p53 Alteration in Gastric Precancerous Lesions - NCBI

American Journal ofPathology, Vol. 144, No. 3, March 1994Copyright © American Societyfor Investigative Patbology

p53 Alteration in Gastric Precancerous Lesions

Yih-Horng Shiao,* Massimo Rugge,tPelayo Correa,* H. Peter Lehmann,* andW. Douglas Scheer*From the Department ofPathology,* Louisiana StateUniversity Medical Center, New Orleans, Louisiana; andCattedra di Istochimica ed Immunoistochimica,t Universitadegli Studi di Padova-Servizio di Anatomia Patologica,Cittadella, Padua, Italy

It has beenpostulated that chronic atrophic gas-tritis, intestinal metaplasia, and dysplasia areprecancerous stages of stomach tumorigenesis.We investigated the timing ofp53 alterations inthese events ofgastric tumorigenesis. Each of12cases ofarchived tissue containingprecancerousand cancerous lesions were selectedfor the de-tection ofp53 alterations. Accumulation ofp53protein was detected by immunohistochemistry.Exons 5 to 8 ofp53 gene were examinedfor mu-tations by polymerase chain reaction-singlestrand conformation polymorphism andDNA se-quencing. p53 immunoreactivity was detected in60% ofthe dysplasia cases and in 60% ofthe caseswith carcinomas.p53gene alterations werefoundin 37.5% ofthe metaplasia cases, 58.3% ofthe dys-plasia cases, and 66.7% of the cases with carci-nomas. In 71% of the cases, mutations wereshown as G:C -- A:T transitiont We conclude thatmutation ofthep53gene is an early event in stom-ach tumorigenesis. (AmjPathol 1994, 144:511-517)

In 1980, stomach cancer was a leading cancerthroughout the world, with 682,400 new cases peryear.1 It is the second most frequent cancer in de-veloping countries and the fourth most frequent can-cer in developed countries.2 In the United States,cancer of the stomach is the eighth most commoncause of cancer deaths. Although the mortality ratehas decreased from 22.8 per 100,000 in 1950 to 9 per100,000 in the 1980s in white males and from 12.3 to4.3 per 100,000 in white women in the United States,3the overall five-year survival rate is only 5 to 15%.4 Thereason for the low survival rate of stomach canceris related to the fact that the disease is typically

diagnosed at an advanced stage.5 Therefore, an un-derstanding of phenotypic and genotypic events instomach tumorigenesis is important in terms of earlydetection, selecting appropriate treatment, andhence preventing the progression of this disease.The most common gastric cancer in populations at

high risk is the so called intestinal type, in which themalignant tissue resembles glands of the gastroin-testinal tract.6 This type is preceded by a chain ofmorphological events, namely, chronic gastritis, at-rophy, metaplasia of the small intestinal and colonictypes, dysplasia, intramucosal (early) carcinoma andinvasion.7 p53 gene mutations have been reported inthe dysplasial and carcinoma9'10 stages. The accu-mulation of p53 protein has also been detected indysplasia1l and carcinoma.12'13 Patients with p53protein accumulation in gastric carcinoma showeda worse survival than those without p53 accumula-tion.12.13

To investigate the timing of p53 alterations in theprocess of gastric tumorigenesis, areas representingmorphologically normal cells (mucosa, lymphocytes,or muscle), intestinal metaplasia, dysplasia, and car-cinoma from each gastrectomy specimen were mi-crodissected from formalin-fixed, paraffin-embeddedsections and submitted for detection of p53 gene mu-tations by polymerase chain reaction-single-strandconformation polymorphism (PCR-SSCP), and DNAsequencing. p53 protein accumulation was detectedby immunohistochemistry (IHC).

Materials and Methods

Study SubjectsTwelve gastrectomy cases were selected from agroup of Italian patients enrolled in a prospectivestudy on gastric epithelial dysplasia.14 Morphologi-cally normal cells (mucosa, lymphocytes, ormuscle), gastric epithelial dysplasia, and gastric

Supported by NIH grant PO1-CA28842.Accepted for publication November 22, 1993.

Address reprint requests to Dr. Pelayo Correa, Department ofPathology, Louisiana State University Medical Center, 1901 PerdidoStreet, New Orleans, LA 70112.

