Pancreatic cancer in europe: Ki-ras gene mutation pattern shows geographical differences

Int. J. Cancer: 57, 167-171 (1994) 0 1994 Wiley-Liss, Inc.

-

-S

Publication of the internationalUnion Against Cancer Publicationde i Union Internationale Contre le Cancer

PANCREATIC CANCER IN EUROPE: Ki-ras GENE MUTATION PATTERN SHOWS GEOGRAPHICAL DIFFERENCES Aldo SCARPAIB4,Paola CAPELLI],Alberto VILLANUEVA~, Giuseppe ZAMBONI', Felix LLUW,Roberto ACCOLLA' , Gianmario MARIUZZI] and Gabriel CAP ELL;\^ 'Istituto di Anatomia Patologica and RIstitutodi Immunologia e Malattie Infettive, Universita di Verona, Policlinico Borgo Roma, 37134 Verona, Italy; 2Laboratorio de Investicacion Gastrointestinal, Av. Sant Antoni M" Claret, 167, 08025 Barcelona, Spain. Seventy-seven pancreatic adenocarcinomas (60 Spanish and 17 Italian) were tested for Ki-ras gene mutations by analysis of polymerase chain reaction amplified sequences. Mutations involving codon I 2 GGT; gly) were detected in I6 Italian and 46 Spanish cases (80. o in total). All Italian mutations involvedthe second base and were G to A transitions GAT; asp) in 8 cases and G to T transversions (GTT; Val) % in t e i remaining 8. Forty-two Spanish mutations w e 6 characterized. Thirty-eight were at the second and 4 at the fiirst base: asp in 24 cancers, val in 14, arg (CGT) in 2 and cys @G in 2. Previous European studies and our present data show t at I49 of the I86 pancreatic cancers harbored a codon I 2 Ki-ras mutation (8O%), the large majority affecting the second base (73Yo). with a transitions/transversions ratio of I .3: I. However, the mutational pattern of cancers of the different European countries shows remarkable differences, both in the site of the mutation (first or second base) and in the ratio of t:ransitionsover transversions. Moreover, a significant subgroup of pancreatic carcinomas do not harbor Ki-ras mutations. The classification of pancreatic cancers, according to the preseince or absence, and type of Ki-ras mutation, may be of imiportance in epidemiological studies. A critical reappraisal of existing epidemiological data, through a retrospective genotypic study using paraffin-embedded cancer samples, may reveal significant correlations with specific genotoxic agents.

b

4

8 1994 Wiley-Liss,Inc.

Distinct chromosomal and gene alterations are associated with specific tumor types, and geographical differences exist either in the frequence of a specific association (Amakawa et al., 1989; Tsuda et al., 1992) or in the pathogenetic mechanism of the abnormality (Pelicci et al., 1986; Vogelstein and Kinzler, 1992). Analysis of such genetic changes has provided important information for the understanding of the etiology and epidemiology of cancer. Comparative analysis of point mutations in oncogenes or tumor-suppressor genes, for example, has given clues concerning no1 only their endogenous or exogenous origin, but also possible etiologic factors (Tsuda et al., 1992; Vogelstein and Kinzler, 1992). Pancreatic adenocarcinoma is one of the most malignant neoplastic diseases in humans. Large variations in the incidence of this cancer are found among different countries (Warshaw and Castillo, 1992) as well as within the same country. In Italy, for example, regional variations ranging from 2 to 10 per 100,000 individuals can be observed. Epidemiological studies have failed to strictly correlate, in a significantly statistical manner, specific environmental factors such as diet and occupational exposure to particular chemicals (Warshaw and Castillo, 1992). One reason for such failures may be linked to the fact that pancreatic cancer cannot be efficiently subclassified on the basis of morphologic criteria alone. The possibility exists that specific genetic abnormalities found in pancreatic malignancies of the same or different ethnic groups may be a useful tool to better classify such cancers and to design more stringent epidemiological studies. Among the different gene alterations described in pancreatic cancer (Horii et al., 1992; Scarpa et al., 1993b), the high frequency of Ki-ras point mutations represents a constant feature (Almoguera et al., 1988; Griinewald et al., 1989; Mariyama er al., 1989; Motojima et al., 1991; Nagata et al., 1990; Smith et al., 1988). In the present paper, the pattern of

