Galectin-8 binding to integrins inhibits cell adhesion and induces apoptosis.

The interaction of cells with the extracellular matrix regulates cell adhesion, motility, growth, survival and differentiation through integrin-mediated signal transduction. Here we demonstrate that galectin-8, a secreted mammalian (beta)-galactoside binding protein, inhibits adhesion of human carcinoma (1299) cells to plates coated with integrin ligands, and induces cell apoptosis. Pretreatment of the cells with Mn(2+), which increases the affinity of integrins for their ligands, abolished the inhibitory effects of galectin-8. The inhibitory effects of galectin-8 were specific and were not mimicked by plant lectins or other galectins (galectin-1 and galectin-3). In accordance with its anti-adhesive effects, transfection of galectin-8 cDNA into 1299 cells significantly reduced (by 75%) colony formation, when compared to the number of colonies formed by cells transfected with an empty vector. Affinity chromatography over immobilized galectin-8 indicated that few membrane proteins interacted with galectin-8 in a sugar-dependent manner. Microsequencing and western immunoblotting revealed that (alpha)(3)(beta)(1 )integrin derived from 1299 as well as other cells (e.g. HeLa and human endothelial cells) is a major galectin-8 binding-protein. Furthermore, immunoprecipitation and immunohistochemical studies suggested that endogenous galectin-8, secreted from 1299 cells, forms complexes with (alpha)(3)(beta)(1) integrins expressed on the surface of 1299 cells. Galectin-8 also interacts with other members of the integrin family, like (alpha)(6)(beta)(1 )integrins. In contrast, galectin-8 only minimally interacts with (alpha)(4 )or (beta)(3 )integrins. We propose that galectin-8 is an integrin binding-protein that interacts to a different extent with several, but not all members of the integrin family. Binding of galectin-8 modulates integrin interactions with the extracellular matrix and thus regulates cell adhesion and cell survival.

Inhibition of oncogene-mediated transformation by ectopic expression of p21Waf1 in NIH3T3 cells.

The p53-regulated p21Waf1 protein is a universal inhibitor of cyclin-dependent kinases (CDKs). To study the potential tumor-suppressive properties of CDK inhibitors, the ability of p21Waf1 to interfere with oncogene-mediated cellular transformation was analysed in the NIH3T3 cell system. Cotransfection of waf1 together with activated ras or several other oncogenes into NIH3T3 cells potently inhibited the formation of transformed foci in a dose-dependent manner. Expression of the CDK-binding N-terminal half of p21Waf1 (N-p21Waf1) was necessary and sufficient to inhibit Ras-induced focus formation. In contrast, expression of the C-terminal domain (C-p21Waf1) had no effect on Ras-induced focus formation. Immunofluorescence analysis revealed that ectopically expressed p21Waf1 and C-p21Waf1 were localized in the nucleus, while N-p21Waf1 was found in the cytoplasm, with the tendency to accumulate around the nuclear membrane. Surprisingly, stable NIH3T3 transfectants expressing ectopic p21Waf1 grew at the same rate and displayed similar cell cycle distribution as NIH3T3 cells transfected with the same vector containing no insert. However, ectopic p21Waf1 expression did inhibit Ras-mediated anchorage-independent colony formation, indicating that p21Waf1 can selectively interfere with oncogene-mediated transformation without affecting NIH3T3 cell growth, at least at the levels of p21Waf1 expression achieved in these experiments. Transient transfection of waf1 into NIH3T3 cells inhibited Ras-induced transcription from a E2F-responsive element but not from a serum-responsive element, indicating that p21Waf1 acts downstream of early transcriptional events induced by Ras but upstream of E2F-controlled gene transcription. These results provide evidence that p21Waf1 potently suppresses oncogene-mediated cellular transformation of NIH3T3 cells and that it may do so by inhibiting E2F-driven transcription of S phase genes.

The tumor suppression function of p21Waf is contained in its N-terminal half ('half-WAF').

The Waf-1 encoded protein, p21, mediates p53 suppression of tumor cell growth. Overexpression of p21 in the H1299 tumor cell line suppresses colony formation similar to that resulted from p53 overexpression. In an effort to localize the tumor suppression function within the structure of p21 we utilized vectors constructed with systematic truncations of p21 and tested their efficiency in suppressing tumor cell growth. We demonstrate that the N-terminal half of the p21 molecule (residues 1-80 and 1-89) shows better tumor cell growth suppression than the entire p21 molecule whereas the C-terminal half of p21 does not show this effect. These results may have implications for gene therapy of cancer.

