Absence of Tapasin Alters Immunodominance against a Lymphocytic Choriomeningitis Virus Polytope

Denise S. M. Boulanger, Roberta Oliveira, Lisa Ayers, Stephen H. Prior, Edward James, Anthony P. Williams and Tim Elliott
J Immunol January 1, 2010, 184 (1) 73-83; DOI: https://doi.org/10.4049/jimmunol.0803489

Abstract

Tapasin edits the peptide repertoire presented to CD8+ T cells by favoring loading of slow off-rate peptides on MHC I molecules. To investigate the role of tapasin on T cell immunodominance we used poxvirus viral vectors expressing a polytope of lymphocytic choriomeningitis virus epitopes with different off-rates. In tapasin-deficient mice, responses to subdominant fast off-rate peptides were clearly favored. This alteration of the CD8+ T cell hierarchy was a consequence of tapasin editing and not a consequence of the alteration of the T cell repertoire in tapasin-deficient mice, because bone marrow chimeric mice (wild-type recipients reconstituted with tapasin knockout bone marrow) showed the same hierarchy as the tapasin knockout mice. Tapasin editing is therefore a contributing factor to the phenomenon of immunodominance. Although tapasin knockout cells have low MHC I surface expression, Ag presentation was efficient and resulted in strong T cell responses involving T cells with increased functional avidity. Therefore, in this model, tapasin-deficient mice do not have a reduced but rather have an altered immune response.

CD8+ T cells recognize antigenic peptides presented by MHC I molecules on the surface of APCs or target cells. Folding of MHC I molecules and loading with an appropriate peptide require the assistance of a number of chaperones present in the endoplasmic reticulum (ER), including protein disulfide isomerase, calnexin, calreticulin, ERp57, tapasin (Tpn), and the TAP, the latter four forming the peptide loading complex. The peptide repertoire loaded on a specific MHC I allele will depend on 1) the sequence of the peptide that must contain the anchor residues specific for each MHC I allele to allow binding with a biologically significant affinity, 2) efficient proteasomal cleavage of newly synthesized proteins in the cytosol, 3) efficient transport of resulting peptides to the ER by TAP, 4) optimal trimming by the endoplasmic reticulum aminopeptidase associated with Ag processing in the ER, and 5) for most alleles Tpn editing (reviewed in Refs. 1–3). Crystal structures of MHC I and molecular dynamics simulation studies (4, 5) suggest that Tpn stabilizes the peptide-receptive MHC I conformation, allowing peptide exchange and increasing the probability of loading slow off-rate peptides. Confirming these predictions, Tpn has been shown to favor loading of slow off-rate variants of the OVA-SIINFEKL epitope expressed as minigenes in transfected cells (6). In the absence of Tpn, MHC I molecules can be loaded with fast off-rate suboptimal peptides that can be lost before reaching the cell surface. A larger number of empty or less stable suboptimally loaded MHC I are therefore present on the cell surface in Tpn-deficient cells. As a result, not only the peptide repertoire presented on the cell surface is modified but also the MHC I surface expression is reduced. Grandea et al. (7) and Garbi et al. (8) showed that the low surface expression of MHC I molecules resulted in 1) impaired positive and negative thymic T cell selection, resulting in a reduction of CD8+ T cell number in the thymus and periphery, and 2) reduced Ag presentation to T cells, resulting in a poor in vivo CTL response to influenza virus. However, the CTL response to vaccinia virus and lymphocytic choriomeningitis virus (LCMV) were normal, and Tpn knockout (KO) mice were able to clear a Listeria monocytogenes infection as efficiently as wild-type (wt) mice.

Using DNA vaccines, Thirdborough et al. (9) showed that the hierarchy of T cell responses generated in mice to variants of the H-2Kb restricted ovalbumin peptide SIINFEKL, administered individually, matched the kinetic stability observed by Howarth et al. (6) in a Tpn-dependent manner. However, these epitopes were overexpressed, and because each of the vaccines was administered in different mice, there was no direct competition between epitopes. Also, all of the variant epitopes were likely to be recognized by the same TCR. In this study, we used poxvirus vectors expressing a polytope of various LCMV epitopes of different off-rates to investigate the role of Tpn in the presentation of competing epitopes in an in vivo viral vaccine setting (competition between the epitopes expressed within the polytope and also with viral epitopes). We also wanted to determine whether the variation of level of T cell response observed in the SIINFEKL variants system is sufficient to alter immunodominance in a more complex biological system. We also wanted to further study the role of Tpn in vivo to understand the apparent paradox between the reduced Ag presentation level and a normal CTL response to some Ags as observed by Grandea et al. (7) and Garbi et al. (8). A limitation to the use of Tpn KO mice is the difficulty to differentiate the effect of the reduced number of CD8+ T cells and the reduced efficiency of Ag presentation on the T cell responses. We therefore generated bone marrow chimeric (BMC) mice (C57BL/6 mice transplanted with Tpn KO BM) that would contain a normal T cell repertoire selected on Tpn+/+ cells but Tpn-deficient APCs.

Materials and Methods

Cells

Primary chicken embryo fibroblasts were prepared using 10-d-old embryonated eggs from Rhode Island or Brown Leghorn chicken and grown in MEM containing 10% FCS (Globepharm, Surrey, UK). RMA-S, TAP2-deficient mouse T cell line, was cultured in RPMI 1640 (Lonza, Verviers, Belgium) supplemented with 10% FCS and 2 mM glutamine (Lonza,). Kb-transfected L cells (K89) and Db-transfected L cells (DbL), kindly provided by N. Shastri, California, were cultured in RPMI 1640 supplemented with 10% FCS, 2 mM glutamine, 1% HEPES, 1 mM pyruvate, and 50 μM β-mercaptoethanol.

Peptides

Peptides CSANNSHHYI, SGVENPGGYCL, FQPQNGQFI, KAVYNFATM (Cys41 was substituted into Met to avoid dimerization (10)), ISHNFCNL, and YTVKYPNL (Peptide Protein Research, Fareham, U.K.) and YAHINALEY, GLLSDHKSN, KAVYNFATCGI, AAFEFINSL, TSYKFESV, and VSLDYINTM (GL Biochem, Shangai, China) were synthesized by fluorenylmethoxycarbonyl chemistry and were >95% pure by HPLC and mass spectrometry.

Tetramers

Tetramers were manufactured essentially as described (11), with minor modifications. Mouse MHC class I and mouse β2-microglobulin (β2m) expression vectors were kindly provided by Prof. Dirk Busch (Munich, Germany), cloned into pET-3a (Novagen, Madison, WI) and transformed into BL21(DE3) CodonPlus RIPL (Stratagene, La Jolla, CA) cells.

β2m was refolded by slow dilution to a final concentration of 2 μM in the presence of 10 μM peptide. The heavy chain was then refolded in the presence of β2m and peptide to a final concentration of 1 μM. The refolded complex was then concentrated and applied to a HiLoad 26/60 Superdex 200 column (GE Healthcare, Buckinghamshire, England). The refolded complex was biotinylated by incubation at 16°C with 50 μM d-biotin and 1 μg/ml biotin protein ligase (Avidity, Aurora, CO). The biotinylated complex was then applied to a HiLoad 26/60 Superdex 200 column and then to a Mono-Q HR 5/5 column (GE Healthcare). The biotinylated monomer was tetramerized by incubation with 1:4 molar ratio of R-PE-labeled streptavidin (Invitrogen, Paisley, U.K.) at 4°C.

