Analysis of monoclonal antibodies to prostate-specific antigen: reactivity with native and recombinant prostate-specific antigen.

Epitope mapping analysis was performed on 53 antibodies submitted to the ISOBM TD-3 PSA Workshop. Western blotting and N-terminal amino acid sequencing, using both native and recombinant human prostate-specific antigen (rPSA), identified four different epitope groups for native PSA. Under reducing conditions native PSA was not recognized by 18/53 antibodies suggesting they reacted with conformation-dependent epitopes. Nine other antibodies reacted with a rPSA polypeptide doublet of 34-35 kD corresponding to different rPSA glycoforms. From sequence mapping studies 22/53 antibodies bound epitopes within amino acid residues 25-85, 3/53 antibodies bound to epitopes within residues 86-220, while 10/53 antibodies bound epitopes within residues 221-261. These results indicate that there are multiple immunogenic epitopes localized on the PSA molecule.

Summary report of the TD-3 workshop: characterization of 83 antibodies against prostate-specific antigen.

Twelve research groups participated in the ISOBM TD-3 Workshop in which the reactivity and specificity of 83 antibodies against prostate-specific antigen (PSA) were investigated. Using a variety of techniques including cross-inhibition assays, Western blotting, BIAcore, immunoradiometric assays and immunohistochemistry, the antibodies were categorized into six major groups which formed the basis for mapping onto two- and three-dimensional (2-D and 3-D) models of PSA. The overall findings of the TD-3 Workshop are summarized in this report. In agreement with all participating groups, three main antigenic domains were identified: free PSA-specific epitopes located in or close to amino acids 86-91; discontinuous epitopes specific for PSA without human kallikrein (hK2) cross-reactivity located at or close to amino acids 158-163; and continuous or linear epitopes shared between PSA and hK2 located close to amino acids 3-11. In addition, several minor and partly overlapping domains were also identified. Clearly, the characterization of antibodies from this workshop and the location of their epitopes on the 3-D model of PSA illustrate the importance of selecting appropriate antibody pairs for use in immunoassays. It is hoped that these findings and the epitope nomenclature described in this TD-3 Workshop are used as a standard for future evaluation of anti-PSA antibodies.

Transcriptional control of the human biliary glycoprotein gene, a CEA gene family member down-regulated in colorectal carcinomas.

Biliary glycoprotein (BGP) isoantigens are derived by alternative splicing from a single gene and are the human homologs of rat C-CAM and the mouse Bgp species. These glycoproteins represent a family of cell-adhesion molecules. The mouse Bgp isoforms also act as receptors for the hepatitis viral capsid-protein. BGP is a member of the carcinoembryonic antigen (CEA) gene family, which belongs to the immunoglobulin supergene family, yet it displays restricted expression patterns and unique functions. Since the loss or reduced expression of BGP is associated with human colorectal carcinomas, the elements in its upstream regulatory region were analyzed. A cluster of transcriptional initiation sites and the minimal promoter, located within 150 bp upstream of the major transcriptional start site, were active in human colon carcinoma and hepatoma cells. Unlike the CEA gene, BGP gene transcription was not modulated by a silencer region; repetitive elements in the BGP upstream region were not involved in activation or repression. Footprinting experiments identified two cis-acting elements and mobility-shift assays demonstrated that these elements bound several transcription factors, among them, USF, HNF-4 and an AP-2-like factor. In cotransfection experiments, both the USF and HNF-4 transcription factors transactivate the BGP gene promoter and compete for the same regulatory element. The Sp1 transcription factor, shown to be involved in CEA gene transcriptional regulation, does not bind to the BGP gene promoter. We, therefore, propose that the relative distributions and interactions of these transcription factors mediate distinct transcriptional regulation of the BGP gene in colon and liver; this regulation could be distorted during the oncogenic process.

Nonspecific cross-reacting antigen 50/90 is elevated in patients with breast, lung, and colon cancer.

A total of 22 genes have been identified in the carcinoembryonic antigen (CEA) gene family. The protein products of this family are highly homologous and include CEA, biliary glycoprotein, nonspecific cross-reacting antigen 50/90 (NCA 50/90), NCA 95, and pregnancy-specific beta-glycoprotein. We used a monoclonal antibody with high affinity to develop a specific enzyme-linked immunosorbent assay (ELISA) method for NCA 50/90 in serum and plasma. Our calibrators were based on affinity-purified recombinant protein from a baculovirus expression system. No significant reactivity with purified CEA, recombinant NCA 95, or recombinant biliary glycoprotein was found by Western blot analysis or in the ELISA method. Only 1 of 15 sera from pregnant women (chorionic gonadotropin > 1000 ng/ml) was positive in the NCA 50/90 ELISA, suggesting that this method does not detect pregnancy-specific glycoprotein. A cutoff value of 18 ng/ml was established based on the 95% value of serum and plasma from 147 healthy volunteers. Only 3 of 31 serum and plasma samples from patients with clinically inactive breast cancer were elevated above the cutoff value, but 44% of 136 samples from patients with clinically active breast cancer were positive. NCA 50/90 measurements were elevated in 7 of 25 patients with active breast cancer whose CEA and CA 15-3 values were below cutoff, and NCA 50/90 values do not correlate with CEA in breast cancer. In addition, we found sensitivities of 70, 39, and 42% for lung cancer, colon cancer, and leukemia, respectively. The sensitivity for non-small cell lung cancer was 85%, however, compared to 50% for small cell lung cancer. Serum from leukemia patients showed an overall sensitivity of 43%, but 71% (10 of 14) sera from patients with chronic myelogenous leukemia were positive compared to, for example, chronic lymphocytic leukemia where 0 of 7 sera had NCA 50/90 values above the cutoff. These studies suggest that NCA 50/90 may have clinical utility in the management of patients with a variety of cancers.

