MSU RETURNING MATERIALS: Place in book drop to LIBRARIES remove this checkout from ”- your record. FINES wm be charged if book is returned after the date stamped below. POTENTIAL FOR ONCOGENE EXPRESSION IN THE LIVER AND IN SPONTANEOUS AND CHEMICALLY-INDUCED HEPATOMAS OF THE B6C3F1 MOUSE 3? Roeeann Lorraine Vorce A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Phnrmacology and Toxicology 1989 5o7492( POTENTIAL FOR ONCOGENE EXPRESSION IN THE LIVER AND IN SPONTANEOUS AND CHEMICALLY-INDUCED HEPATOMAS ‘ OF THE B6C3F1 MOUSE by Roseann Lorraine Vorce The male hybrid 86C3F1 mouse exhibits a 30% spontaneous hepatoma incidence, whereas the paternal C3H/I-Ie strain and the maternal C57BL/6 strain exhibit a 60% and a negligible incidence, respectively. The I-Ia-ras, Ki-ras, and myc oncogenes have been implicated in a variety of solid tumors. Specifically, Ha- and, less frequently, Ki-ras have been reported to be activated in 36C3F1 mouse liver tumors. The first objective of this study was to examine two possible points of transcriptional control of Hadras, Ki-ras, and myc in all three mouse strains, the hypothesis being that these oncogenes may be primed for expression in the nascent liver of those strains exhibiting a. high spontaneous hepatoma incidence. A positive correlation has been established between gene expression and both hypomethylation and the presence of DNase I hypersensitive sites. It was found that Ha-ras is hypomethylated in a site-specific manner in BéC3F1 and C3H/He mouse liver as compared to C57BL/6 mouse liver. DNase I hypersensitive sites were observed in the Ha-ras and myc oncogenes in the three mouse strains. However, Ha-ras appears to possess an additional site in BGCBFI and C3H/He as compared to‘ C57BL/6. Similarly, the Ki-rss oncogene exhibited a DNase I hypersensitive site only in 36C3F1 and C3H/He mouse liver. These results indicate that the hepatoma-prone strains (36031“. and C3H/I-Ie) may have a Rose ann Lorraine Vorce greater potential for Ha- and Kidras expression than does the non-hepatoma-prone strain (C57BL/6). This may explain, in part, the high propensity of BbCBFI and C3H/I-Ie mice toward hepatoma development. It was also hypothesized that Ha-ras, Ki-ras and myc have an increased potential for expression in 86C3Fl mouse liver tumors. Therefore, the methyla- tion states of these genes was examined in spontaneous liver tumors and in tumors induced by three diverse hepatocarcinogens: benzidine, phenobarbital, and chloroform. Ha-ras was found to be hypomethylated in all tumors examined, whereas Ki-ras was sometimes hypomethylated. The methylation state of myc usually was unaltered, although this gene appeared to be amplified in tumors. These results suggest that a component of the mechanism by which these oncogenes are activated in 36C3F1 mouse liver tumors involves an increased potential for expression, via hypomethylation of the ras oncogenes and amplifi- cation of myc. Therefore, it appears that common mechanisms may underlie the development of both spontaneous and chemically-induced 86C3F1 mouse liver tumors. ACKNOWLEDGEMENTS I sincerely thank Dr. Jay Goodman for his guidance and encouragement throughout this project. Our many and varied discussions have been vital to my development both as a scientist and as a person. The conscientious participation of Drs. Hsing-Jien Kung, James Trosko, Robert Roth, and Ken Moore as members of my guidance committee is very much appreciated. I also thank Dr. John Burch for his suggestions regarding the DNase I experiments; Laurie Mead Varner for technical assistance; and Diane Hummel for her expertise during the preparation of this dissertation. The award of a competitive fellowship by The Pharmaceutical Manufactur- ers Association Foundation, Inc., is gratefully acknowledged. The friendship of Karen Mudar, Jean Ray, Beth VandeWaa, Debbie MacKen- zie-Taylor, Donna Lehman, Candy, Ginger, and Murray has been invaluable to me during the course of this project. Finally, I thank my husband, Paul Stemmer, for his unwavering love, patience, and support. TABLE OF CONTENTS Page LIST OF TABLES vi LIST OF FIGURES vii LIST OF ABBREVIATIONS xi INTRODUCTION 1 1. The B6C3Fl mouse in carcinogen bioassays 1 2.. Oncogene involvement in tumorigenesis 4 A. Oncogenes as normal components of the genome 4 B. Oncogenes in neOplastic growth 5 C. Activated oncogenes in B603Fl mouse liver tumors 7 3. Methylation and transcriptional activity of genes 10 A. Methylation as a regulation point of gene transcriptions 10 B. Gene activation through induction of a hypomethylated state 12 C. Hypomethylation and cancer 15 D. Use of restriction endonucleases to assess gene methyla- tion state 17 4. Deoxyribonuclease I hypersensitivity and transcriptional acti- vity of genes 19 5. Hypothesis and experimental objectives 23 MATERIALS AND METHODS 26 1. Animah: Maintenance and carcinogen treatment 26 2. Isolation of DNA: Marmur method 27 3. Restriction enzyme digestion, agarose gel electrophoresis and Southern transfer of DNA to nitrocellulose paper 28 A. Restriction enzyme digestion 28 B. Agarose gel electrophoresis 28 C. Southern transfer 29 4. Amplification and isolation of plasmids containing pRSA l3 and myc Pat I fragment 30 5. Nick translation of the Ha-ras BS-9 probe 31 6. Hybridization of DNA affixed to nitrocellulose membranes 33 TABLE OF CONTENTS (continued) 7. 10. 11. 12. I3. 14. 15. RESULTS 1. 2. 3. 4. 5. Simultaneous isolation of DNA and RNA: CsCI method A. DNA isolation B. RNA isolation Isolation of DNase I-treated DNA Restriction enzyme digestion, electrophoresis, and Southern transfer of DNA to Gene Screen Plus A. Restriction enzyme digestion B. Agarose gel electrophoresis C. Southern transfer Labelling of probes with 3213: Random primers method Hybridization of DNA affixed to Gene Screen Plus with labelled probes Electrophoresis and Northern transfer of DNA A. Preparation of RNA B. Electrophoresis and Northern transfer Hybridization of Northern blots Assessment of RNA integrity Autoradiograpby 32P_ Methylation status of the Ha-ras, Ki-ras, and myc oncogenes in B6C3F1, C3H/He, and C57BL/6 mouse liver Oncogene methylation state in B6C3F1 mouse liver tumors Assessment of oncogene methylation state with Hha I DNase I studies RNA studies DISCUSSION 1. Differential potential for expression of oncogenes in B6C3F1, C3H/He, and C57BL/6 mouse liver A. Hypomethylation of Ha-ras in B6C3F1 and C3H/He mouse liver B. Methylation state of the Ki-ras and myc oncogenes in B6C3F1, C3H/He, and C57BL/6 mouse liver C. DNase I hypersensitive sites in oncogenes Increased potential for oncogene expression in B6C3F1 mouse liver tumors A. Hypom ethylation of the Ha-ras oncogene B. Hypomethylation of the Ki-ras oncogene C. Amplification and hypomethylation of the myc oncogene 33 34 34 35 36 36 37 37 38 38 39 39 40 41 42 43 57 74 110 120 120 120 125 125 127 127 130 132 TABLE OF CONTENTS (continued) Page 3. Level of Ha-ras mRNA in benzidine-induced B603F1 mouse liver tumors 134 4. Implication of increased potential for oncogene expression 136 SUMMARY AND CONCLUSIONS 140 BIBLIOGRAPHY 142 LIST OF TABLES Table Page 1 Methylation state of the Ha-ras in three mouse strains 46 2 Ha-ras mRNA levels in benzidine-induced B6C3F1 mouse liver tumors and adjacent non-tumor tissue 1 19 Figure 10 11 12 13 14 LIST OF FIGURES Recognition sites of the restriction endonucleases Msp I, Hpa II, and Hha I Schematic diagram of chromatin alterations involved in gene activation and the relationship of these changes to sensitivity of genes to DNase I digestion Methylation state of the Ha-ras oncogene in B6C3F1, C3H/He, and C57BL/6 mouse liver using the 88-9 clone Methylation pattern of the Ha-ras oncogene in control mice Methylation pattern of the Ha-ras oncogene in mice of the opposite sex Effect of the omission of the depurination step on the transfer of large DNA fragments Methylation status of the Ki-ras oncogene in control mice Methylation status of the myc oncogene in control mice Methylation status of the Ki-ras oncogene in mice of the opposite sex Methylation status of the myc oncogene in mice of the opposite sex Methylation status of the serum albumin gene in control mice Methylation status of the Ha-ras oncogene in benzidine- induced hepatic tumors and adjacent non-tumor tissue Lack of effect of depurination on the methylation assessment of Ha-ras in benzidine-induced tumor and non-tumor tissue Methylation status of the Ha-ras oncogene in phenobarbital- and chloroformdinduced hepatic tumors 18 20 47 49 50 51 52 55 56 58 59 60 LIST OF FIGURES (continued) Figure 15 16 I7 18 19 20 21 22 25 26 27 Methylation status of the Ha~ras oncogene in spontaneous hepatic tumors Methylation status of the Ki-ras oncogene in benzidine- induced hepatic tumors and adjacent non-tumor tissue Methylation status of the Ki-ras oncogene in phenobarbital- and chloroform-induced hepatic tumors Methylation status of the Ki-ras oncogene in spontaneous hepatic tumors Methylation status of the Ha-ras oncogene following partial hepatectomy Methylation status of the Ki-ras oncogene following partial hepatectomy Methylation status of the myc oncogene in benzidine-induced hepatic tumors and adjacent non-tumor tissue Methylation status of the myc oncogene in phenobarbital- and chloroform~induced hepatic tumors Methylation status of the myc oncogene in spontaneous hep a- tic tumors Methylation status of the serum albumin gene in benzidine- induced hepatic tumors and adjacent non-tumor tissue Methylation status of the serum albumin gene in phenobarbi- tal- and chloroform-induced hepatic tumors Methylation status of the serum albumin gene in spontaneous hepatic tumors Hha I assessment of the methylation state of the Ha-ras oncogene in B6C3F1, C3H/He, and C57BL/6 mouse liver Hha I assessment of the methylation state of the Ha-ras oncogene in benzidinedinduced tumors and adj acent non-tumor tissue 62 63 64 66 69 70 71 72 73 75 76 77 79 82 LIST OF FIGURES (continued) Figure 29 30 31 32 33 34 35 36 37 38 39 41 Hha I assessment of the methylation state of the Ha-ras oncogene in phenobarbital-induced, chloroform-induced, and spontaneous hepatic tumors (short exposure time) Hha I assessment of the methylation state of the Ha-ras oncogene in phenobarbital-induced, chloroform-induced, and spontaneous hepatic tumors (long exposure time) Hha I assessment of the methylation state of the Ki-ras oncogene in B6C3F1, C3H/He, and C57BL/6 mouse liver \ Hha I assessment of the methylation state of the Ki-ras oncogene in benzidine-induced tumors and adjacent non-tumor tissue Hha I assessment of the methylation state of the Ki-ras oncogene in phenobarbital-induced, chloroform-induced, and spontaneous hepatic tumors Hha I assessment of the methylation state of the myc onco- gene in B6C3F1, C3H/I-Ie, and C57BL/6 mouse liver Hha I assessment of the methylation state of the myc onco- gene in benzidine-induced tumors and adjacent non-tumor tissue Hha I assessment of the methylation state of the myc onco- gene in phenobarbital-induced, chloroform-induced, and spon- taneous hepatic tumors Hha I assessment of the methylation state of the serum albumin gene in B6C3F1, C3H/He, and CS7BL/6 mouse liver Extent of DNA digestion in DNase I-treated nuclei Assessment of the Ha-ras oncogene for the presence of DNase I hypersensitive sites in B6C3F1 mouse liver using EcoRl Assessment of the Ha-ras oncogene for the presence of DNase I hypersensitive sites in C3H/He mouse liver using EcoRl Assessment of the Ha-ras oncogene for the presence of DNase I hypersensitive sites in C57BL/6 mouse liver using EcoRl 85 86 88 90 92 95 96 99 101 103 104 LET OF FIGURES (continued) Figure 42 43 45 46 47 48 49 50 51 52 53 Assessment of the Ha-ras oncogene for DNase I hypersensitive sites in B6C3F1 mouse liver using Hind 1:: Confirmation of the presence of a Ha-ras band at 1.9 kb in DNase I-treated B6C3F1 mouse liver DNA digested with Hind m Assessment of the firms oncogene for the presence of DNase I hypersensitive sites in C3H/He mouse liver using Hind III Assessment of the Ha-ras oncogene for the presence of DNase I hypersensitive sites in CS7BL/6 mouse liver using Hind III Assessment of the Ki-ras oncogene for the presence of DNase I hypersensitive sites in B6C3F1 mouse liver Assessment of the Ki-ras oncogene for the presence of DNase I hypersensitive sites in C3H/He mouse liver Assessment of the Ki-ru oncogene for the presence of DNase I hypersensitive sites in C57BL/6 mouse liver Assessmant of the myc oncogene for the presence of DNase I hypersensitive sites in B6C3F1 mouse liver Assessment of the myc oncogene for the presence of DNase I hypersensitve sites in C3H/He mouse liver Assessment of the myc oncogene for the presence of DNase I hypersensitive sites in C57BL/6 mouse liver Assessment of RNA integrity Levels of Ha-ras mRNA in benzidine-induced tumors and adjacent non-tumor tissue Page 105 106 107 108 109 111 112 113 114 115 116 118 SAM LIST OF ABBREVIATIONS diethylnitrosamine deoxyribonucleic acid deoxyribonuclease I Harvey-ras Kirsten-ras messenger RNA ribonucleic acid ribonuclease S-adenosylmethionine INTRODUCTION 1. The B6C3Fl Mouse in Carcinogen Bioassays ' The sbcam mouse (male C3H/He x female C57BL/6) was developed by the National Cancer Institute in the 1960s for use in carcinogen bioassays. Several advantages exist for the choice of this strain in these long-term studies (Cameron ,5 31., 1985). As a hybrid, the B6C3F1 mouse is genetically heterogeneous, a trait shared with the human population, and hybrids exhibit increased hardiness and longevity over inbred