Minor histocompatibility antigen
Stem cell transplantation
Acute lymphoblastic leukemia
Single nucleotide polymorphism
Lymphoblastoid cell line
Human minor histocompatibility antigens (mHag) are target antigens of the graft-versus-leukemia response observed after allogeneic HLA-identical stem cell transplantation. We previously defined the molecular nature of the B cell lineage-specific mHag HB-1. The CTL epitope was identified as the decamer peptide EEKRGSLHVW presented in the context of HLA-B44. The HB-1 antigen is encoded by a locus of yet unknown function on chromosome 5q32. A single nucleotide polymorphism within this locus results in an amino acid change from histidine (H) to tyrosine (Y) at position P8 within the CTL epitope. Based on genomic information, we have developed a PCR-RFLP assay to perform HB-1 typing at the DNA level. We determined that the allelic frequency for the H and Y variant is 0.79 and 0.21, respectively. From these data, we calculated that the expected recipient disparity between HLA-B44-matched sibling pairs for HB-1H is 2.8%, whereas recipient disparity for HB-1Y is expected to be 12.4%. Therefore, we addressed whether the HB-1Y peptide is reciprocally immunogenic. We revealed that both peptide variants bind equally efficient to HLA-B44 molecules and that the H/Y substitution has noinfluence on formation of epitope precursor peptides by 20 S proteasome-mediated degradation. More directly, CTL recognizing the naturally presented HB-1Y peptide could be generated from a HB-1H homozygous donor using peptide-pulsed dendritic cells. Using a set of synthetic structurally related peptide variants, we found that the H/Y substitution has a major impact on TCR recognition by CTL specific for either of the HB-1 allelic homologues. HB-1 is the first human mHag described that induces bi-directional allogeneic CTL responses that may contribute to a specific graft-versus-leukemiaresponse following allogeneic stem cell transplantation.
Disparities in minor histocompatibility antigens (mHag) between HLA-matched donor/recipient pairs can lead to vigorous T cell responses involved in graft-versus-host disease (GVHD) and graft-versus-leukemia (GVL) reactivity after allogeneic stem cell transplantation (SCT) 1. mHag are peptides derived from polymorphic cellular proteins presented at the cell surface by MHC molecules 2, 3. More than 50 different mHag loci have been defined among inbred strains of mice 4. However, the number of mHag in humans is unknown and knowledge about the structure, genetics and tissue expression of their encoding genes is scarce. Identification of human mHag has been achieved using specific T cells isolated from SCT recipients 5–14. By this way a number of male-specific HY mHag 5–10 and four autosomal mHag, known as HA-1, HA-2, HB-1 and HA-8have been identified 11–14. Human mHag characterized so far have shown to be unidirectional due to differential MHC binding or antigen processing 12, 14, 15. Reciprocal CTL responses have been found for mouse mHag 16–19.
We previously reported that a mismatch for HB-1 could induce donor-derived CTL reactive against B lineage acute lymphoblastic leukemia (ALL) cells after HLA-matched SCT 20. By cDNA library screening, using HB-1-specific CTL as reagent, the antigen was identified as peptide EEKRGSLHVW presented in the context of HLA-B44 molecules (i.e. HB-1H.B44 epitope) 13. The HB-1-coding cDNA exists in two allelic variants, designated as HB-1H and HB-1Y. A single nucleotide polymorphism (SNP) results in an amino acid exchange from histidine (H) to tyrosine (Y) at position P8 of the HB-1H.B44 epitope. Interestingly, expression of HB-1 is restricted to cells of the B lymphoid lineage including tumor cells of B lineage ALL patients 13. Therefore, disparity for HB-1 might be involved in selective GVL reactivity against B lymphoid malignancies after SCT.
In this study, we characterized the genomic structure of mHag HB-1 that allowed us to develop a typing method at the genomic level. Using this method, we determined the allele and genotype distribution for the H and Y variant. From these data, we calculated that the expected recipient disparity for HB-1H between HLA-B44-matched sibling pairs is around 2.8%. The reciprocal recipient disparity for HB-1Y is significantly higher which is 12.4%. Therefore, we investigated whether HB-1 induces bi-directional CTL responses. We demonstrate that the HB-1Y allelic variant is equally immunogenic.