Viruses

Modified vaccinia virus Ankara (MVA)-multiple epitope 1 (ME1) and FWPV-ME1 (FP9 strain) recombinant viruses, provided by Oxxon Therapeutics/Oxford Biomedica (Oxford, U.K.), express a synthetic protein containing the following H-2b restricted LCMV epitopes: KAVYNFATCGI (gp33-43), FQPQNGQFI (np396-404), SGVENPGGYCL (gp276-286), YTVKYPNL (np205-212), and CSANNSHHYI (gp92-101). Recombinant viruses and MVA-wt and fowlpox virus (FWPV) were grown on chicken embryo fibroblasts in MEM supplemented with 2% FCS and titrated by plaque assay under agar overlay containing 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside or neutral red, respectively. Titers were expressed as PFU per milliliter.

Mice

Generation of Tpn-deficient mice, kindly provided by Luc Van Kaer, is described in Ref. 7. Homozygote mice, backcrossed six times with C57BL/6 mice on arrival, were backcrossed twice more in house, and the heterozygote offspring were intercrossed to generate F2 Tpn KO homozygotes (N8). Animal welfare and experimentation were conducted in accordance with the United Kingdom Coordinating Committee for Cancer Research (London, U.K.) guidelines with approval from the University of Southampton Ethical Committee and under U.K. Home Office License.

Generation of BMC mice

BMC mice were generated following (12). BM cells from 6- to 15-wk-old Tpn KO (N8) mice were harvested, and red blood cells were eliminated through a Ficoll cushion. To eliminate any mature T cells selected on the Tpn KO thymic stroma, donor Thy1.2+ BM mature T cells were depleted before the transfer through a MACS LD column containing CD90 (Thy1.2) microbeads (Miltenyi Biotec, Auburn, CA) following the manufacturer’s recommendations. Recipient C57BL/6 mice were irradiated with 10 Gy and reconstituted ∼4 h later by i.v. injection with 6 × 106 Thy1.2-negative cells resuspended in PBS. To avoid rejection of donor low MHC I Tpn KO cells by host NK cells, recipient mice also received an i.p. injection of 250 μg of anti-NK1.1 mAb diluted in PBS 2 h after the transfer and 4 d later. Chimeras were rested for 4–6 mo after reconstitution to allow for complete elimination of host-derived professional APCs.

Brefeldin A decay assay

RMA-S cells were incubated overnight at 26°C to maximize MHC I surface expression. After being washed, 5 × 105 cells were pulsed at different time intervals with peptides at a final concentration of 20 μM for 1 h at 26°C. After washing with medium, de novo transport of MHC I to the cell surface was blocked by the addition of BFA, and peptide-loaded RMA-S cells were incubated at 37°C to allow decay of unstable molecules. Cells were stained with Y3 or B22 primary monoclonal Abs (anti-H-2Kb and anti-H-2Db, respectively) and goat anti-mouse IgG-PE (Abcam, Cambridge, U.K.) secondary Abs to detect peptide-loaded MHC I molecules. Samples were analyzed by flow cytometry on a Beckton Dickinson (Oxford, U.K.) FACSCalibur flow cytometer, and data were analyzed with the Cell Quest software.

Immunization and intracellular cytokine (IFN-γ) staining

Mice were primed and boosted at a 2-wk interval by i.p. injection of 107 PFU of recombinant or wt virus. Spleens were harvested 7 d after the boost. Red blood cells were eliminated by centrifugation through a Ficoll cushion. Intracellular cytokine staining was performed using the BD Cytofix/Cytoperm Plus Fixation/Permeabilization kit with GolgiPlug (BD Biosciences, San Jose, CA). Splenocytes (2 × 106 splenocytes from immunized mice or 106 splenocytes from immunized mice + 106 splenocytes from naive mice) were incubated with peptides at the concentration indicated in the figure legends. After 2 h of incubation at 37°C, BFA (GolgiPlug) was added. After 2 h at 37°C, the plates were transferred at 4°C and stored overnight before staining. Dead cells were stained with ethidium monoazide bromide (Molecular Probes, Eugene, OR) simultaneously to block with Fcγ block (2.4G2; BD Pharmingen, San Diego, CA). Extracellular staining was performed for 30 min on ice with an anti-CD8a-APC (BD Pharmingen) Ab. Cells were permeabilized with Cytofix/Cytoperm and stained with an anti-IFN-γ-PE-Cy7 Ab (BD Pharmingen) for 30 min on ice. Cells were fixed in 1% paraformaldehyde solution and analyzed using a Beckton Dickinson FACSCanto and the Diva software. In some experiments, surface staining included an anti-TCRβ-PE Ab (eBiosciences, Hatfield, U.K.).

Tetramer staining and tetramer decay

Splenocytes were prepared as above. After the blocking step, the cells were stained with a predetermined saturating concentration of tetramer for 45 min on ice. Anti-CD8-APC Abs were added and incubated for a further 15 min. After three washings in FACS buffer (PBS, 2% FCS, 0.02% NaN3), the cells were either fixed in 1% PFA or incubated at room temperature for different periods of time in the presence of 100 μg/ml Y3 mAb. After a final wash, the cells were fixed, and samples were analyzed as described above. Total fluorescence was determined by multiplying the mean fluorescence intensity (MFI) of the gated tetramer-positive population by the number of events within the gate (13).

Hybridoma generation

TSYK- and KAV-specific hybridoma were generated according to Sanderson and Shastri (14). Splenocytes harvested from C57BL/6 mice, immunized as described above, were restimulated with each peptide in vitro for 4 d in the presence of 20 U/ml IL-2 before being fused with BWZ.36/CD8α fusion partners. The SGVE-specific hybridoma was kindly provided by Dr. Marcus Groettrup, Konstanz, Germany.

Epitope presentation assay

Splenocytes harvested from C57BL/6 or Tpn KO mice were stimulated with 10 μg/ml Escherichia coli LPS (Sigma-Aldrich, St. Louis, MO) for 48 h. Approximately 105 or 106 cells per well of a 96 well plate were then infected for 0, 2, 4, or 6 h with MVA-wt or MVA-ME1 at a multiplicity of infection (moi) of 3 PFU per cell. After being washed to eliminate any residual virus, appropriate T cell hybridoma (105 cells per well) were added and incubated at 37°C O/N. After freeze–thawing, IL-2 production was measured by ELISA.

Reverse phase HPLC fractionation of peptides

LPS-blasts (108 cells) were infected with recombinant MVA at an moi of 3 PFU per cell for 6 h. Surface peptides were eluted by incubation for 1 min in 0.13 M citrate phosphate buffer (pH 3.1) supplemented with 1% BSA. After centrifugation, the supernatants were filtered through 10-kDa cut-off Microcon Ultracell YM-10 filters (Millipore, Bedford, MA). Filtrates were fractionated by HPLC on a C18 column (Vydac, Kinesis, St Neots, Cambs, U.K.) using a 0.1% trifluoroacetic acid 15–50% acetonitrile/water gradient. Fractions were collected in 96-well plates and dried by vacuum centrifugation.

T cell activation assay

DbL or K89 were added to the dried fractionated samples and cocultured with the appropriate T cell hybridoma. The lacZ activity induced upon hybridoma activation was measured by the conversion of chlorophenolred-β-d-galactopyrannoside (CPRG) (Roche, Mannheim, Germany) to chlorophenol red by its absorbance at 595 nm with 655 nm as the reference wavelength (14).