Ki-ras gene mutations in 60 Spanish and 17 Italian pancreatic cancers was studied and compared to that of cancers of other European countries. Remarkable differences were evident in the spectrum of point mutations of cancers arising in the various European countries. MATERIAL AND METHODS

The methods used to detect Ki-ras gene mutations varied according to the specific materials, experience and facilities available. Italian samples were frozen biopsies, from which high-molecular-weight DNA could b e obtained. In these cases, Ki-ras mutations were sought for in both exons 1 and 2 by the single-strand conformation polymorphism (SSCP) method of polymerase chain reaction (PCR)-amplified fragments (Scarpa et al., 1993b). The positive cases were further analyzed by single-allele direct sequencing of PCR-amplified fragments (Scarpa et al., 19938). Spanish samples were formalin-fixed and paraffin-embedded cytological fine-needle aspirates, from which only degraded DNA is obtainable. These cases were thus analyzed only for mutations in Ki-ras codon 12 by the artificial RFLP (restriction fragment length-polymorphism) method using a nested PCR and restriction enzyme analysis (Cappel16 et al., 1991) (see below).

Tumor samples and controls Italian samples. Seventeen biopsies were collected at the time of surgery from the same number of patients with pancreatic adenocarcinomas (8 males, 9 females; median age 60years). Before D N A extraction, 5 k m H. and E.-stained sections were examined to evaluate the ratio of neoplastic to normal tissue. All samples contained a proportion of cancer cells ranging from 5 to 80%. In 13 cases, normal tissue from the same patients was also available and used as an internal control. Spanish samples. Sixty fine-needle aspirates, in which the neoplastic cells accounted for at least 10% of the entire cell content, were selected for the study. The aspirates were obtained percutaneously under ultrasound or computerized tomography scan guidance from the same number of patients with pancreatic cancer (34 males, 26 females; median age 54 years). The samples were formalin-fixed and paraffin-embedded, then 6-pm H. and E.-stained section were examined to evaluate the ratio of neoplastic to normal cells. D N A was extracted from 3 to 5 adjacent sections. DNA preparation High-molecular-weight D N A was extracted from Italian frozen samples according to standard procedures. Paraffinembedded Spanish samples were deparaffinized with 500 p1 of xylene, washed twice in ethanol, dried, incubated overnight at 37°C in 50 pl of digestion buffer (100 mM Tris p H 8.2, lOmM EDTA, 100 kgiml proteinase K), boiled for 10 min and &Towhom correspondence and reprint requests should be sent, at the Istituto di Anatomia Patologica, Universita di Verona, Policlinico Borgo Roma, 1-37134Verona, Italy. Fax: (39) (45) 80 98 136.

Received: October 8, 1993 and in revised form December 7,1993.

168

SCARPA ET AL. Ki-ras exon 1

centrifuged. One microliter of the supernatant was used for PCR amplification.

