Rat Hepatocarcinogenesis - NCBI

8 downloads 137 Views 3MB Size Report
Jun 10, 1987 - and JOSEPH LOCKER, PhD, MD. After long-term feeding of a choline-devoid diet to rats, the authors analyzed rasK, rasH, and rasNtranscripts.
American Journal of Pathology, Vol. 129, No. 2, November 1987 Copyright X American Association of Pathologists

Analysis of Ras Genes and Linked Viral Sequences in Rat Hepatocarcinogenesis NALINI CHANDAR, PhD, BENITO LOMBARDI, MD,

From the Department ofPathology, University of Pittsburgh,

School of Medicine, Pittsburgh, Pennsylvania

WOLFGANG SCHULZ, PhD, and JOSEPH LOCKER, PhD, MD After long-term feeding of a choline-devoid diet to rats, the authors analyzed rasK, rasH, and rasNtranscripts and gene structure in livers and liver tumors. They controlled their analysis by studying cell lines derived from chemically induced hepatomas. Transcripts from all three genes were elevated in all tumors, but not in the livers from which they arose. The transcript elevations may represent an effect of active cell proliferation in the tumors. Clone HiHi-3, derived from the Kirsten murine sarcoma virus, detected a large number of hybridization bands, most of which were not part of

the rasK-p21 gene. Most tumors had an altered band at 2.6 kb; some had other altered bands. No alterations were seen in liver DNA, and none of the cell lines showed the 2.6 kbband. At low stringency, a rasHprobe, which contains a short segment of a similar viral sequence, also detected altered bands in tumors and a single treated liver. These changes in endogenous viral sequences ofthe rat genome appear to be characteristic of carcinogenesis by a choline-devoid diet. (Am J Pathol 1987, 129:232-241)

ONCOGENES, transforming genes found in retroviruses and cancers, are altered versions of cellular regulatory genes termed proto-oncogenes.' In cell transformation, retroviruses can provide exogenous oncogenes, or proto-oncogenes can mutate to transforming activity. Such mutations alter the sequences encoding protein structure or the sequences regulating the timing or level oftranscription in the cell. The increase to transforming levels of expression can result from gene amplification or translocation of exons to regions with stronger transcription controls. Viral transduction is a special case of the latter mechanism. Viral sequences are inserted into the gene, which is then transcribed under the control of a viral promoter and the viral long-terminal-repeat enhancer, superseding normal proto-oncogene regulatory elements.5 Oncogenes of the ras family have been frequently detected in tumors.6 Two distinct ras genes were discovered as the transforming genes of the rat retroviruses Harvey murine sarcoma virus (HaMSV) and Kirsten murine sarcoma virus (KiMSV). The latter virus resulted from the passage of a nontransforming virus (Moloney murine leukemia virus) in rats7 and represents a complex recombinant that contains leukemia virus elements, endogenous nontransforming rat virus elements, and a cellular gene (the ras gene), which has further mutated.8'9 The rasH oncogene

probably has a similar origin. In contrast, the rasN oncogene was isolated from a human tumor,6 not a retrovirus. Each ras oncogene encodes a distinct transforming p21 protein.10 Cloned probes from all three detect distinct cellular proto-oncogenes which encode nontransforming p21 proteins. Cellular proto-oncogenes play an important role in carcinogenesis. Enhanced ras expression, for example, has been observed in chemical hepatocarcinogenesis.t 1-14 Chronic feeding of a choline-devoid (CD) diet to rats, in the absence of exogenous carcinogens, causes the appearance of a fatty liver, hepatocellular necrosis with increased hepatocyte turnover, hepatofibrosis, and eventually hepatocellular carcinoma.I5 18 In previous studies, we have analyzed the effects of the CD-diet on methylation of total liver DNA19 and on expression and methylation of the a-fetoprotein gene.20 This paper describes the use of ras probes to study the CD-diet effects on ras gene expression and Supported by Grants 1638 from the Council for Tobacco Research, CA23449 from the National Cancer Institute, and BC471 from the American Cancer Society. Accepted for publication June 10, 1987. Address reprint requests to Joseph Locker, University of Pittsburgh, School of Medicine, Department of Pathology, 781 Scaife Hall, Pittsburgh, PA 15261.

