T cell responses to bluetongue virus are directed against multiple and identical CD4 + and CD8 + T cell epitopes from the VP7 core protein in mouse and sheep

Page 1

TC

JC

a

ARRAA

KBT

1

vowtcgatapsRcpia

0d

Vaccine 29 (2011) 6848– 6857

Contents lists available at ScienceDirect

Vaccine

j ourna l ho me pag e: www.elsev ier .com/ locate /vacc ine

cell responses to bluetongue virus are directed against multiple and identicalD4+ and CD8+ T cell epitopes from the VP7 core protein in mouse and sheep

osé-Manuel Rojas, Teresa Rodríguez-Calvo, Lourdes Pena, Noemí Sevilla ∗

entro de Investigación en Sanidad Animal (CISA-INIA), Instituto Nacional de Investigación Agraria y Alimentaria, Valdeolmos, Madrid, Spain

r t i c l e i n f o

rticle history:eceived 19 May 2011eceived in revised form 15 July 2011ccepted 16 July 2011vailable online 30 July 2011

eywords:TV

cell epitopes

a b s t r a c t

Bluetongue virus (BTV), an economically important orbivirus of the Reoviridae family, is a non-enveloped,dsRNA virus that causes a haemorrhagic disease mainly in sheep, but little is known of the cellular immu-nity elicited against BTV. We observed that vaccination of interferon type I-deficient mice (IFNAR(−/−)),or inoculation of the wild type C57BL/6 strain with BTV-8, induced a strong T cell response. Therefore,we proceeded to identify some of the T cell epitopes targeted by the immune system. We selected, usingH-2b-binding predictive algorithms, 3 major histocompatibility complex (MHC)-class II-binding peptidesand 7 MHC-class I binding peptides from the BTV-8 core protein VP7, as potential T cell epitopes. Peptidebinding assays confirmed that all 7 MHC-class I predicted peptides bound MHC-class I molecules. ThreeMHC-class I and 2 MHC-class II binding peptide consistently elicited peptide-specific IFN-� production(as measured by ELISPOT assays) in splenocytes from C57BL/6 BTV-8-inoculated mice and IFNAR(−/−)-vaccinated mice. The functionality of these T cells was confirmed by proliferation and cytotoxicity assays.

+

Flow cytometry analysis demonstrated that CD8 T cells responded to MHC-class I binding peptides andCD4+ T cells to MHC-class II binding peptides. Importantly, these 5 epitopes were also able to inducedIFN-� production in sheep inoculated with BTV-8. Taken together, these data demonstrate the activationof BTV-specific T cells during infection and vaccination. The characterisation of these novel T cell epitopesmay also provide an opportunity to develop DIVA-compliant vaccination approach to BTV encompassinga broad-spectrum of serotypes.

. Introduction

Bluetongue virus (BTV) is a non-contagious, insect-transmittediral pathogen of domestic and wild ruminants, considered onef the most economically important arbovirus [1]. The virus, forhich 24 serotypes have been described so far, is transmitted to

he vertebrate host by some species of biting midges from the Culi-oides genus [2]. BTV is the prototype member of the Orbivirusenus of the Reoviridae family. This double-stranded RNA virus has

10 segments genome that is enclosed by a complex capsid struc-ure consisting in a nucleocapsid containing 5 proteins (major, VP3nd VP7; minor, VP1, VP4 and VP6). VP3 and VP7 are conservedroteins of hydrophobic nature, playing an important role in thetructure integrity of the virus core [3]. VP1, VP4 and VP6 haveNA transcriptase- and RNA-modifying properties [4]. The outerapsid layer structure is composed of 2 additional major structural

roteins (VP2 and VP5) [5–7]. VP2 is responsible for receptor bind-

ng, hemagglutination and eliciting serotype-specific neutralizingntibodies [8–10]. VP5 forms trimers in globular motifs of the outer

∗ Corresponding author. Tel.: +34 91620 2300; fax: +34 91 6202247.E-mail address: [email protected] (N. Sevilla).

264-410X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2011.07.061

© 2011 Elsevier Ltd. All rights reserved.

layer of the BTV particle [11]. In addition to the structural proteins,4 non-structural (NS) proteins NS1, NS2, NS3 and NS3A participatein the control of BTV replication, maturation and export from theinfected cell [12,13].

BTV is capable of infecting cells of the immune system likemonocytes [14], dendritic cells [15] and �� T cells [16] and theseare likely routes for dissemination of the virus within the host (andto the vector). Animals which recover from the disease develop along-lasting immunity to the virus. Both neutralizing antibodies[17] and cytotoxic T lymphocytes (CTL) play a role in protectiveimmunity to BTV [18], although cellular immunity is likely to becrucial as BTV protection can exist in the absence of neutraliz-ing antibodies [19,20]. Moreover, adoptive transfer of lymphocytescould at least partially protect monozygotic sheep from subsequentBTV challenge [18]. Interestingly, BTV infection and vaccinationinduces CTLs in sheep capable of cross-reaction with differentBTV serotypes [16,21–23]. Therefore vaccination designed to elicitCTL responses can potentially protect animals against a broad-spectrum of BTV serotypes.

Immunodominant serotype cross-reactive T-cell determinantshave been located within the structural proteins of BTV-cores [24].VP7 is a major BTV group reactive antigen [25] and sheep vaccinatedwith a capripox virus encoding VP7 showed clinical protection

Page 2

cine 2

a[pooTi

2

2

msppw

2

SgoC8so(as(tCc

2

saaf(kw

2

mRhRicasTsta

J.-M. Rojas et al. / Vac

gainst heterotypic challenge, although the virus still replicated26]. These data led us to define T cell epitopes in VP7. Thus, in theresent report we identify novel CD4 and CD8 T cell epitopes rec-gnized in mouse model from the VP7 core protein of BTV-8. Somef these epitopes are also recognized by T cells form infected sheep.his work underlines the necessity to further characterise anti-BTVmmunity in order to develop more effective vaccinations.

. Materials and methods

.1. Virus

BTV stocks were prepared from infection of MC57 cells withultiplicity of infection (MOI) of 1. Virus titre was determined by

tandard plaque assays in Vero cell line. The virus (1 × 106 PFUrior to inactivation) was inactivated by incubation with freshlyrepared 3 mM BEI for 24 h at 37 ◦C, and the reaction was stoppedith 0.02 M sodium thiosulphate.

