Supplementary Table S1 from An Inducible Mouse Model of Melanoma Expressing a Defined Tumor Antigen

Central tolerance toward tissue-restricted Ags is considered to rely on ectopic expression in the thymus, which was also observed for tumor Ags encoded by cancer-germline genes. It is unknown whether endogenous expression shapes the T cell repertoire against the latter Ags and explains their weak immunogenicity. We addressed this question using mouse cancer-germline gene P1A, which encodes antigenic peptide P1A35–43 presented by H-2Ld. We made P1A-knockout (P1A-KO) mice and asked whether their anti-P1A35–43 immune responses were stronger than those of wild-type mice and whether P1A-KO mice responded to other P1A epitopes, against which wild-type mice were tolerized. We observed that both types of mice mounted similar P1A35–43-specific CD8 T cell responses, although the frequency of P1A35–43-specific CD8 T cells generated in response to P1A-expressing tumors was slightly higher in P1A-KO mice. This higher reactivity allowed naive P1A-KO mice to reject spontaneously P1A-expressing tumors, which progressed in wild-type mice. TCR-Vβ usage of P1A35–43-specific CD8 cells was slightly modified in P1A-KO mice. Peptide P1A35–43 remained the only P1A epitope recognized by CD8 T cells in both types of mice, which also displayed similar thymic selection of a transgenic TCR recognizing P1A35–43. These results indicate the existence of a minimal tolerance to an Ag encoded by a cancer-germline gene and suggest that its endogenous expression only slightly affects diversification of the T cell repertoire against this Ag.

Central tolerance toward tissue-restricted Ags is considered to rely on ectopic expression in the thymus, which was also observed for tumor Ags encoded by cancer-germline genes. It is unknown whether endogenous expression shapes the T cell repertoire against the latter Ags and explains their weak immunogenicity. We addressed this question using mouse cancer-germline gene P1A, which encodes antigenic peptide P1A 35–43 presented by H-2L d. We made P1A-knockout (P1A-KO) mice and asked whether their anti-P1A 35–43 immune responses were stronger than those of wild-type mice and whether P1A-KO mice responded to other P1A epitopes, against which wild-type mice were tolerized. We observed that both types of mice mounted similar P1A 35–43-specific CD8 T cell responses, although the frequency of P1A 35–43-specific CD8 T cells generated in response to P1A-expressing tumors was slightly higher in P1A-KO mice. This higher reactivity allowed naive P1A-KO mice to reject spontaneously P1A-expressing tumors, which progressed in wild-type mice. TCR-Vb usage of P1A 35–43-specific CD8 cells was slightly modified in P1A-KO mice. Peptide P1A 35–43 remained the only P1A epitope recognized by CD8 T cells in both types of mice, which also displayed similar thymic selection of a transgenic TCR recognizing P1A 35–43. These results indicate the existence of a minimal tolerance to an Ag encoded by a cancer-germline gene and suggest that its endogenous expression only slightly affects diversification of the T cell repertoire against this Ag. M any tumor Ags identified on the basis of their recognition by autologous T lymphocytes are derived from unmutated self-proteins (1, 2). Some are differentiation Ags that are shared between tumor and untransformed tissue, as is the case for melanocyte differentiation Ags. Spontaneous or vaccine-induced immune responses to these Ags in cancer patients can lead to autoimmune aggression of normal tissue, as typified by the vitiligo that affects some melanoma patients. A second type of tumor Ags is encoded by cancer-germline genes (also called cancer-testis genes). These genes, usually located on chromosome X, are expressed in many tumors, but not in normal cells except male germ cells and placental cells. The latter cells express very low classical MHC class I molecules at their surface (3, 4). Ags encoded by cancer-germline genes were thus considered as optimal T cell targets for cancer immunotherapy, as an immune response unhampered by natural tolerance would also not lead to aggression of normal tissue. In agreement with that notion, cyto-lytic T lymphocytes (CTL) …

