Search

Type IV collagen trimer is a major component of basement membranes (BMs). It is composed of polypeptides named alpha1(IV)-alpha6(IV) chains. Chains alpha1,2(IV) are widely expressed in BMs while alpha3(IV)-alpha6(IV) are more restricted in human tissues. We have now studied by immunohistochemical means the distribution of collagen IV chains in fetal and adult human kidney, in oncocytomas, in renal cell carcinomas (RCCs) and their metastases and in experimental xenografts of human tumors. alpha1,2(IV) chains were found in all BMs of fetal and adult kidney as well as of renal tumors, while alpha3(IV)-alpha6(IV) chains were found in BMs of distal segments of developing and mature tubules. alpha3(IV)-alpha5(IV) chains were seen also in BMs of developing fetal glomeruli after the capillary loop stage. Most of the RCCs and their metastases showed occasional expression of alpha3(IV)-alpha6(IV) with papillary variants showing only expression of alpha5(IV) chain. There was a distinct expression of alpha3(IV)-alpha5(IV) chains in BMs of 3 oncocytomas. In 2 of them a variable expression of the alpha6(IV) chain was seen. In 3 of 4 xenografts, immunoreactivity for human-specific monoclonal antibody (MAb) for alpha1,2(IV) was seen in the BM-like structures. No alpha3-alpha6(IV) was seen in any of the xenografts, while polyclonal antiserum for type IV collagen presented immunoreactivity in BMs of all xenografts. Our results show that oncocytomas and most of the RCCs express scarce variants of type IV collagen containing alpha3(IV)-alpha6(IV) chains. In experimental xenograft tumors, both implanted RCC cells and host stromal cells have a capacity to produce type IV collagen.

We studied the expression of laminin (Ln) chains (alpha1-alpha3, beta1-beta3, gamma1) in human renal-cell carcinomas (RCC), papillary renal neoplasms (PRN) and oncocytomas, in RCC cell lines and their xenografts. In RCCs the basement membranes (BM) showed immunoreactivity for chains of Ln-1 (alpha1-beta1-gamma1). Only in well-differentiated RCCs could vessel BMs be distinguished from those of carcinoma cell islets. RCCs and oncocytomas also exhibited an abundant immunoreactivity for Ln beta2 chain in both vessel and tumor cell BMs, while Ln alpha2 chain was not seen in any renal tumors. In distinction from RCCs, PRNs presented a strong BM immunoreactivity for Ln alpha3 and beta3 chains and for Ln-5, as well as lack of Ln beta2 chain. A more variable reactivity for Ln-5 was seen in oncocytomas. As PRNs and oncocytomas have been suggested to originate from collecting ducts, it is notable that in normal human kidney, we could detect immunoreactivity for Ln-5 and its chains only in BM of the tubules of the loop of Henle. In immunoprecipitation experiments, an abundant production of Ln-1, but not of Ln-5, was seen in cultured RCC cells, while in xenografts of the same cells BM-confined immunoreactivity for both Ln-1 and Ln-5 was seen. Ln beta2 chain was produced by 2 of the 4 RCC cell lines in culture but was found only in 1 of the xenografted tumors.

E- and N-cadherins are proteins involved in intercellular adhesion and are localized, e.g., in the adherens junctions of epithelial cells. Kidney tubules express these molecules in a distinctive pattern, the expression of N-cadherin being restricted to proximal tubules and that of E-cadherin to distal tubules and collecting ducts. Renal cell carcinomas (RCCs) and oncocytomas are considered to originate from these tubular epithelia. To find out whether cadherins could serve as markers for a cellular origin of these tumors, we studied the expression of E- and N-cadherins in RCCs and oncocytomas, in cell lines derived from RCCs as well as in tumors grown in nude mice. Most RCCs co-expressed E- and N-cadherins, as did 2 of the 4 cell lines studied. The expression pattern did not correlate with the histological grade of the tumors, and even the least differentiated tumors, as well as metastases, showed expression of cadherins. Renal oncocytomas expressed E-cadherin but not N-cadherin, which is in line with previous studies that have proposed a collecting duct origin for these tumors. Papillary renal neoplasms, a separate entity usually not classified as RCC, expressed neither of the cadherins studied despite the abundant expression of beta-catenin. Our results suggest that most RCCs co-express the characteristic adhesion molecules of both proximal and distal tubules, which makes it questionable whether the origin of these tumors can be reliably located to any distinct part of the renal tubule. Our results also suggest that in RCCs the increased histological grade is not directly associated with changes in the expression of either of the cadherins, indicating other mechanisms underlying the deficient capacity to form polarized tubular structures.

