Recoverin, a photoreceptor-specific calcium-binding protein, is expressed by the tumor of a patient with cancer-associated retinopathy.

Recoverin is a member of the EF-hand family of calcium-binding proteins involved in the transduction of light by vertebrate photoreceptors. Recoverin also was identified as an autoantigen in the degenerative disease of the retina known as cancer-associated retinopathy (CAR), a paraneoplastic syndrome whereby immunological events lead to the degeneration of photoreceptors in some individuals with cancer. In this study, we demonstrate that recoverin is expressed in the lung tumor of a CAR patient but not in similar tumors obtained from individuals without the associated retinopathy. Recoverin was identified intially by Western blot analysis of the CAR patient's biopsy tissue by using anti-recoverin antibodies generated against different regions of the recoverin molecule. In addition, cultured cells from the biopsy tissue expressed recoverin, as demonstrated by reverse transcription-PCR using RNA extracted from the cells. The immunodominant region of recoverin also was determined in this study by a solid-phase immunoassay employing overlapping heptapeptides encompassing the entire recoverin sequence. Two linear stretches of amino acids (residues 64-70, Lys-Ala-Tyr-Ala-Gln-His-Val; and 48-52, Gln-Phe-Gln-Ser-Ile) made up the major determinants. One of the same regions of the recoverin molecule (residues 64-70) also was uniquely immunopathogenic, causing photoreceptor degeneration upon immunization of Lewis rats with the corresponding peptide. These data demonstrate that the neural antigen recoverin more than likely is responsible for the immunological events associated with vision loss in some patients with cancer. These data also establish CAR as one of the few autoimmune-mediated diseases for which the specific self-antigen is known.

Construction of a 2.8-megabase yeast artificial chromosome contig and cloning of the human methylthioadenosine phosphorylase gene from the tumor suppressor region on 9p21.

Many human malignant cells lack methylthioadenosine phosphorylase (MTAP) enzyme activity. The gene (MTAP) encoding this enzyme was previously mapped to the short arm of chromosome 9, band p21-22, a region that is frequently deleted in multiple tumor types. To clone candidate tumor suppressor genes from the deleted region on 9p21-22, we have constructed a long-range physical map of 2.8 megabases for 9p21 by using overlapping yeast artificial chromosome and cosmid clones. This map includes the type IIFN gene cluster, the recently identified candidate tumor suppressor genes CDKN2 (p16INK4A) and CDKN2B (p15INK4B), and several CpG islands. In addition, we have identified other transcription units within the yeast artificial chromosome contig. Sequence analysis of a 2.5-kb cDNA clone isolated from a CpG island that maps between the IFN genes and CDKN2 reveals a predicted open reading frame of 283 amino acids followed by 1302 nucleotides of 3' untranslated sequence. This gene is evolutionarily conserved and shows significant amino acid homologies to mouse and human purine nucleoside phosphorylases and to a hypothetical 25.8-kDa protein in the pet gene (coding for cytochrome bc1 complex) region of Rhodospirillum rubrum. The location, expression pattern, and nucleotide sequence of this gene suggest that it codes for the MTAP enzyme.

Dietary lipids and calorie restriction affect mammary tumor incidence and gene expression in mouse mammary tumor virus/v-Ha-ras transgenic mice.

We have studied the effects of food restriction (FR) and substitution of fish oil (FO; omega 3) for corn oil (CO; omega 6) on breast tumor incidence and survival in mouse mammary tumor virus/v-Ha-ras transgenic (Onco) mice. The diets were as follows: group 1, 5% (wt/wt) CO fed ad libitum (AL); group 2, 5% CO, restricted calories (40% fewer calories than AL; FR); group 3, 20% CO fed AL; and group 4, 20% FO fed AL. After 3 years, 40% of FR Onco (group 2) mice were alive, whereas there were no survivors in the other three groups. Similarly, tumor incidence was reduced to 27% (5 out of 18) in FR animals (group 2), whereas it was 83% (11 out of 13) in group 1 mice, 89% (16 out of 18) in group 3 mice, and 71% (10 out of 14) in group 4 mice. These protective effects of FR on survival and tumor incidence were paralleled by higher expression of the tumor suppressor gene p53 (wild type) and free-radical scavenging enzymes (catalase and superoxide dismutase) in breast tumors. Immunoblotting showed less ras gene product, p21, and increased p53 levels in the tumors of FR mice. In addition, FR decreased RNA levels of c-erbB-2, interleukin 6, and the transgene v-Ha-ras in tumors. In contrast, analysis of hepatic mRNA from tumor-bearing FR mice revealed higher expression of catalase, glutathione peroxidase, and superoxide dismutase. Survival and tumor incidence were not influenced significantly by dietary supplementation with FO in place of CO. Taken together, our studies suggest that moderate restriction of energy intake significantly inhibited the development of mammary tumors and altered expression of cytokines, oncogenes, and free-radical scavenging enzymes.