Kt

Artifcial RFLP method for detection of Ki-ras codori 12 mutations Analysis of Ki-ras codon 12 mutations was performed by a modified artificial RFLP method (Cappelli et a/., 1991). The method is based on PCR amplification of a Ki-ras sequence, including codon 12. A specific mutation, alone o r in conjunction with an artificial mutation introduced by using a mutated primer, creates a new restriction enzyme site in the amplified sequence. After appropriate digestion, the pattern of the bands visualized by ethidium bromide staining of a 12% polyacrylamide gel is diagnostic for the presence and type of mutation. A nested PCR approach, using the primers listed in Table 11, has been used to maximize the sensitivity. Since all the described codon 12 (-coding for glycine) Ki-ras mutations in pancreatic cancer involve the first or second base, the strategy was as follows. Samples were first screened for the presence of mutations (Fig. 1). A 217-bp D N A sequence of the first coding exon of the Ki-rus gene was amplified in the first PCR using K1 and K7 primers for 20 cycles (92°C for 64 sec, 50°C for 35 sec and 72°C for 85 sec). One microliter of the PCR product was used as template for the second 35 PCR cycles (92°C for 15 sec, 50°C for 15 sec and 72°C for 30 sec), using the nested primers K2 and K3. Five milliliters of the PCRamplified samples were digested with HphI restriction enzyme, run in a 12% polyacrylamide gel and visualized by ethidium bromide staining. The presence of mutations was determined after consistent results were obtained in at least 2 independent experiments. In positive cases, mutations were characterized by using the appropriate restriction enzymes on the amplified

Amplified

Codons

rreion

analwed

Kla. GGCCTGCTGAAAATGACTGA Klb, GTCCTGCACCAGTAATATGC exon 1 4 to 37 (162 bP) ma, TTCCTACAGGAAGCAAGTAG K2b, CACAAAGAAAGCCCTCCCCA exon 2 46 to 74 (128 bD)

KI, AACCTTATGTGTGACATGTTCTA K2; CCTGGTGAAAATGACTGAAT K3, AGGCACTCTTGCCTACGTCA K4. ACTTGTGGTAGTTGGAGGTG

'The base underlined represents the mutation introduced.

Q

K2

-G-

5'

-GCTGA

3'-

K7

12

CGACT

-

+T

K3

8B Fi

-GGTGA -CCACT

DIGESTION

CCA CT

Hph I

Hph I

65

16

49 40

16

U M N

9

FIGURE 1 - Diagram of artificial RFLP approach for detection of Ki-ras codon 12 mutations. After a first amplification (PCR 1)of the first coding exon of the Ki-ras gene using K l and K7 primers (Table 11), an aliquot is reamplified using nested mutant primers K2 and K3 (PCR 2). The K3 down-stream primer, which contains an A instead of a G at the second position of codon 13, generates a new restriction site for HphI that will be destroyed whenever a mutation occurs in the first 2 Gs of codon 12. Also, the K2 up-stream primer is modified in order to introduce an HphI site, which is used as an internal control to test the completion of digestion. After digestion of the 65-bp amplified fragment, the samples are run in a 12% acrylamide gel and stained with ethidium bromide. The pattern of bands is diagnostic for the presence of mutations. U is the uncut 65-bp DNA-amplified product, M is the 49-bp DNA fragment that identifies the mutant allele, and N is the 40-bp fragment that represents the normal allele.

IN FIGURE 1

Mutation

Primers1

$

~

Ser ( A G T ) K2. K3 66 bo A r g i m ) K2: K3 66 bb Ala K2, K3 66 bp Val (m) K2, K3 66 bp ASD(GAT) K4. K5 109 bD Cy;(mjK6, K5 112 bp

(a)

~

$

~

Rertriction site

~

Expected

~ bands

Mae1 (CTAG) 43 and 23 Sac1 (GAGCTC) 38 and 28 40 and 26 BbvI (GCAGC) HincII (GTYRAC 43 and 23 HDhI (GGTGA(NIK\93 and 16 Hind111 (AAGC'fi) 94 and 17

'See Table 11. products from the same I(2 and K3 primers to detect serine, arginine, alanine and valine mutations; K4 and K5 to detect aspartic acid mutation; K6 and K5 to detect cysteine mutation (Table 111).

Statistical analysis Statistical analysis was performed by comparing the confidence intervals of the observed frequencies of Ki-ras mutations (Table IV) and of specific amino-acid substitutions in the diverse countries (Table V). The statistical significance for specific mutations among all countries has been evaluated by the x2 test for multiple comparisons, and in the case of comparisons for a specific mutation between 2 countries by the x2 test with Yates' correction for continuity (Glantz, 1987).