511

512 Shiao et alAJP March 1994, Vol. 144, No. 3

carcinoma were obtained from all 12 patients. Intes-tinal metaplasia was identified only in eight cases.After performing a 5-p section for hematoxylin andeosin staining, five serial 10-p sections were cut forDNA extraction. An additional 5-p section was pre-pared for p53 immunohistochemical determination.

IHC

Sections were dewaxed in xylene (twice for 5 min-utes each) and rehydrated through serial ethanol(100%, 95%; two changes of 3 minutes each) todistilled water (twice for 5 minutes each). A poly-clonal antibody against both wild-type and mutantp53 protein (CM-1; Signet Laboratories, Inc.,Dedham, MA)15 was used at a working concentra-tion of 1:25 (v/v) dilution at room temperature for 20minutes. Hydrogen peroxide, normal goat blockingserum, biotinylated immunoglobulins, avidin-biotincomplex, and 3-amino-9-ethylcarbazole substratesolution were used according to the instructionsfrom the detection kit (Signet ELITE avidin-biotin de-tection system; Signet Laboratories, Inc.). Sectionswere lightly counterstained with Mayer's hematoxy-lin, mounted with aqueous mounting medium (Crys-tal/Mount; Biomeda, Foster City, CA), and post-mounted with Postmounting Crystal/Mount inPermount (Biomeda). Positive control sections fromformalin-fixed, paraffin-embedded colonic cancerswere cut and prepared in the same manner as thespecimens. In serial sections from each case tis-sue, primary antibody was omitted as the negativecontrol. In each lesion the nuclear p53 immunoreac-tivity, identified as a red stain, was determined byan experienced pathologist (MR). Immunoreactivityin any nuclei was considered as positive for p53protein accumulation.

DNA Extraction

Five 10-p sections were dewaxed using a routinehistological procedure. Unstained section was over-laid on a hematoxylin-and-eosin-stained section, inwhich the precancerous and cancerous areas hadbeen circled with marking pen. Each lesion was mi-crodissected from the unstained sections followingthe mark and were combined in a 1.5-ml microcen-trifuge tube for proteinase K digestion as describedby Wright and Manos16 with some modifications;use of 1% (w/v) sodium dodecyl sulfate instead of0.5% (v/v) tween 20 solution and incubation with en-zyme for 2 days instead of overnight. After heat in-

activation of the proteinase K, DNA was extractedby adding 160 pl saturated NaCI solution to the pro-teinase K digests, mixing thoroughly, and centrifug-ing for 15 minutes at 12,000 rpm in a bench-top Ep-pendorf Centrifuge at room temperature. The upperaqueous layer was transferred to a fresh 1.5-ml ep-pendorf tube using a wide-bore pipette to avoid theshearing of DNA. The DNA was then precipitatedwith 3 volumes of cold absolute ethanol, placed onice for 30 minutes and centrifuged (12,000 rpm) for15 minutes at room temperature to pellet the DNA.After careful removal of the supernatant, 500 pl of70% ethanol was added to wash the DNA. Thetubes were centrifuged at 12,000 rpm for 15 min-utes and the supernatant was discarded. After dry-ing in an RC 10.10 Concentrator (Jouan, Winches-ter, VA), DNA was dissolved in 100 pl of doubledistilled water at 4 C overnight. Finally, the amountof dissolved DNA was quantified on the TKO-100Minifluorometer according to the manufacturer'sprocedure (Hoefer Scientific Instruments, San Fran-cisco, CA) and transferred to a 0.5-ml microcentri-fuge tube for long-term storage at -20 C. A nega-tive control containing only reagents and no tissuewas run in parallel for each DNA extraction.

PCR-SSCP

The primers for amplification of p53 exon 5 to exon8 (Figure 1) from genomic DNA were selected usingthe "OLIGO" software program.17 The PCR mixture,modified from Prior,18 contains 67 mmol/L Tris-base (pH 8.8), 16.6 mmol/L ammonium sulfate[(NH4)2SO4], 6.7 pmol/L disodium ethylenediamine-tetraacetic acid (EDTA), 0.5 pg/pl nuclease-free bo-vine serum albumin, 1.5 mmol/L MgCI2, 200 pmol/Ldeoxyribonucleoside triphosphates, 0.5 pmol/Lprimers, 1.25 units of Taq polymerase (Cetus-PerkinElmer, Norwalk, CT), and 20 ng of DNA template ina total 50 pl reaction mixture. After adding 50 to 100pl of mineral oil to cover the top of each reactionmixture, the reactions were carried out in EricompTwinblock thermal cycler (Ericomp, San Diego, CA),which had been prewarmed to 90 C. The tempera-ture profile, as described by Wright and Manos,16denatures the DNA at 94 C for 5 minutes in the firstcycle followed by 30 seconds annealing at 55 C, 2minutes extension at 72 C, and 30 seconds at 94 Cfor a total of 40 reaction cycles. The extension timefor the last cycle was increased to 5 minutes to en-sure complete extension. A negative control con-taining no DNA template was run in parallel for eachamplification reaction.