2.1 Genomic structure of the HB-1 coding sequence
Typing for mHag HB-1 can be performed by functional assays using CTL MP1. However, to determine HB-1 disparity by typing at the genomic level, we unraveled the genomic structure surrounding the SNP. In the Human Genome Project the HB-1 coding sequence has been mapped to contig NT-006489 of ∼1 Mb on chromosome 5q32, which we confirmed by FISH analysis. To define exon-intron boundaries, we compared consensus sequences between the previously cloned HB-1 cDNA fragment (AF103884) and contig NT-006489. This analysis revealed that the HB-1H antigenic peptide is encoded by two exons (provisionally named A and B) which are separated by an intron of ∼8.2 kb (Fig. 1). A splice donor site could be identified thirteen nucleotides after the first HB-1 peptide coding triplet and a splice acceptor site could be identified just before the second peptide coding region (Fig. 2). The C/T SNP is located at position 9 in exon B. These data reveal that the first four residues, EEKR, of the HB-1 peptide sequence are encoded by exon A and the remaining six residues including the polymorphic histidine residue, GSLHVW, are derived from exon B.
2.2 Genomic typing for the HB-1 H and Y polymorphism
Knowledge of the genomic organization surrounding the HB-1 SNP allowed us to design a PCR-RFLP method for genomic typing. PCR primers (represented by horizontal arrows in Fig. 1) were chosen to amplify a DNA fragment containing a part of the intronic sequence and a part of the allele-specific sequence of exon B. Amplification with these primers resulted in a 153-bp fragment and subsequent NlaIII restriction analysis rendered an easily interpretable H/Y allele assignation. To determine the population frequency of the HB-1 alleles and genotypes, we typed a large number of unrelated individuals. In total, we have typed 162 individuals of which 32 were Dutch and 130 Spanish (Table 1). The genotype distribution in the Spanish cohort was not statistically different from the genotype distribution in the Dutch cohort (Chi-square test, p=0.20). The overall genotype frequency was 63% HH, 33% HY and 4% YY, and the H and Y alleles were present at frequencies of 0.79 and 0.21, respectively (Table 1). Between siblings, the frequency of recipient disparity for mHag can be calculated from the formula p×q3 + 0.75×p2×q2, where p and q represent allele frequencies 21. The predicted recipient disparity for HB-1H is 2.8% (p=0.79 and q=0.21), whereas recipient disparity for HB-1Y is expected to be 12.4% (p=0.21 and q=0.79).
|Spanish (n=132)||Dutch (n=30)||Total (n=162)|
|H||77 %||87 %||79 %|
|Y||23 %||13 %||21 %|
|HH||60 %||77 %||63 %|
|HY||36 %||20 %||33 %|
|YY||4 %||3 %||4 %|
2.3 Differential recognition of the HB-1H peptide by CTL MP1
Immunogenicity of mHag is due to differential antigen processing and presentation or due to TCR discrimination of the variant peptides. The HB-1H peptide is recognized by CTL MP1, whereas the HB-1Y peptide is not, even when exogenously added to HLA-B44+ target cells (Fig. 3a). This suggests that P8 is a dominant residue for CTL recognition. To address this, we investigated which amino acids of the HB-1H.B44 epitope are engaged in HLA binding and/or TCR interactions. Variant peptides were synthesized in which the native amino acids were replaced by an alanine (A) residue. Not surprisingly, substitutions at HLA-B44 anchor positions 2 and 10 diminished CTL recognition (Fig. 3b). Furthermore, substitutions at position 3, 4, 5, 6, 7 and 8 completely abolished CTL recognition. These data show that the polymorphic amino acid at position P8 is one of the key determinants within the HB-1H.B44 epitope for recognition by CTL MP1.
2.4 Effect of HB-1 polymorphism on MHC class I binding
To determine whether the HB-1Y peptide is reciprocally antigenic, we first assessed whether the H/Y substitution influences peptide binding to HLA-B44 molecules. Known anchors for optimal binding to HLA-B44 are a glutamic acid (E) at P2 and a Y, phenylalanine (F) or tryptophan (W) at P10. The H/Y polymorphism is at P8, suggesting that binding is probably not influenced. To determine whether both peptides bind to HLA-B44 molecules, we measured their relative binding affinity using a competition-based peptide-binding assay (Fig. 4). Peptide EEKRGSLHVW and EEKRGSLYVW inhibited binding of a Fl-labeled EBNA3c.B44 peptide by 50% (IC50 value) at 8.5 μM and 4.0 μM, respectively. These data demonstrate that both HB-1 peptides bind efficiently to HLA-B44 molecules and, therefore, the H/Y amino acid change does not result in differential peptide binding to HLA-B44.