We studied the expression of tenascin (Tn) and isoforms of fibronectin (Fn) in human renal cell carcinomas (RCC) and oncocytomas, in RCC cell lines and in their s.c. implanted xenografts in nude mice. In well-differentiated RCCs and oncocytomas extra-domain A (EDA)-Fn and Tn immunoreactivities were confined to the basement membranes and blood vessels, while in less-differentiated RCCs they were also widely seen in the stroma, correlating with the morphological differentiation of the tumor. Expression of EDB-Fn and oncofetal (onc)-Fn was very scarce in most of the RCCs and oncocytomas. Western blotting results demonstrated the predominance of the M(r) 190,000 Tn subunit in most RCCs. Among the 4 RCC cell lines, 3 showed Tn in the extracellular matrix. As xenografts, they formed moderately or poorly differentiated tumors, with abundant Tn. Three of the RCC cell lines also showed secretion of EDA-Fn and 2 of them secretion of onc-Fn and EDB-Fn into the culture medium, while in xenografts there was a strong expression of all Fn isoforms. In xenografts, the expression of Tn closely recapitulated that seen in clinical tumors and in the cell lines in vitro, while the expression of Fn isoforms in cultured cells and their xenografts was highly discordant.

The expression of 17 beta-hydroxysteroid dehydrogenase (17-HSD) enzyme protein was studied in benign and malignant human breast tissue using the time-resolved immunofluorometric assay (IFMA), immunoblotting and immunohistochemistry. The presence and distribution of estrogen and progestin receptors was also analyzed immunohistochemically. Cytosolic 17-HSD concentrations in malignant breast specimens were highly variable (less than or equal to 0.2-311 ng/mg protein). As was previously found for the placental enzyme, the molecular weight of the 17-HSD expressed in malignant breast tissue was 35 kDa, estimated following polyacrylamide gel electrophoresis and immunoblotting. The cellular distribution of 17-HSD was further studied by immunohistochemistry. Immunostaining for 17-HSD was observed in 71% of the benign breast lesions (fibroadenomas and cases of mastopathia chronica) and in 47% of the cancer specimens (intra-ductal carcinomas, invasive ductal carcinomas). In benign lesions, the staining was exclusively localized in the cytoplasm of epithelial cells, with no immunoreactivity in the stromal cells. The staining in the cancer specimens was also detected only in the cytoplasm of malignant epithelial cells. A strong or moderate expression of 17-HSD was related to the presence of PR in the specimen (chi 2 = 4.657, p = 0.031). However, the expression of PR was not a prerequisite for expression of 17-HSD in all the cancer specimens. Our data suggest that, in addition to the reported regulation of 17-HSD by progestins, other factors are also involved in this process in breast tissue.

Transferrin is required for proliferation of most cells in culture. This effect is presumably mediated by the binding of transferrin to its receptor, a surface glycoprotein which is preferentially expressed by actively growing cells. Here we show that normal human fibroblasts cultured in serum, and other media containing transferrin express transferrin receptors in a distinctly non-random way; punctate foci of the receptor were seen only at the leading lamellae of the cells, whereas cells grown without serum, or in transferrin-depleted serum showed a random distribution of the receptor. In contrast, malignant fibrosarcoma cells showed the receptor uniformly throughout the cell surface in all media tested, including those containing transferrin. The data suggest that the ligand causes a directional lateral movement of the receptor in normal but not in malignant cells. Application of the receptor antibody caused a rapid internalization of the receptor in both cell types.