.2. Mice, sheep and inoculations

C57BL/6 mice were purchased from Harlan Interfauna IbéricaL and IFN-�/�R0/0 (IFNAR(−/−)) mice on a C57BL/6 genetic back-round were kindly provided by Professor R. Zinkernagel (Institutef Experimental Medicine, Zurich) and bred in our animal facility.57BL/6 mice (8–12 week old) were infected with 100 PFU BTV-

subcutaneously (s.c.) three times at 10–11 day intervals beforeacrifice 3 days after the last infection. IFNAR(−/−) mice (8–16 weekld) were inoculated s.c. twice with 100 �l of BEI-inactivated BTV-8equivalent to 1 × 106 PFU prior to inactivation) at a week intervalnd challenged s.c. with 10 PFU BTV-8 a week later. IFNAR(−/−) wereacrificed at day 6 post-challenge. Three month old female sheepMallorquina breed) were inoculated with 1 × 105 PFU BTV-8. Allhe procedures herein described were carried out under Europeanommunity guidelines and approved by the local ethical reviewommittee.

.3. Peptides

Peptides sequences from the VP7 core protein of BTV-8 (acces-ion number: ACJ06230) were selected according to predictivelgorithms and motifs for binding to H-2 Db/Kb/Ab moleculesvailable on the web [27–30]. VP7 peptides were purchasedrom Altabiosciences (Birmingham, UK). The gp(33–41) peptideKAVYNFATC) from lymphocytic choriomeningitis virus (LCMV)nown to bind Db and Kb molecules was used as irrelevant peptidehere mentioned.

.4. Peptide binding assays

Peptide binding assay were performed following a modifiedethods from [31]. The transporter-associated protein-deficient

MA-S cell line was kindly provided by Dr McArdle (The Notting-am Trent University, UK). These cells were culture in completePMI + 10% FCS. For binding assays, 2–5 × 105 cells were incubated

n serum-free RPMI for 1 h at 37 ◦C in the presence of varyingoncentration of peptides. Cells were stained with anti-Db-FITCnd anti-Kb antibodies or the appropriate isotype controls (eBio-ciences). Cells were analysed on a BD FACSCalibur flow cytometer.

he binding was calculated as the ratio of mean fluorescence inten-ity (MFI) for Db or Kb molecules detected in the presence of peptideo the MFI detected in the absence of peptide. Data are presenteds mean of 3 independent experiments.

9 (2011) 6848– 6857 6849

2.5. Splenocyte cultures and IFN-� ELISPOT assays

Splenocytes from infected C57BL/6 or vaccinated IFNAR(−/−)

were obtained by mechanical disruption and cultured inT cell media (RPMI + 10%FCS + 4 mM l-glutamine + 10 mMHEPES + 1% 100X non-essential amino-acids + 1 mM sodiumpyruvate + 100 U/ml penicillin/100 �g/ml streptomycin + 50 nM(-mercaptoethanol). IFN-� ELISPOT assays were performed accord-ing to the manufacturer protocol (Diaclone, France). As negativecontrol, cells were cultured either without stimuli, with DMSO(equivalent volume to that added with peptide), or with a lysateof the MC57 cell line (used in the virus preparation). All cultureswere performed in triplicates and results are presented as averagenumber of spots for each mouse. Assays were considered validonly when IFN-� spot counts in control wells were below 20 for2 × 105 cells, and standard deviations in positive wells below 10%of the average. A positive control of splenocytes activated with0.5 �g/ml Concanavalin-A was always included to validate theELISPOT assay.

2.6. Intracellular cytokine staining and flow cytometry

1 × 106 splenocytes were stimulated either with 20 �g/ml ofpeptide, BTV-8, BEI-BTV or the appropriate negative controls for5 h (at 37 ◦C 5%CO2) in U-bottom 96-well plates, in presence of1 �g/ml brefeldin-A. Cells were harvested, washed twice withPBS + 1% FCS + 0.02% sodium azide and stained with anti-mouseCD4-FITC and anti-mouse CD8-PerCP antibodies (BD pharmingen).After washing cells were fixed and permeabilised with PBS contain-ing 4% paraformaldehyde and 0.1% saponin. Cells were stained withanti-mouse IFN-�-PE (BD pharmingen). The analyses was done ona BD FACSCalibur flow cytometer.

2.7. Proliferation assays

Splenocytes (2 × 105 per well) were plated in U-bottom 96-wellplates in triplicate in presence of stimuli (either 20 �g/ml pep-tide, BEI-BTV or the appropriate negative controls). 3H Thymidine(Hartmann Analytic, Germany) was added at a final concentrationof 5 �Ci/ml, and cells were cultured for 72–96 h at 37 ◦C 5%CO2.Using a cell harvester, cells were transferred onto UniFilter-96microplates (Perkin-Elmer) and counted on a 1450 MicroBeta Triluxcounter (Perkin-Elmer).

2.8. Splenocyte restimulation, LPS blast generation and CTLassays

Splenocytes (4–5 × 106 per well) were cultured for 5–6 days in24-well plates in presence of 10 �g/ml of stimulating peptide inT cell media. These cells were used as effector cells in CTL assays.Depending on the experiment, RMA-S cells or syngeneic LPS blastwere used as target cells. LPS blasts were generated from syngeneicsplenocytes cultured at a density of 1.5 × 106/ml with 25 �g/mlLPS (Sigma) and 7 �g/ml Dextran Sulphate (Sigma) for 48 h [32].Target cells (1 × 107) were harvested and labelled with an isotonicsolution of Na2

51CrO4 (50 �Ci) (Hartmann Analytic, Germany) for1 h at 37 ◦C in serum-free RPMI. Target cells were then pulsed for1 h at 37 ◦C, 5%CO2 with 20 �g/ml VP7 peptide, irrelevant peptidegp(33–41) or BEI-BTV (equivalent to 1 × 105 PFU). Effector cells

(E) and target cells (T) were plated in triplicates in U-bottom 96well plates at different ratios (E:T) and incubated for 5 h at 37 ◦C,5%CO2. Maximum release was measured using the supernatantsof target cells lysed with 0.1%SDS and spontaneous release by the

Page 3

6 cine 2

sl

P

2I

ibPbctwapt

2

dwEtP

3

3i

1sisimfIiSetccwCwmCtIvc

3H

ir

850 J.-M. Rojas et al. / Vac

upernatants of target cells cultured alone. The specific percentageysis was calculated using the following formulae:

ercentage cytotoxicity = (Experimental release − Spontaneous release)(Maximum release − Spontaneous release)

× 100

.9. PBL isolation from sheep, peptide stimulation and ovineFN-� measurement by ELISA

Peripheral blood leukocytes (PBL) were obtained 28 days post-nfection by standard gradient centrifugation methods from thelood of 8 BTV-8 infected and 2 control sheeps (inoculated withBS). PBL (2–5 × 105 per well) were cultured in duplicates in U-ottom 96-well plates with either VP7 peptides, DMSO as negativeontrol, VP7(181) peptide as peptide control, and BEI-BTV as posi-ive control for T cell responses to the virus. After 96 h, supernatantsere collected and the presence of ovine IFN-� was assessed using

commercially available ELISA kit (Mabtech, Sweden). Cytokineroduction was normalised to 1 × 106 PBL. The detection limit ofhe assay was 20 pg/ml IFN-�.