2.5 Effect of HB-1 polymorphism on proteasome cleavage
To address whether antigen processing allows antigenicity of the HB-1Y peptide, we investigated whether the H/Y substitution changes proteasome-mediated cleavage of epitope precursor peptides. Therefore, we performed in vitro proteasome digestions with 22-mer synthetic peptides containing either the HB-1H or HB-1Y peptide sequence flanked by its natural amino acids (Table 2). Proteasome-mediated cleavage generally determines the exact C-terminal residue of the CTL epitope, whereas the N terminus is usually elongated requiring trimming by amino peptidases 22–24. The 20 S proteasome-mediated proteolysis of both peptides yielded a heterogeneous mixture of peptides. Digestion of the HB-1H peptide resulted in the generation of three epitope-containing fragments: i.e. the 15-mer QPECREEKRGSLHVW, the 11-mer REEKRGSLHVW, and the 9-mer EKRGSLHVW. The proteasome precisely determines the C-terminal W of the HB-1H epitope. The minimal 10-mer epitope, EEKRGSLHVW, could not be found in this digest. The most likely relevant epitope precursor peptide is the 11-mer fragment, which requires N-terminal trimming to generate the minimal epitope. However, this fragment can also be recognized as efficiently as the minimal epitope without further trimming (Fig. 5). CTL MP1 also recognizes the 9-mer peptide fragment when loaded on Raji.B*;4403 cells, but a high peptide concentration is required for lysis (Fig. 5). The 20 S proteasome-mediated digestion of the HB-1Y peptide resulted in formation of the following fragments: the 15-mer QPECREEKRGSLYVW, the 11-mer REEKRGSLYVW, and the 5-mer SLYVW. Again the proteasome properly cleaves after the C-terminal W. Furthermore, the Y amino acid does not introduce an extra cleavage site indicating that the H/Y substitution within mHag HB-1 does not lead to differential processing by the proteasome.
2.6 The HB-1Y peptide forms an immunogenic CTL epitope
To prove that mHag HB-1 is reciprocally immunogenic, we attempted to generate CTL with specificity for the HB-1Y peptide. Therefore, CD8+ T cells isolated from a HB-1H homozygous donor were stimulated with HB-1Y.B44 peptide-pulsed mature DC. After three rounds of stimulation, peptide specificity of the initiated T cell culture was tested. The generated T cell culture (CTL KU) produced 905 pg/ml IFN-γ upon stimulation with HB-1Y peptide-pulsed Raji.B*;4402 cells, whereas upon stimulation with unpulsed or HB-1H peptide-pulsed Raji.B*;4402 cells IFN-γ production was 3 and 31 pg/ml, respectively. After repeated stimulation with peptide-pulsed autologous EBV-LCL, >95% of CTL KU were positive for TCRBV21S3 (data not shown). As shown in Fig. 6a and b, CTL KU recognized Raji.B*;4402 cells pulsed with the HB-1Y peptide more efficiently than pulsed with the HB-1H peptide. Moreover, CTL KU also recognized the naturally processed and presented HB-1Y.B44 peptide on the cell surface of Raji.B*;4402 cells stably transfected with the HB-1Y allele (Raji.B*;4402.HB-1Y). These results are in line with our findings that the HB-1Y peptide is not destructed by proteasomal cleavage and binds to HLA-B44 molecules. In contrast to CTL MP1 which recognizes only the HB-1H peptide (Fig. 3a), the anti-HB-1Y CTL KU cross-reacts with the HB-1H antigen when naturally processed on Raji.B*;4402.HB-1H cells (Fig. 6a). This difference in specificity may be due the lower avidity of the in vitro-induced anti-HB-1Y CTL that has been induced using a high concentration of synthetic peptide. Nevertheless, our data clearly demonstrate that disparity for mHag HB-1 can result in bi-directional CTL epitope formation.
