I Clin Pathol: Mol Pathol 1996;49:M85-M90

p53 gene alterations and protein accumulationin colorectal cancer

R Bertorelle, G Esposito, C Belluco, L Bonaldi, A Del Mistro, D Nitti, M Lise,L Chieco-Bianchi

AbstractAim-To correlate immunohistochemicalstaining with single strand conformationpolymorphism (SSCP) analysis of the p53gene in colorectal cancer in order to un-derstand how the findings provided by thetwo techniques complement each other indefining p53 functional status.Methods-Frozen tumour tissue from 94patients with colorectal cancer was studiedfor p53 protein accumulation and genemutations. Accumulation of p53 proteinwas detected by immunohistochemistryusing PAbl801 and BP53-12-1 monoclonalantibodies. The findings were then com-pared with SSCP analysis of exons 5 to 8of the p53 gene. All cases with a positiveresult by SSCP analysis were confirmedby sequencing.Results-Nuclear staining was observed in51 (54.2%) cases. SSCP analysis of theDNA amplified by PCR revealed that theelectrophoretic pattern had shifted in 30cases; sequence analysis confirmed the oc-currence of a mutation in 29 cases and ofa polymorphism in one. In 27 cases bothassays gave a positive result, and in 40both were negative; therefore, concord-ance between PCR-SSCP and immuno-histochemistry was seen in 72% of cases.Conclusion-The data indicate that pos-itive immunostaining corresponds withthe presence of a mutation in most, butnot all, cases studied; other mechanismscould be responsible for stabilisation andaccumulation of p53 protein in the nuc-leus. Nonsense mutations which do notconfer stability on the protein will not bedetected by immunohistochemistry andfalse negative results can also occur withSSCP analysis.(J7 Clin Pathol: Mol Pathol 1996;49:M85-M90)

Keywords: p53, colorectal cancer, immunohisto-chemistry, PCR-SSCP.

Several genetic alterations have been shown toplay a significant role in tumorigenesis. By andlarge, the most frequently observed molecularchanges occur in the p53 gene.' The p53 geneencodes a protein involved in control of thecell cycle, which is capable of binding to DNAand acts as a negative regulator in the cell'sresponse to damaged DNA.2

Functional alteration of p53 protein canoccur through several mechanisms: point muta-tions, deletions, and rearrangements in the p53

gene; binding with viral proteins (for example,SV40 T antigen, adenovirus Elb, papil-lomavirus E6); binding with cellular proteins(for example, mdm-2 oncoprotein, heat shockproteins); and oligomerisation.' Wild-type p53protein has a very short half-life, whereas mut-ated p53 is stable and can accumulate at highconcentrations in the nuclei of tumour cells. Asa consequence, immunohistochemical stainingwith specific antibodies can be used to detectmutant p53 protein. However, p53 over-expression can also reflect the accumulation ofwild-type protein that has been stabilised via amechanism other than mutation- for example,by binding with SV40 T antigen, which greatlyincreases the half-life of the protein.4The detection of p53 protein by immuno-

histochemistry is simple, can be applied to bothfrozen and archival tissues, and has been usedextensively in the study of tumour pathology;correlation of immunohistochemical findingswith anatomopathological variables and clinicaloutcome revealed that accumulation of p53protein in tumour cells is of prognostic import-ance. As pointed out by Wynford-Thomas5 andHall and Lane,6 immunohistochemical as-sessment of p53 should be carried out in astandardised manner to enable comparison ofresults obtained in different laboratories and,in this context, quantitation of p53 over-expression is particularly important. Antigendetection is influenced by many variables, suchas the absolute amount of the protein, thecharacteristics of the antibody (sensitivity,specificity, concentration, duration of in-cubation), the type of material under study(frozen versus fixed/embedded tissue), and theeffects of fixation and other treatments. De-tecting the precise alteration occurring at thegenetic level is much more laborious and costly.The most widely used molecular approach issingle strand conformation polymorphism(SSCP) analysis of DNA fragments amplifiedby PCR,7 with subsequent sequence analysis.The sensitivity of SSCP is influenced by theexperimental conditions, and by the length ofthe amplified fragment under study.A number of investigations have addressed

the relation between p53 protein accumulationand p53 gene alterations in different tumourtypes.8-17 Several antibodies were used in theseanalyses, PAb 1801 and DO-7 being the mostcommon. The SSCP technique utilised by vari-ous workers is based mainly on the classicmethod of Orita et a17 with modifications inthe percentages of glycerol or acrylamide used.The concordance rates between immuno-

