Differences in nucleolar antigens of rat liver and Novikoff hepatoma ascites cells.

Antisera to nucleoli of Novikoff hepatoma ascites and normal rat liver cells were produced in rabbits by injection of whole, isolated nucleoli. These antisera have been used to compare the nucleolar antigens that were partially fractionated by differential solubilization from nucleoli. Fourteen antigens were detected by these antisera; ten of these antigens were detected by both antisera. Ouchterlony double diffusion analysis of soluble extracts from normal rat liver and Novikoff hepatoma ascites nucleoli and fetal rat liver nuclei provided evidence for antigens found only in liver extracts, only in tumor extracts, or only in tumor and fetal extracts. Antisera preabsorbed to remove antibodies to common antigens of liver and tumor provided confirmatory evidence for one nucleolar antigen in liver that was not found in tumor or fetal rat liver, one antigen in tumor that was not found in adult or fetal rat liver, and three antigens in both tumor and fetal rat liver that were not found in adult rat liver. In addition, the antitumor nucleolar antiserum preabsorbed with liver nuclear extracts still produced positive nucleolar fluorescence in Novikoff hepatoma ascites cells but not in liver cells. Conversely, anti-liver nucleolar antiserum preabsorbed with tumor nucleolar extracts did not produce detectable tumor nucleolar fluorescence but did produce positive fluorescence in liver nucleoli.     

Nucleolar phosphoproteins of normal rat liver and Novikoff hepatoma ascites cells.

The nucleolar acid-soluble proteins from normal rat liver and Novikoff hepatoma ascites cells were labeled in vivo for 2 hr after the injection of [32P]orthophosphate and in vitro with [gamma-32P]adenosine triphosphate in systems containing 0.25 M sucrose, 5 mM MgCl2, 12.5 mM NaCl, and 0.05 M Tris-HCl buffer, pH 7.3, AT 37 DEGREES. Two-dimensional polyacrylamide gel electrophoresis and autoradiography showed that approximately 40 and 20 protein spots were labeled in vivo and in vitro, respectively. There were some marked differences in labeling in vivo between normal rat liver and Novikoff hepatoma acid-soluble nucleolar proteins. By 32P analysis of gel slices, the proportion of the total 32P incorporated into protein Spot A1-4 was greater in normal liver, and the proportion of 32P incorporated into some high-molecular-weight protein spots, such C23-24 and C26-27, was greater in the Novikoff hepatoma ascites cells. With the in vitro incubation system, the 32P uptake per mg protein was about twice as high in Novikoff hepatoma nucleolar proteins as in normal rat liver nucleolar proteins but, generally, the same proteins were labeled in both tissues.     

Comparison of gene expression in preneoplastic and neoplastic rat liver to adult, fetal, regenerating, and tumor-promoted liver

Computer-assisted analysis was performed on the in vitro translation products of polyadenylated RNA samples isolated from normal adult Fischer rat liver and from preneoplastic and neoplastic rat liver samples which were generated by the Solt Farber technique (Solt, D. and Farber, E., Nature 263:701-703, 1976). The vast majority of the differences in translation products observed throughout the progressive development of hepatocellular carcinoma was quantitative in nature. Importantly, this quantitative heterogeneity first became prevalent at the very early preneoplastic stage of hepatoma formation. Only 3 consistent qualitative alterations in translation products were observed to be associated with the hepatocarcinogenesis process. The appearance of two new polypeptides of molecular weight and isoelectric point of 32/5.2 and 43/5.1 appeared to be related to an early preneoplastic event in hepatoma development and the transition from a preneoplastic to a neoplastic state, respectively. Importantly, these new polypeptides were not observed in the in vitro translation products generated from fetal or regenerating liver samples or from liver samples which were chronically treated with phenobarbital or terachlorodibenzo-p-dioxin. One translation product (located at 35/6.6) of normal adult, fetal, and regenerating liver RNA samples was undetected in all preneoplastic, neoplastic, phenobarbital-, and terachlorodibenzo-p-dioxin-treated liver RNA translation products. The possibility exists that the specific loss of this gene product may promote the development of the transformed phenotype.     

