Lack of expression of glutathione-S-transferase P, gamma-glutamyl transpeptidase, and alpha-fetoprotein messenger RNAs in liver tumors induced by peroxisome proliferators

Many structurally unrelated nonmutagenic peroxisome proliferators induce altered areas, neoplastic nodules, and hepatocellular carcinomas in rats. Unlike the lesions induced by genotoxic hepatocarcinogens, these lesions do not stain positively for the phenotypic markers gamma-glutamyl transpeptidase (GGT) and glutathione-S-transferase P (GST-P). To ascertain whether the absence of immunocytochemically detectable GST-P and GGT proteins in peroxisome proliferator-induced neoplastic lesions is due to the absence of specific mRNAs, we analyzed the total RNA isolated from hepatocellular carcinomas induced by three different peroxisome proliferators (ciprofibrate, Wy-14643, and BR-931) and the genotoxic carcinogens, 2-acetylaminofluorene and aflatoxin B1 (AFB), for the presence of GST-P, GGT, and alpha-fetoprotein (AFP) mRNAs. Northern and dot blot analysis of total RNA isolated from liver tumors induced by three different peroxisome proliferators revealed no detectable GST-P, GGT, and AFP mRNAs. GST-P mRNA was also not detected in a transplantable hepatocellular carcinoma established from a liver tumor induced by ciprofibrate. In contrast, GST-P mRNA levels were high in primary liver tumors induced by both 2-acetylaminofluorene and AFB and the two transplantable hepatocellular carcinomas established from such tumors. By immunoblot method, GST-P protein was found to be abundant in both primary and transplantable liver tumors induced by genotoxic carcinogens but not in those derived from peroxisome proliferator treatment. The GGT and AFP mRNAs were also not found in all 18 liver tumors induced by peroxisome proliferators that were analyzed and also in the ciprofibrate-derived transplantable liver tumor. The expression of GGT and AFP genes in liver tumors induced by 2-acetylaminofluorene and AFB was variable. These studies with peroxisome proliferators show that the GST-P and GGT gene derepression is not essential for the hepatocarcinogenesis or successful tumor transplantation. Further characterization of the molecular basis for the differential expression, particularly of the GST-P gene in liver tumors, may help identification of the critical event(s) in hepatocarcinogenesis by genotoxic carcinogens and nongenotoxic peroxisome proliferators.

Immunocytochemical localization of liver-specific proteins in pancreatic hepatocytes of rat

Hepatocytes are induced in the pancreas of rats maintained first on a copper-deficient diet for 8 weeks and then on normal rat chow. These cells are morphologically identical to parenchymal cells of the liver. These hepatocytes contain two liver-specific proteins: carbamyl phosphate synthetase I, a mitochondrial matrix protein that participates in the conversion of ammonia to carbamyl phosphate; and urate oxidase, an enzyme that catalyzes the oxidation of uric acid to allantoin. In addition, we also present evidence indicating that dietary administration of ciprofibrate induces peroxisomal beta-oxidation pathway enzymes, while the levels of catalase are unaltered in pancreatic hepatocytes. These observations along with the previously published results further establish the identity of pancreatic hepatocytes to parenchymal cells of liver and clearly indicate that transdifferentiation of pancreatic cells to hepatocytes is associated with activation of several liver-specific genes.

Immunohistochemical studies on local antitumor effects of streptococcal immunopotentiator, OK-432, in human solid malignant tumors

Immunocytochemical techniques were used to clarify the local inhibitory effects of a streptococcal immunopotentiator, OK-432, against solid malignant tumor growth. Natural killer (NK) cells and fibronectin were chosen as immunostaining markers to demonstrate the antitumor effects. Immunocytochemical staining was performed by the avidin-biotin-peroxidase complex method. These investigations demonstrated that (1) local administration of OK-432 seems to promote a marked induction of NK cells and fibroblasts around or entering into the cancerous lesions and (2) the cancer cell-killing effect of NK cells and the fibronectin-enriched stromal reaction augmented by the injection of OK-432 suggest at least the possibility of protection against neoplastic growth with invasion and the spread of distant or nodal metastases of solid carcinomas.