Istituto di Oncologia,Universita degli Studidi Padova, e Serviziodi CitodiagnosticaMolecolareOncologica,Azienda Ospedalieradi Padova, ItalyR BertorelleG EspositoL BonaldiA Del MistroL Chieco-Bianchi

Clinica Chirurgica II,Universita degli Studidi Padova, ItalyC BellucoD NittiM Lise

Correspondence to:Dr Roberta Bertorelle,Istituto di Oncologia,Via Gattamelata, 64,35128 Padova,Italy.Accepted for publication5 December 1995

M85

on 11 July 2018 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.2.M85 on 1 A

pril 1996. Dow

nloaded from

Bertorelle, Esposito, Belluco, Bonaldi, Del Mistro, Nitti, et al

histochemistry and SSCP (both negative orboth positive) fall within a relatively wide range(50 to 86%), but this probably depends ondifferences both in methodology and tumourtype.The present study correlates immunohisto-

chemical staining of p53 protein with SSCPanalysis of DNA amplified by PCR.

MethodsACCUMULATION OF P53 PROTEINFrozen sections of 94 colorectal carcinomaswere stained, as described previously,'8 withtwo monoclonal antibodies that react with boththe wild-type and mutant forms of human p53protein; PAb 1801 (Oncogene Science, Union-dale, New York, USA) recognises an epitopebetween amino acids 32 and 79,19 and BP53-12-1 (Biogenex, San Ramon, California, USA)reacts with the N-terminal region between am-ino acids 1 and 45.20 Briefly, sections wereincubated with primary antibody overnight at4°C (for PAb 1801) or for 90 minutes at roomtemperature (for BP53-12-1). Using 3-3'diaminobenzidine (DAB) as substrate, theavidin-biotin peroxidase complex techniquewas carried out as described by the manu-facturer (ABC kit, Vector Laboratories, Bur-lingame, California, USA), and the sectionswere then counterstained with Mayer's haema-toxylin.

Paraffin wax sections were also examined in20 cases because few cells were stained onthe frozen sections or they contained too fewtumour cells. Antigen was retrieved from theformalin fixed, paraffin wax embedded materialby pre-heating the tissue sections in a micro-wave oven.2122 The amount of tumour tissuein each biopsy specimen was evaluated onfrozen sections stained with haematoxylin andeosin, and compared with the amount of nor-mal tissue on the same slide.

Sections with nuclear staining in less than10% of the tumour cells were regarded asnegative, those with staining in 10% to 25%of cells were scored +, those with staining in25% to 75% were scored + +, and those withstaining in over 75% were scored + ++.

P53 MUTATION ANALYSISExtraction ofDNATo obtain genomic DNA, sections of frozentissue 20 .tm thick were digested overnight at37°C with proteinase K in TNE buffer (10mM

Table 1 Comparison of reactivity of PAb 1801 and BP53-12-1 antibodies in cases withequivocal results

Immunohistochemistry

PAb 1801 BP53-12-1

Case Frozen Paraffin Frozen Paraffin

B.I. - - + + + +B.A. - + + + +P.M.V. - + - +C.E. - + - + +V.1. - + + + + + +S.U. - - +++ +++

Tris, 100 mM NaCl, 1 mM EDTA) and so-dium dodecylsulphate, and then extracted withphenol/chloroform; DNA was precipitated with1/10 volume 3 M sodium acetate (pH 5 0) andtwo volumes of cold absolute ethanol (- 20°C),washed with 70% ethanol, air dried, and re-dissolved in distilled water.

In some cases DNA was extracted from form-alin fixed, paraffin wax embedded material asdescribed by Levi et al.23

PCR-SSCPp53 gene mutations were detected b, PCR-SSCP analysis of exons 5 to 8 as describedpreviously.'8 Briefly, 25 amplification cycleswere carried out in a DNA thermal cycler 90r%(Perkin Elmer, Norwalk, Connecticut, USA)followed by five further "hot" cycles with 1 uCialpha "P-dATP. Annealing temperatures were60°C (exons 5 and 6), 560C (exon 7) and51°C (exon 8). The amplified products weredenatured in formamide loading buffer (95%formamide, 20 mM EDTA, 0 05% xylene cy-anol, 0 05% bromophenol blue) and run at30W on a 6% polyacrylamide gel containing5% glycerol. The gel was then dried and ex-posed to x ray film at -800C for 12 to 48hours.