Homology between rat liver RNA populations during development, regeneration, and neoplasia

To investigate the degree of homology which may exist between rat liver RNA populations during development, regeneration, and neoplasia, we hybridized polyadenylated RNAs from (a) normal adult, (b) 24-h regenerating, (c) 20-day fetal livers, and (d) the transplantable Morris hepatoma 5123tc to homologous and heterologous complementary DNAs and to cDNAs enriched for sequences preferentially transcribed in either adult or fetal liver. We also compared the in vitro translation products of these RNAs. Analyses of normal adult, regenerating, and fetal liver RNA populations and their translation products show that the overall pattern of gene expression during liver regeneration differs little from that of normal adult rat liver and that mature hepatocytes do not appear to revert to an "immature" state upon reentering the cell cycle. Comparisons between fetal, normal adult, and tumor RNA populations revealed that RNA populations from fetal liver and the 5123tc tumor lack sequences normally expressed in the mature adult liver. However, the tumor does not "reexpress" sequences which are preferentially expressed in fetal livers.     

Comparison of nuclear nonhistone phosphoproteins of rat liver and Novikoff hepatoma.

As part of a continuing comparison of nuclear proteins of tumors and other tissues, 32P-labeled nuclear proteins were extracted successively with 0.15 and 0.35 m NaC1 from the nuclei of normal, regenerating, and thioacetamide-treated rat liver as well as Novikoff hepatoma 3 hr after injection of 32Pi into rats. Separation of proteins of these fractions with aqueous phenol was carried out before two-dimensional electrophoresis on polyacrylamide gels. By autoradiography many common spots were found, but four 32P-labeled protein spots, CU', C13p, C21p, and CMp, were found in the Novikoff hepatoma and not in the various liver samples studied. Two spots, B6 and B10, were found in the liver patterns and not in the tumor. Sot B33 was very dense in regenerating liver but was only a faint spot in thioacetamide-treated liver. The greater density of Spots CU', C13p, C21p, and CMp in the tumor patterns is consistent with the increased density reported earlier for spots of the C-region of a variety of tumors.     

Chromatin-associated glycoproteins of normal rat liver and Novikoff hepatoma ascites cells.

Glycoproteins were demonstrated in the 0.6 M NaCl extract of chromatin of normal liver and Novikoff hepatoma cells by periodic acid-Schiff staining of sodium dodecyl sulfate-polyacrylamide gels. In chromatin extracts of Novikoff hepatoma cells, six major periodic acid-Schiff-positive bands coincident with Coomassie Blue-stained protein bands were found with molecular weights of 30,000, 54,000, 64,000, 75,000, 104,000, and 127,000. A corresponding extract from normal rat liver chromatin revealed the presence of four major periodic acid-Schiff-positive bands that migrated with protein bands at molecular weights of 16,000, 30,000, 54,000, and 75,000. The glycoproteins of these extracts contained either mannose or alpha-D-glucose, fucose, and N-acetylglucosamine as shown by the specificity of lectins in affinoelectrophoresis. One of the glycoproteins detected by affinoelectrophoresis was very basic.     

Differential transglutaminase distribution in normal rat liver and rat hepatoma.

Distribution of transglutaminase activity was determined in normal rat liver, a 3'-methyl-4-dimethylaminoazobenzene-induced primary hepatoma, and the Novikoff hepatoma. Over 90% of the total enzyme activity was found in the 105,000 X g supernatant of normal liver, whereas only 30% was found in this fraction of the hepatomas, the remainder being found in the particulate fraction. The is distribution pattern did not correlate with protein distribution nor did it change during cellular proliferation, since regenerating liver and embryonic tissue had the same pattern as normal liver. Cell protein was a suitable acceptor substrate for the enzyme. Kinetic analyses showed that liver and hepatoma enzymes had a similar Km and Vmax for putrescine incorporation into cell protein. Hepatoma particulate enzyme was more stable than either liver or hepatoma supernatant enzyme. The enzyme may also act as the acceptor molecule.     

Silver staining of nucleolar granules in tumor cells.