232

CD-DIET HEPATOCARCINOGENESIS

Vol. 129 * No. 2

structure. With these probes, we found elevated ras transcript levels, but we also found structural alterations in genomic segments homologous to non-p2 l regions of KiMSV and HaMSV, segments unique to the rat genome (Locker et al., submitted for publication).

Materials and Methods Experimental Animals and Tumors Male Fisher 344 rats (Harlan Sprague-Dawley, Inc., Indianapolis, Ind) weighing 90-100 g were placed on a refined choline-deficient (CD) or cholinesupplemented (CS) diet as described in Yokoyama et al.'8 Liver tumors became apparent at 13-14 months when animals began to lose weight or were palpable in the abdomen. The tumors were separated from surrounding parenchyma and necrotic tissue and frozen immediately. Two large tumors (from animals A and C) and two smaller tumors (from animals B and D) were obtained from an initial series of animals. The letter designations enable comparison of our analyses of individual tumors in two previous studies.'9'20 Tumor-bearing animals from a later series have been designated E-G. Nucleic acids were isolated from the tumors, from the livers that contained them, and from a series of livers obtained after 14 months of feeding a CD or CS diet. In most cases, the number of analyses possible on individual tumors was limited by the quantities of nucleic acids isolated; such limitations were not a factor in analysis ofthe liver samples.

Cell Lines Cell lines McA-HC7777,21 NH-Tu6,22'23 and MH,C,24 were provided by Patricia A. Hoffee, University of Pittsburgh; HTC-SR25 was from Peter Ove, University of Pittsburgh; and Clone 926 and McAHC899421'27 were from the American Type Culture Collection. EOC-ST1269 is an oval cell line established in our laboratories.28'29 All were maintained on Williams E medium (GIBCO, Grand Island, NY) containing 10% fetal calf serum + 2 mM glutamine, except that McA-RH8994 medium contained 15% fetal calf serum and 0.5 mM arginine. Recombinant DNA Clones Plasmids BS-930 (rasH) and HiHi-330 (rasK) were provided by Edward M. Scolnick (Merck, Sharpe, and Dohme Research Laboratories, West Point, Pa). Clone pP485-4,6 containing a portion of the human rasN gene, was provided by Stuart A. Aaronson (Na-

233

tional Institutes of Health, Bethesda, Md). Plasmid HiHi-3 (Figure 1) is a subclone of KiMSV cDNA clone." This plasmid contains the transforming rasK gene and some flanking 5' and 3' sequences. Probes were derived by cutting with appropriate restriction enzymes, resolving on "strand-separating" acrylamide gels, and eluting by diffusion as described by Maxam and Gilbert. 10 The rasH hybridization probe was the entire cDNA insert of BS-9, excised as a 0.5 kb EcoRI fragment, and the rasN probe was the entire genomic insert of pP485-4 excised as a 0.9-kb PvuII fragment. DNA and RNA Purification and Characterization Tissue originally frozen in liquid N2 was powdered with dry ice in a metal blender; the dry ice was then allowed to sublime at -80 C overnight before homogenization. High-molecular-weight DNA was extracted from frozen tissue powder or from tissue culture cells and purified according to protocols that we have previously described.3' RNA was purified from frozen tissue powder by homogenizing in a buffer containing 100 mm NaCl, 10 mm Tris, pH 7.6, 50 mm EDTA, 10 mg/ml aurin tricarboxylic acid.32 Sarkosyl was added to a final concentration of 1%. Tissue culture cells were directly solubilized in growth flasks with lysing solution + 1% Sarkosyl. The mixture was repeatedly extracted with buffer-equilibrated phenol. Solid LiCl was dissolved to a final concentration of 2 M. The RNA was precipitated at 0 C for 16 hours, pelleted at 10,000 g for 10 minutes, redissolved in homogenization buffer, and reprecipitated with 2 M LiCl as above. RNA was quantitated with the orcinol reaction.33 EcoR I Hine H-

I

-K-.oe

HiHi-3

Xxxx

Probes (

Kb

EcoR I XbaI -H4inellA

SstI

_

r 0

Il 0.5

_

_

1 1.0

Figure 1-V-rasK clone HiHi-3 (modified from Ellis et al'). Clone HiHi-3 was constructed by excising a Hincil fragment from a larger KiMSV clone. The excised fragment was ligated to EcoRI linkers and cloned in pBR322. The sequence of the rasK gene within this clone was reported by Tsuchida et aW2 within the sequence of the large Sstll-Hincil fragment. On the basis of this information, we purified individual hybridization probes: Probe 1 was the small EcoRI-Sstil fragment; Probe 2, the Sstll-Xba fragment; Probe 3, the small Xba-EcoRI fragment; and Probe 4, the large Sstil-EcoRI fragment. Probes 2 and 4 contain 24 base-pairs of non-rasK sequences at the 5'-end.