strains. In addition, the B6C3Fl mouse has a low incidence of mammary tumors and leukemia relative to many murine strains. However, the male B6C3F1 mouse has a spontaneous hepatoma incidence of approximately 30% (Becker, 1982; Maronpot gt a_l., 1987); spontaneous hepatoma incidence in females is much lower at 7-8% (Maronpot e_t_ 2.1., 1987). This characteristic is heritable and, in males, intermediate between that of the paternal C3H/He strain and the maternal CS7BL/ 6 strain which exhibit 60% incidence and virtually zero inci- dence, respectively, at 18 months of age (Becker, 1982). Tumors exhibit a low level of malignancy, and neither metastasize nor kill the host. The liver of the B6C3Fl mouse (both males and females) is extremely sensitive to tumor induction by a wide variety of chemicals, including classic tumor initiators such as diethylnitrosamine (DEN) (V esselinovitch and Mihailovich, 1983; Stowers 91 fl» 1988), N-hydroxy-Z-acetylaminofluorene (Wiseman e_t_ a_l., 1986), and other genotoxicants (Ashby and Tennant, 1988). In addition, a number of Ames test-negative chemicals effectively induce liver tumors in this animal (Ashby and Tennant, 1988). An example of a chemical falling into this category is phenobarbital, a non-mutagenic rodent tumor promoter. This chemical is able to increase the hepatoma incidence in B6C3F1 mice to 100% when administered in drinking water at a concentration of 0.05% for one year (Becker, 1982). Furthermore, the B6C3F1 mouse liver appears to be sensitive to hepatoma induction by high doses of some chemicals, including trichloroethylene and other chlorinated ethane derivatives, which rarely produce tumors in other species or at other sites (Clayson, 1987). Another chemical falling into this category is butylated hydroxytoluene (BHT), a widely used food preservative which is not considered mutagenic. BHT has been reported to cause liver tumors in male, but not female, B6C3F1 mice at the maximum tolerated dose of 2% in the diet (Inai g 31., 1988). In fact, a comprehensive survey evaluating 222 chemicals conducted by the U.S. National Cancer Institute/National Toxicology Program revealed that mouse liver accounts for 24% of all chemical/tissue reports of carcinogenicity among the 115 carcinogens plus the 24 equivocal carcinogens evaluated for carcinogenicity in B6C3F1 mice and Fisher 344 rats (Ashby and Tennant, 1988). This same survey found that only 30% of the mouse liver-specific carcinogens are mutagenic in the Ames test. Thus, it is obvious that some characteristic of B6C3F1 mouse liver confers upon it a predisposition to the development of tumors. The high spontaneous hepatoma incidence and sensitivity to chemical induction of liver tumors in the B6C3F1 mouse have fueled the controversy over whether or not mouse liver tumors are a valid endpoint for carcinogen bioassays. The classification of mouse liver-positive chemicals as carcinogens, despite negative results in both a bioassay using another rodent species and the Ames test, is especially disconcerting. It is interesting to note that phenobarbital, were it bioassayed today, would be placed in the Environmental Protection Agency's Category I (probable human carcinogen), with zero tolerance and no threshold. In reality, epidemiological studies have produced no evidence that chronic admini stration of this drug is carcinogenic in humans (Clemmesen and Hjalgrim-Jensen, 1978a,b, 1980, 1981). In addition, Becker (1982) found that phenobarbital is unable to induce liver tumors in the non-hepatoma prone CS7BL/ 6 mouse, although B6C3F1 and C3H/He mice, both of which are hepatoma-prone strains, showed a strong positive response. Thus, the simple extrapolation of the results of the carcinogen bioassay to a wide variety of species based on a positive response in mouse liver alone may be an overly conservative interpretation. Based on the above discussion, data generated using the B6C3F1 mouse in carcinogen bioassays led to the suggestion that this animal exhibits an abnormal response during the promotion phase of tumor develOpment. Supporting this view is the finding by Drinkwater and Ginsler (1986) and Hanigan and coworkers (1988) that the increased susceptibility of the paternal, hepatoma-prone C3H/He mouse over the maternal, non-hepatomadprone C57BL/6 mouse to DEN induction of liver tumors is heritable and largely determined by a single genetic locus which appears to exert its effects during tumor promotion. Specifically, the proliferative rate of both normal and preneOplastic hepatocytes is increased in animals possessing this HCS (hepatocarcinogen sensitivity) locus. Similarly, Dragani e_t a}. (1987) have hypothesized that the increased susceptibility to carcinogenesis induced by DEN / 1,4-bis [ 2-(3,5—dichloropyridyloxy)] benzene (a phenobarbital-er chemical) observed in B6C3F1 mice as compared to a non-hepatoma-prone murine strain is due to a higher proliferative rate of initiated B6C3F1 hepatocytes. Because promotion is comidered to consist of epigenetic alterations (i.e., changes in gene expression) which lead to a proliferative advantage of initiated cells, this study was designed to examine putative control points of transcriptional activity of specific oncogenes believed to be involved in hepatocarcinogenesis. The hepatoma-prone B6C3Fl mouse is an excellent model in which to study the molecular mechanisms by which a phenotypically normal cell is transformed into a malignant cell. In addition, it is hoped that the rational interpretation of carcinogen bioassay data will be facilitated by the results presented herein. 2. Oncogene Involvement in Tumorigenesis A. Oncogenes as normal components of the genome Oncogenes are a group of normal cellular genes which are highly conserved in evolution (Shilo, 1984; Shilo and Weinberg, 1981) and appear to play a vital role in normal growth and differentiation. For example, the myc oncogene is differentially expressed both temporally mid spatially during development of the fetal and neonatal mouse (Slamon and Cline, 1984; Ruppert g; 31., 1986; Zimmerman g _a_l., 1986) and in develOping human placenta (Pfeifer-Ohlsson _e_t_ 51., 1984). The src oncogene has been implicated in the differentiation of neural tissue, and fos appears to be involved in the differentiation of cells. of different hematopoietic lineage (Muller, 1986). The expression of a number of oncogenes has been demonstrated to be cell cycle-dependent, i.e., expression of these genes is minimal or absent in quiescent celb, but greatly elevated in proliferating cells (for review see Kaczmarek, 1986). Oncogenes falling into this category include myc, fos, myb, Ha-ras, Ki-ras, and N-ras. The expression of two of these (myc and myb) have been found to vary in a manner dependent on the different stages of the cell cycle. Although it is clear from the above discussion that oncogenes are somehow involved in cell proliferation and differentiation, the precise biochemi- cal role of oncogene products remains unclear. In general, oncogenes can be classified into one of four categories (Weinberg, 1985; Pitot, 1986; Garrett, 1986). The protein products of a number of oncogenes, including src and ab 1, function as kinases, the unusual substrate being a tyrosine residue of various proteins. It is well established that phosphorylation of many enzymes serves to alter their kinetic properties, so it is likely that this oncogene functions in a similar manner. The products of a second group of oncogenes, such as Ha- and Ki-ras, are located on the plasma membrane and bind GTP. Thus, the function in these oncogene products is speculated to involve transmembrane signalling. The products of a third class of oncogenes binds DNA. The expression of two members of this class, myc and myb, is elevated just prior to DNA replication. These two traits make it likely that such oncogenes are necessary for DNA synthesis. The last class of oncogene products are homologs of growth factors (e.g., sis and platelet-derived growth factor) or growth factor receptors (e.g., erbB and the epidermal growth factor receptor). These oncogenes most likely play a role in the pathway of mitogen-induced cell proliferation. B. Oncogenes in neOplastic growth It has become increasingly obvious in recent years that oncogenes play a causal role in tumorigenesis (Land e_t g, 1983a; COOper and Lane, 1984; Slamon e_t 2.1., 1984; Spandidos, 1985). Two major lines of evidence lead to this conclusion. First, the study of acute transforming retroviruses has revealed that such viruses contain oncogenes (v-onc) which are responsible for the transforma- tion of host cells into malignant cells. The discovery of homologous cellular.(host) oncogenes (c-onc) (DeFeo 51 2.1., 1981) led to the subsequent finding that activation of such genes could occur during tumorigenesis (Der e_t 51., 1982). It is interesting to note that retroviral oncogenes originated from the host and apparently were acquired as a consequence of the ability of retroviruses to reversibly integrate and recombine into the host's DNA. The second line of evidence has demonstrated that transfection of DNA from transformed cells into Nn-I3T3 cells results in their acquisition of the transformed phenotype. It has been shown that the activated oncogenes, primarily from the src and ras families, are the critical entity transferred to NIH3T3 cells in such cases. A number of mechanisms have been discerned by which oncogenes are activated. The ras and src oncogenes are often activated by point mutation. For example, a mutation in the 12th (Reddy gt g_I., 1982; Tabin gt _a_l_., 1982; Taparowsky e_t _a_l., 1982) or 61st (Sekiya gt 51., 1985) codon of the Ha-ras oncogene appears to be a requirement for conferring the transforming capacity on this gene. Mammary tumors induced by N-nitroso-N-methylurea have been reported to contain a specific point mutation, a G to A transition, in codon 12 of the Ha—ras oncogene (Zarbl gt gL, 1985). The p21 protein product of such a mutated gene exhibits altered electrophoretic mobility as compared to the normal protein (Quintanilla e_t a_l., 1986; Harper e_t g_l., 1987), and it is believed that an accompanying alteration in biochemical ftmction contributes to the malignant phenotype. Mutation has also been demonstrated to be the mechanism by which the Ki-ras oncogene is activated (Santos gt 311., 1984; O'Hara e_t _a_l., 1986; Liu e_t 311., 1987). In contrast, overexpression of the myc oncogene is commonly seen in malignant cells. This occurrence may be due to gene amplification (Alitalo gt gL, 1987), rearrangement to the region of the immunoglobulin heavy chain locus (Blick _e_t _a_l., 1986; Murphy gt g1., 1986), or unknown causes (Rothberg gt g_l., 1984; Erisman e_t gl., 1985; Yoshimoto e_t 3.1., 1986). Similarly, an increased transcrip- tion rate of Ha-ras is sufficient for transformation of NIH3T3 cells (Chang gt 51., 1982; DeFeo, 1981). The multistep nature of cancer points to multiple changes in multiple genes, and there is evidence that activation of more than one oncogene is required to effect transformation (Glaichenhaus gt 9, 1985). The ras and myc oncogenes have qualitatively different effects which can act in a complementary fashion in the transformation process: the ras oncogene is potent in inducing refractile morphology, anchorage independence, and growth factor secretion and is weak in its ability to immortalize cells, whereas the myc oncogene is capable of immortalizing cells (Land gt ga1., 1986). Due to these characteristics, an activated (mutated) Ha-ras oncogene is able to transform NIH3T3 cells when transfected into them, whereas transfection of myc, as expected, has no effect in this already immortal cell line. It has been shown that neither Ha-ras nor myc alone is able to transform rat embryo fibroblasts (REFs) in the _ig 3113 transfection assay, whereas simultaneous introduction of these two oncogenes into REFs produces a high degree of transformation (Land gt gl., 1983b; Birrer g 51., 1988; Storer g 51., 1988). Similarly, an activated Ha-ras oncogene is able to transform hamster fibroblasts only if they have previously been immortalized by carcinogen treat- ment (Newbold and Overell, 1983). These results are corroborated by observations 131 zigvg: an activated ras oncogene has been found together with an activated myc oncogene in a promelocytic leukemia and in an American Burkitt lymphoma (Murray gt _a_1., 1983), and two or more oncogenes were found to be transcrip- tionally active in a variety of human malignancies (Slamon e_t gl., 1984). C. Activated oncogenes in B6C3F1 mouse liver tumors The presence of activated oncogenes was first identified in sponta- neous B6C3F1 mouse liver tumors by Fox and Watanabe in 1985. To this end, tumor DNA was transfected into NIH3T3 cells; subsequent foci formation was taken as evidence that an activated oncogene was contained in the tumor DNA. Reynolds and coworkers (1986) identified Ha-ras as the activated oncogene species in the majority (1 1/13) of transfection assay-positive spontaneous tumors examined. The p21 protein product derived such tumors (i.e., those containing an activated Ha-ras oncogene) has been shown to exhibit altered electrophoretic mobility as compared to the normal protein (Reynolds _e_t 51., 1986; Wiseman e_t _a_l., 1986; Reynolds gt 3.1., 1987; Stowers gt gL, 1988), suggesting that activation of Ha-ras is the result of a mutation in this gene. As mentioned in the previous section, activation of the Ha-ras oncogene is often due to a mutation in codon 61. Hybridization of tumor DNA with oligonucleotide