Results

Immune response profile to competing vaccine epitopes in the presence of Tpn

To investigate the role of Tpn on CD8+ T cell immunodominance in an in vivo setting where multiple epitopes are expressed simultaneously, we used poxvirus vectors expressing a polytope containing five LCMV epitopes of different off-rates, described previously (15–17) as either immunodominant (gp33-43, KAV, Db- and Kb-restricted (18), and np396-404, FQPQ, Db-restricted) or subdominant (gp276-286, SGVE, Db-restricted; np205-212, YTVK, Kb-restricted; and gp92-101, CSAN, Db-restricted). The relative kinetic stability (off-rate) of all of the peptides used in this study was determined in a BFA decay experiment (Fig. 1). KAV9 (KAVYNFATM) and FQPQ peptides had a much slower off-rate than SGVE, CSAN, and KAV11 (KAVYNFATCGI) presented on H-2Db. The slowest off-rate LCMV peptide presented on H-2Kb was also KAV9 followed by KAV11, ISHN, and then YTVK.

FIGURE 1.
FIGURE 1.

Decay of peptide-loaded MHC I molecules on RMA-S cells. Db-restricted peptides (top panel) and Kb-restricted peptides (bottom panel) were loaded on RMA-S cells. MHC I decay was assessed by Db or Kb surface staining after different incubation times at 37°C. Results are plotted as percentages of the maximum MFI obtained at time 0 for each peptide. Dashed lines represent the controls. TSYK Kb-restricted peptide was used as negative control for Db stabilization, and DMSO was used as a negative control (top panel); SIINFEKL (SIIN) was used as a positive control for high-affinity Kb-restricted peptide, and DMSO was used as a negative control (bottom panel). TSYK and YAHI are MVA-specific epitopes used in this study to monitor the T cell response against the MVA vector (see text).

The hierarchy of CD8+ T cell response against the different LCMV epitopes was determined by ex vivo intracellular cytokine (IFN-γ) staining of splenocytes harvested from immunized C57BL/6 mice. A single injection with either the fowlpox or the vaccinia MVA recombinant viruses (FWPV-ME1 or MVA-ME1, respectively) did not generate a detectable response. C57BL/6 mice were therefore primed with the fowlpox recombinant virus and boosted with the MVA recombinant virus. T cells specific for three of the LCMV epitopes (KAV, CSAN, and SGVE) were detected (Fig. 2A). An immunodominant response against the KAV epitope was consistently observed. Most mice also developed a detectable response against the CSAN epitope. A response against the SGVE epitope was infrequently observed. Surprisingly, the Db-restricted immunodominant FQPQ epitope was recognized poorly in this system, probably due to an inefficient proteasomal cleavage from the polytope polypeptide compared with the LCMV nucleoprotein. No response against the Kb-restricted subdominant YTVK epitope was ever observed. The overall response profile against the three recognized epitopes (higher T cell response against KAV than against CSAN and SGVE) (Fig. 2A) therefore matched the kinetic stability hierarchy determined by the BFA decay experiment (Fig. 1), consistent with our previous findings (9).

FIGURE 2.
FIGURE 2.

T cell response against an LCMV polytope. A, Hierarchy of T cell response against an LCMV polytope in C57BL/6 mice immunized with 107 PFU of FWPV-ME1 and 107 PFU of MVA-ME1 at 2-wk intervals. Percentage of CD8+ live lymphocytes expressing IFN-γ determined by intracellular cytokine staining after ex vivo stimulation with the LCMV-specific peptides indicated on the x-axis. ISHN and DMSO were used as negative controls. Results from 19 mice tested in 10 individual experiments after stimulation with 4 or 5 × 10−6 M peptide are shown. B, Hierarchy of response in Tpn KO mice. Results from nine mice from six individual experiments are shown. C, Comparison of the T cell response in C57BL/6, Tpn KO, and BMC mice after ex vivo stimulation using 5 × 10−6 M concentrations of the respective peptides. One representative experiment out of five is shown.

Immune response profile in the absence of Tpn

To determine whether the immunodominance of CD8+ T cell responses seen following recombinant virus immunization was under the control of Tpn, we immunized Tpn KO mice with the same prime–boost strategy. In Tpn KO mice, the response against SGVE and CSAN increased, whereas the response against the KAV epitope decreased compared with the response observed in C57BL/6 mice. The hierarchy of CD8+ T cell response became SGVE > CSAN > > KAV (Fig. 2B, 2C). In some mice, the response to CSAN was as high or sometimes slightly higher than the response to SGVE, but the response to KAV was always much lower. Thus, the role of Tpn in favoring presentation and therefore T cell response of slow off-rate peptides was also confirmed in this vaccine setting and was potent enough to induce a shift in immunodominance hierarchy.

Immune response against MVA-specific epitopes

After prime–boost immunization we noted that the percentage of CD8+ T cells with low CD8 surface expression (activated lymphocytes) was always higher in Tpn KO mice than that in C57BL/6 mice, showing that the Tpn KO mice are highly responsive. Because the percentage of CD8+ T cells responding to the LCMV epitopes was low compared with the percentage of activated CD8+ T cells, we suspected that this might be a consequence of responses to competing epitopes from the viral vector and therefore tested the CD8+ T cell response against five vaccinia-specific epitopes: YAHINALEY (A47L 157–165, Db-restricted), GLLSDHKSN (J6R 543–551, Kb-restricted) (19), TSYKFESV (B8R 20–27, Kb-restricted), AAFEFINSL (A47L 138–146, Kb-restricted), and VLSDYINTM (A19L 47–55, Kb-restricted) (20). After two immunizations with 107 PFU of MVA-wt, C57BL/6 and Tpn KO mice reacted strongly and consistently against the TSYK epitope and, in some experiments, against YAHI (Fig. 3A). The TSYK epitope is present in the vaccinia B8R protein, which has no homologue in FWPV. To determine the importance of TSYK in the immune response generated after a heterologous immunization protocol with the recombinant viruses, groups of two C57BL/6 mice were immunized either with two injections of MVA-ME1 or with two injections of both vectors in either order (MVA-ME1 first or FWPV-ME1 first). Both mice injected twice with MVA-ME1 responded strongly against TSYK (22% of IFN-γ–producing CD8+ splenocytes) but very weakly against the LCMV epitopes (maximum 0.9% against KAV9). Both mice primed with MVA-ME1 and boosted with FWPV-ME1 showed a lower response against TSYK (4% or 5%) but did not respond better against the LCMV epitopes (0.8% maximum). Both mice primed with FWPV-ME1 and boosted with MVA-ME1 reacted better against TSYK (6–9%) and also responded to the LCMV epitopes (3–4% against KAV9). TSYK was therefore immunodominant over the LCMV epitopes even in heterologous immunizations (Fig. 3B) and represents one of the major components of the immune response in both C57BL/6 and Tpn KO mice (compare the level of response against TSYK in Fig. 3C and the response against the LCMV epitopes in the same mice in Fig. 2C).

FIGURE 3.
FIGURE 3.

