Human tumor antigens recognized by T lymphocytes: implications for cancer therapy

Transcription

1 Human tumor antigens recognized by T lymphocytes: implications for cancer therapy Rong-Fu Wang and Steven A. Rosenberg Surgery Branch, National Cancer institute, Bethesda, Maryland Abstract: The adoptive transfer of cytotoxic T lymphocytes (CTLs) derived from tumor-infiltrating lymphocytes (TIL) along with interleukin-2 (IL-2) into autologous patients with cancer resulted in the objective regression of tumor, indicating that these CTLS recognized cancer rejection antigens on tumor cells. In the past year, a number of such tumor antigens were isolated by the use ofcdna expression systems and biochemical approaches. The majority of tumor antigens identified to date have been found to be nonmutated, selfproteins. This raises important questions regarding the mechanism of antitumor activity and autoimmune disease. Several tumor-specific mutated tumor antigens have also been recently identified, which include cell cyclin-dependent Idnase 4 (CDK4) and -catenin. For the first time, a novel human cancer antigen was recently found to be generated by the use of an alternative open reading frame of the previously identified tyrosinase-related protein-i (TRP-i) gene. The identification of human tumor rejection antigens provides new opportunities for the development of therapeutic strategies against cancer. The potential clinical applications of these tumor antigens will be discussed. J. Leukoc. Biol. 60: ; Key Words: cyclin-dependens kinase 4. -catenin autoimmune disense iraerleukin-2 INTRODUCTION The immune system plays important roles in immunosurveillance aganist cancer and in tumor regression. The fine specificity and cognitive processes of anti-tumor immune responses can be mediated through B and T cells that recognize tumor antigens expressed on tumor cells. The importance of T cells in the regression of tumors has been demonstrated in experimental animal tumor models [1-3]. In the last few years it has been possible to generate cytotoxic T lymphocytes (CTL) from either peripheral blood lymphocytes (PBL) or tumor-infiltrating lymphocytes (TIL) derived from patients with cancer. The availability of these CTLs has allowed one to evaluate the role of T cells in tumor regression in humans. Early work by Rosenberg and co-workers clearly showed that adoptive transfer of cultured T cells derived from TIL along with interleukin-2 (IL-2) resulted in the objective regression of cancer [4, 5]. To further understand how T cells mediate anti-tumor responses and what molecules are seen by T cells on tumor cells, efforts to identify tumor antigens recognized by these CTLs with antitumor reactivity in vivo and in vitro have resulted in the molecular identification of a number of such tumor antigens in the past year. These studies provided the basis for understanding T cell-mediated antitumor activity, and a rationale for the development of anti-cancer vaccines. This review summarizes the current status and progress toward identifying human tumor antigens and their potential applications to cancer treatment. Approaches to Identification of Human Tumor Antigens The first approach that led to identification of tumor antigens was to stably transfect genomic DNA [6] or cdna libraries [7, 8] into cells expressing the appropriate MHC molecule. Positive clones were isolated on the basis of the ability to stimulate cytokine release from CTL. This method was recently improved by the use of cdna libraries that are transiently transfected into COS or 293 cells along with a cdna encoding the appropriate MHC molecule as shown in Figure 1 [9, 10]. Alternatively, cdna pools were transfected into 293 cells that stably expressed the matched MHC molecule [9]. The majority of genes encoding tumor antigens, which include MAGE [6, 1 1], GAGE [12], BAGE [13], tyrosinase [10, 14], TRP-i [15], CDK4 [16], and -catenin [17], were isolated with this approach. Drawbacks of this approach include the possibility of identifying cdna clones that encode cross-reacting peptides recognized by T cells due to the high expression level in the COS system, and the need to identify the antigenic peptides from the amino acid sequence of the antigenic proteins. Abbreviations: CTL, cytotoxic I lymphocytes; TIL, tumor infiltrating lymphocytes; IL-2, interleukin-2; TRP-1, tyrosine-related protein-i; PBL, peripheral blood lymphocytes; HPLC, high-pressure liquid chromatography; APC, antigen-presenting cells; TCR, T cell receptors; PCR, polymerase chain reaction; APC, adenomatosis polyposes coli gene. Reprint requests: Dr. Rong-Fu Wang, Surgery Branch, National Cancer Institute, Building 10, Room 2B42, Bethesda, MD Received March, 1996; accepted May 2, Journal of Leukocyte Biology Volume 60, September 1996

2 Tumor cells I + MHC-matcbed Andgen4oss cell One Genomic or edna library * Screening for positive clones by CTL DNA sequencing of positive edna clones and IdentIfIcatIon of antigenic epitopes Fig. 1. Gene cloning system for the identification of tumor antigens recognized by T cells. The second approach is based on biochemical purification. It is well known that T cells recognize a peptide bound to the MHC class I molecule. Therefore, peptides could be eluted with acid from either the tumor cell surface or purified peptide-mhc complexes. Different peptides were then separated by high-pressure liquid chromatography (HPLC) and tested for their ability to stimulate cytokme secretion from CTL when pulsed onto MHC-matched antigen-presenting cells (APC). if the amount of peptide in the positive fraction is sufficient to allow sequence determination by Edman degradation, a naturally processed peptide recognized by CTL can be directly identified [18]. Howe er, in most cases, it may be difficult to obtain suffident.smounts of peptides for sequence determination by the Edman degradation method because peptides have low binding affinity to MHC class I molecule or are present at relatively low levels on the cell surface. In those situations, mass spectrometry may be used to obtain the peptide sequence information [19]. The peptide sequence can then be used to search databases in order to find the gene encoding the antigenic peptide. Drawbacks to this approach include the technical complexities associated with the fractionation of the active peptides, the use of instruments such as tandem mass spectrometry, and the need to isolate a cdna encoding the identified peptide if the origin of the peptide is not known. The third approach is to test T cell recognition of peptides derived from a protein whose expression is known to be associated with tumor [20, 21]. To do so, it is necessary to generate CTLs in vitro using peptides derived from a putative tumor antigen [20-24]. However, these CTLs, in many cases, do not recognize the tumor cells [22, 23]. The possible explanation for this observation is that these CTLS may bear the T cell receptors (TCR) with low affinity to the MHC/peptide complexes. Alternatively, the peptide recognized by CTLs in in vitro systems may not be naturally processed and presented on the tumor cell surface. Although significant progress has been made in understanding the rules of peptide binding motifs for different MHC molecules [25], these peptides, which contain MHC binding motif may not, however, be naturally processed and presented on tumor cells at sufficient levels to be Wang and Rosenberg Human tumor antigens 297

3 recognized by T cells. Nevertheless, this approach may be useful to test whether putative tumor proteins such as ras and p53 are true tumor antigens. Tumor Antigens Tumor Antigens Expressed Only in Testes and Difterent Types of Cancer MAGE-1 was the first tumor antigen identified on a human using a genetic approach. A genomic library was made from the tumor cell line, and then was transfected into a MHC-matched, antigen-loss variant [6]. Stable transfectants were screened by CTLs isolated from the patient, who was repeatedly immunized with mutagenized autologous tumor. A gene, MAGE-1, was isolated from the transfectants on the basis of its ability to stimulate cytokine release from the CTLs [6]. MAGE-1 contains three exons and an open reading frame that encodes 309 amino acids. The peptide epitope, EADPTGHSY, was subsequently identified and recognized by CTL in the context of HLA-Ai [26]. Expression of the MAGE-1 gene was found in -40% of s as well as in a significant percentage of other tumors, but not detected in the normal human tissues except testis [6]. A second epitope peptide, SAYGEPRKL, was identified from the MAGE-1 protein and recognized by HLA-Cwi6-restricted CTLS [27]. DNA hybridization analysis indicated that MAGE-1 is a member of a multiple-gene family, among which MAGE-3 encodes a homologue of MAGE-1. A similar peptide, EVD- PIGHLY, derived from MAGE-3 was shown to be recognized by HLA-A1-restricted CTLs [11]. The second epitope peptide was identified by use of HLA-A2-restricted CTLs generated in vitro with a synthetic peptide derived from MAGE-3 [28]. Like MAGE-1, MAGE-3 was not expressed in human normal tissues except testis. Since MAGE-3 expression was detected in --6O% of s, it may represent a good candidate for the development of vaccine strategies. Two additional antigens were identified from a cdna library made from the same MZ2 cell line. A cdna clone encoding BAGE was isolated by use of a class I HLA-Cw16-restncted T cell clone [13]. This gene encodes a polypeptide of 43 amino acids and appears to be expressed in a similar pattern to members of the MAGE gene family. A peptide of nine amino acids was recognized by CTLs in the context of HLA-Cw16. Using the same approach, a gene encoding an antigen recognized by a class I HLA-Cw6-restncted T cell clone was isolated. This led to the identification of a new family of genes named GAGE [12]. Among them, GAGE-i and GAGE-2 encode a peptide that was recognized by CTLs in association with HLACw*O6O1. GAGE genes were found to have the similar expression pattern as the MAGE and BAGE genes [12] and its expression was detected in s as well as in other tumors. Recently, a cdna clone was isolated by screening cdna library using a HLA-A2-restricted T cell clone. The part of this cdna sequence was found to be identical to that of the previously published GnT-V gene encoding for N-acetylglucosaminyltransferase V protein. Interestingly, an antigenic peptide VLPDVFIRC recognized by T cells was encoded by a sequence located in an intron of GnT-V gene [29]. Although the fully spliced GnT-V mrna was detected in a variety of normal and tumoral tissues, the mrna containing the intron region coding for the antigenic peptide was not detected at a significant level in normal tissues. It should be noted that all antigens in this class (MAGE, GAGE, BAGE, and GnT-V) were identified from a cdna library-derived MZ2-mel tumor cell line, the biological significance of this phenomenon is unclear. However, they could be good candidates for the developemt of cancer immunotherapy because they are tumor-specific shared antigens. Shared Differentiation Antigens Four melanocyte proteins, tyrosinase, MART- 1/Melan-A, gploo, and TRP1, have been shown to represent antigens recognized by -reactive T cells. These antigens were isolated by screening edna libraries using TIL, which induced tumor regression when administered to autologous patients along with IL-2. Thus, they may represent tumor rejection antigens. The epitopes recognized by CTLs from corresponding genes were identified. Interestingly, the epitope recognized by T1L586 resulted from translational products of an alternative open reading frame of the same TRP-1 gene. Antigenic Peptides Produced by Normal Open Reading Frames Tyrosinase is the first member of differentiation antigens identified to date. A edna clone encoding tyrosinase as a tumor antigen was isolated by screening a cdna library from the SK29 [9]. Tyrosinase consists of 529 amino acids and is an important enzyme involved in the synthesis of melanin [30]. DNA sequence analysis mdicated that the cdna isolated from the 5K29 library represented a normal, nonmutated gene, suggesting that a normal gene encodes peptides for CTL recognition. Two distinct epitopes were identified on the tyrosinase molecule and recognized by CTLs in the context of HLA-A2: the first peptide (MLLAVLYLL) corresponding to the leader sequence of tyrosinase, and the second, YMNGTMSQV, corresponding to amino acids of the tyrosinase protein [31]. Recently, tyrosinase was shown to be recognized by an HLA-A24-restricted T1L888 derived from the patient 888 [14]. This TIL cell line was previously shown to result in the regression of multiple metastatic lesions when transferred into the autologous patient along with IL-2. The epitope peptide recognized by this TIL remained to be identified. However, a second CTL line derived from patient 1415 was also shown to recognize tyrosinase in the context of 298 Journal of Leukocyte Biology Volume 60, September 1996