AbstractCancer immunotherapy based on vaccination with defined tumor antigens has not yet shown strong clinical efficacy, despite promising results in preclinical models. This discrepancy might result from the fact that available preclinical models rely on transplantable tumors, which do not recapitulate the long-term host-tumor interplay that occurs in patients during progressive tumor development and results in tumor tolerance. To create a faithful preclinical model for cancer immunotherapy, we generated a transgenic mouse strain developing autologous melanomas expressing a defined tumor antigen recognized by T cells. We chose the antigen encoded by P1A, a well-characterized murine cancer germ line gene. To transform melanocytes, we aimed at simultaneously activating the Ras pathway and inactivating tumor suppressor Ink4a/Arf, thereby reproducing two genetic events frequently observed in human melanoma. The melanomas are induced by s.c. injection of 4-OH-tamoxifen (OHT). By activating a CreER recombinase expressed from a melanocyte-specific promoter, this treatment induces the loss of the conditional Ink4a/Arf gene in melanocytes. Because the CreER gene itself is also flanked by loxP sites, the activation of CreER also induces the deletion of its own coding sequence and thereby allows melanocyte-specific expression of genes H-ras and P1A, which are located downstream on the same transgene. All melanomas induced in those mice with OHT show activation of the Ras pathway and deletion of gene Ink4a/Arf. In addition, these melanomas express P1A and are recognized by P1A-specific T lymphocytes. This model will allow to characterize the interactions between the immune system and naturally occurring tumors and thereby to optimize immunotherapy approaches targeting a defined tumor antigen. (Cancer Res 2006; 66(6): 3278-86)

The rev-erbAalpha orphan protein belongs to the steroid nuclear receptor superfamily. No ligand has been identified for this protein, and little is known of its function in development or physiology. In this study, we focus on 1) the distribution of the rev-erbAalpha protein in adult fast- and slow-twitch skeletal muscles and muscle fibers and 2) how the rev-erbAalpha protein influences myosin heavy chain (MyHC) isoform expression in mice heterozygous (+/-) and homozygous (-/-) for a rev-erbAalpha protein null allele. In the fast-twitch extensor digitorum longus muscle, rev-erbAalpha protein expression was linked to muscle fiber type; however, MyHC isoform expression did not differ between wild-type, +/-, or -/- mice. In the slow-twitch soleus muscle, the link between rev-erbAalpha protein and MyHC isoform expression was more complex than in the extensor digitorum longus. Here, a significantly higher relative amount of the beta/slow (type I) MyHC isoform was observed in both rev-erbAalpha -/- and +/- mice vs. that shown in wild-type controls. A role for the ratio of thyroid hormone receptor proteins alpha1 to alpha2 in modulating MyHC isoform expression can be ruled out because no differences were seen in MyHC isoform expression between thyroid hormone receptor alpha2-deficient mice (heterozygous and homozygous) and wild-type mice. Therefore, our data are compatible with the rev-erbAalpha protein playing an important role in the regulation of skeletal muscle MyHC isoform expression.

Expression of myosin isoforms in skeletal muscles in mice lacking the rev-erb Alpha orphan receptor

The MAGED gene subfamily contains three genes in mouse and four in human. The MAGED1, D2, and D3 proteins are highly conserved between mouse and human, whereas paralogues are less conserved between each other. This finding suggests that each MAGED protein exerts a distinct function. To get a better insight into their physiological roles, we have analyzed their expression patterns during embryogenesis and brain development. In the mouse, Maged3 expression is restricted to the central nervous system where it was mostly detected in postmitotic neurons. Maged2 is mainly expressed in tissues of mesodermal origin. The expression pattern of Maged1 roughly summarizes that of Maged2 and Maged3 ; however, contrary to that of Maged3 , it includes the proliferative zones of the nervous system. We observed a discrepancy between Maged1 expression levels of RNA and protein, suggesting that its expression is regulated at a posttranscriptional level during the mouse development. Developmental Dynamics 230:325–334, 2004. © 2004 Wiley‐Liss, Inc.

