To identify shared epitopes for melanoma-reactive CTL restricted by MHC molecules other than HLA-A*0201, six human melanoma patient CTL lines expressing HLA-A1 were screened for reactivity against the melanocyte differentiation proteins Pmel-17/gp100, MART-1/Melan-A, and tyrosinase, expressed via recombinant vaccinia virus vectors. CTL from five of the six patients recognized epitopes from tyrosinase, and recognition of HLA-A1+ target cells was strongly correlated with tyrosinase expression. Restriction by HLA-A1 was further demonstrated for two of those tyrosinase-reactive CTL lines. Screening of 119 synthetic tyrosinase peptides with the HLA-A1 binding motif demonstrated that nonamer, decamer, and dodecamer peptides containing the sequence KCDICTDEY (residues 243–251) all reconstituted the CTL epitope in vitro. Epitope reconstitution in vitro required high concentrations of these peptides, which was hypothesized to be a result of spontaneous modification of cysteine residues, interfering with MHC binding. Substitution of serine or alanine for the more N-terminal cysteine prevented modification at that residue and permitted target cell sensitization at peptide concentrations 2 to 3 orders of magnitude lower than that required for the wild-type peptide. Because spontaneous modification of sulfhydryl groups may also occur in vivo, tumor vaccines using this or other cysteine-containing peptides may be improved by amino acid substitutions at cysteine residues.

CTL recognize peptide epitopes derived from intracellular proteins, presented on the cell surface in association with class I MHC molecules. Purified peptides have proven effective at inducing specific CTL responses and protective immunity against viral infection and virally induced tumors when those peptides represent CTL epitopes (1, 2, 3). Similarly, the identification of shared epitopes for melanoma-reactive CTL may be useful for immunotherapy of cancer. Indeed, melanoma-reactive CTL can be isolated in vitro from melanoma patients, suggesting that an immune response to melanoma occurs in these patients and that its augmentation might be therapeutic. In addition, many melanoma-reactive CTL lines restricted by the class I MHC molecule HLA-A*0201, expressed in nearly half of the melanoma patient population, recognize shared epitopes present on most allogeneic A*0201+ melanomas (4).

A number of gene products have been identified and characterized as sources of immunogenic peptides in human melanoma. These include the MAGE, BAGE, and GAGE gene families (5, 6) and the melanocytic tissue differentiation proteins Pmel-l7/gp100, MART-1/Melan-A, and tyrosinase (7, 8, 9, 10). The tyrosinase gene encodes two peptide sequences that have been demonstrated to be epitopes for two different HLA-A*0201+ CTL clones, and an HLA-B44-restricted tyrosinase peptide has also been reported as a CTL epitope (11, 12, 13). In addition, tyrosinase encodes an epitope recognized by class II MHC-restricted melanoma-reactive CD4+ T cell clones (14). Most of the progress in identifying peptide epitopes for melanoma-reactive CTL has focused on HLA-A*0201-restricted responses; however, over half of the patients do not express HLA-A*0201 (15), and some tumors may lose expression of certain MHC molecules as an immune escape mechanism (16, 17, 18, 19). Thus, it is important to identify immunodominant peptide epitopes for CTL restricted by common MHC molecules other than HLA-A*0201. Two peptide epitopes, derived from MAGE-1 and MAGE-3, have been identified as being presented by HLA-A1 (5, 20, 21), and we have recently reported a Pmel-17/gp100-derived epitope restricted by HLA-A3 (22). However, most of the CTL epitopes for MHC molecules other than HLA-A*0201 remain to be identified.

Pmel-17/gp100, MART-1/Melan-A, and tyrosinase proteins are expressed in the vast majority of human melanomas (8, 23, 24). With the demonstration that peptides from each of these function as shared epitopes for HLA-A*0201-restricted melanoma-reactive CTL, we hypothesized that additional peptides derived from these proteins function as epitopes for melanoma-reactive CTL restricted by other common class I MHC molecules. In the present report we describe reactivities of a panel of six HLA-A1+ CTL lines against these melanocytic tissue differentiation Ags. We have identified a peptide epitope derived from tyrosinase that is recognized by at least two distinct CTL lines, and we have characterized some features of this peptide that have significant implications for its use in clinical tumor vaccine trials and will probably have implications for other MHC-associated peptides epitopes.

The human melanoma cell lines DM6 and DM331 were provided by Drs. H. F. Seigler and T. L. Darrow (Duke University, Durham, NC). The Na8 Mel and Na8 Mel+Tyr cell lines were gifts from Vincent Brichard and Thierry Boon. Na8 Mel is a tyrosinase-negative human melanoma line, and Na8 Mel+Tyr was derived from transfection of Na8 Mel with the tyrosinase cDNA 123.B2 (7). The human melanoma line SkMel24 was obtained from the American Type Culture Collection (Rockville, MD). The human melanoma cell lines VMM12, VMM14, and VMM15 and the fresh tumor cell digests VMM21 and VMM40 were obtained from patients at the University of Virginia (Charlottesville, VA). Immunohistochemical staining of these lines with S-100, HMB-45, and vimentin Abs was characteristic of melanoma. EBV-transformed B lymphoblastoid cell lines (B-LCL)4 were generated from melanoma patients by incubating PBMC in 1 ml of EBV-containing supernatant (from B-958 cells) for 1 h at 37°C and then culturing in RPMI/PHA. The HLA class I transfectants of the C1R cell line (C1R-A1, C1R-A3, C1R-B7, and C1R-B8) were provided by Dr. Peter Cresswell (Yale University, New Haven, CT). The 143B (tk) and CV-1 lines used in the propagation of vaccinia virus were obtained from the American Type Culture Collection. HLA typing was performed by microcytotoxicity assay on autologous lymphocytes when available (GenTrak, Inc., Plymouth Meeting, PA; Olympus Corp., Lake Success, NY; One Lambda, Canoga Park, CA). Expression of HLA-A1 and -A3 by tumor cells was confirmed by staining with the mAbs HA-A1 (One Lambda, Canoga Park, CA) and GAP-A3 (provided by P. Cresswell). Melanoma-reactive CTL from peripheral blood lymphocytes, from tumor-involved nodes, or from tumor-draining nodes were generated in vitro by repeated stimulation with autologous tumor cells following the detailed protocols previously reported (22, 25). The tissue culture medium used for the human cell lines was RPMI 1640 (Sigma Chemical Co., St. Louis, MO) supplemented with 10% FCS, glutamine, and antibiotics, hereafter referred to simply as RPMI.

