Induction of a Cross-Reactive CD8⁺ T Cell Response following Foot-and-Mouth Disease Virus Vaccination

Abstract

Foot-and-mouth disease virus (FMDV) causes a highly contagious infection in cloven-hoofed animals. Current inactivated FMDV vaccines generate short-term, serotype-specific protection, mainly through neutralizing antibody. An improved understanding of the mechanisms of protective immunity would aid design of more effective vaccines. We have previously reported the presence of virus-specific CD8⁺ T cells in FMDV-vaccinated and -infected cattle. In the current study, we aimed to identify CD8⁺ T cell epitopes in FMDV recognized by cattle vaccinated with inactivated FMDV serotype O. Analysis of gamma interferon (IFN-γ)-producing CD8⁺ T cells responding to stimulation with FMDV-derived peptides revealed one putative CD8⁺ T cell epitope present within the structural protein P1D, comprising residues 795 to 803 of FMDV serotype O UKG/2001. The restricting major histocompatibility complex (MHC) class I allele was N*02201, expressed by the A31 haplotype. This epitope induced IFN-γ release, proliferation, and target cell killing by αβ CD8⁺ T cells, but not CD4⁺ T cells. A protein alignment of representative samples from each of the 7 FMDV serotypes showed that the putative epitope is highly conserved. CD8⁺ T cells from FMDV serotype O-vaccinated A31⁺ cattle recognized antigen-presenting cells (APCs) loaded with peptides derived from all 7 FMDV serotypes, suggesting that CD8⁺ T cells recognizing the defined epitope are cross-reactive to equivalent peptides derived from all of the other FMDV serotypes.

Foot-and-mouth disease virus (FMDV) is a member of the family Picornaviridae, genus Aphthovirus. The FMDV particle consists of a positive-strand RNA molecule of approximately 8,500 nucleotides, enclosed within an icosahedral capsid. The genome encodes a unique polyprotein from which four structural proteins (P1A, P1B, P1C, and P1D; also referred to as VP4, VP2, VP3, and VP1, respectively) and nine nonstructural proteins are cleaved by viral proteases (48). FMDV shows a high genetic and antigenic variability, which is reflected in the seven serotypes and multiple subtypes reported so far (13). The virus causes a highly contagious infection in cloven-hoofed animals which is characterized by the formation of vesicles on the mouth, tongue, nose, and feet. In addition, most infected animals develop viremia.

The virus elicits a rapid humoral response in both infected and vaccinated animals (26). Virus-specific antibodies protect animals in a serotype-specific manner against reinfection or against infection in the case of vaccination, and protection is generally correlated with high levels of neutralizing antibodies (38). Control of the disease is achieved by vaccination with a chemically inactivated whole-virus vaccine emulsified with adjuvant; however, this provides only short-term, serotype-specific protection (2). The introduction of this vaccine has been very successful in areas of the world where the disease is enzootic. However, one of the major difficulties in implementing vaccination is the inability to distinguish vaccinated animals from infected/recovered animals, which may still be shedding virus. Currently, a number of assays specifically developed for this purpose are being validated (29, 41), and the success of these assays is dependent on the use of purified vaccine antigen. A strategy using replication-deficient adenovirus 5 expressing FMDV antigens has been shown to provide early protection against homologous challenge (39).

The identification and characterization of T cell epitopes are important for understanding protective immunity mediated by CD8⁺ and CD4⁺ T lymphocytes. Such T cell responses are pathogen specific and are restricted by major histocompatibility complex (MHC) class I and class II molecules, which present foreign peptides to the immune system (55, 56). The role of cellular immunity in the protection of animals from FMDV is still a matter of some controversy. Specific T cell-mediated antiviral responses have been observed in cattle and swine following either infection or vaccination (3, 7, 24). CD4⁺ T cell responses are suggested to play an important role in protection against FMDV, and published studies demonstrate the presence of FMDV-specific MHC class II-restricted responses in cattle and pigs (22, 24). CD4⁺ epitopes within both P1A and P1D proteins have recently been identified in cattle (23). We have recently reported the presence of FMDV-specific, MHC class I-restricted CD8⁺ T cell responses in cattle following infection or vaccination. Despite these observations, the significance of cell-mediated immune responses in protective immunity to FMDV remains unclear.