Alteration in Gastric Precancerous Lesions 513AJP March 1994, Vol. 144, No. 3

1 3 5 7 9

3 EA T SB +B E7 T E81

3 E5A +2- E5B6A 4 E6B 6 E7 8 E8 10 5

E5A(152 bp) 1: 5'-TTCCTCTTCCTGCAGTACTC-3'

2: 5'-TCCGTCATGTGCTGTGACTG-3'

E5B6A(167 bp) 3: 5'-GCCATCTACAAGCAGTCACA-3'

4: 5'-GCCAGACCTAAGAGCAATCA-3'

E6B(132 bp) 5: 5'-TTAGGTCTGGCCCCTCCTCA-3'

6: 5'-AGTTGCAAACCAGACCTCAG-3'

E7(136 bp) 7: 5'-TTGTCTCCTAGGTTGGCTCT-3'

8: 5'-CAAGTGGCTCCTGACCTGGA-3'

E8(149 bp) 9: 5'-TGGTAATCTACTGGGACGGA-3'

10: 5'-CTGCTTGCTTACCTCGCTTA-3'

Before carrying out SSCP, 10 pl of PCR productwas electrophoresed in 2% (w/v) agarose and visu-alized with ethidium bromide stain (0.5 pg/ml) toconfirm the absence of contamination and to ensurethat the PCR product was a single band of the ap-propriate size. The SSCP method was modified fromAinsworth et al.19 One microliter of each PCR prod-uct was mixed with 5 pl of a denaturing solution of95% (v/v) formamide, 20 mmol/L disodium EDTA,0.05% (w/v) xylene cyanole, and 0.05% (w/v) bro-mophenol blue. Immediately before electrophoresis,the samples were heated to 95 C for 5 minutes in awater bath. After heat denaturation, the tubes wereimmediately placed on ice to prevent renaturation.A 2-pI aliquot of each denatured sample wereloaded onto 0.75-mm-thick, 12% (w/v) polyacryl-amide gel (29:1 ratio of acrylamide to bisacryl-amide) gel with 22.5 mmol/L Tris-borate (pH 8.4)and 2 mmol/L EDTA in Miniprotein 11 Slab cell (Bio-Rad, Richmond, CA). Electrophoresis of gel wascarried out at room temperature at 5 mA constantcurrent for 4 hours.

Single stands of DNA were visualized with silverstain (Bio-Rad) according to the instructions pro-vided by the manufacturer. The band pattern ofeach precancerous or cancerous lesion was com-pared to that of morphologically normal cells fromthe same patients. Any extra band(s) present in thesample was considered as positive for a mutationand DNA sequencing was performed on thosesamples.

Figure 1. Locations and sequences of fiveprimer sets (E5A, E5B6A, E6B, E7, and E8) forthe amplification ofp53 exons 5 to 8. Arrowprimer. Number in parenthesis: base pairs ofamplified product.

DNA Sequencing

Twenty microliters of PCR product was electropho-resed in 1% (w/v) agarose gel. The target productwas excised from the gel and eluted in Tris-EDTAbuffer at 55 C for at least 10 hours. After elution,DNA was precipitated with absolute ethanol,washed with 70% (v/v) ethanol, and dried in the RC10.10 Concentrator (Jouan). DNA was redissolvedinto 20 pl double-distilled water and 5 pl of this so-lution was used for DNA sequencing.A dideoxy DNA sequencing method was per-

formed using CircumVent Thermal Cycle Sequenc-ing Kit as described by the manufacturer's instruc-tions (New England BioLabs, Beverly, MA). Thedetection signal was revealed by incorporation of[a-35S]dATP into the DNA sequence. The sequenceladder was resolved in 7.7 mol/L urea and 6% (w/v)polyacrylamide gel using a 35 x 40 cm electropho-resis chamber (GIBCO, Bethesda Research Labo-ratories, Gaithersburg, MD). After electrophoresis,the gel was dried in a gel dryer (Bio-Rad) and ex-posed to x-ray film for 3 to 5 days at room tempera-ture.