Expression and sequence of the gene for tissue inhibitor of metalloproteinases in patients with abdominal aortic aneurysms.

PURPOSE: The tissue inhibitors of metalloproteinases (TIMPs) are major inhibitors of several enzymes that are destructive to connective tissue, and TIMP-1 has been reported to be deficient in the wall of abdominal aortic aneurysms. This deficiency could represent failure of expression resulting from either local tissue conditions or mutation in the primary structure of the gene (or one of its regulatory elements). METHODS: Southern blotting techniques were used to examine the possibility of global deletions or inserts in the gene (14 patients); Northern blot techniques were performed to examine the expression of the mRNAs in cultured fibroblasts under basal conditions (six patients); and sequence analysis of the cDNA derived from fibroblast mRNAs was done after amplification by polymerase chain reaction (six patients). RESULTS: The Southern blots revealed a normal distribution of the known alleles of the gene without unique restriction-length polymorphisms. Fibroblast expression of TIMP mRNA was normal under basal conditions. Sequence analysis of the cDNAs revealed an identical-point polymorphism in two of the six patients (a single base pair substitution of C-->T at the third position in codon 101), but the amino acid was conserved. CONCLUSION: The studies reported here do not support the hypothesis that deficiency of TIMP-1 in specimens of aorta of patients with abdominal aortic aneurysms results from a primary genetic defect.

Human biliary glycoprotein gene: characterization of a family of novel alternatively spliced RNAs and their expressed proteins.

Eight different human biliary glycoprotein (BGP) isoantigens, structurally related members of the carcinoembryonic antigen family, CD66/67 family, and immunoglobulin superfamily, are derived by alternative splicing from a single genomic transcription unit. Novel BGP isoforms have been identified by polymerase chain reaction amplification and by DNA sequencing of amplified cDNA segments. In addition to verifying previously documented BGPs, we describe four new forms, two of which have unusual nonimmunoglobulin exons contributed by inverted Alu repeats. Determination of the genomic DNA sequence encompassing most of the known extracellular and intracellular domains demonstrates that the translatable Alu-like sequences are encoded in bona fide exons. The third novel BGP isoform contains none of the extracellular disulfide-linked immunoglobulin-like domains typical of these molecules but retains N-terminal and intracellular domains, suggesting distinct functions for N-terminal versus other disulfide-linked domains. cDNAs coding for each identified isoform have been transfected into COS7 monkey cells, and the resulting polypeptides are heavily N glycosylated but can be deglycosylated to their expected primary sizes. Many of these deglycosylated forms can be correlated with unique patterns of BGP expression in different cell lines, while in granulocytes, some previously undescribed or alternatively modified forms may predominate. The BGP family represents a potentially large but unknown source of functional diversity among cells of epithelial and hematopoietic origin. The availability of a defined set of expressed of BGP cDNAs should permit critical definition of their function.

Characterization of two new members of the pregnancy-specific beta 1-glycoprotein family from the myeloid cell line KG-1 and suggestion of two distinct classes of transcription unit.

Pregnancy-specific beta 1-glycoproteins (PSGs) represent a large group (approximately 12-15) of proteins, related to members of the carcinoembryonic antigen family, that are abundant in placental tissue and in the sera of pregnant women. We describe the isolation and characterization of two additional PSG cDNAs, PSG9 and PSG10, whose transcripts are largely expressed in placental tissue and to a lesser extent in some other cell types, including myeloid cells differentiated to granulocytes. PSG9 and PSG10 are representatives of two distinct classes of PSG protein that have N-termini with or without the Arg-Gly-Asp motif implicated in adhesion. In addition to this distinction at the amino acid level, our analysis of several PSG cDNAs suggests that the transcription units encoding these proteins may be further distinguished in their 3' untranslated sequences, thus suggesting possibilities for transcriptional regulation of the two major protein classes.

Human pregnancy-specific beta 1-glycoproteins are coded within chromosome 19.

