(CANCER RESEARCH 52, 4507-4513, August 15, 1992]

K-raj Mutation is an Early Event in Pancreatic Duct Carcinogenesis in the SyrianGolden Hamster1

Wendy L. Cerny, Kathy A. Mangold, and Dante G. Scarpelli2

Department of Pathology, Northwestern University Medical School, Chicago, Illinois 60611

ABSTRACT

K-ra.v oncogene mutation has been shown to be a frequent event inpancreatic ductal adenocarcinomas induced by the carcinogen V-ni-

troso-bis(2-oxopropyl)amine in the hamster. The present study examines the mutational status of the K-r«.voncogene in lesions that precede

the appearance of invasive ductal adenocarcinomas. Syrian golden hamsters (80-100 g) received 12 weekly doses of 15 mg/kg A'-nitroso-bis-

(2-oxopropyl)amine and were serially sacrificed at 8, 12, 14, 16, or 24weeks following the initiation of treatment. Ten ¿im-thick sections offormalin-fixed paraffin-embedded pancreas were examined for hyper-plasia, papillary hyperplasia, carcinoma in situ, and invasive and met-

astatic ductal carcinoma. Marked lesions of interest were scraped fromthe slide, subjected to polymerase chain reaction-mediated amplificationof the first exon of the K-ras gene, and probed by oligonucleotide-

specific hybridization for mutations at codon either 12 or 13. Of 186samples assayed, K-ras codon 12 mutations were detected in 26% of

hyperplasias, 46% of papillary hyperplasias, 76% of carcinoma in situ,80% of adenocarcinomas, and 43% of lymph node métastases. Codon 12mutations were exclusively G to A changes at the second position.Codon 13 mutations were only detected in 9 of 168 samples. Theseresults suggest that K-ras activation is an early event in W-nitroso-bis-(2-oxopropyl)amine-induced pancreatic carcinogenesis in the hamster.

INTRODUCTION

Despite numerous reports in the literature concerning thedetection of activated oncogenes in a wide range of humantumors, little is known about the actual roles these genes play inneoplastic transformation. One approach toward elucidatingtheir role is to determine when during multistage carcinogenesis a particular oncogene is activated. By identifying criticalstages during the neoplastic process, it may be possible to unravel the sequence of events required for a particular cell type tobecome neoplastic. Such analyses ideally involve the identification of the preneoplastic lesions that precede frank tumor formation and have resulted in the identification of several discretesteps in the progression of colonie adenomas to adenocarcinomas, involving both oncogene activation (K-ras mutation) andtumor suppressor loss (1,2). Similiarly in mouse skin, analysisof benign papillomas that are believed to precede the development of invasive carcinomas has revealed the activation of 2different oncogenes (fas and H-ras) (3, 4). Mutation of the K-and H-ras genes prior to the onset of neoplasia has been reported in rat mammary tissue by Kumar et al. (5). These activated oncogenes were detected in morphologically normalmammary cells within 12 days after carcinogen exposure andmonths before tumor development.

In the present study, we have used this approach to determinewhether K-ras activation is an early or late event in the devel-

Received 3/9/92; accepted 6/8/92.The costs of publication of this article were defrayed in part by the payment of

page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This study was supported in part by NIH Grants CA34051, CA09560, andDK08574, and by the Marie A. Fleming and Edith M. Patterson Cancer ResearchFunds.

2 To whom requests for reprints should be addressed, at Department of Pathology, Northwestern University Medical School, 303 East Chicago Avenue, Chicago,IL 60611.

opment of BOP3-induced pancreatic ductal adenocarcinoma inthe Syrian golden hamster. This animal model has proven to beparticularly relevant in the study of pancreatic ductal carcinogenesis due to the fact that the tumors induced are biologicallyand morphologically strikingly similar to those encountered inhumans (6). Recently, it has been shown by our laboratory (7)and others (8,9) that these tumors frequently contain a mutatedK-ras oncogene and, therefore, also share identity with theirhuman counterparts at the molecular level (10-15). Since themajority of patients with pancreatic ductal adenocarcinomapresent with advanced disease, early or precursor lesions havevery rarely been seen in humans. However, using the Syrianhamster model we were able to observe pancreatic lesions atvarious stages following BOP treatment and directly assess thestatus of K-ras activation in focal areas of hyperplasia, papillaryhyperplasia, and CIS, as well as in adenocarcinomas and lymphnode métastases.Our finding that K-ras activation can be anearly event in BOP-induced pancreatic carcinogenesis in thehamster is the first step toward understanding how this oncogene is involved in the neoplastic transformation of pancreaticducts.

MATERIALS AND METHODS

Carcinogen Administration. Fifty weanling Syrian golden hamsters(80-100 g; Charles River) received s.c. injections weekly with 15 mg/kgof BOP for 12 weeks. Animals were killed at 8, 12, 16, or 24 weeks(experiment I), or 14 weeks (experiment II) following the initiation oftreatment, and the pancreas was removed and fixed in 10% bufferedformalin.