T cell response against MVA-specific epitopes. A, CD8+ T cell response against five vaccinia-specific epitopes. T cell responses in C57BL/6 and Tpn KO mice immunized twice with 107 PFU of MVA-WT (m7m7) were tested by intracellular cytokine staining against YAHINALEY, GLLSDHKSN, TSYKFESV, AAFEFINSL, and VLSDYINTM. KAV9 was used as a negative control. B, Comparison of the IFN-γ response of C57BL/6 mice immunized with two injections of MVA-ME1 recombinant virus (M7M7) or primed with MVA-ME1 and boosted with FWPV-ME1 (M7F7) or primed with FWPV-ME1 and boosted with MVA-ME1 (F7M7) (all immunizations with 107 PFU of virus). Both mice primed with FWPV-ME1 reacted to a high extent against TSYK and gave the strongest response against the LCMV epitopes. This immunization strategy was therefore used throughout this study. C, Comparison of the IFN-γ response of C57BL/6, Tpn KO, and BMC mice primed with FWPV-ME1 and boosted with MVA-ME1 and stimulated ex vivo with 10−9 M TSYK. The T cell response was expressed as the percentage of CD8+ live lymphocytes expressing IFN-γ. D, Same as C expressed as numbers of CD8+ live lymphocytes expressing IFN-γ per spleen.

The higher response of Tpn KO mice against TSYK (Fig. 3C) as well as against CSAN and SGVE (Fig. 2C) was not an artifact due to the low percentage of CD8+ T cells found in naive Tpn KO mice, because immunized mice also had a higher total number of IFN-γ–producing CD8+ T cells per spleen (Fig. 3D). This is remarkable, because naive Tpn KO mice have half the percentage of CD8+ T cells compared with C57BL/6 mice. Therefore, at most a similar or more likely a lower number of naive T cell precursors in Tpn KO mice compared with C57BL/6 mice must have been primed, activated, and proliferated more efficiently until homeostasis constraints were reached.

Generation of BMC mice and comparison of their immune response to C57BL/6 and Tpn KO mice

The difference in the hierarchy of response against the LCMV epitopes seen in Tpn KO mice could be due either to a difference in the hierarchy of epitope presentation or to a difference in the T cell repertoire altered by impaired thymic selection. To assess the respective influence of the editing function of Tpn and the altered T cell repertoire on the T cell response of Tpn KO mice, we generated Tpn KO chimeric mice by transplanting BM cells harvested from Tpn KO mice into irradiated C57BL/6 recipient mice. All hematopoietic cells, including dendritic cells, in the BMC mice should derive from the Tpn KO BM transplant and should not express Tpn. The percentage of CD8+ splenocytes in the BMC mice was normal, and Tpn could not be detected in either splenocytes or BM dendritic cells by RT-PCR (Supplemental Table I, Supplemental Fig. 1A). MHC I levels on BMC hematopoietic cells were the same as those on Tpn KO cells and were 10–20% of the surface expression seen on C57BL/6 cells (Supplemental Fig. 1B), confirming the donor-origin of circulating lymphocytes and restoration of normal positive selection in BMC mice.

After heterologous prime–boost immunization, BMC mice developed a strong response against SGVE and CSAN and a reduced response against KAV as in Tpn KO mice (Fig. 2C). BMC mice also developed a strong response against TSYK after either homologous prime–boost immunizations with MVA-wt or heterologous prime–boost immunizations with both recombinant viruses (Fig. 3C). Because BMC mice share the same APCs as Tpn KO mice but a T cell repertoire educated on high MHC I like C57BL/6 mice, the BMC T cell profile showed that the hierarchy of T cell response was essentially a consequence of Tpn editing and not a difference in the T cell repertoire.

Quantitation of relative amounts of peptides presented on the cell surface

Because TSYK has a similar off-rate as KAV9 (Fig. 1) and on the basis of the published data on Tpn function, we would have expected the response to TSYK to be lower in Tpn KO mice than that in C57BL/6. However, the response against TSYK in Tpn KO mice was consistently higher. It is not clear why TSYK is not presented poorly like KAV in Tpn KO mice. This could be due to a difference in abundance, processing efficiency, duration of exposure (21), or pathway of presentation.

To answer this question and to confirm Tpn editing of the peptide repertoire in this model, we attempted to quantitate the relative amount of each peptide presented on the surfaces of infected cells. C57BL/6 and Tpn KO LPS-blasts infected for 2, 4, or 6 h were cocultured with T cell hybridomas specific for TSYK (3E1), KAV presented by H-2Kb (D9 or G10), KAV presented by H-2Db (5E2), or SGVE. We used LPS-blasts to maximize MHC I expression. Although LPS treatment increased MHC I expression on both C57BL/6 and Tpn KO LPS-blasts, the latter still expressed less surface MHC I than C57BL/6 LPS-blasts (Fig. 4A). Fig. 4B shows that KAV was presented efficiently by H-2Kb on C57BL/6 LPS-blasts but was not detected on Tpn KO LPS-blasts (Fig. 4B). Interestingly, KAV was presented only very weakly by H-2Db on C57BL/6 LPS-blasts (note the difference of sensitivity of the 5E2 H-2Db-restricted hybridoma compared with the D9 H-2Kb-restricted hybridoma in Fig. 4C) and not at all on Tpn KO blasts. Because SGVE was neither detected on C57BL/6 nor on Tpn KO LPS-blasts in this assay (data not shown), we therefore infected a larger number of LPS-blasts, acid-eluted peptides from MHC I surface molecules, and fractionated them by HPLC. Using this method, we could show that approximately two times more SGVE was recovered from Tpn KO than C57BL/6 LPS-blasts (6 × 10−10 M compared with 3 × 10−10 M, respectively) (Fig. 5B). When the relative level of H-2Db MHC I expression on C57BL/6 compared with Tpn KO LPS-blasts was taken into consideration (Fig. 4A), this translated into approximately six times more on Tpn KO on a per-cell basis. KAV was detected only in one extract from C57BL/6 LPS-blasts (data not shown), but the amount was too low to be quantified (maximum lower than 10−10M). These data therefore confirm that the subdominant epitope SGVE, which has a faster off-rate from H-2Db than the immunodominant epitope KAV, is presented better in Tpn KO than in wt cells, whereas the reverse is true for KAV. These semiquantitative data therefore correlate well with the response hierarchies observed following recombinant virus infection.

FIGURE 4.
FIGURE 4.

Epitope presentation on LPS-blasts. A, C57BL/6+/+ and Tpn KO−/− LPS-blasts were stained for H-2Db (left panel) and H-2Kb (right panel) surface molecules. Samples were analyzed by flow cytometry using a Beckton Dickinson FACSCalibur flow cytometer, and data were analyzed with FlowJo 7.5. Filled histograms show the unstained controls. B, Approximately 105 LPS-blasts were infected with MVA-ME or MVA-wt at an moi of 3 PFU per cell for 2 h (light gray histograms), 4 h (dark gray), or 6 h (black histograms). Uninfected controls are shown as white histograms. LPS-blasts were cocultured with the following hybridoma cells: 3E1 hybridoma specific for TSYK (left panels), D9 hybridoma specific for KAV presented by H-2Kb (middle panels), 5E2 hybridoma specific for KAV presented by H-2Db molecules (right panels). Hybridoma activation could not be tested by CPRG staining, because the MVA-ME recombinant virus expresses β-galactosidase. Therefore, overnight IL-2 production was measured by ELISA. (C) KAV peptide titrations. Approximately 105 DbL or K89 cells were incubated with two step dilutions of peptide and 105 hybridoma cells (D9 on K89 or 5E2 on DbL). After overnight incubation hybridoma activation was revealed with CPRG. (D) TSYK peptide titration. As in C, K89 cells were incubated with peptide and with the 3E1 hybridoma.