With the aid of a simple silver-staining procedure, large numbers and unusual arrays of nucleolar argyrophilic granules were found in Novikoff hepatoma, KB, and HeLa cells. Some of these arrays consisted of linearly arranged discrete granules, and others were in two to three rows each containing three to five granules. Corresponding formations were not found in either the normal or regenerating liver nucleoli which contained an argyrophilic network in which the dark granules were apparently associated with the less dark argyrophilic fibrils of a reticulum. The nucleolar argyrophilic granules were readily identifiable in the separated daughter nuclei of the tumor cells in telophase, suggesting that the increased nucleolar activity of the G1 phase begins in these cells even before cell division has been completed.     

mRNA levels for human nucleolar protein P120 in tumor and nontumor cells

A monoclonal antibody to a human tumor nucleolar 120 kD protein was developed by Freeman et al. (Cancer Res. 48: 1244-1251, 1988). Its complementary DNA (cDNA) has been isolated and sequenced (Fonagy et al., submitted). To determine the relative messenger RNA (mRNA) level for protein p120, cellular mRNA was extracted, slot-blotted onto nitrocellulose filters, and hybridized to radioactive p120 cDNA fragments. Human tumor cells contained 15-60 times more p120 mRNA than human term placenta. The rat Novikoff hepatoma ascites cell mRNA hybridized to the p120 cDNA probes, but the p120 monoclonal antibody did not react with the Novikoff hepatoma proteins. Novikoff hepatoma mRNA contained 8 times as much p120 mRNA as normal rat liver. As a control, a cDNA was used for protein B23, an abundant nucleolar protein; there were 3.5, 29, and 14 times more B23 mRNA than p120 mRNA in normal rat liver, Novikoff hepatoma ascites cells, and HeLa cells, respectively. Whereas the increased levels of the mRNA and protein B23 reflect increased activity of the nucleolus for any increment of nucleolar function, the increased levels of p120 mRNA and the p120 protein reflect the activity of the G1 phase of the cell cycle. The elevated level of p120 mRNA in tumors may reflect the heightened G1 cascade in transformed cells.     

A comparison of polysomal messenger ribonucleoprotein particles from normal and neoplastic rat liver.

Free polysomes were isolated from normal and regenerating rat liver and from Morris hepatomas 7777, 7800, 5123C and 9618A. Sucrose gradient analysis ruled out the possibility of any significant messenger RNA degradation in these polysome preparations. The ethylenediaminetetraacetate-disrupted polysomes were fractionated on oligodeoxythymidylic acid-cellulose columns. The column-bound polyriboadenylic acid-containing messenger ribonucleoprotein particles were eluted with a formamide buffer, precipitated with ethanol, and subjected to polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. The messenger RNA-associated proteins from the different tissues were qualitatively similar, but two proteins with molecular weights of 66,000 and 109,000 found as minor proteins in normal liver appeared on gels as major protein bands when hepatoma messenger ribonucleoprotein particles were examined. The 66,000- and 109,000-molecular-weight proteins in these particles from regenerating liver appeared quantitatively similar to those isolated from normal liver.     

Comparison of thioredoxin reductases from Novikoff ascites hepatoma cells and normal liver of rats.

Adult rat liver contained variant forms of thioredoxin reductase with isoelectric points at pH 4.9 and at approximately pH 4.7 compared to pH 5.1 for the enzyme from Novikoff ascites hepatoma. Fetal and regenerating liver contained only the form with the isoelectric point at pH 4.9. All three enzymes precipitated with and were inhibited by a rabbit antibody to purified enzyme from Novikoff tumor.     

Cytokeratin and nonhistone protein antigenic changes in rat liver during azo dye but not hepatotoxin feeding

Chromosomal proteins from rat liver have been assessed immunologicallyfor changes in specific cytokeratins or primary tumor nonhistoneproteins during treatment with an azo dye hepatocarcinogen (3'-methyl-4-dimethylaminoazobenzene),(3'-Me-DAB) or the hepatotoxin a-naphthyl-iso-thiocyanate (-NIT).Fisher rats were fed laboratory diets supplemented with eitheragent and sacrificed sequentially at various intervals. Chromosomalproteins from these livers were electro-phoresed in the presenceof sodium dodecyl sulfate, transferred to nitrocellulose sheetsand reacted with rabbit antisera to Novikoff hepatoma cytokeratinor 3'-Me-DAB induced primary hepatoma dehistonized chroma tin.Livers from carcinogen-fed animals exhibited distinct, sequentialchanges in antigenic nonhistone proteins until the immunologicalspecificity characteristic of the hepatoma was achieved concomitantwith the induction of neoplasia. No antigenic changes were observedto occur in hepatotoxin-fed animals. The rat carcinoma-specificcytokeratin antigens p39 and p49 in Novikoff hepatoma were observedto appear as early as three weeks after the start of carcinogenfeeding and were present maximally in 23 week livers with insitu carcinoma. These cytokeratins were not detected in -NIT-fedanimals. Our results support the concept that the carcinogenicprocess can be related to temporal changes in the expressionof cell-specific cytokeratins in addition to nonhistone chromosomalproteins. Furthermore, these data suggest the expression ofthese antigenic species to not be the direct result of changesin liver structure and cellular composition associated withcarcinogen toxicity; rather, neoplastic transformation is apparentlyrequired.     