234

CHANDAR ET AL

Hybridization Analysis Genomic DNA was digested with restriction enzyme PvuII (New England Biolabs, Beverly, Mass) for 2 hours at 37 C, resolved on agarose gels, and blotted to nitrocellulose. Molecular weights of hybridization bands were calculated by comparison with standard digests run on the same gels. For dot blots, various concentrations of RNA were spotted on nylon membranes (Zetaprobe, Bio-Rad Laboratories, Richmond, Calif) by means of a filtration manifold. Membranes were baked at 80 C for 2 hours in vacuo before hybridization. Hybridization probes were labeled by nick translation according to Rigby et al.34 High-stringency hybridizations were carried out in a mixture that included 10% dextran sulfate and 50% formamide at 37 C; these conditions approximate Tm-20. Low-stringency hybridizations were carried out in 10% dextran sulfate and 35% formamide at 24 C (approximately Tm-40). All washes were carried out at 45 C, below the Tm of the hybrids.

Results Analysis of Transcript Levels We began our analysis by studying the steady-state levels of ras gene transcripts. RNA from 4 tumors, the livers from which they arose, and other CD and CS diet livers was analyzed by hybridization to probes for all three ras genes (Figure 2). We initially used HiHi-3 as a rasK probe, but subsequently used "Probe 2" to specifically measure rasK transcripts. HiHi-3 hybridized much more strongly than ras gene probes to all liver and tumor RNAs, indicating that its non-rasK KiMSV sequences are transcribed. With each of the probes, the 4 tumors had higher transcript levels than livers. Fourteen-month CD diet livers had no significant elevations, compared with CS diet controls. Gene Analysis With HiHi-3 We first utilized HiHi-3 under the assumption that it was a rasK-specific hybridization probe. Despite the complex band patterns detected with this probe, we observed bright altered bands in 2 tumors and more subtle alterations in 2 other tumors (Figure 3). Tumors A and C had, in high-stringency hybridization, a prominent increase in a normally weak hybridization band at 2.6 kb. Tumor A also had a new band at 1.8 kb; tumor B lost a band at 3.0 kb; and tumor D appeared to have a new band at 19.5 kb. Despite careful analysis of these patterns using multiple exposure densities and densitometric scans, we found no alter-

AJP

*

November 1987

ations in CD livers, only in tumors. Subsequent analysis at high stringency showed an increased 2.6-kb band in 2 additional tumors (see below). To rule out the possibility that the 2.6-kb band represented a virus, rather than an integrated genomic element, the DNAs of Figure 3 were blotted and hybridized to HiHi-3 without restriction enzyme digestion (not illustrated). No viral bands were detected. As a control for our DNA analysis of CD diet-induced hepatomas, we analyzed a variety of rat hepatomas from other sources, available as cell lines in our laboratories (Figure 4). The hybridization patterns varied because of the different strains of origin for these hepatomas, but most were indistinguishable from the patterns of normal liver. Only McA8994 cells had an alteration, a band of increased density at 3.2 kb. No cell line showed a prominent 2.6-kb band. This latter alteration may distinguish CD-diet hepatocarcinogenesis from other forms of hepatocarcinogenesis. The complexity of the patterns does not enable us to rule out other altered bands. Gene Analysis With Regions of HiHi-3 HiHi-3 was subdivided into probes representing the rasK gene itself (Probe 2) and sequences 5' (Probe 1) and 3' (Probe 3) to rasK. However, Probe 2 includes 24 bases of 5'-flanking sequences. High- and lowstringency hybridizations to normal liver DNA (Figure 5) indicated simple hybridization patterns with Probes 2 and 3. Bands detected at 5.2 and 4.8 kb represent the rasK gene(s), while a band at 9.5 kb probably represents a unique genomic 3'-flanking region. A 4.2 kb band is detected with Probe 2 at low stringency, but is strikingly absent at high stringency. It comigrates with a band detected by Probe 1 at 4.2 kb. The 5.2 kb band detected with Probe 2 also comigrates with a band detected by Probe 1. This observation may indicate genomic linkage between rasK and viral elements, but it must be clarified by further experiments. It is likely that some of the low-stringency hybridization bands represent homology to the 25 bases of non-rasK sequence at the 5' end of Probe 2. Because our analysis indicated that the hybridization patterns of Probes 2 and 3 could be superimposed without hindering analysis, further experiments were carried out with Probe 4, which demonstrates both rasK gene elements and unique 3'-flanking elements. Probe 1 was used to characterize endogenous KiMSV-related viral sequences. Subsequent experiments used low-stringency hybridization conditions, because they revealed more bands without hindering