~ 5AACAAGATTTACCTCTATTG ;

KO. AAACTTGTGGTAGTTGGAAGC K7. GTCCTGCACCAGTAATATGCA

5. PCR 1

3'

TABLE I - PRIMERS USED TO PCR-AMPLIFY Kt-ms SEQUENCES (5'-3') FOR SSCP ANALYSIS Primers

-

3'

*' '

PCR-SSCP analysis and direct DNA sequencing PCR-SSCP analysis was performed as described (Scarpa et al., 1993b). Briefly, 30 cycles of PCR were performed with [y-7'P]-ATP 5'-end-labeled primers (Table I). A portion of the PCR mixture was denatured and run on 6% polyacrylamide gels, which were dried on filter paper and exposed to X-ray film at -80°C for 1 hr with an intensifying screen. The D N A samples showing mobility shifts of single-stranded D N A fragments were subjected to direct sequencing. In order to enrich the mutated sequences, the shifted bands were eluted from the gel, amplified, and then sequenced as previously described (Scarpa et al., 19930). Wild-type bands of the same cases were also sequenced as an an internal control.

RESULTS

SSCP analysis of amplified D N A fragments from Ki-ras exon 1 showed abnormally migrating bands in 16 Italian cases

169

KI-ras MUTATIONS IN PANCREATIC CANCER TABLE IV - FREQUENCE OF Ki-,as MUTATIONS IN EUROPEAN COUNTRIES Reference

Po\lt~ve/sample~ 7o tP'tPA

Country

Netherlands Austria Italy Spain BelgiumiEngland

28/30 47163 16/17 46160 12/16

93 75 94 77 75

Smitetal., 1988 Griinewald et al., 1989 Thisstudy Thisstudy Lemoine et a!., 1992'

'In this report the exact geographic origin of patients is not indicated. TABLE V - CODON 12 (GLY) Ki-ras MUTATIONS IN DIFFERENT EUROPEAN PANCREATIC CANCERS Netherlands

Val Asp Al a

3 Ser

:)%I

\

-

I

Italy

:$::{ ~I

-

l;j2) -

Austria

-

L8 (38) 14(30) l(2)' L4 (30) -

Spain'

B/E2

24 57) 14)33)

2 17) 3 125)

if:] -

2 (17)

-

asp) in 24 cancers, G to T transversions ( G G T -+GTT; gly -+ Val) in 14 ( E T + E; gly arg) in 2 and (G + E; gly + cys) in 2. The 2 cases with a double mutation showed aspartic acid plus cysteine and aspartic acid plus arginine substitutions, respectively. Our results, together with those of other European studies, are summarized in Tables IV and V. No substantial differences in the overall frequency of Ki-ras mutations among different countries were found when comparing the confidence intervals (Table IV), whereas for specific mutations (Table V) a statistically significant difference was found only for the frequency of arg substitution ( p < 0.001). When different countries were compared in pair-wise combination, statistically significant differences were observed as follows: Spain vs. BelgiumiUK: asp ( p < 0.05); Netherlands vs. Austria: arg @ < 0.05); Austria vs. Spain: arg (p < 0.01); Netherlands vs. Spain: cys ( p < 0.01). --f

-

5 (41) = cys; A x = ser: First base mutations are CGT = arg; second base mutations are GAT = asp; C T = Val; E T = ala.-'Two cases showed a double mutation, asp plus cys and asp plus arg, respectively.-2Belgian/English cases. The exact geographic origin of patients is not indicated in the report.

FIGURE 2 - SSCP analysis of PCR-amplified Ki-rus exon 1 from normal tissue (N) and cancer biopsies (numbers) of 7 Italian cases. Shifted bands in addition to the normal ones (N), corresponding to mutated sequences, are present in 2111 cancers except case 13. Since single-strand sequences assume different conformations, in nondenaturing polyacrylamide gels, according to their base sequence, the same shift at SSCP analysis corresponds to the same sequence. In this example, there are 2 different sequences, both corresponding to a codon 12 mutation, as demonstrated by direct sequencing shown in Figure 3, i e . GGT to GAT in cases 14, 15 and 17, and GGT to GTT in cases 1T;SZandTCCase numbers correspond to those presented in a previous report (Scarpa et ul., 1993~).