Direct DNA sequencingShifted bands were eluted from the gel, amp-lified by PCR incorporating a phosphorylatedprimer, and digested with lambda exonucleaseto convert the double stranded PCR productto a single stranded form. Other reaction com-ponents were removed from the template bothprior to and following digestion by using MicroSpin columns containing Sephacryl S-400 HRresin (Pharmacia Biotech, Piscataway, NewJersey, USA).DNA was then sequenced by the Sanger

dideoxy chain termination method24 using the"fmol DNA sequencing system" (PromegaCorporation, Madison, Wisconsin, USA).

ResultsIMMUNOHISTOCHEMISTRYOverall, 51 (54-2%) of the 94 primary colo-rectal cancers exhibited positive nuclear stain-ing with BP53-12-1; 49 (52-1%) showednuclear reactivity with PAb 1801. No nuclearstaining was detected in normal mucosa witheither antibody. More than 90% of cells wereneoplastic in 55 cases, between 50 and 90% ofcells in 22 cases, and between 10 and 50% ofcells in 12 cases. In five cases less than 5-10%of cells were neoplastic and therefore immuno-histochemistry was also carried out on paraffinwax sections.

In the 20 cases for which frozen and paraffinwax sections were stained with both antibodies,the same results were obtained in both speci-men types in all but six cases (table 1). Onanalysis of frozen sections, all six were negativewith PAb 1801, while four became positive onstaining with the same antibody on paraffinwax sections. When BP53-12-1 was used, two

M86

on 11 July 2018 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.2.M85 on 1 A

pril 1996. Dow

nloaded from

p53 gene alterations and protein accumulation in colorectal cancer

Figure 1 Immunohistochemical detection ofp53 protein by monoclonal antibodies PAb 1801 (A and B) and BP53-12-1 (C and D) on formalin fixed,paraffin wax embedded tissue sections of two different cases (x 185).

cases were negative on analysis of the frozensections while all six were positive on ex-

amination ofparaffin wax sections. All six cases,

therefore, were finally regarded as having p53protein accumulation for concordance analysis.In the case of the two negative results on frozensections (PMV, CE) with both antibodies, thisresult was probably because the material was

unsuitable as both antibodies gave a positiveresult on paraffin wax sections and a largenumber of tumour cells was present. In theother cases the BP53-12-1 antibody seemed tobe more sensitive.By comparing the immunohistochemical

patterns of nuclear reactivity, evaluated as boththe percentage of positive cells/total number ofneoplastic cells, and the intensity of staining,some differences between the two antibodiesemerged. In general, when more than 75% ofthe cells were positive, the antibodies showeda similar pattern (25 cases were + + + withboth antibodies); however, cases that scored +or + + with PAb 1801 often scored + + +with BP53-12-1 (15 cases had a higher score

with BP-53-12-1). Moreover, BP53-12-1 gen-

erally showed a cleaner and more intense stain-ing reaction (fig 1).

P53 GENE MUTATIONSTumour samples from all 94 cases of colorectalcarcinoma were screened by PCR-SSCP anal-ysis for mutations within exons 5-8 of the p53

gene, and compared with normal tissue fromthe same patient. A non-wild-type migrationpattern was found in 30 cases, 27 of whichshowed also p53 protein accumulation. Sub-sequent sequence analysis confirmed the oc-currence ofa mutation in 29 cases and showed apolymorphism at codon 213 in one. Moreover,this same polymorphism was found in oneof the above 29 cases, in association with amutation in the same exon. Eleven casesshowed mutations in exon 5, four in exon 6,eight in exon 7, and six in exon 8; the types ofmutations identified are described in table 2.In summary, there were 17 G to A or C toT transitions, four A to G transitions (twopolymorphisms), seven transversions, two de-letions, and one duplicating insertion. SSCPand sequence analysis of two representativecases are shown in fig 2.