analysis. Low-stringency hybridization with Probe 1 (Figure 6, right) also tested whether the faint band normally

Vol. 129 * No. 2

CD-DIET HEPATOCARCINOGENESIS

235

a-Dot-blot hyFigure 2-Transcript levels. bridizations. Total cellular RNA was filtered onto nylon membranes in 1- and 0.2-pug aliquots. Membranes were hybridized with nick-translated probes. Probe HiHi-3 gave a strong hybridization to all RNAs and thus required a short exposure time. The three ras probes required much longer exposure times to show analyzable signals. RNA was analyzed from livers of 2 animals fed a CS diet, 2 animals without tumors fed a CD diet, and from livers (L) and tumors (T) from 4 tumor-bearing animals fed a CD diet (anib-Quantitation of tranmals A, E, F, and G). scripts. The blot hybridizations from a were quantitated with a densitometer. Values were normalized to the CS diet values, which were defined as 100%. Three bars are shown for each comparison: CS (open bars) represents livers of 2 animals fed the control CS diet; CD (cross-hatchedbars), livertissue from 6 animals, with or without tumors, fed the CDdiet; and T (black bars), 4 tumors. The line above each bar is the standard deviation of that data set.

present at 2.6 kb is amplified in tumors or obscured by another species that comigrates. The normal faint band increases in intensity at lower stringency, while the bright tumor band does not. By densitometric analysis, the tumor 2.6-kb band is in each case about three times more intense than its counterpart in CS and CD diet livers. Presumably, this tumor band rep-

resents a different genomic region, more homologous to the probe than the normal faint 2.6-kb band. Analysis with Probe 4 (Figure 6, left) demonstrated the bands of the rasK gene and its unique 3'-flanking region at 9.5, 5.2, and 4.8 kb at high stringency, and additional bands of 4.2 and 1.1 kb at low stringency. There was no alteration of these bands in tumors or

236

CHANDAR ET AL

C CS CD L T

AJP * November 1987

A

D

LT

L

B T

kb

19.5 15.1 -

-4.88

-

Figure 3-HiHi-3 high-stringency hybridization to genomic blots. Five micrograms of each DNA was digested with Pvull and resolved on a 1% agarose gel. The left panel shows the analysis of animals A and C, which had large tumors, and CS and CD diet controls. The right panel shows the analysis of animals B and D, which had small tumors. The molecular weights of some major species and of some bands that show alterations are listed in the center. The 4.8-kb band represents a major rasK species; the other bands are found in HiHi-3 but not its rasK-specific regions. On inspection, all CS, CD, and L lanes show identical pattems. Tumor A shows increased intensity of bands at 2.6 and 1.8 kb; C shows increased intensity of a band at 2.6 kb; B shows reduced intensity of a 3.0 kb band; and D shows a possible new band at 19.5 kb. Other changes could be obscured by the complexity of the pattems.