ResultsThe results of p53 alterations in precancerous andcancerous lesions detected by IHC, PCR-SSCP,and DNA sequencing are shown in Table 1.

514 Shiao et alAJP March 1994, Vol. 144, No. 3

Table 1. p53 Alterations Detected by IHC, SSCP, and DNA Sequencing

Normal cells IM GED GCCase IHC SSCP IHC SSCP IHC SSCP IHC SSCP DNA sequencing Codon

1 - - NP. - *E7 - *E7 *CCATCCTCA deletion 250-2532 - _ NP. + *E7 + *E7 *CGG-CAG (missense) 2483 _ _ - - + - - - Not tested4 - *E6 - *E6B *E6B *E6B *CGA-*CGG (polymorphism) 2135 - - - - - - + *E7 *TGC-TCC (missense) 2426 - I.T. E5B6A lT. E7 IT. *E5A *CCC-TCC (missense) 1517 - - - *E5B6A - *E5B6A + *E5B6A *T-*C (Intron5)8 - - NP. + *E5B6A - *E5B6A *CGCOCAC (missense) 1759 - - + + Not tested10 - - IT. - IT. - IT. - Not tested11 - *E6B NP. + *E5B6A + *E5B6A *CGCOCAC (missense) 175

*E6B *E6B *CGA.--CGG (polymorphism) 21312 - *E7 - *E7 + *E7 + *E7 *AAC-*AAT (same A.A.) 247

IM: intestinal metaplasia; GED: gastric epithelial dysplasia; GC: gastric cancer; A: adenosine; C: cytidine; G: guanosine; T: thymidine; A.A.:amino acid; N.P.: lesion not present; IT.: insufficient tissue.

* Mutation confirmed by DNA sequencing.

Immunohistochemical results were not obtainedin two cases (6 and 10) because of insufficient tis-sue. A positive p53 immunoreactivity was observedin the stage of dysplasia and carcinoma (Figure 2).p53 immunoreactivity was detected in 60% (6 of 10)of the dysplasia cases, and 60% (6 of 10) of thecarcinoma cases. A consistent stain, either positive

Figure 2. p53 immunohistochemical stain with CM-1 antibody. N.normal mucosa; D: dysplasia; arrow: positive immunoreactivity innuclei (250X).

or negative, in both precancerous and cancerousstages was observed in 60% of cases. However,there were two cases (3 and 8) having a positivep53 stain in dysplasia but not in carcinoma.

p53 gene alterations detected by PCR-SSCP onlywere found in 25% (3 of 12) of the normal, 50% (4of 8) of the metaplasia, 66.7% (8 of 12) of the dys-plasia, and 75% (9 of 12) of the carcinoma cases.When the degree of dysplasia was considered,there were p53 gene alterations detected by PCR-SSCP in two cases of mild dysplasia (cases 5 and7), one case of moderate dysplasia (case 6), andfive cases of severe dysplasia (cases 1, 2, 8, 11,and 12). All the alterations were located in the ex-ons 5 to 7 region. No mutations were detected inexon 8. Once a mutation of a particular exon pre-sented in a precancerous stage, the same alterationwas also detected in all subsequent stages, includ-ing carcinoma (Figure 3), except in case 6. In thiscase, different mutations were present in metapla-sia, dysplasia, and carcinoma.The gene alterations detected by PCR-SSCP

were confirmed by DNA sequencing (Figure 4).Samples with the same electrophoretic pattern inSSCP showed the same mutation by DNA sequenc-ing. Case 1 had a 9-base deletion between codons250 to 253. Five cases (2, 5, 6, 8, and 11) had mis-sense mutations. Case 7 presented a mutation at in-tron 5. Three cases (4, 11, and 12) had basechanges without substitution of amino acids in allstages of tumorigenesis. Base change at codon213 in cases 4 and 1 1 was a natural polymorphism.G:C -> A:T mutation was detected in 71% (5 of 7) ofpoint mutations.The majority of cases with negative IHC and posi-

tive SSCP were found in cases with gene deletion

Alteration in Gastric Precancerous Lesions 515AJP March 1994, Vol. 144, No. 3

Case 1 Case 2N D C N D C

Case 12N M D C

Figure 3. PCR-SSCP analysis of exon 7 in cases1, 2, and 12. N: normal cells; M: metaplasia; D:dysplasia; C: carcinoma; arrow: extra band rep-resenting a mutation.