CORRELATION BETWEEN PCR-SSCP ANDIMMUNOHISTOCHEMICAL FINDINGSThe concordance of results between immuno-histochemistry and PCR-SSCP was 72-0%(both positive in 27 and both negative in 40cases). Twenty four cases were positive onimmunohistochemistry but negative on SSCPanalysis, while only two cases were positive onSSCP analysis and negative on immuno-histochemistry (table 3). The case showing thepolymorphism at codon 213 and a negativeresult on immunohistochemistry was excluded

M87

41-

-AW

on 11 July 2018 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.2.M85 on 1 A

pril 1996. Dow

nloaded from

Bertorelle, Esposito, Belluco, Bonaldi, Del Mistro, Nitti, et al

Table 2 Nuclear accumulation ofp53 protein in cases with p53 gene mutations

p53 mutations p53 accumulation

SSCP Sequence Immunohistochemistry

Case Exon Codon Base change PAb 1801 BP53-12-1

1. A.A. 5 157 GTCCTC + + + +2. S.E. 5 161 GCC-oACC + + + + + +3. V.I. 5 181 CGC-CAC + + + + +4. G.A. 5 132 AAG-.AGG + +5. B.A. 5 143 GTG-oATG + + + + + +6. G.C. 5 175 CGC-oCAC + + + + +7. M.G. 5 161 GCC-oACC + + + + +8. V.R. 5 157-158 Insert GTC CGC + ++ + + +9. A.A. 5 165 CAG-+TAG -

10. B.D. 5 157 GTC-oCTC + + + + + +11. M.P. 5 141 TGC-oTAC ++ + ++12. P.O. 6 191 3 bp deletion + + + + + +13. B.L. 6 208-213 16 bp deletion + + + + + +14. S.P. 6 213 CGA-oCGG* + + + + +

216 GTG-oCTG15. M.G. 6 213 CGA-oCGG* -16. S.N. 6 220 TAT--TGT + +++17. B.R. 7 249 AGG-AGT + + + +18. L.R. 7 248 CGGTGG + + + + + +19. C.A. 7 245 GGC-+AGC + + + + +20. L.O. 7 237 ATG- ATT + + + + + +21. S.U. 7 244 GGCTGC -+ + +22. T.D. 7 238 TGT-TAT + + + + + +23. R.L. 7 248 CGG-oTGG + + + + + +24. C.L. 7 244 GGC-oGAC + + + + + +25. F.G. 8 282 CGG-TGG + + +26. L.T. 8 282 CGG-oTGG + + + + + +27. M.L. 8 282 CGG-oTGG + + + + + +28. B.R. 8 276 GCC-oACC + + + + + +29. A.A. 8 278 CCT-oACT + + + + + +30. H.W. 8 306 CGA-oTGA -* Polymorphism; bp =base pairs.

from concordance analysis. No correlation wasobserved between the immunohistochemicalscore and the presence and type of mutation.

V

N .T

Table 3 Relation between p53 accumulation detected byimmunohistochemistry and p53 mutations by SSCPanalysis in colorectal carcinomas

Ininiunohistochenzisti-y

Positive Negative Total

SSCP positive 27 (29%) 2 (2%/X) 29 (31%)SSCP negative 24 (26%) 40 (430/,,) 64 (69%)Total 51 (55%) 42 (45%'Y,) 93 (100%)*

* The case showing the polymorphism only was excluded fromconcordance analysis.

.* . 4ow;-4

exon 5

V.1.

A-4

A~~~ C Go\S

S_.

B.D.

5' 5

.:c-..400V

.-W

-dft

Im.

A C GTl

G C

T

C

Figure 2 Upper panel: SSCP analysis in normal (N) and tumour (T) samples of twopatients. Lower panel: Sequencing of tumour sample from patient VI. showing a G to Atransition at codon 181 and in patient B.D. showing a G to C transversion at codon 157.

DiscussionIn this study 94 primary colorectal carcinomaswere evaluated for p53 protein expression andp53 gene mutations by immunohistochemistryand PCR-SSCP. At present, the use ofimmunohistochemistry to detect p53 muta-tions is still controversial, as accumulation ofp53 protein might also occur through sta-bilisation by mechanisms other than mutation.That accumulation of p53 protein is an indexof cell cycle deregulation is still a matter ofdebate. In this regard, it is important to high-light some methodological aspects of both theimmunohistochemical and PCR-SSCP assays.