Cattle MHC (bovine leukocyte antigen [BoLA]) is relatively complex, with variable haplotypes expressing one, two, or three of the six classical class I genes (6, 15). At present, about 60 full-length validated cattle MHC class I cDNA sequences have been identified (www.ebi.ac.uk/ipd/mhc/bola), and the haplotypes commonly found in the Holstein breed are well characterized. We have previously identified amino acid motifs present in peptides binding to BoLA class I alleles N*02101, N*02201, and N*01301 (20). More recently, a number of Theileria parva CD8⁺ T cell epitopes presented through these and additional class I alleles have been described (25). Identification of such epitopes allows detailed analysis of cellular immune responses to vaccination and infection.

In the present study, we aimed to identify MHC class I-restricted CD8⁺ T cell epitopes within the FMDV capsid protein. Using a panel of overlapping peptides, we have identified a BoLA A31-restricted epitope that is similar in all FMDV serotypes.

MATERIALS AND METHODS

Viruses and vaccines.

Fowlpox recombinants expressing the polyprotein P1A, P1B, P1C, and P1D (VP4, VP2, VP3, and VP1, respectively) and including the nonstructural protein 2A of FMDV serotypes O and A were constructed as described previously (28). The O₁ Manisa and A₂₂/Iraq vaccines were prepared from antigen concentrate stored over liquid nitrogen, which is being held at a commercial facility as part of a new United Kingdom strategic reserve. In accordance with the European Pharmacopoeia Monograph (http://www.edqm.eu), this commercially produced oil adjuvant vaccine had been shown to have a 50% protective dose (PD₅₀) value of 18.

Animal experiments.

For the vaccination study, Holstein-Friesian cattle (Bos taurus) were vaccinated intramuscularly with one bovine dose of either serotype O or serotype A inactivated vaccine (Merial, Pirbright, United Kingdom) and boosted 11 weeks later. All of the animals used in this study (Table 1) were from a partially inbred herd, in which MHC class I haplotypes have been characterized at the level of expressed genes (16, 17). Details of cattle MHC haplotypes, alleles, and nomenclature can be found at the website www.ebi.ac.uk/ipd/mhc/bola/. Animals were typed by PCR using sequence-specific primers (PCR-SSP), reference strand conformational analysis (RSCA) (1, 6), or by indirect immunofluorescence (14, 21).

TABLE 1.

Animals used for studying CD8⁺ T cell responses to FMDV

Animal	MHC haplotypea	Vaccine FMDV serotype
FMD9	A18/A11	O1 Manisa
FMD17	A14/A31	O1 Manisa
FMD18	A10/A31	O1 Manisa
FMD46	A31/A31	A22/Iraq

^aThe A31 haplotype consists of two genes, N*02101 and N*02201.

For the infection study, six male Holstein calves aged between 6 and 9 months and weighing approximately 150 kg were exposed for 24 h to calves that had been infected 24 h previously by subepidermo-lingual injection with approximately 1 × 10⁵ 50% tissue culture infective dose (TCID₅₀) of FMDV serotype O UKG/34/2001. Clinical signs and rectal temperatures were recorded over an 8-day period, and the animals were maintained in isolation. The calves infected by injection were not used in this study.

Separation and preparation of lymphocyte subsets.

Heparinized venous blood was centrifuged at 3,000 × g over Histopaque 1086 (Sigma, Poole, United Kingdom), and the mononuclear cells were washed three times in phosphate-buffered saline (PBS). The cells were either used immediately or frozen in fetal calf serum (FCS) containing 10% dimethyl sulfoxide (DMSO) (Sigma, Poole, United Kingdom). CD14⁺ cells were purified by magnetic antibody cell sorting (MACS) using anti-human CD14⁺ microbeads (Miltenyi, Surrey, United Kingdom) (53) following the manufacturer's instructions. CD8⁺ α/β T cells were positively selected by incubating peripheral blood mononuclear cells (PBMCs) with the monoclonal antibody CC58 (37) and then with anti-mouse IgG1 microbeads (Miltenyi, Surrey, United Kingdom). Alternatively, CD3⁺ CD8⁺ CD4⁻ γδ TCR⁻ cells and CD3⁺ CD8⁻ CD4⁺ γδ TCR⁻ cells were separated using a FACSAria cell sorter and FACSDIVA v5 software (BD, Germany) using the following monoclonal antibodies: GB21A (anti-bovine γδ T cell receptor [TCR]; VMRD, Pullman, WA) (11), MM1A (anti-bovine CD3; VMRD) (12), and CC8 (anti-bovine CD4) (31, 40). Autologous antigen-presenting cells (APCs) were generated as described previously (28).