.10. Statistical analysis

The statistical analysis was done using unpaired two-tailed Stu-ent’s t-test. A two-tailed non-parametric Mann–Whitney U testas used to compare average IFN-� production by ELISPOT and

LISA between control group and VP7 peptide groups in all miceested in this study. Data handling analyses was performed usingrism 5.0 (GraphPad Software Inc. San Diego, CA, USA).

. Results

.1. Infection of C57BL/6 mice and vaccination of IFNAR(−/−) micenduces T cell responses

As previously described [33], C57BL/6 mice infected with0–100 PFUs of BTV-8 did not succumb and showed little to noign of infection. By contrast, IFNAR(−/−) mice died at day 6–7 post-nfection with BTV-8 and only vaccinated mice with inactivated BTVurvived the viral challenge. Since T cells have been reported to bemportant in anti-BTV immunity [18], we studied their response in

ice with a range of immunological assays. By ELISPOT, splenocytesrom infected C57BL/6 and vaccinated IFNAR(−/−) mice producedFN-� specifically when BTV-8 was present in the culture (eithernactivated BTV-8 (BEI-BTV) or infectious BTV-8) (Fig. 1A and B).imilarly, splenocytes from both strains were capable of prolif-rating to the inactivated virus in culture (Fig. 1C and D). Thus,hese data showed that T cells from infected C57BL/6 mice or vac-inated IFNAR(−/−) mice were responding specifically to BTV. Toharacterise the T cell population responding specifically to BTV,e used flow cytometry analysis for intracellular IFN-� staining. In57BL/6 mice, the average production of IFN-� for C57BL/6 miceas 1.23 ± 0.6% of CD4+ and 3 ± 1.5% of CD8+ T cells. In IFNAR(−/−)

ice, both T cell subsets were also producing IFN-� (CD4+, 1 ± 0.4;D8+, 2.6 ± 1) (Fig. 1E and F). In summary, these data indicatehat following infection and vaccination with BTV of C57BL/6 andFNAR(−/−) mice, anti-BTV CD4+ T cells and CD8+ T cells are acti-ated and, thereby represent suitable models to study anti-BTV Tell immunity.

.2. Peptide prediction from the VP7 core protein and binding to-2 Db and H-2 Kb molecules

In order to further characterise the T cell response elicited dur-ng BTV protection, we searched T cell epitopes involved in thisesponse. We looked for epitopes in the VP7 core protein from

9 (2011) 6848– 6857

BTV as it is well-preserved among serotypes and therefore likelyto cross-react with several BTV serotypes. Using a combination ofpredictive MHC peptide-binding algorithms available on the web(SYFPEITHI, Pro-Pred-I, NetMHC I and II) [27–30], 10 peptides fromVP7 were selected on the basis of their putative affinity for H-2 I-Ab/Db and Kb molecules (Table 1). Peptide VP7(327) was selectedin spite of its relatively weak binding scores to Db and Kb moleculesas it is contained in the MHC class II binding peptide VP7(324).

To confirm the binding to MHC class I molecules, we performedbinding assays using RMA-S cells. These cells are deficient in thetransporter-associated protein and express little to no MHC classI molecules (Db or Kb) on the cell surface [34]. By adding exoge-nous peptide, these molecules are stabilised and the increasedexpression can be measured by flow cytometry using either anti-Db or anti-Kb antibodies. As predicted, all 7 peptides selected forDb and/or Kb binding stabilised these molecules on the surfaceof RMA-S cells to varying degree. Binding was ranked accord-ing to the control peptide gp(33–41) of moderate binding affinity[35,36]. Thus, 2 VP7 peptides (VP7(283) and VP7(72)( were strongbinders of Db molecules, whereas another 2 VP7 peptides (VP7(175)and VP7(327( were moderate binders of Db (Fig. 2A and B). Pep-tide VP7(245) (binding Kb molecules) showed no binding to Db

molecules. Similarly, 2 VP7 peptides, (VP7(35) and VP7(80)( werestrong binders of Kb molecules. VP7(245) peptide showed low bind-ing to Kb (Fig. 2A and B), whereas VP7(327) peptide showed nobinding. Taken together, these data confirm that the predictivealgorithms used in the present study were able to select MHC classI-restricted peptides from VP7 capable of binding either Db or Kb

molecules.

3.3. Several peptides from VP7 are recognized by T cells after BTVchallenge

Splenocytes from infected C57BL/6 and vaccinated IFNAR(−/−)

mice were cultured in presence of these VP7 peptides, and thespecific production of IFN-� was assessed by ELISPOT. All VP7peptides were recognized by splenocytes from infected C57BL/6mice (Fig. 3A). Thus, specific IFN-� production to the putativeMHC class II-binding peptides VP7(139), VP7(324) and VP7(181)were detected respectively, in 7, 5 and 3 out of 7 mice tested.Although IFN-� production could be detected to all 7 MHC classI-binding peptides, it appeared that peptides VP7(283), VP7(72),VP7(327) and VP7(80) were consistently recognized by spleno-cytes from C57BL/6 mice infected with BTV-8. Importantly, IFN-�production was statistically significant in cultures with the I-Ab binding peptides VP7(139) and VP7(324), the Db bindingpeptides VP7(72), VP7(283) and VP7(327) and the Kb bindingpeptide VP7(80). In vaccinated IFNAR(−/−) mice, IFN-� produc-tion to VP7(139) and VP7(324) peptides was detected in mostanimals (6 and 5 out of 7 mice tested, respectively) (Fig. 3B). Inter-estingly, no specific IFN-� production was detected to VP7(181)peptide. Varying levels of specific IFN-� production were alsodetected to all VP7 MHC class I-binding peptides, although pep-tides VP7(283), VP7(327), VP7(35) and VP7(80) were consistentlyrecognized by vaccinated IFNAR(−/−) mice. In general, IFN-� pro-duction in IFNAR(−/−) mice was statistically significant with theI-Ab binding peptides VP7(139) and VP7(324), the Db bindingpeptides VP7(283) and VP7(327) and the Kb binding peptidesVP7(35) and VP7(80). These data suggest that BTV infection trig-gers the activation of T cells recognizing multiple epitopes fromthe virus.