probes capable of detecting a single base change in codon 61 has revealed that this codon is mutated in 50/57 chemically-induced (Wineman gt g.1., 1986; Reynolds gt a_l., 1987; Stowers gt _a_l., 1988) and 15/15 spontaneous (Reynolds gt gl., 1987) B6C3F1 mouse liver tumors containing an activated Hadras oncogene; mutations in codons 13 or 17 of Ha-ras have been seen in 7/57 chemically-induced tumors (Reynolds e_t gl., 1987). In addition, an activated Ki- ras oncogene, as well as other oncogenes, occasionally has been observed (Reynolds e_t 51., 1987). These results indicate that the presence of an activated oncogene in B6C3F1 mouse liver tumors is frequently associated with mutation of Ha-ras, usually within codon 61, regardless of whether the tumors were chemical- ly-induced or spontaneous. It must be noted that DNA from non-tumor B6C3F1 mouse liver has not been demonstrated to contain activated oncogenes in the aforementioned studies, suggesting that the frequently-observed mutation in codon 61 of Ha~ras is an acquired characteristic. Therefore, this mutation cannot be responsible for the predisposition of the B6C3F1 mouse to hepatoma development. In addition, a number of spontaneous and chemically-induced tumors have tested negative in the NIH3T3 transfection assay. This result suggests that a mechanism other than mutation of Ha-ras, either through disruption of transcriptional regulation of Ha- ras or activation of another oncogene, underlies hepatoma development in these cases. The transformation of a phenotypically normal cell into a malignant cell is considered to consist of an initial mutagenic event, termed initiation, followed by epigenetic events leading to a selective growth advantage of initiated cells (promotion), putatively through changes in gene expression patterns. It has been demonstrated that mutational activation of Ha-ras is an early event in B6C3F1 mouse liver carcinogenesis (Wiseman gt g1., 1986), as well as mammary (Zn-bl g gl_., 1985) and skin (Quintanilla gt _a_l., 1986; Bizub _e_t a_.l., 1986; and Felling e_t 91., 1986) carcinogenesis. Once a critical mutation occurs, deregulation of gene expression becomes pivotal in determining whether or not the cell, proceeds to a phenotypically malignant state. The Ha-ras oncogene in the nascent liver of the B6C3F1 mouse may be primed for expression, leading to an increased possibility that an activating mutation will be expressed and will thereby result in the phenotypic alterations associated with malignancy. In addition, a compro- mised ability to control the transcriptional activity of Han-as may be the heritable factor responsible for the high spontaneous hepatoma incidence and the sensitivity to chemical induction of liver tumors in the B6C3F1 mouse. Supporting this view is the fact that B6C3F1 mouse liver appears to respond abnormally during tumor promotion (see Section 1). In addition, it has been demonstrated that DEN-induced hepatomas in Fisher 344 rats do not possess activated Ha-ras oncogenes (0/28), whereas Ha-ras was activated in 14/14 DEN- induced hepatomas in B603F1 mice (Stowers gt gL, 1988). Since it is unlikely that DEN produces different mutation frequency and/or patterns in these two species, it seems reasonable to suspect that a critical mutation in Ha-ras of B6C3F1 mouse liver is more likely to exert phenotypic effects due to an increased probability of Ha-ras expression in this species. Therefore, the studies which comprise this thesis examined two parameters of gene expression with regard to the oncogenes putatively involved in B6 C3F 1 mouse hepatocarcinogenesis. First, the methylation state of these genes was examined due to the established correlation between hypomethylation of a gene and its potential for expression. 10 Second, these genes were assessed for the presence of deoxyribonuclease I (DNase I) hypersensitive sites, as transcriptionally active genes are known to contain sites which are exquisitely sensitive to the action of this enzyme. 3. Methylation and Transcriptional Activity of Genes A. Methylation as a regulation point of gene transcription In mammals, a number of regulatory mechanisms participate in the control of gene expression. At the DNA level, one of these mechanisms has been identified as the methylation state of a gene (Riggs and Jones, 1983; Jones, 1985). In general, a relatively low degree of gene methylation (hypomethylation) is associated with gene expression, whereas a relatively high degree of gene methylation (hypermethylation) acts to block transcription. The only naturally-occurring methylated base found in mammalian DNA is S-methylcytosine (S-MC). Approximately 3% of all cytosine residues exist ‘ in the form of S-MC, and at least 90% of these methylated bases occur in the sequence 5‘-CG-3' (termed I'CpG islands"). Depending on the species and tissue, between 70 and 90% of such CG sequences are methylated. Furthermore, these sites are methylated in a symmetrical fashion, i.e., cytosine residues are methylated on both strands: Me suc‘c, -3. 3'-G('3 ~5' Me It has been hypothesized that S-MC exists at all sites that will ever be methylated early in embryonic development (Singer gt 9, 1979; Resin and Riggs, 1980). Tissue-specific patterns of gene expression then are established during the DNA replication accompanying cell division through selective inhibition of methylation of genes destined to be transcriptionally active or through active demethylation of specific genes (Ruin-e_t $1., 1986). Once a methylation pattern 11 is established within a cell's genome, it is prOpagated through successive rounds of cell division, as described below. Replication of DNA containing a S-MC residue results in the formation of hemimethylated DNA, i.e., DNA that is methylated on only the parental strand. It has been demonstrated that a maintenance methylase activity exists in mammalian cells which recognizes hemimethylated sites. Using S-adenosyl methionine (SAM) as the methyl group donor, this enzyme methylates the newly synthesized strand of DNA, thereby forming the fully methylated site. Thus, once a pattern of gene methylation is established, maintenance methylase activity allows somatic heritability of this trait. Hypomethylation of a gene has been associated with its transcriptional activity in a variety of systems. The serum albumin gene is hypomethylated in normal rat hepatocytes, but hypermethylated in non-parencbymal cells (Vorce and Goodman, 1985), conditions consistent with expression in the former cell type and non-expression in the latter. Similarly, Ichinose and coworkers (1988) have demonstrated a correlation between hypomethylation of the pepsinogen gene and its expression in developing rat stomach. Conversely, it has been observed that CpG islands within housekeeping genes on the inactive mammalian X chromosome are methylated; methylation has therefore been proposed as the mechanism by which the inactivity of such genes is maintained (Bird, 1986). Supporting this hypothesis is the finding that demethylation of three CpG islands within the glucose-b-phosphate dehydrogenase gene on the inactive X chromosome is asso- ciated with activation of this gene (Toniolo gt 5.1., 1988). Methylation of certain regions of a gene appears to be more critical than others in inhibiting gene expression. Specifically, methylation of the 5' end of a number of genes appears to regulate transcription. One such example is the hamster adenine phosphoribosyltransferase gene. When the body of this gene is 12 artificially methylated, there is no inhibition of transcription, but when a region near the 5' end is methylated, transcription is significantly reduced (Keshet gt 3.1., 1985). A similar situation has been found in other systems: absence of methylation in the 5' region of the calcitonin gene has been correlated with its expression in a medullary thyroid carcinoma cell line (Baylin gt _a_l., 1986), and a region of the albumin gene extending from the 5' end to the middle is hypomethyl- ated in a hepatoma cell line which synthesizes albumin, but methylated in a nonproducing variant (Orlofsky and Chasin, 1985; Ott _et gl_., 1982). However, in other genes, different patterns of methylation appear to be necessary for gene expression. In S49 lymphoma cells, the metallothionine gene is heavily methyl- ated and not expressed; when induced to synthesize metallothionine by UV irradiation, a region of one allele spanning this gene becomes demethylated (Lieberman gt _a_l., 1983). A variant of the mouse hepatoma cell line Hepa-l has a high constitutive level of cytochrome P1450 expression which appears to ’be due to hypomethylation of a specific site in the middle of the P1450 gene (Peterson gt gl_., 1986). Hypomethylation of regions near the 3' end of a gene has also been found to be necessary for expression. The myc oncogene is methylated at a site near the 3' end in normal cultured fibroblasts, but hypomethylated in three of five tumor cell lines (Cheah gt g_l., 1984). Thus, hypomethylation of regions of genes near the 5' end, 3' end, and middle has been correlated with gene expression. B. Gene activation through induction of a hypomethylated state The cytidine analog S-asacytidine (S-asaCR) is capable of inducing a hypomethylated state in proliferating cells by virtue of its ability to inhibit maintenance methylase irreversibly following its incorporation into DNA (Taylor gt _a_l., 1984; Jones, 1985). S-AzaCR treatment of a variety of cell types has been demonstrated to activate previously quiescent, methylated genes (Jones, 1985). For example, S-azaCR is able to induce expression of the hypoxanthine guanine l3 phosphoribosyl tramferase gene contained on the inactive X chromosome (Mohan- das gt a_l., 1981; Graves, 1982; Lester _e_t a_l., 1982). Other genes whose expression has been induced by 5-azaCR treatment include prolactin, growth hormone, globin, and thymidine kinase (Jones, 1985). 5-AzaCR also has been shown to induce cellular differentiation (Jones, 1985). New phenotypes induced by 5-azaCR treatment of various cell lines include muscle cells, adipocytes, chrondrocytes, and others. Similarly, immuno- resistance acquired by tumor cells via loss of tumor antigen expression can be eliminated through 5-azaCR treatment (Altevogt gt _a_l., 1986). These results indicate that inhibition of methylation by S-azaCR results in derepression of transcription in a variety of genes. S-AzaCR treatment has been shown to induce transformation of established cell lines (Yasutake gt 9, 1987; Hsiao _e_t 51., 1985) and primary rat tracheal epithelial cultures (Walker and Nettesheim, 1986). An increased ability to form experimental metastases was observed after S-asaCR treatment of 316 melanoma cells (Trainer gt gL, 1985). Furthermore, there is evidence that 5- azaCR is carcinogenic in mice (Cavaliere gt a_l., 1987). It thus appears that alteration of gene expression via S-uaCR-induced gene hypomethylation plays a role in the transformation process. It follows that derangement of methylation patterns may play a role in tit !i_vo_ carcinogenesis (Riggs and Jones, 1983; Jones, 1985; Boehm and Drahovsky, 1983). Support for this hypothesis comes from dietary studies in which animals are fed a methyl-deficient diet. It has been observed that rats fed a diet deficient in methionine and choline develop liver tumors (Yokoyama gt 9., 1985; Ghoshal and Farber, 1984; Mikol e_t a_lw 1983). It has been demonstrated that a methyl-deficient diet results in a decrease in available methyl groups, i.e., the ratio of S-adenosyl methionine to S-adenosyl homocysteine is significantly l4 lowered in animals fed the choline/methionine-deficient diet as compared to controls (Wilson gt a_l., 1984). Furthermore, the S-MC content of hepatic DNA, as assessed by high performance liquid chromatography, is significantly decreased in rats fed this methyl-deficient diet for 22 weeks (Wilson gt _a_l., 1984). A similar decrease in S-MC content of hepatic DNA was seen in rats fed a choline-devoid diet for 14 months (Locker gt gL, 1986). In this case, restriction endonuclease analysis (using enzymes that can distinguish between methylated and immethyl- ated recognition site) showed hepatic DNA to be hypomethylated in both tumor and non-tumor tissue of animals fed the methyl-deficient diet as compared to hepatic DNA from animals fed the control diet. These results suggest that a methyl-deficient diet may be carcinogenic by virtue of its ability to transcrip- tionally activate critical genes through the induction of a generalized state of DNA hypomethylation. However, reversible alkali-labile lesions, indicating DNA damage and repair (Rushmore gt gl., 1986), and free-radical production (Rushmore gt 3., 1987) have been observed as early events resulting from commencent of a methyl-deficient diet. Thus, it is possible that mutations also result from this dietary regimen and may thereby contribute to carcinogenesis. It is interesting to note that a choline-devoid diet causes a high level of cell proliferation in the liver which persists after the reintroduction of choline, although the proliferative rate decreases with time (Chandar and Lombardi, 1988). This result suggests that DNA hypomethylation may activate genes involved in mitogenesis. Indeed, hypo- methylation of Ha-ras and Ki-ras has been observed in preneoplastic and neo- plastic livers of rats maintained on a methyl-deficient diet (Bhave gt g_l., 1988), and ras gene transcription is enhanced during rat hepatocarcinogenesis induced by choline deficiency (Yaswen 2 gL, 1985). 