Tissue Preparation. Two 10-itm-thick sections of formalin-fixedparaffin-embedded pancreas were stained with hematoxylin-eosin andexamined for the presence of preneoplastic and neoplastic duct lesions,which included hyperplasia, papillary hyperplasia, CIS, and invasiveand metastatic ductal carcinoma. Ductal areas that contained a combination of lesion types were always classified as the most advanced lesionpresent. (For example, areas where hyperplasia and CIS were contiguous were classified as CIS). Selected lesions were marked and photographed.

Tissue Scraping and Solubilization. Slides were first soaked in xy-lene to remove coverslips, then rehydrated in successive washes of absolute ethanol, 95, 80, and 70% ethanol, respectively. Sections weredestained in 1% hydrochloric acid in 70% ethanol before a final rinse inwater. Marked lesions of interest were scraped from the slide using a22G needle with special care taken to limit the amount of surroundingnormal stromal and acinar tissue. Scraped tissue fragments were placedin 25 til of buffer (50 HIMTris, pH 8.5, 1 min EDTA, 0.5% Nonidet P-40detergent, 200 Mg/ml proteinase K). Samples were incubated for 16 to24 h at 40°C.All samples were heated to 95°Cfor 10 min prior to being

used in the PCR.Polymerase Chain Reaction. PCR-mediated amplification of a Ill-

base pair product from the first exon of the hamster K-ras gene inextracted samples was achieved using primers specific for the codon 12and 13 regions of the human K-ras gene (Clontech Laboratories, Inc.,Palo Alto, CA). Forty-five cycles of primer annealing at 37°C(2 min),primer extension at 74°C(2 min), and denaturation at 94°C(1.5 min)

3 The abbreviations used are: BOP, /V-nitroso-bis(2-oxopropyl)amine; CIS, car-cinoma-i'n-siru; PCR, polymerase chain reaction.

4507

Research. on October 9, 2020. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

K-ras IN HAMSTER PANCREATIC CARCINOGENESIS

were performed using Taq DNA polymerase (Stratagene, La Jolla, CA)and a Perkin-Elmer Cetus Thermal Cycler (Norwalk, CT). Aliquots ofPCR were electrophoresed through an 8% polyacrylamide gel, followedby ethidium bromide staining, for analysis.

Oligonucleotide-specific Hybridization. Successful PCR were applied to multiple Zeta-Probe nylon membranes (Bio-Rad Laboratories,Richmond, CA) using a Bio-Dot SF apparatus (Bio-Rad) and a 0.4 MNaOH transfer solution. A set of 7 replicate blots was prepared, oneblot for each oligonucleotide probe. A panel of 20mer probes, eachspecific for a single base change at either the first or second position ofhuman K-ras codon 12 or 13 (Clontech Laboratories), was used forsequence-specific hybridization of the transferred PCR. Oligonucleotide probes were end-labeled with [7-32P]ATP in a T4 polynucleotidylkinase labeling reaction. Following hybridization at 37°C,the blots

were subjected to stringent washes in a 3 M tetramethylammoniumchloride (Aldrich Chemicals, Milwaukee, WI) solution [3 Mtetramethylammonium chloride, 50 mm Tris (pH 8.0), 2 HIMEDTA, 0.1% sodium dodecyl sulfate] at 68°C.Blots prepared for codon 12 analysis

were stripped (0.1 x standard saline-citrate, 0.5% sodium dodecyl sulfate; 95°Cfor 20 min x 2) and reprobed for codon 13 mutations.

RESULTS

The development of pancreatic ductal adenocarcinoma in thehamster following BOP treatment is marked by a series ofwell-characterized changes in ductal morphology (16-18)(Fig. 1). The untreated normal ductal epithelium in weanlinganimals consists of a single layer of flattened cells surroundedby a connective tissue layer (Fig. la). The earliest change observed in response to carcinogen exposure is hyperplasiathroughout the ductal network (Fig. le). Such hyperplasia canbecome more severe focally and precedes the development ofpapillary hyperplasia, which is characterized by epithelial outgrowths, some of which are dysplastic, extending into the lumen (Fig. Id). These changes are observed at the level of themain duct, as well as its smaller tributaries including ductules.Evidence of elongated epithelial fronds with bridging across thelumen in ducts, which maintain an intact basement membrane,is diagnostic of CIS (Fig. le). The majority of carcinomas thatdevelop are classified as well-differentiated adenocarcinomas(Fig. If).

Examination of pancreata collected at 8-, 12-, 14-, 16-, and24-week intervals following the initiation of BOP treatmentrevealed changes in duct morphology that progressed in degreeand complexity with time. Organs collected at 8 weeks weremarked by mildly hyperplastic ducts with occasional papillarychange. Of the 50 lesions marked on sections cut from pancreata taken from 8 of 11 hamsters sacrificed at 12 weeks, approximately half were classified as hyperplasias (26 of 50), withpapillary hyperplasias and CIS samples compromising 24%(12 of 50) and 18% (9 of 50) of the samples, respectively. Threeprimary carcinomas were identified in sections from this timepoint and metastatic lesions were absent. By 24 weeks, both thenumber and complexity of the lesions had increased. Examination of tissue sections prepared from 14 animals sacrificed at 24weeks identified a total of 144 lesions of interest. Notable at thistime point was the frequent appearance of lymph node métastases (24 of 144, or 17%) and the increased incidence of carcinomas (24 of 144, or 17%). Preneoplastic lesions (hyperplasia,papillary hyperplasia, and CIS) occurred with approximatelyequal frequency (31 of 144, 34 of 144, and 31 of 144, respectively).