3.6 3.2 3.0 2.6 2.1 M-,

1.8 ----

- 1.1

a b

c d

e f

livers, nor in the tissue culture cell lines (not illustrated). RasH Gene Analysis At high stringency, the rasH probe detected bands at 1.6 and 1.3 kb. None of the tumors or CD or CS diet livers showed changes in these rasH gene specific bands. Of the tissue culture cell lines, HTC showed

-

9

h i J

doubled intensity of the 1.6 kb band; other cell lines showed no apparent changes (data not illustrated). Low stringency also demonstrated bands at 4.7 and 2.1 kb (Figure 7), which probably are not part of the rasH gene. Unlike the 4.7-kb band, the 2.1 -kb band is variably demonstrated in low-stringency hybridizations, suggesting that it is an abundant species with low homology to the probe. It does not correspond to any major species demonstrated with HiHi-3. The

CD-DIET HEPATOCARCINOGENESIS

Vol. 129 * No. 2

bands were detectable at low stringency. Representative analysis is shown in Figure 8. None of 6 tumors, 10 CD livers, 4 CS livers, or 7 tissue culture cell lines showed alterations in these patterns.

I) (0

0)

0

co

e)

w

N_ r-

=F

237

=)

I-

_

Discussion

kb -4.8 -4.2 -3.6 -3.2 -3.0 -2.6

a

b

c

d

e

f

g

h

Figure 4-HiHi-3 hybridization to cell line DNA digests. High-stringency hybridizations was carried out as in Figure 1. Lane d shows normal liver DNA from a Sprague-Dawley rat, the source of cell line EOC ("ethionine oval cells"29 n). Other strains have patterns like F344 rats,21 despite their varied strain attributions. Cell lines McA-HC7777, McA-HC8994, MH,CI (from Morris hepatoma 779525), and HTC (from Morris hepatoma 7288Cm2 are derived from Morris hepatomas and were induced with various carcinogens.22 Tu 623,24 is derived from Novikoff hepatoma, which was induced with 4-dimethylaminoazobenzene. Clone 927 (CI.9), a spontaneously immortalized line of rat liver "epithelial" cells, is not the result of carcinogen treatment. Note that none of the lines show an increased band at 2.6 kb. Une 8994 has an increased density band at 3.2 kb. Other apparent differences in the lanes reflect variation in the amounts of DNA loaded in the digests.

2.1-kb band, however, was much stronger in one CD liver and in tumor F. Tumor E showed a new band at 4.5 kb. RasN Gene Analysis RasN-specific sequences were detectable at high stringency as 4.4-, 3.8-, and 2.8-kb bands. Other faint

Ras Genes and Transcript Levels Our results show an increase in the tumor transcript levels of all three ras genes. Other studies have shown elevated levels of various ras gene transcripts not only in hepatic tumors but also in carcinogentreated livers. 1013 RasK and rasH transcripts were elevated in livers of rats treated with ethionine and a CD diet, and in oval cells isolated from these livers."I Liver rasH transcripts were reported to be elevated during treatment with 3'-methyl-4-dimethylaminoazobenzene'0" 2 and in tumors induced by this agent. 13 Increased rasH transcripts were found in aflatoxin-induced tumors.13 Several Morris hepatomas had increased rasH and rasK transcripts.'2 These rasgene transcript elevations probably reflect active cell proliferation in the tumors, because elevated rasK and rasH transcripts occur after partial hepatectomy.35 Studies in our laboratories have shown that CD dietinduced tumors have elevated levels of other oncogene transcripts, including those of the myb, c-myc, mycN, erb-B, and rafgenes (J. Kandala et al., manuscript in preparation), in addition to the ras transcript elevations. Because ras oncogenes generally differ from ras proto-oncogenes by point mutations,3"10 the transcript elevations do not indicate transforming activity. We found no structural alterations of ras genes detectable by Southern blot analysis, ie, translocations, deletions, and amplifications. Although our data do not not rule out the presence of transforming ras genes, it seems most likely that elevated ras (and other) transcripts result from altered control of ras proto-oncogenes in the tumor cells, perhaps because the tumors consist of proliferating cell populations.

Genomic Alterations Detected With Clones HiHi-3 and BS-9 Despite the lack of detectable alterations in protooncogenes, we observed other genomic alterations with clones HiHi-3 and BS-9. These alterations were in regions of the KiMSV that have an endogenous counterpart in the rat genome. We did not demonstrate genomic linkage between these endogenous viral elements and the ras genes, despite their linkage in the probes. Altered bands were not detected in an extensive series of CD- and CS- livers, but only in the

238

CHANDARETAL

AJP * November 1987

LOW STRINGENCY

HIGH STRINGENCY

kb -9.5 -5.2 -4.8 -4.2

Figure 5-Probe hybridization panterns and stringency effects. Five micrograms Pvull digests of F344 DNA were resolved on a 1% agarose gel and blotted to nitrocellulose. Hybridization to Probes 1-3 was compared at low and high stringency (see Materials and Methods). Bands apparent mainly at low stringency are marked with asterisks. Molecular weights are indicated for bands detected with probes containing rasK sequences.