(94%). Representative results art: shown in Figure 2. All cases were negative at SSCP analysis for the Ki-ras exon 2. All of the shifted bands at SSCP analysis were due to mis-sense point mutations affecting the second base of codon 12 (Fig. 3). The mutations were G to A transitioris ( E T + GAT; gly + asp) in 8 cases, and G to T transversions ( G G T + T T ; gly + Val) in all the others. No double mutations were observed. The somatic nature of the mutations could be demonstrated in 10 cases by their absence in the available normal tissues from the same patients. In Spanish cases, only codon 12 was analyzed (Fig. 4). Mutations were detected in 46 of the 60 pancreatic carcinomas (76.6%). In all, 48 mutations were detected, since 2 cases showed double mutations. In 6 cases it was impossible to characterize the mutations detecled by the diagnostic artificial RFLP approach using K2 and K3 primers and HphI digestion, since the characterization approach using different primers and restriction enzymes gave no clear results. Thirty-eight mutations were at the second base and the remaining 4 at the first base. They were G to A transitions ( G G T + GAT; gly -+

DISCUSSION

Different genetic alterations may be detected in pancreatic carcinomas both at the chromosomal (Johansson et a/., 1992) and the gene level (Horii et al., 1992; Lemoine et al., 1992; Scarpa et al., 19936). However, only the somatic mutational activation of the Ki-ras gene represents a constant feature of the large majority of pancreatic cancers. Indeed, a number of studies provided evidence of a high frequency of Ki-ras mutations, irrespective of the method used. There is only one discordant study (Gonzales-Cadavid et al., 1989), in which analysis by PCR direct sequencing resulted in a low frequency (17.6%) of Ki-ras mutations. In all the positive cases in which the entire exon 1 was analyzed (Almoguera et al., 1988; Mariyama et al., 1989), the mutations were located at codon 12. Additional reports focused attention on codon 12 alone or in conjunction with codons 13 and 61 (Smit et al., 1988; Griinewald et al., 1989; Nagata et al., 1990; Motojima et al., 1991; Lemoine et al., 1992). No mutations in codons 13 and 61 have been reported, except in one case in which point mutations in both codons 12 and 13 were detected (Nagata et al., 1990). Our data on Italian cancers extend the knowledge on Ki-ras gene mutations in pancreatic cancer and reasonably exclude the presence of mutations outside of codon-12. This conclusion is based on the results of SSCP analysis, which was selected because of its ability to detect all possible mutations in the first and second bases of codon 12 (Scarpa et al., 19936), as well as many mutations in exons 1 and 2 of the ras gene family, including codons 13, 18 and 61 (see Scarpa et a/., 19936). Specific codon 12 Ki-ras mutations are regarded as critical, possibly the initial event in pancreatic oncogenesis, due to their extremely high frequency and to their presence in pre-invasive pancreatic cancer (Lemoine et aZ., 1992). Experimental data in animal model systems strongly support the notion that a carcinogen may be the mutagen which, after hitting the rus gene, initiates the oncogenic process (Balmain and Brown, 1988). Furthermore, the introduction in the germ line of a mutated H a m s gene, placed in front of the pancreasspecific elastase gene promoter, leads to the development of pancreatic tumors immediately after expression of the chimeric gene at the late stage of embryogenesis (Quaife et al., 1987). However, the possible role of chemical carcinogens in the induction of ras mutations in human pancreatic cancers has yet to be determined. In this context, it is worth noting that examples of reproducibly defined mutations in a specific gene associated with a specific human tumor do exist. This is the case of p53 gene mutations found in 2 types of human cancer (Vogelstein and Kinzler, 1992): (i) liver cancers of subjects exposed to food-contaminating aflatoxin BI, which showedp.53 mutations similar to those caused by this substance in mutagenesis experiments and (ii) skin cancer in light-exposed areas

170

SCARPA ETAL.