Immunostaining of normal human fibroblasts with a monoclonal antibody (MAb) (V22AC12) revealed typical cytoplasmic arrays of vimentin filaments in both mitotic and interphase cells. In human A8387 fibrosarcoma cells and SV40-virus-transformed human fibroblasts, the same antibody showed positivity only in mitotic cells and in interphase cells only after treatment of the fixed cells with alkaline phosphatase. Upon immunoblotting with the MAb, an Mr 57,000 vimentin polypeptide was seen in normal fibroblasts. In fibrosarcoma cells the same polypeptide was revealed by this antibody only after treatment with alkaline phosphatase. The Mr 57,000 vimentin polypeptide was a major cytoskeletal protein in both fibroblasts and fibrosarcoma cells. Inclusion of Ca2+ into the cytoskeleton extraction medium brought about a somewhat increased degradation of vimentin in fibroblasts. In fibrosarcoma cells, such treatment caused a quantitative disappearance of the Mr 57,000 protein with a concomitant appearance of 3 distinct, low-molecular-weight degradation products in the detergent-soluble fraction. Another Ca2+-induced change in the polypeptide profile of fibrosarcoma cells was the disappearance of the Mr 240,000 non-erythroid alpha-spectrin and the concomitant appearance of a prominent Mr 140,000 degradation product. Inclusion of proteolysis inhibitors in the Ca2+-supplemented extraction medium inhibited degradation of both vimentin and alpha-spectrin polypeptides. The results suggest differences in the composition of the cytoskeletons of normal fibroblasts and fibrosarcoma cells, manifested in the differential Ca2+-susceptibility of vimentin and non-erythroid alpha-spectrin. Results with MAb V22AC12 suggest that differential phosphorylation of vimentin could account for at least part of this difference.

Monospecific antibodies and indirect immunofluorescence microscopy were used to investigate the presence of cytoskeletal intermediate filaments of the keratin, vimentin and desmin types in 43 soft-tissue sarcomas. The results showed that vimentin was present in the neoplastic cells of all types of soft-tissue sarcomas studied, whereas keratin was absent, the only exception being the epithelial-like cells in biphasic synovial sarcoma. The presence of desmin was confined to leiomyosarcomas and rhabdomyosarcomas, which usually showed desminpositivity. In addition, malignant fibrous histiocytomas occasionally displayed some desmin-positive cells. Thus, determination of intermediate filaments is of aid in the histogenetical diagnosis of tumors, because keratin-positivity largely excludes the possibility of a sarcoma, and desmin-positivity aids in recognizing leiomyosarcomas and rhabdomyosarcomas.

Affinity of iodinated fibronectin (Fn) and its defined proteolytic fragments to electrophoretically separated polypeptides of normal and malignant cells was studied in an overlay assay. Cellular 125I-Fn and a major 125I-Fn fragment (Mr 120,000-140,000), containing the cell-binding site, revealed in fibroblasts Mr 170,000, Mr 140,000, and Mr 47,000 Fn-binding polypeptides of which the first two could also be found in the plasma membrane preparations. Binding of 125I-Fns to Mr 170,000 and Mr 140,000 polypeptides was inhibited by the synthetic peptide Arg-Gly-Asp-Ser and to all 3 polypeptides by Fns and Mr 120,000-140,000 fibronectin fragment. Both fibrosarcoma cells and SV40-virus-transformed fibroblasts appeared to lack the Mr 140,000 Fn-binding polypeptide. Binding was similar when Fn from normal fibroblasts or fibrosarcoma cells was used in the assay, while plasma 125I-Fn had weaker affinity towards the Mr 140,000 polypeptide. Instead, proteolytic Fn-fragments, lacking the cell binding site, did not bind to any proteins in the assay. Radioactive cell-surface labelling showed differences in the corresponding surface polypeptide profiles of normal and malignant cells. The results suggest that the failure of pericellular matrix deposition in malignant cells could be due to either defective surface exposition or defective binding property of the Fn-receptor-like polypeptides.

Intermediate filaments (IF) are ubiquitous cytoplasmic structures which, by virtue of their cell- and tissue-type-specific characteristics, are widely used as markers of tissue derivation and as differential diagnostic aids in surgical pathology. In contradistinction to other IFs, vimentin filaments, characteristic of mesenchymal cells, may be co-expressed with other cell-type specific IFs--cytokeratin filaments, desmin filaments, glial filaments and neurofilaments--in some tumor cells, embryonic cells, and cells in vitro. In this study we describe a novel type of IF co-expression which does not involve vimentin-filaments, viz. the presence of both cytokeratin filaments and neurofilaments in human bronchial carcinoid tumor cells.