3.4. T cells elicited after BTV challenge are functional

We focused our attention on the VP7 peptides consistentlyrecognized by splenocytes from BTV-8 infected C57BL/6 mice

Page 4

J.-M. Rojas et al. / Vaccine 29 (2011) 6848– 6857 6851

Fig. 1. T cell response to BTV-8 in C57BL/6 and IFNAR(−/−) mice. (A) Production of IFN-� by splenocytes from C57BL/6 mice challenged with BTV-8 (measured by ELISPOT) inresponse to BTV-8 or inactivated virus (BEI-BTV). Controls are cells cultured without stimuli (alone) or with a lysate from the cell line used to prepare the viral stock (MC57).(B) Production of IFN-� by splenocytes from IFNAR(−/−) mice vaccinated with BEI-BTV in response to BTV-8 or inactivated BTV. ELISPOT data are representative of 7 mice ineach group. (C) Proliferation by tritiated-thymidine uptake assays of splenocytes from BTV-8 infected C57BL/6 mice when the inactivated-virus was present in the culture.( red wf CD8+

i metry

aI(ffoV

D) Proliferation assay of splenocytes from vaccinated IFNAR(−/−) mice when cultuor each strain. (E) Intracellular cytokine staining (IFN-�) of BTV-specific CD4+ andndicate the amount of IFN-� positive cell within the T cell compartment. Flow cyto

nd BTV-vaccinated IFNAR(−/−) mice. Proliferation assays to the-Ab binding peptides VP7(139), VP7(181), VP7(324) and VP7(327)putative core binding sequence of VP7(324) peptide) were per-

ormed using splenocytes from infected C57BL/6 mice (Fig. 4A) androm IFNAR(−/−) vaccinated mice (Fig. 4B). With both models, webserved a significant proliferation of the splenocytes to VP7(139),P7(324) and VP7(327) peptides but not to the control (DMSO) or

ith inactivated-virus. Proliferation assay data are representative of at least 3 miceT cells in C57BL/6-challenged mice or (F) vaccinated IFNAR(−/−) mice. Percentages

data is representative of 5 mice in each group. Data are presented as mean ± SD.

VP7(181) peptide. These data indicate that these T cells are capableof proliferating when they re-encounter the antigen. To determinethe lytic activity of the T cells specific for the immunogenic Db and

Kb binding peptides VP7(283), VP7(327) and VP7(80), CTL assayswere carried out. In both C57BL/6 (Fig. 4C) and IFNAR(−/−) (Fig. 4D)mice, CTLs specific for VP7(283), VP7(80) and VP7(327) peptidewere produced but not for VP7(175) peptide. All 3 VP7 peptides

Page 5

6852 J.-M. Rojas et al. / Vaccine 29 (2011) 6848– 6857

Table 1VP7 epitope prediction in H-2b haplotype according to predictive algorithms. Ten peptides from VP7 were selected on the basis of their putative affinity for H-2 I-Ab/Db andKb molecules.

Peptide name Position Sequence Allele binding prediction Score SYFPEITHI Score ProPred I Score NetMHC

VP7(139) 139–153 GRWFMRAAQAVTAVV I-Ab N/A N/A 0.626VP7(324) 324–338 RPEFAIHGVNPMPGP I-Ab N/A N/A 0.568VP7(181) 181–195 PMMIYLVWRRIENFA I-Ab N/A N/A 0.101VP7(72) 72–80 AAGINVGPI Db 26 6.37 0.695VP7(175) 175–183 FQGRNDPMM Db 24 5.34 0.615VP7(283) 283–291 TAILNRTTL Db 30 9.74 0.905VP7(327) 327–335 FAIHGVNPM Db/Kb (I-Ab)* 17/8 3.07/0.18 0.709/0.436VP7(35) 35–43 IAINRYNGL Kb 23 3.27 0.795VP7(80) 80–88 ISPDYTQHM Kb 16 2.25 0.654VP7(245) 245–253 IQVVFYISM Kb 17 3.09 0.703

N/A: not available.* Putative core binding sequence for I-Ab of VP(324) peptide.

Fig. 2. VP7 peptide binding to MHC class I H-2b molecules. RMA-S cells were incubated in serum-free RPMI with different peptide concentrations for 1 h at 37 ◦C, 5% CO2. Cellswere subsequently stained with anti-Db-FITC and anti-Kb-PE antibodies to evaluate the stabilisation of MHC class I molecules by flow cytometry. (A). Results are expressedas the average mean fluorescence intensity (MFI) ratio between the cells incubated with and without peptide of 3 independent experiments. Db . Two VP7 peptides (VP7(283)and VP7(72)) were strong binders of Db molecules, whereas another 2 VP7 peptides (VP7(175)and VP7(327) were moderate binders of Db. Peptide VP7(245) (binding Kb

molecules) showed no binding to Db molecules, whereas the reported gp(33–41) peptide from LCMV showed moderate binding to this molecule. Kb . Two VP7 peptides,(VP7(35)and VP7(80)) were strong binders of Kb molecules, VP7(245) peptide showed low binding to Kb, whereas VP7(327) peptide showed no binding. Negative controlp b b ontro b

r g at thm betw

cdvtscVep(n

eptide VP7(175) (binding D molecules) showed no binding to K , and positive canked according to the ratio between VP7 peptide and gp(33–41) peptide bindinoderate binder [35,36], thus ratios above 1.2 were defined as strong binding, ratios

apable of inducing CTLs, were also presented endogenously, asemonstrated by the lysis of target cells pulsed with the inacti-ated virus (Fig. 4E). CTLs elicited by BTV infection have thereforehe potential to lyse BTV-infected cells. To determine which T cellubset (CD4+ or/and CD8+) can recognize VP7 epitopes, spleno-ytes from infected C57BL/6 mice were stimulated for 5 h withP7 peptides, and subsequently stained for IFN-�, CD4 and CD8. As

xpected, peptide VP7(139) elicited a CD4+ T cell response, whereaseptides VP7(283) and VP7(80) elicited a CD8+ T cell responseFig. 5). Similar data were obtained from IFNAR(−/−) mice (dataot shown). Remarkably, peptides VP7(324) and VP7(327) elicited

l peptide gp(33–41) showed intermediate binding to K . (B). Binding affinity wase saturating concentrations of 20 (g/ml. Peptide gp(33–41) could be defined as a

een 1.2 and 0.8 as moderate binding, and ratios between 0.8 and 0.4 as low binding.

a CD4+ and a CD8+ T cell response. This confirms that VP7(327)peptide is likely to represent the core binding sequence to I-Ab

molecules, and that peptide VP7(324) contains both CD4 and CD8epitope.

