Paracrine control of gastric epithelial cell growth in culture by transforming growth factor-alpha.

Studying primary cultures of replicating canine oxyntic mucosal cells, we found evidence for modulation of cell growth by endogenous factors. [3H]thymidine incorporation into DNA was rapid with cells cultured in medium free of serum or added growth factors, and growth rates of these cultures were markedly dependent on plating density, indicating mitogenic control by soluble endogenous growth factors. Data indicated that endogenous transforming growth factor-alpha (TGF-alpha) exerted mitogenic control under the following conditions. 1) TGF-alpha was detected in the cultured cells by radioimmunoassay and immunohistochemistry. 2) TGF-alpha-like immunoreactivity and receptor reactivity were present in the medium in concentrations sufficient to exert mitogenic control. 3) Receptors for TGF-alpha and epidermal growth factor (EGF) were present in the cultures. 4) Immunoabsorption by a TGF-alpha-specific antisera reduced [3H]thymidine incorporation. TGF-alpha was localized to parietal cells by immunohistochemistry and cell separation. In contrast, combined [3H]thymidine autoradiography and immunohistochemistry with anti-TGF-alpha did not detect TGF-alpha in dividing cells. We conclude that parietal cell TGF-alpha exerts paracrine control of mucosal cell growth in vitro, and we speculate that this is an important paracrine mechanism in vivo.

Expression of neu protein, epidermal growth factor receptor, and transforming growth factor alpha in breast cancer. Correlation with clinicopathologic parameters.

The major objectives of this study were twofold: to determine 1) if growth factors or growth factor receptors were expressed similarly or differently in a clinically well-characterized group of breast cancer patients and 2) if these phenotypic characteristics were associated with any of the commonly used prognostic parameters. Formalin-fixed paraffin-embedded tumor tissue from 51 node-positive breast cancer patients were analyzed for the expression of neu, epidermal growth factor-receptor (EGF-R), and transforming growth factor alpha (TGF alpha) using immunoperoxidase staining. Positive membranous staining for neu was observed in 15 (29%) tumors. Over-expression of neu was observed in high-grade, estrogen-receptor-negative tumors (P less than 0.05). Epidermal growth factor receptor was expressed in 22 (43%) of the tumors analyzed and found to a greater degree in estrogen-receptor-negative and high-grade tumors (P less than 0.025). A significant correlation between neu and EGF-R expression was also noted. Tumors expressing membranous staining of neu had a greater than 70% chance of expressing EGF-R (P less than 0.01). Expression of TGF alpha was found in 68% of tumors and TGF alpha was detected in grade 1 and 2 tumor to a greater degree than EGF-R. The authors conclude that assaying tumors for these antigens may give additional phenotypic characteristics that can give further insight into the biology of breast cancer.

Immunological identification and distribution of parathyroid hormone-like protein polypeptides in normal and malignant tissues.

Monoclonal and polyclonal antibodies recognizing human parathyroid hormone-like protein (PTHLP) have been produced using a series of recombinant and synthetic PTHLP peptides. These antibodies have been used to develop a two-site immunometric enzyme immunoassay which detects PTHLP[1-87] and PTHLP[1-141] but not PTH. The immunoassay detected PTHLP in extracts of squamous carcinomas and normal tissues at concentrations from 7-515 ng PTHLP[1-87]/mg protein. Immunoblotting of the extract which showed the highest immunoreactivity, a squamous carcinoma of the lung from a patient with hypercalcemia, revealed a major band having an apparent mol wt of 26,500 and several other higher mol wt bands. Similar polypeptides were observed by immunoblotting cell extracts from a cell line, SCaBER, which secretes immunoreactive PTHLP into its medium and also from tumors in nude mice derived from this cell line. Chaotropic agents did not alter the immunoblotting pattern, and antibodies to three different epitopes of PTHLP recognized these bands, indicating PTHLP expression in the extracts. Immunohistochemical staining of normal human tissue with these antibodies revealed several PTHLP-containing tissues and confirmed the results of the immunoassay, suggesting a paracrine role for PTHLP. Staining was observed in several neoplastic tissues including squamous cell carcinomas, lung carcinoma, bladder carcinoma, osteogenic sarcoma, and adenocarcinoma of the colon.

Preparation and characterization of monoclonal antibodies specific for human transforming growth factor alpha.