The full-length Pmel-17 cDNA in pcDNA1/neo (Invitrogen, San Diego, CA) was provided by Dr. S. N. Wagner, University of Essen (Essen, Germany). A tyrosinase cDNA was provided by Thierry Boon, Ludwig Institute (Brussels, Belgium). The MART-1/Melan-A gene was PCR cloned and sequenced in our laboratories from the human melanoma line DM6 using standard PCR conditions and the primers (5′ to 3′) GCG CGG TAC CCT GAC CCT ACA AGA TGC CA and GCG CAA GCT TGT CTC AGG TGT CTC GCT GGC. Each of these three gene sequences was subcloned into the vaccinia cloning vector pSC11.3 (26) adjacent to the vaccinia P7.5 early/late promoter and sequenced to verify proper insertion. Recombinant vaccinia viruses were then produced from these modified vectors using standard methods (27). Purified viral stocks were titrated and tested for proper expression of these three Ags using HLA-A*0201-restricted CTL reactive against the Pmel-17/gp100, MART-1/Melan-A, and tyrosinase epitopes YLEPGPVTA, AAGIGILTV, and YMDGTMSQV, respectively. These CTL lysed nonmelanoma target cells infected with the vaccinia recombinants at levels comparable to the levels of lysis of HLA-A*0201-positive tumor cells, while uninfected targets or targets infected with a vaccinia encoding an irrelevant Ag were not lysed (data not shown).

Peptides were synthesized by standard F-moc chemistry using a Gilson model AMS422 peptide synthesizer (Gilson, Middleton, WI). Biologically active peptides identified at initial screening were purified to >98% by reverse phase HPLC on a Vydac C-4 column with 0.05% trifluoroacetic acid/water and an acetonitrile gradient and then re-evaluated in cytolysis assays.

Cell-mediated lysis of target cells was determined using a standard 4-h 51Cr release assay, as described previously (25). Transient expression of individual melanoma differentiation Ags in nonmelanoma cells was accomplished by infecting target cells with recombinant vaccinia viruses using 50 plaque-forming units (titrated on CV-1)/cell for 5 to 7 h at 37°C before 51Cr labeling. Peptide-pulsed targets were generated by diluting peptides in RPMI and preincubating with 51Cr-labeled target cells for 2 h before the addition of CTL. Assay wells containing peptide and target cells but no CTL were used as controls to rule out toxicity of the peptides themselves. Positive controls for specific cytolytic activity consisted of autologous tumor or allogeneic melanoma tumor matched for MHC class I expression. Negative controls consisted of target cells infected with irrelevant recombinant vaccinia viruses or mock infected. Cold target inhibition assays were performed as follows. Labeled or unlabeled target cells were preincubated with peptides (20 μg/ml) for 1 h and then washed. T cells were then preincubated with unlabeled (cold) targets for 1 h before incubation with labeled (hot) targets for 4 h.

Melanoma-reactive CTL lines were generated from human melanoma patients and tested for reactivity against autologous and allogeneic melanomas and autologous and allogeneic B-LCL following transient expression of the melanocyte differentiation proteins Pmel-17/gp100, MART-1/Melan-A, and tyrosinase. These CTL lines were tested as soon as residual NK and lymphokine-activated killer activities were low, and reactivity against autologous melanoma could be demonstrated, typically at 25 to 40 days in culture. As shown in Table I, reactivity against tyrosinase-derived epitopes was observed for five of the six HLA-A1+ CTL lines. The one HLA-A1+ CTL line that did not demonstrate reactivity against tyrosinase was derived from a patient whose tumor line, DM331, does not express tyrosinase or any other defined melanocyte differentiation proteins.5 Thus, tyrosinase-derived epitopes were recognized by CTL from all HLA-A1-positive patients whose tumors express tyrosinase.

Table I.

Recognition of melanocyte differentiation Ags by human melanoma patient CTL lines

CTL LineMHC Typing of CTL DonorMHC Class I Molecules EvaluatedAg Reactivity Observeda
Pmel-17/ gp100MART-1/ Melan-ATyrosinase
DM331 -A1, 2;-B15, B62 -A1, 2 − − − 
VMM12 -A1, A3, -B7, B14 *b − − 
VMM14 -A1, 25;-B8, B48 -A1, 25;-B8 − − 
VMM15 -A1, 25;-B8, B18 − 
VMM21 -A1, 2;-B7, 37 − 
VMM40 -A1, 2;-B27, 57 − − 
CTL LineMHC Typing of CTL DonorMHC Class I Molecules EvaluatedAg Reactivity Observeda
Pmel-17/ gp100MART-1/ Melan-ATyrosinase
DM331 -A1, 2;-B15, B62 -A1, 2 − − − 
VMM12 -A1, A3, -B7, B14 *b − − 
VMM14 -A1, 25;-B8, B48 -A1, 25;-B8 − − 
VMM15 -A1, 25;-B8, B18 − 
VMM21 -A1, 2;-B7, 37 − 
VMM40 -A1, 2;-B27, 57 − − 
a

Positive values indicate that target cells expressing the Ag listed, via a vaccinia construct, were lysed (at an E:T ratio of 30:1) at a level at least 10% higher than background lysis of the target cell infected with an irrelevant vaccinia construct. Representative vaccinia data are shown in Figure 1.

b

* indicates that all MHC were evaluated: these CTL were evaluated for recognition of autologous EBV-B cell lines infected with each of the vaccinia constructs listed.

Of the HLA-A1+ CTL lines recognizing tyrosinase epitopes, VMM12 CTL and VMM15 CTL lines were sufficiently available to permit detailed analysis. To determine which of the class I MHC alleles was responsible for presentation of tyrosinase epitopes to these two CTL lines, the tyrosinase-vaccinia constructs were evaluated in C1R cells transfected with one of the class I MHC alleles expressed on the patient’s tumor. Following infection with the recombinant vaccinia virus expressing full-length tyrosinase, the HLA-A1 transfectant of C1R was recognized and lysed by VMM12 CTL and VMM15 CTL, but HLA-A3, -B7, and -B8 C1R transfectants were not recognized under identical conditions (Fig. 1). Thus, HLA-A1 is a restriction element for the recognition of tyrosinase by CTL lines VMM12 and VMM15.

FIGURE 1.

Recognition of tyrosinase by VMM12 and VMM15 CTL is restricted by HLA-A1. C1R transfectants expressing HLA-A1, -A3, -B7, and -B8 were infected with a recombinant vaccinia virus expressing full-length tyrosinase or mock infected, as indicated. These cells and VMM12 and VMM15 tumor cells were subsequently labeled with 51Cr. Labeled target cells were cocultured with VMM12 CTL (A) or VMM15 CTL (B) in a standard chromium release cytotoxicity assay. VMM12 expresses HLA-A1, -A3, -B7, and -B14. VMM15 expresses HLA-A1, -A25, -B8, and -B18. VMM38 EBV share HLA-B18 with VMM15.