FIGURE 5.
FIGURE 5.

HPLC fractionation of acid-eluted peptides from LPS-blasts. A, B , Acid-eluted peptides from 108 LPS-blasts infected with MVA-ME at an moi of 3 PFU/ml were fractionated by HPLC (top panels). To confirm the identity of the eluted peptides, 10 pmol of synthetic peptides were run through the HPLC under identical conditions (bottom panels). Samples were assayed with the 3E1 hybridoma (A) or the SGVE hybridoma (B) and K89 (A) or DbL (B) as APCs. C, Titration curves of 3E1 and SGVE hybridoma reactivity after incubation with the appropriate APCs and serial dilutions of the respective peptides. After overnight incubation, hybridoma activation was revealed with CPRG.

Fig. 4B also shows that C57BL/6 LPS-blasts presented slightly more TSYK than Tpn KO LPS-blasts. In acid elution experiments, ∼50 times more TSYK was recovered from C57BL/6 than from Tpn KO LPS-blasts (maximum concentrations of 2 × 10−8 M and 4 × 10−10 M, respectively, average from 4 experiments, one of them shown in Fig. 5A). This increase could not be accounted for by the fivefold increase of H-2Kb MHC I expression on C57BL/6 LPS-blasts compared with Tpn KO LPS-blasts. We conclude therefore that like KAV the high-affinity peptide TSYK is also subject to Tpn-mediated editing but that its high abundance is sufficient to breach a threshold for priming even in the absence of tapasin.

Quantitative aspects of the CD8+ T cell response

The intracellular cytokine staining assays described above were performed after stimulation with a standard concentration of 4 or 5 × 10−6 M of peptide. However, as shown in Fig. 1, the peptides used in this study have different affinities. Therefore we tested responses at saturating concentrations of peptide (Fig. 6A, 6B). The intensity of the response depended on the concentration of peptide used in the ex vivo stimulation. The response to KAV and TSYK was already maximal at 10−6 M, but the response to CSAN and SGVE could be increased at higher concentrations, reaching a maximum at 10−4 M. Interestingly, Tpn KO CD8+ T cells seemed to be activated by lower concentrations of peptides (Fig. 6A, 6B). Because Tpn KO T cells have been selected on a low MHC I background, we hypothesized that T cells with a higher avidity would have been selected in Tpn KO mice compared with C57BL/6 mice. We therefore compared the T cell responses of C57BL/6, Tpn KO, and BMC mice by intracellular cytokine staining using decreasing concentrations of peptides for the ex vivo stimulation, referred to as the functional avidity. The concentrations necessary to obtain 50% of the maximum IFN-γ response varied greatly between peptides (Fig. 6C–E) and were lower in Tpn KO than those in C57BL/6 for TSYK (30-fold), CSAN (4-fold), and SGVE (5-fold). This implied that TSYK-, CSAN-, and SGVE-specific T cells have a higher avidity in Tpn KO mice compared with those in C57BL/6 mice. To confirm that the lower peptide requirement for activating Tpn KO T cells was not due to the different quality of APCs, responders from immunized mice were mixed with naive C57BL/6 and Tpn KO splenocytes in the presence of different concentrations of peptide. Although Tpn KO splenocytes improved the efficiency of the stimulation compared with C57BL/6 (Fig. 6F) a difference in functional avidity was still clearly observed between C57BL/6 and Tpn KO responders. Similarly, the functional avidity of the BMC T cells was higher than that in C57BL/6 but lower than that in Tpn KO mice although they share the same Tpn−/− APCs (Fig. 6C). Therefore the higher Tpn KO T cell response and functional avidity was only partly due to the enhanced capacity of Tpn−/− APCs to load with extracellular peptides.

FIGURE 6.
FIGURE 6.

Functional T cell avidity. A, Comparison of the CD8+ T cell response in C57BL/6 and Tpn KO mice after heterologous immunizations with FWPV-ME1 and MVA-ME1 (F7M7). Splenocytes were stimulated ex vivo with different concentrations of LCMV peptides. TSYK was used as an MVA-specific control. B, Comparison of the CD8+ T cell response against the TSYK MVA-specific epitope in C57BL/6, Tpn KO, and BMC mice immunized twice with 107 PFU of wt MVA (m7m7) and titration of the response with decreasing concentrations of peptide. C, Functional avidity of TSYK-specific CD8+ T cells harvested from C57BL/6, Tpn KO, and BMC mice determined by intracellular cytokine staining in the presence of decreasing amounts of peptide added to 2 × 106 splenocytes from immunized mice (physiological conditions). To estimate functional avidity of CD8+ T cells regardless of their different numbers in the different types of mice, the percentage of CD8+ live lymphocytes expressing IFN-γ after stimulation with a range of peptide dilutions was expressed as the percentage of the maximum value. D, E, Functional avidity of CSAN- and SGVE-specific T cells compared in C57BL/6 and Tpn KO. F, To normalize the quality of the APCs in the different types of mice, 106 splenocytes from naive C57BL/6 or Tpn KO mice (less optimally loaded Tpn KO APCs would load exogenous stabilizing peptides more efficiently than C57BL/6 APCs) were added to 106 splenocytes from immunized mice. Addition of naive Tpn KO splenocytes to the C57BL/6 and to a lesser extent to the Tpn KO responder splenocytes from immunized mice reduced peptide requirement to achieve 50% of the maximum response.

Another explanation for the increased functional avidity in Tpn KO mice could result from the selection of T cells bearing a higher TCR affinity. TCR avidity of C57BL/6 and Tpn KO TSYK-specific T cells was therefore compared in a “tetramer decay” assay. Splenocytes from immunized C57BL/6 and Tpn KO mice were stained with a saturating concentration of fluorescent tetrameric complexes of H-2Kb:TSYK, determined by titration (Fig. 7A, 7B), washed, and incubated for increasing periods of time at room temperature in the presence of anti-H-2Kb blocking Ab (Fig. 7C, 7D). Using this method, we found that there was no difference in TCR avidity between C57BL/6 and Tpn KO mice (Fig. 7D).

FIGURE 7.
FIGURE 7.

Tetramer staining and TCR avidity. A, TSYK-Tetramer titration on splenocytes from naive (N) or F7M7 immunized C57BL/6 and Tpn KO mice. Tetramer concentration was expressed as percentage v/v. Results are expressed as the percentage of tetramer-positive cells within the CD8+ lymphocyte population. B, TSYK-Tetramer titration expressed as a percentage of the maximum response. C, TSYK-Tetramer decay from Tetramer+ CD8+ cells of one typical C57BL/6 and one Tpn KO mouse. Staining is shown after 0, 3, or 30 min of incubation at room temperature in the presence of blocking Y3 Ab. Staining without tetramer and staining of naive mice with tetramer are shown as controls. D, Tetramer decay from Tetramer+ CD8+ T cells from immunized C57BL/6 and Tpn KO mice expressed as the percentage of the maximum total fluorescence corresponding to time 0.