We have isolated and characterized cDNAs that code for apoproteins having amino acid sequences highly similar to pregnancy-specific beta 1-glycoproteins (PS beta G). cDNAs coding for PS beta Gs, as well as the cDNA clone reported here, are members of the carcinoembryonic antigen (CEA) gene family. The previous localization of CEA-related genes to human chromosome 19, and the high level of DNA sequence conservation in the CEA family, suggested that the PS beta G genes are also located on this chromosome. We demonstrate here that chromosome 19 is indeed the site of PS beta G sequences. Our finding is in contrast to the recently reported indication that pregnancy-specific glycoproteins are encoded in chromosomes X and 6.

Carcinoembryonic antigens: alternative splicing accounts for the multiple mRNAs that code for novel members of the carcinoembryonic antigen family.

The recent cloning of complete cDNAs encoding carcinoembryonic antigen (CEA) and non-specific cross-reacting antigen has revealed the existence of a new gene family belonging to the immunoglobulin gene superfamily. We have reported the isolation of a partial CEA cDNA and of L-cell transfectant cell lines that express human antigens cross-reactive with commercial antibodies directed to native CEA (Kamarck, M., J. Elting, J. Hart, S. Goebel, P. M. M. Rae, J. Nedwin, and T. Barnett. 1987. Proc. Natl. Acad. Sci. USA. 84:5350-5354). In this study, we describe the identification and cloning of 3.9-, 3.7-, 2.2-, and 1.8-kb cDNAs and a 23-kb genomic transcription unit, which code for new members of the CEA gene family. DNA sequence analysis of these cloned DNAs establishes the existence of a set of four alternatively spliced mRNAs which are expressed in several tumor cell lines, in human fetal liver, and in L-cell transfectants. Deduced amino acid sequences of the encoded isoantigens show extensive similarity to CEA and nonspecific cross-reacting antigens, but in addition demonstrate transmembrane and cytoplasmic domains. We designate members of this antigen family transmembrane CEAs. The transmembrane CEA isoantigens share general structural characteristics with members of the immunoglobulin gene superfamily and can be specifically compared to the cell adhesion molecules, N-CAM (neural cell adhesion molecule) and MAG (myelin-associated glycoprotein).

Carcinoembryonic antigen family: expression in a mouse L-cell transfectant and characterization of a partial cDNA in bacteriophage lambda gt11.

Genomic DNA and mRNA from the adenocarcinoma cell line LoVo were used to generate L-cell transfectants and a bacteriophage lambda gt11 cDNA clone that express epitopes of carcinoembryonic antigen (CEA). Primary and secondary L-cell transfectants expressing CEA were selected with a fluorescence-activated cell sorter (FACS). These transfectants, including some clones that were selected for high-level CEA expression by multiple rounds of FACS sorting, express a surface protein of 150 kDa that reacts with all anti-CEA antibodies tested. In parallel, a cDNA library of LoVo poly(A)+ RNA was constructed in lambda gt11 and fusion proteins were screened with polyclonal antisera against CEA. One positive clone, lambda cLV7, was identified that hybridized specifically to transfectant DNA. The nucleic acid sequence of the cDNA insert (cLV7) contained two regions of extensive internal homology, with greater than 70% identity at the amino acid level. cLV7 hybridized to three mRNA species of LoVo cells and to a predominant mRNA of the CEA-expressing transfectants. Hybridization of cLV7 to restriction endonuclease-digested genomic DNA of colon carcinoma cells, normal human cells, and human-mouse somatic cell hybrids revealed the presence of multiple hybridizing bands, one of which was present in transfectant cells. These CEA-related sequences are not rearranged in tumors and, by somatic cell hybrid analysis, were mapped to human chromosome 19.

Factors influencing the yield of satellite DNA in extractions from Drosophila virilis and Drosophila melanogaster adults and embryos.

The application of different DNA extraction methods to identical batches of Drosophila virilis and Drosophila melanogaster flies or embryos has revealed that the ionic strength of a homogenization medium is of critical importance if chloroform extractions are performed. The low yield of satellite DNA after homogenization in low salt buffers is less severe if EDTA is included in the buffer. Phenol extraction procedures result in no such differential behavior of satellite and main band DNA, but under certain circumstances a particular satellite fraction of Drosophila virilis DNA may be lost.

DNA sequence organization in the genomes of five marine invertebrates.

The arrangement of repetitive and non-repetitive sequence was studied in the genomic DNA of the oyster (Crassostrea virginica), the surf clam (Spisula solidissima), the horseshoe crab (Limulus polyphemus), a nemertean worm (Cerebratulus lacteus) and a jelly-fish (Aurelia aurita). Except for the jellyfish these animals belong to the protostomial branch of animal evolution, for which little information regarding DNA sequence organization has previously been available. The reassociation kinetics of short (250-300 nucleotide) and long (2,000-3,000 nucleotide) DNA fragments was studied by the hydroxyapatite method. It was shown that in each case a major fraction of the DNA consists of single copy sequences less than about 3,000 nucleotides in length, interspersed with short repetitive sequences. The lengths of the repetitive sequences were estimated by optical hyperchromicity and S1 nuclease measurements made on renaturation products. All the genomes studied include a prominent fraction of interspersed repetitive sequences about 300 nucleotides in length, as well as longer repetitive sequence regions.