Selection of immature CD4CD8 double-positive (DP) thymocytes for CD4 or CD8-lineage commitment is controlled by the interaction of the TCR with stromal cell-expressed peptide/MHC. We show that thymocyte-intrinsic genes influence the pattern of expression of a MHC class I-restricted transgenic (tg) TCR so that in DBA/2 mice, DP thymocytes with a characteristically high expression of tg TCR, infrequently transit to CD8 single-positive thymocytes. In contrast, in B10.D2 mice, the same tg TCR is expressed at lower levels on a subpopulation of DP thymocytes that more frequently transit to CD8 single-positive thymocytes. These characteristics were not influenced by thymic stromal components that control positive selection. Radiation chimeras reconstituted with a mixture of BM from tg TCR mice of the two genetic backgrounds revealed that the relative frequency of transit to the CD8 lineage remained thymocyte-intrinsic. Identifying the gene products whose polymorphism controls CD8 T cell development may shed new light on the mechanisms controlling T cell commitment/selection in mice other than the most studied “C57BL/6”-based strains.

Selection of immature CD4CD8 double-positive (DP) thymocytes for CD4 or CD8-lineage commitment is controlled by the interaction of the TCR with stromal cell-expressed peptide/MHC. We show that thymocyte-intrinsic genes influence the pattern of expression of a MHC class I-restricted transgenic (tg) TCR so that in DBA/2 mice, DP thymocytes with a characteristically high expression of tg TCR, infrequently transit to CD8 single-positive thymocytes. In contrast, in B10.D2 mice, the same tg TCR is expressed at lower levels on a subpopulation of DP thymocytes that more frequently transit to CD8 single-positive thymocytes. These characteristics were not influenced by thymic stromal components that control positive selection. Radiation chimeras reconstituted with a mixture of BM from tg TCR mice of the two genetic backgrounds revealed that the relative frequency of transit to the CD8 lineage remained thymocyte-intrinsic. Identifying the gene products whose polymorphism controls CD8 T cell development may shed new light on the mechanisms controlling T cell commitment/selection in mice other than the most studied "C57BL/6"-based strains.

BACKGROUND. Testicular germ cell tumors are the most common malignancy in young males, and the frequency of these tumors has risen dramatically over the last century. Because it is known that the MAGE genes are expressed in a wide variety of tumors but are express ed only in the mitotic spermatogonia (germ cells) and in the primary spermatocytes in the normal testis, the authors screened the expression of MAGE-A4 in a panel of testicular germ cell tumors.METHODS. Monoclonal antibody 57B raised against MAGE-A4 was tested immunohistochemically on 12 classical seminomas, 5 anaplastic seminomas, 10 various specimens of nonseminomatous germ cell tumors (NSGCTs), 2 combined tumors containing seminoma components, 1 Sertoli cell tumor, 2 Leydig cell tumors, and 15 carcinomas in situ (CIS). In addition, monoclonal antibody 57B was tested on embryonic gonad (age 8 weeks) and fetal gonads (ages 15 weeks, 17 weeks, and 28 weeks).RESULTS. Classical seminomas uniformly and specifically expressed MAGE-A4 compared with anaplastic seminomas and NSGCTs, which were negative for this antigen. Specific expression of MAGE-A4 also was seen in subpopulations of CIS cells, providing additional evidence for heterogeneity of the phenotype of these cells, in which it is I believed that differentiation and proliferation generate seminomas and NSGCTs. Finally, MAGE-A4 was expressed in the fetal precursors of the stem germ cells from 17 weeks of gestation onward, in accordance the fact that CIS can arise from prespermatogonia in the fetus.CONCLUSIONS. MAGE-A4 can be considered a potential specific marker for normal premeiotic germ cells and germ cell tumors and can be used to characterize classical seminomas. Cancer, 2001;92:2778-85. (C) 2001 American Cancer Society.