Localization of the MAGE1 gene encoding a human melanoma antigen to chromosome Xq28.

MAGE1 encodes a tumor specific antigen MZ2-E that elicited a cytotoxic T lymphocytic response (CTL) in the patient from whom it was derived. In this study, cDNA and genomic probes have been used to localize this gene by Southern analysis of a human-rodent somatic cell hybrid panel. The probes detect a small multigene family, and both MAGE1 and several other members of this family are located on the long arm of the human X chromosome. A cosmid with a 12-kb insert including the entire MAGE1 gene was biotinylated and used to further localize the gene to Xq28 by in situ hybridization of metaphase spreads. The function of this antigen in normal cells and tumor cells currently remains unclear.

KIT ligand (mast cell growth factor) inhibits the growth of KIT-expressing melanoma cells.

Previous studies in vivo and in vitro show that KIT kinase promotes normal melanocyte development and growth. However, the role of the KIT proto-oncogene in neoplastic melanocytes is not certain. We therefore examined KIT expression and function in human melanomas. Our results show that KIT mRNA was expressed in 12 of 28 melanoma cell lines (approximately 40%), mainly in those originating from pigmented tumors. Surprisingly, activation of KIT with mast cell growth factor (MGF) in melanoma cells produced biological responses opposite to those elicited in normal melanocytes. MGF inhibited rather than stimulated the growth of metastatic melanoma cell lines. The opposite effects may be due to aberrant signal transduction by KIT in melanoma cells in response to MGF. The in vitro inhibition of melanoma cells by MGF suggests that growth in vivo of this tumor is not promoted by KIT kinase activation, but rather that transformed melanocytes might regress when MGF is expressed in their immediate environment.

Differential expression of MAGE-1, -2, and -3 messenger RNA in transformed and normal human cell lines.

The MAGE-1 gene codes for a tumor-specific antigen, MZ2-E, that elicited a cytotoxic T-lymphocyte response in the melanoma patient from whom it was derived. We have developed a simplified method, using polymerase chain reaction amplification of exon 3 followed by restriction enzyme pattern analysis, to distinguish expression of the MAGE-1 gene from MAGE-2 and MAGE-3, other members of this gene family. MAGE-1 mRNA was expressed in 53% of 17 melanoma lines, two of seven Epstein-Barr virus-transformed B-cell lines, and 2 of 5 breast cell lines including a line established form normal breast epithelium. MAGE-1 is not likely to be the common melanoma antigen recognized by the other HLA-A1- or HLA-A2-restricted cytotoxic T-lymphocytes examined in this study, but the fact that it is expressed in about 50% of melanoma cell lines makes it a reasonable target for the immunotherapy of patients bearing HLA-A1.

Recognition of shared melanoma antigens by human tumor-infiltrating lymphocytes.

We have established that melanomas express shared tumor antigens (Ags) that can be recognized by T cells if presented in the context of self-MHC molecules. Tumor-infiltrating lymphocytes (TILs) from six melanoma patients were tested for lysis of large panels of HLA-matched or unmatched targets representing a variety of tissue types. Lysis was specific for allogeneic melanomas sharing at least one HLA-A, -B, or -C Ag with TILs, and demonstrated commonly expressed tumor Ags. Similar findings were obtained when cytokine secretion by TILs was used to indicate specific Ag recognition. Transfection of the HLA-A2.1 gene into HLA-A2- melanoma lines conferred susceptibility to lysis by HLA-A2 restricted melanoma TILs, demonstrating expression of common tumor Ags among patients of diverse HLA types. These findings have important implications for developing broadly applicable cancer immunotherapies such as vaccines.

c-Kit-kinase induces a cascade of protein tyrosine phosphorylation in normal human melanocytes in response to mast cell growth factor and stimulates mitogen-activated protein kinase but is down-regulated in melanomas.

The proto-oncogene c-Kit, a transmembrane receptor tyrosine kinase, is an important regulator of cell growth whose constitutively active oncogenic counterpart, v-kit, induces sarcomas in cats. Mutations in murine c-kit that reduce the receptor tyrosine kinase activity cause deficiencies in the migration and proliferation of melanoblasts, hematopoietic stem cells, and primordial germ cells. We therefore investigated whether c-Kit regulates normal human melanocyte proliferation and plays a role in melanomas. We show that normal human melanocytes respond to mast cell growth factor (MGF), the Kit-ligand that stimulates phosphorylation of tyrosyl residues in c-Kit and induces sequential phosphorylation of tyrosyl residues in several other proteins. One of the phosphorylated intermediates in the signal transduction pathway was identified as an early response kinase (mitogen-activated protein [MAP] kinase). Dephosphorylation of a prominent 180-kDa protein suggests that MGF also activates a phosphotyrosine phosphatase. In contrast, MGF did not induce proliferation, the cascade of protein phosphorylations, or MAP kinase activation in the majority of cells cultured from primary nodular and metastatic melanomas that grow independently of exogenous factors. In the five out of eight human melanoma lines expressing c-kit mRNAs, c-Kit was not constitutively activated. Therefore, although c-Kit-kinase is a potent growth regulator of normal human melanocytes, its activity is not positively associated with malignant transformation.