Synthetic peptides.

The protein sequence derived from FMDV serotype O UKG/35/2001 was used to synthesize 138 peptides, 15 amino acids in length and overlapping by 10 amino acids, spanning the structural polyprotein P1A, P1B, P1C, and P1D (VP4, VP2, VP3, and VP1) and including the nonstructural protein 2A as described previously (21). Peptide pools consisted of 10 peptides/pool. Short peptides (9-mers and 10-mers) were synthesized by fluorenylmethoxycarbonyl (fMOC) chemistry using an APEX 396 automated peptide synthesizer (Advanced ChemTech) or provided by Peptide Protein Research Ltd. (Wickham, United Kingdom). All peptides were purified by high-performance liquid chromatography (HPLC) and resuspended at 10⁻³ M in PBS with up to 1% DMSO.

BoLA MHC class-I P815 transfectants.

Mouse P815 cells expressing BoLA class I genes were prepared as described previously (19, 34). Transfectants were used as target cells to identify MHC restriction.

IFN-γ ELISPOT assay.

Cultured gamma interferon (IFN-γ) enzyme-linked immunospot (ELISPOT) assays were performed to assess the frequency of antigen-specific CD8⁺ T cell responses as described previously (28). APCs were prepared by either infecting CD14⁺ cells with recombinant fowlpox virus expressing the structural polyprotein P1 and the nonstructural protein P2A of FMDV (rFPV-FMD-P12A) (multiplicity of infection [MOI] of 5 PFU/cell) or directly loading CD14⁺ cells with peptides for 1 h at 37°C followed by two washes with PBS. Alternatively, mouse P815 cells expressing individual BoLA class I genes were either infected with rFPV-FMD-P12A (MOI of 5 PFU/cell) or directly loaded by adding peptides for 1 h at 37°C followed by two washes with PBS. APCs infected with wild-type fowlpox virus (FPV-wt) (MOI of 5 PFU/cell) or loaded with 13Y, a BoLA N*02201 self-peptide (S. A. Ellis, unpublished observations) were used as negative controls. A total of 5 × 10⁴ CD8⁺ T cells per well were stimulated with autologous virus-infected APCs (prepared as described above) for 7 days in 24-well plates (TPP Techno Plastic Products, Switzerland) and in the presence of 10 U/ml recombinant human interleukin 2 (IL-2) (Roche, Mannheim, Germany) at a ratio of 5 CD8⁺ T cells to 1 APC. CD8⁺ T cells were harvested, and their purity was analyzed by flow cytometry and restimulated in vitro in a MAHA 1540 (Millipore, United Kingdom) plate as described previously (18). Three wells in each plate were exposed to 20 μl of the mitogen concanavalin A at 1 μg/ml (Sigma, Poole, United Kingdom) as a positive control.

IFN-γ ELISPOT assay to assess ex vivo antigen-specific CD8⁺ T cells was performed exactly as described previously (28) using 2.5 × 10⁵ CD8⁺ T cells/well and using a ratio of 5 CD8⁺ T cells to 1 APC as described above.

Nonradioactive cytotoxicity assay.

To analyze antigen-specific killing of target cells, a nonradioactive cytotoxicity assay initially described for assessing virotoxin-induced cytotoxicity (27) was used. Target cells (1 × 10⁴ autologous APCs or BoLA P815 transfectants) were plated in a 96-well round-bottom plate (Costar) and either loaded with peptide or infected with recombinant fowlpox virus or wild-type virus (MOI of 5 PFU/cell) and incubated for 2 h at 37°C. The cells were washed once with PBS and resuspended in 100 μl of RPMI 1640 medium containing 5% FCS (Invitrogen), and equal volumes of effector cells (MACS- or fluorescence-activated cell sorting [FACS]-sorted CD8⁺ T cells cultured for 7 days as described above) were added to target cells at effector-to-target cell ratios of 10:1, 1:1, and 0.1:1. The plates were centrifuged at 200 × g for 1 min and incubated at 37°C in 5% CO₂ for 24 h. The cells were then washed once with PBS, and adherent cells removed with PBS-based cell dissociation buffer (Invitrogen) following the manufacturer's instructions. The cells were washed once in serum-free RPMI 1640 medium and resuspended in 250 μl of RPMI 1640 medium containing 4 mM calcein AM and 4 mM ethidium homodimer (EtD-1) (Invitrogen). NP-40 in PBS (final concentration of 0.5%) was added to wells containing target cells to generate dead-cell controls. The cells were incubated for 15 min in the dark and analyzed by flow cytometry.