FIGURE3 - Anti-sense sequences of the mutated bands of cases 11 and 15, eluted from the gel shown in Figure 2. N is the sequence of the SSCP upper normal band of case 11, eluted from the same gel. In both cases the mutation involves the second base of codon 12.

FIGURE4 - Artificial RFLP approach for detection of Ki-rm codon 12 mutations. The samples were run in a 12% acrylamide gel and stained with ethidium bromide. U is the uncut 65-bp DNA amplified product; M is the 49-bp DNA fragment that identifies the mutant allele; and N is the 40-bp fragment that represents the normal allele. DNA samples are respectively: P23, a benign pancreatic mass; P25, P30, P24 and P27, fine-needle aspirates of pancreatic carcinomas showing a codon 12 Ki-ras mutation. H20 is a negative control corresponding to the PCR reaction mixture with no addition of DNA.

which harbor p53 gene mutations similar to those caused by ultraviolet light. Analysis of the specific nucleotide changes at codon 12 of the Ki-ras gene in human pancreatic carcinoma reveals mutations of a heterogeneous type. The heterogeneity of the Ki-ras mutations is not consistent with the interpretation that a single specific carcinogen is the causative factor. Rather, the distinct mutations are probably due to different exogenous or endogenous carcinogens. Only well-conducted epidemiological studies taking into account the working history, diet and presence and type of ras mutation may help in resolving this issue.

Taken together, previous European studies (Griinewald et al., 1989; Lemoine et a/., 1992; Smit et al., 1988) and our present data, demonstrate that 149 of the 186 pancreatic cancers analyzed showed a codon 12 Ki-ras mutation (80%). The large majority of these mutations affected the second base of codon 12 (109/149) (73%), and the ratio of transitions/ transversions was slightly in favor of the former (61:48) (1.27:l). However, remarkable differences are evident when the patterns of mutations in cancers of the different European countries are compared (Table V), both in the site of the mutation (first or second base) and in the ratio of transitions over transversions. In Italian cases, 2 types of mutation were observed, asp and Val, whereas in cancers from 2 other European countries a third type of mutation was detected with a high frequency, arg in Austrian (Griinewald et al., 19893 and cys in Dutch (Smit et al., 1988) cases, each of which was present in about one-third of the cases. In another European report, in which Belgian and English cases were studied, the presence of ser mutations at high frequency has been reported (Lemoine et al., 1992). The only other ethnic group in which ras mutations in pancreatic cancer have been extensively studied is Japanese (Mariyama et al., 1989; Motojima et al., 1991; Nagata et aL, 1990; Scarpa et al., 1993a). Among 135 Japanese pancreatic cancers, 121 (89.6%) showed codon 12 Ki-ras mutations, most of them at the second base (109/121) (90%) with a transition/ transversion ratio of 2.3:l (76:33). The specific mutations were asp, val and arg in 67%, 26% and 7% of cases, respectively. The different methods used to detect mutations do not, in our opinion, account for the above-mentioned differences. In fact, the results obtained by the different authors in Japanese cases (the largest homogeneous ethnic group analyzed to date) are concordant, in both frequency and type of mutations, in spite of the different methods or materials (frozen o r formalinfixed) used (Mariyama et ul., 1989; Motojima el al., 1991; Nagata et al., 1990; Scarpa et al., 1993a). It is conceivable, therefore, that international differences in the pattern of mutations may reflect ethnic pecularities associated with distinct environmental and/or genetic factors. Moreover, the fact that a significant proportion of pancreatic carcinomas do not harbor mutations in any of the 3 ras genes (Griinewald et al., 1989; Smit et aL, 1988) suggests that a subgroup of these cancers might also develop through a pathway of genetic alterations which d o not involve mutations at codon 12 of the Ki-ras gene.