Lesions from every stage provided suitable template DNA foruse in PCRs, resulting in the amplification of a Ill-base pairproduct specific for the codon 12,13 region of the K-ras gene

(Fig. 2). Factors influencing the successful amplification of aparticular lesion included the size and cellularity of the tissuesample, as well as the length of formalin fixation prior to paraffin embedding. Although size alone was not always a reliableindicator of the suitability of a sample for PCR, very smalllesions (3-4 mm2) of the same stage were generally pooled from

a single slide if possible. The frequency of PCR success for eachcategory of lesions was different and, in general, reflected thecellularity of that lesion. For example, hyperplasias demonstrated the lowest frequency of PCR success, with only 46% ofthe areas scraped resulting in a K-ras specific PCR product. Ofall the lesions studied, however, this category had the fewestcells per total area, with the ductal lumen accounting for mostof the area within the region marked. On the other hand, papillary hyperplasias, which typically exhibited an increased degree of cellularity as the epithelial outgrowths extended into theductal lumen, had a higher rate of PCR success, with 66% ofthese lesions resulting in K-ras specific product. The highest

frequency of PCR positive lesions occurred in the CIS andinvasive carcinoma categories, where the cells usually filledmore of the field of interest. Tissue scraped in these groupsresulted in a K-ras specific product following PCR amplifica

tion in 74 and 78% of the samples, respectively. PCR successdecreased for metastatic lesions (50%), probably due to thecystic nature of some of these samples and the consequentialdecrease in cellularity. The fact that the frequency of PCRpositive samples is less than 100% in all cases can also beattributed to the very small size of some of the areas scraped(<5 mm2) and suggests that there is a lower limit to the samplesize for formalin-fixed paraffin-embedded tissue that is suitable

for PCR amplification. However, the most important factoraffecting the ability of tissue to provide suitable PCR templateDNA was the length of formalin fixation. In our experience,tissue that remained in formalin for more than 2 weeks provedto be unsuitable for use in the PCR amplification reaction andconsistently failed to generate any product. Tissue fixed for 1 to7 days worked well in the amplification reaction and as a result,shorter fixation times were chosen for samples collected at 14weeks in experiment II. Unfortunately, tissue collected in experiment I at the 8- and 16-week time points remained in formalin longer than 2 weeks and therefore was not available forPCR amplification and the subsequent molecular analysis.Overall, 64% of the lesions scraped from the 12-, 14-, and24-week slides resulted in a K-ras specific PCR product.

Lesions that generated a PCR product were subjected toanalysis of K.-ras mutational activation at codons 12 or 13using oligo-specific hybridization. Approximately equivalentamounts of PCR product (as judged by the intensity of ethidiumbromide staining) were loaded on slot blots and hybridized witha panel of 7 different oligomers, each specific for a single basechange at either the first or second position of codon 12 orcodon 13. Since DNA sequencing has revealed that this regionof the hamster K-ras gene is very highly conserved and closelymatches the human sequence (7, 8), we were able to use humanK-ras gene probes to screen our hamster PCR products. A

representative set of blots from such an analysis is shown inFig. 3. The blot hybridized with the probe specific for the G toA change at the first position of codon 12 (serine) is shown inFig. 3Cas an example of the level of background hybridizationthat was routinely observed. The serine probe usually demonstrated the highest degree of nonspecific binding within theoligomer panel. Fig. 3Äshows an example of a blot hybridized

4508

Research. on October 9, 2020. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

K-ras IN HAMSTER PANCREATIC CARCINOGENESIS

Fig. I. Photomicrographs of hematoxylin and eosin-stained sections of hamster pancreatic tissue from untreated control and BOP-treated animals. Lesions arerepresentative of those scraped and probed for K-ros mutations, a, normal pancreatic duct from an untreated animal lined by flat epithelial cells (x 188); b, "normal"pancreatic duct from a BOP-treated animal lined by cuboidal epithelium (x 188); c, hyperplastic duct in which the epithelium consists oftall basophilic cells (x 135);d, papillary hyperplasia with epithelial outgrowths of intensely basophilic and crowded cells (x 125); e, CIS (x 170);/, invasive well-differentiated adenocarcinoma(x 120); g, ductal adenocarcinoma metastatic to a lymph node (x 50).

with the aspartate probe for codon 12, and is representative of in the PCR products generated from the focal lesions. Strin-the range of signal intensities observed. gent wash conditions eliminated most background binding