Lymphocyte proliferation.

MACS- or FACS-sorted lymphocyte subsets were labeled with carboxy-fluorescein diacetate, succinimidyl ester (CFDA-SE) (Vybrant CFDA-SE Cell Tracer kit; Molecular Probes, Eugene, OR) as described previously (46). Antigen-presenting cells were generated as described above and loaded with peptide or infected with rFPV-FMDV-P12A in 6-well plates (Costar) in RPMI 1640 medium containing 5% FCS (Invitrogen). CFDA-SE-labeled lymphocytes were added at a ratio of 10 T cells to 1 APC and incubated at 37°C in 5% CO₂ for 5 days. The cells were harvested, fixed with 2% paraformaldehyde, and stained as described above with monoclonal antibodies CC8, CC58, and GB21A conjugated to Alexa Fluor 647 (Alexa Fluor 647 protein labeling kit; Invitrogen, Paisley, United Kingdom). Cell purity and proliferation were measured by flow cytometry.

Flow cytometry.

To analyze cytotoxicity, target and effector cells labeled with calcein AM or EtD-1 were gated independently using forward scatter (FSC) and side scatter (SSC) to differentiate cell populations. Specific killing was measured as described previously (27) using a BD FACSCalibur flow cytometer. Data were analyzed using CellQuest Pro v5.2 (BD, San Jose, CA).

Data analyses.

IFN-γ ELISPOT results are shown as the mean of triplicate spots/10⁶ CD8⁺ T cells ± standard deviations (SD). CD8⁺ T cell responses are considered to be positive when the means are at least twice the negative control or greater than 2 SD from the negative control, whichever is highest. The cytotoxicity of target cells is shown as the mean percentage of dead cells of three replicate spots ± SD. Calculation of descriptive statistics (geometric statistics and standard deviations), nonparametric statistical analyses, and graphs were generated using GraphPad Prism for Windows v5.01 (GraphPad, San Diego, CA).

Protein sequences.

FMDV protein sequences were obtained from NCBI (Table 2) and aligned with VectorNTi v10 (Invitrogen) using the ClustalW algorithm (54).

TABLE 2.

Overlapping peptides used to identify the putative epitope from peptide 252 (O/UK_790-804)

Peptide	Size	Sequence
Full-length peptide	15-mer	790LRTATYYFADLEVAV804
Overlapping peptides	9-mer	LRTATYYFA
	9-mer	RTATYYFAD
	9-mer	TATYYFADL
	9-mer	ATYYFADLE
	9-mer	TYYFADLEV
	9-mer	YYFADLEVA
	9-mer	YFADLEVAV
	10-mer	LRTATYYFAD
	10-mer	RTATYYFADL
	10-mer	TATYYFADLE
	10-mer	ATYYFADLEV
	10-mer	TYYFADLEVA
	10-mer	YYFADLEVAV

RESULTS

Ex vivo detection of FMDV-specific CD8⁺ T cells following infection and vaccination.

We have previously reported the identification of FMDV-specific CD8⁺ T cells following infection and vaccination using a cultured ELISPOT assay approach. To assess the magnitude of the response ex vivo, we analyzed samples taken at weekly intervals during the first 21 days following infection and vaccination. We identified FMDV-specific CD8⁺ T cells after 14 days in both cases (Fig. 1). The frequency of FMDV-specific CD8⁺ T cells in the vaccinated animals (0.020%) was similar to that in infected animals (0.016%). The responses quickly diminished and were undetectable by ex vivo ELISPOT assay at day 21 (data not shown). However, these responses could be amplified using a cultured ELISPOT assay, and subsequent experiments were performed by this method.

FIG. 1.

Infection with FMDV or vaccination using inactivated FMDV vaccine induces FMDV-specific CD8⁺ T cells. MACS-sorted CD8⁺ T cells from infected (A) or vaccinated (B) cattle were tested for antigen-specific IFN-γ release ex vivo 14 days postinfection or vaccination using APCs infected with rFPV-FMDV-P12A. Each symbol represents the value for an individual animal. The horizontal line indicates the median value for the group of animals. SFU, spot-forming units.

Induction of IFN-γ release by CD8⁺ T cells in response to peptide pools.