3.5. VP7 T cell epitopes identified in murine models can also berecognized by T cells from the natural host

We next determine whether these VP7 epitopes are conservedin sheep, a natural host of BTV. PBL from eight BTV-8 infectedsheep were obtained from blood 28 days post-infection. Cells were

Page 6

J.-M. Rojas et al. / Vaccine 29 (2011) 6848– 6857 6853

200 300

BA

**

RANFI6/LB75C (-/-)

100

150

200

**

**

***

**

**

**

*

0

50

Ave

rage

IFN

-γ s

pots

for

1x10

6 ce

lls

Ave

rage

IFN

-γ s

pots

for

1x10

6 ce

lls

0

100

***

*

0

Con

trol

Pept

139

Pept

324

Pept

181

Pept

283

Pept

72

Pept

175

Pept

327

Pept

35

Pept

80

Pept

245

0

Con

trol

Pept

139

Pept

324

Pept

181

Pept

283

Pept

72

Pept

175

Pept

327

Pept

35

Pept

80

Pept

245

Fig. 3. IFN-� production to VP7 peptides in C57BL/6 and IFNAR(−/−) mice. (A) C57BL/6 mice were inoculated with 100 PFU of BTV-8 three times at 10–11 day intervals. Threedays after the last inoculation mice were sacrificed and splenocytes were obtained to measure IFN-� production to the VP7 peptides by ELISPOT. Data are presented asaverage IFN-� production for each peptide in each mouse and (—) indicates the median IFN-� production. As negative control (�), cells were cultured with a volume of DMSOequivalent to that added with the peptides. Specific IFN-� production to the putative MHC class II-binding peptides (�) VP7(139), VP7(324) and VP7(181) were detectedrespectively, in 7, 5 and 3 out of 7 mice tested. (B) IFNAR(−/−) mice were vaccinated with BEI-inactivated virus twice at a week interval and subsequently challenged with10 PFU of BTV-8. Vaccinated animals were sacrificed 6–7 day-post challenge and splenocytes were obtained to measure the specific IFN-� production to the VP7 peptides.I s (6 ant to all

U contro

cAVictsVacoTs

4

Mi2Vttppgo

nAt

FN-� production to VP7(139) and VP7(324) peptides was detected in most animalo VP7(181) peptide. Varying levels of specific IFN-� production were also detected

test was used to compare IFN-� production between the peptide groups and the

ultured in the presence of VP7 peptides or inactivated virus.s control, cells were cultured with the non-stimulating peptideP7(181). PBL from all inoculated sheep produced IFN-� to the

nactivated-virus, demonstrating that BTV-8 infection activates Tell immune responses to the virus (Fig. 6A). By contrast, pep-ide specific IFN-� production was detected in 4 out of 8 infectedheep for VP7(139), in 5 out of 8 for VP7(324), in 3 out of 8 forP7(327), in 4 out of 8 for VP7(283) and in 4 out of 8 for VP7(80),nd just peptides VP7(324), VP7(283) and VP7(80) were signifi-antly capable of inducing IFN-� production (Fig. 6A). The groupf naïve animals were not producing IFN-� in any case (Fig. 6B).herefore, 3 VP7 epitopes identified in mouse are conserved inheep.

. Discussion

In this work, we have characterised, for the first time, MHC-I andHC-II T cell epitopes in the VP7 protein of BTV that are recognized

n mouse models and in natural hosts. A total of 3 CD8 epitopes and CD4 epitopes have been identified that map at different places onP7 core protein. Analysis by BLASTP on the NCBI website revealed

hat several VP7 proteins of other BTV serotypes have identical pep-ide sequences to the VP7 T cell epitopes of BTV-8 described in theresent study (Fig. 7). These T cell epitopes are therefore likely to beresent in multiple BTV serotypes and represent an attractive tar-et for vaccination capable of protecting against a broad-spectrumf serotypes.

The use of predictive algorithms has successfully identifiedovel T cell epitopes from virus [37,38] and tumour antigens [39].ll 7 predicted peptides were capable of binding their respec-

ive MHC class I molecules, and were recognized by T cells from

d 5 out of 7 mice tested, respectively). No specific IFN-� production was detectedVP7 MHC class I-binding peptides. *p < 0.05, **p < 0.01, A two-tailed Mann–Whitneyl group.

infected/vaccinated mice. This observation suggested that anti-BTVT cell immune responses are likely to be directed at several epi-topes, questioning the existence of a strong immunodominant biasduring the development of T cell responses to BTV. This data con-firms a similar observation in sheep [21], where multiple epitopesfrom the NS1 protein appeared to induce CTLs. Whether this is aparticularity of BTV immune responses or a competition mecha-nism through which the virus evades the immune system remainsto be determined. In any case, protection from the disease is likelyto require the activation of multiple T cell populations specific forseveral BTV epitopes.

The elimination of BTV infection is likely to be mediatedby a combination of cell and antibody-mediated immunity.However, neutralizing antibodies are probably the first line ofdefense when re-infection occurs. Due to the variability of BTVouter capsid, re-infection is therefore likely to occur with avirus of different serotypes. The partial cross-protection betweenserotypes observed with CTL in several reports is thereforelikely due to this phenomenon [23,26]. Indeed, naive popu-lations are particularly susceptible to BTV infection, whereasendemic populations can cope very well with the disease. Thisimplies that a certain degree of immunity to BTV exists inthe endemic population. Since T cells are more likely to cross-react with other serotypes [22,23,26], the establishment of avaccine stimulating T cells responses which overlap a widerange of BTV serotypes is in turn likely to protect naïve pop-ulations. However, it is possible that cross-reactive CTL alone

would not be sufficient to protect animals. A T cell-basedvaccine should also include a component capable of stimu-lating CD4+ T cells, which are central to the developmentand maturation of effective cellular and humoral immunity.

Page 7

6854 J.-M. Rojas et al. / Vaccine 29 (2011) 6848– 6857

Fig. 4. Recognition of VP7 peptides by splenocytes from infected/vaccinated mice. (A and B). Splenocyte proliferation to VP7 I-Ab-restricted peptides. Splenocytes from (A)BTV-8-challenged C57BL/6 mice or (B) vaccinated-IFNAR(−/−) mice were cultured for 72–96 h in the presence of VP7 peptides and [3H] thymidine to measure proliferation. Dataare reported as counts per minute (cpm) ([3H] thymidine incorporation) (mean ± SD) and are representative of at least 3 mice. *p < 0.01 unpaired Student’s t-Test (peptidesVP7(139)/VP7(324)/VP7(327) vs. Control (DMSO) and VP7(181)). (C) CTL assays for VP7 peptides. Splenocytes from BTV-8-challenged C57BL/6 mice or (D) from vaccinatedIFNAR(−/−) mice were cultured with a VP7 peptide for 5–6 days. The cytotoxicity of these cells was tested against peptide-pulsed 51Cr labelled target cells (either syngeneicLPS blast or RMA-S cells). (E) CTLs (from C57BL/6) capable of lysing cells presenting VP7(283), VP7(80) or VP7(327) peptides were also able to lyse target cells (LPS-blast)p cesseo ith reli

Itv

wuVVmvc

ulsed with inactivated virus, demonstrating that these peptides are naturally prof at least 3 mice for each stain. *p < 0.01 unpaired Student’s t-test (cytotoxicity wrrelevant peptide (�) gp(33–41)).

n that respect, activating cross-reactive CD4+ T cells is likelyo be central towards the development of a universal BTVaccine.