Oligonucleotide-specific hybridization proved to be a sue- (Fig. 3C), and positive signals for mutations were identifiedcessful and reliable method for evaluating K-ras mutations after assessing the loading concentration of the particular

4509

Research. on October 9, 2020. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

K-i-oj IN HAMSTER PANCREATIC CARCINOGENESIS

1234

Fig. 2. PCR-mediated amplification of the first exon of the hamster K-ras gene(111 base pairs) from normal pancreatic duct and various BOP-induced preneoplastic and neoplastic lesions. A IO ni aliquot from a loo «IPCR was loaded ineach lane. Lanes I and 1, untreated pancreatic duct; Lanes 3 and 4, hyperplasia;Lanes 5 and 6, papillary hyperplasia; Lanes 7 and 8, CIS; Lanes 9 and 10,adenocarcinoma; Lanes 11 and 12, metastatic carcinoma; Lane 13, PCR carryovercontrol (no template DNA added); Lane 14, positive control (hamster pancreasDNA); Lane 15, *X174-RF Hae\\\ digest (500 ng).

sample as evidenced in the blot probed with the wild-type oli-gonucleotide (Fig. 3,4), and considering the degree of background binding for that particular set of blots. Backgroundbinding was estimated by comparing the intensity of the normalpancreas sample (slot 1A) signal on the mutated oligonucle-otide blots to its wild type oligonucleotide signal. This ratio wasthen taken as the threshold of background binding, and all othersample signal ratios had to exceed this number to be consideredpositive. The only mutations detected in all of the samplesanalyzed (regardless of stage) were G to A changes at the secondposition of codon either 12 or 13, with codon 12 mutationsoccurring much more frequently than those in codon 13. Ineither case, this mutation results in the substitution of an as-partate residue for a glycine in the ras p21 protein. Positivesignals were confirmed by laser densitometry.

Results from the codon 12 oligonucleotide-specific hybridization assay are summarized in Table 1, and are expressed asthe percentage of lesions scraped at each stage that were foundto contain a mutated K-ras oncogene. Within the group ofnormal duct lesions, a distinction is made as to the age of theanimals from which the samples were taken, although all of thespecimens in this category came from hamsters that were notexposed to carcinogen. The first group of 28 samples is from15 weanling hamsters, while the group designated as "age-matched" represents 16 samples from 10 hamsters approxi

mately 8 months old. These 2 different time points were chosento assess the background status of K.-ras activation in untreatedpancreatic ducts in animals of approximately the same ages asthose used in the BOP protocol at the beginning (weanlinghamsters) and end (8-month-old hamsters) of the study. Regardless of the age of the hamster from which the ductal samplewas taken, K-ras mutations were not detected in any of theuntreated ducts. The group of 15 samples referred to as BOP"normal" are areas scraped from ducts that exhibited a rela

tively normal morphology in animals that had received thecomplete regimen of BOP dosing. It should be noted that verymild hyperplasia is usually observed uniformly throughout the

pancreatic ducts as a result of BOP exposure, and that thesesamples are representative of such a morphology (Fig. Ib).Of 15 BOP "normal" samples assayed (from 12- and 14-weeksamples), only one demonstrated a K-/-M.Vmutation.

The remaining categories in Table 1 represent the classes ofpreneoplastic and neoplastic lesions previously described, andinclude samples from 12-, 14-, and 24-week specimens. Among

A

B

6

B

Fig. 3. Oligonucleotide-specific hybridization to detect K-ros codon 12 mutations. Slot blots were loaded with samples of PCR product amplified from normal, preneoplastic, and neoplastic lesions as follows: Row 1: A, normal hamsterpancreas; B, well-differentiated hamster transplantable adenocarcinoma; C,poorly differentiated hamster transplantable adenocarcinoma; D, extractionbuffer (no template DNA). Row 2: A, untreated pancreatic duct; C-E, hyperplasia.Row 3: A-E, papillary hyperplasia. Row 4: A-E, CIS. Row 5: A-E, adenocarcinoma. Row 6: A-E, metastatic carcinoma. Blot I was probed with an oligonucleotide specific for the wild type codon 12 sequence (GGT; glycine). Blot B wasprobed with an oligonucleotide specific for a G to A base change in the secondposition of codon 12 (GAT; aspartate). Samples in blot B that probed positive forthe aspartate mutation were: I B, 2E, 3B. 3E, 4A, 4C. 4D, 5A, SB, 5C, SE, 6A, and6D. Blot C was probed with an oligonucleotide specific for a G to A base changein the first position of codon 12 (AGT; serine). Blot C is shown as an example ofthe degree of background hybridization observed for the other oligonucleotideprobes, since it consistently gave the most intense background signal.