Immunocytochemical localization of urate oxidase, fatty acyl-CoA oxidase, and catalase in bovine kidney peroxisomes

We investigated the localization of urate oxidase, peroxisomal fatty acyl-CoA oxidase, and catalase in bovine kidney by immunoblot analysis and protein A-gold immunocytochemistry, using the respective polyclonal monospecific antibodies raised against the enzymes purified from rat liver. By immunoblot analysis, these three proteins were detected in bovine kidney and bovine liver homogenates. Subcellular localization of these three enzymes in kidney was ascertained by protein A-gold immunocytochemical staining of Lowicryl K4M-embedded tissue. Peroxisomes in bovine kidney cortical epithelium possessed crystalloid cores or nucleoids, which were found to be the exclusive sites of urate oxidase localization. The limiting membrane, the marginal plate, and the matrix of renal peroxisomes were negative for urate oxidase staining. In contrast, catalase and fatty acyl-CoA oxidase were found in the peroxisome matrix. These results demonstrate that, unlike rat kidney peroxisomes which lack urate oxidase, peroxisomes of bovine kidney contain this enzyme as well as peroxisomal fatty acyl-CoA oxidase.

Immunohistochemical studies on the localization of fibronectin in human thyroid neoplastic tissues

Fibronectin (FN) has been considered to be involved in the malignant transformation of cells. It was thus of interest to morphologically study the cell surface distribution of FN in various types of thyroid neoplasms employing immunohistochemical techniques. Immunohistochemical staining using the specific anti-human FN antibody was performed by the avidin-biotin-peroxidase complex procedure. Compared to the staining pattern of normal thyroid tissues (control), differences in staining in malignant thyroid lesions and even in fetal and embryonal adenomas were evident, often on the lateral and basal aspects of the cell membrane. Colloid adenomas, however, showed FN staining on the basement membrane, basically similar to the controls. Moreover, in thyroid carcinoma, different staining patterns relating to each of the histological types were observed. The present investigation demonstrates the heterogeneity of the thyroid neoplasms in terms of FN binding.

Tissue specificity and species differences in the distribution of urate oxidase in peroxisomes

The localization of urate oxidase in different tissues of rat and in the livers of selected mammalian species was investigated by immunoblot analysis and protein A-gold immunoelectron microscopy. Urate oxidase was purified from rat liver and used as an antigen to generate polyclonal antibodies in the rabbit. The antibodies were found to be monospecific by immunodiffusion and immunoblot analyses. By immunoblot analysis, urate oxidase was detected in the livers of rat, two strains of mice, hamster, dog, cat, and cow, but not in the Cynomolgus monkey and human liver. Urate oxidase was not detected by immunoblot method in rat kidney, jejunal mucosa, adrenal gland, testis, and pancreas. The subcellular localization of urate oxidase was ascertained by the protein A-gold immunocytochemical staining of the Lowicryl K4M embedded tissues. Urate oxidase was localized exclusively in the crystalloid core of the peroxisome in hepatic parenchymal cells of rat, mouse, hamster, dog, cat, and cow. The limiting membrane and the matrix of hepatic peroxisomes in these species were negative for the staining. The marginal plates of feline, canine, and bovine hepatic peroxisomes were also negative for urate oxidase. This enzyme was also not detected within the peroxisomes of human and monkey livers by the immunocytochemical technique. Peroxisomes (microperoxisomes) in extrahepatic rat tissues did not stain positively for urate oxidase by the protein A-gold immunocytochemical method, although they were positive for catalase. Fatty acyl-CoA oxidase was present in peroxisomes of jejunal mucosa, Leydig cells of test-is and pancreas but not in adrenal gland. Administration of a hepatic peroxisome proliferator, ciprofibrate or Wy-14643, failed to induce urate oxidase in rat liver. These results indicate that urate oxidase is a liver specific protein in rat and its localization within the liver peroxisomes of six mammals, excluding man and a nonhuman primate, and that its localization is limited exclusively to the crystalloid core. Unlike fatty acyl-CoA oxidase, urate oxidase does not appear to be inducible significantly by peroxisome proliferator treatment in the rat liver.