We found that I-Ab binding peptides VP7(139) and VP7(324)ere consistently recognized by infected and vaccinated animalssed in the present study. MHC class I-binding peptides VP7(80),P7(283) and VP7(327) were also recognized. Intriguingly, peptide

P7(72) appeared to be only recognized by T cells from infectedice, whereas peptide VP7(35) was only recognized by T cells from

accinated mice. This observation suggests that infection and vac-ination may activate a different profile of T cell specificities. If this

d by antigen presenting cells (E:T ratio = 66:1). Data presented are representativeevant peptide (�) (VP7(283)/VP7(80)/VP7(327)) or inactivated virus (BEI-BTV) vs.

observation was to be confirmed it may represent the basis for aDIVA diagnostic test of BTV.

We demonstrated that the T cells specific for these peptideswere functional. I-Ab binding peptides VP7(139), VP(324) andVP7(327) induced CD4+ T cells to produce IFN-� and prolifer-ate. Whereas MHC class I-binding peptides VP7(80), VP7(283) andVP7(327) induced CD8+ T cells to produce IFN-� and differentiate

into CTL capable of lysing cells pulsed with virus. To the best ofour knowledge, these VP7 peptide sequences are the first descrip-tion of CD4+ and CD8+ T cell epitopes from the VP7 core protein ofBTV in any host. The functional characterisation of these cells also

Page 8

J.-M. Rojas et al. / Vaccine 29 (2011) 6848– 6857 6855

Fig. 5. Intracellular IFN-� staining of VP7 peptide-specific CD4+ and CD8+ T cells. Splenocytes from C57BL/6 mice were cultured in presence of VP7 peptides for 5 h. PeptidesVP7(324) and VP7(327) induced both CD4+ and CD8+ T cell responses, confirming that they are likely to bind H-2 Db and I-Ab molecules. Peptides VP7(283) and VP7(80)induced IFN-� production by CD8+ T cells, whereas peptide VP7(139) induced CD4+ T cell responses. Data presented are representative of at least 4 mice.

BA 500 500

pg/m

l)

300

400

500**

**

**

300

400

500

pg/m

l)

IFN

-γ(p

100

200**

100

200

IFN

-γ(p

trol p

eptid

e

BEI-BTV

VP7(139

)

VP7(324

)

VP7(327

)

VP7(283

)

VP7(80)

0

ntrol p

eptid

e

BEI-BTV

VP7(139

)

VP7(324

)

VP7(327

)

VP7(283

)

VP7(80)

0

ContrCont

Fig. 6. IFN-� production to inactivated BTV and VP7 peptides in infected sheep. (A). PBLs from infected sheep were collected 28 days post-infection and cultured for 96 hwith inactivated virus, VP7 peptides, or the non-immunogenic peptide VP7(181) as control. Culture supernatants were then harvested, and IFN-� production was measuredby ELISA. (B). PBLs from PBS inoculated sheep as control were collected and cultured and stimulated as PBLs from BTV-8 infected sheep. Data are presented as average IFN-�production (pg/ml) for each peptide in each animal and (—) indicates the median IFN-� production. (**) Indicates significant IFN-� production by Mann–Whitney test.

Page 9

6856 J.-M. Rojas et al. / Vaccine 29 (2011) 6848– 6857

BTV8 IAINRYNGL AAGINVGPISPDYTQHM GRWFMRAAQAVTAVV FQGRNDPMM IQVVFYISM TAILNRTTL RPEFAIHGVNPMPGP

VP7(35) VP7(72) VP7(80) VP7(139) VP7(175) VP7(245) VP7(283) VP7(324 )

BTV8 …IAINRYNGL…..…AAGI NVGPI SPDYTQHM………… ….. GRWF MRAAQ AVTAVV…FQGRN DPMM……… IQVVF YISM…TAILNRT TL ……..RPEFAIHGVNPM PGP…BTV1 ……………………………..…………… …………… ………………… …………. .……………… …………… ………………… …………… …………….… …………… ………………… …………… …… ……… ……………………………… …………BTV3 ……………………………..…………… …………… ………………… …………… ……..……… …………… ………………… …………… …………….… …………… …V…………… …………… …… ……. ……………………………… …………BTV4 ……………………………..…………… …………… ………………… …………. .……………… …………… ………………… …………… …………….… …………… ………………… …………… …… ….. ……………………………… …………BTV6 ……………………………..…………… …………… ………………… …………. .……………… …………… ………………… …………… ………………. .………… …V…………… …………… …… ….. ……………………………… …………BTV9 ……………………………..…………… …………… ………………… …………. .……………… …………… ………………… …………… ………………… …………… ………………… …………… …… ….. ……………………………… ………… BTV10 …………………………… ..…………… …………… ………………… …………. .……………… …………… ………………… …………… ………………… …………… ………………… …………… …… ….. ……………………………… …………

R21VTB D IQBTV12 …………………………… ..…………… ………………R…………… …………. .……………… …………… ………………… …………… ………………… …………… ………………… …………… …… ……. ………… D………………… IQ……BTV13 …………………………… ..…………… …………… ………………… …………. .……………… …………… ………………… …………… ………………… …………… …V…………… …………… …… ….. ……………………………… …………BTV17 …………………………… ..…………… …………… ………………… …………. .……………… …………… ………………… …………… ………………. …………… ………………… …………… …… ….. ……………………………… …………BTV18 …………………………… ..…………… …………… ………………… …………. .……………… …………… ………………… …………… ………………… …………… ………………… …………… …… ….. ……………………………… …………BTV19 …………………………… ..…………… …………… …………Q…… ……... .……………… …………… ………P…I… …………… …………….… …………… ………………… …………… …… ……. …………V………IA…I V…Q…BTV21 …………………………… ..…………… …………… ………………… …………. .……………… …………… ………………… …………… ………………… …………… ………………… …………… …… ….. …………D………………I Q………BTV23 …………………………… ..…………… …………… ………………… …………. .……………… …………… ………………… …………… ………………… …………… ………………… …………… …… ….. ……………………………… …………

V52VTB ABTV25 …………………………… ..…………… …………… …………V……..... .……………… …………… ………………… …………… ………………… …………… ………………… …………… …… ….. …………………………A……………

VP7(327)

Fig. 7. Locations of the 9 T cell epitopes in VP7 of BTV-8. The amino acid of the VP7 protein of 15 BTV serotypes (8, 1, 3, 4, 6, 9, 10, 12, 13, 17, 18, 19, 21, 23 and 25) containingt oxes:d ences

cc

bIpemgVisNTsrv

bappir

taftmwswowp

ThaIit

[

he described T cell epitopes were aligned. The identified sequences are shown in bots indicate the BTV-8 matches among amino acids. (For interpretation of the refer

onfirms that anti-BTV immunity is likely to be driven by several Tell epitopes.