Increased functional avidity could also result from an increased expression of the TCR, increased expression of adhesion and activation molecules upon activation, or a decreased activation threshold (22) to counterbalance the low MHC I expression on APCs involved in T cell thymic selection or in the adoptive immune response. In immunized mice, the MFI of the TCR surface expression on the CD8+ population was usually lower in Tpn KO mice (Fig. 8A), reflecting the higher percentage of activated cells in Tpn KO mice. However, the MFI of the IFN-γ–expressing cells was identical in both types of mice. Interestingly, the TCR expression was slightly lower on naive Tpn KO CD8+ T cells (p < 0.0001) (Fig. 8B). Therefore, the increased functional avidity of Tpn KO T cells is not due to an increase of TCR expression. In addition, we have been unable to identify a difference in activation potential between wt and Tpn KO cells using markers such as differences in Lck phosphorylation and calcium mobilization, and this is a subject of ongoing investigation.

FIGURE 8.
FIGURE 8.

Comparison of TCR expression in naive and immunized C57BL/6 and Tpn KO mice. A, TCR surface staining MFI of splenocytes from C57BL/6 (black) or Tpn KO (gray) immunized mice stimulated in vitro with TSYK, or DMSO as a negative control. Cells were gated on CD8+ lymphocytes after stimulation with TSYK (CD8+ TSYK) or DMSO (CD8+ DMSO) or gated on CD8+ cells producing IFN-γ after stimulation with TSYK (IFN+ TSYK). B, Comparison of the TCR surface staining MFI of CD8+ lymphocytes from naive C57BL/6 or Tpn KO mice. Immunized controls were included in the same experiment.

Discussion

Previous work using variants of the OVA 323-331 peptide SIINFEKL characterized by different off-rates, expressed as minigenes in fibroblasts transfected with DNA plasmids, showed that in vitro the presence of Tpn favors presentation of the slow off-rate peptides (6). This hierarchy correlated with the in vivo CD8+ T cell response hierarchy obtained against the same variants administered to C57BL/6 mice individually as DNA immunogens expressing either of the variant epitopes (9). We show that in C57BL/6 mice immunized with poxvirus recombinants expressing an LCMV polytope the T cell response against the slow off-rate KAV epitope was higher than those against CSAN and SGVE, whereas in Tpn KO mice the T cell response was higher against the faster off-rate CSAN and SGVE epitopes, resulting in a different immunodominance hierarchy. The difference in the T cell response hierarchy in Tpn KO mice could either be due to the presence of Tpn-deficient APCs and lack of Tpn editing or be due to a different T cell repertoire affected by an abnormal thymic T cell selection. The former interpretation is favored by our demonstration that BMC mice showed a T cell response hierarchy similar to that of Tpn KO mice. Furthermore, peptide elution from C57BL/6 and Tpn KO LPS-blasts showed that Tpn KO cells present approximately six times more SGVE peptide than C57BL/6 cells. Tpn editing of the peptide repertoire is therefore an important mechanism underpinning T cell immunodominance. Processing, transport, and affinity for MHC have all been identified as important factors influencing immunodominance. All of these factors converge at the point of peptide loading and repertoire editing.

It was surprising that the response against KAV was lower than the response against CSAN and SGVE in Tpn KO mice (and BMC mice). In the absence of Tpn function, it is likely that loading is simply controlled by peptide on-rate. The peptide repertoire should therefore depend on the relative abundance of available peptides (above a certain association-rate threshold). Because KAV, CSAN, and SGVE are expressed from a polytope, we expected similar amounts of peptides to be generated and presented on the cell surface. However, differential processing has been shown to lead to presentation of very different amounts of peptides. KAV can be presented by both H-2Db and H-2Kb MHC I molecules on LCMV-infected cells. However, our data (Fig. 4B) show that KAV expressed from the MVA recombinant vector is presented mostly on H-2Kb molecules. Due to the very low abundance, we have not been able to quantitate the amount of KAV presented on either H-2Kb or H-2Db molecules, but our data (not shown) suggest that KAV is presented at even lower amounts than SGVE. It is therefore possible that KAV is indeed processed less efficiently than CSAN and SGVE. Another explanation for the low response against KAV would be the deletion of KAV-specific T cells from the T cell repertoire of Tpn KO mice, but this is unlikely because BMC mice also showed a reduced response against KAV. Whatever the reason for the poor T cell response against KAV in the absence of Tpn, efficient presentation, at least on H-2Kb molecules (Fig. 4B), was restored in the presence of Tpn, contributing to the alteration of the immunodominance hierarchy.

Interestingly the response against TSYK, an MVA-specific epitope of a similar half-life as KAV on H-2Kb molecules, was actually immunodominant in both C57BL/6 and Tpn KO mice and was higher in Tpn KO mice than that in C57BL/6 mice. Tpn editing therefore shapes the hierarchy of T cell response only to a certain extent, because some epitopes such as TSYK seem to not be affected by the presence or absence of Tpn. We hypothesized that the truncated B8R protein (soluble IFN-γ receptor), which yields the TSYK epitope, might be expressed at higher levels than the polytope, thereby competing effectively with the LCMV epitopes in the presence or absence of Tpn (i.e., superabundance may overcome differences observed in the presence or absence of Tpn when the Ag is limiting). The fact that we recovered more TSYK than KAV supports this theory. Together with the fact that TSYK-specific T cells have a higher functional avidity than KAV-specific T cells (data not shown), the amount of TSYK presented on the surfaces of both C57BL/6 and Tpn KO APCs would be sufficient to induce an efficient T cell response, whereas the amount of KAV in Tpn KO mice might be under the threshold level necessary to induce an efficient T cell response. Therefore, Tpn editing and T cell properties (such as avidity, burst kinetics, interclonal competition, and so on) contribute together to shape the immunodominance hierarchy.

As a consequence of the low MHC I expression on Tpn KO cells, positive thymic selection is impaired (7) and results in a lower CD8 T cell number. We hypothesized that selected T cells in Tpn KO mice would have a higher avidity to compensate for the lower MHC I density. This hypothesis was supported by our observation that Tpn KO T cell responses were consistently higher than those in C57BL/6; T cells were activated with lower peptide concentrations, and their functional avidity was higher than those of BMC and C57BL/6 T cells supplemented with Tpn KO APCs. However, TCR avidity, determined by tetramer decay assay, was identical in Tpn KO and C57BL/6 T cells. Furthermore, the level of TCR expression was the same in Tpn KO IFN-γ–producing T cells from immunized mice. TCR expression was not increased either on naive Tpn KO T cells; on the contrary, naive Tpn KO T cells had a slightly lower TCR surface expression than C57BL/6 T cells that might be essential to avoid self-reactivity.

Tpn is genetically conserved in species as distant as fish (23), birds (24), and mammals (25), suggesting that it has an evolutionary survival advantage. The appearance in evolution of receptor diversity provided both the opportunity to recognize many pathogens in a specific way but also provided the means for extensive autoreactivity. One function of Tpn might therefore be to ensure some focusing of the presentable repertoire to minimize autoreactivity. Although Tpn KO mice do not show any obvious signs of autoimmunity, these mice are naive or sacrificed after one or two rounds of immunization. It would be interesting to know whether multiple successive infections with different Ags would generate cross-reactive T cells against self-Ags. Also, it is noteworthy that the cumulative CD8+ T cell response of Tpn KO mice against TSYK, KAV, CSAN, and SGVE (tested at saturating concentrations) accounts for 100% of the CD8+ lymphocyte population, whereas it accounts for ∼30% in C57BL/6 mice at most (Fig. 6A). Therefore, a substantial number of naive T cells would still be available after successive infections in C57BL/6 mice but maybe not in Tpn KO mice.