2.7 Effect of HB-1 polymorphism on TCR discrimination
The TCR of the HB-1H-specific CTL MP1 includes the TCRBV6S1 β-chain, whereas the TCR of the HB-1Y-specific CTL KU is composed of the TCRBV21S3 β-chain. To investigate whether differential recognition of the HB-1H and HB-1Y epitope is the result of TCR discrimination, we have compared TCR specificity of both CTL using a set of synthetic structurally related epitope variants (Fig. 7a and b). Substitution of the polymorphic residue with an A abrogates lysis of both CTL. Interestingly, the HB-1H-specific CTL MP1 also recognizes the R analogue, which closely resembles the H residue, whereas the F analogue, which is more structurally related to the Y residue, is not recognized. Vice versa, the HB-1Y-specific CTL KU recognizes the F analogue and not the R analogue. These data indicate that the H/Y antigenic disparity results from differences in TCR interactions probably due to conformational changes of the polymorphic epitopes in the context of HLA-B44.
The HB-1-encoding gene of yet unknown function is located on chromosome 5q32. HB-1 consists of at least two allelic variants, HB-1H and HB-1Y, which encode the HLA-B44-restricted peptides EEKRGSLHVW and EEKRGSLYVW, respectively. These peptides are generated from two different exons separated by an intronic region of ∼8.2 kb. Based on this genomic information, we have developed a method to perform HB-1 typing at the DNA level. Using this assay, we determined that the allelic frequency for the H and Y polymorphism is 0.79 and 0.21, respectively. From these frequencies we calculated that the expected recipient disparity for HB-1H is 2.8%, whereas recipient disparity for HB-1Y is expected to be 12.4%. So anti-tumor alloresponses in B lineage ALL patients after SCT are more likely to occur in the HB-1Y direction than in the HB-1H direction.
Disparity for mHag between SCT recipients and donors results from genetic polymorphisms that either affect MHC presentation, antigen processing or TCR recognition of self-peptides 3. For instance, immunogenicity of mHag HA-1 is a clear example where the amino acid change has effect on peptide binding so that the negative allelic peptide is not presented at the cell surface by HLA-A*;0201 molecules 12. In contrast, Brickner et al. 14 reported that disparity for mHag HA-8 results from altered antigen processing rather than from differences in interaction with HLA. They showed that the negative allelic peptide of HA-8 is poorly transported into the ER by TAP molecules. Another possibility is that allelic variant peptides of mHag are cleaved differentially by the proteasome resulting in either creation or loss of a cleavage site. Beekman et al. 22 showed that single amino acid substitutions in variant viral sequences can lead to premature destruction of the CTL epitope by proteasome-mediated digestion. Here, we describe that disparity for mHag HB-1 is neither the result from altered peptide binding to MHC nor from differential antigen processing, but results from differences in TCR recognition. Other examples where both allelic variant peptides bind with similar affinity to the relevant HLA molecule are the SMCY-derived mHag HY.B7 and HY.A2 5, 6, the UTY-derived HY.B60 mHag 9 and the DFFRY-derived HY.A1 mHag 7. Like these mHag, HB-1 stimulates T cells with the appropriate TCR specificity that discriminates mHag+ from mHag– target cells. It has been shown that such discriminative T cell recognition can be further affected by post-translational modification of the polymorphic residue 6, 7. However, a reciprocal CTL response against different allelic peptides of human mHag has not been formally demonstrated. We found that the HB-1Y variant peptide is also able to stimulate CTL recognizing the naturally processed and presented peptide on HB-1Y expressing target cells. These results indicate that allelic discrimination of HB-1 epitopes results from different conformations adopted in the same HLA molecule leading to differential T cell triggering. Interestingly, HB-1 is the first described human mHag inducing bi-directional CTL responses.
The development of methods to type donor/recipient pairs for mHag makes it possible to establish whether recipient disparity for molecular-defined mHag is associated with GVHD and GVL responses. It has been documented that recipient disparity for HA-1 and HA-8 is associated with acute GVHD 25–27. Because HA-8 is expressed ubiquitously, HA-8-specific donor CTL responses will be directed against both cells of hematopoietic as well as non-hematopoietic origin. But linkage of HA-1 incompatibility with GVHD was surprising, since HA-1 is selectively expressed in hematopoietic cells. However, HA-1 is highly expressed in dendritic cells (DC) and it has been shown that host-derived DC play a crucial role in the initiation of GVHD 28. Thus, inflammatory responses mediated by HA-1-specific CTL against recipient DC in tissues might cause cytokine-mediated damage or attracts other donor CTL that recognize mHag on non-hematopoietic cells of the host. In contrast, HB-1 is selectively expressed on immature B lymphoblasts including B lineage ALL cells and is not endogenously expressed by DC. Therefore, HB-1 probably induces a selectiveGVL response without collateral damage of normal tissues. Whether incompatibility for hematopoietic cell-restricted mHag, such as HA-1, HA-2 and HB-1, is associated with a reduced risk of relapse is so far unknown.