FIGURE 1.

Recognition of tyrosinase by VMM12 and VMM15 CTL is restricted by HLA-A1. C1R transfectants expressing HLA-A1, -A3, -B7, and -B8 were infected with a recombinant vaccinia virus expressing full-length tyrosinase or mock infected, as indicated. These cells and VMM12 and VMM15 tumor cells were subsequently labeled with 51Cr. Labeled target cells were cocultured with VMM12 CTL (A) or VMM15 CTL (B) in a standard chromium release cytotoxicity assay. VMM12 expresses HLA-A1, -A3, -B7, and -B14. VMM15 expresses HLA-A1, -A25, -B8, and -B18. VMM38 EBV share HLA-B18 with VMM15.

Close modal

VMM15 CTL were also evaluated for recognition of a panel of target cells. As shown in Figure 2, lysis depended on the expression of HLA-A1 and the expression of tyrosinase. In particular, there is no lysis of the tyrosinase-negative melanoma Na8 Mel, but there was significant lysis of the tyrosinase-transfected cell line Na8 Mel+Tyr. Both Na8 Mel and Na8 Mel+Tyr express HLA-A1 (data not shown). The melanoma line DM331, which is HLA-A1+ and tyrosinase, is not lysed, whereas the HLA-A1+/tyrosinase+ melanomas HT144 and VMM14 are lysed. VMM12 tumor cells variably lose expression of HLA-A1 and are not lysed in this assay, but have been lysed by VMM15 CTL in other assays when HLA-A1 is expressed (data not shown). DM6 is a tyrosinase+/HLA-A1 melanoma and is not lysed. In several other assays, SkMel24, another tyrosinase/HLA-A1+ melanoma, was not lysed by this CTL line (data not shown). Thus, tyrosinase expression was both necessary and sufficient for lysis of HLA-A1+ target cells by VMM15 CTL.

FIGURE 2.

Target cell recognition by VMM15 CTL correlates with expression of tyrosinase. VMM15 CTL were assayed for cytolytic activity in a 4-h 51Cr release assay. All cells lysed at 10% or greater in this assay had titratable lysis over an E:T cell ratio range of 5 to 40, whereas none of the others were lysed above 5% at any E:T cell ratio tested. Spontaneous release was less than 20% for all target cells reported. The class I MHC expression and tyrosinase expression for each target cell line are: VMM15 tumor and VMM15-EBV (A1, A25, B8, B18; tyrosinase+ for tumor only), Na8 Mel and Na8 Mel+Tyr (A1; tyrosinase+ for Na8 Mel+Tyr only), HT144 (A1, A24, B13, B15, C3; tyrosinase+), DM6 (A2, B12, B13 or -35, C1, C2; tyrosinase+), VMM12 (A1 variably expressed, A3, B7, B14; tyrosinase+), VMM14 (A1, A25, B8, B48; tyrosinase+), and DM331 (A1, A2, B15, B62, tyrosinase).

FIGURE 2.

Target cell recognition by VMM15 CTL correlates with expression of tyrosinase. VMM15 CTL were assayed for cytolytic activity in a 4-h 51Cr release assay. All cells lysed at 10% or greater in this assay had titratable lysis over an E:T cell ratio range of 5 to 40, whereas none of the others were lysed above 5% at any E:T cell ratio tested. Spontaneous release was less than 20% for all target cells reported. The class I MHC expression and tyrosinase expression for each target cell line are: VMM15 tumor and VMM15-EBV (A1, A25, B8, B18; tyrosinase+ for tumor only), Na8 Mel and Na8 Mel+Tyr (A1; tyrosinase+ for Na8 Mel+Tyr only), HT144 (A1, A24, B13, B15, C3; tyrosinase+), DM6 (A2, B12, B13 or -35, C1, C2; tyrosinase+), VMM12 (A1 variably expressed, A3, B7, B14; tyrosinase+), VMM14 (A1, A25, B8, B48; tyrosinase+), and DM331 (A1, A2, B15, B62, tyrosinase).

Close modal

To identify the HLA-A1-restricted peptide epitope(s) recognized by VMM12 CTL and VMM15 CTL, we scanned the tyrosinase protein sequence for peptides matching described HLA-A1 binding motifs (30, 31, 32, 33, 34). For nine-residue peptides, these motifs can be summarized as follows: threonine, serine, or methionine at position 2; aspartate, glutamate, alanine, or serine at position 3; and tyrosine at the carboxyl terminus (position 9). Also included in this evaluation were peptides containing more than nine residues. Since the C-terminal tyrosine residue appears to be invariant for HLA-A1 binding peptides, we predicted that longer peptides would retain the C-terminal tyrosine. In a previous study of HLA-A*0201-restricted epitopes, we identified a post-translational modification of tyrosinase resulting in the conversion of an asparagine residue in an N-linked glycosylation consensus sequence into aspartate (12). Therefore, in addition to testing the wild-type version of potential HLA-A1 binding peptides that contained N-linked glycosylation sites (-N-x-[S/T]-), we also tested variants of these peptides with aspartate (D) substituted for asparagine (N). With these considerations, peptides matching this motif and ranging in length from 8 to 12 amino acids were synthesized and tested for their ability to reconstitute the epitope recognized by VMM15 CTL. None of the first set of peptides tested had activity; so additional peptides synthesized included a number that fit the binding motif loosely: containing only threonine or serine at position 2, only an acidic residue at position 3, or only the C-terminal tyrosine. A total of 119 peptides were synthesized and evaluated. In one experiment, 3 of 48 peptides tested were recognized: KCDICTDEY, DAEKCDICTDEY, and EKCDICTDEY, all of which contained the core nonamer sequence KCDICTDEY (Fig. 3). These three peptides containing the KCDICTDEY sequence will, collectively, be referred to as KCDICTDEY-related peptides.

FIGURE 3.

VMM15 CTL recognize synthetic peptides from tyrosinase containing the sequence KCDICTDEY. C1R cells were labeled with 51Cr and pulsed with the indicated crude synthetic peptides at approximately 20 μg/ml for 2 h before CTL coculture. CTL were used at an E:T cell ratio of 15:1. Open bars indicate lysis in the presence of CTL. As a control for possible direct peptide toxicity, lysis was also measured in the absence of CTL (filled bars).

FIGURE 3.

