Immunogold and immunogold/silver staining in the ultrastructural localization of target molecules identified by monoclonal antibodies.

A new method for demonstrating the binding and internalization of target molecules identified by two different monoclonal antibodies (MAbs) is described. This double staining technique utilizes, in a pre-embedding procedure, an immunogold/silver staining and a MAb that recognized cell surface antigens and a post-embedding technique where only immunogold is used to identify intracytoplasmic antigens. We could demonstrate that immunogold and gold followed by silver enhancement are two highly sensitive and accurate techniques. Colloidal gold particles are versatile tracers at the electron microscopic level when used at two different particle sizes (5 and 20 nm in diameter) in a double labelling method for the simultaneous identification of antigenic sites on the same section; the combined use of immunogold and gold/silver staining for the simultaneous localization of breast cancer-associated antigens on the same thin section is more accurate. These techniques have also been used to visualize the internationalization of antigen-antibody complexes by incubating MAb-gold labelled tissue sections for 30 min at 37 degrees C. We conclude that the use of both colloidal gold and gold/silver constitutes a further improvement in immunoelectron microscopy techniques and can help visualize the relative pattern of reactivity of two MAbs on the same cells. Further applications of these techniques will have an important impact on the development and use of MAbs in oncology.

An in vitro model for cancer invasion: 1. Grading and quantitation of aggressiveness.

Invasion, destruction and replacement of normal tissues by cancer cells is the first critical step in the metastatic cascade and is the result of complex interactions between tumor and host factors. In an attempt to understand the complex mechanism of local invasion, we have developed a simple, reliable and generally applicable in vitro system using the confrontation of precultured heart fragments (PHF) with a constant known number of potentially invading (10(5)) cells. This grading system is based essentially on quantitative data standardized both for the portion of the explant invaded by the neoplastic cells and the type of invasion demonstrated. Using this system we could measure the ability of an aggressor cell to (a) adhere or attach to PHF (Grade 1), (b) invade the outer fibroblastic layers (Grade II) and/or the cardiac muscle cells (Grade III) and (c) destroy and completely replace the PHF (Grade IV). It was at once apparent from these experiments that, irrespective of their invasive capacity and/or their neoplastic state, both malignant (SKBR-2 III, BT-20, MCF-7, ZR-75-30 LoVo and YAC-1) and non-malignant (HBL-100) cells attach to the PHF. Only malignant cells, however, showed substantial local invasion of both the outer fibroblastic layers and/or the cardiac muscle cells. Most important for the actual problems of invasion in vivo, is the fact that malignant cells from different cell lines demonstrate a wide range in their invasion capacity: three different patterns of invasion were thus established; a highly invasive, a slowly invasive and a poorly invasive pattern. We also show that invasion and proliferation, as defined for the purposes of this study, are two different and independent properties of a given cell line. SKBR-2 III and YAC-1 are here shown to possess the most aggressive potential; they both invade the PHF very early and completely. The rapid proliferation of these aggressive cells and the destruction of the host tissue lead to the rapid disappearance of the myoblasts and their complete replacement by the invading cells. Non-malignant epithelial cells attached to but could not invade even the outer fibroblastic layers of the PHF.

Affinity purification of a high molecular weight human breast cancer-associated antigen identified by the BCD-B4 monoclonal antibody.

This study reports the purification and characterization of a high molecular weight human breast cancer-associated antigen identified by a previously described (1,2) murine monoclonal antibody, BCD-B4. Immunohistochemical analysis indicated that BCD-B4 recognizes an antigen expressed in an altered form on the human breast carcinoma cell line, BT-20, compared to the non-malignant human mammary epithelial cell line, HBL-100. Chemical treatments and enzymatic digestions suggested that the recognized moiety was a protein. The antigenic determinant was resistant to neuraminidase and periodate treatments but was sensitive to trypsin and proteinase K. The antigen was purified by affinity chromatography and its molecular weight, determined by SDS-PAGE analysis under non-reducing conditions, was proven to be 250 Kd. Under reducing conditions, the molecule dissociated into two polypeptides of 125 and 45 Kd, respectively. Both subunits could be isolated from normal HBL-100 and neoplastic BT-20 cellular protein extracts by affinity chromatography. The higher molecular weight subunit showed; however, qualitative and quantitative differences between the two cell lines: it was expressed in greater quantity on BT-20 cells and its molecular weight was 15 Kd higher. Both subunits could also be identified by immunoblots of BT-20 cells.

Establishment and characterization of a thymic medullary epithelial cell clone.

Thymic stromal cells were cultured in conditions which select for epithelial cells. These were then transformed in vitro by contact with N-methyl-N'-nitro-N-nitrosoguanidine and cloned at limit dilution. One of the clones was characterized as being of medullary origin on the basis of its reactivity with a battery of antibodies previously shown to distinguish cortical from medullary thymic epithelial cells. The importance of this clone lies in the potential it offers to delineate how various T cell subpopulations acquire their distinct markers and function within the thymus.

Specific adherence of thymocytes to a thymic medullary epithelial cell line.

TMF, a glycoprotein found preferentially at the contact points between thymocytes and thymic medullary epithelial cells, was investigated for its participation in adherence between these two cell types. It was found that thymocytes adhere specifically to a TMF+ cell line, that anti-TMF antibody can inhibit the adherence and that extraneous TMF can effectively compete with the adherence reaction. It is concluded that TMF is indeed involved in the temporary contact which occurs between thymocytes and medullary epithelial cells.

Immunocytochemical localization of a medullary thymic epithelial glycoprotein.