Animals FMD9, FMD17, and FMD18 were vaccinated and boosted against FMDV serotype O₁/Manisa, and animal FMD46 was vaccinated against serotype A₂₂/Iraq. Freshly isolated CD8⁺ T cells were stimulated in vitro with autologous APCs infected with a recombinant fowlpox virus expressing the structural polyprotein P1 and the nonstructural protein 2A of FMDV (rFPV-FMD-P12A). After a 7-day incubation, CD8⁺ T cells were harvested, and their activation was analyzed by an IFN-γ ELISPOT assay, in response to APCs loaded with pools of peptides. The N*2201 self-peptide 13Y was used as a negative control. An element present in pool 11 containing peptides 245 to 254 induced a statistically significant greater number of IFN-γ-producing cells (P < 0.005) than the negative-control peptide in animals FMD46, FMD17, and FMD18 (Fig. 2). The common MHC class I alleles in these three animals are BoLA N*02101 and N*02201 present in the A31 haplotype (Table 1). We then tested IFN-γ response to the 10 individual components of peptide pool 11 by using the cultured ELISPOT assay. Peptides 252 (LRTATYYFA DLEVAV; O/UK_790-804) and 253 (YYFADLEVAVKHEGN; O/UK_795-809) induced a statistically significant greater number of IFN-γ-producing CD8⁺ T cells (P < 0.005) than any of the other peptides or the negative controls (Fig. 3).

FIG. 2.

CD8⁺ T cells from FMDV-vaccinated animals recognize an epitope present in FMDV-derived peptide pools. CD8⁺ T cells from vaccinated cattle (FMD46, FMD17, FMD18, and FMD9) were cultured for 7 days with APCs infected with rFPV-FMDV-P12A, harvested, and then tested by a cultured ELISPOT assay for antigen-specific IFN-γ release in response to APCs loaded with peptide pools containing 10 15-mer peptides. Values are means plus standard deviations (SD) (error bars) of three replicate samples. Values which are statistically significantly different (P < 0.005) from the values for APCs loaded with N*02201 self-peptide 13Y are indicated by asterisks.

FIG. 3.

CD8⁺ T cells from FMDV-vaccinated animals recognize an epitope present within peptides 252 (LRTATYYFADLEVAV) and 253 (YYFADLEVAVKHEGN). APCs were loaded with 15-mer peptides, and autologous CD8⁺ T cells from vaccinated cattle were tested for antigen-specific IFN-γ release by a cultured ELISPOT assay. Values are means plus standard deviations (SD) (error bars) of three replicate samples. Values that are statistically significantly different (P < 0.005) from the values for APCs loaded with N*02201 self-peptide 13Y are indicated by asterisks.

Identification of a FMDV 9-mer CD8⁺ T cell epitope restricted by BoLA N*02201.

We predicted 13 possible 9-mer and 10-mer sequences (Table 2) derived from the 15-mer peptide 252 (O/UK_790-804). We tested the ability of these peptides to induce IFN-γ release from CD8⁺ T cells using a cultured ELISPOT assay (data not shown). We identified one peptide, peptide YYFADLEVA (O/UK_795-803), which showed a statistically significant response (P < 0.005) in the three animals expressing the A31 haplotype (Fig. 4). To confirm the restriction element responsible for the presentation of the putative epitope, CD8⁺ T cells were stimulated in vitro with autologous APCs infected with rFPV-FMD-P12A and harvested 7 days later, and IFN-γ release was measured by an ELISPOT assay using P815 transfectants expressing BoLA N*02101 or N*02201 loaded with 10⁻⁶ M FMDV O/UK_790-804 peptide (15-mer) or FMDV O/UK_795-803 peptide (9-mer) as targets. There was no response to P815 transfectants loaded with the 15-mer peptide (data not shown); however, there was a statistically significant response (P < 0.005) to P815-N*02201 loaded with FMDV O/UK_795-803 peptide compared with P815-N*02101 loaded with the same peptide (Fig. 5).

FIG. 4.

Identification of putative FMDV epitopes. Antigen-presenting cells were loaded with 10⁻⁹ M of peptides and IFN-γ released by autologous CD8⁺ T cells was measured by a cultured ELISPOT assay. 13Y is the N*02201 self-peptide. O/UK_790-804 is peptide 252 (15-mer). Values which are statistically significantly different (P < 0.005) from the values for 13Y are indicated by an asterisk.

FIG. 5.