The real impact of mHag in GVL responses may be answered in approaches using hematopoietic cell-restricted mHag as targets for immunotherapy 29. HB-1 is differentially expressed in cells of the B lymphoid lineage including tumor cells of all B lineage ALL subtypes (pro-B-ALL, common-B-ALL, pre-B-ALL and B-ALL). Such highly restricted tissue expression renders HB-1 an ideal target for anti-tumor immunotherapy. HB-1H- and HB-1Y-specific T cells could be generated ex vivo and used for adoptive immunotherapy post-transplant. Studies in mice showed that the adoptive transfer of mHag-specific CTL has the ability to eradicate leukemic cells 30. Recipient disparity for HB-1H (∼2.8% in the HLA-B44+ population) and for HB-1Y (∼12.4% in the HLA-B44+ population) is low which limits the potential therapeutic value of HB-1 mHag-specific T cells. We have preliminary results that the HB-1 amino acid sequence harbors another potential mHag (SLH/YVWKSEL) that binds with intermediate affinity to HLA-A2.1 molecules as determined in a competition-based peptide-binding assay (data not shown). Furthermore, peptide fragments containing this sequences with the C-terminal L residue are generated by 20 S proteasome-mediated digestion (Table 2). Currently, we are investigating whether these HLA-A2.1-restricted polymorphic peptides are able to stimulate CTL responses. Molecular characterization of additional hematopoietic cell-restricted mHag will be required for treatment of substantial numbers of leukemia patients by adoptive immunotherapy with mHag-specific CTL.
4 Materials and methods
4.1 Cell culture
Cell lines were cultured in IMDM (Gibco BRL, Gaithersburg, MD) containing 10% FCS. EBV-LCL were generated by transformation with B95–8 EBV supernatant, and cultured in IMDM/10% FCS. The cell line Raji was transfected with pCR3.B*;4402 or pCR3.B*;4403 by electroporation. Raji.B44 transfectants were selected with 1.5 mg/ml G418 (Gibco BRL) and cloned by limiting dilution. One clone of Raji.B*;4402 and Raji.B*;4403 were subsequently co-transfected with either pcDNA3-zeo.HB-1H or pcDNA3-zeo.HB-1Y, selected with 0.5 mg/ml zeocin (Invitrogen, San Diego, CA) and cloned by limiting dilution. CTL MP1, generated from common-ALL patient MP (HLA-B*;4403+) after HLA-identical SCT, was grown in IMDM/10% human serum (HS), irradiated EBV-LCL of patient MP (2.5×105/ml), irradiated allogeneic PBMC (5×105/ml), 100 IU/ml IL-2 (Eurocetus, Amsterdam, The Netherlands), and 1 μg/ml PHA (Boehringer, Mannheim, Germany). CTL MP1 recognizes the HB-1H.B44 epitope, EEKRGSLHVW, in association with either HLA-B*;4403 or HLA-B*;4402 13, 20.
4.2 Genotyping of HB-1 mHag polymorphism
DNA was isolated using the Puregene method (Gentra Systems). HB-1 DNA amplification was performed by PCR using 0.5 μM HB1–5B forward primer (5′-GAGAAGTTGTAAGCTCAAGTCTCAGC-3′), 0.5 μM HB1–3B reverse primer (5′-CAAACCTCAACAGCAGTGTCAAAG-3′), 0.25 mM dNTP and 2.5 U Taq Gold polymerase (Perkin Elmer). A fragment of 153 bp was obtained by 35 cycles of the following PCR protocol: 1 min at 95 °C, 1 min at 60 °C, and 1 min at 72 °C. Ten microliters of the amplification product was digested with 10 U NlaIII for 2 h at 37 °C and analyzed on a 2% agarose gel. The HB-1H allele produced 95- and 58-bp bands, whereas the HB-1Y allele produced a single 153-bp band.