4 HLA-A24 [32]. Exonuclease deletion experiments demonstrated that TIL 1415 and TIL 888 recognized different epitopes derived from the tyrosinase molecule. Subsequent studies showed that TIL1415 recognized the overlapping 9-mer and lo-mer peptides AFLPWHRLF and AFLPWHRLFL, corresponding to amino acids and of the tyrosinase sequence [32]. In addition, peptide SEIWRDIDF derived from the tyrosinase protein was recently shown to be presented to T cells by HLA-B44 [33]. These results indicate that multiple epitopes can be generated from the same tyrosinase protein and presented to T cells by different MHC class I-restricted molecules. While the role of CD8 T cells in tumor rejection has been clearly demonstrated, CD4+ T cells may also play an important role in the immune response against tumor. First, a significant proportion of cell lines express MHC class II molecules. Therefore, cells can potentially present the antigenic peptide to CD4 T cells by their MHC class II molecules. Alternatively, professional antigen-presenting cells can present peptides to T cells by the uptake of antigens. Second, CD4 TIL derived from both and breast cancers have been shown to respond to the autologous tumor cells [34, 35]. In support of the notion that CD4 T cells indeed recognize an antigenic peptide presented by MHC class II molecules, evidence was presented that CD4+ reactive TIL1O88 recognized an antigenic peptide derived from the tyrosinase protein in a HLA-DR4- restricted fashion [36, 37]. Because the TIL1O88 cell line derived from patient 1088 was shown to recognize extracts derived from multiple and melanocyte cell lines, this suggests that TIL 1088 recognizes a nonmutated tyrosinase gene product. Two distinct peptides (QNILLS- NAPLGPQFP, DYSYLQDSDPDSFQD) were recently identified to be recognized by the CD4 TIL1O88 [38]. Therefore, the tyrosinase gene product contains multiple CD8 T cell epitopes as well as CD4 T cell epitopes. The existence of other MHC class Il-restricted tumor antigens from and breast cancers was recently reported [34, 35, 39]. The MART-i ( antigen recognized by T cells- 1) gene was recently isolated following the screening of a cdna library with the HLA-A2-restricted -reactive TIL 1235 [7]. MART-i is identical to Melan-A, which was independently cloned by another group [10], and encodes a protein of 1 18 amino acids. The MART-i edna isolated from cells did not contam any mutation or other alterations. Northern blot analysis indicated that MART-i was expressed only in s, melanocytes as well as retina, but not in normal human tissues and other types of cancer cell lines including colon, breast, neuroblastoma, sarcoma, and renal cell lines [7]. These observations were confirmed by immunohistochemical studies using a monoclonal antibody [Y. Kawakami and S. A. Rosenberg, unpublished data]. The MART-i gene product was found to be an immunodominant antigen recognized by the majority of HLA- A2-restricted reactive CTL established from TIL in the Surgery Branch of the National Cancer Institute [7, 40] as well as a large percentage of -reactive clones derived from the PBL of HLAA2+ patients [41]. Out of 21 HLA-A2 restricted reactive TIL were tested for recognition, only 3 failed to recognize MART-i [40]. By screening a number of synthetic peptides with the HLA-A2 binding motif, one 9-mer peptide, AAGIGILTV (M9-27), and two overlapping 10- mer peptides, EAAGIGILTV and AAGIGILTVI, were found to be recognized by HLA-A2-restricted specific CTL [40]. The M9-27 peptide was found to have an intermediate binding affinity to HLA-A2.i, whereas both of the overlapping io-mer peptides had low binding affinities. T2 cells could be sensitized with lower concentrations of the M9-27 peptide than with the overlapping 1O-mer peptides, suggesting that M9-27 may represent the predominant peptide from MART-i recognized on the cell surface in the context of HLA-A2. Using the same approach, a cdna clone was isolated and shown to be almost identical to the two genes, gpioo and Pmeli7, which had previously been cloned using anti antibodies [8, 42, 43]. One of the gpioo cdna clones that had previously been isolated based on its recognition by antibody [43, 44] was capable of stimulating cytokine release from HLA-A2-restricted TIL when cotransfected into COS cells along with the HLA-A2 cdna [8, 42]. Northern blot analysis demonstrated that gploo was expressed in neonatal cultured melanocyte lines, most cell lines and retina, but not expressed in non cancer cell lines including breast, colon, lung, or cancers of neuroectodermal origin [8]. The gploo molecule appears to represent a highly immunogenic antigen, since it was found to be recognized by 8 out of 21 HLA- A2-restricted reactive TIL established from different patients in the Surgery Branch of the National Cancer Institute. Some CTL lines could recognize multiple gploo epitopes. Five gploo epitopes were identified by CTLs derived from different patients [8, 45, 46]. A biochemical approach was also used by other groups to identify antigenic peptides recognized by CTLS. One gpioo epitope, YLEPGPVTA, was independently isolated from an HPLC-purified fraction of peptides obtained from HLA-A2 molecules on cells as a antigen [19]. This peptide was recognized by five out of five HLA-A2-restricted -specific CTLs raised from peripheral blood lymphocytes. The importance of this observation was considerable because the peptide identified by two different approaches (i.e., genetic and biochemical) represented a naturally processed peptide on tumor cells. When HLA-A2-binding peptides were isolated from cells and fractionated on a reversed-phase HPLC column, three distinct fractions were found to stimulate a single MART-i-specific T cell clone [47]. One of these fractions may have contained the same peptide as AAGIGILTV (M9-27) derived from MART-i, suggesting that this peptide is also naturally processed and presented by MHC class I molecules to T cells [47, 48]. A predominant fraction was also found to be reactive to CTLS and Wang and Rosenberg Human tumor antigens 299

5 may contain a peptide corresponding to the MART-i peptide ILTVILGVL (M9-32), which overlaps with the four carboxyl terminal residues of M9-27 and has a similar HLA-A2 binding affinity [47]. However, the M9-32 peptide was recognized only when the peptide was pulsed onto the target cells in the presence of anti-hla-a2 antibody and 32-microglobulin. Antigenic Peptides Translatedfrom an Alternative Open Reading Frame Recent studies demonstrated that T1L586 derived from patient 586 recognized one or more antigens in the context of HLA-A31 [49]. When a cdna expression libraty was then transiently transfected into COS along with the gene encoding HLA-A3i, a partial cdna clone was isolated [15], which was identical to that of the previously reported TRP-i gene [50, 51]. The protein encoded by TRP-i/gp75 was originally identified as an antigen recognized by immunoglobulin G antibodies in the serum of a patient with [52]. Northern blot analysis mdicated that the TRP-i gene was expressed in, normal melanocyte cell lines, and retina, but not in other normal tissues tested [15]. This was in agreement with results obtained with monoclonal antibodies, which mdicated that the gp75 protein was found to be one of the most abundant intracellular glycoproteins in melanocyte-lineage cells, but was not detected in non-melanocytic cell types [53, 54]. The gp75 molecule has recently been shown to have DHI-2-carboxylic acid oxidase activity involved in the synthesis of melanin [55]. To determine the identity of the epitope in gp75, small cdna fragments generated by exonuclease deletion and polymerase chain reaction (PCR) were identified to confer recognition by T cells when cotransfected with the HLA- A3i cdna. A number of synthetic peptides were made based on the predicted amino acid sequence of gp75 and previously identified peptide binding motif for HLA-A3i [56]. However, none of the peptides tested were found to stimulate cytokine release from TIL586 when pulsed onto HLA-A3i-positive EBV-transformed B cells [57]. One possible explanation for this is that the epitopes recognized by TILS86 may result from the gene product translated from an alternative reading frame of gp75 because a small DNA fragment, which did not contain any ATG codons in the normal gp75 open reading frame (ORF1), was still capable of stimulating cytokine secretion from TILS86. Two ATG codons in a relatively good context were, however, present in different open reading frames relative to the normal gp75 ORF1 reading frame (Fig. 2). To explore this possibility, three peptides derived from ORF2 and two peptides from ORF3 were selected and synthesized on the basis of the HLA-A3i binding motif. Surprisingly, the peptide MSLQRQFLR, which was derived from ORF3 was capable of stimulating cytokine release from T1L586 when pulsed onto HLA-A3 1-positive EBV-transformed B cells (Fig. 2) [57]. A number of studies showed that protein translation can he initiated at more than one AUG codon [58, 59]. In rare cases, translation can be initiated at both AUG and non- AUG codons, such as CUG; however, in those situations, the same reading frame is utilized for translation of the different gene products [60]. To our knowledge the use of overlapping open reading frames to translate two distinct polypeptides from a single eukaryotic cellular mrna had never been reported, although translation of overlapping reading frames have been reported in viral mrnas [61-63]. Another possibility is that the epitope peptide recognized by TILS86 might be generated from a frame shift as a result of a deletion or insertion. A recent report indicated that the T cell epitope peptides derived from the frameshift of the mutated adenomatosis polyposis coli (APC) gene in colon cancer were recognized by CTLs generated from vaccinated BALB/c mice [64]. However, site-directed mutagenesis and DNA sequence analysis ruled out the possibility that the antigenic peptide recognized by T1L586 was derived from a frameshift product of gp75. Indeed, the ATG start codon of ORF3 was required for generating an antigenic peptide recognized by T1L586 [57]. These results demonstrated a novel mechanism by which a human tumor antigen can be generated from an alternative open reading frame and presented to T cells by an MHC class I molecule. Recently, several related examples were reported of the use of alternative open reading frames. For instance, pi6 was demonstrated to encode an inhibitor of the cyclin D-dependent kinases CDK4 and CDK6 [65]. Thus, its function is to block the phosphorylation of the retinoblastoma protein (prb) and to prevent cell cycle progression from the Gi to S phase [66]. pi6 probably acts as a tumor suppressor since deletions and mutations of this gene were found in a variety of tumors [66]. Interestingly, a dissimilar protein was recently identified from an alternative open reading frame of the mouse pi6(4a gene to encode a i9-kda (p19arf) protein [67]. The ectopic expression of p19arf in fibroblasts was found to induce Gi and G2 phase arrest. Therefore, the p16 gene encodes two dissimilar proteins, both of which may be required for cell cycle regulation. A second example of this phenomenon is an antigenic peptide recognized by CTLS obtained from the acid extract of mouse RLoi tumor cells. Sequence analysis by Edman degradation indicated that the CTL-recognizing RLo1 peptides were derived from the 5 untranslated region of c-akt oncogene [68], suggesting that c-akt may encode a peptide translated from a cryptic initiation codon of the same transcript. In another study, a truncated cdna clone was isolated from a mouse spleen cdna library using lacz-inducible T cells as a probe [69]. This cdna clone was shown to encode a truncated form of a-tubulin. Deletion and mini-gene constructs with an ATG start codon in frame with the a-tubulin failed to identify the epitope recognized by T cells. However, a peptide derived from a non-atg-defined alternative open reading frame was found to be recognized by T cells. A mini-gene con- 300 Journal of Leukocyte Biology Volume 60, September 1996