Our results show that in the absence of Tpn Ag presentation on reduced MHC I still occurs efficiently and leads to a strong T cell response, favoring a response toward subdominant epitopes. Therefore, Tpn KO mice do not have a reduced but rather have an altered immune response. This can explain why Grandea et al. (7) and Garbi et al. (8) observed a reduced Ag presentation and CTL response against immunodominant epitopes such as flu NP366 but still a normal response against overall vaccinia virus, LCMV, and Listeria.

Tpn can therefore have an important role in the immune response and consequently in vaccination efficiency, especially for peptides expressed at low concentrations. Downregulation of MHC I expression on human tumors is usually a marker of poor prognosis and can result from a downregulation of MHC I itself, β2m, TAP, or Tpn expression (26–31). Our results predict that presentation of immunodominant epitopes would be reduced on such tumor cells but that presentation of subdominant epitopes would be increased. Re-expressing Tpn in Tpn-deficient tumors has been shown to increase MHC I surface expression in vitro, and C57BL/6 mice injected with cancer cells infected with an adenovirus vector expressing TAP1 and Tpn had a greater survival rate than mice receiving untreated tumor cells (32). Vaccination in the absence of Tpn could be useful to induce or boost an immune response against subdominant epitopes. Examples where this approach would be beneficial are, for instance, avoiding or reducing competition by immunodominant epitopes from a viral vector, increasing or broadening the immune response against subdominant epitopes included in a polytope vaccine, or targeting immune escape Tpn-deficient tumors by vaccination against subdominant epitopes. This could possibly be achieved in the future by pharmacological inhibition of Tpn in transfused APCs, with either drugs or inhibitory molecules such as adenovirus E19 protein (33).

Acknowledgments

We thank Oxxon Therapeutics/Oxford Biomedica for providing us with the recombinant FWPV and MVA. We thank D. Busch for providing us with plasmids H2-Kb-pET3a and mβ2m-pET3a, L. Van Kaer for providing us with Tpn knockout mice, M. Groettrup for providing us with the SGVE hybridoma, N. Shastri for providing us with K89, DbL, and BWZ.36/CD8α, and the animal house staff for their excellent assistance.

Disclosures The authors have no financial conflicts of interest.

Footnotes

  • This work was supported by Cancer Research UK Grant C7056/A2739.

  • The online version of this article contains supplemental material.

  • Abbreviations used in this paper:

    β2m
    β2-microglobulin
    BMC
    bone marrow chimeric
    CPRG
    chlorophenol red-β-D-galactopyrannoside
    ER
    endoplasmic reticulum
    FWPV
    fowlpox virus
    KO
    knockout
    LCMV
    lymphocytic choriomeningitis virus
    ME1
    multiple epitope 1
    MFI
    mean fluorescence intensity
    moi
    multiplicity of infection
    MVA
    modified vaccinia virus Ankara
    Tpn
    tapasin
    wt
    wild-type.

  • Received October 17, 2008.
  • Accepted October 28, 2009.