4.3 Peptide synthesis
Peptides were synthesized by solid-phase strategies on an automated multiple peptide synthesizer (Abimed AMS 422). Peptides were purified by reversed phase-HPLC and lyophilized. Finally, peptides were dissolved in DMSO, diluted in IMDM to a peptide concentration of 1 mM and stored at –20 °C before use.
4.4 HLA-B44 peptide-binding assay
We designed an HLA-B44 peptide-binding assay using 721.221.B*;4403 cells similar to the assay described by Van der Burg et al. 31. Briefly, naturally bound peptides were eluted by exposing 721.221.B*;4403 cells for 90 s to ice-cold citric acid buffer with pH 3.1 (1:1 mixture of 0.263 M citric acid and 0.123 M Na2HPO4). Cells were immediately buffered with ice-cold IMDM/2% FCS, washed, and resuspended at a concentration of 0.5×106/ml in IMDM/2% FCS containing 1.5 μg/ml β2-microglobulin (Sigma). Subsequently, the stripped cellswere incubated at 5×104/well in a V-bottom 96-well plate with 1.25 μM of a fluorescein (Fl)-labeled HLA-B44-binding reference peptide and a titrated concentration of competitor test peptide. The sequence of the Fl-labeled was EENLLDFVRF (EBNA3c 281–290) wherein we substituted the aspartic acid (D) with a cysteine (C) to tag a Fl group to the peptide: EENLLC(Fl)FVRF. After 24 h incubation at 4 °C, cells were washed three times with PBS/1% BSA, fixed with 0.5% paraformaldehyde, and analyzed on an Epics XL flow cytometer (Beckman Coulter, Fullerton, CA). The percentage inhibition of Fl-labeled reference peptide binding was calculated using the following formula: % inhibition = [1–(mean fluorescence intensity (MFI)reference and competitor peptide – MFIno reference peptide)/(MFIreference peptide – MFIno reference peptide)]×100.
4.5 Proteasome digestion assay
The 20 S proteasomes were purified from EBV-LCL as described elsewhere 32. HB-1 peptides (22-mer, 20 μg) were incubated with 1 μg of purified proteasomes in 300 μl digestion buffer at 37 °C for 8 and 18 h. Trifluoroacetic acid (30 μl) was added to stop the digestion. Peptide digestions were analyzed by electrospray ionization mass spectrometry as described previously 22, 28. Briefly, digestion solutions were diluted five times in water-methanol-acetic acid (95:5:1, v/v/v) and 1 μl was trapped on the precolumn (MCA-300–05-C8; LC Packings). The precolumn was washed for 3 min to remove the buffers present in the digests. Subsequently, the trapped analytes were eluted with a steep gradient going from 70% Bto 90% B in 10 min, with a flow of 250 nl/min [A, water-methanol-acetic acid (95:5:1, v/v/v); B, water-methanol-acetic acid (10:90:1, v/v/v)]. Mass spectra were recorded from mass 50–2,000 Da everysecond with a resolution of 5,000 full width/half maximum (FWHM). The intensity of the peaks in the mass spectra was used to estimate the relative amounts of peptides generated after proteasome digestion. The relative amounts of the peptides are given as a percentage of the total amount of peptide digested by the proteasome at the indicated time. Fragments with intensities less than three times the noise peak in a spectrum were ignored in the determination of sequences of peptides generated after proteasome digestion.