Overexpression of strong transcriptional activators like herpes simplex virion protein 16 (VP16) may lead to non-specific inhibition of gene expression as a result of the titration of transcription factors. Here we report that a fusion between the homeoprotein Hoxa2 and the VP16 activation domain inhibits transcription from the strong promoter/enhancers of cytomegalovirus (CMV) and Rous sarcoma virus (RSV). A similar fusion involving a Hoxa2 mutant protein that is defective in DNA binding has no effect on the CMV promoter but increases, rather than inhibits, the RSV promoter activity. This suggests that depending on its ability to bind DNA, the VP16 activator can interact with different sets of cofactors, giving rise to distinct transcriptional effects.

The first human members of the MAGE gene family that have been described are expressed in tumor cells but silent in normal adult tissues except in the male germ line. Hence, they encode strictly tumor-specific antigens that represent attractive targets for cancer immunotherapy. However, other members of the family were recently found to be expressed in normal cells, indicating that the family is larger and more disparate than initially expected. We therefore performed a database screening to identify all of the recorded members of both classes of human MAGE genes. This report provides an overview of the MAGE family and proposes a general nomenclature for all of the MAGE genes identified thus far. We found that the MAGE-D genes were particularly well conserved between man and mouse, suggesting that they exert important functions. In addition, the genomic structure of the MAGE-D genes indicates that one of them corresponds to the founder member of the family, and that all of the other MAGE genes are retrogenes derived from that common ancestral gene. Intriguingly, the COOH-terminal domain of MAGE-D3 was found to be identical to trophinin, a previously described protein believed to be involved in embryo implantation.

The rev-erbA(alpha) gene, belonging to the steroid receptor superfamily of transcription factors, is highly conserved during evolution but little is known so far about its functions in development or in adult physiology. Here, we describe genetically altered mice lacking the rev-erbA(alpha) gene. These animals do not show any obvious phenotype in either fat tissue or skeletal muscle, despite the known regulation of rev-erbA(alpha) expression during adipocyte and myotube differentiation in vitro. However, during the second week of life, the cerebellum of rev-erbA(alpha) mutants presents several unexpected abnormalities, such as alterations in the development of Purkinje cells, delay in the proliferation and migration of granule cells from the external granule cell layer and increased apoptosis of neurons in the internal granule cell layer. Interestingly, the expression pattern of rev-erbA(alpha) suggests that the abnormalities observed in the external granule cell layer could be secondary to Purkinje cell alterations. Taken together, our data underline the importance of rev-erbA(alpha)expression for the appropriate balance of transcriptional activators and repressors during postnatal cerebellar development.

The GAGE-1 gene was identified previously as a gene that codes for an antigenic peptide, YRPRPRRY, which was presented on a human melanoma by HLA-Cw6 molecules and recognized by a clone of CTLs derived from the patient bearing the tumor. By screening a cDNA library from this melanoma, we identified five additional, closely related genes named GAGE-2-6. We report here that further screening of this library led to the identification of two more genes, GAGE-7B and -8. GAGE-1, -2, and -8 code for peptide YRPRPRRY. Using another antitumor CTL clone isolated from the same melanoma patient, we identified antigenic peptide, YYWPRPRRY, which is encoded by GAGE-3, -4, -5, -6, and -7B and which is presented by HLA-A29 molecules. Genomic cloning of GAGE-7B showed that it is composed of five exons. Sequence alignment showed that an additional exon, which is present only in the mRNA of GAGE-1, has been disrupted in gene GAGE-7B by the insertion of a long interspersed repeated element retroposon. These GAGE genes are located in the p11.2-p11.4 region of chromosome X. They are not expressed in normal tissues, except in testis, but a large proportion of tumors of various histological origins express at least one of these genes. Treatment of normal and tumor cultured cells with a demethylating agent, azadeoxycytidine, resulted in the transcriptional activation of GAGE genes, suggesting that their expression in tumors results from a demethylation process.