(case 1), intron mutation (case 7), or base changeswithout substitution of amino acids (cases 4, 11,and 12).

DiscussionMutations of the p53 gene have been widelystudied in primary gastric carcinoma by im-munohistochemical and molecular genetic tech-niques 12,13,20,21 The presence of p53 alterations inprecancerous stages has also been reported8 inisolated lesions from different patients. In this study,detection of p53 alterations in samples containingprecancerous and cancerous lesions from the samepatient provided an opportunity to study the spec-trum of p53 alterations in gastric tumorigenesis.

The high prevalence of immunoreactivity, by IHC,in dysplasia and carcinoma suggests that p53 pro-tein accumulation is associated with gastric tumori-genesis. Observation of a positive p53 stain startingfrom dysplasia demonstrates that p53 protein accu-mulation occurs at a late precancerous stage.

Examination of the p53 gene by PCR-SSCP andDNA sequencing revealed mutations that were notdetected by immunohistochemistry (cases 1 and 7).This indicates that alterations of the p53 gene maynot always result in protein accumulation. Detectionof the same type of mutation in precancerousstages and in carcinoma provide evidence for themonoclonal expansion of mutated cells (cases 1, 2,7, 8, 11, and 12). In case 6, different mutations weredetected by SSCP in metaplasia, dysplasia, andcarcinoma. In addition, no base change was identi-fied by DNA sequencing in metaplasia and dyspla-sia. The repeated experiment also showed thesame result. This suggests that the instability of thep53 gene during tumor progression22 could initiategene alterations in a small population of cells, whichmay not be detected by DNA sequencing. A pos-sible false positive base misincorporation duringPCR by Taq polymerase is low because the frag-ment of PCR product was short.23 The chance forerror of amplification in three different lesions from

Case 12

3 1

__C -> T-A

51

Case 11

A C G T

3'

-C-G -> A

5'

Figure 4. Dideoxy DNA sequencing on cases 12 and 11. Case 12:C -* T mutation in codon 247; case 11: G A mutation in codon175.

the same patient is unlikely. Nucleotide A-,G transi-tion at codon 213 in cases 4 and 11 has beenknown as a rare polymorphism.24 Base changewithout amino acid substitution at codon 247 incase 12 may be a background mutation during evo-lution.

Seven out of the nine cases on which DNA se-quencing was carried out, gene alterations were lo-calized in exon 5 and exon 7. In addition, mutations

516 Shiao et alAJP March 1994, Vol. 144, No. 3

in codons 175 and 248, which have been frequentlydetected in gastric cancer,9,25 were observed inthree cases (2, 8, and 11). This suggests that thesame carcinogen may have been involved, or a par-ticular conformation of sequence-dependent DNAin this region is susceptible to the binding of a spe-cific carcinogen. Seventy-one percent (5 of 7) of thedetected mutations were G:C -- A:T transition,which has been reported as a major type of muta-tion in gastric cancer.2627 This type of mutation hasbeen linked to exposure to nitric oxide.28 29 Nitrosocompounds have been experimentally used to in-duce gastric cancer.3031 Inflammatory cells, in thegastric mucosa, associated with Helicobacter pyloriinfection have been proposed as source of nitricoxide-related mutagens in gastric carcinogenesis.6

Applications of PCR-SSCP to cases (3 and 9) thatshowed positive IHC revealed that the detectableaccumulation of p53 protein may not result from mu-tations of p53 gene or that mutations were locatedoutside the exons 5 to 8. It has been shown thatagents that damage DNA32 and viral proteins33 caninduce the accumulation of wild-type p53 protein. Intwo cases (3 and 8), positive p53 immunoreactivitydetected in dysplasia was not detected in carci-noma. This may represent a false negative result,which has been reported in formalin-fixed, paraffin-embedded tissue.34 Another possibility is loss ofp53 expression in the progression to carcinoma. Anegative p53 immunohistochemical stain was ob-served in cases with gene deletion, intron mutation,base change without substitution of amino acid,and no detected mutation. This may be caused byeither loss of p53 expression or the short half-life ofwild-type p53 protein. A positive immunoreactivitydetected in dysplasia or carcinoma for cases 7 and12, which have intron mutation and base changewithout amino acid substitution, respectively, couldbe a result of mutations outside exons 5 to 8. Allsamples with missense mutations showed a positivep53 stain, suggesting that missense mutations re-sult in a stable p53 protein.