4.6 In vitro CTL response induction
DC used as stimulator cells were generated from thawed PBMC of an HLA-B44+ donor. Briefly, PBMC (5×106/ml) were resuspended in IMDM/2% HS and incubated for 2 h at 37 °C in 75-cm2 tissue culture flasks. Nonadherent cells were removed and adherent cells were subsequent cultured in IMDM/5% HS supplemented with 800 U/ml GM-CSF (Schering-Plough, Amstelveen, TheNetherlands) and 1,000 U/ml IL-4 (Schering-Plough). After 6 days, cells were washed and cultured for 3 days in IMDM/5% HS supplemented with 800 U/ml GM-CSF, 1,000 U/ml IL-4 and 1 μg/ml recombinant trimeric CD40L (kindly provided by Immunex Corporation). On day 9, DC were harvested and pulsed with 50 μM HB-1Y.B44 peptide for 4 h at room temperature in the presence of 3 μg/ml human β2-microglobulin (Sigma). Subsequently, peptide-pulsed DC were irradiated (30 Gy), washed, and resuspended at 2×105 cells/ml in IMDM/10% HS. CD8+ T cells were isolated by positive selection using CD8 mAb-coupled magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany). Primary stimulation was performed in 24-well plates. CD8+ T cells (106/well) were cocultured with autologous peptide-pulsed mature DC (2×105/well) and autologous irradiated PBMC as feeder cells (105/well) in 2 ml IMDM/10% HS supplemented with 1,000 U/ml IL-6 (Sandoz, Basel, Switzerland) and 10 ng/ml IL-12 (Hoffman-La Roche, Nutley, NJ). At day 7 and 14, CTL were harvested, washed and restimulated with peptide-pulsed DC at an effector/stimulator ratio of 10:1 in the presence of 100 U/ml IL-2 and 5 ng/ml IL-7 (Genzyme, Cambridge, MA). From day 21 on, CTL were maintained by weekly stimulation with irradiated adherent PBMC or EBV-LCL pulsed with 10 μM peptide in IMDM/10% HS containing 100 U/ml IL-2 and 5 ng/ml IL-7.
4.7 Antibodies and immunofluorescence analysis
The following mAb were used for immunofluorescence analysis: TS2/18 (CD2), SPV-T3b (CD3), M310 (CD4), DK25 (CD8), CRI304.3 (TCRBV6S1), IG125 (TCRBV21S3), MAB104 (CD80), HB15A (CD83), HA5.2B7 (CD86), W6/32 (HLA-class I), Q5/13 (HLA-DR/DP), and SPV-L3 (HLA-DQ). Immunofluorescence analysis was performed by direct or indirect labeling. FITC-conjugated goat F(ab′)2 anti-mouse IgG and IgM (Tago Immunologicals, Camarillo, CA) was used for secondary staining. Cells were analyzed on an Epics XL flow cytometer (Beckman Coulter).
4.8 Chromium-release assay
Chromium-release assays were performed to assess cytolytic activity of CTL. Briefly, 106 target cells were incubated with 100 μCi 51Cr (Amersham, Buckinghamshire, GB) for1 h at 37 °C. Labeled target cells were mixed in V-bottom microtiter plates (103/well) with various numbers of effector cells in a total volume of 150 μl IMDM with 10% FCS. In peptide recognition assays, target cells were preincubated with various concentrations of peptide for 30 min at room temperature in a volume of 100 μl prior to the addition of effector cells. After 4 h of incubation at 37 °C, 100 μl supernatant was collected and radioactivity was measured by a gamma counter. The mean percentage specific lysis of triplicate wells was calculated using the following formula: % specific lysis = [(experimental release – spontaneous release)/(maximal release – spontaneous release)]×100.
4.9 IFN-γ secretion assay
IFN-γ-producing CTL were detected by the IFN-γ secretion assay (Miltenyi Biotec). Briefly, 0.5×106 CTL were incubated in a 24-well plate with 0.5×106 irradiated Raji.B*;4402 cells in a total volume of 2 ml IMDM/10% HS. Peptides were added to a final concentration of 10 μM. After 16 h of incubation at 37 °C, cells were harvested, washed with PBS/0.5%FCS/5 mM EDTA and labeled at a concentration of 108 cells/ml with 50 μg/ml Ab–Ab conjugates directed against CD45 and IFN-γ for 10 min on ice. Subsequently, cells were diluted withIMDM/10% FCS at 1×105 cells/ml and allowed to secrete for 45 min at 37 °C. After the cytokine-capturing period, cells were collected, resuspended at a concentration of 108 cells/ml in PBS/0.5% FCS/5 mM EDTA and stained with 5 μg/ml PE-conjugated anti-IFN-γ mAb and FITC-conjugated anti-CD8 mAb for 20 min at 4 °C. Cells were analyzed on an Epics XL flow cytometer (Beckman Coulter).
We thank G. F. M. Merkx for FISH analysis of the HB-1 gene and Dr. J. W. Drijfhout for generously providing the fluorescent-labeled HLA-B44-restricted EBNA3c peptide. This work was supported by grants from the Dutch Cancer Society (KUN 97–1508) and the Fondo de Investigacion Sanitaria, Ministry of Health, Spain (00/0363).