6 ORF]. ORF1 ORF1 ORF1 ORF1 ORF2 ORF3 GCXTAAACTCCTCTCTGGGCTGTATCTI CTICCCTrGCTACT1 P1 1 CAG 60 MSAPKLLSLGCIFFPLLLFQ 120 QARAQFPRQCATVEALRSGM 180 CCPDLSPVSGPGTDRCGSSS 240 GRGRCEAVTADSRPHSPQYP 300 H D G RD DR E VW P L R F F N R T CR MI G R S G P CA S SI G HVT NIL ORF1 0RF2 ORF3 C NGNF S G HNC G TC R P GWR GA A TA IS Q D T TV G RAV LAG EEL 0 ROT LRT Q LW DV PSW L ER SC.360 ORF1 ORF2 ORF3 ORF1 420 A CD Q RVL IV R RN L L DL SK E E P V T R G F S STOP L STOP Apal AAGAACCACT1 Jx:CtrrGGATAToGcAAAGc:GCACAACTCACCCT...ATA 1584 KNHFVRALDMAKRTTHP Fig. 2. The nucleotide, amino acid sequence, and open reading frames of the gp75 gene. The partial nucleotide and amino acid sequences of the first 157 amino acids are shown from the start codon for translation of ORF1 (gp75). The DNA fragment that conferred the ability to stimulate GM-CSF release from T1L586 is underlined. Two putative start codons, ATG ( ) and ATG ( ), are in bold and may result in the translation of ORF2 and ORF3, respectively. The peptide sequence recognized by T1L586 from ORF3 is in bold and underlined. struct with a methionine codon for translation followed by a sequence encoding ASVVEFSSL was found to be at least 100-fold more active than the parental construct. This ohservation was further confirmed using a synthetic peptide pulsed on H-2 Kbpositive cells, suggesting that a cryptic translation product may be presented to T cells by an MHC class I molecule. Tumor Antigens Expressed in a Variety of Tissues The majority of human antigens identified thus far are tissue specific. However, one -specific T cell line was found to recognize an antigen whose expression is ubiquitous in normal tissues. Using cdna expression cloning, a cdna clone, termed p15, was isolated and recognized by TIL129O in an HLA-A24-restricted fusion [70]. This gene was not previously described in the available databases and appeared to be expressed in a wide variety of normal tissues based on Northern blot analysis. A non-mutated peptide, AYGLDFYIL, was identified as a T cell epitope derived from the p15 protein. At the present time it is not clear whether this peptide represents a true T cell epitope or this peptide is cross-reactive with the product derived from an unknown gene but is recognized by T1L1290 on tumor cells. Alternatively, the p15 gene is overexpressed in tumor cells. The low level of expression in normal tissues may not generate a sufficient amount of MHC/peptide complex on the cell surface to be recognized by T cells. HER-2/neu gene was recently identified as a shared tumor antigen recognized by T cells in breast and ovarian cancers, suggesting that ubiquitously expressed antigens can serve as an immune target. The HER-2/neu proto-oncogene encodes a tyrosine kinase protein whose expression has been shown to be increased in a number of breast and ovarian cancers. In breast cancer, HER-2/neu overexpression is associated with aggressive disease. Cytotoxic T lymphocytes isolated from tumor-associated lymphocytes can specifically recognize a synthetic peptide corresponding to amino acids of the HER-2/neu protein [71]. This demonstrated that CTLs isolated from human tumors recognized HER-2/neu as an ovarian tumor antigen. To investigate whether the identified peptide ( ) from HER-2/neu can induce anti-tumor CTLS from PBL, peptide-specific CTLS were generated in vitro using this peptide and shown to lyse HLA-A2 and HER-2/neu ovarian tumor cells but not K562 cells [72], suggesting that the peptide epitope ( ) may be naturally processed and presented by the HLA-A2 molecule on tumor cells. Recently, four ovarian tumor-reactive CTLs were established from different HLA-A2 patients and were capable of recognizing both freshly isolated HER-2/neu tumor Wang and Rosenberg Human tumor antigens 301

7 cells and non-hla-a2 ovarian tumor lines transfected with HLA-A2 cdna [21]. To evaluate whether these CD8 CTL recognized the same epitope or different HER-2/neu peptides, 19 synthetic peptides with the HLA-A2 peptide binding motif were tested for lysis of T2 cells by these CTLs. Interestingly, a common epitope (amino acid : KIFGSLAFL) was found to be recognized by four out of four CTL lines [21]. Another epitope peptide ( ) was also found to be recognized by two out of four CTL lines. These results were in agreement with data obtained by other groups [73, 74]. Recognition and lysis of ovarian cancer cells by CTLS derived from patients with ovarian cancer were showed to be correlated with the expression level of HER-2/neu in the tumor cells [73]. Most importantly, it was shown that breast and ovarian cancerspecific CTLS recognized the same epitope peptide (GP2; amino acids ) derived from the HER-2/neu protein in the context of HLA-A2 [75, 76]. It appears that the GP2 peptide represents a common epitope shared by different epithelial tumors because it was recognized by CTL lines derived from breast, ovarian, non-small lung, and pancreatic cancers [76]. Another tumor antigen in this class is the mucin-associated epitope from the muc-i gene product. Its expression has been shown to be associated with breast and pancreatic adenocarcinomas [77, 78]. The muc-i gene is expressed on epithelial cells, fibroblasts, and B cells, and can serve as a target for T cell recognition. However, T cell recognition of the muc-i gene product appeared to be non-mhc restricted [79]. The epitopes for T cell recognition were found in the tandem repeat of the muc-i protein [80]. Tumor-Specific Antigens Cancer cells may arise from the accumulation of genetic alterations in a single clone of cells. Many mutations have been found in tumor suppressor genes such as ras and p53 as well as genes such as p16 involved in cell cycle control in tumor samples. The mutated proteins or peptides could be immunogenic and be seen as foreign by the host immune system. Mutated peptides have been identified in several mouse tumors as T cell antigens [18, 81, 82]. For example, in the mouse Lewis lung carcinoma, a peptide recognized by CD8+ CTLs was isolated and found to correspond to a mutated peptide of connexin 37 [18]. This mutation led to a change of one amino acid in the coding region and enhanced the binding of the mutated peptide to H-2 Kb, but the normal peptide failed to do so. In another study, CD4+ T cells appeared to recognize a unique antigen in a mouse UV-induced tumor. Using a biochemical approach, a purified protein was found to stimulate the CD4 T cells when loaded in antigen-presenting cells. A 1 Ii.. in ]#{149}: 11.hI.I.Ip Negative Clonal selection: deletion p Positive selection No selection: Apoptosis due to negligible affinity Fig. 3. T cell selection in the thymus and peripheral based on the avidity of self peptides and T cell activation required for immune response in the periphery. The avidity of seif-peptides is determined by many factors, but the binding affinity of self-peptide to MHC and its concentration or density are major determinants. T cells specific for high-affinity seif-peptide for MHC may be clonally deleted or anergized, and T cells for self-peptide with no or little affinity for MHC cannot survive. Only those T cells reactive with seif-peptides with low, but high enough affinity for their MHC are positively selected and escape self-tolerance. These T cells migrate into the periphery, some die by apoptosis or anergy due to improper stimulation, others are activated by increase in avidity of seif-peptides due to overexpression of self-antigens (tumor antigens), efficient antigen processing, and presentation or immunization by specific self-antigens. CD4 T cells also play important roles in modulating the immune system. Activation of self-reactive T cells may result in tumor destruction and/or autoimmune response. 302 Journal of Leukocyte Biology Volume 60, September 1996

