we showed that epitope-specific T cells were able to efficiently lyse peptide-pulsed antigen-presenting target cells in a flow cytometry-based cytotoxicity assay (Fig. 2C).
In contrast to the high frequencies of MuV epitope-specific T cells observed in mumps patients, only very low frequencies of MuV epitope-specific T cells could be detected in young adults that were recently vaccinated best rank ---> lower rank %rank predicted binding to HLA-B*07:02 tested positive in T cell assay Figure 1. Predicted binding affinity of the candidate epitopes to HLA and capability to induce cytokine production by T cells Mumps virus (MuV) candidate epitopes were selected based on their predicted binding (threshold rank < 2%) to HLA-B*07:02 molecule (left panel; n = 12 peptides) and HLA-A*01:01 molecule (right panel; n = 10 peptides). Selected peptides were used to stimulate peripheral blood mononuclear cells (PBMC) of respectively three HLA-B*0702 positive and three HLA-A*01:01 positive mumps patients. CD8 + T effector cell lines from these patients were stimulated with the selected peptides and production of interferon (IFN)-γ and tumor necrosis factor (TNF) by CD8 + T cells was measured by flow cytometry. Peptide-reactive T cells were identified as a twofold increase of IFNγ + /TNF + cells within the CD3 + /CD4 -/CD8 + live gate in comparison with medium control. Peptides that were tested positive induced a response in all three HLA-B*07:02 and HLA-A*01:01 positive mumps patients. Filled bars (gray) indicate the peptides that tested positive. www.nature.com/scientificreports/ with a third dose of the MMR vaccine (Fig. 3). Despite the fact that epitope-specificity varied between individuals, in all tested PBMC samples of HLA-B7-and HLA-A1-positive vaccinated persons, small amounts of MuV epitope-specific CD8 + T cells could be visualized with the p:HLA dextramers.

Discussion
Identification of more functional CD8 + T cell epitopes that cover a broad repertoire of T cells against mumps virus (MuV) is important for monitoring and functionally characterizing the T cell response following mumps infection or vaccination. Based on HLA-I binding prediction, most of the set of 41 MuV peptides, from an earlier identified panel 16 , bound well to the HLA-B*07:02 molecule, followed by the HLA-A*02:01 and HLA-A*01:01 molecules, i.e. belonging to the most common HLA-I alleles 17 . In addition to the previously identified HLA-A*02:01 epitopes, we now confirmed three HLA-B*07:02-restricted T cell epitopes (i.e. APIQGTNLL (V/P/I 27-35 ), KPRTSTPVTEF (V/P/I 142-152 ), IPNARANL (NP 115-122 )), each inducing a strong T cell response in expanded effector T cells obtained from three different HLA-B*07:02-positive mumps patients. Interestingly, the first described MuV-specific CD4 + T-helper cell epitope, GTYRLIPNARANLTA (NP 110-124 ), which we recently discovered 18 , contains the newly identified HLA-B*07:02-restricted CD8 + T cell epitope IPNARANL (NP 115-122 ). This opens the possibility of using synthetic peptides, containing both epitope sequences, to stimulate CD4 + and CD8 + T cell responses simultaneously. In addition to the HLA-B*07:02-restricted epitopes, HLA-A*01:01restricted T cell epitopes (i.e. YSDPNNHEVY (L 1336-1343 ) and VTDSNLIY (L 1983-1992 )) showed to induce a T cell   19 . Thus, in general, the identified immunogenic MuV peptides were predicted to bind more strongly to HLA-I than the non-immunogenic peptides did (Fig. 1), although not all strong HLA-binding peptides appeared to be immunogenic. However, it should be taken into account that expanded CD8 + T-effector cell lines were used for immunogenicity testing of the eluted epitope candidates. It may therefore be that, due to relatively higher frequencies of T cells reactive against the immunodominant T cell epitopes, subdominant epitopes against which only few T cells were reactive could not be not identified; T cells recognizing subdominant epitopes may be overgrown by the more dominant T cell clones represented in higher frequencies.