VMM15 CTL recognize synthetic peptides from tyrosinase containing the sequence KCDICTDEY. C1R cells were labeled with 51Cr and pulsed with the indicated crude synthetic peptides at approximately 20 μg/ml for 2 h before CTL coculture. CTL were used at an E:T cell ratio of 15:1. Open bars indicate lysis in the presence of CTL. As a control for possible direct peptide toxicity, lysis was also measured in the absence of CTL (filled bars).

Close modal

These synthetic KCDICTDEY-related peptides were further purified by HPLC and evaluated in a dose titration assay (Fig. 4). Recognition of C1R-A1 cells pulsed with each of these peptides was indistinguishable by VMM15 CTL, and at higher peptide concentrations, targets pulsed with each of these three peptides were recognized at levels comparable to those of the autologous tumor cells (Fig. 4). Similarly, we found that lysis of target cells pulsed with DAEKCDICTDEY (50 μg/ml) can be inhibited by competition with the wild-type VMM15 tumor cells in a cold target inhibition assay (data not shown). Thus, the epitope reconstituted by a KCDICTDEY-related peptide corresponds to the epitope presented naturally on VMM15 tumor cells. However, the concentration of peptide required for reconstitution was high, with half-maximal lysis observed at 2 μg/ml (∼2 μM). These peptides are distinguished from most other CTL epitopes by containing two cysteine residues. Since most other defined epitopes for melanoma-reactive CTL can be reconstituted at nanomolar concentrations or lower (10, 21, 35), this raised the question of whether reconstitution of the A1-associated epitope by these KCDICTDEY-related peptides was complicated by modifications at the sulfhydryl groups.

FIGURE 4.

VMM15 CTL recognize, at comparable doses, 9-, 10-, and 12-residue peptides containing the amino acid sequence KCDICTDEY in association with HLA-A1. These three synthetic peptides were purified by HPLC and evaluated in a dose-titration assay. C1R transfectants expressing HLA-A1 (C1R-A1) were labeled with 51Cr and then pulsed with the peptides at various concentrations, as indicated, for 2 h before CTL coculture. Cytotoxicity was measured in a standard chromium release assay. TPMFNDINIY is an irrelevant peptide derived from tyrosinase that also fits the HLA-A1 binding motif. Lysis of nonpeptide-pulsed C1R cells was 19%, and lysis of VMM15 tumor cells (positive control) was 50%. An E:T cell ratio of 25:1 was used.

FIGURE 4.

VMM15 CTL recognize, at comparable doses, 9-, 10-, and 12-residue peptides containing the amino acid sequence KCDICTDEY in association with HLA-A1. These three synthetic peptides were purified by HPLC and evaluated in a dose-titration assay. C1R transfectants expressing HLA-A1 (C1R-A1) were labeled with 51Cr and then pulsed with the peptides at various concentrations, as indicated, for 2 h before CTL coculture. Cytotoxicity was measured in a standard chromium release assay. TPMFNDINIY is an irrelevant peptide derived from tyrosinase that also fits the HLA-A1 binding motif. Lysis of nonpeptide-pulsed C1R cells was 19%, and lysis of VMM15 tumor cells (positive control) was 50%. An E:T cell ratio of 25:1 was used.

Close modal

By direct evaluation of peptides eluted from HLA-A1 molecules, it was not possible to determine definitively which of the KCDICTDEY-related peptides was naturally processed and presented in association with HLA-A1 (data not shown), suggesting that they are present in low quantity or have been modified. Subsequent studies thus addressed potential modifications of both the nonamer KCDICTDEY and the dodecamer DAEKCDICTDEY. Spontaneous modification of free cysteine residues by disulfide bond formation will occur in medium containing cystine (C-C) or glutathione, such as RPMI or human serum (36). Thus, we examined the effect of substituting serine or alanine for one or both cysteine residues in the synthetic KCDICTDEY-related peptides in an effort to prevent the formation of disulfide bonds.

The synthetic nonamer with the wild-type sequence (KCDICTDEY) induced half-maximal lysis at 2 to 20 μM. However, when the cysteine closer to the N-terminus (position 2) was replaced with serine or alanine (KSDICTDEY or KADICTDEY), half-maximal lysis was observed after incubation of targets with 100- to 1000-fold lower concentrations (2 and 20 nM, respectively; Fig. 5 A). On the other hand, peptides with alanine or serine substituted for the cysteine distal to the N-terminus, at position 5 (KCDISTDEY or KCDIATDEY) were not discernibly more effective than the wild-type peptide. Nonamer peptides with alanine or serine substituted for both cysteine residues were actually less effective than the wild-type peptide.

FIGURE 5.

Alanine or serine substitution at the more amino-terminal cysteine greatly enhances VMM15 CTL recognition. C1R transfectants expressing HLA-A1 (C1R-A1) were labeled with 51Cr and then pulsed with the peptides at various concentrations, as indicated, for 2 h before CTL coculture, at which time the peptide concentration decreased twofold. TPMFNDINIY is derived from tyrosinase and fits the HLA-A1 binding motif. Cytotoxicity was measured in a standard chromium release assay. A, VMM15 CTL were used at an E:T cell ratio of 10:1. Lysis of autologous tumor was 70%, and lysis of non-peptide-pulsed C1R-A1 was 4%. B, VMM15 CTL were used at an E:T cell ratio of 25:1. Lysis of autologous tumor was 44%; lysis of non-peptide-pulsed C1R-A1 was 3%.

FIGURE 5.

Alanine or serine substitution at the more amino-terminal cysteine greatly enhances VMM15 CTL recognition. C1R transfectants expressing HLA-A1 (C1R-A1) were labeled with 51Cr and then pulsed with the peptides at various concentrations, as indicated, for 2 h before CTL coculture, at which time the peptide concentration decreased twofold. TPMFNDINIY is derived from tyrosinase and fits the HLA-A1 binding motif. Cytotoxicity was measured in a standard chromium release assay. A, VMM15 CTL were used at an E:T cell ratio of 10:1. Lysis of autologous tumor was 70%, and lysis of non-peptide-pulsed C1R-A1 was 4%. B, VMM15 CTL were used at an E:T cell ratio of 25:1. Lysis of autologous tumor was 44%; lysis of non-peptide-pulsed C1R-A1 was 3%.

Close modal

Similarly, the dodecamer with the wild-type sequence (DAEKCDICTDEY) required >10 μM for half-maximal sensitization of target cells for lysis. However, when the cysteine closer to the N terminus (position 5) was replaced with alanine (DAEKADICTDEY), half-maximal lysis was observed after incubation of targets with approximately 100 nM (Fig. 5 B). On the other hand, alanine substitution for the cysteine closer to the C-terminus, at position 8 (DAEKCDIATDEY), prevented recognition at every concentration tested. These data demonstrate that the more C-terminal cysteine (residue 247 of tyrosinase) is critical to the formation of the CTL epitope, raising the possibility that spontaneous modifications of that residue by disulfide bond formation may be integral to the actual epitope.