8 TABLE 1. Human Tumor Anticens Recoanized by T Lvmohocvtes Antigens MHC element Tumor Epitope Ref. Antigens expressed in many tumor types MAGE-1 HLA-A1 HLA-Cw16 MAGE-3 HLA-A1 HLA-A2 BAGE HLA-Cw16 GAGE-i HLA-Cw6 GnT-V HLA-A2 EADVFGHSY SAYGEPRKL EVDPIGHLY FLWGPRALV AARAVFLAL YRPRPRRY VLPDVFIRC [26] [27] [11] [28] [13] [12] [29] Differentiation antigens expressed in melanocyte lineage cells A. Epitopes derived from the normal proteins MART-i/Melan-A HLA-A2 HLA-A2 gpioo/pmeli7 HLA-A2 HLA-A2 HLA-A2 HLA-A2 HLA-A2 tyrosinase HAL-A2 HAL-A2 HLA-B44 HLA-A24 HLA-DR4 HLA-DR4 B. Epitopes derived from an alternative reading frame TRP-1/gp75 HLA-A31 Tumor-specific antigens -Catenin HLA-A24 CDK4 HLA-A2 MUM-i HLA-B44 AAGIGILTV ILTVILGVL KTWGQYWQV ITDQVPFSV YLEPGPVTA LLDGTATLRL VLYRYGSFSV MLLAVLYLL YMNGTMSQV SEIWRDIDF AFLPWHRLF QNILLSNAPLGPQP DYSYLQDSDPDSFQD MSLQRQFLR SYLDSGIHE ACDPHSGHFV EEKLIVVLF [40] [47,48] [45] [45] [19] [8] [45] [31] [31] [33] [32] [38] [38] [57] [17] [16] [88] Tumor antigens ubiquitously expressed HER-2/neu HAL-A2 breast/ovarian HAL-A2 breast/ovarian HAL-A2 breast/ovarian p15 HLA-A24 MUC-1 Non-MHC breast/ovarian KIFGSLAFL ELVSEFSRM IISAVVGIL AYGLDFYIL [21] [21,71] [75,76] [70] [78] After cloning and sequencing of the gene encoding the peptide, it was found that the CD4 T cells recognized a mutated form of ribosomal protein L9. It is of interest to test whether mutated ras and p53 are tumor-specific antigens. Several groups have raised CTLs against normal or mutated peptides from the ras proto-oncogene [83] and p53 tumor suppressor gene [20, 84-87]. However, in most cases these CTLs failed to recognize tumor cells. One possibility is that peptides selected based on the peptide binding motif are not naturally processed or that the density of MHC/peptide complexes is too low on tumor cells to allow recognition by these CTLS. It should be noted that, although many mutations have been identifled in ras and p53, in reality, only a very small subset of mutations enhance, if they are processed, the MHC binding affinity or T cell recognition. Recently, mutated gene products as tumor-specific antigens recognized by CTL derived from patients have been reported in human cancers by screening cdna libraries using reactive CTLs. A cdna clone was recently isolated following the transient transfection of COS cells with HLA-B44 and pools of cdnas derived from the LB33 cell line. This gene was named MUM-i (-ubiquitous mutated). The peptide epitope, EEKUYVLF, was found to be recognized by CTL [88]. By comparison of the DNA sequence encoding the peptide in the tumor cells to that in the normal lymphocytes of the patient, a point mutation was found in the sequence of the cdna isolated from the tumor, which led to a change of one amino acid (Ser to lie) at position 5 of the peptide. The normal gene product was not recognized by the CTL, suggesting that this mutation was responsible for recognition of peptide by T cells. Because both the normal and mutated peptides bound efficiently to the class I HLA-B44 molecule, but only the mutated form could be recognized by T cells, suggesting that this mutation may have an effect on T cell recognition. Further analysis indicated that the antigenic peptide spanned the intro-exon boundary of an incompletely spliced message. A second mutated gene was isolated and shown to be recognized by three class I HLA-A2-restncted T cell clones isolated from patient SK29 [16]. This Wang and Rosenberg Human tumor antigens 303

9 gene encodes cyclin-dependent kinase 4 (CDK4), an enzyme involved in cell cycle regulation. A point mutation (a C to T transition) led to a substitution of a cysteine for an arginine residue at codon 24. The io-mer normal peptide ACDPHSGHFV and mutated peptide ARDPHSGHFV were synthesized and tested for their ability to sensitize target cells for lysis. It was found that the mutated peptide was capable of sensitizing the target cells at 100- to fold lower peptide concentration than the normal peptide. The CDK4 protein normally forms a complex with cyclin Di and phosphorylates the prb protein, and promotes the cell cycle from Gi to S phase [66]. However, assembly of CDK4 with cyclin Di as well as its kinase activity were inhibited by pi611 mentioned above. When the mutated CDK4 protein was expressed in insect cells along with cyclin Dl and other CDK binding proteins such as pl645, it was found that the mutated CDK4 failed to bind to pi6h!4a, implying that the mutation in the CDK4 gene may lead to a loss of cell cycle control. A third mutated gene was also recently identified and shown to be recognized by T1L1290, which was derived from a recurrent tumor in patient 888 [17]. Partial cdna sequence analysis indicated that this gene encoded a mutated -catenin gene product. A point mutation was found in the sequence of the cdnas isolated compared to that previously published. This mutation led to a change of serine to phenylalanine in the coding region. The mutated -catenin peptide SYLDSGIHF was found to induce half-maximal lysis of target cells at a concentration of 1 pm, whereas a 1 mm concentration of the normal peptide (SYLDSGIHS) was required for an equivalent level of sensitization. Thus, it appeared that the mutated peptide was recognized by TIL129O about one million-fold more efficiently than the normal peptide. Competitive peptide binding assay indicated that this substitution resulted in an increase in binding affinity of the mutated peptide to HLA- A24 molecule. The 3-catenin protein has been shown to be a cytoplasmic protein that interacts with the cellular adhesion molecule E-cadherin [89]. A number of mutations have been found in the -catenin gene product from different tumors [90, 91]. Loss of cell adhesion molecules may play a role in the metastatic #{149} process [92]. Generation of Tumor-Specific CTL in vitro The availability of the identified epitope has made possible the generation of CTLs from PBL in vitro using synthetic peptides. It was demonstrated that CTLS generated from PBL in vitro using the synthetic peptide derived from MAGE-3 were capable of recognizing HLA-Ai tumor cells expressing MAGE-3 [24]. A similar approach was taken to generate CTLS from HLA-A2 PBL in vitro using a peptide FLWGPRALV derived from MAGE-3 [28]. Interestingly, these CTLS were found to lyse tumor cell lines only after the tumor cells were treated with interferon-y. This treatment may have resulted in an increase in HLA class I gene expression. Peptide-specific CTLs were also generated using PBL from a patient and a peptide derived from the MACE-i protein [93]. When the epitope peptide derived from tyrosinase was used to generate CTLS in vitro using PBL from normal donors, it was found that the majority of specific CTL lines generated were capable of recognizing HLAA2+ s [24, 94]. In general, some CTLS raised in vitro may not necessarily recognize cells simply because the peptide itself may not be naturally processed or presented on the cell surface in a relatively low level. Alternatively, these CTLs bear low affinity of TCR for MHC/peptide complex. Peptide-speciflc CTLs were also induced from PBL with the M9-27 peptide derived from the MART-i protein, which possessed intermediate HLA-A2 binding affinity. CTLs induced with the M9-27 peptide recognized HLA-A2 cells which confirmed that the M9-27 peptide was naturally processed and presented on tumor cells [95]. Melanoma-reactive CTL could be raised from seven of eight patients PBL and two of four normal donor PBL by stimulating with the M9-27 peptide. The induction of -reactive CTL from PBL of patients occurred more rapidly than from normal donors, suggesting that there was an increase in the CTL precursor frequency in patients [95]. Although -reactive CTL could be induced by in vitro stimulation with gpioo peptides, the induction was not as efficient as with the MART-i M9-27 peptide [95, 96]. The in vitro induction of -reactive CTL with gpioo peptides required more restimulations than were needed with the MART-i M9-27 peptide, indicating that the precursor frequency of CTL, which were specific for the gpi0o epitopes may be lower than the frequency of MART-i-specific T cells. The epitope peptides identified thus far from antigens generally have intermediate or relatively low binding affinity, but it has been demonstrated that modified peptides can enhance the binding affinity of peptides to MHC molecules and therefore improve their immunogenecity [97, 98]. To improve the binding affinity, attempts were made to modify the anchor residues at P2 and P9 of the peptides [99]. Following the screening of many modified peptides, peptides were identified which possessed a significantly better HLA-A2 binding affinity than the parental peptides but were still recognized by -reactive TIL specific for the native epitopes. When tested for their ability to stimulate CTL responses in vitro, some of the modified epitopes were found to be more efficient in generating CTLS than were the natural epitopes [99]. These results indicated that some of the modified peptides were more immunogenic in vitro than the native peptides, which may also be true in vivo when patients are immunized with these modified peptides. The ability to generate CTLS in vitro using synthetic peptides has two important implications: (1) these antigenspecific CTLs could be used for the treatment of patients with cancer (see below); (2) CTLs could be generated with synthetic peptides that are derived from a putative tumor 304 Journal of Leukocyte Biology Volume 60, September 1996