CD8⁺ T cells recognizing O/UK_795-803 are restricted by N*02201. P815 cells expressing BoLA N*02101 or N*02201 were loaded with 10⁻⁶ M FMDV peptide O/UK_795-803, and IFN-γ release by CD8⁺ T cells was measured by a cultured ELISPOT assay. Values are means ± standard deviations (error bars) of three replicate samples. Values that are statistically significantly different (P < 0.005) from the value for cells expressing BoLA N*02101 are indicated by an asterisk.

To confirm that CD8⁺ αβ T cells and not CD4⁺ or γδ TCR⁺ cells were responding, we used fluorescence-activated cell sorting (FACS) to purify CD3⁺ CD8⁺ γδ TCR⁻ T cells and CD3⁺ CD4⁺ γδ TCR⁻ T cells from animals expressing the A31 haplotype. Following a 7-day in vitro peptide stimulation, we confirmed that both populations were >99% pure (Fig. 6B). CD8⁺ T cells proliferated when stimulated with both FMDV O/UK_790-804 and FMDV O/UK_795-803 peptides. In contrast, CD4⁺ T cells proliferated only upon stimulation with the 15-mer FMDV O/UK_790-804 peptide (Fig. 6A).

FIG. 6.

Peptide-specific proliferation of CD3⁺ CD8⁺ and CD3⁺ CD4⁺ T cells. CD3⁺ CD8⁺ γδ TCR⁻ and CD3⁺ CD4⁺ γδ TCR⁻ T cells from animals expressing the A31 haplotype were sorted by FACS and stimulated with 10⁻⁶ M FMDV O/UK_790-804, FMDV O/UK_795-803, control peptide 13Y, or concanavalin A (ConA). The purity of the T cells following stimulation was analyzed by flow cytometry (B), and proliferation was determined by the intensity of CFSE staining (A). The figure shows one representative sample of three independent experiments from one of three different animals.

The sequence of putative CD8⁺ T cell epitope O/UK_795-803 is relatively conserved among FMDV serotypes.

Animal FMD46 responded to peptide O/UK_795-803 despite having been vaccinated against serotype A₂₂/Iraq (Fig. 4). The corresponding sequence in this serotype differs from that of serotype O UKG/35/2001 by 2 amino acids. A peptide encoding the sequence for A₂₂/Iraq_796-804 (YYFSDLEIV) was synthesized (Table 3), and we tested the ability of cattle vaccinated with serotype O to recognize serotype A peptide and vice versa. Figure 7 shows that both peptides, FMDV O/UK_795-803 and A₂₂/Iraq_796-804, were recognized by all the animals expressing the A31 haplotype. P815-N*02201 transfectants were used to confirm MHC restriction as described above with similar results (data not shown).

TABLE 3.

FMDV-derived and control peptides

FMDV serotype or peptide	Peptide sequencea	Positions	GenBank accession no.
O1/BFS	YYFADLEIA	795-803	AAT01758
O/UK	.......V.	795-803	AAT01779
A22/Iraq	...S....V	796-804	AAT01707
ASIA-1	...S...V.	793-801	AAT01743
SAT-1	...S...V.	796-804	AAR92410
SAT-2	...C....T	796-804	AAT01792
SAT-3	...C...V.	794-802	AAT01796
Type C	...S.....	794-802	AAT01753
13Y (self-peptide)	HL.FPEPLFY

^aThe entire sequence for FMDV serotype O1/BFS is shown. Amino acids which are identical to the amino acid at that position in the O1/BFS sequence are indicated by a period. Only the amino acids which are different from the amino acid at that position in the O1/BFS sequence are shown.

FIG. 7.

IFN-γ responses to related FMDV epitopes. Antigen-presenting cells were loaded with 10⁻⁹ M concentrations of peptides, and IFN-γ produced by autologous CD8⁺ T cells was measured by a cultured ELISPOT assay. 13Y is the N*02201 self-peptide. The animals were vaccinated with FMDV serotype A (Vacc. A) or serotype O (Vacc. O). Values are means ± standard deviations (error bars) of three replicate samples. Values that are statistically significantly different (P < 0.005) from the value for 13Y are indicated by an asterisk.

Using a ClustalW protein alignment, we show that FMDV O/UK_795-803 is similar but not identical to the same region in representative isolates of all seven FMDV serotypes (Table 2). We therefore synthesized an additional five peptides, and using cultured ELISPOT assays, we tested the ability of CD8⁺ T cells from animals vaccinated with FMDV serotypes A and O to recognize autologous APCs loaded with synthetic peptides derived from different FMDV serotypes. This panel included the two most divergent peptides that were derived from the SAT-2 and SAT-3 peptides. Figure 8 shows that vaccinated cattle carrying the A31 haplotype have statistically significant responses to all peptides tested.