4510

Research. on October 9, 2020. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

K-ras IN HAMSTER PANCREATIC CARCINOGENESIS

Table 1 Detection ofK-ras codon 12 mutations in amplified DNA from variousstages of pancreatic ductal tumorigenesis

Table 2 Detection of K-ras codon 13 mutations in amplified DNA from variousstages of pancreatic ductal tumorigenesis

StageNormal

duct from untreated hamsters, weanlingNormal duct from untreated hamsters, 8 mo. old"Normal" duct from BOP-treated hamsters

HyperplasiaPapillary hyperplasiaCISPrimary adenocarcinomaMetastatic carcinomaNo.

oflesions

analyzed28

16152326343014No.

ofG to A

mutationsdetected(%)0

01(7)

6 (26)12 (46)26 (76)24 (80)

6 (43)

these groups, the frequency of K-ras mutations detected increased as the stage of the lesion became more advanced. Twenty-six % of hyperplastic lesions were found to contain anactivated K-ras gene, as compared to 46% of the papillary hyperplasia samples. The percentage of mutated K-ras positivelesions rose to 76% for the CIS group, and reached a maximumof 80% in the primary adenocarcinoma samples. Metastaticlesions demonstrated a K-ras mutation frequency of 43%. Thefrequencies for each group were found to be statistically significant from that of the untreated controls (X2 test, P < 0.01).

Within each category of lesions, mutations were found to occurin samples from various time points. Although a majority of thelesions in each group are from 24-week samples (70% hyper-

plasias, 68% of CIS, 60% of adenocarcinomas, and 86% ofmétastases),aspartate mutations were found with comparablefrequency in the 12- and 14-week lesions.

Table 2 summarizes the results from the oligonucleotide-specific hybridization assays for codon 13 changes. Of the 142preneoplastic and neoplastic lesions analyzed, only 9 werefound to contain a K-ras gene mutated at codon 13. All codon13 mutations were G to A transitions at the second position.Codon 13 mutations were detected in samples of hyperplasia(2 lesions), papillary hyperplasia (4 lesions), and CIS (3 lesions). Of these 9 lesions, 4 were found to have a K-ras mutation at the codon 12 position as well (1 hyperplasia, 1 papillaryhyperplasia, and 2 CIS samples).

DISCUSSION

Adenocarcinoma of pancreatic ducts is distinguished by having the highest reported frequency of ras activation of all humantumors studied thus far (19). Over the past 3 years, severalinvestigators have reported mutational activation of the K-rasgene in 80-90% of the human pancreatic ductal carcinomasanalyzed (10-15). Work in our laboratory (7) and of others(8, 9) has shown that BOP-induced pancreatic ductal carcinomas of the hamster also demonstrate K-ras gene mutations. Thefrequent activation of the ras oncogene in human and carcinogen-induced experimental animal pancreatic ductal adenocarcinomas has stimulated further interest in this oncogene and itsrole in the process of neoplastic transformation. The use of theexperimental hamster model permitted us to examine the ductal lesions that precede the development of gross tumors. Byapplying PCR amplification to DNA solubili/al from theselesions, we were able to examine the status of K-ras activationthroughout the transformation process, beginning with smallhyperplastic foci and proceeding through the more advancedstages of papillary hyperplasia and CIS. While the changes thatare observed in pancreatic ductal epithelium in response toBOP treatment certainly are distributed along a continuum andinclude numerous examples of mixed lesions, we imposed these

StageNormal

duct from untreated hamsters, weanlingNormal duct from untreated hamsters, 8 mo. old"Normal" duct from BOP-treated hamsters

HyperplasiaPapillary hyperplasiaCISPrimary adenocarcinomaMetastatic carcinomaNo.

oflesions

analyzed27

16IS20

20292712No.

ofG to A

mutationsdetected0

0024300

categories as a way to define those morphological changes thatappeared to be most consistently associated with progression tocarcinoma.

Our ability to successfully amplify the first exon region of theK-ras gene from focal scrapings was dependent upon severalfactors. The most critical one that affected all lesions was thelength of formalin fixation. Tissue that remained in fixative for2 weeks or longer proved to be unsuitable for PCR amplification. In our experience, the ideal length of fixation ranged from24 to 48 h, although tissue fixed for up to 1 week was usable.This effect has been noted by other investigators (20) and isbelieved to be due to the cumulative effect of nucleoprotein andnucleic acid cross-linking (21).

Analysis of the data summarized in Table 1 suggests that theK-ras mutation can be an early event during the process ofneoplastic transformation in the hamster pancreas. Detectionof mutated K-ras in hyperplasia samples demonstrates that thismutation accompanies the earliest detectable morphologicalchange following carcinogen exposure. The fact that the mutated oncogene is only found in 26% of the hyperplasia samplescan be interpreted in several ways. First, it is important torecognize that although the oligonucleotide hybridizationmethod of identifying mutations is very sensitive, it does have afinite limit of detection. It is possible that we were not able todetect K-ras mutations where they existed in only a very fewcells within a lesion. In preliminary experiments (data notshown), we determined that we could detect a mutated copy ofthe K-ras gene when it was present in approximately 5% of the