Interestingly, the VP7(324) and the VP7(327) peptides elicitedoth a CD4+ and a CD8+ T cell response. According to the NetMHC-

I predictive algorithm [30] the core binding sequence of VP7(324)eptide is the VP7(327) peptide, and our data support this hypoth-sis as the VP7(327) sequence was capable of binding to H-2 Db

olecules, inducing IFN-� production by both T cell subsets andenerating CTL activity. Importantly, no direct stabilisation withP7(324) peptide of MHC class I molecules was observed in bind-

ng assays (data not shown), indicating that CD8+ T cell responseseen in these experiments required cross-presentation (and likely- and C-terminal degradation) of this peptide onto Db molecules.aken together, these data demonstrate that the VP7(324) peptideequence contains both a CD4 and a CTL epitope. This particularegion of VP7 may therefore represent an interesting target foraccination.

These T cells epitopes identified in mouse were also recognizedy cells from the natural host. Peptides VP7(283) and VP7(324)ppear to be the most cross-reactive in sheep as they induce IFN-�roduction in 4 and in 3 animals respectively. PBL from all sheeproduced IFN-� specifically to the inactivated virus, confirming the

mportance of T cell responses in BTV immunity [18]. These dataepresent the first description of BTV T cell epitopes in sheep.

Analysis of the presence of these peptide sequences in VP7 pro-ein expressed in other BTV serotypes, showed that these 2 CD4nd 3 CTL epitopes are likely to be present in most serotypesor which sequences are available. T cells specific for these epi-opes are therefore likely to react to other BTV serotypes, and

ay provide protection. This is in line with observations in sheep,here anti-BTV CTLs were capable of cross-reacting with other BTV

erotypes [16,21,22]. Recently, a study in which sheep vaccinatedith BEI-inactivated BTV-1 induced cross-reactive T cells against

ther serotypes has been reported [23]. Activation of T cells forell-preserved protein sequences from BTV is therefore likely torotect the host against a wide spectrum of BTV serotypes.

In conclusion, our work demonstrates that both CD4+ and CD8+

cells are activated following BTV infection and vaccination. Weave previously reported that in the absence of IFN-�� receptor the

ntibody response is strong enough to clear the viral infection [40].n this new report, the protection we observed with the vaccine andn the absence of IFN-�� receptor is expected to rely on T cell activa-ion since CD4+ T cell is central for activating humoral response. We

[

[

black boxes indicate MHC-I epitopes and red boxes indicate MHC-II epitopes. Theto color in this figure legend, the reader is referred to the web version of the article.)

were also able to characterise 2 CD4 epitopes and 3 CTL epitopesfrom the VP7 core protein of BTV-8. Importantly, we confirmedthat BTV might induce T cell responses specific for several epitopes,and thereby successful vaccination is likely to require activation ofanti-BTV CD4+ and CD8+ T cells. Understanding the T cell responseand characterising other T cell epitopes from well-conserved BTVproteins will therefore provide the tools to develop more effectivevaccination capable of targeting multiple BTV serotypes.

Acknowledgements

This work was funded by grants AGL2009-07353 fromMinisterio de Ciencia e Innovación (MCINN) and 228394-NADIR-Integrating Activities 7th EU program. T.R was supported by acontract from Comunidad Autónoma de Madrid.

References

[1] Wilson AJ, Mellor PS. Bluetongue in Europe: past, present and future. PhilosTrans R Soc Lond B Biol Sci 2009;364(September (1530)):2669–81.

[2] Mellor P, Baylis M, Mertens PP. Bluetongue. London, UK: Elsevier, AcademicPress; 2009.

[3] Wilson WC, Ma HC, Venter EH, van Djik AA, Seal BS, Mecham JO. Phylogeneticrelationships of bluetongue viruses based on gene S7. Virus Res 2000;67(April(2)):141–51.

[4] Roy P, Mertens PP. Orbiviruses and coltiviruses: molecular biology. In: Gra-noff A, Webster RG, editors. Encyclopaedia of virology. 2nd ed. London, UnitedKindom: Academic Press; 1999. p. 1062–74.

[5] Grimes JM, Burroughs JN, Gouet P, Diprose JM, Malby R, Zientara S, et al.The atomic structure of the bluetongue virus core. Nature 1998;395(October(6701)):470–8.

[6] Mertens PP, Burroughs JN, Anderson J. Purification and properties of virus par-ticles, infectious subviral particles, and cores of bluetongue virus serotypes 1and 4. Virology 1987;157(April (2)):375–86.

[7] Roy P. Bluetongue virus proteins. J Gen Virol 1992;73(December (Pt12)):3051–64.

[8] Mertens PP, Pedley S, Cowley J, Burroughs JN, Corteyn AH, Jeggo MH, et al.Analysis of the roles of bluetongue virus outer capsid proteins VP2 and VP5 indetermination of virus serotype. Virology 1989;170(June (2)):561–5.

[9] Maan S, Maan NS, Samuel AR, Rao S, Attoui H, Mertens PP. Analysis and phy-logenetic comparisons of full-length VP2 genes of the 24 bluetongue virusserotypes. J Gen Virol 2007;88(February (Pt 2)):621–30.

10] Hassan SS, Roy P. Expression and functional characterization of bluetonguevirus VP2 protein: role in cell entry. J Virol 1999;73(December (12)):9832–42.

11] Nason EL, Rothagel R, Mukherjee SK, Kar AK, Forzan M, Prasad BV, et al.Interactions between the inner and outer capsids of bluetongue virus. J Virol2004;78(August (15)):8059–67.

12] Huismans H, Van Dijk AA. Bluetongue virus structural components. Curr TopMicrobiol Immunol 1990;162:21–41.

Page 10

cine 2

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

J.-M. Rojas et al. / Vac

13] Han Z, Harty RN. The NS3 protein of bluetongue virus exhibits viroporin-likeproperties. J Biol Chem 2004;279(October (41)):43092–7.