171

KI-m\ MUTATIONS IN PAh'CR€-.AIIC CANCER

The classification of pancreatic cancers into subgroups, according to the presence or absmce as well as to thc type of Ki-rus mutation, may be of importance in epidemiologic studies. It is possible that a critical reappraisal of existing cpiderniological data. after a retrospcctive genotypic study o f paraffin-embedded cancer samples retrieved from the archives of pathology departments, may reveal significant correlations with specific genotoxic agents. Future investigations employing newly developed molecular biological methods to identify genetic alterations, combined with methods of characterizing exposure to endogenous and exogenous carcinogens and their early cffccts (Wogan, 1992), have great potential for further elucidating the role of gcnotoxic agents in the etiology of

human pancreatic cancers and for the development of strategies for their prevention.

ACKNOWLEDGEMENTS

This work was supported by (a) the Associazione Italiana Ricerca Cancro (AIRC), Milan, Italy (grants to L.F.-D. and to R.A.); the Italian Ministry of Scientific Research, Rome, Italy (MURST, 40% and 60%); and CNR, Rome, Italy (PF Biotecnologic c Biostrumentazione to R.A.) and (b) the Spanish Ministry of Sciences (MEC), Madrid, Spain (Grant PM900057 from DGICYT).

REFERENCES A1 MOClUEKA,

c.,SHIBAIA.D.. FORRtsI'EK. K.. .WARTIN. J., ARSIIEIM. MOTOJIMA,K., URASO,T., NACAI'A,Y.. SHIKU,H.. TSUXODA, T. and

1 1 0 . M.. Most human carcinomas of the exocrine KASEMATSU. T., Mutations in the Kirsten-ras oncogene are common n mutant c-K-rus genes. Cell, 53,549-554 (1988). but lack correlation with prognosis and tumor stage in human carcinoma. Amer. J. Gusfroenferol.,86, 1784-1788 (1991). ALIAKAWA. R.. FLKUHAKA, S., OHSO. H., DOI. S., OWMA,S.. T A ~ A H E pancreatic . S.. Y/\MAHI:, 11.. EDAMLJKA, S.. TOMONO. N., NASIJ.K.. KOKAKA. Y., NAGATA, Y., AHE,M., MOTOSHIMA, K., NAKAYAMA. E. and SHIKL,It., SHIOMUKA. T.. ABF.. M.H.W. and U~:IIINO, II., Involvement of hcl-2 Frequent glycine-to-aspartic acid mutations at codon 12 of c-Ki-ras gene in Japanese follicular lymphoma. Blood. 73,787-791 (1989). gene in human pancreatic cancer in Japanese. /up. J. Cancer Res., 81, 135-140 (1990). BALMAIS,A. and BROWS, K.. Oncogene activation in chemical PEI.ICCI,P., KSOWLES,D., MAGRATH,I. and I)AI.I.A FAVERA.R., carcinogenesis. Advanc. Cancer Res.. 51, 147-182 (1988). CAPPELLA, G., CRONALL.K-MITRA. S., P E I S A I )M. ~ , and PERUCHO, M.. Chromosomal breakpoints and structural alterations of the c-mn,yc Frequency and spectrum of mutations at codon 12 and 13 of the locus differ in endemic and s oradic forms of Burkitt lymphoma. Proc. c-K-rusgene in human tumors.Enviro/z. IIlthPtmp., 93, 125-131 (1991). nut. Acud. Sci. (Wash.), 83,2&4-2988 (1986). QUAIFE, C . , PINKERT, C., ORsrrz, D., PALMITER. R. and