Nuclear protein phosphokinases in normal and neoplastic tissues.

Protein phosphokinases were isolated from the nuclei of normal and fetal liver and neoplastic tissues. Chromatography on phosphocellulose columns resolved the normal and fetal liver kinases into five reproducible fractions. Each of the fractions differed in optimal divalent cation and substrate requirements. Hepatic proliferation was accompanied by quantitative changes in the kinase activity profiles (with endogenous phosphoprotein as natural substrate). An additional phosphoprotein kinase activity stimulated by Mn2+ was found in the nuclei of malignant cells. This tumor-specific kinase could not be detected either in tumor cytoplasm or in fetal or regenerating liver nuclei. Mn2+-dependent phosphoprotein kinase from Novikoff hepatoma phosphorylated only one major protein band detectable by polyacrylamide gel electrophoresis. This substrate could not be detected in chromatin of normal tissues.     

Propionibacterium acnes-mediated humoral immune responses to tumor-specific antigens on rat liver cells transformed in vitro by chemical carcinogens.

Effects of Propionibacterium acnes on production of antibodies against tumor-specific membrane antigens were investigated in syngeneic inbred BD IV and BD VI rats. BD rat liver cell lines transformed in vitro by chemical carcinogens were used as target cells for tumor-specific antigens. By membrane immunofluorescence, antibodies against these rat liver cell lines were detected in syngeneic BD rat sera. Antibodies were produced in syngeneic rats under the adjuvant effect of heat-killed P. acnes only. In assays with various target cells and absorption experiments, the antibodies reacted with a tumor-specific individual antigen or tumor-specific cross-reacting antigen on the surfaces of transformed BD rat liver cells. No antibodies against these antigens were found in the sera obtained from syngeneic rats immunized with either the transformed cell lines or P. acnes but not with both. Freund's complete adjuvant did not induce antibodies against these tumor-specific antigens.     

A mutation in subunit B of the DNA polymerase alpha-primase complex from Novikoff hepatoma cells concomitant with a conformational change and abnormal catalytic properties of the DNA polymerase alpha-primase complex

Mutated constituents of the DNA replication complex might contribute to the mutational load of the genome during tumor development by impairing DNA synthesis as well as cell cycle-related control of DNA replication. To prove or disprove this hypothesis, we looked for mutations in the cDNA sequences of the four subunits of DNA polymerase alpha-primase from both highly malignant Novikoff hepatoma cells and regenerating normal rat liver and compared physicochemical and catalytic properties of the DNA polymerase alpha-primase complexes purified from both sources. Sequence analysis showed two mutations in subunit B from Novikoff cells: one in nucleotide position 855 (CCG-->CCA) that did not result in an amino acid exchange and one in position 862 (GTG-->ATG) that caused a change of valine to methionine in codon 288. No mutation was found in the three other subunits. The wild-type and mutated sequences of subunit B were cloned and expressed in vitro. Sedimentation analysis of the expressed polypeptides revealed different sedimentation constants, indicating that the amino acid exchange affected the conformation of subunit B. The analysis of the purified DNA polymerase alpha-primase complexes showed a sedimentation value that was significantly higher for the enzyme complex from normal liver than for that from Novikoff cells. In addition, DNA polymerase alpha-primase complexes from Novikoff cells showed higher sensitivity to camptothecin, topotecan, and structurally related compounds (such as (R,S)-7-ethyl-10-hydroxy camptothecin, 9-aminocamptothecin, and 10-hydroxycamptothecin) than the enzyme from normal rat liver. Thus, the amino acid change found in subunit B appears to result in a conformational change of the DNA polymerase alpha-primase complex from Novikoff hepatoma cells. Whether this mutation influences genetic instability or tumor development needs to be explored.     