In immunohistochemistry there are no setstandards for the specificity or sensitivity of thedifferent antibodies available, nor the minimalpercentage of positive cells required to score asection as positive. As the detection of a givenamount of antigen by immunohistochemistryis influenced by the affinity and concentrationof the antibody, as well as by the sensitivity ofthe detection system, it is imperative that theconditions used are defined for diagnostic andprognostic purposes. In our experience, the useof two different monoclonal antibodies op-timally defines the specificity of the reactivity

M88

-B. -- ..

on 11 July 2018 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.2.M85 on 1 A

pril 1996. Dow

nloaded from

p53 gene alterations and protein accumulation in colorectal cancer

observed. In fact, while the nuclear stainingwas confirmed by both antibodies in the samecases (discordant results were an occasionalfinding) in the present study, cytoplasmic re-activity was a rare finding and could be in-terpreted as an artefact (rather than a realaccumulation of p53 protein outside the nuc-leus).

It is important to establish the minimumnumber of stained neoplastic cells needed toregard a tumour sample as positive. Because thenumber of stained cells required for a positiveresult varies from study to study, the resultsobtained by different groups are difficult tocompare. Indeed, for some workers one or afew positive cells are indicative of a positivesample, while we and others" 1519 require aminimum of 5-10% stained cells for a positiveresult to avoid the possibility of false positivereactions due to background staining or arte-facts. In the present study, many tumoursamples which scored as + or + + on stainingwith PAb 1801 antibodybecame + + or + + +with BP53-12-1. Thus, the phenotypic patternof the tumour could depend on the sensitivityand specificity of the antibody used, and there-fore correlation of the phenotypic pattern withthe presence and type of mutation, or evenwith patient prognosis and outcome would bearbitrary.A comparison of PCR-SSCP and immuno-

histochemical findings showed concordant res-ults in 72-0% of our cases. In particular,27 (93- 1 %) of the 29 cases showing a p53 genemutation were also positive for p53 proteinaccumulation on immunohistochemistry, while24 (37 5%) of the 64 cases without mutationswere positive on immunohistochemistry. In 26(28 0%) cases the immunohistochemical andSSCP results were discordant; negativeimmunohistochemistry/positive SSCP in two,and positive immunohistochemistry/negativeSSCP in 24. On sequencing, the two caseswith a p53 mutation on SSCP analysis and noimmunohistochemical staining had a nonsensemutation which resulted in the substitution ofa coding triplet by a stop signal at codons 165(exon 5) and 306 (exon 8), respectively. Inboth cases, the encoded protein was truncatedat the C-terminal region, and lacked the nuclearlocalisation signal I (residues 312-323) es-sential for its transport to the nucleus.2526 Se-quence analysis of DNA from normal andtumour tissue in the case with the poly-morphism at codon 213 in exon 6 showed asilent CGA to CGG transition (arginine toarginine), and the normal p53 protein produceddid not accumulate in the cell.27 As the SSCPscreening method does not distinguish poly-morphisms from mutations, sequence analysisis necessary to clarify the significance of thealtered electrophoretic pattern.Accumulation of p53 protein with no evid-

ence of gene mutation may either reflect anaccumulation of a non-mutated protein, or afalse negative SSCP result (the p53 gene ismutated but not detected by the SSCP tech-nique). The sensitivity of SSCP analysis is stilla matter of debate; not only does it depend onthe experimental conditions (that is, tem-

perature, ionic strength and the presence orabsence of glycerol), but also on the length ofthe amplified fragment being analysed.28Sheffield et al29 reported that SSCP sensitivitydramatically decreases with increasing frag-ment size; the percentage of single base sub-stitutions detected by SSCP is 90% when theamplified fragment is 135 base pairs (bp), and70% when it is 200bp. When the fragmentsize exceeds 300bp, sensitivity is even lower(67%).30 As the length ofthe amplified productswe analysed was 245 bp for exon 5, 184 bp forexon 6, 110 bp for exon 7, and 213 bp for exon8, SSCP sensitivity for exons 5 and 8 mighthave been less than 70%. With regard to theexperimental conditions, Condie et al" re-ported that SSCP detected 90% of the p53gene mutations when the electrophoretic runwas carried out at 4°C, with 0 5 x TBE, but itssensitivity decreased to 55% under differentanalytical conditions (room temperature, 5%glycerol).