15 C. Hypomethylation and cancer It has been proposed that modifications of DNA methylation are involved in the inheritance of epigenetic defects, including those seen in carcino- genesis (Holliday, 1987 a,b). A number of studies have shown that the S-MC content of DNA is lower in malignancies than in normal tissue (Riggs and Jones, 1983). A reduction in 5-MC was observed in both benign and malignant human colon tumors (Feinberg gt a_l., 1988) and hamster kidney tumors (Lu gt _a_l., 1988) as quantitated by HPLC analysis. Similar results were obtained by restriction enzyme analysis: hypomethylation was observed in a variety of genes in DNA derived from benign and malignant human colon neoplasms relative to the methylation state of these genes in normal tissue (Gaels gt gL, 1985). Genomic hypomethylation was also observed in a mimber of metastatic variants selected from a poorly metastatic human melanoma cell lines (Liteplo and Kerbel, 1987). These results indicate that generalised DNA hypomethylation is a common phenomenon in transformed cells. This condition may facilitate aberrant gene expression, including those genes involved in the carcinogenic process. Hypometbylation of specific oncogenes has been observed in numerous neoplasms. The Ha- and Ki-ras oncogenes were found to be hypomethylated in various human tumors (Feinberg and Vogelstein, 1983), and the 5' region of the Ha-ras oncogene is undermethylated in a leukemic cell line (Barbieri gt g_l., 1987). Hypomethylation of the myc oncogene, but not two other genes, has been observed in human hepatocellular carcinomas (Kaneko .e_t gL, 1985). Furthermore, a specific site in the (third exon of the myc oncogene is hypomethylated in human hepatocellular carcinoma (Nambu gt g_l., 1987) and in a set of human tumor cell lines (Cheah e_t g_l., 1984). Thus, it appears that oncogenes thought to be involved in carcinogenesis often possess an increased potential for expression in tumors 16 versus non-tumor tissue. Furthermore, site-specific hypomethylation may be sufficient to permit transcriptional activity of certain genes. Treatment of cultured cells with a number of carcinogens results in a decrease in genomic S-MC content (Wilson and Jones, 1984; Wilson e_t _a_l., 1987a: Boehm _et g1” 1983). in gitrg, the transfer of a methyl group from SAM to hemimethylated DNA is inhibited by a range of ultimate carcinogens (Wilson and Jones, 1983), a result suggesting that the activity of maintenance methylase may be inhibited by carcinogenic species. Paradoxically, methylase activity is higher in tumorgenic as compared to non-tumorigenic cell lines (Kmtiainen and Jones, 1986); the activity of this enzyme is also higher in target rat tissues after _ig gigs; treatment with N-methyl-N-nitrosourea (Pfohl-Leskowicz and Dirheimer, 1986). In agreement with this observation, gg gggg methylation is increased following treatment of cells with N-acetoxy-N-Z-acetylaminofluorene, even though this chemical causes an initial dose-dependent decrease in maintenance methylase activity (Boehm g gL, 1983). The physical and catalytic properties of methylase derived from normal rat liver and a transplantable hepatocellular carcinoma have been compared and found to be indistinguishable (Ruchirawat gt gL, 1985). Therefore, although many carcinogens possess the ability to inhibit methylation of DNA, the mechanism does not appear to involve long-term depression of methyl- ase activity. However, transient inhibition of methylase by carcinogens during periods of hyperplasia could result in propagated DNA hypomethylation since the methylation state of genes is somatically heritable. It is also possible that the rebound increase in methylase activity results from increased transcription of the methylase gene due to its hypomethylation. Although methylation of genes appears to be a control point of gene expression, hypomethylation alone is not considered sufficient to produce tran- scriptional activity. This is illustrated by the fact that some hepatoma cell line l7 variants contain a hypomethylated serum albumin gene yet do not produce albumin. In addition, it has been shown that the stimulation of prolactin and growth hormone production by S-azaCR is a function of time, i.e., maximal stimulation occurs 3 weeks after drug exposure (Laverriere _e_t gL, 1986). These observations indicate that further epigenetic alterations beyond hypomethylation are necessary for transcription to proceed. D. Use of restriction endonucleases to assess gene methylation state To assess the methylation state of a gene, the restriction endo- nucleases Msp I and Hpa H may be utilized. The Msp I/Hpa H isoschizomers cleave double-stranded DNA at the following recognition sequece: 5'-CCGG-3' (Figure 1). However, Msp I, but not Hpa H, is able to perform this function when the internal cytosine residue is methylated. Conversely, Hpa H, but not Msp I, will cleave this sequence when the external cytosine is methylated. Since more than 90% of 5—methylcytosine occurs in the sequence of 5'-CG-3' (Riggs and Jones, 1983), the methylation state of a gene may be assessed by comparing the restriction patterns produced by digestion with Msp I and Hpa H. If a gene is hypomethylated, very similar restriction patterns will be produced by digestion with either enzyme. However, if a gene is hypermethylated, digestion with Msp I will produce more bands of a smaller size than will digestion by Hpa H. In addition, the restriction enzyme Hha I can be used for methylation state assess- ment (Figure 1). This restriction enzyme cleaves the sequence 5'-GCGC-3' only when the internal cytosine residue is unmethylated. Since this recognition site also contains the 5'-CG-3' sequence at which most 5-MC occurs, Hha I will more readily cleave hypomethylated regions of DNA. Because an isoschizomer is not available which cleaves this site when methylated, Hha I analysis can only be used to compare the methylation state of a gene between samples. 18 Me I Msp I 5'-CCGG-3' 5'-CC GG-3' 3'-GGCC-5' 3'-GGC'C-5' Me e Hpa H 5'-CCGG-3' 5'-CC GG—3' 3'-GGCC-5' 3'-GGCC'3 -5' Me Hha I 5'-GCGC-3' 3'-CGCG—5' Figure 1. Recognition sites of the restriction endonucleases Msp I, Hpa H, and Hha I. l9 4. Deoxyribonuclease I Hypersensitivity and Transcriptional Activity of Genes Deoxyribonuclease I (DNase I) is a non-specific endonuclease that splits phosphodiester linkages of DNA, producing 5' phosphate terminated polynucleo tides. Although purified DNA is randomly degraded by this enzyme, transcrip- tionally active regions of chromatin are approximately lO-fold more sensitive to DNase I digestion than are quiescent regions of chromatin (Gross and Garrard, 1988). This property is a consequence of the DNA/RNA/protein interactions which determine the conformation of DNA within chromatin. As can be seen in Figure 2A, chromatin in the region of quiescent genes is normally condensed. However, as a gene becomes prepared for transcription, the structure of chroma- tin becomes altered locally (stage l). Full activation involves commencement of DNA transcription in this area of decondensed chromatin (stage 2). Figure 2B illustrates the fact that both genes actively transcribing RNA and genes with the potential for transcriptional activity are preferentially sensitive to DNase I digestion, supposedly because such regions are more accessible to this enzyme. It has been discovered that the 5' end of genes with the potential for transcription (whether transcribing RNA or not) contains sites which are exqui- sitely semitive to DNase I digestion (Elgin, 1981, 1982; Gross and Garrard, 1988). Such sites are roughly two orders of magnitude more sensitive to DNase I than is bulk chromatin and are thought to represent areas which allow transcription factors (such as RNA polymerase and topoisomerases) to gain access to regulatory (e.g., promoter) regions of genes. These DNase I hypersensitive sites appear to be nucleosome-free and lack histone proteins, a condition which would be expected to promote DNA/transcription factor interactions, as well as DNA/DNase I interactions. However, not all genes possessing a DNase I hypersensitive site are actively engaged in RNA transcription. 20 Figure 2. Schematic diagram of chromatin alterations involved in gene activation and the relationship of these changes to sensitivity of genes to DNase I digestion. (A) Chromatin becomes decondensed as a gene becomes readied for transcription (Stage 1), and this change in conformation becomes more pronounced as the gene is actively transcribed (Stage 2). (B) Due to the decondensed conformation of chromatin in the region of genes with the potential for transcription and those genes actively transcribing DNA, DNase I can more readily degrade DNA in these areas than in highly condensed regions of chromatin containing quiescent genes. (Adapted from Molecular Biologlof the Cell, Garland Publishing Company, New York, 1983.) 21 W STAGE I ALTERS CHIOMATIN STRUCTURE W STAGE 2 TURNS ON GENE sum Figure 2 ZZ Depending on the gene, DNase I hypersensitive sites may be constitutive, inducible, developmental, or tissue-specific (Gross and Garrard, 1988). For example, DNase I hypersensitive sites appear in the metallothionein gene when cells are induced to express this enzyme by cadmium treatment (MacArthur and Lieberman, 1987). Similarly, a site at the 5' end of the 8-major globin gene was observed after dimethylsulfoxide induction (Balcarek and McMorris, 1983). This site appears prior to observed increases in globin mRNA, suggesting that alterations in chromatin are necessary to allow transcription to commence. DNase I hypersensitive sites have been found to be regulated developmentally in the a-fetoprotein gene (Turcotte gt _a_l., 1986), and tissue-specific DNase I hypersensitive sites have been observed in the cardiac myosin light chain gene (Winter and Arnold, 1987). DNase I hypersensitive sites can be induced in the major chicken vitellogenin gene by hormone treatment, but only in specific tissues (Burch and Weintraub, 1983). Again, the appearance of hypersensitive sites in the gene precedes its expression. Few studies have examined oncogenes with regard to DNase I hypersensiti- vity. DNase I hypersensitive sites have been observed 5' to the myc oncogene in I-H.-60 cells (Tuan and London, 1984) and a Burkitt lymphoma (Siebenlist g £11., 1984) expressing myc. In some Burkitt lymphomas, translocation of myc to the immimoglobulin region is associated with the appearance of new DNase I hypersensitive sites, suggesting that sites norm ally involved in the regulation of the immunoglobulin gene may now exert influence over myc transcription (Dyson and Rabbits, 1985). A similar result was obtained in myc-transfected plasma- cytoma cells (Feo gt gL, 1986) using 81 nuclease, an enzyme which closely resembles DNase I in its ability to recognize hypersensitive sites. A 81 nuclease- sensitive site was also detected in the myc oncogene in HL-60 celh upon chemical induction of differentiation, a regimen known to increase the expression of myc 23 (Grosso and Pitot, 1985). Conversely, induction of differentiation in HL-60 cells results in a decrease in transcriptional initiation of myc expression; this phenome- non was accompanied by a concurrent loss of two DNase I hypersensitive sites near the myc promoter (Siebenlist _ei _a_l., 1988). Thus, there is precedent for the association of nuclease hypersensitive sites with oncogene expression. 5. Hypothesis and Experimental Objectives The hypothesis underlying this study consists of two related parts. First, because the liver of the B6C3F1 mouse exhibits a high spontaneous tumor incidence and exceptional sensitivity to chemical induction of tumors, I have hypothesized that certain oncogenes possess a relatively high potential for transcriptional activity in the nascent liver of this animal. The oncogenes examined include Ha-ras and Ki-ras, both of which have been identified as activated oncogenes in B6C3F1 liver tumors, and myc, which has been shown to cooperate with ras oncogenes to effect transformation of normal cells. In addition, because of the heritable nature of hepatoma deveIOpment, the potential for expression of these three oncogenes was assessed in the liver of the two parental strains, C3H/He and CS7BL/6. In this manner, it is possible to compare the potential for expression of each oncogene among the three mouse strains with respect to their spontaneous hepatoma incidence. I predicted that those strains having a high spontaneous tumor incidence (B6C3Fl and C3H/He) possess an elevated potential for Ha-ras, Ki-ras, or myc expression versus the strain displaying a low spontaneous tumor incidence (C 57BL/ 6). Two different parameters of gene expression were examined for each oncogene: the methylation state of the gene and the presence or absence of DNase I hypersensitive sites. Restriction enzyme analysis was used to assess the methylation state of a gene, and an assay for DNase I hypersensitive sites was 24 developed. Hypomethylation and the presence of DNase I hypersensitive sites appear to be necessary, but not sufficient, for gene expression. It follows that a state of relative hypomethylation of and/or DNase I hypersensitive sites in an oncogene in one mouse strain as compared to another strain would be indicative of a higher potential for expression of that gene in the first strain. This increased potential for transcriptional activity may facilitate aberrant gene expression, which may, in turn, contribute to hepatoma development. This approach may identify a heritable factor underlying the