FIG. 8.

CD8⁺ T cells derived from animals vaccinated with FMDV serotype A or O respond to epitopes derived from heterologous FMDV serotypes. APCs were loaded with 10⁻⁹ M concentrations of peptides derived from representative samples of each FMDV serotype, and autologous CD8⁺ T cells from vaccinated cattle were tested for antigen-specific IFN-γ produced by a cultured ELISPOT assay. Values are means ± standard deviations (error bars) of three replicate samples.

N*02201-restricted CD8⁺ T cell epitope O/UK_795-803 induces killing of target cells.

Since IFN-γ release is not necessarily a measure of target killing, we used a nonradioactive cytotoxicity assay to analyze peptide-specific killing of target cells. In vitro-stimulated CD8⁺ T cells were cultured for 24 h in the presence of autologous APCs loaded with 10⁻⁶ M peptide FMDV O/UK_795-803, A₂₂/Iraq_796-804, 13Y, or rFPV expressing VP1 and nonstructural protein 2A (rFPV-FMD-VP12A), and target cell death was measured by flow cytometry. Cells from animals expressing the A31 haplotype showed statistically significant killing when loaded with FMDV O/UK_795-803 or A₂₂/Iraq_796-804 or infected with rFPV-FMD-VP12A but not with the control peptide (P < 0.005). In contrast, cells from animal FMD9 that do not express the A31 haplotype showed killing of target cells infected with rFPV-FMD-VP12A only (Fig. 9). Peptide-specific cytotoxicity and MHC restriction were confirmed using P815-BoLA transfectants with similar results (data not shown).

FIG. 9.

Killing of APCs loaded with FMDV peptides by autologous CD8⁺ T cells. APCs from FMDV-vaccinated animals were loaded with 10⁻⁶ M FMDV O/UK_795-803 (FMD9, FMD17, and FMD18), A₂₂/Iraq_796-804 (FMD46), or 13Y or infected with rFPV-FMD-P12A. Killing by autologous CD8⁺ T cells was measured by flow cytometry. The effector-to-target cell (E:T) ratio is shown on the x axes. Values are means ± SD (error bars) of three replicate samples.

DISCUSSION

We have previously reported the identification of virus-specific memory CD8⁺ T cells following infection with FMDV or vaccination with inactivated virus (28). In that study, we were unable to identify effector CD8⁺ T cells ex vivo at week 6 or later after infection or vaccination. In the current study, we identified FMDV-specific responses at day 14 after vaccination and infection, but these responses quickly diminished, and we were unable to detect circulating FMDV-specific CD8⁺ T cells above background at day 21 or later in both cases. Although the frequency of these responses was low (0.016 to 0.020% of circulating CD8⁺ T cells), it was significantly greater than that before infection (P = 0007) or before vaccination (P = 0.0297). However, these responses could be amplified 10-fold (0.156%) using the cultured ELISPOT assay approach.

It has been reported that in humans, the average CD8⁺ T cell response against influenza virus ex vivo is 0.3% but can be as low as 0.04% (45). In another study, in volunteers vaccinated with canarypox recombinant expressing HIV epitopes, the mean frequency of Gag-specific CD8⁺ T cells was 0.053%, expanding to an average of 9.8% following exposure to HIV (5). In the present study, we did not challenge the vaccinated animals, so we are unable to analyze the expansion of FMDV-specific CD8⁺ T cells following infection.

Although the role of CD8⁺ T cells in FMDV infection is not known, in the coxsackievirus mouse model of picornavirus infection, virus-specific CD8⁺ T cells have been shown to contribute to the control of virus infection, reducing virus titers up to 50-fold (30, 51). The reported presence of cytotoxic CD8⁺ T cells specific for poliovirus (57), another member of the family Picornavirinae, adds to the increasing evidence that CD8⁺ T cells may play a role in contributing to disease control in these related pathogens.