Shared human melanoma antigens. Recognition by tumor-infiltrating lymphocytes in HLA-A2.1-transfected melanomas.

Three predominantly CD8+ CTL lines, TIL 501, TIL 620, and TIL 660, were generated from three HLA-A2+ melanoma patients by culturing tumor-infiltrating lymphocytes in 1000 U/ml IL-2. These tumor-infiltrating lymphocytes lysed 12 of 18 HLA-A2+ autologous and allogeneic melanomas, but none of 20 HLA-A2-negative melanomas. They also did not lyse the MHC class I negative lymphoma-leukemia cell lines, Daudi, K562, or HLA-A2+ non-melanoma cell lines including PHA or Con A-induced lymphoblast, fibroblast, EBV-transformed B cell, Burkitt's B cell lymphoma, and colon cancer cell lines. Autologous and allogeneic melanoma lysis was inhibited by anti-CD3, by anti-MHC class I, and by anti-HLA-A2 mAb, indicating recognition of shared tumor Ag among melanoma cell lines in a TCR-dependent, HLA-A2-restricted manner. Six HLA-A2-negative melanoma cell lines obtained from five HLA-A2-negative patients were co-transfected with the HLA-A2.1 gene and pSV2neo. All 17 cloned transfectants expressing cell surface HLA-A2 molecules, but none of 12 transfectants lacking HLA-A2 expression, were lysed by these three HLA-A2-restricted, melanoma-specific CTL. Lysis of the HLA-A2+ transfectants was inhibited by anti-CD3, by anti-MHC class I, and by anti-HLA-A2 mAb, indicating recognition of shared tumor Ag on transfectants in a TCR-dependent, HLA-A2-restricted manner. These results identify the HLA-A2.1 molecule as an Ag-presenting molecule for melanoma Ag. They also suggest that common melanoma Ag are expressed among melanoma patients regardless of HLA type. These findings have implications for the development of melanoma vaccines that would induce antitumor T cell responses.

Frequent generation of nonrescuable reorganized Moloney murine sarcoma viral genomes.

Nonproducer transformants infected with wild-type Moloney murine sarcoma virus were screened for the generation of variants with reorganized genomes. Seven of 20 lines contained such viral genomes, 5 of which were found to be nonrescuable. Blotting analysis indicated that viral RNA molecules transcribed from these variant genomes could not be encapsidated into virions. The nonrescuable genomes were molecularly cloned and all were found to have suffered deletions or deletions/inversions involving the 5' long terminal repeat as well as some adjacent sequences. Nucleotide sequence analysis suggested that the long terminal repeat or the tetranucleotides G-G-T-C and G-A-C-C (or both) were involved in the generation of these mutants. Transfection studies showed that the cloned DNAs of the 5 mutants transformed NIH/3T3 monolayers. Removal of the 3' long terminal repeat from the genomes that lacked the 5' long terminal repeat or carried it in an inverted orientation abolished or considerably reduced the transforming activity.

The nucleotide sequence of the rat cytoplasmic beta-actin gene.

The nucleotide sequence of the rat beta-actin gene was determined. The gene codes for a protein identical to the bovine beta-actin. It has a large intron in the 5' untranslated region 6 nucleotides upstream from the initiator ATG, and 4 introns in the coding region at codons specifying amino acids 41/42, 121/122, 267, and 327/328. Unlike the skeletal muscle actin gene and many other actin genes, the beta-actin gene lacks the codon for Cys between the initiator ATG and the codon for the N-terminal amino acid of the mature protein. The usage of synonymous codons in the beta-actin gene is nonrandom, and is similar to that in the rat skeletal muscle and other vertebrate actin genes, but differs from the codon usage in yeast and soybean actin genes.

Evolutionary aspects of immunoglobulin heavy chain variable region (VH) gene subgroups.