differential incidence of spontaneous hepatomas in the B6C3Fl, C3H/He, and C57BL/6 mouse strains. The second part of my hypothesis is that an increased potential for expression is a component of the mechanism by which the Ha-ras, Ki-ras, and myc oncogenes are activated in B6C3F1 mouse liver tumors. Experimental evidence suggests that a point mutation in the Ha-ras oncogene is often present in B6C3F1 mouse liver tumors. However, phenotypic effects of such a mutation cannot occur imless the gene is also expressed; the critical step then becomes deregula- tion of gene expression. Therefore, to determine whether or not the Ha-ras, Ki- ras, and myc oncogenes possess an increased potential for expression in B6C3F1 mouse liver tumors, the methylation state of each gene was assessed by restriction enzyme analysis. It was also important to determine whether or not oncogene activation through an increased potential for expression is qualitatively similar in sponta- neous and chemically-induced BbC3Fl mouse liver tumors. To this end, tumors were chemically-induced in three groups of mice, and spontaneously-arising tumors were harvested from a fourth mouse group. The three carcinogens used included: 1) benzidine, a mutagenic complete carcinogen; 2) chloroform, a non- mutagen which is capable of producing liver tumors in the B6C3Fl mouse; and 3) phenobarbital, a non-mutagenic rodent liver tumor promoter. This approach 25 provided a group of tumors arising from diverse treatments, and thereby allows the determination of whether or not tumor development in the different groups shares common mechanisms of oncogene activation. Lastly, a determination of transcriptional activity of each oncogene was made in control liver as well as in hepatoma tissue. To accomplish this, the amount of messenger RNA for Ha-ras, Iii-ras, and myc was assessed by Northern blot analysis. This portion of the study allows comparisons to be made between the potential for transcription displayed by each gene with the RNA product of transcription. From these experiments, it is believed that the molecular mechanisms underlying hepatoma development in the B6C3F1 mouse, as well as carcinogenesis in general, has been further elucidated. In addition, insight has been gained regarding heritable factors underlying the propensity of the B6C3Fl mouse to deve10p spontaneous and chemically-induced hepatomas. MATERIALS AND METHODS 1. Animals: Maintenance and Carcinogen Treatment Primarily, male B6C3Fl male C3H/He, and female C57BL/6 mice were used in these studies. Where indicated, the opposite sex of these animals (female B6C3F1, female C3H/He, and male C57BL/6 mice) were employed. Young adult mice (18-19 g) were purchased from Charles River Laboratories (Portage, MI), housed in constant temperature and humidity conditions with a 12-hr light/dark cycle, and provided with food and water 113 §b_it_ugg. Male B603F1 mice were used for carcinogen treatment. The benzidine- treated animals were given 120 ppm benzidine in drinking water for one year prior to sacrifice; this treatment resulted in an 85% tumor incidence. For the phenobarbital-treated group, animals received 0.05% (w/v) phenobarbital in drink- ing water for one year and were sacrificed six months later. Tumors were apparent in 68% of these animals. Tumors were induced with chloroform by administering 200 mg/kg of this chemical in a corn oil vehicle by gavage twice weekly for one year, a regimen resulting in a tumor incidence of 80%. A group of animals was allowed 24 months for the development of spontaneous tumors in the absence of any chemical treatment. Partial hepatectomies were performed on one group of control B6C3F1 mice. Methoxyfluorene or ketamine was used for anesthesia, and surgery was performed as described by Higgins and Anderson (1931). “I 27 2. Isolation of DNA: Marmur Method Hepatic nuclei were prepared by the method of Blobel and Potter (1966) as follows. Fresh liver was homogenized in 3 volumes of ice-cold 0.25 M sucrose, 50 mM Tris (pl-I 7.5), 25 mM KCl, 5 mM MgCl2 (STEM) and filtered through cheese cloth. A 10-ml aliquot of homogenate was mixed with 2 ml of 20% Triton X-100, diluted with 8 m1 STKM, and centrifuged at 750 x g for 10 min at 4°C. The resulting nuclear pellet was washed once with the same volumes of STKM and Triton X-100 md once with STKM alone. . High molecular weight DNA was isolated from nuclei by a modification of the method of Marmur (1961). Nuclei from 10 ml homogenate were suspended in 15 ml 10 mM Tris (pH 7.9), 0.1 M NaCl, 5 mM EDTA, 0.5 M NaClO4, and 0.5% sodium dodecyl sulfate and incubated at 37°C for 40 minutes. An equal volume of chloroform:3% isoamyl alcohol was used to deproteinize the suspension, followed by centrifugation at 400 x g for 9 minutes; this wash was repeated once. DNA was precipitated from the upper (aqueous) fraction with 2 volumes of ice-cold 95% ethanol and pelleted by centrifugation at 12,100 x g for 10 minutes at 4°C. The DNA was redissolved in 10 mM Tris (pH 7.9), 5 mM EDTA, and ribonuclease A (Sigma; heat-treated at 80°C for 10 min to destroy DNase activity) was added to a final concentration of 200 pg/ml. After incubation at 37°C for 40 min, the NaCl concentration was adjusted to 0.1 M and 3 mg/ml protease (Sigma; heat-treated at 80°C for 10 minutes) was added. Incubation was continued at 37°C for 90 minutes. An equal volume of chloroform:3% isoamyl alcohol was used to deproteinize the solution, and this was followed by a wash with an equal volume of redistilled phenol. One-half volume of 7.5 M ammonium acetate and 2 volumes 95% ethanol were used to precipitate the DNA, which was stored overnight at -20°C. The DNA was dissolved in 5 ml 10 mM Tris (pH 7.8), 5 mM EDTA, reprecipitated with 0.5 volumes ammonium acetate and 2 volumes 95% ethanol, and again stored at 28 -20°C overnight. The DNA was then dissolved in 250 111-1000 1.31 5 mM Tris (pH 7.8), 0.5 mM EDTA. Absorbance at 260 nm was used to quantitate the concentration of DNA, as one A260 unit equals 50 pg DNA/ml. The A260/A'280 ratio was used to assess its purity, and the ratios obtained were approximately 1.7. (The ratio of a pme DNA solution is approximately 2.0, and this value decreases with increasing protein contamination.) 3. Restriction Enzyme Digestion, Agarose Gel Electrophoresis and Southern Transfer of DNA to Nitrocellulose Paper A. Restriction enzyme digestion Forty microgram aliquots of DNA were digested to completion with 5 units/pg Msp I or Hpa H (BRL, Gaithersburg, MD) at 37°C for 2 hours in a reaction volume of 120 pl. Reaction buffers were supplied as a 10X concentrate by BRL, and the final reaction mixture contained 50 mM Tris (pH 8.0), 10 mM MgCl2 for Msp I, and 20 mM Tris (pH 7.4), 10 mM MgC12 for Hpa H. Following digestion, 10 ul of marker dye (composed of 50% glycerol, 0.25% bromophenol blue, and 0.25% xylene cyanole FF) was added. Samples, along with 2 pg lambda phage Hind IH fragments (BRL, Gaithersburg, MD) diluted to 120 1.11 with running buffer, were heat-treated at 65°C for 10 minutes to dissociate "sticky ends“ formed by restriction enzymes digestion, and cooled on ice. B. Agarose gel electrophoresis ElectrOphoresis was carried out in a model HO/Hl horizontal electro- phoresis apparatus using TBE (89 mM Tris, pH 8.3, 89 mM boric acid, 2.5 mM EDTA) as the running buffer. Wick gels consisting of 1.4% agarose in TBE were poured and covered with TBE to a depth of approximately 2.5 cm. A 0.9% agarose gel was prepared in TBE by heating the solution to 100°C to dissolve the agarose, followed by cooling with stirring to 55-60°C. The edges of the gel support tray were sealed to the wick gels by pipetting liquid agarose into this space and 29 allowing it to solidify. The remainder of the agarose was then poured, the well former was inserted, and the gel was allowed to solidify for at least one hour. After careful removal of the well former, the samples were loaded into the wells using a Pasteur pipette; unused wells were filled with TBE. Electrophoresis was carried out at 40-50 V for 16 hours. The gel was then cut at the ends of the gel support tray and placed in a solution of 0.5 ug/ml ethidium bromide for 15 minutes in order to stain the DNA. The gel was destained in glass distilled water for 10 minutes. The DNA was visualized on a UV light box (Fotodyne, New Berlin, WI) and photographed through a Kodak No. 9 Wrattan gelatin filter at an aperture of 8 for 1 second using Type 667 black and white Polaroid film. A ruler was placed along the gel to assess the distance migrated by the lambda Hind HI fragments. The molecular weights of these fragments, which range from 2.2 to 23.1 kilobases in size, were plotted on the ordinate of semilog paper as a function of distance. C. Southern transfer After being photographed, the gel was soaked in 0.25 N HCl for 10 minutes to fragment the DNA by random depurination; this facilitates the transfer of high molecular weight DNA (Wahl gt gL, 1979). Otherwise, the transfer procedure was performed according to the method of Southern (1975). The gel then was soaked in 1.5 M NaCl, 0.5 M NaOH for 2 hours with occasional agitation in order to denature the DNA. This solution was decanted and replaced with 1 M Tris (pH 8.0), 1.5 mM NaCl. Neutralization was carried out for 2 hours with occasional agitation. The gel was inverted and placed on a piece of 3 MM paper (Whatman) prewet in 10X SSC (1.5 M NaCl, 0.5 M sodium citrate, pH 7.0) with the ends of the 3 MM paper immersed in a tray of 10X SSC to serve as wicks. A piece of nitrocellulose paper (Schleicher and Schuell) was cut to the exact size of the gel and wet in glass distilled water. In the event that the nitrocellulose 30 membrane failed to wet completely, it was covered with boiling water. The membrane then was immersed briefly in 10X SSC before being placed on the gel. Several blotting pads (BRL) and approximately 2 inches of paper towels (both cut to the same size as the gel) were stacked on top of the membrane. A 600 g weight was placed on top of the paper towels, and the entire assembly was covered with plastic wrap to prevent drying of the wick. Transfer of the DNA to the nitrocellulose by capillary action was carried out for 20-24 hours, during which time the wet paper towels were removed several times. At the end of the transfer period, the paper towels and blotting pads were discarded and the nitrocellulose was carefully peeled off the gel and air-dried on 3 MM paper. The DNA was affixed to the membrane by baking in a vacuum oven for 2 hours at 80°C. 4. Amplification and Isolation of Plasmids Containing pRSA 13 and myc Pet 1 Fragment The pRSA 13 clone of serum albumin (Sargent g gL, 1981) and the Pat 1 fragment of myc (Vennstrom gt 9, 1981) had been cloned into the pBR322 plasmid leaving an intact tetracycline resistance gene. The plasmids were amplified and isolated as described by Maniatis _e_t gl_. (1982). Ten milliliters of 2P04’ 0.5% glucose 0.5% Casamino acids, .002% thiamine-HCI, 1 mM MgC12) plus 0-015 M-9 glucose medium (0.6% NazHPO4, 0.3% KH 0.05% NaCl, 0.1% NH4C1, mg/ml tetracycline was innoculated with _E_I. c_ol_i_ strain HB101 containing the pBR322 plasmid. The innoculum was incubated overnight at 37°C with constant shaking; it was sometimes necessary to allow the incubation to proceed for an additional 24 hours. Following incubation, the entire 10 ml were added to 1 liter of M—9 medium, and the mixture was incubated and shaken at 37°C. When the absorbance at 600 nm reached 0.5-0.6, 150 mg of chloramphenicol (Sigma Chemical) was added per liter of culture, and the plasmid was allowed to amplify 31 overnight. Cells were chilled on ice for 5 minutes, then centrifuged at 5,000 rpm, 10 minutes at 5°C. Cells were suspended in 40 ml of 10 mM Tris (pH=8.0) and 1 mM EDTA (washing buffer) and centrifuged at 5000 rpm, 10 minutes, 5°C. Cells were resuspended in 4 ml freshly prepared lysozyme (2 mg/ml), 50 mM glucose, 10 mm EDTA, and 25 mM Tris (pH=8.0) and incubated at 0°C for 30 minutes. Two volumes of alkaline SDS (0.2 N NaOH, 1% SDS) were added and the mixture was incubated at 0°C for 5 additional minutes. Following the addition of 1.5 volumes of 3 M sodium acetate (pH=4.8), the suspension was incubated at 0°C for 1 hour, followed by centrifugation at 15,000 rpm for 20 minutes. The supernatant was precipitated with 2 volumes of ethanol overnight at -20°C followed by centrifuga- tion at 10,000 rpm for 15 minutes. The pellet was resuspended in 15 ml of sterile 10 mM Tris (pH=7.4) and 1 mM EDTA, and 15.8 g cesium chloride and 7.5 mg ethidium bromide was added. The suspension was centrifuged in a type 60 Ti fixed angle rotor at 22°C, 40,000 rpm for 48 hours. The supercoiled plasmid DNA (lower UV-visible band) was removed and the ethidium bromide was extracted with an equal volume of water-saturated butanol until the pink color of the top layer disappeared. The top layer (approximately 8 ml) was dialyzed overnight against 1 liter of 10 mM Tris, 0.1 mM EDTA, and the dialysis buffer was changed once. The dialysate was washed once with redistilled phenol and once with chloroform. The plasmid DNA was precipitated by the addition of 0.1 volume of 20% sodium acetate (pH 5.2) and 2 volumes of 95% ethanol and pelleted by centrifugation at 10,000 rpm for 20 minutes, 5°C. After decanting the supernatant, the pellet was solubilized in 500 pl of 10 mM Tris, 0.1 mM EDTA and stored at 20°C. 5. Nick Translation of the Ha-ras BS-9 Probe The BS-9 clone of Ha-ras (Ellis gt _a_l., 1981) is a 450 base clone correspond- ing to the 5' region of the gene, commencing approximately 