Because of the similar effects on CTL recognition observed with amino acid substitutions for the more N-terminal cysteine residue in both the dodecamer DAEKCDICTDEY and the nonamer KCDICTDEY, and because of the similar dose-titration curves observed for CTL recognition (Fig. 4), we hypothesized that residues 243 to 251 (KCDICTDEY) of both peptides adopt a similar conformation when bound to HLA-A1. A cold target inhibition assay was performed with peptide-pulsed target cells to test that hypothesis. As shown in Figure 6, unlabeled target cells pulsed with the dodecamer effectively inhibited lysis of 51Cr-labeled targets pulsed with the nonamer, at a level equivalent to the inhibition induced by cold targets pulsed with the nonamer itself. Recognition of labeled targets pulsed with the dodecamer DAEKSDICTDEY was similarly inhibited by unlabeled targets pulsed with the nonamer KSDICTDEY (data not shown). These data suggest that the conformation of the peptide-MHC complex within the region recognized by the TCR of VMM15 is, therefore, similar or identical for the nonamer and dodecamer peptides.

FIGURE 6.

The dodecamer peptide DAEKSDICTDEY competes with the nonamer peptide KSDICTDEY for recognition by VMM15 CTL. B-LCL derived from patient VMM15 (VMM15 EBV) were 51Cr labeled, pulsed with the peptide KSDICTDEY at a final concentration of 20 μg/ml for 1 h, and then washed. Unlabeled (cold) VMM15 EBV were similarly treated with KSDICTDEY, DAEKSDICTDEY, or no peptide, as indicated. Increasing numbers of these various unlabeled targets were preincubated with VMM15 CTL for 1 h at 37°C. Labeled (hot) targets were then added for an additional 4-h coculture at a final effector to labeled target ratio of 10:1.

FIGURE 6.

The dodecamer peptide DAEKSDICTDEY competes with the nonamer peptide KSDICTDEY for recognition by VMM15 CTL. B-LCL derived from patient VMM15 (VMM15 EBV) were 51Cr labeled, pulsed with the peptide KSDICTDEY at a final concentration of 20 μg/ml for 1 h, and then washed. Unlabeled (cold) VMM15 EBV were similarly treated with KSDICTDEY, DAEKSDICTDEY, or no peptide, as indicated. Increasing numbers of these various unlabeled targets were preincubated with VMM15 CTL for 1 h at 37°C. Labeled (hot) targets were then added for an additional 4-h coculture at a final effector to labeled target ratio of 10:1.

Close modal

We have shown previously that the melanocyte differentiation protein Pmel-17/gp100 is a source of epitopes for melanoma-reactive CTL restricted by HLA-A*0201 and HLA-A3 (10, 22). Using vaccinia constructs encoding this protein and CTL lines from patients with varied MHC profiles, we identified the Pmel-17/gp100 peptide ALLAVGATK as an epitope for A3-restricted CTL (22). We now provide evidence that tyrosinase-derived peptides are frequently recognized by HLA-A1+ CTL. Among HLA-A1+ patients, CTL recognition of tyrosinase epitopes was evident in all cases, except when the autologous tumor itself failed to express tyrosinase. By screening over 100 synthetic peptides derived from the tyrosinase sequence, we found that the peptides KCDICTDEY, EKCDICTDEY, and DAEKCDICTDEY were all recognized by A1-restricted melanoma-reactive CTL.

The naturally processed epitope may be a nonamer or may be as long as a dodecamer. Peptides containing 12 residues are uncommon among described class I MHC-associated peptides (37), and we are unaware of previously reported examples of dodecamer peptides representing CTL epitopes. Dodecamer peptides have been identified in association with cells containing Ag-processing defects, where a TAP-independent pathway has been implicated in their processing and presentation (38). Also, in cells with normal Ag-processing machinery, dodecamers associated with HLA-A1 and HLA-A11 have been described (37). A dodecamer from cytochrome c oxidase that is associated with HLA-A1 contains putative anchor residues at positions 3 and 12 (YTDYGGLIFNSY) (37). Thus, that peptide is predicted to bind to the HLA-A1 molecule with a kink between residues 3 and 12 to allow for the extra length. The synthetic DAEKCDICTDEY peptide, however, could potentially bind to HLA-A1 in either of two conformations. While the tyrosine residue at the C terminus is almost certain to function as the C-terminal anchor residue, the motif requirement for an acid residue at position 3 could be satisfied either by the glutamic acid (E) at position 3 or by the aspartic acid (D) at position 6. In the former case, the central portion of the peptide would have to be kinked and would extend out of the plane of the binding groove, a conformation that could be stabilized or induced by a disulfide bond between the two cysteine residues. If, however, the aspartic acid residue at position 6 functions as an anchor residue, the three N-terminal residues (DAE) would be expected to extend up and out of the binding groove, while the rest of the peptide interacts in the groove in a conformation resembling that of the nonamer peptide.

If the dominant conformation of the dodecamer is kinked, then it would be anticipated that CTL recognition would be sensitive to the peptide structure between residues 5 and 12, and that CTL recognizing that peptide would not cross-react with target cells pulsed with the synthetic nonamer lacking the N-terminal DAE residues. However, CTL reactive against the dodecamer were equally reactive against both the decamer EKCDICTDEY and the nonamer KCDICTDEY. While these studies were performed with a CTL line, which probably contained distinct CTL subpopulations, our cold target inhibition assays provide evidence that both the nonamer and dodecamer peptides are recognized by the same CTL. We have also observed that dodecamers containing glycine substituted for alanine or glutamate at positions 2 and 3 were recognized at levels comparable to those of the wild-type peptide (data not shown). Therefore, the nonamer and dodecamer peptides probably exist in similar conformations when bound to HLA-A1, such that the three amino-terminal residues of the dodecamer should extend out of the binding groove, and residues 4 through 12 would be expected to lie flat, in a manner analogous to the configuration expected for the nonamer peptide.