In a comparison of 52 different mutations,Moyret et al32 found that SSCP consistentlyfailed to detect mutations in two regions, thefirst of which was localised between codons151 and 154, and the other between codons272 and 273. Interestingly, samples from fourof the five cases that escaped SSCP screeningwere of colon carcinoma. We did not observeany alterations on SSCP analysis which, aftersequencing, revealed a mutation in these re-gions.Although our study focused on exons 5 to

8, where the majority of the mutations arethought to be localised,"33 mutations in otherexons could account for some negative SSCPcases. Indeed, recent studies by Greenblatt etal34 and Hartmann et al15 indicated that 13%and 22%, respectively, of the p53 gene mut-ations reported fell outside exons 5-8. A falsenegative SSCP result may also occur when thetumour is not sufficiently represented in thetissue from which the DNA is extracted. Mor-phological analysis of our discordant cases re-vealed that at least five samples were composedof less than 5-10% tumour cells.Accumulation of wild-type p53 protein can

be detected by immunohistochemistry becausethe antibodies are not specific for the mutantform, and the wild-type protein can be sta-bilised by interacting with other nuclear, cellu-lar and viral proteins.'6'8 In any case, bindingwith these proteins results in the functionalinactivation of the p53 gene. Moreover, con-centrations of wild-type p53 protein in thecell may also increase either in response tocontinuous genetic damage, which may occurmore frequently in tumour cells than in normaltissue, or because another component of thep53 suppression pathway (for example, WAFl/cipl) is altered."9

In general, the environment (mutations, pro-tein interactions, or stimuli derived from ac-tivation of oncogenes, such as c-myc and ras,as reported by Lu et al'0) seems to be thedetermining factor for accumulation of p53protein at high concentrations." This ac-cumulation results from the loss of rapid p53protein turnover and might be a signal of its

M89

on 11 July 2018 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.2.M85 on 1 A

pril 1996. Dow

nloaded from

Bertorelle, Esposito, Belluco, Bonaldi, Del Mistro, Nitti, et al

altered function; loss of function of this cellcycle checkpoint gene could contribute to gen-

etic instability and tumorigenesis. This ob-servation is supported by the studies ofDonehower et al42 in p53 gene deficient mice,in which the loss of normal p53 protein was

shown to predispose the animals to neoplasticdisease.The search for p53 gene mutations by mo-

lecular assays is important for defining thefunctional alterations that derive from thedifferent mutations. In this regard, it might beadvisable to study all of the exons of the p53gene. Accumulation of p53 protein may bemore meaningfully regarded as an index of thefunctional status of the p53 gene, rather than an

index of gene mutation, providing the analyticalprocedures are standardised and the ap-

propriate controls are used.In our opinion, immunohistochemistry is a

simple and rapid method for the evaluation ofp53 protein expression as a prognostic markerin many tumours. Its association with mo-

lecular analysis of the p53 gene might providea more precise definition of the changes/muta-tions resulting in the accumulation of the p53protein.

We thank Mrs. Patricia Segato for help in preparing themanuscript; Emma D'Andrea for critical discussion; and CatiaCavestro, Dina Pozza and Rossana Trevisan for technical as-

sistance. This work was supported in part by a grant from theAssociazione Italiana per la Ricerca sul Cancro (AIRC), Milano,and also in part from the Regione Veneto, Ricerca SanitariaFinalizzata, Italy.

1 Hollstein M, Sidransky D, Vogelstein B, Harris CC. p53mutations in human cancers. Science 1991;253:49-53.

2 Kastan MB, Onyekwere 0, Sidransky D, Vogelstein B, CraigR. Participation of p53 protein in the cellular response toDNA damage. Canicer Res 1991;51:6304-11.

3 Prokocimer M, Rotter V. Structure and function of p53 in

normal cells and their aberrations in cancer cells: pro-jection on the hematologic cell lineages. Blood 1994;84:2391-411.

4 Sarnow P, Ho YSH, WilliamsJ, Levine AJ. Adenovirus Elb-58 kd tumor antigen and SV40 large tumor antigen are

physically associated with the same 54 kd cellular proteinin transformed cells. Cell 1982;28:387-94.