BKINSTER, R.. Gi.AsT%,S.A..Primer ofhiosfafisfics.hlcGraw-liill, New York (1987). Pancreatic neoplasia induced by ras ex ression in acinar cells of GONZALES-CADAVID, N.F., ZHOC:.D.,BAI-I-IFOKA, H., BAR-ELI,M. transgenic mice. Cell, 48, 1023-1034 (1987f .. and CLINE,M.J.. Direct sequencing analysis of exon I of the c-K-ras SCAKPA, A., CAPELLI, P., MLKAI.K., ZAMBOSI, G.. ODA,T,, IAC:OSO, gene shows a low frequency of mutations in human pancreatic C . and I ~ W H A S H IS., , Pancreatic adenocarcinomas frequently show carcinomas. Clncogene. 4, 1137-1 140 (1989). p53 gene mutations. Amer. J. Pufhol., 142, 15341.543 (19930). C;KcUEWAI.I>. K.. LYONS, J., FKOHLICIi, A,, FEICHTIN(iER, H., W ~ G E K , P., ZAMBONI, G., ODA,T.. MCJKAI, K., Bos~rri, R.A., SCIIWAB,G., JANSSF.S,J.W.G. and BARTRAM.C.R.. lligh SCAKPA. A,. CAPELLI. S.. frequency of Ki-rus codon 12 mutations in pancreatic adenocarcino- F., MAKI'IGNOSI,G.. IACONO,C., SERIO, G. and HIKUHASHI, Neoplasia of the ampulla of Vater: Ki-rus andp53 mutations.Amer. J. mas. In,. J. Cuncer, 43, 1037-1041 (19K9). P ~ f h o l 142,1163-1172 .. (19936). HOKII,A,. NAWISCRU IYOSHI. Y., Ictill, S., NAGASE. H.. ANDO, S ~ 1 . r V.. . BOOT, A., SMI.IS, A,, FLEUKEN. G.. CORNEI.ISSt., C.J. and Bos, H., YANAGISAWA, A , , IYA, E.. KATO, Y. and S A K A MY.. ~~RA. J.L., K-rus codon 12 mutations occur very frequently in pancreatic Frequent somatic mutations of the APC gene in human pancreatic adenocarcinomas. Niccl. Acid7 Res., 16,7773-7782 (1988). cancer. Chncer Rex., 52,66964698 (1992). S., BIIUDHISAWASDI. V., KJHANA, T., SUGIMURA, JOIIANSSON. B., BAKDI.G . , I I E I M . s.. MAS1)AHL. N., MERTESS,F.. TSLDA,[I., SATARUG, and HIKOHASIII. S., Cholangiocarcinomas in Japanese and Thai B A K - J E N S EE., ~ , ASI>KL~S-SASDHERG, A. and MITELMAS,F., Nonran- T. patients4ifference in etiology and incidence of point mutations of dom chromosomal rearrangements in pancreatic carcinomas. Cuncer, the c-Ki-ras protooncogene. Mol. Carcinogen., 6,266-269 (1992). 69,1674-1681 (1992). L ~ U O I N N., E . J A I NS., , HUGIIES, C., !hADDON, S., MAII.I.ET,R., HALL, VO(iEISTEIS, B. and KINZLLK,K., Carcinogens leave fingerprints. G., Ki-ras oncogene ixtivation in preinvasive pancre- Nature (Lond. ), 355,209-210 (1992). P. and KL~PPEI.. atic cancer. Gusrroenfero/oQp.102, 230-236 (1992). WARSHAW, A . and CASTII.l.0. C.F.-D.. Pancreatic carcinoma (review). MARIYAMA, M., KISHI.K., NAKAMLKA, K.. OBAIA,11. and NISHIMLKA,New Engl. J. Med., 326,455-465 (1992). S.. Frequency and types of point mutation at the 12th codon of the WOCAS, G., Molecular epidemiology in cancer risk assessment and c-Ki-ra gene found in pancreatic cancers from Japanese patients. Jup. r n t i o n - r e c e n t progress and avenues for future research. Exp. J. Cuncer ReA., 80, 622-626 (1989). LO/.Med.. 98. 167-178 (1992).