Activated CD8+ T cells detect virally infected cells and tumor cells by recognition of major histocompatibility complex class I-bound peptides derived from degraded, endogenously produced proteins. In contrast, CD8+ T cell activation often occurs through interaction with specialized antigen-presenting cells displaying peptides acquired from an exogenous cellular source, a process termed cross-priming. Here, we observed a marked inefficiency in exogenous presentation of epitopes derived from signal sequences in mouse models. These data indicate that certain virus- and tumor-associated antigens may not be detected by CD8+ T cells because of impaired cross-priming. Such differences in the ability to cross-present antigens should form important considerations in vaccine design.

The antigen processing pathway generates the peptides displayed by MHC I molecules on the cell surface. Whether these peptides are generated in the cytosol or from longer intermediates transported into the ER is unclear, because peptides other than those bound to MHC I have been difficult to find. Using a novel assay, we show that N-terminally extended antigenic analogs were associated with high-molecular weight material in the cytosol and were transported by TAP. In the ER, a nonapeptide was predominant that was converted to the final octapeptide only in presence of the appropriate MHC I molecule. The existence of extended peptides and their MHC I-dependent trimming suggest a mechanism for efficiently satisfying the distinct sequence preferences of polymorphic MHC I molecules.

Previous reports have described antigens that are recognized on human melanoma cells by autologous cytolytic T lymphocytes (CTL). The genes coding for a number of these antigens have been identified. Here we report the cloning of a gene that codes for an antigen recognized by autologous CTL on a human renal carcinoma cell line. This antigen is presented by HLA-B7 and is encoded by a new gene that we have named RAGE1. No expression of RAGE1 was found in normal tissues other than retina. RAGE1 expression was found in only one of 57 renal cell carcinoma samples, and also in some sarcomas, infiltrating bladder carcinomas, and melanomas. This represents the first identification of an antigen recognized by autologous CTL on a renal tumor.

Summary The expression of the gp100 antigen is generally thought to be confined to cells of the melanocytic lineage, which makes the protein a suitable melanoma-specific marker. Strikingly, after screening a panel of normal tissues, tumour samples and cell lines of nonmelanocytic origin, we found transcripts encoding gplOO in virtually every tissue and cell line tested. In contrast, tyrosinase and MART-1/MelanA transcripts were detected only in cells of the melanocytic lineage. However, no gpl 00 protein could be detected by either Western blotting or cytotoxicity assays. Therefore, at the protein level, gp100 remains exclusive for cells of melanocytic origin despite its transcription in many cell types. The major implication of this finding is that screening of patient material for gp100 expression should preferrably be performed by antibody staining. Reverse transcriptase polymerase chain reaction (RT-PCR) can be employed, provided that it is performed in a tightly controlled, semiquantitative setting.Keywords: gp100; RT-PCR; melanoma; renal cell carcinoma; tumour-associated antigens The molecular cloning of tumour-associated antigens has provided new tools for the immunotherapy of cancer (reviewed in Van den Eynde and Brichard, 1995). Now it has become feasible to immunize cancer patients against these antigens to stimulate specifically a cellular anti-tumour response (Marchand et al, 1995). To select patients eligible for immunotherapeutic protocols, the antigenic profiles of the patients' tumours must be characterized. This is usually achieved by reverse transcriptase polymerase chain reaction (RT-PCR) using RNA obtained from tumour samples when available. An alternative source of tumour cells is whole blood, which often contains numerous circulating residual tumour cells (Brossart et al, 1995;Hoon et al, 1995). A suitable antigen to target in melanoma is gplOO because it is thought to be specific for cells of melanocytic lineage and it is adequately expressed in melanoma cells. Expression in non-melanocytic cells, at least as measured by antibody reactivity (Vennegoor et al, 1988) or Northern analysis (Kawakami et al, 1994), is virtually absent. In a routine RT-PCR screening of a number of human tumour samples and normal tissues, we noticed to our surprise that gplOO transcripts were present in almost all materials tested, whereas the protein was not detectable. The renal cell carcinoma cell lines (RCC) LE-9104-RCC, LE-921 l-RCC and LE-9415-RCC and the melanoma cell line Mel 603 were established in our laboratory. The RCC cell line SK-RC-7 was kindly provided by Dr E Oosterwijk (Department of Urology, Nijmegen University, The Netherlands). The RCC cell lines MZ-1851-RCC and Camejo and melanoma cell lines MZ-2-mel and MZ-7.4-mel were generously provided by Dr B Seliger (J Gutenberg University, Mainz, Germany). The breast carcinoma cell lines (BRCA) MCF-7 and SK-BR-3 were a gift from Dr

The identification of novel tumor antigens is of extreme importance for effective immunotherapy against cancer. A major obstacle in this field is the limited life span of tumorspecific cytotoxic T lymphocytes (CTLs) in vitro. Therefore we searched for a method to isolate the tumor specificity of these CTLs, i.e., their T-cell receptors (TCRs) and transfer it to an immortalized T-cell line. For this purpose, a TCRnegative Jurkat T-cell line was equipped with a nuclear factor of activated T cells ( Key words: T-cell receptor; tumor antigen; NFAT; CAMEL; cDNA expression cloning; CD8; luciferase; retrovirus; JurkatSince the characterization of the first human tumor-specific antigen in 1991, MAGE-1, 1 a considerable number of other targets for tumor-specific cytotoxic T cells (CTLs) have been identified. Most of these antigens have been cloned from melanoma cells 2 and less frequently from other types of cancer. 3,4 For the development of effective immunotherapy against cancer, the identification of a larger variety of tumor-associated antigens is of great importance, to broaden the range of different tumors that can be treated as well as to prevent the outgrowth of antigen-loss tumor cells after vaccination.Several strategies have led to the identification of tumor-specific T-cell epitopes. Most frequently, cDNA expression cloning has been employed, in which the gene encoding the antigen is identified by transfecting pools of tumor-derived cDNA into recipient cells, followed by screening with the tumor-specific CTL. [5][6][7][8] In another approach, the peptide epitope for the tumor-specific CTL is identified, either by peptide elution and mass spectrometry 9 or by mimicry from designated peptide libraries, 10 although with the latter method no tumor antigens have been identified yet. Several tumor antigens have been found by the method of "reverse immunology." These antigens were initially cloned on the basis of their differential expression in tumors versus healthy tissues and antigenic peptides were subsequently identified by performing in vitro T-cell inductions. 11,12 For efficient elucidation of novel physiologically relevant tumor-specific antigens, large quantities of tumor-specific CTLs are required to allow for multiple screening rounds. In practice, the limited amount of CTLs available often constitutes the major obstacle for the cloning of new tumor antigens. We have sought to circumvent this issue by cloning and transferring the T-cell receptor (TCR) of a tumor-specific T cell to an immortalized T-cell line. In this way, an unlimited source of tumor-specific T cells can be obtained, which facilitates the identification of novel tumor antigens. The feasibility of TCR cloning and transfer has already been demonstrated by others. [13][14][15][16][17] We studied the properties of a TCR, isolated from a characterized CAMEL (CTL-recognized antigen on melanoma)-specific CTL clone, 8 which was transferred into a TCR-negative human leukemia Jurkat T-cell line. To detect TCRmediated activation, the cells were tr...