Four out of the five novel MuV epitopes that were all derived from the genotype G outbreak strain (MuVi/ Utrecht.NLD/40.10) appeared to be conserved among the two Jeryl Lynn vaccine strains (JL2 and JL5), and therefore are useful for monitoring cellular immune responses after natural infection with the genotype G outbreak strain and after vaccination. Interestingly, one of the HLA-B*07:02-restricted epitopes, i.e. APIQGTNLL (V/P/ I 27-35 ) of the genotype G outbreak strain (MuVi/Utrecht.NLD/40.10) differed from the peptide sequences in the two Jeryl Lynn vaccine strains (JL2 (APIQGTNSL) and JL5 (TPIQGTNSL)) and also showed to have sequence differences among various circulating mumps strains, including genotype G, H and C strains. Future experiments are planned to explore the impact of this mismatch on the T cell recognition of the various MuV strains, i.e. vaccine versus circulating MuV strains. Using peptide-HLA class I (p:HLA-I) dextramers, high frequencies of T cells against all five MuV epitopes were ex vivo detected in PBMCs of all three HLA-B*07:02-and three HLA-A*01:01-positive mumps patients. Individual differences in epitope dominance of the T cell responses were observed, where one patient had a stronger T cell response to (a) particular epitope(s), another patient showed to have a stronger T cell response to (an)other epitope(s). Functional characterization of these specific T cells revealed a cytotoxic phenotype with expression of degranulation marker CD107a, and secretion of granzyme-A/B, granulysin, perforin, sFas(L). With a flow cytometry-based cytotoxicity assay, epitope-specific T cells showed indeed to be able to efficiently lyse MuV peptide-pulsed target cells. Thus, we for the first time describe that MuV epitope-specific CD8 + effector T cells isolated from mumps patients have a cytotoxic phenotype which is a critical feature for mediating viral clearance following acute MuV infection. Although T cells against the novel MuV epitopes were also detectable in recently vaccinated persons, only very low frequencies of these specific CD8 + T cells were observed, which is in line with our previous findings 9 . Following respiratory virus infection, a robust CD8 + T cell response can be induced to mediate viral clearance 20 . The question is whether an attenuated virus vaccine strain with lower infection efficiency will be able to induce a comparable strong CD8 + T cell response. Based on our previous findings, where the mumps-specific CD4/CD8 T cell ratio was higher after vaccination than after natural infection, MuV vaccination seems to induce relatively more CD4 + than CD8 + T cells 9 . Both CD4 + and CD8 + T cells are involved in protection against viral disease. Prevention of infection can be achieved by induction of CD4 + T cells providing help to B cells to produce protective virus-neutralizing antibodies, whereas vaccine-induced CD8 + T cells support disease attenuation and protection against complications after infection 21 .
In order to be able to accurately measure the low CD8 + T cell response after vaccination, we propose that not a single peptide but rather a pool of confirmed immunogenic peptides representing a broad HLA-I coverage should be used. Such a peptide pool would provide a useful tool to monitor and functionally characterize the CD8 + T cell response to MuV to better understand its role after natural infection and vaccination. In addition, MuV epitopes with sequence variations between vaccine and circulating MuV strains can be used to explore the impact of this mismatch on immunity of vaccinated persons to antigenic different circulating MuV strains.

Methods
Subjects and blood sample handling. Peripheral blood samples of mumps patients (19-40 years) used for this study were collected in 2012, during a mumps epidemic (genotype G) that spread across multiple locations within the Netherlands. Sampling of the mumps patients took place 1-2 months after disease onset as part of the previously described clinical study VAC-263 (NL37852.094) 22 . In addition, peripheral blood samples were used that were collected from young adults (18-30 years) four weeks after they received a third dose of the MMR vaccine as part of our MMR3 clinical study 6 or from our blood donor system 23 . Written informed consent was obtained for all participants. The clinical studies were approved by the medical ethical committee (METC Noord-Holland) and performed in accordance with the Declaration of Helsinki. HLA class I typing (B*07:02 and HLA-A*01:01) was performed by high resolution next-generation sequencing (University Medical Center Utrecht, the Netherlands) or by flow cytometry using anti-HLA class I A1-biotin antibody (8.L.104, , AB31563, Abcam) with streptavidin PE-CF594 (562,318, BD) and anti-HLA class I B7-PE antibody (BB7.1, 372,403, Biolegend) utilizing LSRFortessaX20 (BD) and analyzed with FlowJo (V10/Treestar). Peripheral blood mononuclear cells (PBMCs) were isolated from the blood samples by density gradient centrifugation according to manufacturer's instruction ( (Lymphoprep, 1,114,547, Progen) or Cell Preparation Tubes (CPT, BD))and cryopreserved at − 135 °C before thawing for use. Epstein-Barr virus-transformed B-lymphoblastoid cell lines (BLCL) as antigen-presenting cells, were generated from PBMCs as described before 23 .
Expansion of MuV epitope-specific CD8 + T cells from mumps patients. Epitope-specific CD8 + T cell lines were generated by sorting dextramer-positive cells (FACSMelody sorter (BD)) from HLA-B*0702and HLA-A*0101-positive mumps patients and subsequently further expanded as described above for effector CD8 + T cell lines.
Testing for cytotoxicity capacity. A flow cytometry-based killing assay was used to determine the cytotoxic capacity of the various epitope-specific T cell lines 18 . For this purpose, BLCL as antigen-presenting cells were labeled with 0.1 µM CellTrace violet (C34557, Thermo Fisher Scientific) and pulsed with one of the five selected peptides (5 µM) or medium or non-immunogenic MuV peptide as negative control. Subsequently, epitopespecific T cell lines and BLCL were co-cultured at different effector/target (E:T) ratios in AIM-V supplemented with 2% human AB serum. After 4 h, cells were stained with Fixable Viability Stain 780 (BD) and fixated. Cells were acquired using a LSRFortessa X20. Measurements were made in triplicate and averaged.