The thymic microenvironment is known to play a key role in T-cell differentiation, but the exact nature of the interactions between epithelial and lymphoid cells has not been fully elucidated. With a monoclonal antibody to a thymic epithelial glycoprotein, we report the localization of an antigen specific for medullary epithelial cells of the mouse thymus. This antigen is found in the Golgi apparatus of epithelial cells, and on their borders with adjoining lymphocytes. This location is compatible with the previously reported observation that differentiation signals transmitted to thymic lymphocytes by thymic epithelial cells require actual contact between these two cell types.

Specific targeting of membrane nodulins to the bacteroid-enclosing compartment in soybean nodules.

Rhizobium bacteroids in nodule cells are surrounded by the peribacteroid membrane (pbm), which is derived from the host plasma membrane during infection. The pbm was purified from R. japonicum 61A76-induced soybean nodules and analyzed by comparing it with the host cell plasma membrane for the presence of nodulins, nodule-specific plant proteins. Nodulins were found in pbm by reacting Western blots with a nodule-specific antiserum raised against the pbm. Peribacteroid fluid (the fluid enclosed in the pbm) was also found to contain several nodulins. The pbm nodulins were confirmed to be of plant origin by in vitro translation of poly(A) nodule mRNA followed by immunoprecipitation by the nodule-specific antiserum. Antibodies raised against a synthetic peptide corresponding to a repeated domain in nodulin-24, a pbm nodulin, and the nodule-specific pbm antiserum reacted exclusively with the pbm. The absence of pbm-nodulins in the plasma membrane suggests that the infected cells direct the intracellular transport of the pbm nodulins exclusively to this de novo synthesized subcellular compartment essential for symbiotic nitrogen fixation.

Ultrastructural localization of factor VIII procoagulant antigen in human liver hepatocytes.

Factor VIII is generally believed to circulate in blood as a multimeric complex of two glycoproteins which are physiologically and immunologically distinct. One component of the factor VIII complex is factor VIII procoagulant activity (FVIII:C) which is associated with factor VIII/procoagulant antigen (FVIII:Ag, formerly FVIII/CAg). The second, larger unit of the complex is factor VIII/von Willebrand factor (vWF:Ag, formerly factor VIII-related antigen or FVIIIRAg). FVIII:C has anti-haemophilic activity and is defective or deficient in patients with classical haemophilia, and vWF:Ag is absent in patients with von Willebrand disease. FVIII:Ag was demonstrated recently in endothelial cells lining hepatic sinusoids, by using immunoperoxidase staining and light microscopy, whereas biochemical data had indicated its presence predominantly in the hepatocyte fractions and in lesser amounts in endothelial cells. Moreover, recent hybridization experiments detected FVIII:C messenger RNA in liver and kidney tissues. Despite several efforts, the cells responsible for FVIII:C synthesis have not been unequivocally identified. Here we use protein A-gold complex labelling to demonstrate the ultrastructural localization of FVIII:C in human liver cells; the results indicate that hepatocytes may synthesize FVIII:Ag.

Primary structure of the soybean nodulin-35 gene encoding uricase II localized in the peroxisomes of uninfected cells of nodules.

Nodulin-35 (N-35), a subunit of nodule-specific uricase (uricase II) of soybean (Glycine max), is shown to be preferentially synthesized on free polysomes during nodule development and is localized in peroxisomes of the uninfected cells of this tissue. A cDNA clone, isolated by using mRNA from immunoprecipitated polysomes, revealed the primary structure of this protein with a molecular mass of 35,100. That this clone represents N-35 was confirmed by comparing the deduced amino acid sequence with the partial sequence of a CNBr-cleaved peptide of purified N-35. Southern blot hybridizations with genomic DNA suggest that there are several EcoRI fragments containing N-35 sequences. Three of these sequences were isolated from a genomic library of soybean. Nucleotide sequence analysis showed that the complete gene extends almost 5000 base pairs on two EcoRI fragments and the coding region (309 codons) is interrupted by seven introns ranging in size from 154 to 1341 base pairs. Lack of a signal sequence and its translation on free polysomes suggest that N-35 is posttranslationally transported to the peroxisomes. Furthermore, there is no cross-hybridization of N-35 cDNA with RNA from young (3- to 4-day) roots and leaves, indicating that the observed "uricase" activity in these tissues is due to the product of a different gene.

Isolation and expression of Rhizobium japonicum cloned DNA encoding an early soybean nodulation function.

A first visible step in the nodulation of legumes by Rhizobium spp. is the deformation and curling of root hairs. We have identified and cloned DNA sequences encoding this function from two strains of Rhizobium japonicum (USDA 122 and USDA 110) with a weakly homologous probe from Rhizobium meliloti. Root hair curling encoded by the cloned DNA fragments was examined on soybeans (Glycine soja ) after conjugative transfer of these sequences in broad-host-range vectors to various bacterial genera. Pseudomonas putida gave unambiguous expression of the root hair curling genes. This enabled us to identify the 8.7-kilobase EcoRI fragments encoding root hair curling from each strain. The phenotypes encoded by the plasmids pBS1 (derived from strain USDA 122) and pBS2 (derived from strain USDA 110) are distinct and represent a phenotype characteristic of their parent R. japonicum strains. Subclones of pBS1 and pBS2 were generated in single and multicopy vectors, and their expression was analyzed in P. putida. We established that a 4.2-kilobase internal Sa/I fragment of pBS1 and a 3.5-kilobase SstI -EcoRI fragment of pBS2 are sufficient to confer root hair curling on soybeans.