50 base pairs 32 upstream from the N-terminus of the p21 protein (Dahr g gl_., 1982). This probe was 32P-labelled using a nick translation kit purchased from BRL. On ice, 1 ug BS-9-containing plasmid DNA was mixed with 8 ul each a-3zP-labelled dATP, dCTP, dGTP, and dTTP (New England Nuclear; 10 mCi/ml; 800 Ci/mmol) and glass distilled water to a volume of 45 111. After a brief mixing, 5 ul of a 1011 reaction buffer containing DNA polymerase I and DNase I was added. Final composition of the reaction solution was 2.08 1.1M in each of the four radiolabelled nucleotides, 0.04 U/pl DNA polymerase I, 4 pg/vl DNase I, 5 mM Tris (pH 7.5), 0.5 mM magnesium acetate, 0.1 mM 2-mercaptoethanol, 0.01 mM PMSF (phenylmethyl- sulfonyl fluoride), 10 11 g/ml bovine serum albumin and 5% glycerol. Nick translation was carried out at 15°C for one hour. The reaction vial then was placed on ice, and 38 ul STE and 12 ul of a bromophenol blue solution was added (final volume = 100 pl). Separation of newly labelled probe from unincorporated nucleotides was accomplished by spun column chromatography (Maniatis gt gL, 1982). Columns were prepared by packing Sephadex G-50 Fine in a 1-ml syringe plugged with glass wool. To this end, Sephadex was equilibrated in STE (10 mM Tris, pH 8.0, 1 mM EDTA, 0.1 M NaCl), and transferred to the column with a Pasteur pipette. The column was placed in a 15-ml Corex tube and centrifuged (IEC Centre 7-R) at 2000 rpm for 8 minutes. Sephadex was added to the column and recentrifuged until the post-centrifugation volume reached 0.9 ml. Columns were washed at least twice with 100 pl STE before use. An Eppendorf tube (cap removed) was placed in a 15-ml Corex tube, and the column was placed in the tube such that the tip was within the Eppendorf tube. The entire 100 pl nick translation mixture was applied to the top of the column, which was centrifuged at 2000 rpm for 8 minutes. The column, which retains unincorporated nucleotides along with bromophenol blue, was discarded. The 33 labelled DNA contained in the Eppendorf tube was diluted with 100 pl STE, and 2 ul aliquots were counted in a Packard Tricarb Model 460C scintillation counter using Safety Solve (RPI, Mount Prospect, IL) scintillation cocktail. 6. Hybridization of DNA Affixed to Nitrocellulose Membranes Membranes were wet in a minimum volume of hybridization buffer (50% formamide, 10 mM Hepes (pH 7.4), 3X SSC, 1 mg/ml yeast tRNA, 1X Denhardts buffer (0.02% each Ficoll 400, polyvinyl pyrollidone, and bovine serum albumin), and 100 ug/ml denatured salmon sperm DNA). After wrapping in plastic wrap, 32 the membranes were prehybridized at 42°C for approximately 16 hours. The P- labelled probe (prepared by nick translation) was denatured by boiling for 5 minutes in a water bath and 106 cpm were added to a small amount of hybridization buffer. The probe solution was spread on plastic wrap, and the nitrocellulose membrane was placed on it DNA side down. The membrane was rewrapped with plastic wrap and massaged to spread the probe evenly across the membrane. Hybridization was carried out at 42°C for 48 hours. Membranes were washed 4 times with 2X SSC-0.1% SDS at room tempera- ture for 5 minutes each, twice with 1X SSC-0.1% SDS at 55°C for 10-15 minutes each and, finally, 4 times with 0.1x SSC-0.1% SDS at 55°C for 10-15 minutes each. Membranes were allowed to air dry and wrapped in plastic wrap for autoradiography. 7. Simultaneous Isolation of DNA and RNA: CsCl Method DNA and RNA were isolated simultaneously by an adaptation of the method described by Chirgwin _e_t gl__. (1979). For this procedure, livers were excised and immediately frozen in liquid nitrogen. Tissue was either used immediately or stored at -85°C. Frozen tissue was ground to a fine powder in liquid nitrogen with 34 a mortar and pestle. The frozen powder was then stirred into at least 24 volumes of 4.0 M guanidinium isothiocyanate, 0.5% sarcosyl, 1 M sodium citrate (pH 7.0), and 0.5% 8 emercaptoethanol until a particle-free viscous liquid was produced. This was then layered onto a CsCl step gradient (3 ml 5.7 M CsCl, 2 ml 3.0 M CsCl; CsCl made in 0.1 M EDTA, treated overnight with 0.1% diethylpyrocarbo- nate, and autoclaved) and ultracentrifuged at 29,000 rpm for 22 hours at 20°C in a Beckman SW41 rotor. At the end of the centrifuge run, the upper guanidium- containing layer was removed and discarded. A. DNA isolation The DNA, which migrates to just below the 5.7 M/3.0 M CsCl inter- face, was removed with a wide-mouth pipette. RNase A (Sigma; preboiled 10 min) was added to the DNA-containing solution to a final concentration of 50 ug/ml and the DNA was dialyzed against 1 liter of 10 mM Tris/1 mM EDTA/0.1% SDS (pH 7.5) for one hour at room temperature. Proteinase K (BMB, Indianapolis, N) was added to a final concentration of 50 ug/ml, the dialysis buffer was changed, and dialysis was continued for 1 hour. The DNA was then dialyzed against 2 liters of 10 mM Tris/1 mM EDTA (pH 7.5) at 4°C for 16-20 hours. An equal vohime of ultra-pure phenol (equilibrated with 0.5 M Tris, pH 8.0 until the aqueous layer reached pH 7.6) was used to extract proteins, and this was followed by a wash with an equal volume of chloroform:3% isoamyl alcohol. The DNA was then precipitated with 0.5 volumes of 7.5 M NH 4OAc and 2 volumes of ethanol. After DNA was dissolved in TE (10 mM Tris/l mM EDTA, pH 8.0), it was reprecipitated with NH 4OAc and ethanol and again dissolved in TE. DNA was quantitated by the A260 absorbance, and the A260/A280 ratio routinely approached 1.8. B. RNA isolation After removal of the DNA fraction, the remaining CsCl solution was decanted by rapid inversion of the centrifuge tube, and the tube was allowed to 35 drain for several minutes. A razor blade was used to cut the tube approximately 1.5 cm from the end, thus forming a 'cup", the bottom of which contains the RNA pellet. The RNA was dissolved in diethyl pyrocarbonate-treated water (RNA from. 1 g. of liver in 500 111 water), and washed twice with an equal volume of chloroform :butanol (4:1). The phases were separated by centrifugation in a Brinkman microfuge at 14,000 rpm for 1 minute. The top, RNA-containing layer was removed to an RNase-free lS-ml Corex centrifuge tube. One-tenth volume 2.0 M potassium acetate (pH 5.2; DEPC-treated and autoclaved) and 2.5 volumes 100% ethanol were used to precipitate the RNA, which was then stored at -20°C. 8. Isolation of DNase I-Treated DNA The methods of Burch and Weintraub (1983) were followed for preparation of nuclei, DNase I treatment, and DNA isolation. Preliminary studies identified 15°C as the optimum temperature for DNase I digestion as the control sample (no added DNase D remains viscous after incubation at this temwature for 10 minutes. Fresh mouse liver (approximately 1 g) was chopped finely in ice-cold SSCT (1X SSC, 10 mM Tris, pH 7.4). The liver pieces were washed several times with ice-cold SSCT to remove contaminating blood elements, and briefly (5 seconds) centrifuged at 1500 rpm in an IEC Centre-7R centrifuge at 4°C to pellet the tissue. After decanting the supernatant, the liver was homogenized in 10 ml RSB (10 mM Tris, pH 7.4, 10 mM NaCl, and 3 mM MgC12) plus 0.5% Nonidet P-40 (NP- 40) and 1 mM PMSF. Nuclei were pelleted by centrifugation at 1500 rpm for 10 minutes at 4°C (IEC Centre-7R centrifuge) and washed twice in RSB plus 0.5% NP-40. Nuclei from the livers of 5-6 mice were pooled in 7-ml RSB plus 0.1 mM CaClz. One milliliter aliquots of the nuclei suspension were added to tubes containing 0 (2 tubes), 28, 33, 38, 43, or 47 units DNase L DNase I-treated 36 samples were incubated at 15°C for 10 minutes; one control tube was incubated at 0°C, and one control tube was incubated at 15°C to assess intrinsic nuclease activity. The reaction was halted by the addition of an equal volume of stop buffer (1% SDS, 0.6 M NaCl, 20 mM Tris, pH 7.5, 10 mM EDTA, and 400 ug/ml proteinase K). Samples were incubated for 2 hours at 37°C and diluted to a volume of 4-5 ml for deproteination. A wash with an equal volume of phenol (neutral pH) was followed by a wash with an equal vohime of chloroform:3% isoamyl alcohol; separation of the phases was accomplished by centrifugation at 400 x g for 15 minutes for the first wash and 10 minutes for the second wash. DNA was preciptiated by the addition of 0.5 volumes 7.5 M ammonium acetate and 2 volumes ethanol. The DNA was rinsed in 70% ethanol, allowed to dry briefly, and dissolved in 1 ml TE (10 mM Tris, pH 8.0, 1 mM EDTA). DNA concentration was assessed by absorbance at 260 nm. To visualize the extent of DNase I digestion, 1 ug samples were diluted to a volume of 20 111 and 4 pl of 5X marker dye (15% Ficoll 400, 0.125% bromophenol blue, 0.125% xylene cyanode ff, and 5X TBE) was added. DNA was loaded onto a submerged 0.9% agarose baby gel in TBE (model H6, BRL) and electrophoresed for 2 hours at 30 V. The gel was stained with 1 ug/ml ethidium bromide for 15-30 minutes, destained overnight, and photographed as described in Section 3B. 9. Restriction Enzyme Digestion, Electrophoresis, and Southern Transfer of DNA to Gene Screen Plus A. Restriction enzyme digestion Restriction enzyme digestion of DNA was performed as described in Section 3A with the following modifications. Twenty microgram aliquots of DNA were digested with restriction enzyme in a reaction volume of 60 111. A Ficoll- based marker dye was used (5X = 15% Ficoll 400, 0.125% bromophenol blue, 0.125% xylene cyanole ff, and 5X TBE). 37 Reaction buffers for Msp I and Hpa H were as described in Section 3A. For the other enzymes used, the reaction buffers (all supplied as a 10X concentrate by BRL) were as follows: Hha I, Hind HI, and Pst I - 5 mM Tris (pH 8.0), 1 mM MgC12, and 5 mM NaCl; Eco R1 - 5 mM Tris (pH 8.0), 1 mM MgC12, and 10 mM NaCl. DNA was digested with 5 units Hha I per ug DNA, and 3 units HindIH,PstI,orEco R1 per ug DNA. B. Agarose gel electrophoresis Preparation of the agarose gel was performed as described in Section 3B with the following modifications. A 0.9% agarose gel (0.8% where indicated) in 300 ml TBE was prepared and poured in a gel support tray which had the ends taped to contain the agarose. After solidification and tape removal, the gel was placed in a H0/Hl electrophoresis unit and covered with TBE to a depth of approximately 5 mm. Samples were loaded into the wells using a Pipetman (Rainin). Electrophoresis, ethidium bromide staining, and photogrq>hy were performed exactly was described in Section 3B. C. Southern transfer Transfer of DNA to Gene Screen Plus was performed as suggested by the manufacturer (NEN/DuPont). All gel soaks were performed with constant agitation on a Orbit Shaker (LabLine). DNA was depurinated in 0.25 N HCl for 10 minutes and denatured in 0.4 N NaOH, 0.6 M NaCl for 30 minutes. The gel then was neutralized in 1.5 M NaCl, 0.5 M Tris (pH 7.5) for 30 minutes. The transfer was set up as described in Section 3C except that 5-6 pieces of 3 MM paper were used in place of blotting pads. At the end of the transfer, the Gene Screen Plus membrane was peeled off the gel and briefly immersed in 0.4 N NaOH to denature the DNA; it was then 38 neutralized by dipping in 0.2 M Tris, 2X SSC (pH 7.5). The membrane was allowed to air dry, and, in some cases, it was baked at 80°C for 2 hours in a vacuum oven. 10. Labelling of Probes with 32 P: Random Primers Method Probes for Ha-ras, Ki-ras and myc were purchased from Oncor (Gaithers- burg, MD). The Oncor Ha-ras probe differs from the BS-9 clone in that it is approximately 66% larger, extending about 300 base pairs further in the 3' direction than does the BS-9 clone. The 1.5 kb Pst I fragment of myc (Vennstrom gt gl., 1981) and the pRSA 13 clone of serum albumin (Sargent e_t gl_., 1981) were both utilized as inserts in pBR322. A random primers labelling kit was purchased from BRL. For each reaction, 25 ng of DNA was denatured by boiling for 5 minutes and immediately placed on ice. dATP, dGTP, and dTTP were added to a final concentration of 0.02 mM each, along with a 3.3x buffer concentrate (containing 18 A260 units/ml oligodeoxyribonucleotide hexamers, 0.67 M Hepes, 0.17 M Tris, 17 mM MgC12, 33 mM 2-mercaptoethanol, and 1.33 mg/ml BSA, pH 6.8), and cow to 49 ul. Approximately so vCi of 21-32 P dCTP (10 mCi/ml; 3000 Ci/mmol) was added and mixed briefly prior to the addition of 3 units of the Klenow fragment of DNA polymerase I. Labelling was performed at 25°C for one hour. The volume was increased to 100 111 by the addition of 38 ul STE and 12 pl of a bromophenol blue solution. Separation of newly labelled probe from unincorporated nucleotides was accomplished by spun column chromatography as described in Section 5. 11. Hybridization of DNA Affixed to Gene Screen Plus with 3ZP—Labelled Probes Prehybridization and hybridization were carried out in Scotch-pak (Kapak) heat-sealable bags placed in a 65°C water bath with constant agitation (as 39 recommended by NEN/DuPont for Gene Screen Plus). Membranes were prehybri- dized for at least 15 min with approximately 50 ill/cmz of a solution containing 10% dextran sulfate, 1% SDS, and 1 M NaCl, preheated to 65°C. This solution was sometimes buffered with 50 mM Tris, pH 7.5. Denatured probe (1x105 cpm per ml hybridization solution; prepared by random primers method) and sonicated salmon sperm DNA (100 vg/ml) in 500 pl STE were then added and mixed thoroughly with the prehybridization solution by pulling the bag back and forth over an edge. Hybridization was carried out for approximately 16 hours. Membranes were washed twice with 2X SSC for 5 minutes at room temperature, twice with 2X SSC/1% SDS for 15 minutes each at 65°C, and twice with 0.1x SSC for 15 min each at room temperature (250 ml solution per wash), all with constant agitation. Membranes were then blotted with 3 mm paper (Whatman) and wrapped in Handi- Wrap (Dow Chemical Co.) in such a manner as to prevent drying. Membranes that were to be rehybridized were stripped of probe after autoadiography (Section 15) by pouring boiling 10 mM Tris/1 mM EDTA/ 1% SDS, pH 7.5 over them and agitating until cool; this procedure was then repeated. Autoradiography was performed on the stripped membranes to ensure complete probe removal prior to rehybridization, which was performed as described above. 