5 Wynford-Thomas D. p53 in tumor pathology: can we trustimmunocytochemistry? 7Pathol 1992;166:329-30.

6 Hall PA, Lane DP. p53 in tumor pathology: can we trustimmunohistochemistry?-Revisited! Pathol 1994;172:1-4.

7 Orita M, Suzuky Y, Sekita T, Hayashi K. Rapid and sensitivedetection of point mutations and DNA polymorphismsusing the polymerase chain reaction. Genontics 1989;5:874-9.

8 Cripps KJ, Purdie CA, Carder PJ, White S, Komine K,Bird CC, et al. A study of stabilisation of p53 proteinversus point mutation in colorectal carcinoma. Oncogene1994;9:2739-43.

9 Dix B, Robbins P, Carrello S, House A, lacopetta B. Com-parison of p53 gene mutation and protein overexpressionin colorectal carcinomas. BrJ_ Cancer 1994;70:585-90.

10 Cordon-Cardo C, Dalbagni G, Saez GT, Oliva MR, ZhangZ-F, Rosai J, et al. p53 mutations in human bladder cancer:

genotypic versus phenotypic patterns. Irnt _J Cancer 1994;56:347-53.

11 Esrig D, Spruk CH III, Nichols PW, Chaiwun B, StevenK, Groshen S, et al. p53 nuclear protein accumulationcorrelates with mutations in the p53 gene, tumor grade,and stage in bladder cancer. AmJPathol 1993;143:1389-97.

12 Umekita Y, Kobayashi K, Saheki T, Yoshida H. Nuclearaccumulation of p53 protein correlates with mutations inthe p53 gene on archival paraffin-embedded tissues ofhuman breast cancer. Jpn Cancer Res 1994;85:825-30.

13 Allred DC, Clark GM, Elledge R, Fuqua SAW, Brown RW,Chamness GC, et al. Association ofp53 protein expressionwith tumor cell proliferation rate and clinical outcome innode-negative breast cancer. _7 Natl Cancer Inst 1993;85:200-5.

14 Hong SI, Hong WS, Jang JJ, Lee DS, Cho NS, Jung ME,et al. Alterations of p53 gene in primary gastric cancer

tissues. Anticancer Res 1994;14:1251-6.

15 Scarpa A, Capelli P, Mukai K, Zamboni G, Oda T, IaconoC, et al. Pancreatic adenocarcinomas frequently show p53gene mutations. AmjIPathol 1993;142:1534-43.

16 Baas 10, Mulder J-WR, Offerhaus GJA, Vogelstein B,Hamilton R. An evaluation of six antibodies for immuno-histochemistry of mutant p53 gene product in archivalcolorectal neoplasms. 7 Pathol 1994,172:5-12.

17 Kikuchi-Yanoshita R, Konishi M, Ito S, Seki M, Tanaka K,Maeda Y, et al. Genetical changes of both p53 allelesassociated with the conversion from colorectal adenomato early carcinoma in familial adenomatous polyposis andnon-familial adenomatous polyposis. Canicer Res 1992;52:3965-71.

18 Bertorelle R, Esposito G, Del Mistro A, Belluco C, NittiD, Lise M, Chieco-Bianchi L. Association ofp53 gene andprotein alterations with metastases in colorectal cancer. An

_7 Surg Pathol 1995;19:463-71.19 Banks L, Matlashewski G, Crawford L. Isolation of human-

p53-specific monoclonal antibodies and their use in thestudies ofp53 expression. Eiir.Biochem 1986; 159:529-34.

20 Bartek J, Bartkova J, Vojtesek B, Staskova Z, Lukas J, RejtharA, et al. Aberrant expression of the p53 oncoprotein is acommon feature of a wide spectrum of human malig-nancies. Oncogene 1991;6: 1699-703.

21 Shi S-R, Kev ME, Kalra KL. Antigen retrieval in formalin-fixed, paraffin-embedded tissues: an enhancement methodfor immunohistochemical staining based on microwaveoven heating of tissue sections. Y Histochemi Cvtochenm 199 1;39:741-8.

22 Cuevas EC, Bateman AC, Wilkins BS, Johnson PA, WilliamsJH, Lee AHS, et al. Microwave retrieval in immuno-histochemistry: a study of antibodies. 7 Clin Pathol 1994;47:448-52.