Regardless of the precise length and conformation of the peptides bound to the MHC, the -KCDICTDEY-related peptides are unusual in containing two cysteine residues. In general, sulfhydryl groups on cysteine residues are susceptible to derivatization with sulfhydryl groups on other biologic molecules such as cystine or glutathione, both of which are present in human serum and RPMI (36). Especially since formation of disulfide bonds is favored at mildly basic pH and in the presence of oxygen, both the in vivo setting (pH 7.4, pO2 40–100 mm Hg) and the standard culture conditions in vitro would favor modification of cysteine residues with exposed side chains. To assess the roles of the two cysteine residues in the KCDICTDEY-related peptides with respect to T cell recognition, a series of experiments was performed comparing the naturally occurring peptides with peptides containing amino acid substitutions. Substitution of serine or alanine for the more N-terminal cysteine prevents potential modifications of that residue, including formation of an internal disulfide bond. After making such a substitution, both the dodecamer and the nonamer were capable of reconstituting CTL epitopes at concentrations 2 to 3 logs lower than the unsubstituted peptides. Thus, we hypothesize that disulfide bonds involving this more N-terminal cysteine residue may negatively affect binding of exogenous peptide to the HLA-A1 molecule, and that in the naturally processed peptide, the cysteine residue at that position (position 2 of KCDICTDEY, residue 244 of tyrosinase) is not modified, perhaps being protected within the binding groove of HLA-A1. Similarly, the more C-terminal cysteine would be anticipated to point upward and to have an exposed side chain. In cystine- and glutathione-rich medium, such as RPMI (in vitro) or human serum (in vivo), the more C-terminal cysteine may well be modified by cysteinylation or by derivatization with another molecule containing a sulfhydryl group. T cell recognition of that epitope may depend on such a modification. Indeed, as shown in Figure 5, recognition by VMM15 CTL is acutely sensitive to the presence of a cysteine residue at this position (KSDICTDEY vs KSDISTDEY).

Peptides load onto MHC molecules in the oxidative environment of the endoplasmic reticulum (ER), where modification of cysteine residues by reversible formation of disulfide bonds would be expected (39). However, at least one enzyme in the ER, protein disulfide isomerase, can catalyze dissociation of disulfide bonds (40). We predict that those peptides not containing internal disulfide bonds or other modifications of the more N-terminal cysteine are those capable of binding to the nascent MHC molecules for presentation at the cell surface. Whether there are specific molecular chaperones favoring this process in the ER remains to be demonstrated.

Aside from what happens during normal Ag processing and presentation, synthetic peptides free in solution are susceptible to modifications of cysteine residues by formation of disulfide bonds, especially in the presence of oxygen or in slightly basic conditions (e.g., exposed to air or in vivo). Because of the proximity of the two cysteine residues, the likelihood of interaction between them probably exceeds the likelihood of interaction with another molecular species in solution. In fact, intramolecular disulfide bonds between cysteine residues in short peptides containing the sequence -CXYC- are highly favored at concentrations up to 200 μM in neutral pH solutions and are facilitated by air oxidation of the cysteine residues (28). The resulting cyclized form of KCDICTDEY-related peptides may not be capable of binding to the HLA-A1 molecule or, if bound, would probably result in a molecular conformation different from that of the epitope recognized by CTL. The fact that cyclization and other cysteine modifications occur readily may well explain the high concentration of KCDICTDEY-related peptides required for sensitization of C1R-A1 target cells in vitro and would imply similar results in vivo if a tumor vaccine using the wild-type peptide were administered.

On the other hand, peptides containing substitutions of serine or alanine for the more N-terminal cysteine induced half-maximal lysis of target cells at 1 to 20 nM, which compares favorably to peptide concentrations required for reconstitution of the MAGE-1 epitope EADPTGHSY on HLA-A1 (half-maximal lysis at ∼10 nM) (5) and the MART-1 peptide AAGIGILTV on HLA-A*0201 (half-maximal lysis at ∼1–100 nM) (35). We believe that replacement of the more N-terminal cysteine probably enables purified peptides to bind to cell surface MHC molecules in a manner that more accurately mimics the configuration of peptides binding to nascent MHC molecules in the ER. Ideally, we would like to be able to evaluate directly the HLA-A1 binding affinity for the KCDICTDEY-related peptides and for the substituted peptides. However, these peptides are so susceptible to modification in medium by cyclization, dimerization, or polymerization that the effective concentration of free peptide cannot be expected to match the total measured concentration. Thus, existing methods for measuring binding affinity of these peptides can be expected to be misleading, and specialized methods for working with these peptides are required.

In summary, at least three peptides containing the KCDICTDEY sequence are capable of reconstituting an epitope for HLA-A1-restricted melanoma-reactive CTL. Modifications of the sulfhydryl group of the more C-terminal cysteine residue by cysteinylation or by derivatization with glutathione are predicted to occur in vitro and in vivo, and these modifications may well be required for T cell recognition. Modifications of the more N-terminal cysteine residue, however, are not believed to occur in vivo during normal processing and presentation. However, they may occur when the peptide is free in solution, and modifications such as intramolecular disulfide bond formation leading to cyclization of the peptide probably inhibit markedly the loading of that peptide on HLA-A1 at the cell surface.

As demonstrated by our data with the wild-type and substituted forms of KCDICTDEY-related peptides, cysteine residues can both positively and negatively affect MHC binding and/or T cell recognition, a finding corroborated by another recent study (29). This suggests that potential problems with using cysteine-containing peptides as immunogens may be overcome by modifications of those cysteine residues, where those modifications are based on an understanding of interactions with the medium components, with the MHC, and with the TCR. In this particular example, immunization with a peptide containing serine or alanine in place only of the more N-terminal cysteine may result in more effective loading onto MHC molecules of APC.

As more peptide epitopes for melanoma-reactive CTL are identified, there is growing evidence for the important role of those peptides derived from the melanocytic tissue differentiation proteins Pmel-17/gp100, MART-1, and tyrosinase. The biochemical interactions of these peptides with other molecules in vitro and in vivo may affect interactions with the MHC molecules and with the TCR. Design of peptide-based immune therapies must take into account these potential interactions, which are probably even more complex in vivo than they are in vitro.

We acknowledge Thierry Boon (Ludwig Institute, Brussels, Belgium) for providing the cDNA for tyrosinase, Stephan N. Wagner (University of Essen, Essen, Germany) for providing cDNA for Pmel-17/gp100, Donna Deacon and Suzanne Morrison for CTL growth and HLA typing, Peter Dyke for generation of EBV-transformed B cell lines, Karen Frye for assistance with CTL culture, and Eric Huczko for providing HLA-A*0201-restricted murine anti-tyrosinase CTL lines.