12. Electrophoresis and Northern Transfer of RNA A. Preparation of RNA Water used in the preparation of RNA was treated with 0.1% DEPC (diethyl pyrocarbonate), allowed to sit overnight and autoclaved. Concentrated running buffer was autoclaved before use. RNA preparation and electrophoresis was performed according to the method suggested by NEN/DuPont. RNA was pelleted by centrifugation at 17,600 x g for 20 minutes at -10°C. Pellets were washed with 1 ml absolute ethanol at room temperature and recentrifuged as 40 above. RNA was dried under a stream of ethanol and dissolved in 100-250 pl water. The concentration of RNA was quantitated by absorbance at 260 nm (1 A260 unit = 40 pg RNA/ml). RNA was then diluted to a concentration of 4.4 pg/pl. A 3.7-pl aliquot was removed to an Eppendorf tube and stored at -85°C until assessment of RNA integrity was performed (Section 14). A 4.5-pl aliquot of RNA (20 pg) was denatured in a solution of 1X MOPS (5X MOPS = 0.2 M morpholinopropanesulfonic acid, 50 mM sodium acetate and 5 mM EDTA, pH 7.0), 2.2 M formaldehyde and 50% deionized formamide in a volume of 20 pl. Samples were incubated at 60°C for 10 minutes, and 5 pl of sterile loading buffer (50% glycerol, 1 mM EDTA, 0.4% bromophenol blue and 0.4% xylene cyanode ff) was added. B. Electrophoresis and Northern transfer RNA samples were loaded onto a 1.0% agarose gel prepared in 2.2 M formaldehyde and 1X MOPS. Gels were submerged in 1X MOPS, and electropho- resis was carried out at 30 V for 18-20 hours. Prior to transfer of RNA to Gene Screen Plus, the gel was rinsed 5 times in an equal volume of glass distilled water (1 minute per wash). The transfer was set up exactly as described in Section 3C, except that blotting pods were omitted; transfer was carried out for approximately 22 hours. After the membrane was removed from the gel, it was rinsed in 2X SSPE (0.3 M NaCl, 20 mM NaHzPO and 2 mM EDTA, pH 7.4), allowed to air dry, and baked at 80°C in 4! a vacuum oven. 13. Hybridization of Northern Blots Northern blots were hybridized in heat-sealable plastic bags (Scotch-pak, Kapak Corp.) according to the method described by NEN/DuPont. Membranes 41 were prehybridized in a minimum volume (approximately 50 pl/cmz) of prehybri- dization solution (5X SSPE, 50% deionized formamide, 5X Denhardt's solution, 10% dextran sulfate and 1% SDS) for 2-4 hours in a 42°C water bath with constant agitation. To hybridize, 4x105 cpm/ml denatured, 32'P-labelled probe (random primers method) was added ad the bag was resealed. Mixing of the probe into the prehybridization solution was accomplished by pulling the bag back and forth over an edge several times. Hybridization was carried out for 16-18 hours at 42°C with constant agitation. Membranes were washed twice with 250 ml 2X SSPE at room temperature, twice with 400 ml 2X SSPE/2% SDS at 65°C, and twice with 250 ml 0.1x SSPE at room temperature; all washes were 15 minutes in duration. After blotting on 3 MM paper, membranes were wrapped in Handi-Wrap in preparation for autoradio- graphy (Section 15). The resulting autoradiograph was scanned with an LKB UltroScan XL densitometer to assess the relative degree of hybridization of probe to RNA samples. The resulting peaks were cut out and weighed to quantitate this relationship. 14. Assessment of RNA Integrity RNA samples were thawed and denatured in a mixture of 50% DMSO, 0.01 M NaHzPO4 (pH 7.0), and 6.8% deionized glyoxal at 50°C for 1 hour in a volume of 16 pl as described in Maniatis (1982). Four microliters of RNA marker dye were added, and a 13 pl (13 pg) aliquot of each sample was loaded onto a 1% agarose gel prepared and submerged in 0.01 M NaHZPO4 (pH 7.0). Electrophoresis was carried out at 30 V for 2 hours; running buffer was changed every 30 minutes. Gels were stained with 1 pg/ml ethidium bromide for 15 minutes and destained 42 overnight in glass distilled water. Photography of the gel was performed as described in Section BB. 1 5. Autoradiography The Handi-Wrap-wrwped, hybridized membrane was taped to a Cronex Lightning Plus intensifying screen in a film holder. hi complete darkness, a sheet of X-OMAT AR film was placed over the membrane, and another intensifying screen was placed over the film before the film holder was closed. The film holder was then wrapped in aluminum foil and clipped between two clipboards. Autoradiogrqahy was carried out at -85°C for 1-7 days. Prior to film development, the autoradiographic apparatus was allowed to equilibrate to room temprature. In complete darkness, the film was removed from the film holder, and immersed in GBX developer (Kodak) for 5-6 minutes. The film was then rinsed in water for one mimlte, and immersed in GBX fixer (Kodak) for 5-6 minutes before being rinsed in water and allowed to dry. RESULTS 1. Methylation State of the Ha-ras, Ki-ras, and myc Oncogenes in B6C3F1, C3H/He, and C57BL/6 Mouse Liver The degree of methylation of Ha-ras in male B6C3Fl, male C3H/He, and female C57BL/6 mouse liver was assessed by comparison of the restriction patterns obtained following digestion with either Msp I or Hpa H. If the patterns were similar the gene examined was considered relatively hypomethylated. If the patterns were dissimilar (resistant to digestion by Hpa ID, the gene was considered relatively hypermethylated. In some cases, an intermediate degree of methyla- tion was noted. The methylation state of the Ha-ras oncogene was assessed using two different probes. The BS-9 clone, described by Ellis gt g1. (1981), covers approximately 450 base pairs toward the 5' end of the Ha-ras gene. The Oncor clone is approximately 66% larger, extending approximately 300 base pairs in the 3' direction. The results of hybridizing Msp I- and Hpa H-digested DNA with the BS-9 probe for the Ha-ras oncogene are shown in Figure 3. When B6C3F1 mouse hepatic DNA is digested with Msp I (lane 1), a number of bands are produced, ranging in size from < 1.9 kilbases (kb) to 8.8 kb. Hpa H digestion of the same DNA (lane 2) produces some fragments of comparable size (i.e., 1.9 kb, 2.7 kb, 8.8 kb). In addition, hybridization is clearly evident at a molecular weight of 23.1 kb. These results indicate that the Ha-ras oncogene is methylated to an intermediate degree in the liver of the B6C3Fl mouse. Digestion of C3H/He hepatic DNA with Msp 1 (lane 3) and Hpa H (lane 4) produces very similar restriction maps, with A“ k_b 23.1 - Figure 3. Methylation state of the Ha-ras oncogene in B6C3F1, C3H/He, and C57BL/6 mouse liver using the BS-9 clone. DNA was isolated by the method of Marmur from B6C3F1 (lanes 1 and 2), C3H/He (lanes 3 and 4), C57BL/6 (lanes 5 and 6) mouse liver and digested with Msp I (M) or Hpa H (H). The resulting fragments were electrophoresed, and transferred to nitrocellulose paper. The DNA was then hybridized to the 2'P-labelled BS-9 probe for the Ha-ras oncogene, and regions of homology were visualized by autoradiography. 45 homologous bands occurring at 23.1 kb, 5.8 kb, and two sizes < 2.4 kb. It is thus evident that the Ha-ras oncogene is hypomethylated in the liver of the C3H/He mouse as compared to the B6C3F1 mouse. hi contrast, hepatic DNA from the C57BL/6 mouse shows a different restriction pattern when digested by Msp I (lane 5) as compared to Hpa H (lane 6). Msp I digestion produces three major bands of a relatively low molecular weight, whereas Hpa H digestion produces several bands, two at relatively high molecular weights (23.1 kb and 5.8 kb), and only one corresponding to a band produced by Msp I digestion (2.4 kb). Because Msp I cleaves the Ila-ras oncogene in CS7BL/6 liver DNA to much greater extent than does Hpa H, it is concluded that the Ha-ras oncogene is hypermethylated in the liver of this mouse. The results of restriction enzyme analysis performed on a number of animals showed a consistent trend (Table 1). Ha-ras was either hypomethylated or methylated to an intermediate degree in 4/4 B6C3F1 mice and 5/7 CBH/He mice. In C57BL/6 mice Ha-ras was hypermethylated in 4 animals and methylated to an intermediate degree in 2 animals. The results of experiments using the Oncor probe for methylation assess- ment are presented in Figure 4. In young adult animals, there is one difference in the methylation state of firms among the three strains. An additional 15 kb band is apparent in Msp I-digested C57BL/6 DNA (lanes 9 and 11; see arrow), but not in Msp I-digested DNA from B6C3F1 and C3H/He mice. This band is probably due to the relatively rare occurrence of a methyl group on the external cytosine residue of a 5'-CCGG-3' recognition site, thus preventing Msp I from completely cleaving this fragment. Because male mice of the B6C3Fl and C3H/He strains and female C57BL/6 mice were used for this experiment, it is possible that the site-specific methylation of the Ha-ras gene seen in C57BL/6 mouse liver results from sex differences. To address this possibility, hepatic DNA from female 46 TABLE 1 METHYLATION STATE OF THE Ha-ras ONCOGENE IN THREE MOUSE STRAH‘JS Hypo- Intermediate Hyper- methylation Methylation methylation B6C3Fl 2"I 2 0 C3H/He 2 3 2 C57BL/ 6 0 2 4 I"Numbers refer to the number of individual animals displaying the indicated level of methylation. 47 l 2 3 4 5 6 7 8 9 1011 12 Figure 4. Methylation pattern of the Ha-ras oncogene in control mice. Hepatic DNA was isolated from male B6C3F1 (lanes 1-4), male C3H/He (lanes 5-8), and female C57BL/6 (lanes 9-12) mice on a CsCl gradient, digested with Msp I (M) and Hpa H (H), and electrOphoresed in a 0. 9 agarose gel. DNA was then transferred to Gene Screen Plus, hybridized to a P-labelled probe for the Ha-ras oncogene (Oncor), and visualized by autoradiogrmhy. The arrow points to a 15 kb band present in Msp I-digested C57BL/6 DNA but absent in B6C3F1 and C3H/He DNA. 48 B6C3F1, female C3H/He, and male C57BL/6 (opposite sex) was subjected to restriction enzyme assessment of Ha-ras methylation state. As seen in Figure 5, a 15-kb band was observed in Msp I-digested C57BL/6 DNA, but not in Map I- digested DNA from the other two strains, when probed with Ha-ras (see arrow). This result confirms the finding that the Ha-ras oncogene is hypomethylated in a site-specific manner in B6C3F1 and C3H/He mouse liver as compared to C57BL/6 mouse liver. In other respects, the Ha-ras oncogene appears to be relatively hyper- methylated in liver of mice of all three strains (Figures 4 and 5). This conclusion is based on the fact that Msp I digestion (M) is extensive, producing a number of bands ranging from a few hundred base pairs to ~9 kb, whereas the bulk of hybridization to Hpa H-digested DNA (H) occurs at very large (:23 kb) fragments, with 2 bands occurring at smaller molecular weights. The importance of complete transfer of high molecular weight DNA in the determination of gene methylation state by Msp I/Hpa H analysis is illustrated in Figure 6. Although DNA samples used in Figures 4 and 6 were digested, electrophoresed, and transferred to nylon membranes concurrently, the depurina- tion step of gel processing was omitted in the experiment shown in Figure 6. A high degree of hybridization at the region _>_ 23 kb is seen in Hpa H-digested DNA that underwent depurination (Figure 4, even-numbered lanes), whereas little hybridization is seen in this region when depurination was omitted (Figure 6, even- numbered lanes). It must be noted that the 15 kb band seen in Msp I-digested C57BL/6 DNA is not apparent in Figure 6. Thus, in order to assess accurately the methylation state of genes using restriction enzyme analysis, efficient transfer of large DNA fragments must be assured through inclusion of a depurination step. The methylation state of the Ki-ras (Figm-e 7) and myc (Figure 8) oncogenes was also assessed in the three mouse strains. Comparison of the restriction 49 B6C3F1(female) c3H/He(fema1e) C57BL/6(lna1e) 1 2 3 4 5 6 7 8 9101112 NM a s Oh! Figure 5. Methylation pattern of the Ha-ras oncogene in mice of the opposite sex. Hepatic DNA from female B6C3F1 (lanes 1-4), female C3H/He (lanes 5-8) and male C57BL/6 (lanes 9-12) was isolated and treated as described in Figure 4. The arrow points to a 15 kb band present in Msp I-digested C57BL/6 DNA but absent in B6C3F1 and C3H/I-1e DNA. 50 Figure 6. Effect of the omission of the depurination step on the transfer of large DNA fragments. DNA samples were Msp I (M)- or Hpa H (H)-digested, electro- phoresed, and transferred to Gene Screen Plus concurrently with the samples shown in Figure 4, except that the HCl soak prior to Southern transfer was omitted. DNA was derived from male B6C3F1 (lanes 1-4), male C3H/He (lanes 5- 8), and female C57BL/6 (lanes 9-12) mouse liver. 51 l 2 3 1+ 3 6 7 8 9 1‘) ll 12 23.1 — 9.4 — 6.6 — 4.4 - 2.3 2.0 - Figure 7. Methylation status of the Ki-ras oncogene in control mice. The fiembrane shown in Figure 2 was stripped of the Ila-ras probe and rehybridized to P-labelled probe for Ki-ras. Autoradiography shows the methylation pattern of the Ki-ras oncogene to be similar in all three mouse strains. 