10 TABLE 2. Cancer Therapies Based on the Molecular Identification of Cancer Antigens 1. Active immunotherapy with: a. Immunodominant peptides 1) combined with adjuvants 2) linked to helper peptides, lipids or liposomes 3) pulsed onto antigen-presenting cells such as dendritic cells b. Modified antigenic peptides c. Proteins alone or combined with adjuvants d. Naked DNA encoding cancer antigens 1) gene gun for intradermal injection 2) intramuscular injection 3) linked to lipids e. Recombinant vaccinia, fowlpox, or adenovirus encoding 1) cancer antigens alone 2) cancer antigens plus genes encoding cytokines, costimulatory molecules, or other genes to enhance the immune response f. Recombinant bacteria such as BCG, salmonella, or listeria encoding cancer antigens alone or other molecules 2. Active immunotherapy (above) followed by the administration of imrnunostimulatory cytokines 1) IL-2 2) IL-6 3) IL-b 4) IL-i2 5) IL-iS 3. Passive immunotherapy with anti-tumor lymphocytes raised by in vitro sensitization of IlL or PBL to 1) immunodominant peptides pulsed onto antigen-presenting cells 2) antigenic proteins coincubated with antigen-presenting cells (exogenous antigen-presenting pathway to raise CD4 cells) antigen and then tested for their ability to recognize the tumor cells. In this way, a tumor antigen might be identified. TCR Usage of Antigen-Specific CTLs The molecular definition of tumor antigens provide opportunities to study the mechanism of T cell recognition of MHC-peptide complexes. Direct assessment of the TCR repertoire in CTLS that recognized the same tumor antigen peptide may have important implications for understanding the interaction between a TCR receptor and a given MHCpeptide complex, and for monitoring the immune response against cancer following vaccine treatment. The TCR usage of three CTL clones that recognized the M9-27 peptide in association of HLA-A2 was analyzed and V7 was found to be utilized by two of the CTL clones [100, 101]. In another study, analysis of eleven HLA-A2-restricted CTLS that recognized MART-i and were derived from different patients showed that Va2, V37, and Vi4, particularly the combination of Va2/Vfi4, was frequently used by TCR [41]. A similar study was also conducted to analyze the TCR usage in CTL clones that recognized MACE-i in association with HLA-Ai. It was found that three clones expressed different TCRs and they were differentially responsive to amino acid substitutions in variant synthetic peptides [102]. Thus, different TCR are capable of recognizing a single peptide that is presented by the same MHC molecule. Analysis of CTLS that recognized either a peptide from MACE-i or a peptide from BAGE in an HLA-Cwi6-restricted fashion indicated that these T cells shared TCRV3 usage [27]. A study was carried out to determine the TCR usage of CTL clones that recognize HER2/neu. It was found that TCRV33 and V36 were predominantly used in these T cell clones [i03j. Data on TCR usage is still limited, general conclusions cannot be drawn at the present time. Further studies may provide the basis for better understanding the mechanisms that regulate in vivo anti-tumor immune responses. In addition, it may be possible to generate therapeutic cells by the transfer of TCR genes from tumor-reactive T cell clones to alternative effector cells or to hematopoietic stem cells [101]. Implications for Cancer Therapies The identification of tumor antigens marks a major step forward in the understanding of T cell-mediated antitumor activity and provides an opportunity for the development of new strategies for the treatment of cancer. However, the majority of tumor antigens identified thus far are self-proteins, raising a fundamental question regarding the mechanism of how the immune system breaks self-tolerance. It has been proposed that the recognition of self-peptide bound to MHC molecules in the thymus is critical for the process of positive selection. Evidence for this has been presented in recent studies. Using transgenic mice deficient in either the transport-associated antigen processing (TAP1) or -microglobulin, several groups have successfully demonstrated that MHC-bound peptide is required for T cell selection and proposed a differential avidity model Wang and Rosenberg Human tumor antigens 305

11 of thymic selection [104, i05]. In accordance with this model, low-avidity recognition of self-peptide results in a positive selection, whereas high avidity interactions proyoke the induction of tolerance [106]. The low avidity interaction could result from the expression of TCR with a low affinity to seif-peptide. Alternatively, if T cells bear a high-affinity TCR, the low avidity recognition could still be achieved by the poor presentation of seif-peptide due to either quantitatively sparse self-peptide with high affinity for MHC or abundant self-peptides with negligible affinity for their MHC molecules (Fig. 3). Therefore, it is predicted that self-peptides recognized by T cells as tumor antigens may display low affinity for their MHC because these T cells can escape self-tolerance. This is supported by recent findings obtained by analysis of the identified nonmutated cancer peptides. The majority of these peptides as shown in Table 1 contain non-dominant anchor residues at positions 2 or 9 compared to MHC-peptide binding motifs and exhibit intermediate or low affinity for their MHC. Using normal or transgenic mice, it was found that T cell tolerance is efficiently induced with high-affinity seif-peptides (dominant self-determinants) for MHC molecules [107, 108]. Some self-peptides in Table 1 display high affinity for MHC but these peptides may fail to induce self-tolerance due to their crypicity or subdominance of these sell-peptides. Therefore, an ideal tumor antigen for use in a cancer vaccine should have a high enough binding affinity to MHC for efficient presentation, but sufficiently low affinity to preclude the induction of self-tolerance. Because positive or negative selection in the thymus is incomplete, peripheral selection is very important for generating self-reactive T cells. The avidity required for peripheral T cell activation is much higher than that required for T cell selection in the thymus [109]. In normal situations, these self-reactive peripheral T cells do not respond to self-peptides and tissues that express these self-peptides because the density of self-peptide/mhc complex is below the threshold required for activation. These resting T cells may be deleted or become anergy. However, many factors including inflammation, tissue destruction at tumor sites, antigen processing, up-regulation of MHC or co-stimulatory molecules, cytokine release, and immunization with peptide vaccines may significantly increase the density of self-peptide/mhc complex on the cell surface [107, 110]. These self-reactive T cells are activated and may mount an antitumor response as shown in Figure 3. For example, overexpression of tumor antigens such as HER2/neu protein in breast and ovarian cancers may result in HER2/neu-reactive T cells in patients. In addition to an increase in peptide concentrations, modification of self peptide to enhance their MHC binding affinity but still being recognized by T cells is an alternative way to increase the avidity of self-peptides. In general, immunogenicity has been shown to be correlated with the MHC binding affinity [98, 99, 107]. Covalent linkage of an antigenie peptide to its MHC molecule by genetic means may also improve the MHC binding affinity [X. Q. Kang, unpublished data]. A potential consequence of active immunization with cancer peptides is the development of autoimmune disease [iii, ii2]. Development of vitiligo has been found to be correlated with good prognosis or clinical responses to immuno-chemotherapy in patients [1 13-i 15]. The adoptive transfer of TIL into patients with along with IL-2 resulted in the objective regression of tumor, but only occasionally resulted in depigmentation of skin of the treated patients, suggesting that treatment of patients with anti-self T cells does not necessarily cause tissue destruction in normal organs [8, 49]. Recently, Houghton and co-workers have demonstrated that passive immunization with a mouse monoclonal antibody (TA99) against gp75 induced protection against and rejection of the gp75 Bi6FiO in syngeneic mice [116]. There was no evidence of decrease in pigmentation, inflammation, or changes in cellular morphology or tissue architecture in the eyes of mice treated with antibody [1 16], suggesting that the threshold required for tumor regression is lower than that for normal tissue destruction. Identification of new tumor antigens has made it possible to induce tumor-specific immune responses against cancer. Strategies for the development of effective anti-tumor immunotherapies are summarized in Table 2. Active immunotherapy involves the direct immunization of cancer patients with cancer antigens in an attempt to boost immune responses against the tumor. The immunodominant peptides derived from tumor antigens could readily be synthesized in vitro and used for immunization either alone or in a form intended to improve their immunogenicity such as in combination with adjuvant, linkage to lipids/liposomes or helper peptides, or pulsed onto antigenpresenting cells. It has recently been reported that peptide-pulsed dendritic cells induce antigen-specific antitumor immune responses in mice [117]. Modification of the immunodominant peptides to improve binding efficiency to MHC antigens can potentially increase immunogenicity and induce a stronger antitumor activity. The most effective forms of cancer vaccine involve the incorporation of genes encoding tumor antigens into recombinant plasmid or viruses such as vaccinia, fowlpox, or adenovirus. For example, recombinant adenoviruses encoding either MART-i or gpio0 have been generated [1 i8] and are being used in clinical trials at the Surgery Branch, National Cancer Institute. Because tumor cells transduced with genes encoding cytokines or costimulatory molecules have been shown to elicit in vivo antitumor activity, the combination of cancer antigen genes with other genes encoding cytokines such as IL-2, costimulatory molecules such as B7.i may enhance the immune response following infection. Alternatively, cancer vaccines using recombinant virus encoding tumor antigens can be enhanced by the exogenous administration of immunostimulatory cytokines [ii9]. Naked DNA vaccines are an alternative to the use of recombinant viruses. It has been 306 Journal of Leukocyte Biology Volume 60, September 1996