In this study, we aimed to identify BoLA class I-restricted CD8⁺ αβ T cell epitopes present in the structural proteins P1A, P1B, P1C, and P1D (also called VP4, VP2, VP3, and VP1, respectively) of FMDV. We identified a 9-mer peptide present within structural protein P1D that stimulates both IFN-γ production and cytotoxicity using CD8⁺ T cells from animals vaccinated with either inactivated FMDV serotype A or O vaccine. We identified a high level of conservation within this putative epitope in representative examples of all seven FMDV serotypes. With the exception of the P1A protein which has been shown to be relatively conserved, other T cell epitopes identified in structural proteins of FMDV are highly variable in different FMDV serotypes and isolates (23).

The O/UK_795-803 epitope is presented by the N*02201 MHC class I allele, expressed on the A31 haplotype, which was present in three of the four animals used in this study. Although a large number of peptides were screened, this was the only epitope detected, which suggests that it is immunodominant. This MHC haplotype often includes the class II allele BoLA DRB3*0701 (9), and this has previously been shown to present O/UK_790-804 (the 15-mer that includes O/UK_795-803) to CD4⁺ T cells (23). Although not widely reported, overlapping CD4⁺ and CD8⁺ T cell epitopes have been observed previously (14, 35, 42). To our knowledge, this is the first report of a BoLA class I-restricted CD8⁺ T cell epitope in FMDV.

All three A31⁺ animals recognized both O/UK_795-803 and A₂₂/Iraq_796-804 whether vaccinated with serotype O or A vaccine. An alignment of the putative epitope from 103 FMDV sequences (7) shows that this region is similar, but not identical, across FMDV serotypes (summarized in Table 3). Synthetic peptides derived from representative samples of these 7 serotypes were recognized by MACS-sorted CD8⁺ T cells from A31⁺ cattle vaccinated with serotype A and O vaccine (Fig. 7). Although we do not suggest that the epitope reported here is present in all FMDV isolates, an analysis of 103 complete genome sequences representing all FMDV serotypes shows that the putative epitope is similar (YYFXDXEXX) (8). ELISPOT and cytotoxicity responses to the two most divergent peptides, YYFCDLEIA (SAT-3) and YYFCDLEIT (SAT-2), are similar to those from peptides derived from the vaccine sequences. Previous analysis of the peptide-binding characteristics of N*02201 (16; also unpublished data) suggests that it binds well to peptides containing a phenylalanine (F) or similar residue at position 3. MHC class I molecules usually require peptides to be anchored additionally at the carboxyl-terminal end (18). The peptides predicted to have bound N*02201 in this study (Table 3) all have F at position 3, but they do not share a amino acid at the C terminus.

Our data indicate that O/UK_795-803 presented by N*02201 induces IFN-γ release, cytotoxicity, and proliferation of CD8⁺ T cells. Twenty-four months after vaccination, all animals carrying the A31 haplotype had O/UK_795-803-specific circulating CD8⁺ T cells with the phenotype CD3⁺/αβ CD8⁺/CD4⁻/γδ TCR⁻/CD45RO⁺ (data not shown), demonstrating that vaccination with the conventional FMDV vaccine induces circulating memory CD8⁺ T cells which, given the appropriate stimulus, can be expanded and are cytotoxic. Although the frequency of peptide-specific CD8⁺ T cells obtained in this study is low, the frequencies of memory responses to influenza virus HLA-restricted epitopes following infection or vaccination are similar (47, 52). It is important to stress that all of the CD8⁺ T cells detected in this study recognized the same epitope, and there is a strong likelihood that CD4⁺ responses in these animals would prove to be focused on the longer epitope, the O/UK_790-804 epitope (as indicated in Fig. 5).

Although the importance of T cell responses in protection to FMDV is still unclear, there is evidence to suggest that the quantity of IFN-γ produced postvaccination correlates with the capacity of the vaccinated animals to control pharyngeal virus replication (44), and there are also reports of FMDV-vaccinated sheep that were protected from infection in the absence of FMDV-specific antibodies at the time of challenge (10, 43). We have now established that vaccination of cattle with the current commercially available inactivated FMDV vaccine induces a CD8⁺ T cell response which is long lived and, depending on the genetic background of the animals, cross-reactive. The ability of FMDV-vaccinated animals to produce cross-reactive neutralizing antibodies has been demonstrated before in cattle, sheep (32, 49), pigs (36), and mice (4) using a number of vaccination strategies. Our data support previously published evidence which indicates that the genetic background of the animals is important in establishing cross-reactive immune responses to FMDV (33, 50). Effective vaccination strategies, along with appropriate breeding programs of economically important species which take into account the ability of the animals to mount appropriate immune responses, will be required to combat livestock infectious diseases in the future.