*

-1.1 *

PROBE:

I

2 3

tumors. Four of6 tumors had a specific alteration in a 2.6-kb band. This 2.6-kb band alteration was not detected in carcinogen-induced tumors as represented

by the cell lines. The changes thus appear to be specific for and common in hepatocellular carcinomas induced by choline deficiency. Hybridization with clone BS-9 demonstrated some altered tumor (and 1 CD liver) bands, but only at low stringency. The altered bands were not the rasH-specific bands detectable at high stringency. These altered bands appear to represent the same kind of endogenous viral sequences detectable with HiHi-3.

1

2

3

RasKProbe 2 contains a 25 base-pair segment 5' to the p21 initiation codon36; and under our low-stringency conditions, this 25-base homology is sufficient to detect some non-rasK hybridization bands, especially a 4.2-kb band (see Figure 3). Clone BS-9 contains 64 base-pairs of sequence 5' to its p21 initiation site.37 This 5'-end region represents a segment ofHaMSV in a position analogous to that of Probe 1 in KiMSV. Our comparison of published HaMSV and KiMSV sequences ofthese regions shows 2 HaMSV segments (each 17 bases) with 54% and 56% homology to the KiMSV segment. Thus, while clone BS-9 is subopti-

239

CD-DIET HEPATOCARCINOGENESIS

Vol. 129 * No. 2

E

CS CS CD CD

L

T

T

F

E

F

L

CS CS CDCD T

L

-

-

T

L

9.5

5.2 4.8

---4.2---

3.6 3.2 3.0 2.6

-

-

1.1

a

b

c

d

e

f

g

h

a

b

c

d

e

f

g

h

Figure 6-Low-stringency hybridization to genomic blots using rasK and non-rasK regions of HiHi-3. The left panel shows hybridization with Probe 4, specific for the rasK p21 gene and its 3'-flanking sequences. Thepanel on the left shows hybridization with Probe 1, which represents the 5' region of HiHi-3, a segment of the KiMSV upstream from the p21 gene. These represent 2 CS diet animals, 2 CD diet animals, and 2 tumor-bearing animals (E and F), all different from those analyzed in Figure 2. Low-stringency hybridization is illustrated here to contrast with the high-stringency hybridization of Figure 2. The superimposition of the left and right patterns is equivalent to hybridization with the intact HiHi-3, as illustrated in Figure 2, except for the difference in stringency. Other slight differences from the pattern of Figure 2 are a result of better band resolution in this blot. For example, the 3.2-kb band of Figure 2 has been resolved to two fainter bands. Note that both tumors have increased intensity of the 2.6-kb band.

mal for detecting such bands, it may detect altered bands of the same kind as those detected with HiHi-3. Hypotheses Several models should be the basis for further analysis:

1) The altered bands

may

represent endogenous

virus activation or infection with a related exogenous virus. Choline deficiency could be the activating agent or predispose to such infections. The absence of such changes in the CD livers, however, makes this hypothesis unlikely. 2) The endogenous viral sequences may be preferred sites of gene translocation or amplification. These arrangements may reflect a special feature of

240

CHANDAR ET AL

E CS CS CD CD

T

L

T

-i~ ;.X

translocation. Alternatively, the altered bands may represent the general instability of the tumor genomes. The latter explanation seems unlikely, because the hepatoma cell lines show only limited similar alterations despite years in culture. 3) The endogenous viral sequences are widespread

F L

AJP * November 1987

E

CS CS CD CD

-4.7 -4.5

-2.1

-1.6 -1.3

a

b

c

d

e

f g

h

*-:.