Monoclonal antibodies reacting with normal rat liver cells as probes in hepatocarcinogenesis

A series of four monoclonal antibodies was raised against suspensions of normal adult WAB/Not rat hepatocytes. An immunoperoxidase-staining technique was used to examine the distribution of determinants detected by these antibodies on frozen sections of fetal, neonatal, adult, and regenerating liver and on a range of 4-dimethylaminoazobenzene-induced liver lesions, including a panel of 32 primary liver carcinomas. Three of the antibodies were directed against hepatocytes, while a fourth antibody stained stromal elements within the liver. The determinants detected by the anti-hepatocyte monoclonal antibodies arose in a specific sequence during normal liver development and, when assessed in conjunction, characterized several phenotypes associated with stages in normal hepatocyte differentiation. These same antibody-defined phenotypes were expressed by the primary liver carcinomas, and the distribution of phenotypes among the tumors revealed a heterogeneity which was not evident from a conventional morphological classification. Primary liver tumors expressed a total of four antibody-defined phenotypes, whereas gamma-glutamyl transpeptidase-positive foci of hepatocytes and neoplastic nodules expressed, respectively, only one or 2 antibody-defined phenotypes. We suggest that monoclonal antibodies directed against normal liver cell components may provide a means to establish lineage relationships between cell populations involved in hepatocarcinogenesis.     

A comparative study of the proteins of rat plasma, liver and hepatoma by agarose immunoelectrophoresis

A convenient and selective microtechnique, agarose immunoelectrophoresis, was applied to a comparison of the antigens in rat liver and plasma, and in primary hepatoma induced in male Fischer strain rats with N-2-fluorenylacetamide or N-hydroxy-N-2-fluorenylacetamide. Of the many soluble proteins antigenic in this system, 3 were singled out for detailed study. The albumin in liver and hepatoma had a higher mobility than that in plasma. Extraction of the soluble fraction of rat liver with ether, or treatment with absorbing charcoal, yielded an albumin band with a mobility identical to that in plasma, suggesting that liver albumin carries absorbed molecules with electronegative charges. The transferrin arc from liver, plasma and hepatoma had identical mobility. One protein with low mobility was present in higher concentration in the soluble liver fraction of male than of female rats, but it was reduced in hepatoma of male rats. The “h₂” proteins of liver were found in the cathodic region as 5 arcs, some of which were reduced, while others were not detectable in hepatoma.     

Retinyl palmitate, retinyl phosphate, and dolichyl phosphate of postnuclear membrane fraction from hepatoma, host liver, and regenerating liver: marginal vitamin A status of hepatoma tissue

The retinyl palmitate content of the postnuclear membrane fraction from 10 Morris hepatomas, their host rat livers, one acetylaminofluorene-induced rat liver hepatoma, and the host liver and of regenerating rat liver was measured by reverse-phase high-pressure liquid chromatography of the chloroform:methanol extracts. Membranes from the hepatoma tissue contained less than detectable levels of retinyl acyl esters, whereas membranes from host liver tissue and regenerating liver contained levels of retinyl palmitate within normal ranges. The amount of cellular retinol-binding protein was also decreased considerably in cytosols from 9618 and 7777 hepatomas. The ratio of endogenous retinyl phosphate to the polyisoprenoid dolichyl phosphate available for mannosylation in an assay containing postnuclear membranes and guanosine dephospho[14C] mannose was decreased by a factor of 3 to 10 in hepatoma tissue. Such change in ratio was not attributable to specific changes in retinyl phosphate mannose-synthesizing activity, but it appeared to be related to the vitamin A deficiency condition of the membrane from tumors. As for membranes from vitamin A-deficient liver tissue, postnuclear membranes from rat cystic hepatocarcinoma, Morris 7777, 3924A1-1, and 5123D-1-2 transplantable rat hepatomas and guinea pig line 10 hepatoma all synthesized a mannolipid with intermediate hydrophobic properties between retinyl phosphate mannose and dolichyl phosphate mannose and not normally found in liver tissue. These alterations in patterns of lipid intermediates may be responsible for altered glycosylation of glycoproteins in neoplastic cells. In conclusion, the present investigation establishes that hepatoma cell membrane is in a status of vitamin A and of retinyl phosphate depletion, while dolichyl phosphate contents appear similar to host liver membrane.