We isolated and determined the sequences of two human germ-line heavy chain variable region (VH) genes and compared them with mouse VH genes. The results show that the human VHI subgroup is evolutionarily related to the mouse VHII subgroup. Evolutionary preservation of homologies in VH genes of the same subgroup includes not only the coding region but also intron size and homology in noncoding regions. This suggests that a VH gene subgroup constitutes a multigene family that undergoes concerted evolution. The homology between genes of the same subgroup in different species is greater than that between genes of different subgroups within a species. One of the VHII genes contains, in complementarity-determining region 2 (CDR2), a 13-base-pair previously shown to be in CDR2 of a VHIII gene and in a heavy chain diversity region gene, DH [Wu, T. T. & Kabat, E. A. (1982) Proc. Natl. Acad. Sci. USA 79, 5031-5032], suggesting the insertion of diversity region gene sequences into the VH gene. One of the human VH genes is a pseudogene because of a terminator, which, together with our previous results, shows that the VH gene repertoire contains 40% pseudogenes. In one of the VH genes, direct and inverted repeats at both 5' and 3' ends of the gene suggest a potential transposable element that encompasses the entire VH gene. It is possible that such a structure may facilitate saltatory replication and rapid expansion of VH gene families.

Analysis of recombinant DNA clones specific for the murine p53 cellular tumor antigen.

Three cDNA clones, corresponding to two non-overlapping regions of the mRNA coding for the mouse p53 cellular tumor antigen, were isolated and characterized. In hybridization-selection assays, these clones were capable of selectively binding p53 mRNA, as demonstrated by in vitro translation and immunoprecipitation with anti-p53 monoclonal antibodies. The p53 mRNA appeared to be the only messenger species specifically selected by these clones. The size of the p53 mRNA was found to be approximately 2 kb, and its levels to vary substantially among different types of transformed cells. Evidence was found for the existence of two distinct p53-specific genes in mouse genomic DNA. Two partially overlapping recombinant phage clones were obtained, both derived from the same p53-specific genomic DNA region. The orientation of the various cDNA clones relative to that of the p53 mRNA was established by S1 analysis and the relationship between the cDNA clones and the genomic ones was determined by comparative restriction enzyme mapping and nucleic acid hybridization.

The DNA sequence of the H-2kb gene: evidence for gene conversion as a mechanism for the generation of polymorphism in histocompatibilty antigens.

We have determined the DNA sequence of the H-2Kb gene of the C57B1/10 mouse. Comparison of this sequence with that of the allelic H-2Kd shows surprisingly that the exons have accumulated more mutations than their introns. Moreover, many of these changes in the exons are clustered in short regions or hot spots. Additional comparison of these sequences with the H-2Ld and H-2Db sequences shows that, in several cases, the altered sequence generated at the hot spot is identical to the corresponding region of a non-allelic H-2 gene. The clustered changes are responsible for 60% of the amino acid differences between the H-2Kb and H-2Kd genes and suggest that micro-gene conversion events occurring within the exons and involving only tens of nucleotides are an important mechanism for the generation of polymorphic differences between natural H-2 alleles.

Nucleotide sequence of the rat skeletal muscle actin gene.

The actins constitute a family of highly conserved proteins found in all eukaryotic cells. Their conservation through a very wide range of taxonomic groups and the existence of tissue-specific isoforms make the actin genes very interesting for the study of the evolution of genes and their controlling elements. On the basis of amino acid sequence data, at least six different mammalian actins have been identified (skeletal muscle, cardiac muscle, two smooth muscle actins and the cytoplasmic beta- and gamma-actins). Rat spleen DNA digested by the EcoRI restriction enzyme contains at least 12 different fragments with actin-like sequences but only one which hybridized, in very stringent conditions, with the skeletal muscle cloned cDNA probe. Here we describe the sequence of the actin gene in that fragment. The nucleotide sequence codes for two amino acids, Met-Cys, preceding the known N-terminal Asp of the mature protein. There are five small introns in the coding region and a large intron in the 5'-untranslated region. Comparison of the structure of the rat skeletal muscle actin gene with available data on actin genes from other organisms shows that while the sequenced actin genes from Drosophila and yeast have introns at different locations, introns located at codons specifying amino acids 41, 121, 204 and 267 have been preserved at least from the echinoderm to the vertebrates. A similar analysis has been done by Davidson. An intron at codon 150 is common to a plant actin gene and the skeletal muscle acting gene.

Organization and evolution of immunoglobulin VH gene subgroups.