Associations of p53 protein accumulation withpoor survival12'13 have been shown, but a contraryresult has also been reported.35 In addition, as wehave pointed out that alterations of the p53 genemay not result in the accumulation of p53 protein,and vice versa. To clarify the relationship betweenp53 alterations and survival requires a study con-sidering both p53 protein accumulation and genemutations in a larger population.We conclude that the mutation of p53 gene is an

early event in stomach tumorigenesis, although its

detection by immunohistochemical stain is first seenat the stage of dysplasia. This also indicates thatPCR-SSCP is a sensitive technique for detectinggenetic alterations. Early detection of p53 alter-ations in precancerous gastric lesions will be usefulinformation in terms of prevention and may be rel-evant in understanding its role in the natural historyof the disease.

References1. Parkin DM, Stjerhsward J, Muir CS: Estimates of the

world-wide frequency of twelve major cancers. BullWHO 1984, 62:163

2. Parkin DM, Laara E, Muir CS: Estimates of the world-wide frequency of sixteen major cancers in 1980. Int JCancer 1988, 41:184-197

3. Silverberg E, Lubera JA: Cancer statistics. CA 1986,36:9-25

4. Antonioli DA: Current concepts in carcinoma of thestomach. Contemporary Issues in Surgical Pathology.Edited by Appelman HD. New York, Churchill Living-stone, 1988, p 121

5. Boddie AW Jr, McBride CM, Balch CM: Gastric can-cer. Am J Surg 1989, 157:595-606

6. Correa P: Human gastric carcinogenesis: a multistepand multifactorial process. First American Cancer So-ciety Award Lecture on Cancer Epidemiology and Pre-vention. Cancer Res 1992, 52:6735-6740

7. Correa P: A human model of gastric carcinogenesis.Cancer Res 1988, 48:3554-3560

8. Tohdo H, Yokozaki H, Haruma K, Kajiyama G, TaharaE: p53 gene mutations in gastric adenomas. VirchowsArch [B] 1993, 63:191-195

9. Yokozaki H, Kuniyasu H, Yasuhiko K, Nishimura K,Todo H, Ayhan A, Yasui W, Ito H, Tahara E: p53 pointmutations in primary human gastric carcinomas. JCancer Res Clin Oncol 1992, 119:67-70

10. Matozaki T, Sakamoto C, Suzuki T, Matsuda K, UchidaT, Nakano 0, Wada K, Nishisaki H, Konda Y, NagaoM, Kasuga M: p53 gene mutations in human gastriccancer: wild-type p53 but not mutant p53 suppressesgrowth of human gastric cancer cells. Cancer Res1992, 52:4335-4341

11. Rugge M, Shiao Y-H, Correa P, Baffa R, Di Mario F: Im-munohistochemical evidence of p53 overexpression ingastric epithelial dysplasia. Cancer Epidemiol Biomar-kers Prevention 1992, 1:551-554

12. Martin HM, Filipe Ml, Morris RW, Lane DP, Silvestre F:p53 expression and prognosis in gastric carcinoma.Int J Cancer 1992, 50:859-862

13. Starzynska T, Bromley M, Ghosh A, Stern PL: Prognos-tic significance of p53 overexpression in gastric andcolorectal carcinoma. Br J Cancer 1992, 66:558-562

14. Rugge M, Farinati F, Di Mario F, Baffa R, Valiante F,Cardin F: Gastric epithelial dysplasia: a prospective

Alteration in Gastric Precancerous Lesions 517AJP Marcb 1994, Vol. 144, No. 3

multicenter follow-up study from the interdisciplinarygroup on gastric epithelial dysplasia. Hum Pathol1991, 22:1002-1008

15. Midgley CA, Fisher CJ, Bartek J, Vojtesek B, Lane DP,Barnes DM: Analysis of p53 expression in human tu-mours: an antibody raised against human p53 ex-pressed in Escherichia Coli. J Cell Sci 1992, 101:183-189

16. Wright KD, Manos MM: Sample preparation from par-affin embedded tissue. PCR Protocol: A Guide toMethods and Applications. Edited by Innis MA, et al.San Diego, Academic Press Inc., 1990, pp 153-158

17. Rychlik W, Rhoads RE: A computer program forchoosing optimal oligonucleotides for filter hybridiza-tion, sequencing and in vitro amplification of DNA.Nucleic Acid Res 1989, 17:8543-8551