52 B6C3Fl CBH/He C57BL/6 l 2 3 4 7 my kb 23.1 ."s. " Figure 8. Methylation status of the myc oncogene in control mice. The m brane shown in Figure 7 was stripped of the Ki-ras probe and, rehybridized to P-labelled probe (Oncor) for myc. Autoradiography shows the methylation pattern of myc to be similar in all three mouse strains. 53 patterns produced by Msp I digestion of male B6C3FI, male C311, and female CS7BL/6 hepatic DNA shows no differences for Ki-ras (Figure 7, lanes 1, 3, 5, 7, 9, and 11) or myc (Figure 8, lanes 1, 3, 5, 7, 9, and ll); the same is true for Hpa II- digested DNA hybridized to Ki-ras (Figure 7, lanes 2., 4, 6, 8, 10, and 12) and myc (Figure 8, lanes 2., 4, 6, 8, 10, and 12). These results indicate that no differences exist‘between the three mouse strains with regard to the methylation state of the Iii-ras and myc oncogenes. Furthermore, the increased extent of digestion by Msp I versus Hpa II seen in both figures indicates that Ki-ras and myc are hypermethylated in male B6C3F1, male CSH/He, and female CS7BL/6 mouse liver. Msp I/Hpa II analysis using hepatic DNA from mice of the opposite sex mice (female B6C3F1, female C3H/I-Ie, and male CS7BL/6) produced an identical result for Ki-ras (Figure 9) and myc (Figure 10), indicating that these two oncogenes are hypermethylated to a similar extent in the liver of mice of the opposite sex. Within each strain, Msp I and Hpa II restriction patterns were observed to be very consistent when probed with Ila-ras; this was also true for Ki- ras and myc. The serum albumin gene is known to be expressed in normal liver. There- fore, this gene was used as a positive control for restriction enzyme analysis of gene methylation state. Figure 11 shows that Msp I and Hpa II digestion of 36C3F1 (lanes 1-4), C3H/He (lanes 5-8) and C57BL/6 (lanes 9—12) DNA produces very similar restriction patterns when probed with the serum albumin clone pRSA 13 as evidenced by a major band at 6.0 kb in Msp I-digested samples (lanes 1, 3, 5, 7, 9, and 11) and 6.4 kb in Hpa II-digested samples (lanes 2, 4, 6, 8, 10, and 12). A similar result was seen in DNA from mice of the opposite sex (data not shown). These results not only indicate the expected result of hypomethylation of this gene, but also provide evidence that the conditions of restriction enzyme reactions were appropriate for complete digestion. 54 B6C3F1 C3H/He C57BL/6 female female male 23.1 Figure 9. Methylation status of the Ki-ras oncogene in mice of the opposite sex. The membsgne shown in Figure 5 was stripped of the Ha-ras probe and rehybri- dized to a P-labelled probe for Ki-ras. 55 gases—mesa u «on: o»:— uou ocean cosmos—um c. 632% 05.5808 2:. J3: onion—mo one if, NH fiH as m a mans o\qmamo wm o 3 unease—€59. use 030.:— eoung 05 no coma—tau ass a 0.5m...“ no.8 o. ouomouoo or: 2: «6 353a sagas—>502 .3 9:63 . y m H.nN b o m e n N H oaoamu oaoeou m=\=no apnoea 56 4} p I Figure 11. Methylation status of the serum albumin gene in control mice. The fiembrane shown in Figure 8 was stripped of the myc probe and rehybridized to a P-labelled probe for serum albumin. 57 2. Oncogene Methylation State in B6C3F1 Mouse Liver Tumors Figure 12 shows the results of hybridizing the probe for Ila-ras to benzidine- induced tumor (lanes 3, 4, 7, 8) and adjacent non-tumor (lanes .1, 2, 5, 6) DNA that has been digested with Msp I and Hpa II. A striking difference in the restriction pattern is obvious between Hpa II-digested tumor and non-tumor DNA. Hpa II digestion of tumor DNA produced intense bands at 4.4, 3.2, and 2.6 kb (lanes 4 and 8; see arrows) which are not apparent in non-tumor DNA (lanes 2 and 6). Furthermore, the smaller two of these three bands correspond to bands produced by Msp I digestion of tumor and non-tumor DNA, indicating that these bands result from a loss of methylation at 55063436? sites. This result was observed in all 4 tumor/non~tumor pairs of samples examined from the benzidine-treated group of mice. Note that DNA derived from adjacent non-tumor liver tissues (lanes 1, 2, 5, and 6) produces restriction patterns nearly identical to those seen in control (young adult) animals (see Figure 4, lanes 1-4). Because Southern transfer of DNA from the benzidine treatment group was carried out without depurination, this experiment was repeated for Ila-ras with inclusion of the HCl treatment (Figure 13). Again, bands are seen in Hpa Il-digested tumor DNA (lanes 4 and 8) which are not apparent in non-tumor DNA (lanes 2 and 6). Therefore, the results of experiments presented in Figures 12 and 13 are qualitatively similar, despite the lack of high molecular weight DNA transferred in the former. A similar outcome for Ila-ras was produced by Msp I/Iipa II analysis of tumor DNA from the other three treatment groups. DNA derived from young adult male 36C3F1 mouse liver was used for controls in these experiments. The restriction patterns produced by Msp I and Hpa II digestion of control DNA were essentially identical to those seen in Figure 4 (lanes 1-4) in all cases. Hpa II digestion of phenobarbital-induced tumor DNA (Figure 14, lanes 4 and 6), chloroform-induced tumor DNA (Figure 14, lanes 8, 10, 12, and 14), and 58 Figure 12. Methylation status of the I-Ia-ras oncogene in benzidine-induced hepatic tumors and adjacent non-tumor tissue. DNA was isolated from non-tumor hepatic tissue (lanes 1, 2, 5 and 6) or adjacent tumors (lanes 3, 4, 7 and 8) on a CsCl gradient and digested with Msp I (M) or Hpa 11 (H). Electrophoresis and Southern transfer to Gene Screen Plus were performed as described in Methods, except the depurination step was omitted. Lanes l-4 represent DNA isolated from one animal, and lanes 5-8 represent DNA isolated from a second animal. Note the increased intensity of the bands marked by arrows in the Hpa I-digested tumor DNA (lanes 4 and 8) as compared to non-tumor DNA (lanes 2 and 6). 59 Figure 13. Lack of effect of depurination on the methylation assessment of Ha- ras in benzidine-induced tumor and non-tumor tissue. Tumor (T) and non-tumor (N)-derived DNA was digested with Msp I (M) or Hpa 11 (H), electrophoresed and transferred to Gene Screen Plus as described in Methods with inclusion of the depurinaton step. Note that extra bands are again observed in Hpa III-digested tumor DNA (lanes 4 and 8) as compared to non-tumor DNA (lanes 2 and 6). 60 AM one: <29 .oficoo peamommprfl an: E «comma «as one £033 A: can .2 A: .m .c .v moss: 502 .5 ensur— 65 2 «Ema ca nu ma HA OH 0 Q h w m v m WI L -\ ZKOLOKOJIU A<9Hmm<002mzm JOKPZOU 66 .8 an»: 5.5 35:8 c3393“ 2.: a. as: .2 as .m .o .v and: 502 .3 9:53 . h‘ 3' I- 1?}, ‘u . ,. g ’ w)“ "f. ".I- . “ rel try, 4‘5, | ~ Mn" to 5%, .‘. 1Q , ,s' x c. IR"..- - —. 32")" ‘.: I‘ ‘ v... x 0.0 - 7:... To mi H.n~ '1... D _ fl anHNHHdOHmmhonvn.NH 528830 A‘BHflgOg 693.200 77 .5825.» 833 new 0:9:— cozgannm can .30:— ub8 2: no 60933.. can 3 @0335??— 6305 3.76m 05 no 6252* m a 3 “533.5550.- 5:95 33 3 95mg 3 uBon—m 0.55808 2F .3083 3:52— naoonuuaono 2 SEN 582:“ Bur—on 05 mo £53m 9333502 .oN 0.5m?“ 78 3 2am...— v.¢ v.0 l «.m l H.n~ OH m m b w m c m N H mMOSDB mDOW24920mm AOmBZOU 79 .haauuuomvuuouas up wanna—a? 0.33 33080: no «common can $5090. unmmouco 3.73.— 22 .Su mac:— vozvpfium a 3 6033.5»: 33 nu mzxzmu dgfimc: examnmo w:\:mo HumUom 81 all cases (Figure 2.8). The high molecular weight band produced by Hha I digestion of tumor DNA (lanes 2., 4, 6, and 8) was observed to exhibit consistently greater electrophoretic mobility than the corresponding band produced in non-tumor tissue by this enzyme (lanes 1, 3, 5, and 7). This indicates that Hha I is able to cleave DNA in the area of Ila-ras to a greater extent in benzidine-induced tumors due to hypomethylation of sites in this tissue as compared to adjacent non-vtumor tissue. Similar results were obtained when DNA from phenobarbital-induced (Figure 29, lanes 2-3), chloroform-induced (Figm'e 29, lanes 4-7) and spontaneous (Figure 29, lanes 8-11) tumors were digested with Hha 1. Compared to control DNA (lanes 1 and 12), the major band produced by Hha I-digested tumor DNA sometimes exhibited greater electrophoretic mobility (lanes 2. and 9) and a doublet was frequently produced (lanes 3, 6, 7, 8, and 10). An increased exposure time of the membrane shows a much greater intensity of band at 2.6 and 6.2 kb in all tumors as compared to controls (Figure 30). The greater ability of Hha I to cleave tumor DNA than control DNA indicates that the degree of methylation of Ha-ras is often decreased in tumor tissue as compared to control liver. Hha I analysis was also used to examine the methylation state of Ki-ras (Figure 31) in male 36C3Fl (lanes 1-4), male C3H/He (lanes 5-8), female C57BL/6 (lanes 9-12), female 36C3F1 (lanes 13-14), female C3H/He (lanes 15-16) and male CS7BL/6 (lanes 17-18) mouse liver. Two very high molecular weight bands (:23 kb) are produced by Hha I digestion of each sample indicating that no strain or sex differences exist in the methylation state of the Ki-ras oncogene. When DNA samples from the benzidine group were analyzed with Hha I (Figure 32), greater mobility of the lower band was observed in two tumor samples (lanes 2 and 6) as compared to non-tumor samples (lanes 1, 3, and 5) when hybridized to Ki-ras. This result suggests that the Ki-ras oncogene is sometimes 82 ._23 kb band may be indicative of myc hypomethylation in some tumors from the three treatment groups depicted in this figure. Hybridization of the serum albumin probe to Hha I-digested control and opposite sex DNA is shown in Figure 37. In each sample, bands at 10 and 8 kb are produced. These results indicate that serum albumin is methylated to the same extent in male and female B6C3Fl, C3H/He, and C57BL/6 mouse liver. A simlar 9O .fia use w 0080: <29 3.3800 3 «on «an 8" use p .e .N 0080: <75 .883 3 8000 00830800 a“ 85 53308 03009393020 0300.5 no £83 a 3 3303 BOP—4 .uuuumvu 03 0:9:— v0=0nch 0 3 6082583.— ecc 0&9:— magnum 05 «o 008:5» ask on 0863 8 8393 080.5808 05. .8083 030m... 850838090 88 6003.8..8030830 .wvoscfiAsuEusn—oc0nn 8 0:0moono 00.78% 05 «o 3.30 80301508 05 no 808000000 H 85 .mm 0.8m?“ 91 1....a.. 8 28E e... van 8. m U mDOmZ: 0858 QAQWmU H080 .0H.H\H.Hm0 .3393 3 080M028 058 05 «O 330 80301508 05 Ho 80830000 H 8H: .3” 08mg 93 «m 28E I v.0 ‘b‘..&1blriflzsz HM ma hm 0H ma vH nH NH ”H an m m h w m w n N H OHNI 0H080u 0Hna0u 0Hca0u 0HeE 0HOE o\qnan0 0=\=n0 «snows m\qmam0 ozxzno “snows Figure 35. Hha I assessment of the methylation state of the myc oncogene in benzidine-induced tumors and adjacent non-tumor tissue. The membrsfie shown in Figure 32 was stripped of the Ki-ras probe and rehybridized to a P-labelled probe for myc (Pst I fragment). 95 PHENO- C BARBITAL CHIDROFORH SPONTANEOUS 6 7 B 91011 2.3 1.0 Figure 36. Hha I assessment of the methylation state of the myc oncogene in r‘ ‘ and spontaneous hepatic tumors. The “15$“. shown in Figure 33 was stripped of the Ki-ras probe and rehybridized to a labelled probe for myc (Pst I fragment). 96 a Cu 605 2.55:0 use .0H.H\ .— can 0 o :8 .5: n a use u U 5 o on 3 080m 8H85nHHHevaHH—Wuue qu wm 0.5m; .385: 0 a 05 so 33.. screen—“mm." mask—Land.” sou 32.— e020 80830000 «68 34 m H 8.3 . mu hm 03mm .m 97 5:55:55 we as m wImm OHcE 0HOEQH : ansoo 0:\:m0 0H050u ammoom S 28E NH HH as m m S o m w m m H 0H050H 0HOE 0HOE oxqmemo 0=\:mO HanOom 98 result was observed when the serum albumin probe was hybridized to Hha I- digested DNA from all four groups of tumors (data not shown). These results indicate that the methylation state of the serum albumin gene is not altered in tumor tissue as compared to normal liver. 4. DNase I Studies A photograph of an ethidium-bromide-stained baby gel upon which was electrophoresed a set of DNase I—treated B6C3F1 DNA samples is presented in Figure 38. In this figure, as in the following DNase I figures (Figures 39-5l), the 0°C control sample is in lane 1, the 15° control sample is in lane 2, and lanes 3-7 represent DNA treated with increasing concentrations of DNase I (28-47 units/ ml). As can be seen in lane 1, electrophoresis of the 0°C control sample results in a band of very high molecular weight DNA. Similarly, the IS’C control DNA (lane 2) also produces a single band of high molecular weight, although DNA in this sample exhibits slightly greater electrophoretic mobility as compared to the O’C control. The DNase I-treated samples show evidence of smaller DNA fragments in a concentration-dependent manner, as evidenced by a faint smear of DNA down the gel. These results indicate that the 0°C control DNA is undegraded, whereas the 15°C control DNA exhibits minimal degradation. Furthermore, it appears that a limited DNase I digestion has been achieved in the DNase I-treated samples. Figure 39 shows the result of hybridizing EcoRl-digested DNase I-treated B6C3F1 DNA to the firms probe. It can be seen that two bands at 4.4 and 3.0 ltb (see arrows) appear in the DNase I-treated samples (lanes 3-7) in a concentration- dependent manner; these bands are extremely faint (probably due to some endogenous nuclease activity) or not present in the controls (lanes 1-2). The appearance of these bands indicate the presence of DNase I hypersensitive sites in the firms oncogene B6C3F1 mouse liver. 99 60.30538:— Hune 03895 83250 £33 c0830 ask How 0.5 H80 £02302 8 H00£0000H0 as 000002395020 0008 308020 80.8238 080 588:8. Stan 3. In «000: H 0.3an Ho 8385:0250 9580008 5:» UomH as so 3 080: UomH so .H 080: Do: as H 032G «so—:3 6008—508 H035: 89: 00.288 003 42D .H0H08— 00000.5..H 00075 am 80300ch