total DNA blotted. Therefore, we would not detect mutations inany sample where copies of the mutated K-ras alíelefell belowthis level, and our reported frequencies could represent an underestimate of the true K-ras mutation frequencies. One mightexpect that this would be most pronounced in the earliest lesions where, hypothetically, a very small percentage of cellswould harbor the mutations presumably caused by carcinogentreatment. Another interpretation of the initial low frequency ofK-ras mutations is that the hyperplastic response itself mayrepresent nonspecific cell replication in response to cell deathassociated with carcinogen treatment and only the mutationpositive lesions will proceed along a neoplastic pathway andeventually result in tumor development. Support for this ideacomes from the increasing incidence of K-ras mutations observed in each subsequent stage of ductal transformation examined in this study, suggesting that lesions that harbor a mutatedK-ras gene are being preferentially expanded. One possible scenario would propose that the K-ras gene becomes mutated atthe second position of codon 12 in response to BOP exposureand that these cells now exhibit some selective advantage thatallows for their expansion and enrichment throughout the neoplastic progression towards the development of a pancreaticductal adenocarcinoma.

4511

Research. on October 9, 2020. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

K-roj IN HAMSTER PANCREATIC CARCINOGENESIS

The detection of mutated K-ras in a single BOP "normal"

sample (1 of 15) raises the possibility that this change may beinvolved in the first steps of transformation, perhaps initiation,which precedes abnormalities in duct cell morphology; however, a larger number of samples would have to be analyzed toconfirm this finding. Unfortunately, these lesions are rare anddifficult to isolate in animals that have received the completeschedule of BOP treatment. The absence of mutations in 44negative control samples (ducts from untreated hamsters in 2different age groups) argues against the introduction of mutations as a result of Tag enzyme errors generated during PCRamplification.

The observed trend of increasing frequency of K-ras codon 12mutation accompanying advancing stages of neoplastic progression does not apply to the category of metastatic lesions.For these samples, a decrease in K-ras mutation frequency wasfound relative to the level seen in invasive carcinomas. Thisdecline may be due to several factors. The cystic nature andflorid lymphocyte population of these lesions could contributeto the dilution of cells harboring K-ras mutations to a pointbelow the sensitivity of the oligonucleotide hybridization technique. The lower incidence of K-ras codon 12 mutations mayalso be due to the random alleile loss of the mutated K-ras inmetastatic lesions, and would suggest a passive role for mutatedK-ras in metastasis.

K-ras mutations at the codon 13 position were rarely detected(Table 2). Four of the 9 samples that tested positive for codon13 mutations also contained codon 12 mutations. Because oftheir very low overall frequency and complete absence in anycarcinoma samples, we believe that this mutation is of littleconsequence in the BOP model of hamster pancreatic carcino-

genesis.Although the detection of an activated K-ras in samples of

ductal hyperplasia implicates this mutation as an early event inthe process of neoplastic transformation, the assignment ofsuch qualifiers as "early" or "late" are relative to what else is

known about the spectrum of molecular changes that are involved in this process. As the first study to address such aquestion in the hamster model of pancreatic carcinogenesis,this larger frame of reference is certainly lacking. However, wecan look to other models in which more extensive research hasresulted in the elucidation of molecular changes that accompany neoplastic progression to see that the activation of the rasoncogene as an early event is not unprecedented. H-ras mutations have been detected in a majority of premalignant papillo-mas arising in the mouse model of dimethylbenzanthracene-induced skin carcinogenesis (3), while K-ras mutations arefrequently observed in early to intermediate adenomas that precede colon carcinoma appearance in humans (1). In both ofthese cases, subsequent molecular events such as additionaloncogene activations or suppressor loss have been identified asrequirements for the development of the fully transformed phe-notype. Both H- and K-ras mutations were found in the mammary tissue of rats exposed to methylnitrosourea within 2weeks following carcinogen administration, yet these cells appeared to remain latent until exposure to estrogens (5). Therefore, ras activation is postulated to be an early, but not a singularly sufficient, event in transformation.

Recent studies that examine another tumor of the humanpancreas, the rare intraductal papillary neoplasm, report conflicting results on the status of K-ras mutation in these lesions.Tada et al. (22) found K-ras mutations in 3 of 5 cases of intraductal papillary neoplasm, and in 100% of adenocarcinomas

(19 of 19) (22). However, Lemoine et al. (23) failed to detectany K-ras activations in the 5 cases of intraductal neoplasm thatthey examined, although they also reported a high frequency ofK-ras codon 12 mutation in invasive adenocarcinomas (12 of16) and intraductal carcinomas with an invasive component(5 of 6) (23). This group also studied 9 cases of ductal papillaryhyperplasia, 5 taken from patients diagnosed with chronic pancreatitis and 4 from patients presenting with invasive adenocar-cinoma, and failed to detect any K-ras mutations. The authorsconclude that the absence of K-ras mutations in the papillaryhyperplasia lesions suggests that they may not represent a pre-neoplastic stage in the development of pancreatic ductal adenocarcinomas in humans, but acknowledge the importance ofthe mutations in the process of tumorigenesis. Fine needle aspirates of pancreatic masses were screened for K-ras mutationsby Shibata et al. (24). Their study detected activated K-rasoncogenes in 72% of patients with malignant cytology (18 of25), and 25% of patients with atypical cytology (2 of 8), butfound no mutations in patients with benign lesions or benignpancreatic disease. Although the detection of mutated K-rasproved to be a good marker for malignant disease, the study islimited to advanced lesions.