1

This work was supported by National Institutes of Health Grants CA57653 (to C.L.S.), AI21393 and AI20963 (to V.H.E.), and AI 33993 (to D.F.H.); by American Cancer Society Grants IN149H and IM-768 (to C.L.S.); by the Cancer Center (National Institutes of Health Grant P30CA44579) at the University of Virginia; and by the Cancer Research Institute’s Elaine R. Shepard Clinical Investigator Award in Cancer Immunology (to C.L.S.). Fellowship support was provided by the Cancer Research Foundation of America (to D.J.K.), the Cancer Research Institute (to J.C.S.), and the National Institutes of Health Interdisciplinary Training Program in Immunology Grant AI07496 (to T.A.C.). Some of the human cell lines used in this study were obtained with the support of the Tissue Procurement Facility, funded in part by the Cancer Center at the University of Virginia (National Institutes of Health Grant P30CA44579) and by the Pratt Fund at the University of Virginia.

4

Abbreviations used in this paper: B-LCL, Epstein-Barr virus-transformed B lymphoblastoid cell line; ER, endoplasmic reticulum.

5

C. L. Slingluff, Jr., T. A. Colella, D. D. Graham, J. C. A. Skipper, L. Brinkerhoff, D. J. Kittlesen, D. H. Deacon, N. L. Harthun, E. L. Huczko, T. L. Darrow, and V. H. Engelhard. Immune escape due to concordant loss of multiple melanocytic differentiation proteins may be overcome by CTL recognizing unique Ags. Submitted for publication.