The organization and evolution of immunoglobulin variable region genes was studied by comparing human and mouse heavy chain variable region (VH) genes. We show that a VH gene subgroup constitutes a physically linked multigene family separated from another VH subgroup. We mapped the VHIII gene subgroup to be 3' to the VHII gene subgroup based on deletion of VH genes after V-D-J rearrangement. The results indicate that the human VHIII gene subgroup underwent a significant gene expansion as compared to the mouse VHIII subgroup. Amino acid sequence data indicate that human VHIII genes correspond to only a small subset of mouse VHIII genes. Human VHIII genes contain a shorter intron and are two codons shorter than most BALB/c mouse VHIII genes. The nature of nucleotide substitutions between VH genes within a species (human) is similar to that between genes of different species (human/mouse). Both contain approximately 50% silent substitutions.

Simple DNA sequences in homologous flanking regions near immunoglobulin VH genes: a role in gene interaction?

Five closely related immunoglobulin VH genes (subgroup II) were compared by sequencing of several kb of DNA. In three of the genes homology greater than 75% was found along an area of 4 kb that includes the coding region. The homology in flanking regions is only slightly lower than that in the coding sequences. Two other genes, which are located on the same EcoRI fragment, show high homology to the first three genes in the coding and immediately flanking regions. In more distant flanking regions no homology is found with the first three genes. This indicates that their evolutionary history differs from that of the other three genes. A region of simple DNA sequence composed of repetitive TCC and TCA elements was found at a distance of approximately 380 bp upstream from the initiator ATG of these VH genes. This region is the site where the two sets of genes abruptly start to diverge. The structure of the simple DNA sequence in the various VH genes suggests that it may be involved in gene interaction. We propose that both simple DNA sequences and homology in flanking regions serve a function in the correction of VH genes, which seem to be rather free to diverge and drift into pseudogenes. A correction mechanism may help this gene family to maintain its two major features, multiplicity and diversity.

Isolation and characterization of rat skeletal muscle and cytoplasmic actin genes.

Southern blots of rat genomic DNA indicate the existence of at least 12 EcoRI DNA fragments containing actin gene sequences. By using specific probes and stringent conditions of hybridization, it was found that only one of these fragments contains sequences of the skeletal muscle alpha-actin gene. Recombinant bacteriophages originating from eight different actin genes were isolated from rat genomic DNA libraries. One of them, Act 15, contains the skeletal muscle actin gene. Another clone, Act I, contains a gene coding for a cytoplasmic actin, identified tentatively as the beta-actin gene. Both genes have a large intron very close to the 5' end of their transcribed region, followed by several small introns. DNA sequence analysis and comparison with the available data on actin genes in other organisms indicated an interesting relationship between the positions of introns and the evolutionary relatedness. Several intron sites are conserved from at least the echinoderms to the vertebrates; others appear to be present in some actin genes and not in others.

Analysis of myogenesis with recombinant DNA techniques.

Recombinant phages containing the rat skeletal muscle alpha-actin gene and the cytoplasmic beta-actin gene were isolated and the structure of these genes was determined. Both genes contain a large intron in the 5' untranslated region and smaller introns at codons 41, 267 and 327. In addition, the alpha-actin contains introns at codons 150 and 204 not present in the beta-actin gene, whereas the beta-actin gene contains an intron at codon 121. The evolutionary aspects of these findings are discussed. Active genes are organized in chromatin in a conformation which renders them preferentially sensitive to digestion with nucleolytic enzymes. The DNAase I sensitivity of genes programmed to be expressed during myogenesis was tested in a cloned cell population of a myogenic cell line. It was found that these genes are not preferentially sensitive to DNAase I in the chromatin of proliferating mononucleated cells. They become DNAase I sensitive during terminal differentiation.

Polymorphism of germ-line immunoglobulin VH genes correlates with allotype and idiotype markers.

The polymorphic nature of the immunoglobulin VH genes was investigated by Southern blot analysis of liver DNA of sixteen different mouse strains and hybridization with VH probes. Differences in restriction enzyme pattern (REP) were observed and six different patterns of restriction fragments were found for the sixteen strains analyzed. No equivalent polymorphism was observed in another multigene family, the actins. The six patterns correlate with immunoglobin constant region allotypes (Igh-1). Experiments with Igh-1-congenic strains suggest that the VH REP is linked to immunoglobulin constant region haplotype. Mouse strains which share inherited idiotypes also share identical VH restriction pattern. This provides a structural basis for the genetic linkage between idiotypes and allotypes. It also indicates that different strains carry different VH gene repertoires, which may be the basis for the expression of different inherited idiotypes in various strains. We propose that a VH group in a set of linked genes that are coinherited as a cluster with the constant region genes and that VH and Ch can be regarded as an extended haplotype.