23 Levi S, Urbano-Ispizua A, Gill R, Thomas DM, GilbertsonJ, Foster C, et al. Multiple K-ras codon 12 mutations incholangiocarcinomas demonstrated with a sensitive poly-merase chain reaction technique. Canicer Res 1991;51:3497-502.

24 Sanger F, Nicklen S, Coulson AR. DNA sequencing withchain terminating inhibitors. Proc NVatlAcald Sci USA 1977;74:5463-7.

25 Addison C, Jenkins JR, Sturzbecher HW. The p53 nuclearlocalization signal is structurally linked to a p34cdc2 kinasemotif. Oncogene 1990;5:423-6.

26 Shaulskv G, Goldfinger N, Ben-Ze'Ve A, Rotter V. Nuclearaccumulation ofp53 protein is mediated by several nuclearlocalization signals and plays a role in tumorigenesis. MolCell Biol 1990;10:6565-77.

27 Carbone D, Chiba I, Mitsudomi T. Polymorphism at codon213 within the p53 gene. Oncogene 1991,6:1691-2.

28 Hayashi K, Yandell DW. How sensitive is PCR-SSCP? Hin;Mutat 1993;2:338-46.

29 Sheffield VC, Beck JS, Kwitek AE, Sandstrom AW, StoneEM. The sensitivity of single-strand conformation poly-morphism analysis for the detection of single base sub-stitutions. Genomics 1993;16:325-32.

30 Hayashi K. PCR-SSCP: a simple and sensitive method fordetection of mutations in the genomic DNA. PCR MethodsAppl 1991 l1:34-8.

31 Condie A, Eeles R, Borresen A-L, Coles C, Cooper C,Prosser J. Detection of point mutations in the p53 gene:comparison of single-strand conformation polymorphism,constant denaturant gel electrophoresis, and hydroxyl-amine and osmium tetroxide techniques. Huwn Muttat1993;2:58-66.

32 Moyret C, Theillet C, Laurent-Puig P, Moles J-P, ThomasG, Hamelin R. Relative efficiency of denaturing gradientgel electrophoresis and single strand conformation poly-morphism in the detection of mutations in exons 5 to 8of the p53 gene. Oncogene 1994;9:1739-43.

33 Caron De Frommentel C, Soussi T. TP53 tumor suppressorgene: a model for investigating human mutagenesis. GenesChnronosonm Cancer 1992;4:1-15.

34 Greenblatt MS, Bennet WP, Hollstein M, Harris CC. Muta-tions in the p53 tumor suppressor gene: clues to canceretiology and molecular pathogenesis. Cancer Res 1994;54:4855-78.

35 Hartmann A, Blaszyk H, McGovern RM, Schroeder JJ,Cunningham J, De Vries EMG, et al. p53 gene mutationsinside and outside of exons 5-8: the patterns differ inbreast and other cancers. Oncogene 1995;1O:681-8.

36 Momand J, Zambetti GP, Olson DC, George D, Levine AJ.The mdm-2 oncogene product forms a complex with thep53 protein and inhibits p53-mediated transactivation.Cell 1992;69:1237-45.

37 Sheffner M, Wemess BA, Huibregtse JM, Levine AJ, HowleyPM. The E6 oncoprotein encoded by human papillomavirus type 16 and 18 promotes degradation of p53. Cell1990;63:1129-36.

38 Levine AJ, Momand J, Finlay CA. The p53 tumour sup-pressor gene. Nature 1991;351:453-6.

39 El-Deiry WS, Tokino T, Velculescu VE, Levy DB, ParsonsR, Trent JM, et al. WAFI, a potential mediator of p53tumor suppression. Cell 1993;775:817-25.

40 Lu X, Park SH, Thompson TC, Lane DP. ras-inducedhyperplasia occurs with mutation of p53, but activatedras and myc together can induce carcinoma without p53mutation. Cell 1992;70:153-61.

41 Lane DP. The regulation of p53 function: Steiner AwardLecture. Int 3' Cancer 1994;57:623-7.

42 Donehower LA, Harvey M, Slagle BL, McArthur MJ,Montgomery CA Jr, Butel JS, et al. Mice deficient for p53are developmentally normal but susceptible to spon-taneous tumours. Natuire 1992;356: 215-21.

M90

on 11 July 2018 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.49.2.M85 on 1 A

pril 1996. Dow

nloaded from