12 shown that injection of plasmid DNA encoding model tumor antigens directly into muscle or into the skin resulted in both cellular and humoral immune reactions [i20]. Because many epitopes derived from tumor antigens have been identified, it is now relatively easy to generate peptide-specific CTL in vitro as described above. For example, by repeated in vitro sensitization with the MART-i peptide, peptide-specific CTLs were generated with 50-iOO times greater potency on a per cell basis than TILs derived from patients using standard techniques [95]. The adoptive transfer of these CTL may be more effective in mediating tumor regression in vivo compared to conventionally grown TIL. Techniques are currently being developed to generate large numbers of cells with specific antitumor reactivity for use in the adoptive immunotherapy of human cancer. Finally, since successful vaccination requires both CD4 and CD8 T cells, it is important to develop systems for cloning MHC class Il-restricted tumor antigens. Utilization of MHC class I and II restricted cancer peptides combined with cytokines and costimulatory molecules may mount an improved antitumor response against human cancer. ACKNOWLEDGMENTS We would like to thank Drs. Y. Kawakami and P. F. Robbins for their critical reading of this manuscript and helpful discussion. REFERENCES 1. Rosenberg, S. A., Spiess, P., Lafreniere, R. (1986) A new approach to the adoptive immunotherapy of cancer with tumor-infiltrating lymphocytes. Science 233, Kast, W. M., Offnnga, R., Peters, P. J., Voordouw, A. C., Meloen, R. H., Van der Eb, A. J., Melief, C. J. M. (1989) Eradication ofadenovirus El-induced tumors by E1A specific cytotoxic I lymphocytes. Cell 59, Greenbeerg, P. D. (1991) Adoptive T cell therapy of tumors: mechanisms operative in the recognition and elimination of tumor cells. Adv. Immunol. 49, Rosenberg, S. A., Packard, B. S., Aebersold, P. M., Solomon, D., Topalian, S. L, by, S. 1., Simon, P., Lotze. M. 1., Yang, J. C., Seipp, C. A., Simpson, C., Carter, C., Bock, S., Sehwartzentruber, D., Wei,J. P., White, D. E. (1988) Use of tumor infiltrating lymphocytes and interleukin-2 in the immunotherapy of patients with metastatic. Preliminary report. N. EngI. J. Med. 319, Rosenberg, S. A. (1992) The immunotherapy and gene therapy of cancer. J. Clin. Oncol. 10, Van der Bruggen, P., Traversed, C., Chomez, P., Lurquin. C., DePlaen, E., Van Den Eynde, B., Knuth, A., Boon, T. (1991) A gene encoding an antigen recognized by cytolytic I lymphocytes on a human. Science 254, Kawakami, Y., Eliyahu, S., Delgaldo, C. H., Robbins, P. F., Rivoltini, 1., Topalian, S. L., Miki, T., Rosenberg, S. A. (1994) Cloning ofthe gene coding for a shared human antigen recognized by autologous I cells infiltrating into tumor. Proc. Nail. Acad. Sci. USA 91, Kawakami, Y., Eliyahu, S., Delgado, C. H., Robbins, P. F., Sakaguchi, K., Appella, E., Yannelli, J. R., Adema, G. J., Miki, T., Rosenberg, S. A. (1994) Identification ofa human antigen recognized by tumor infiltrating lymphocytes associated with in vivo tumor rejection. Proc. Na1. Acad. Sci. USA 91, Brichard, V., Van Pel, A., W#{246}lfel,T., W#{246}lfel,C., Dc Plaen, E., Leth#{235}, B., Coulie, P., Boon, T. (1993) The tyrosinase gene codes for an antigen recognized by autologous cytolytic I lymphocytes on HLA.A2 s. J. Exp. Med. 1 78, Coulie, P. G., Brichard, V., Van Pd, A., Wolfel, T., Schneider, J., Traversari, C., Mauei, S., Dc Plaen, E. D., Lurquin, C., Szikora,J-P., Reauld,J-C., Boon, 1. (1994) A new gene coding for a differentiation antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 s. J. Exp. Med. 180, Gaugler, B., Van Den Eynde, B., Van der Bruggen, P., Romero, P.. Gaforio, J. J., Dc Ptaen, E., Lethe, B., Brasseur, F., Boon, T. (1994) Human gene MAGE-3 codes for an antigen recognized on a by autologous cytolytic T lymphocytes. J. Exp. Med. 179, Van Den Eynde, B., Pesters, 0., Dc Backer, 0., Gaugler, B., Lucas, S., Boon, 1. (1995) A new family of genes coding for an antigen recognized by autologous cytolytictlymphocytes on a human.). Exp. Med. 182, Boel, P., Wildmann, C., Sensi, M. L., Brasseur, R., Renauld, J. C., Coulie, P., Boon, 1., Van der Bruggen, P. (1995) BAGE: a new gene encoding an antigen recognized on human s by cytolytic T lymphocytes. Immunity 2, Robbins, P. F., El-Gamil, M., Kawakami, Y., Stevens, E., Yannelli, J., Rosenberg, S. A. (1994) Recognition of tyrosinase by tumor infiltrating lymphocytes from a patient responding to immunotherapy. Cancer Res. 54, Wang, R. F., Robbins, P. F., Kawakami, Y., Kang, X. Q., Rosenberg, S. A. (1995) Identification of a gene encoding a tumor antigen recognized by HLA-A31-restricted tumor-infiltrating lymphocytes. J. Exp. Med. 181, Wolfel, 1., Hauer, M., Schneider, J., Serrano, M., Wolfel, C., Klehmann- Hieb, E., Dc Plaen, E., Hankeln,T., MeyerZum Buschenfelde, K.-H., Beach, D. (1995) A pl6lnk4a-insensitive CDK4 mutant targeted by cytolytic I lymphocytes in a human. Science 269, Robbins, P. F., El-Gamil, M., Li, Y. F., Kawakami, Y., Loftus, D., Appella, E., Rosenberg, S. A. (1996) A mutated (-catenin gene encodes a specific antigen recognized by tumor infiltrating lymphocytes. J. Exp. Med. 183, Mandelbolm, 0., Berke, C., Fridkin, M., Feldman, M., Eisenstein, M., Eisenbach, L. (1994) CTL induction by a tumor-associated antigen octapeptide derived from a murine lung carcinoma. Nature 369, Cox, A. L, Skipper, J., Cehn, Y., Henderson, R. A., Darrow, I. L, Shabanowitz, J., Engelhard, V. H., Hunt, D. F., Slingluff, C. L. (1994) Identification of a peptide recognized by five -specific human cytotoxic I cell lines. Science 264, Cheever, M. A., Disis, M. L, Bernhard, H., Gralow, J. R., Hand, S. L, Huseby, E. S., Qin, H. L., Takahashi, M., Chen, W. (1995) Immunity to oncogenic proteins. Imrnunol. Rev. 145, Fisk, B., Blevins, I. L., Wharton, J. T., toannides, C. G. (1995) Identification of an immunodominant peptide of HER-2/neu protooncogene recognized by ovarian tumor-specific cytotoxic I lymphocyte lines. J. Exp. Med. 181, Jung, S., Schluesener, H. J. (1991) Human I lymphocytes recognize a peptide of single point-mutated, oncogenic ras proteins. J. Exp. Med. 173, Skipper, j., Stauss, H. J. (1993)Identification oftwo cytotoxic Tlymphocyterecognized epitopes in the Ras protein. J. Exp. Med. 1 77, Celis, E., Tsai, V., Crimi, C., DeMars, R., Wentworth, P. A., Chesnut, R. W., Grey, H. M., Sette, A., Serra, H. M. (1994) Induction of anti-tumor cytotoxic T lymphocytes in normal humans using primary cultures and synthetic peptide epitopes. Proc. Nat!. Acad. Sci. USA 91, Rammensee, H. G., Friede, I., Stevanoviic, S. (1995) MHC uganda and peptide motifs: first listing. Immunogenetics 41, Traversari, C., Van der Bruggen, P., Luescher, I. F., Lurquin, C., Chomez, P., Van Pci, A., Dc Plaen, E., Amar-Costesec, A., Boon, I. (1992) A nonapeptide encoded by human gene MACE-i is recognized on HLA-A1 by cytolytic I lymphocytes directed against tumor antigen MZ2-E.J. Exp. Med. 176, Van der Bruggen, P., Szikora, J. -P., Boel, P., Wildmann, C., Somville, M., Sensi, M., Boon, T. (1994) Autologous cytolytic I lymphocytes recognize a MAGE-1 nonapeptide on s expressing Cw1601. Eur. J. Immunol. 24, Van der Bruggen, P., Bastin,J., Gajewski, T., Coulie, P. G., Boel, P., de Smet, C., Traversari, C., Townsend, A., Boon, I. (1994) A peptide encoded by human gene MAGE-3 and presented by HLA-A2 induces cytolytic T lymphocytes that recognize tumor cells expressing MAGE-3. Eur. J. Immunol. 24, Guilloux, Y., Lucas, L, Brichard, V. G., Van Pet, A., Viret, C., Dc Plaen, E.. Brasseur, F., Lethe, B., Jotereau, F., Boon, I. (19%) A peptide recognized by human cytolytic I lymphocytes on HLA-A2 is encoded by an intron sequence of the N-acetylglucosaminyltmnsferase V gene. J. Exp. Med. 183, Bouchard, B., Fuller, B. B., Vijayasaradhi, S., Houghton, A. N. (1989) Induction of pigmentation in mouse fibroblasts by expression of human tyrosinase edna. J. Exp. Med. 169, Wolfel, I., Van Pd, A., Brichard, V., Schneider, J., Seliger, B., Meyer Zum Buschenfelde, K.-H., Boon, T. (1994) Iwo tyrosinase nonapeptides recognized on HLA-A2 s by autologous cytolytic I lymphocytes. Eur. J. Immunol. 24, Kang, X. -Q., Kawakami, Y., Sakaguchi, K., El-Gamil, M., Wang, R.-F., Yannelli, J. R., Appella, E., Rosenberg, S. A., Robbins, P. F. (1995) Identification of a tyrosinase epitope reocognized by HLA-A24 restricted tumor-infiltrating lymphocytes. J. Immunol. 155, Wang and Rosenberg Human tumor antigens 307