18. Prior WT, Friedman KJ, Highsmith WE, Perry TR, Sil-verman LM: Molecular probe protocol for determiningcarrier status in Duchenne and Becker Muscular Dys-trophies. Clin Chem 1990, 36:441-445

19. Ainsworth PJ, Surh LC, Coulter-Mackie MB: Diagnos-tic single strand conformation polymorphism (SSCP):a simplified non-radioisotopic method as applied to aTay-Sachs Bi variant. Nucleic Acids Res 1991, 19:405-406

20. Tamura G, Kihana T, Nomura K, Terada M, SugimuraT, Hirohashi S: Detection of frequent p53 gene muta-tions in primary gastric cancer by cell sorting and po-lymerase chain reaction single-strand conformationpolymorphism analysis. Cancer Res 1991, 51:3056-3058

21. Yamada Y, Yoshida T, Hayashi K, Sekiya T, Yokota J,Hirohashi S, Nakatani K, Nakano H, Sugimura T,Terada M: p53 gene mutations in gastric cancer me-tastases and in gastric cancer cell line derived frommetastases. Cancer Res 1991, 51:5800-5805

22. Perry ME, Levine AJ: Tumor suppressor p53 and cellcycle. Curr Opin Genet Dev 1993, 3:50-54

23. Koop BF, Rowan L, Chen W-Q, Deshpande P, Lee H,Hood L: Sequence length and error analysis of Se-quenase and automated Taq cycle sequencing meth-ods. BioTechniques 1993, 14:442-447

24. Carbone D, Chiba I, Mitsudomi T: Polymorphism atcodon 213 within the p53 gene. Oncogene 1991,6:1691-1692

25. Kim J-H, Takahashi T, Chiba I, Park J-G, Birrer MJ,Roh JK, Lee HD, Kim J-P, Minna JD, Gazdar AF:

Occurrence of p53 gene abnormalities in gastric car-cinoma tumors and cell lines. J Natl Cancer Inst 1991,83:938-943

26. Imazeki F, Omata M, Nose H, Ohto M, Isono K: p53gene mutations in gastric and esophageal cancers.Gastroenterology 1992, 103:892-896

27. Renault B, van den Broek M, Fodde R, Wijnen J, Pel-legata NS, Amadori D, Khan PM, Ranzani GN: Basetransitions are the most frequent genetic changes atp53 in gastric cancer. Cancer Res 1993, 53:2614-2617

28. Nguyen T, Brunson D, Crespi CL, Penman BW, Wish-nok JS, Tannenbaum SR: DNA damage and mutationin human cells exposed to nitric oxide in vitro. ProcNatI Acad Sci USA 1992, 89:3030-3034

29. Wink DA, Kasprzak KS, Maragos CM, Elespuru RK,Misra M, Dunams TM, Cebula TA, Koch WH, AndrewsAW, Allen JS: DNA deaminating ability and genotoxic-ity of nitric oxide and its progenitors. Science 1991,254:1001-1003

30. Sugimura T, Fujimura S, Baba T: Tumor production inthe glandular stomach and alimentary tract of rat byN-methyl-N'-nitro-N-nitrosoguanidine. Cancer Res1970, 30:455-465

31. Tatematsu M, Takahashi M, Fukushima S, HananouchiM, Shirai T: Effects in rats of sodium chloride on ex-perimental gastric cancers induced by N-methyl-N-nitro-N-nitrosoguanidine or 4-nitroquinoline-1-oxide. JNatl Cancer Inst 1975, 55:101-106

32. Fritsche M, Haessler C, Brandner G: Induction ofnuclear accumulation of the tumor suppressor proteinp53 by DNA-damaging agents. Oncogene 1993,8:307-318

33. Sarnow P, HO YS, Williams J, Levine AJ: AdenovirusEl b-58 Kd tumor antigen and SV 40 large tumor anti-gen are physically associated with the same 54 Kdcellular protein in transformed cells. Cell 1982, 28:387-394

34. Wynford-Thomas D: p53 in tumour pathology: can wetrust immunohistochemistry? J Pathol 1992, 166:329-330

35. Sasano H, Date F, Imatani A, Asaki S, Nagura H:Double immunostaining for c-erbB-2 and p53 in hu-man stomach cancer cells. Hum Pathol 1993, 24:584-589