Absence of gamma-glutamyl transpeptidase activity in neoplastic lesions induced in the liver of male F-344 rats by di-(2-ethylhexyl)phthalate, a peroxisome proliferator

Male F-344 rats were fed a diet containing 2% di-(2-ethylhexyl)phthalate (DEHP) for 95 weeks. Liver nodules and/or hepatocellular carcinomas (HCC) developed in 6/10 rats fed DEHP and none were found in controls (P less than 0.005 by chi 2 test). All the nodules and HCC were negative for gamma-glutamyl transpeptidase. In the non-tumorous portions of liver, the hepatocytes contained an increased number of peroxisomes and extensive accumulation of lipofuscin. By immunocytochemical analysis, the liver peroxisomes in rats treated chronically with DEHP had visually detectable decrease in the H2O2-degrading catalase and increase in H2O2-producing fatty acyl-CoA oxidase. These results show that higher dietary level of DEHP, which causes substantially greater degree of peroxisome proliferation than the 1.2% dietary level used in the National Toxicology Program bioassay (1982, Publication no. NTP-80-37, Tech. Report Series No. 217), can induce liver tumors in male rats.

Purification, properties, and immunocytochemical localization of human liver peroxisomal enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase

A molecular understanding of genetic disease in which peroxisomal functions are impaired depends on analysis of the structure of normal and mutant enzymes of peroxisomes. We report experiments describing the isolation, characterization, and immunocytochemical localization of enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase bifunctional enzyme (PBE) of the peroxisomal fatty acid beta-oxidation system from normal human liver and compared it with that of rat liver enzyme. The human enzyme, purified approximately equal to 2300-fold by ion-exchange chromatography, is homogeneous as judged by NaDodSO4/PAGE. This PBE is localized exclusively in the matrix of peroxisomes in liver cells by the protein A/gold immunocytochemical method. The human PBE is similar to rat enzyme in size (Mr, approximately equal to 79,000), isoelectric point (pI, 9.8), pH optima, molecular structure as observed by rotary shadowing, and peptide pattern on NaDodSO4/PAGE after proteolytic digestion with Staphylococcus aureus V8 protease. The human and rat enzymes differed in their immunological properties by having partial identity with each other; this is reflected in their slightly dissimilar composition of the amino acids aspartic acid, threonine, glutamic acid, tyrosine, and glycine. COOH-terminal amino acid were similar for both the enzymes: -Gly-Ser-Leu-Ile-COOH. These results suggest that the human and rat liver PBE may be different in their amino acid sequences at their antigenic sites.

Immunocytochemical localization of D-amino acid oxidase in the central clear matrix of rat kidney peroxisomes

Light and electron microscopic localizations of D-amino acid oxidase (DAO) in rat kidney was investigated using immunoenzyme and protein A-gold techniques. The enzyme was purified from rat kidney homogenate and its antibody was raised in rabbits. By Ouchterlony double-diffusion analysis and immunoblot analysis with anti-(rat kidney DAO) immunoglobulin, the antibody was confirmed to be monospecific. The tissue sections (200 micron thick) of fixed rat kidney were embedded in Epon or Lowicryl K4M. Semi-thin sections were stained for DAO by the immunoenzyme technique after removal of epoxy resin for LM, and ultra-thin sections of Lowicryl-embedded material were labeled for DAO by the protein A-gold technique for EM. By LM, fine cytoplasmic granules of proximal tubule were stained exclusively. Among three segments of proximal tubules, and S2 and S3 segments were heavily stained but the S1 segment only weakly so. By EM, gold particles indicating the antigenic sites for DAO were exclusively confined to peroxisomes. Within peroxisomes, the gold particles were localized in the central clear matrix but not in the peripheral tubular substructures. The results indicate that D-amino acid oxidase in rat kidney is present exclusively in peroxisomes in the proximal tubule and that within peroxisomes it is found only in central clear matrix and not in the peripheral tubular substructures.

[Distribution pattern of fibronectin in human lung cancer tissue]

The distribution pattern of fibronectin in human lung cancer tissues has been studied using the indirect immunoperoxidase technique. In adenocarcinoma fibronectin is observed in the basal side of cancer cells and tumor stroma. In squamous cell carcinoma, it is seen in the basal side of cancer cells, surface of cancer cells and tumor stroma. The difference between adenocarcinoma and squamous cell carcinoma in invasion and metastasis was investigated to have some relation to the distribution pattern of fibronectin.