1
Scalzo, A. A., S. L. Elliott, J. Cox, J. Gardner, D. J. Moss, A. Suhrbier.
1995
. Induction of protective cytotoxic T cells to murine cytomegalovirus by using a nonapeptide and a human-compatible adjuvant (Montanide ISA 720).
J. Virol.
69
:
1306
2
Ada, G..
1991
. Strategies for exploiting the immune system in the design of vaccines.
Mol. Immunol.
28
:
225
3
Feltkamp, M. C., H. L. Smits, M. P. Vierboom, R. P. Minnaar, B. M. deJongh, J. W. Drijfhout, J. ter Schegget, C. J. Melief, W. M. Kast.
1993
. Vaccination with cytotoxic T lymphocyte epitope-containing peptide protects against a tumor induced by human papillomavirus type 16-transformed cells.
Eur. J. Immunol.
23
:
2242
4
Darrow, T. L., C. L. Slingluff, H. F. Seigler.
1989
. The role of HLA class I antigens in recognition of melanoma cells by tumor-specific cytotoxic T lymphocytes: evidence for shared tumor antigens.
J. Immunol.
142
:
3329
5
Traversari, C., P. van der Bruggen, I. F. Luescher, C. Lurquin, P. Chomez, A. Van Pel, E. De Plaen, A. Amar-Costesec, T. Boon.
1992
. A nonapeptide encoded by human gene MAGE-1 is recognized on HLA-A1 by cytolytic T lymphocytes directed against tumor antigen MZ2-E.
J. Exp. Med.
176
:
1453
6
Van den Eynde, B., O. Peeters, O. De Backer, B. Gaugler, S. Lucas, T. Boon.
1995
. A new family of genes coding for an antigen recognized by autologous cytolytic T lymphocytes on a human melanoma.
J. Exp. Med.
182
:
689
7
Brichard, V., A. Van Pel, T. Wolfel, C. Wolfel, E. De Plaen, B. Lethe, P. Coulie, T. Boon.
1993
. The tyrosinase gene codes for an antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas.
J. Exp. Med.
178
:
489
8
Coulie, P. G., V. Brichard, A. Van Pel, T. Wolfel, J. Schenider, C. Traversari, E. DePlaen, C. Lurquin, J.-P. Szikora, J.-C. Renauld, T. Boon.
1994
. A new gene encoding for a differentiation antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas.
J. Exp. Med.
180
:
35
9
Bakker, A. B., M. W. Schreurs, A. J. de Boer, Y. Kawakami, S. A. Rosenberg, G. J. Adema, C. G. Figdor.
1994
. Melanocyte lineage-specific antigen gp100 is recognized by melanoma-derived tumor-infiltrating lymphocytes.
J. Exp. Med.
179
:
1005
10
Cox, A. L., J. Skipper, Y. Chen, R. Henderson, T. L. Darrow, J. Shabanowitz, V. H. Engelhard, D. F. Hunt, C. L. Slingluff, Jr.
1994
. Identification of a peptide recognized by five melanoma-specific human cytotoxic T-cell lines.
Science
264
:
716
11
Wolfel, T., A. Van Pel, V. Brichard, J. Schneider, B. Seliger, K. H. Meyer zum Buschenfelde, T. Boon.
1994
. Two tyrosinase nonapeptides recognized on HLA-A2 melanomas by autologous cytolytic T lymphocytes.
Eur. J. Immunol.
24
:
759
12
Skipper, J. C. A., R. C. Hendrickson, P. H. Gulden, V. Brichard, A. VanPel, Y. Chen, J. Shabanowitz, T. Wolfel, C. L. Slingluff, Jr, T. Boon, D. F. Hunt, V. H. Engelhard.
1996
. An HLA-A2 restricted tyrosinase antigen on melanoma cells results from post-translational modification.
J. Exp. Med.
183
:
527
13
Brichard, V. G., J. Herman, A. Van Pel, C. Wildmann, B. Gaugler, T. Wolfel, T. Boon, B. Lethe.
1996
. A tyrosinase nonapeptide presented by HLA-B44 is recognized on a human melanoma by autologous cytolytic T lymphocytes.
Eur. J. Immunol.
26
:
224
14
Topalian, S. L., L. Rivoltini, M. Mancini, N. R. Markus, P. F. Robbins, Y. Kawakami, S. A. Rosenberg.
1994
. Human CD4+ T cells specifically recognize a shared melanoma-associated antigen encoded by the tyrosinase gene.
Proc. Natl. Acad. Sci. USA
91
:
9461
15
Tiwari, J. L., P. I. Terasaki.
1985
. The HLA complex. J. L. Tiwari, Jr, and P. I. Terasaki, Jr, eds.
HLA and Disease Associations
10
Springer-Verlag, New York.
16
Elliott, B. E., D. A. Carlow, A. M. Rodricks, A. Wade.
1989
. Perspectives on the role of MHC antigens in normal and malignant cell development.
Adv. Cancer Res.
53
:
181
(Rev.).
17
Doyle, A., W. J. Martin, K. Funa, A. Gazdar, D. Carney, S. E. Martin, I. Linnoila, F. Cuttitta, J. Mulshine, P. Bunn, J. Minna.
1985
. Markedly decreased expression of class I histocompatibility antigens, protein, and mRNA in human small-cell lung cancer.
J. Exp. Med.
161
:
1135
18
Lassam, N., G. Jay.
1989
. Suppression of MHC class I RNA in highly oncogenic cells occurs at the level of transcription initiation.
J. Immunol.
143
:
3792
19
Lehmann, F., M. Marchand, P. Hainaut, P. Pouillart, X. Sastre, H. Ikeda, T. Boon, P. G. Coulie.
1995
. Differences in the antigens recognized by cytolytic T cells on two successive metastases of a melanoma patient are consistent with immune selection.
Eur. J. Immunol.
25
:
340
20
Celis, E., V. Tsai, C. Crimi, R. DeMars, P. A. Wentworth, R. W. Chesnut, H. M. Grey, A. Sette, H. M. Serra.
1994
. Induction of anti-tumor cytotoxic T lymphocytes in normal humans using primary cultures and synthetic peptide epitopes.
Proc. Natl. Acad. Sci. USA
91
:
2105
21
Gaugler, B., B. Van den Eynde, P. van der Bruggen, P. Romero, J. J. Gaforio, E. De Plaen, B. Lethe, F. Brasseur, T. Boon.
1994
. Human gene MAGE-3 codes for an antigen recognized on a melanoma by autologous cytolytic T lymphocytes.
J. Exp. Med.
179
:
921
22
Skipper, J. C. A., D. J. Kittlesen, R. C. Hendrickson, D. D. Deacon, N. L. Harthun, S. N. Wagner, D. F. Hunt, V. H. Engelhard, C. L. Slingluff, Jr.
1996
. Shared epitopes for HLA-A3 restricted melanoma-reactive human CTL include a naturally processed epitope from Pmel-17/gp100.
J. Immunol.
157
:
5027
23
Adema, G. J., A. J. de Boer, A. M. Vogel, W. A. Loenen, C. G. Figdor.
1994
. Molecular characterization of the melanocyte lineage-specific antigen gp100.
J. Biol. Chem.
269
:
20126
24
Chen, Y. T., E. Stockert, S. Tsang, K. A. Coplan, L. J. Old.
1995
. Immunophenotyping of melanomas for tyrosinase: implications for vaccine development.
Proc. Natl. Acad. Sci. USA
92
:
8125
25
Slingluff, C. L., Jr, A. L. Cox, R. A. Henderson, D. F. Hunt, V. H. Engelhard.
1993
. Recognition of human melanoma cells by HLA-A2.1-restricted cytotoxic T lymphocytes is mediated by at least six shared peptide epitopes.
J. Immunol.
150
:
2955
26
Hahn, Y. S., V. L. Braciale, T. J. Braciale.
1991
. Presentation of viral antigen to class I major histocompatibility complex-restricted cytotoxic T lymphocytes: recognition of an immunodominant influenza hemagglutinin site by cytotoxic T lymphocyte is independent of the position of the site in the hemagglutinin translation product.
J. Exp. Med.
174
:
733
27
Mackett, M., G. L. Smith, B. Moss.
1984
. General method for production and selection of infectious vaccinia virus recombinants expressing foreign genes.
J. Virol.
49
:
857
28
Moroder, L., D. Besse, H-J. Musiol, S. Rudolph-Bohner, F. Siedler.
1996
. Oxidative folding of cystine-rich peptides vs. regioselective cysteine pairing strategies.
Biopolymers
40
:
207
29
Meadows, L., W. Wang, J. M. M. Den Haan, E. Blokland, C. Reinhardus, J. W. Drijfhout, J. Shabanowitz, R. Pierce, A. I. Agulnik, C. E. Bishop, D. F. Hunt, E. Goulmy, V. H. Engelhard.
1997
. The HLA-A*0201-restricted H-Y antigen contains a posttranslationally modified cysteine that significantly affects T cell recognition.
Immunity
6
:
273
30
DiBrino, M., K. C. Parker, J. Shiloach, R. V. Turner, T. Tsuchida, M. Garfield, W. E. Biddison, J. E. Coligan.
1994
. Endogenous peptides with distinct amino acid anchor residue motifs bind to HLA-A1 and HLA-B8.
J. Immunol.
152
:
620
31
DiBrino, M., T. Tsuchida, R. V. Turner, K. C. Parker, J. E. Coligan, W. E. Biddison.
1993
. HLA-A1 and HLA-A3 T cell epitopes derived from influenza virus proteins predicted from peptide binding motifs.
J. Immunol.
151
:
5930
32
Hobohm, U., A. Meyerhans.
1993
. A pattern search method for putative anchor residues in T cell epitopes.
Eur. J. Immunol.
23
:
1271
33
Kast, W. M., R. M. P. Brandt, J. Sidney, J.-W. Drijfhout, R. T. Kubo, H. M. Grey, C. J. M. Melief, A. Sette.
1994
. Role of HLA-A motifs in identification of potential CTL epitopes in human papillomavirus type 16 E6 and E7 proteins.
J. Immunol.
152
:
3904
34
Celis, E., J. Fikes, P. Wentworth, J. Sidney, S. Southwood, A. Maewal, M.-F. Del Guercio, A. Sette, B. Livingston.
1994
. Identification of potential CTL epitopes of tumor-associated antigen MAGE-1 for five common HLA-A alleles.
Mol. Immunol.
31
:
1423
35
Kawakami, Y., S. Eliyahu, K. Sakaguchi, P. F. Robbins, L. Rivoltini, J. R. Yannelli, E. Appella, S. A. Rosenberg.
1994
. Identification of the immunodominant peptides of the MART-1 human melanoma antigen recognized by the majority of HLA-A2-restricted tumor infiltrating lymphocytes.
J. Exp. Med.
180
:
347
36
Stofer-Vogel, B., T. Cerny, A. Kupfer, E. Junker, B. H. Lauterburg.
1993
. Depletion of circulating cyst(e)ine by oral and intravenous mesna.
Br. J. Cancer
68
:
590
37
Engelhard, V. H..
1994
. Structure of peptides associated with MHC class I molecules.
Curr. Opin. Immunol.
6
:
13
38
Henderson, R. A., H. Michel, K. Sakaguchi, J. Shabanowitz, E. Appella, D. F. Hunt, V. H. Engelhard.
1992
. HLA-A2.1-associated peptides from a mutant cell line: a second pathway of antigen presentation.
Science
255
:
1264
39
Hwang, C., A. J. Sinskey, H. F. Lodish.
1992
. Oxidized redox state of glutathione in the endoplasmic reticulum.
Science
257
:
1496
40
Freedman, R. B..
1984
. Native disulphide bond formation in protein biosynthesis: evidence for the role of protein disulphide isomerase.
Trends Biochem. Sci.
9
:
438