13 33. Brichard, V. C., Hennan, J., Van Pet, A., Wildmann, C., Gaugler, B., Wolfel, I., Boon, I., Lethe, B. (1996) A tyrosinase nonpeptide presented by HLA- 844 is recognized on a human by autologous cytolytic I lymphocytes. Eur. J. Immunol. 26, Markus, N. R., Rosenberg, S. A., Toplian, S. L. (1995) Analysis of cytokine secretion by -specific CD4 Tlymphocytes.J. Interferon Cytokine Res. 15, Dadmarz, R.4 Sgagias, M. K., Rosenberg, S. A., Schwartzentruber, D. J. (1995) CD4 I lymphocytes infiltrating human breast cancer recognize autologous tumor in an MHC-class II restricted fashion. Cancer Immunol. Immunother. 40, Topalian, S. L., Rivoltini, L, Mancini, M., Markus, N. R., Robbins, P. F., Kawakami, Y., Rosenberg, S. A. (1994) Human CD4 I cells specifically recognize a shared -associated antigen encoded by the tyrosinase gene. Proc. Neil. Acad. Sci. USA 91, Topalian,S. L., Rivoltini, L.,Mancini, M.,Ng,J.,Hartzman, R.J., Rosenberg, S. A. (1994) Melanoma-specific CD4 I lymphocytes recognize human antigens processed and presented by Epstein-Barr-transformed B cells. mi. J. Cancer 58, Topalian, S. L., Gonzales, M. I., Parkhurst, M., Li, Y. F., Southwood, S., Sette, A., Rosenberg, S. A., Robbins, P. F. (1996) Melanoma-specific CD4 I cells recognize nonmutated HLA-DR-restricted tyrosinase epitopes. J. Exp. Med. In press. 39. Takahashi, T., Chapman, P. B., Yang, S. Y., Hare, I., ijayasaradhi, S., Houghton, A. N. (1995) Reactivity of autologous CD4 I lymphocytes against human. Evidence for a shared antigen presented by HLA.DR15. J. Immunol. 154, Kawakami, Y., Eliyahu, S., Sakaguchi, K., Robbins, P. F., Rivoltini, L., Yannelli, J. B., Appella, E., Rosenberg, S. A. (1994) Identification of the immunodominant peptides of the MART-l human antigen recognized by the majority of HLA-A2 restricted tumor infiltrating lymphocytes. J. Exp. Med. 180, Sensi, M., Traversari, C., Radrizzani, M., Salvi, S., Maccalli, C., Mortarini, R., Rivoltini, L., Farina, C., Nicolini, G., Wolfel, T., Brichard, V., Boon, T., Bordignon, C., Anichini, A., Parmiani, C. (1995) Cytotoxic I lymphocyte clones from different patients display limited T-cell receptors variable gene usage in HLA-A2 restricted recognition of Melan/MART-1 antigen. Proc. Nail. Aco4. Sci. USA 92, Bakker, A. B. H., Schreurs, M. W. J., de Boer, A. J., Kawakami, Y., Rosenberg, S. A., Adema, C. J., Figdor, C. G. (1994) Melanocyte lineagespecific antigen gploo is recognized by melanocyte-derived tumor-infiltrating lymphocytes..1. Exp. Med. 1 79, Adema, C. J., de Boor, A. J., vant Hullenaar, R., Denijn, M., Ruiter, D. J., Vogel, A. M., Figdor, C. G. (1993) Melanocyte lineage-specific antigens recognized by monoclonal antibodies NK1-beteb,HMB-50, and HMB-45 are encoded by a single edna. Am. J. Pathol. 143, Adema, G. J., Boer, A. J., Vogell, A. M., Loenen, W. A. M., Figdor, C. G. (1994) Molecular characterization ofthe melanocyte lineage-specific antigen gploo. J. Biol. Chem. 269, Kawakami, Y., Eliyahu, S., Jennings, C., Sakaguchi, K., Kang, X., Southwood, S., Robbins, P. F., Sette, A., Appella, E., Rosenberg, S. A. (1995) Recognition of multiple epitopes in the human antigen gploo by tumor-infiltrating I lymphocytes associated with in vivo tumor regression.). Immunol. 154, Bakker, A., Schreurs, M., Tafazzul, G., de Boor, A., Kawakami, Y., Adema, G., Figdor, C. (1995) Identification of a novel peptide derived from the melanocyte-specific gploo antigen as the dominant epitope recognized by an HLA-A2.1-restricted anti- ClLline. Mt. J. Cancer 62, Castelli, C., Storkus, W. J., Maeurer, M. J., Martin, D. M., Huang, E. C., Pramanik, B. N., Nagabhushan, T. L., Parmiani, C., Lotze, M. T. (1995) Mass spectrometric idenjification of a naturally processed peptide recognized by CD8 cytatoxic I lymphocytes. J. Exp. Med. 181, Wolfel, T., Schneider, J., Meyer Zum Buschenfekle, K.-H., Rammensee, H.-G., Rotzschke, 0., Falk, K. (1994) Isolation of naturally processed peptides recognized by cytolytic T lymphocytes (CTL) on human cells in association with HLA-A2.l. Ins. J. Cancer 57, Iopalian, S.,Solomon, D., Avis, F. P., Chang, A. E., Freeksen, D. L., Linehan, W. M., Lotze, M. I., Robertson, C. N., Seipp, C. A., Simon, P., Simpson, C. G., Rosenberg, S. A. (1988) Immunotherapy ofpatients with advanced cancer using tumor infiltrating lymphocytes and recombinant interleukin-2: a pilot study. J. Clin. Oncol. 6, Cohen, T., Muller, R. M., Tomita, Y., Shibahara, S. (1990) Nucleotide sequence of the edna encoding human tyrosinase-related protein. Nucleic Acids Res. 18, Vijayasyradhi, S., Bouchard, B. B., Houghton, A. N. (1990) The antigen gp75 is the human homologue of mouse b (brown) locus gene. J. Exp. Med. 171, Mattes, M. J., Thomson, I. M., Old, L. J., Lloyd, K. 0. (1983) A pigmentation-associated, differentiation antigen of human defined by a precipitating antibody in human serum. Ira. J. Cancer 32, Thomson, I. M., Real, F. X., Murakami, S., Cordon-Cardo, C., Old, L J., Houghton, A. N. (1988) Differentiation antigens of melanocytes and : analysis of melanosome and cell surface markers of human pigmented cells with monoclonal antibodies. J. invest. Dermatol. 90, Thomson, T. M., Mattes, J. M., Roux, L, Old, L J., Uoyd, K. 0. (1985) Pigmentation-associated glycoprotein of human and melanocytes: definition with a mouse monoclonal antibody. J. Invest. Dermiuol. 85, Jimenez-Cervantes, C., Solano, F., Kobayashi, I., Urabe, K., Hearing, V. J., Lozano, J. A., Garcia-Borron, J. C. (1994) A new enzymatic function in the melanogenic pathway. J. Biol. Chem. 269, Falk, K., Rotzschke, 0., Takiguchi, M., Grahovae, B., Gnau, V., Stevanovic, S., Jung, G., Rammensee, H.-G. (1994) Peptide motifs of HLA-A1, -All, -A31, and -A33 molecules. Immunogenetics 40, Wang, R. F., Parkhurst, M. R., Kawakami, Y., Robbins, P. F., Rosenberg, S. A. (1996) Utilization of an alternative open reading frame of a normal gene in generatinga novelhuman cancerantigen.j. Exp. Med. 183, Descombes, P., Schibler, U. (1991) Aliver-enriched transcriptional activator protein, LAP, and a transcriptional inhibitory protein, LIP, are translated from the same mrna. Cell 67, Ossipow, V., Descombes, P., Schibler, U. (1993) CCAAT/enhancer-binding protein mrna is translated into multiple proteins with different transcription activation potentials. Proc. Nail. Aco4. Sci. USA 90, Hann, S. R. (1995) Regulation and function of non-aug-initiated proto-oncogenes. Biochimie 76, Shaw, M. W., Choppin, P. W., Lamb, R. A. (1983) A previously unrecognized influenza B virus glycoprotein from a bicistronic mrna that also encodes the viral neuraminidase. Proc. Nail. Acad. Sci. USA 80, Spiropoulou, C. F., Nichol, S. I. (1993) A small highly basic protein is encoded in overlapping frame within the P gene of vesicular stomatitis virus. J. Virol. 67, Schwartz, S., Felber, B. K., Pavlakis, G. N. (1992) Mechanism of translation of monocistronic and multicistronic human immunodeficiency virus type 1 mrnas. Mol. Cell. Biol. 12, Townsend, A., Ohlen, C., Rogers, M., Edwards, J., Mukherjee, S., Bastin, J. (1994) Souce of unique tumor antigens. Nature 371, Scrrano, M., Hannon, G. J., Beach, D. (1993) A new regulatory motif in cell-cycle control causing specific inhibition of cyclin D/CDK4 [see comments]. Nature 366, Sherr, C. J., Roberts, J. M. (1995) Inhibitors of mammalian Cl cyclin-dependent kinases. Genes Dev. 9, Quelle, D. E., Zindy, F., Ashmun, R. A., Sherr, C. J. (1995) Alternative reading frames of the INK4a tumor suppressor gene encode two unrelated proteins capable of inducing cell cycle arrest. Cell 83, Uenaka, A., Ono, I., Akisawa, I., Wada, H., Yasuda, I., Nakayama, E. (1994) Identification ofa unique antigen peptide prl1 on BALB/c RL male 1 leukemia recognized by cytotoxic I lymphocytes and its relation to the Akt oncogene.j. Exp. Med. 180, Malarkannan, S., A1karii, M., Shastri, N. (1995) A rare cryptic translation product is presented by K major histocompatibility complex class I molecule to alloreactive I cells. J. Exp. Med. 182, Robbins, P. F., El-Gamil, M., Li, Y. F., Topalian, S. L., Rivoltini, L., Sakaguchi, K., Appella, E., Kawakami, Y., Rosenberg, S. A. (1995) Cloning of a new gene encoding an antigen recognized by -specific HLA- A24 restricted tumor-infiltrating lymphocytes. J. Immunol. 154, loannides, C. G., Fisk, B., Pollack, M. S., Frazier, M. L., Wharton, J. T., Freedman, R. S. (1993) Cytotoxic I-cell clones isolated from ovarian tumour infiltrating lymphocytes recognize common determinants on nonovarian tumour clones. Scand. J. immunol. 37, Fisk, B., Chesak, B., Pollack, M. S., Wharton, J. T., loannides, C. G. (1994) Oligopeptide induction of a cytotoxic I lymphocyte response to HER-2INeu proto-oncogene in vitro. Cell Immunol. 157, Yoshino, I., Peoples, G. E., Geoedegebauure, P. 5., Mazizrz, R., Eberlein, I. J. (1994) Association of HERzIneu expression with sensitivity to tumor-specific CIL in human ovarian cancer. J. Immunol. 152, Disis, M. L., Alenoff, E. C., McLaughlin, G., Murphy, A. E., Chen, W., Groner, B., Jeschke, M., Lydon, N., McGlynn, E., Livingston, R. B., Moe, R., Cheever, M. A. (1994) Existent T cell and antibody immunity to HER-2/neu protein in patients with breast cancer. Cancer Res. 54, Peoples, C. E., Goedegebuure, P. S., Smith, R., Linehan, D. C., Yoshino, I., Eberlein, T. J. (1995) Breast and ovarian cancer-specific cytotoxic I lymphocytes recognize the same HERz/neu-derived peptide. Proc. NasI. Acad. Sci. USA 92, Linehan, D. C., Goedegebuure, P. S., Peoples, C. E., Rogers, S. 0., Eherlein, I. J. (1995) Tumor-specific and HLA-A2-restricted cytolysis by tumor-associated lymphocytes in human metastatic breast cancer. J. Immunol. 155, Finn, 0. J. (1992) Antigen-specific, MHC-unrestricted I cells. Biotherapy 4, loannides, C. C., Fisk, B., Jerome, K. R., Irimura, I., Wharton, J. T., Finn, 0. J. (1993) Cytotoxic T cells from ovarian malignant tumors can recognize polymorphic epithelial mucin core peptides. J. Immunol. 151, Magarian-Blander, J., Domenech, N., Finn, 0. J. (1993) Specific and effective I-cell recognition of cells transfected with a truncated human mucin cdna. Ann. NYAcad. Sci. 690, Finn, 0. J., Jerome, K. R., Henderson, R. A., Pecher, G., Domenech, N., Barratt-Boyes Magarian-Blander, S. M. (1995) MUC-I epithelial tumor mucin-based immunity and cancer vaccines. Immunol. Rev. 145, Boon, I. (1992) Toward a genetic analysis of tumor rejection antigens. Adv. CancerRes. 58, Journal of Leukocyte Biology Volume 60, September 1996