Purification of membrane polypeptides of rat liver peroxisomes

Peroxisomes were obtained by sucrose density gradient centrifugation from the livers of di(2-ethylhexyl)phthalate-fed rats, and the membranes were prepared by carbonate extraction (Fujiki, Y., Fowler, S., Shio, H., Hubbard, A.L., & Lazarow, P.B. (1982) J. Cell Biol. 93, 103-110). The integrated membrane polypeptides were solubilized with sodium dodecyl sulfate, and purified by repeated polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. Separation of 70 and 68 kDa polypeptides was not attempted in the present study because of their close migration in polyacrylamide gel electrophoresis. Other polypeptides with apparent molecular masses of 41, 27, 26, and 22 kDa were purified to near homogeneity. Antibodies were raised against these purified preparations. The 68 kDa polypeptide is suggested to be produced by the proteolytic modification of 70 kDa polypeptide, since the former increased concomitantly with decrease of the latter when the liver homogenate was incubated, and this change was prevented in the presence of leupeptin during the incubation. The 41 kDa polypeptide was a minor component. The 70 and 68 kDa polypeptides and 41 kDa polypeptide and their antibodies were cross-reactive, but the relation of these polypeptides was not clear. The 27 and 26 kDa polypeptides seemed to be another species of membrane polypeptides, although the relationship of these two polypeptides remains to be clarified. The 22 kDa polypeptide is not related to other membrane polypeptides. The results of immunoblot analysis of subcellular fractions of the liver and an electron microscopic immunocytochemical study to locate the antigenic sites with protein A-gold complex suggest that all of these polypeptides are localized on peroxisomal membranes. On proliferation of rat liver peroxisomes by administration of di(2-ethylhexyl)phthalate, a peroxisome proliferator, all of these polypeptides were markedly increased.

Clinical application of the enzyme immunoassay for pancreatic oncofetal antigen

Pancreatic oncofetal antigen (POA) purified by us has been detected in sera of patients with carcinoma of the pancreas by the micro-Ouchterlony method. In an attempt to improve the sensitivity and clinical usefulness of the serum POA assay, we established an enzyme immunoassay for POA and reported the results with this method. In this study, we investigated serum POA levels in pancreatic cancer and other diseases. The tissue localization of this POA in the pancreas was also studied. For the establishment of the enzyme immunoassay, an anti-POA-F(ab')2 fragment prepared from absorbed antiserum was conjugated with beta-D-galactosidase. The solid-phase "sandwich" principle was used. The normal upper limit of the serum POA level was defined as 400 units/ml. Among 60 patients with pancreatic cancer, 44 had elevated levels (73.3%). Of 22 cases with chronic pancreatitis, 4 had elevated levels (18.2%). In malignant diseases other than pancreatic cancer, elevated levels of serum POA were seen in 17.6% to 48.5% of the patients, most of whom had only slightly elevated levels. These results indicate that enzyme immunoassay for POA is clinically useful for making a diagnosis of pancreatic cancer. Immunoperoxidase staining showed POA to be found at the apical surface of ductular cells in fetal pancreas, at the luminal surface of glandular structures in pancreatic cancer tissue, and also at the luminal surface of the small duct in normal pancreas. Thus it is suggested that a high level of serum POA in patients with pancreatic cancer is derived from pancreatic cancer tissue.

Immunocytochemical demonstration of retinol-binding protein in the lysosomes of the proximal tubules of the human kidney

The localization of plasma retinol-binding protein (RBP) in the human kidney was determined by two immunocytochemical techniques, the PAP method and the protein A-gold technique. By using the affinity purified antibody against RBP obtained from the urine of the patients with cadmium poisoning (Itai-Itai disease), the immunoreactive substances were located by light microscopy in the proximal tubules of the human kidney. By immuno-electron microscopy, the stained organelles were identified as lysosomes in both S1 and S2 segments of the tubules. These data suggested that the reabsorption of low molecular weight plasma proteins like RBP can occur in the two segments. We inferred a similarity between the S1 and S2 segments concerning the reabsorption of RBP.