14] Barratt-Boyes SM, Rossitto PV, Stott JL, MacLachlan NJ. Flow cytometric analysisof in vitro bluetongue virus infection of bovine blood mononuclear cells. J GenVirol 1992;73(August (Pt 8)):1953–60.

15] Hemati B, Contreras V, Urien C, Bonneau M, Takamatsu HH, Mertens PP, et al.Bluetongue virus targets conventional dendritic cells in skin lymph. J Virol2009;83(September (17)):8789–99.

16] Takamatsu H, Jeggo MH. Cultivation of bluetongue virus-specific ovine Tcells and their cross-reactivity with different serotype viruses. Immunology1989;66(February (2)):258–63.

17] Jeggo MH, Wardley RC, Taylor WP. Role of neutralising antibody in passiveimmunity to bluetongue infection. Res Vet Sci 1984;36(January (1)):81–6.

18] Jeggo MH, Wardley RC, Brownlie J. A study of the role of cell-mediated immu-nity in bluetongue virus infection in sheep, using cellular adoptive transfertechniques. Immunology 1984;52(July (3)):403–10.

19] Stott JL, Osburn BI, Barber TL. The current status of research on an experimentalinactivated bluetongue virus vaccine. Proc Annu Meet U S Anim Health Assoc1979;(83):55–62.

20] Stott JL, Barber TL, Osburn BI. Immunologic response of sheep to inactivatedand virulent bluetongue virus. Am J Vet Res 1985;46(May (5)):1043–9.

21] Janardhana V, Andrew ME, Lobato ZI, Coupar BE. The ovine cytotoxic T lympho-cyte responses to bluetongue virus. Res Vet Sci 1999;67(December (3)):213–21.

22] Andrew M, Whiteley P, Janardhana V, Lobato Z, Gould A, Coupar B. Antigenspecificity of the ovine cytotoxic T lymphocyte response to bluetongue virus.Vet Immunol Immunopathol 1995;47(August (3–4)):311–22.

23] Umeshappa CS, Singh KP, Pandey AB, Singh RP, Nanjundappa RH.Cell-mediated immune response and cross-protective efficacy of binaryethylenimine-inactivated bluetongue virus serotype-1 vaccine in sheep. Vac-cine 2010;28(March (13)):2522–31.

24] Takamatsu H, Burroughs JN, Wade-Evans AM, Mertens PP. Bluetongue, AfricanHorse Sickness and related Orbiviurus. In: Walton TE, Osburn BI, editors. Pro-ceedings of the 2nd International Symposium. CRC Press; 1992. p. 491–7.

25] Gumm ID, Newman JF. The preparation of purified bluetongue virus groupantigen for use as a diagnostic reagent. Arch Virol 1982;72(1–2):83–93.

26] Wade-Evans AM, Romero CH, Mellor P, Takamatsu H, Anderson J,Thevasagayam J, et al. Expression of the major core structural protein (VP7) ofbluetongue virus, by a recombinant capripox virus, provides partial protection

of sheep against a virulent heterotypic bluetongue virus challenge. Virology1996;220(June (1)):227–31.

27] Rammensee HG, Bachmann J, Emmerich NPN, Bachor OA, Stevanovic S.SYFPEITHI: database for MHC ligands and peptide motifs. Immunogenetics1999;50:213–9.

[

9 (2011) 6848– 6857 6857

28] Singh H, Raghava GP. ProPred1: prediction of promiscuous MHC Class-I bindingsites. Bioinformatics 2003;19(May (8)):1009–14.

29] Nielsen M, Lundegaard C, Worning P, Lauemoller SL, Lamberth K, Buus S,et al. Reliable prediction of T-cell epitopes using neural networks with novelsequence representations. Protein Sci 2003;12(May (5)):1007–17.

30] Nielsen M, Lund O. NN-align. An artificial neural network-based alignmentalgorithm for MHC class II peptide binding prediction. BMC Bioinformatics2009;10:296.

31] Andersen ML, Ruhwald M, Nissen MH, Buus S, Claesson MH. Self-peptideswith intermediate capacity to bind and stabilize MHC class I molecules maybe immunogenic. Scand J Immunol 2003;52:21–7.

32] Firat H, Garcia-Pons F, Tourdot S, Pascolo S, Scardino A, Garcia Z, et al. H-2class I knockout, HLA-A2.1-transgenic mice: a versatile animal model for pre-clinical evaluation of antitumor immunotherapeutic strategies. Eur J Immunol1999;29(October (10)):3112–21.

33] Calvo-Pinilla E, Rodriguez-Calvo T, Anguita J, Sevilla N, Ortego J. Establishmentof a bluetongue virus infection model in mice that are deficient in the alpha/betainterferon receptor. PLoS One 2009;4(4):e5171.

34] Schumacher TN, Heemels MT, Neefjes JJ, Kast WM, Melief CJ, Ploegh HL. Directbinding of peptide to empty MHC class I molecules on intact cells and in vitro.Cell 1990;62(August (3)):563–7.

35] Gairin JE, Mazarguil H, Hudrisier D, Oldstone MB. Optimal lymphocytic chori-omeningitis virus sequences restricted by H-2Db major histocompatibilitycomplex class I molecules and presented to cytotoxic T lymphocytes. J Virol1995;69(April (4)):2297–305.

36] Hudrisier D, Oldstone MB, Gairin JE. The signal sequence of lymphocytic chori-omeningitis virus contains an immunodominant cytotoxic T cell epitope thatis restricted by both H-2D(b) and H-2K(b) molecules. Virology 1997;234(July(1)):62–73.

37] Wang B, Yao K, Liu G, Xie F, Zhou F, Chen Y. Computational prediction and iden-tification of Epstein-Barr virus latent membrane protein 2A antigen-specificCD8+ T-cell epitopes. Cell Mol Immunol 2009;6(April (2)):97–103.

38] Larsen MV, Lelic A, Parsons R, Nielsen M, Hoof I, Lamberth K, et al. Iden-tification of CD8+ T cell epitopes in the West Nile virus polyprotein byreverse-immunology using NetCTL. PLoS One 2010;5(September (9)):e12697.

39] Rojas JM, McArdle SE, Horton RB, Bell M, Mian S, Li G, et al. Peptide immu-nisation of HLA-DR-transgenic mice permits the identification of a novelHLA-DRbeta1*0101- and HLA-DRbeta1*0401-restricted epitope from p53. Can-

cer Immunol Immunother 2005;54(March (3)):243–53.

40] Calvo-Pinilla E, Rodriguez-Calvo T, Sevilla N, Ortego J. Heterologous prime boostvaccination with DNA and recombinant modified vaccinia virus Ankara protectsIFNAR(−/−) mice against lethal bluetongue infection. Vaccine 2009;28(Decem-ber (2)):437–45.