14 82. Monach, P. A., Meredith, S. C., Siegel, C. T., Schreiber, H. (1995) A unique tumor antigen produced by a single amino acid substitution. Immunol. 2, Van Elsas, A., Nijman, H. W., Van der Minne, C. E., Mourer, J. S., Kast, W. M., Melief, C. J., Schrier, P. I. (1995) Induction and characterization of cytotoxic I-lymphocytes recognizing a mutated p2lras peptide presented by HLA-A let. J. Cancer 61, Houbiers, J. G. A., Nijman, H. W., van der Burg, S. H., Drijfbout, J. W., Kenemans, P., van de Velde, C. J. H., Brand, A., Momburg, F., Kast, W. M., Melief, C. J. M. (1993) In vitro induction of human cytotoxic I lymphocyte responses against peptides of mutant and wild-type p53. Eur. J. Immunol. 23, Nijman, H. W., van der Burg, S. H., Vierboom, M. P., Houbiers, J. G., Kast, W. M., Melief, C. J. (1994) p53, a potential target fortumor-directed I cells. Immunol. Lea. 40, Nijman, H. W., Houbiers, J. G. A., van der Burg, S. H., Vierboom, M. P. M., Kenemans, P., Kast, W. M., Melief, C. J. M. (1993) Characterization of cytotoxic I lymphocyte epitopes of a self-protein, p&3, and a non-self-protein,influenza matrix: Relationship between major histocompatibility complex peptide binding affinity and immune responsiveness to peptides. J. Immunother 14, Theobald, M., Biggs, J., Dittmer, D., Levine, A. J., Sherman, L. A. (1995) Targeting p53 as a general tumor antigen. Proc. NatI. Acad. Sci. USA 92, Coulie, P. C., Lehmann, F., Lethe, B., Herman, J., Lurquin, C., Andrawiss, M., Boon,T. (1995) A mutated intron sequence codes foran antigenic peptide recognized by cytolytic T lymphocytes on a human. Proc. NatI. Acad. Sci. USA 92, Grunwald, C. B. (1993) The structural and functional analysis of cadherin calcium-dependent cell adhesion molecules. Curr. Op. Cell Biol. 5, Hoschuetzky, H., Aberle, H., Kemler, R. (1994). f3-catenin mediates the interaction of the cadherin-catenin complex with epidermal growth factor receptor. J. Cell Biol. 127, Becker, K. -F., Atkinson, M. J., Reich, U., Becker, I., Nekarda, H., Siewert, J. R., Hofler, H. (1994) E-cadherin gene mutations provide clues to diffuse type gastric carcinomas. Cancer Res. 54, Kawanishi, J., Kato, J., Sasaki, K., Fujii, S., Watanabe, N., Niitsu, Y. (1995) Loss of E-cadherin-dependent cell-cell adhesion due to mutation of the b-catenin gene in a human cancer cell line, HSC-39. Mol. Cell Biol. 15, Salgaller, M. L., Weber, J. S., Koenig, S., Yannelli, J. R., Rosenberg, S. A. (1994) Generation of specific anti- reactivity by stimulation of human tumor-infiltratinglymphocytes with MAGE.l synthetic peptide. Cancer immunol. Immunother. 39, los-i Visseren, M. J., Van Elsas, A., van der Voort, E. I., Ressing, M. E., Kast, W. M., Schrier, P. I., Melief, C. J. (1995) CIL specific for the tyrosinase autoantigen can be induced from healthy donor blood to lyse cells. J. Immunol. 154, Rivoltini, L., Kawakami, Y., Sakaguchi, K., Southwood, S., Sette, A., Robbins, P. F., Marincola, F. M., Salgaller, M. L., Yannelli, J. R., Appella, E. (1995) Induction oftumor-reactive CIL from peripheral blood and tumor-infiltrating lymphocytes of patients by in vitro stimulation with an immunodominant peptide of the human antigen MART-l. J. Immunol. 154, Salgaller, M. L., Mshar, A., Marincola, F. M., Rivoltini, L., Kawakami, Y., Rosenberg, S. A. (1995) Recognition of multiple epitopes in the human antigen gploo by peripheral blood lymphocytes stimulated in vitro with synthetic peptides. Cancer Res. 55, Lipford, G. B., Bauer, S., Wagner, H., Heeg, K. (1995) Peptide engineering allows cytotoxic I-cell vaccination against human papilloma virus tumour antigen, E6. Immunol. 84, Lipford, G. B., Bauer, S., Wagner, H., Heeg, K. (1995)In vivo CTL induction with point-substituted ovalbumin peptides: immunogenicity correlates with peptide-induced MHC class I stability. Vaccine 13, Parkhurst, M. R., Salgaller, M., Southwood, S., Robbins, P., Sette, A., Rosenberg, S. A., Kawakami, Y. (1996) Improved induction of reactive CTL with peptides from the antigen gploo modified at HLA-A0201 binding residues. J. immunol. in press Cole, D. J., Weil, D. P., Shamamian, P., Rivoltini, L., Kawakami, Y., Topalian, S., Jennings, C., Eliyahu, S., Rosenberg, S. A., Nishimura, M. I. (1994) Identification of MARI-l-specific T-cell receptors: T cells utilizing distinct I-cell receptor variable andjoining regions recognize the same tumor epitope. Cancer Res. 54, Cole, D.J.,Weil, D. P., Shilyansky,J., Custer, M., Kawakami,Y., Rosenberg, S. A., Nishimura, M. I. (1995) Characterization ofthe functional specificity of a cloned T-cell receptor heterodimer recognizing the MART-i antigen. Cancer Res 55, Romero, P., Pannetier, C., Herman, J., Jongeneel, C. V., Cerottini, i-c., Coulie, P. G. (1995) Multiple specificities in the repertoire of a patient s cytolytic I lymphocytes directed against tumor antigen MAGEl.Al.J. Exp. Med. 182, Peoples, G. E., Yosino, I., Douville, C. C., Andrews,.1. V. R., Goedegebuure, P. S., Eberlein, I. J. (1994) ICR V3 and V36 CIL recognize tumor-associated antigens related to HER-2/neu expression in HLA-A2 ovarian cancers. J. Immunol. 152, Ashton-Rickardt, P. G., Bandeira, A., Delaney,.J. R., Van Kaer, L., Pircher, H. P., Zinkernagel, R. M., Tonegawa, S. (1994) Evidence for a differential avidity model of I cell selection in the thymus [see comments]. Cell 76, Jameson, S. C., Hogquist, K. A., Bevan, M. J. (1994) Specificity and flexibility in thymic selection. Nature 369, Ashton-Rickardt, P. G., Tonegawa, S. (1994) A differential-avidity model for I-cell selection. Immunol. Today 15, Screen, E. E., Lehmann, P. V., Ametani, A., Benichou, G., Miller, A., Moudgil, K. (1993) Dominance and crypticity of I cell antigenic determinants. Annu. Rev. Immunol. 1 1, Moudgil, K. D., Sercarz, E. E. (1994) Can antitumor immune responses discriminate between selfand nonsell? Immunol. Today 15, Pircher, H., Rohrer, U. H., Moskophidis, D., Zinkernagel, R. M., Hengartner, H. (1991) Lower receptor avidity required for thymic clonal deletion than for effector I-cell function. Nature 35 1, Fairchild, P. J., Wraith, D. C. (1996) Lowering the tone: mechanisms of immunodominance among epitopes with low affinity for MHC. Immunol. Today 17, Houghton, A. N. (1994) Commentary. Cancer antigens: Immune recognition of selfand altered self. J. Exp. Med. 180, Nanda, N. K., Sercarz, E. E. (1995) Induction of anti-self-immunity to cure cancer. Cell 82, Nordlund,J. I., Kirkwood, M., Forget, B. M., Milton, G.,Albert, D. M., Lemer, A. B. (1983) Vitiligo in patients with metastatic : Agood prognostic sign. J. Am. Acad. Dermatol. 9, Bystiyn, 1.-C, Rigel, D., Friedman, R. 1., Kopf, A. (1987) Prognostic significance of hypopigmentation in malignant. Arch. Dermosol. 123, Richards, J. M., Mehta, N., Ramming, K., Skosey, P. (1992) Sequential chemoimmunotherapy in the treatment of metastatic. J. Clin. Oncol. 10, Hara, I., Takechi, Y., Houghton, A. N. (1995) Implicating a role for immune recognition of self in tumor rejection: passive immunization against the Brown locus protein. J. Exp. Med. 182, Young,.1. W., Inaba, K. (1996) Dendritic cells as adjuvants for class I major histocompatibility complex-restricted antitumor immunity. J. Exp. Med. 183, Zhai, Y., Yang, J. C., Kawakami, Y., Spiess, P., Wadsworth, S. C., Cardoza, L M., Couture, L A., Smith, A. E., Rosenberg, S. A. (1996) Antigen-specific tumor vaccines: Development and characterization of recombinant adenoviruses encoding MART-i or gploo for cancer therapy.j. Immunol. 156, Bronte, V., Tsung, K., Rao,.1. B., Chen, P. W., Wang, M., Rosenberg, S. A., Restifo, N. P. (1995) IL-2 enhances the function of recombinant poxvirusbased vaccines in the treatment of established pulmonary metastases. J. Immunol. 154, Irvine, K., Rao, J., Rosenberg, S., Restifo, N. (1996) Cytokine enhancement of DNA immunization leads to treatment of established pulmonary metastases. J. Immunol. 156, Wang and Rosenberg Human tumor antigens 309