Figure 7-RasH genomic blots. Clone BS-9 DNA was hybridized at low stringency to Pvull digests of the same series of DNAs as Figure 4. Only the 1.6 and 1.3 bands are visualized at high stringency and represent rasH-specific bands. Despite its brightness, the 2.1-kb band is most variable in its detection in low-stringency blots. Here, in a hybridization internally controlled by flanking lanes of hybridization, this band has increased intensity in several CD diet livers and tumors. Tumor E has a new band at 4.5 kb.

such viral sequences or may be typical of intermediate repetitive sequences. Such rearrangements may activate proto-oncogenes and thus have a causal role in choline deficiency carcinogenesis. However, we have found no evidence for ras gene amplification or

a

UI

..

b c d

-0.8....... ....

T

L

F T

L W

-4A -3.8 -2.8

i °W ~ ~ ~ . . . . :,1.1: .'1. .

e f

g

h

Figure 8-RasN genomic blots. Clone pP485-4 DNA was hybridized at low stringency to Pvull digests of the same series of DNAs as in Figure 4. The 4.4-, 3.8-, and 2.8-kb bands are visualized at high stringency. There are no apparent differences in tumor or liver DNAs.

Vol. 129

*

CD-DIET HEPATOCARCINOGENESIS

No. 2

in the rat genome, and there is a significant probability that one will coincidentally be detected as part of an amplification unit. This would imply that the CD diet predisposes to gene rearrangements and thus differs from chemical carcinogenesis.

References 1. Hunter T, Cooper JA: Protein-tyrosine kinases. Ann Rev Biochem 1985, 54:897-930 2. Burgess A: Growth factors and oncogenes. Immunol Today 1985, 6:107-112 3. Muller R, Verma I: Expression of cellular oncogenes. Curr Topics Microbiol Immunol 1984, 112:73-115 4. Bishop JM: Viral oncogenes. Cell 1985, 42:23-38 5. Hayward WS, Neel BG, Astrin SM: Activation of a cellular onc gene by promoter insertion in ALV-induced lymphoid leukosis. Nature 1981, 290:475-480 6. Eva A, Tronick S, Gol RA, Pierce JH, Aaronson SA: Transforming genes of human hematopoietic tumors: Frequent detection of ras-related oncogenes whose activation appears to be independent of tumor phenotype. Proc Nat Acad Sci USA 1983, 80:4926-4930 7. Kirsten WH, Meyer LA: Morphologic responses to a murine erythroblastosis virus. J Nat Cancer Inst 1967, 39:311-319 8. Norton JD, Avery RJ: Genetic organization and cloning of Kirsten murine sarcoma virus DNA. Biochem Biophys Res Commun 1982, 108:1631-1637 9. Norton JD, Connor J, Avery RJ: Genesis of Kirsten munne sarcoma virus: Sequence analysis reveals recombination points and potential leukemogenic determinant on parental leukemia virus genome. Nucleic Acids Res 1984, 12:6839-6852 10. Tanaka T, Slamon DJ, Battifora H, Cline MJ: Expression ofp21 ras oncoproteins in human cancers. Cancer Res 1986, 46:1465-1470 11. Yaswen P, Goyette M, Shank PR, Fausto N: Expression of c-Ki-ras, c-Ha-ras, and c-myc in specific cell types during hepatocarcinogenesis. Mol Cell Biol 1985, 5:780-786 12. Cote GJ, Lastra BA, Cook JR, Huang D-P,Chiu J-F: Oncogene expression in rat hepatomas and during hepatocarcinogenesis. Cancer Lett 1985, 26:121-127 13. Tashiro F, Morimura S, Hayashi K, Makino R, Kawamura H, Horikoshi N, Nemoto K, Ohtsubo K, Sugimura T, Ueno Y: Expression of the c-Ha-ras and c-myc genes in aflatoxin B1-induced hepatocellular carcinomas. Biochem Biophys Res Commun 1986, 138:858-864 14. Makino R, Hayashi K, Sato S, Sugimura T: Expressions of the c-Ha-ras and c-myc genes in rat liver tumors. Biochem Biophys Res Commun 1984, 119:1096-1102 15. Ghoshal A, Farber E: The induction ofliver cancer by a dietary deficiency of choline and methionine without added carcinogens. Carcinogenesis 1984, 5:1367-1370 16. Mikol YB, Hoover KL, Creasia D, Poirier LA: Hepatocarcinogenesis in rats fed methyl-deficient, amino aciddefined diets. Carcinogenesis 1983, 4:1619-1629 17. Shinozuka H, Katyal SL: Pathology of choline deficiency. In: Nutritional Pathology. Edited by H Sidransky. New York, Marcel Dekker, 1985, pp 279-320 18. Yokoyama S, Sells MA, Reddy TV, Lombardi B: Hepatocarcinogenic and promoting action of a cholinedevoid diet in the rat. Cancer Res 1985, 45:2834-2842