References

  1. ↵
    1. Elliott T.,
    2. A. Williams
    . 2005. The optimization of peptide cargo bound to MHC class I molecules by the peptide-loading complex. Immunol. Rev. 207: 89–99.
    OpenUrlCrossRefPubMed
    1. Hammer G. E.,
    2. T. Kanaseki,
    3. N. Shastri
    . 2007. The final touches make perfect the peptide-MHC class I repertoire. Immunity 26: 397–406.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Purcell A. W.,
    2. T. Elliott
    . 2008. Molecular machinations of the MHC-I peptide loading complex. Curr. Opin. Immunol. 20: 75–81.
    OpenUrlCrossRefPubMed
  4. ↵
    1. Sieker F.,
    2. S. Springer,
    3. M. Zacharias
    . 2007. Comparative molecular dynamics analysis of tapasin-dependent and -independent MHC class I alleles. Protein Sci. 16: 299–308.
    OpenUrlCrossRefPubMed
  5. ↵
    1. Zacharias M.,
    2. S. Springer
    . 2004. Conformational flexibility of the MHC class I α1-α2 domain in peptide bound and free states: a molecular dynamics simulation study. Biophys. J. 87: 2203–2214.
    OpenUrlCrossRefPubMed
  6. ↵
    1. Howarth M.,
    2. A. Williams,
    3. A. B. Tolstrup,
    4. T. Elliott
    . 2004. Tapasin enhances MHC class I peptide presentation according to peptide half-life. Proc. Natl. Acad. Sci. USA 101: 11737–11742.
    OpenUrlAbstract/FREE Full Text
  7. ↵
    1. Grandea A. G., III.,
    2. T. N. Golovina,
    3. S. E. Hamilton,
    4. V. Sriram,
    5. T. Spies,
    6. R. R. Brutkiewicz,
    7. J. T. Harty,
    8. L. C. Eisenlohr,
    9. L. Van Kaer
    . 2000. Impaired assembly yet normal trafficking of MHC class I molecules in Tapasin mutant mice. Immunity 13: 213–222.
    OpenUrlCrossRefPubMed
  8. ↵
    1. Garbi N.,
    2. P. Tan,
    3. A. D. Diehl,
    4. B. J. Chambers,
    5. H. G. Ljunggren,
    6. F. Momburg,
    7. G. J. Hämmerling
    . 2000. Impaired immune responses and altered peptide repertoire in tapasin-deficient mice. Nat. Immunol. 1: 234–238.
    OpenUrlCrossRefPubMed
  9. ↵
    1. Thirdborough S. M.,
    2. J. S. Roddick,
    3. J. N. Radcliffe,
    4. M. Howarth,
    5. F. K. Stevenson,
    6. T. Elliott
    . 2008. Tapasin shapes immunodominance hierarchies according to the kinetic stability of peptide-MHC class I complexes. Eur. J. Immunol. 38: 364–369.
    OpenUrlCrossRefPubMed
  10. ↵
    1. Vabulas R. M.,
    2. H. Pircher,
    3. G. B. Lipford,
    4. H. Häcker,
    5. H. Wagner
    . 2000. CpG-DNA activates in vivo T cell epitope presenting dendritic cells to trigger protective antiviral cytotoxic T cell responses. J. Immunol. 164: 2372–2378.
    OpenUrlAbstract/FREE Full Text
  11. ↵
    1. Altman J. D.,
    2. P. A. Moss,
    3. P. J. Goulder,
    4. D. H. Barouch,
    5. M. G. McHeyzer-Williams,
    6. J. I. Bell,
    7. A. J. McMichael,
    8. M. M. Davis
    . 1996. Phenotypic analysis of antigen-specific T lymphocytes. Science 274: 94–96.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Sigal L. J.,
    2. K. L. Rock
    . 2000. Bone marrow-derived antigen-presenting cells are required for the generation of cytotoxic T lymphocyte responses to viruses and use transporter associated with antigen presentation (TAP)-dependent and -independent pathways of antigen presentation. J. Exp. Med. 192: 1143–1150.
    OpenUrlAbstract/FREE Full Text
  13. ↵
    1. Wang X. L.,
    2. J. D. Altman
    . 2003. Caveats in the design of MHC class I tetramer/antigen-specific T lymphocytes dissociation assays. J. Immunol. Methods 280: 25–35.
    OpenUrlCrossRefPubMed
  14. ↵
    1. Sanderson S.,
    2. N. Shastri
    . 1994. LacZ inducible, antigen/MHC-specific T cell hybrids. Int. Immunol. 6: 369–376.
    OpenUrlAbstract/FREE Full Text
  15. ↵
    1. Gallimore A.,
    2. T. Dumrese,
    3. H. Hengartner,
    4. R. M. Zinkernagel,
    5. H. G. Rammensee
    . 1998. Protective immunity does not correlate with the hierarchy of virus-specific cytotoxic T cell responses to naturally processed peptides. J. Exp. Med. 187: 1647–1657.
    OpenUrlAbstract/FREE Full Text
    1. van der Most R. G.,
    2. K. Murali-Krishna,
    3. J. L. Whitton,
    4. C. Oseroff,
    5. J. Alexander,
    6. S. Southwood,
    7. J. Sidney,
    8. R. W. Chesnut,
    9. A. Sette,
    10. R. Ahmed
    . 1998. Identification of Db- and Kb-restricted subdominant cytotoxic T-cell responses in lymphocytic choriomeningitis virus-infected mice. Virology 240: 158–167.
    OpenUrlCrossRefPubMed
  17. ↵
    1. Kotturi M. F.,
    2. B. Peters,
    3. F. Buendia-Laysa Jr.,
    4. J. Sidney,
    5. C. Oseroff,
    6. J. Botten,
    7. H. Grey,
    8. M. J. Buchmeier,
    9. A. Sette
    . 2007. The CD8+ T-cell response to lymphocytic choriomeningitis virus involves the L antigen: uncovering new tricks for an old virus. J. Virol. 81: 4928–4940.
    OpenUrlAbstract/FREE Full Text
  18. ↵
    1. Hudrisier D.,
    2. M. B. Oldstone,
    3. J. E. Gairin
    . 1997. The signal sequence of lymphocytic choriomeningitis virus contains an immunodominant cytotoxic T cell epitope that is restricted by both H-2Db and H-2Kb molecules. Virology 234: 62–73.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Mathew A.,
    2. M. Terajima,
    3. K. West,
    4. S. Green,
    5. A. L. Rothman,
    6. F. A. Ennis,
    7. J. S. Kennedy
    . 2005. Identification of murine poxvirus-specific CD8+ CTL epitopes with distinct functional profiles. J. Immunol. 174: 2212–2219.
    OpenUrlAbstract/FREE Full Text
  20. ↵
    1. Pasquetto V.,
    2. H. H. Bui,
    3. R. Giannino,
    4. C. Banh,
    5. F. Mirza,
    6. J. Sidney,
    7. C. Oseroff,
    8. D. C. Tscharke,
    9. K. Irvine,
    10. J. R. Bennink,
    11. et al
    . 2005. HLA-A*0201, HLA-A*1101, and HLA-B*0702 transgenic mice recognize numerous poxvirus determinants from a wide variety of viral gene products. J. Immunol. 175: 5504–5515.
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. Radcliffe J. N.,
    2. J. S. Roddick,
    3. F. K. Stevenson,
    4. S. M. Thirdborough
    . 2007. Prolonged antigen expression following DNA vaccination impairs effector CD8+ T cell function and memory development. J. Immunol. 179: 8313–8321.
    OpenUrlAbstract/FREE Full Text
  22. ↵
    1. Slifka M. K.,
    2. J. L. Whitton
    . 2001. Functional avidity maturation of CD8+ T cells without selection of higher affinity TCR. Nat. Immunol. 2: 711–717.
    OpenUrlCrossRefPubMed
  23. ↵
    1. Sültmann H.,
    2. B. W. Murray,
    3. J. Klein
    . 2000. Identification of seven genes in the major histocompatibility complex class I region of the zebrafish. Scand. J. Immunol. 51: 577–585.
    OpenUrlCrossRefPubMed
  24. ↵
    1. Frangoulis B.,
    2. I. Park,
    3. F. Guillemot,
    4. V. Séverac,
    5. C. Auffray,
    6. R. Zoorob
    . 1999. Identification of the Tapasin gene in the chicken major histocompatibility complex. Immunogenetics 49: 328–337.
    OpenUrlCrossRefPubMed
  25. ↵
    1. Paulsson K. M.
    2004. Evolutionary and functional perspectives of the major histocompatibility complex class I antigen-processing machinery. Cell. Mol. Life Sci. 61: 2446–2460.
    OpenUrlCrossRefPubMed
  26. ↵
    1. Atkins D.,
    2. A. Breuckmann,
    3. G. E. Schmahl,
    4. P. Binner,
    5. S. Ferrone,
    6. F. Krummenauer,
    7. S. Störkel,
    8. B. Seliger
    . 2004. MHC class I antigen processing pathway defects, ras mutations and disease stage in colorectal carcinoma. Int. J. Cancer 109: 265–273.
    OpenUrlCrossRefPubMed
    1. Dissemond J.,
    2. T. Kothen,
    3. J. Mörs,
    4. T. K. Weimann,
    5. A. Lindeke,
    6. M. Goos,
    7. S. N. Wagner
    . 2003. Downregulation of tapasin expression in progressive human malignant melanoma. Arch. Dermatol. Res. 295: 43–49.
    OpenUrlCrossRefPubMed
    1. Facoetti A.,
    2. R. Nano,
    3. P. Zelini,
    4. P. Morbini,
    5. E. Benericetti,
    6. M. Ceroni,
    7. M. Campoli,
    8. S. Ferrone
    . 2005. Human leukocyte antigen and antigen processing machinery component defects in astrocytic tumors. Clin. Cancer Res. 11: 8304–8311.
    OpenUrlAbstract/FREE Full Text
    1. Ferris R. L.,
    2. T. L. Whiteside,
    3. S. Ferrone
    . 2006. Immune escape associated with functional defects in antigen-processing machinery in head and neck cancer. Clin. Cancer Res. 12: 3890–3895.
    OpenUrlAbstract/FREE Full Text
    1. Ogino T.,
    2. N. Bandoh,
    3. T. Hayashi,
    4. N. Miyokawa,
    5. Y. Harabuchi,
    6. S. Ferrone
    . 2003. Association of tapasin and HLA class I antigen down-regulation in primary maxillary sinus squamous cell carcinoma lesions with reduced survival of patients. Clin. Cancer Res. 9: 4043–4051.
    OpenUrlAbstract/FREE Full Text
  31. ↵
    1. Ogino T.,
    2. H. Shigyo,
    3. H. Ishii,
    4. A. Katayama,
    5. N. Miyokawa,
    6. Y. Harabuchi,
    7. S. Ferrone
    . 2006. HLA class I antigen down-regulation in primary laryngeal squamous cell carcinoma lesions as a poor prognostic marker. Cancer Res. 66: 9281–9289.
    OpenUrlAbstract/FREE Full Text
  32. ↵
    1. Lou Y.,
    2. G. Basha,
    3. R. P. Seipp,
    4. B. Cai,
    5. S. S. Chen,
    6. A. R. Moise,
    7. A. P. Jeffries,
    8. R. S. Gopaul,
    9. T. Z. Vitalis,
    10. W. A. Jefferies
    . 2008. Combining the antigen processing components TAP and Tapasin elicits enhanced tumor-free survival. Clin. Cancer Res. 14: 1494–1501.
    OpenUrlAbstract/FREE Full Text
  33. ↵
    1. Bennett E. M.,
    2. J. R. Bennink,
    3. J. W. Yewdell,
    4. F. M. Brodsky
    . 1999. Cutting edge: adenovirus E19 has two mechanisms for affecting class I MHC expression. J. Immunol. 162: 5049–5052.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section