The fact that we observe only a single type of K-ras mutationin the ductal lesions, and that this change is found in hyperpla-sias, implicates this carcinogen as the mutagenic agent responsible for the aspartate mutation. BOP has been shown to be aDNA alkylating agent in the hamster pancreas, primarily producing methyl adducts at the N7 and O6 positions of guanine(25, 26). The guanine in the second position of K-ras codon 12(GGT) may be particularly sensitive to mutation due to flankingnucleotide residues. Two studies that analyzed mutations induced by ALmethyl-yV1-nitro-A'-nitrosoguanidine showed that

guanine residues were preferentially mutated if they had a guanine in the adjacent 5'-position (27, 28). This type of DNA

modification is consistent with the observed G to A transitionthat produces the aspartate mutation in the K-ras gene. Thesingular nature of the K-ras mutation found in adenocarcinomas and early lesions in the hamster model stands in contrast tothe range of mutations reported in human pancreatic adenocar-cinoma, where cysteine (8%), arginine (17%), valine (28%), andalanine (3%) substitutions have been reported in addition toaspartate changes (48%) (9-14). These findings reflect the factthat the mutagenic agents that are responsible for the K-raschanges associated with neoplasia and inflammatory diseasefound in the human pancreas are unidentified and probablynumerous. It is striking, however, that although there are a widerange of potential carcinogens that may be responsible for itsactivation, the K-ras gene is an especially sensitive target inboth the human and hamster pancreas. The use of the hamstermodel of pancreatic carcinogenesis allows us the opportunity tocarefully investigate the role of the K-ras oncogene in neoplastictransformation. This report establishes K-ras activation as anearly event in this process.

ACKNOWLEDGMENTS

The authors would like to thank Susan Hubchak and V. Subbarao fortheir expert technical assistance and Nancy Starks for her help in thepreparation of this manuscript.

REFERENCES1. Vogelstein, B., Fearon, E. R., Hamilton, S. R., Kern, S. E., Preisinger, A. C.,

Leppert, M., Nakamura, Y., White, R., Smits, A. M. M., and Bos, J. L.Genetic alterations during colorectal-tumor development. N. Engl. J. Med.,

4512

Research. on October 9, 2020. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

K-ros IN HAMSTER PANCREATIC CARCINOGENESIS

319: 525-532, 1988.2. Fearon, E. R., and Vogelstein, B. A genetic model for eolorectal tumorigen-

esis. Cell, 61: 759-767, 1990.3. Balmain, A., Ramsden, M., Bowden, G. T., and Smith, J. Activation of the

mouse cellular Harvey-ras gene in chemically induced benign skin papillo-mas. Nature (Lond.), 307: 658-660, 1984.

4. Greenhalgh, D. A., Welty, D. J., Player, A., and Yuspa, S. H. Two onco-genes, \-fos and v-raj, cooperate to convert normal keratinocytes to squa-mous cell carcinoma. Proc. Nati. Acad. Sci. USA, 87: 643-647, 1990.

5. Kumar, R., Sukumar, S., and Barbacid, M. Activation of ras oncogenespreceding the onset of neoplasia. Science (Washington DC), 248:1101-1104,1990.

6. Pour, P., Mohr, U., Cardesa, A., Althoff, J., and Kruger, F. W. Pancreaticneoplasms in an animal model—morphological, biological and comparativestudies. Cancer (Phila.), 36: 379-389, 1975.

7. Cerny, W. L., Mangold, K. A., and Scarpelli, D. G. Activation of K-ras intransplantable pancreatic ductal adenocarcinomas of Syrian golden hamsters.Carcinogenesis (Lond.), //: 2075-2079, 1990.

8. Fuji!, H., Egami, H., Chaney, W., Pour, P., and Pelling, J. Pancreatic ductaladenocarcinomas induced in Syrian hamsters by Ar-nitrosobis(2-oxopropyl)-amine contain a c-Ki-ras oncogene with a point-mutated codon 12. Mol.Carcinog., 3: 296-301, 1990.

9. van Kranen, H. J., Vermeulen, E., Schoren, L., Bax, J., Woutersen, R. A., vanlersel, P., van Kreijl, C. F., and Scherer, E. Activation of c-K-ras is frequentin pancreatic carcinomas of Syrian hamsters, but is absent in pancreatictumors of rats. Carcinogenesis (Lond.), 12: 1477-1482, 1991.

10. Almoguera, C., Shibata, D., Forrester, K., Martin, J., Arnheim, N., andPerucho, M. Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell, 53: 549-554, 1988.

11. Smit, V. T. H. B. M., Boot, A. J. M., Smits, A. M. M., Flueren, G. J.,Cornelisse, C. J., and Bos, J. L. KRAS codon 12 mutations occur veryfrequently in pancreatic adenocarcinomas. Nucleic Acids Res., 16: 7773-7783, 1988.