Identification by representational difference analysis of a homozygous deletion in pancreatic carcinoma that lies within the BRCA2 region.

Homozygous deletions have been central to the discovery of several tumor-suppressor genes, but their finding has often been either serendipitous or the result of a directed search. A recently described technique [Lisitsyn, N., Lisitsyn, N. & Wigler, M. (1993) Science 259, 946-951] held out the potential to efficiently discover such events in an unbiased manner. Here we present the application of the representational difference analysis (RDA) to the study of cancer. We cloned two DNA fragments that identified a homozygous deletion in a human pancreatic adenocarcinoma, mapping to a 1-centimorgan region at chromosome 13q12.3 flanked by the markers D13S171 and D13S260. Interestingly, this lies within the 6-centimorgan region recently identified as the BRCA2 locus of heritable breast cancer susceptibility. This suggests that the same gene may be involved in multiple tumor types and that its function is that of a tumor suppressor rather than that of a dominant oncogene.

Mouse mammary tumor viruses with functional superantigen genes are selected during in vivo infection.

Mouse mammary tumor virus (MMTV) encodes a superantigen that is important for viral infectivity in vivo. To determine whether superantigen function was required for infection by milk-borne MMTV, we created HYB PRO/Cla transgenic mice. These mice produced a full-length, packaged viral RNA with a frameshift mutation that caused premature termination of the superantigen protein. Young HYB PRO/Cla mice showed no deletion of their cognate V beta 14+ T cells, although they shed virus in their milk. The nontransgenic offspring of the HYB PRO/Cla mice were infected with this virus, since transgene-specific viral transcripts were detected in their mammary glands. Surprisingly, these offspring demonstrated the progressive deletion of V beta 14+ T cells characteristic of exogenous MMTV (C3H) infection. Sequence analysis demonstrated that these newly acquired viruses had reconstituted superantigen open reading frames resulting from recombination between the HYB PRO/Cla and endogenous Mtv-1 proviral RNAs. Thus, there is selection during the infection process for MMTVs with functional superantigen genes.

Involvement of the ALL-1 gene in a solid tumor.

Translocations involving chromosome band 11q23, found in 5-10% of human acute leukemias, disrupt the ALL-1 gene. This gene is fused by reciprocal translocation with a variety of other genes in acute lymphoblastic and myelogenous leukemias, and it undergoes self-fusion in acute myeloid leukemias with normal karyotype or trisomy 11. Here we report an alteration of the ALL-1 gene in a gastric carcinoma cell line (Mgc80-3). Characterization of this rearrangement revealed a three-way complex translocation, involving chromosomes 1 and 11, resulting in a partial duplication of the ALL-1 gene. Sequencing of reverse transcription-PCR products and Northern blot analysis showed that only the partially duplicated ALL-1 gene was transcribed, producing an mRNA with exon 8 fused to exon 2. This report of ALL-1 gene rearrangement in a solid tumor suggests that ALL-1 plays a role in the pathogenesis of some solid malignancies. The absence of the normal transcript in this cell line, in association with the loss-of-heterozygosity studies on chromosome 11q23 seen in solid tumors, suggests that ALL-1 is involved in tumorigenesis by a loss-of-function mechanism.

Induction of the gene encoding mucosal vascular addressin cell adhesion molecule 1 by tumor necrosis factor alpha is mediated by NF-kappa B proteins.

Mucosal vascular addressin cell adhesion molecule 1 (MAdCAM-1) is involved in trafficking of lymphocytes to mucosal endothelium. Expression of MAdCAM-1 is induced in the murine endothelial cell line bEnd.3 by tumor necrosis factor alpha (TNF-alpha), interleukin 1, and bacterial lipopolysaccharide. Here we show that TNF-alpha enhances expression of a firefly luciferase reporter directed by the MAdCAM-1 promoter, confirming transcriptional regulation of MAdCAM-1. Mutational analysis of the promoter indicates that a DNA fragment extending from nt -132 to nt +6 of the gene is sufficient for TNF-alpha inducibility. Two regulatory sites critical for TNF-alpha induction were identified in this region. DNA-binding experiments demonstrate that NF-kappa B proteins from nuclear extracts of TNF-alpha-stimulated bEnd.3 cells bind to these sites, and transfection assays with promoter mutants of the MAdCAM-1 gene indicate that occupancy of both sites is essential for promoter function. The predominant NF-kappa B binding activity detected with these nuclear extracts is a p65 homodimer. These findings establish that, as with other endothelial cell adhesion molecules, transcriptional induction of MAdCAM-1 by TNF-alpha requires activated NF-kappa B proteins.