241

19. Locker J, Reddy TV, Lombardi B: DNA methylation and hepatocarcinogenesis in rats fed a choline-devoid diet. Carcinogenesis 1986, 7:1309-1312 20. Locker J, Hutt S, Lombardi B: a-Fetoprotein gene methylation and hepatocarcinogenesis in rats fed a choline devoid diet. Carcinogenesis 1987, 8:241-246 21. Morris HP, Meranze DR: Induction and some characteristics of "minimal deviation" and other transplantable rat hepatomas. Rec Results Cancer Res 1972, 44:103-114 22. Hoffee PA, Hunt SW, Chiang J: Isolation of deoxycoformycin-resistant cells with increased levels of adenosine deaminase. Som Cell Genet 1982, 8:465-477 23. Hay R, Macy M, Corman-Weinblatt A, Chen TR, McClintock P: American Type Culture Collection of Cell Lines and Hybridomas. 5th edition. Rockville, American Type Culture Collection, 1985, p 131 24. Richardson IU, Tashjian AH, Levine L: Establishment of a clonal strain of hepatoma cells which secrete albumin. J Cell Biol 1969, 40:236-247 25. Thompson, BE: Liver Cells (HTC). Meth Enzymol 1979, 58: 544-552 26. American Type Culture Collection Catalog of Cell Lines and Hybridomas, 5th edition. p 141 27. de Nechaud B, Becker JE, Potter VR: Effect of glucocorticoids on fetoprotein production by an established cell line from Morris hepatoma 8994. Biochem Biophys Res Commun 1976, 68:8-15 28. Sells MA, Katyal SL, Shinozuka H, Estes LW, Sell S, Lombardi B: Isolation of oval cells and transitional cells from the livers of rats fed the carcinogen DL-ethionine. J Natl Cancer Inst 1981, 355-360 29. Yoshimura H, Harris R, Yokoyama S, Takahashi S, Sells MA, Pan SF, Lombardi B: Anaplastic carcinomas in nude mice and in original donor strain rats inoculated with cultured oval cells. Am J Pathol 1983, 110:322-332 30. Ellis RW, DeFeo D, Shih TY, Gonda MA, Young HA, Tsuchida N, Lowy DR, Scolnick EM: The p21 src genes of Harvey and Kirsten sarcoma viruses originate from divergent members of a family of normal vertebrate genes. Nature 1981, 292:506-510 31. Kunnath L, LockerJ: Variable methylation ofthe ribosomal RNA genes of the rat, Nucleic Acids Res 1982,

10:3877-3892 32. Hallick RB, Chelm BK, Gray PW, Orozco EM: Use of aurin tricarboxylic acid as an inhibitor of nucleases during nucleic acid isolation. Nucleic Acids Res 1977, 4:3055-3064 33. Volkin E, Cohn WE: Estimation of nucleic acids, Methods of Biochemical Analysis. Vol 1. Edited by D Glick. New York, Interscience, 1954, pp 287-303 34. Rigby PWJ, Dieckmann M, Rhodes C, Berg P: Labelling deoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase I. J Mol Biol 1977, 113:237-251 35. Goyette M, Petropoulis CJ, Shank PR, Fausto N: Regulated transcription of c-Ki-ras and c-myc during compensatory growth of rat liver. Mol Cell Biol 1984, 4:1493-1498 36. Tsuchida N, Ryder T, Ohtsuba E: Nucleotide sequence of the oncogene encoding the p21 transforming of Kirsten murine sarcoma virus. Science 1982,217:937-939 37. Dhar R, Ellis RW, Shih TY, Oroszlan S, Shapiro B, Maizel J, Lowy D, Scolnick E: Nucleotide sequence of the p21 transforming protein of Harvey murine sarcoma virus. Science 1982, 217:934-937