12. Grunewald, K., Lyons, J., Fröhlich, A., Feichtinger, H., Weger, R. A.,Schwab, G., Janssen, J. W. G., and Bartram, C. R. High frequency of Ki-rascodon 12 mutations in pancreatic adenocarcinomas. Int. J. Cancer, 43:1037-1041, 1989.

13. Mariyama, M., Kishi, K., Nakamura, K., Obata, H., and Nishimura, S.Frequency and types of point mutation at the 12th codon of the c-Ki-ras genefound in pancreatic cancers from Japanese patients. Jpn. J. Cancer Res., SO:622-626, 1989.

14. Nagata, Y., Abe, M., Motoshima, K., Nakayama, E., and Shiku, H. Frequent

glycine-to-aspartic acid mutations at codon 12 of c-Ki-ras gene in humanpancreatic cancer in Japanese. Jpn. J. Cancer Res., */: 135-140, 1990.

15. Tada, M., Yokosuka, O., Ornata, M., Ohio, M., and Isono, K. Analysis of rasgene mutations in biliary and pancreatic tumors by polymerase chain reactionand direct sequencing. Cancer (Phila.), 66: 930-935, 1990.

16. Pour, P., Althoff, J., and Takahashi, M. Early lesions of pancreatic ductalcarcinoma in the hamster model. Am. J. Pathol., **: 291-308, 1977.

17. Scarpelli, D. G., and Rao, M. S. Experimental Carcinogenesis of exocrinepancreas in the Syrian golden hamster (Mesocricetus Auratus). In: D. G.Scarpelli, J. K. Reddy, and D. S. Longnecker (eds.), Experimental PancreaticCarcinogenesis, pp. 45-64. Boca Raton, FL: CRC Press Inc., 1987.

18. Konishi, Y., Mizumoto, K., Kitazawa, S., Tsujiuchi, T., Tsutsumi, M., andKamano, T. Early ductal lesions of pancreatic Carcinogenesis in animals andhumans. Int. J. Pancreatol., 7: 83-89, 1990.

19. Bos, J. L. ras Oncogenes in human cancer: a review. Cancer Res., 49: 4682-4689, 1989.

20. Rogers, B. B., Alpert, L. C., Hiñe,E. A. S., and Buffone, G. J. Analysis ofDNA in fresh and fixed tissue by the polymerase chain reaction. Am. J.Pathol., 136: 541-548, 1990.

21. Lewin, S. Reaction of salmon sperm deoxyribonucleic acid with formaldehyde. Arch. Biochem. Biophys., 113: 584-602, 1966.

22. Tada, M., Ornata, M., and Ohio, M. Ras gene mutations in intraductalpapillary neoplasms of the pancreas. Cancer (Phila.), 67: 634-637, 1991.

23. Lemoine, N. R., Jain, S., Hughes, C. M., Staddon, S. L., Maillet, B., Hall, P.A., and Klöppel,G. Ki-ras oncogene activation in preinvasive pancreaticcancer. Gastroenterology, 102: 230-236, 1992.

24. Shibata, D., Almoguera, C., Forrester, K., Dunitz, J., Martin, S. E., Cos-grove, M. M., Perucho, M., and Arnheim, N. Detection of c-K-ros mutationsin fine needle aspirates from human pancreatic adenocarcinomas. CancerRes., 50: 1279-1283, 1990.

25. Lawson, T. A., Gingell, R., Nagel, D., Hiñes,L. A., and Ross, A. Methylationof hamster DNA by the carcinogen A'-nitroso-bis(2-oxopropyl)amine. CancerLett., //: 251-255, 1981.

26. Kokkinakis, D. M., and Scarpelli, D. G. DNA alkylation in the hamsterinduced by two pancreatic carcinogens. Cancer Res., 49: 3184-3189, 1989.

27. Burns, P. A., Gordon, A. J., and Glickman, B. W. Influence of neighboringbase sequence on iV-methyl-iV-nitro-A'-nitrosoguanidine mutagenesis in thelad gene of Escherichia coli. ¡.Mol. Biol., 194: 385-390, 1987.

28. Yang, J. L., Hu, M. C., and Wu, C. W. Novel mutational spectrum inducedby A'-methyl-iV-nitro-A'-nitrosoguanidine in the coding region of the hypox-anthine (guanine) phosphoribosyltransferase gene in diploid human libroblasts. J. Mol. Biol., 221: 421-430, 1991.

4513

Research. on October 9, 2020. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

1992;52:4507-4513. Cancer Res   Wendy L. Cerny, Kathy A. Mangold and Dante G. Scarpelli  Carcinogenesis in the Syrian Golden Hamster

Mutation is an Early Event in Pancreatic DuctrasK-

  Updated version

  http://cancerres.aacrjournals.org/content/52/16/4507

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/52/16/4507To request permission to re-use all or part of this article, use this link

Research. on October 9, 2020. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from