A series of monoclonal antibodies (mAbs) against transforming growth factor alpha (TGF alpha) have been produced. The generation of these reagents, as well as their biochemical and immunochemical characterization is described. TGF alpha peptides, mutant recombinant TGF alpha proteins and two-site immunoradiometric assays were used to identify the epitopes recognized by each antibody. This approach has allowed the specific localization of immunodominant domains on the molecule. Certain mAbs were found to be useful for selected procedures. mAb 134A-2B3 was used for immunoblotting both the precursor and mature forms of TGF alpha from conditioned media of tumor cells. One mAb 189-2130.1, which reacted with the carboxyl terminal seventeen amino acids, was able to block TGF alpha binding to the EGF receptor. mAb 213-4.4 was used for immunohistochemical detection of TGF alpha in fixed tumor cells. mAbs 137-178 and 134A-2B3 were used to develop a two-site immunoradiometric immunoassay which was sensitive to 1 ng ml-1 and detected TGF alpha from a variety of tumor cells. A series of mAbs such as these could prove useful in studying the biochemical properties as well as the immunochemical localization of TGF alpha in normal tissues and tumors.

Synthesis of a gene encoding parathyroid hormone-like protein-(1-141): purification and biological characterization of the expressed protein.

PTH-like proteins (PTHLP), which are associated with humoral hypercalcemia of malignancy, have recently been purified. Isolation of their corresponding cDNAs has revealed that they are derived from a single gene. In this report a synthetic gene encoding PTHLP-(1-141), a 141-amino acid protein corresponding to the most abundant PTHLP cDNA detected in human tumors, was expressed in bacteria and purified to homogeneity. Recombinant (r) PTHLP-(1-141) migrates with an aberrantly high mol wt on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, presumably as a result of its unusually basic pI. rPTHLP-(1-141), like PTH, induced hypercalcemia in rats, caused release of 45Ca from fetal rat bones, and stimulated the synthesis of cAMP by rat osteosarcoma cells and canine renal membrane preparations. A comparison of the abilities of rPTHLP-(1-141) and bovine PTH-(1-34) to stimulate cAMP synthesis indicated rPTHLP-(1-141) to be 5-fold more potent in the osteosarcoma assay, while nearly 30-fold less active in the renal membrane adenylate cyclase assay. Although 100-fold less potent than bovine PTH-(1-34) in promoting bone resorption, rPTHLP-(1-141) was a potent calcemic factor in vivo, inducing a rise in serum calcium from 10.4 to 14.5 mg/dl when infused into rats at 1.3 micrograms/h. These results support previous assumptions that PTHLP is the humoral factor responsible for humoral hypercalcemia of malignancy. In addition, they suggest substantial differences between PTHLP and PTH in the regulation of calcium homeostasis.

The effect of fibronectin on the processing of C1q- and C3b/bi-coated immune complexes by peripheral blood monocytes.

In the past several years, it has been demonstrated that plasma fibronectin (Fn) binds to the C1q subunit of the complement system. The effect of Fn on the processing of immune complexes containing C1q and C3b by human peripheral blood monocytes was investigated. Preincubation of monocytes with Fn causes a significant increase in attachment of sheep erythrocytes coated with IgM and C1q (EIg-MC1q), but does not mediate their ingestion. EIg-MC1q attach to the Fn-treated monocytes via the C1q receptor because Fab anti-Fn antibodies do not inhibit their attachment to the monocytes. In addition, Fn-treated monocytes exhibit no change in C1q receptor number or affinity compared with monocytes treated with buffer. Fn mediates the phagocytosis of C3b/bi-coated particles, and C1q can enhance this process in two ways. First, phagocytosis of particles bearing C3b/bi and Fn is enhanced by the presence of C1q on the immune complex. Second, monocytes on Fn-coated surfaces ingest more particles if they are coated with both C3b/bi and C1q, compared with particles coated with either C3b/bi or C1q alone.

C1q, a subunit of the first component of complement, enhances binding of plasma fibronectin to bacteria.

The interaction of plasma fibronectin with C1q of the complement system has been demonstrated in the past several years. In addition, the antibody-independent binding of C1q to bacteria, as well as the binding of plasma fibronectin to bacteria, is well documented. This study examines whether the binding of C1q to bacteria enhances the interaction of C1q and bacteria with plasma fibronectin. Highly purified 125I-C1q bound to several species of bacteria in the absence of antibody. The binding of 125I-C1q to bacteria was saturable and specific since the addition of unlabeled C1q inhibited binding while the presence of bovine serum albumin did not. Bacteria which had been pretreated with either buffer or unlabeled C1q were tested for their ability to bind 125I-fibronectin. When bacteria were preincubated with buffer, Staphylococcus aureus bound fivefold more 125I-fibronectin than did Escherichia coli. However, preincubation of E. coli with C1q increased the binding of 125I-fibronectin by up to 20-fold, whereas pretreatment of S. aureus with C1q increased fibronectin binding by only twofold. These results were confirmed by immunoblotting studies which demonstrated the presence of C1q, as well as an increase in fibronectin antigens on the C1q-treated bacteria as compared with the level of fibronectin on buffer-treated bacteria. In addition, preincubation of 3H-labeled bacteria with C1q enhanced their attachment to fibronectin-coated surfaces but not to albumin-coated surfaces. The biological consequences of these observations are discussed.