P 53 overexpression in epidermoid carcinoma of the head and neck

The theory of field cancerization in tumors of the head and neck reflects the complex oncogenesis that occurs in this region. The mechanisms that control cell proliferation at the molecular levei in epidermoid carcinomas (ECs) of the upper aerodigestive tract are still unclear. Mutations in p53 are the genetic alterations most often detected in ECs of the head and neck and seem to contribute actively to the carcinogenic process triggered by p53 as a tumor-suppressor gene and to its association with tobacco. The objective of the present study was to investigate the expression of p53 protein in epidermoid head and neck carcinomas by immunohistochemistry and its immunohistochemical correlation with other prognostic factors. The study was conducted on 63 consecutive ECs cases not submitted to previous treatment. Specimens of the tumor and of the normal adjacent mucosa were collected during surgery and submitted to immunohistochemical reaction for the determination of the expression of anti-protein p53 antibody (M7001 DAKO AIS, Denmark). Anatomo-c1inical and demographic data were not significantly correlated with the presence of Iymph node metastases or p53 expression in the tumor or in the adjacent normal mucosa. Tumor localization in the larynx was significantly correlated with p53 expression. Histological grading as grades I, 11,111and IV was correlated with significant p53 expression (p = 0.025). Conclusions: 1) in the studied material obtained from 63 cases of head and neck ECs, we detected a 48 percent rate of immunohistochemically detectable p53 overexpression; 2) we did not detect a relationship between demographic patient data and p53 expression in the tumor or in the normal adjacent mucosa; 3) p53 overexpression was significantly more frequent in ECs material from the larynx; and 4) The presence of 12 cases with p53 overexpression in the normal adjacent mucosa and with a p53-negative tumor is in agreement with the theory of field cancerization. Follow-up of this patient series for a longe r period of time will permit a better analysis of these values.


INTRODUCTION E
Pidermoid carcinomas (ECs) represent more than 90 percent of the neoplasias of the upper aerodigestive trace) and their incidence increases among subjects who smoke cigarettes and consume alcoholic beverages. 12ECs correspond to the sixth most frequent form of cancer in the worldY Patients with head and neck ECs in the sanle stage ITIayrespond differently to different forms of treatlnent. 2  Currently recognized prognostic factors are not sufficiently objective or standardized, and do not correspond to the biological behavior of the tumor at the cellular leveI.
The theory of field cancerization in head and neck neoplasias reflects the complex oncogenesis that occurs in this region. 33The frequent occurrence of ITIultiple primary tumors in the upper aerodigestive tract Inay be explained by continuous exposure of the ITIUCOSa to variolls carcinogens.
However, the sequential stages of this complex carcinogenesis have not been flllly elucidated and the molecular mechanisms that control cell proliferation and the aberrant behaviorofECs in the upper aerodigestive tract are still unknown.

MA-TERIAL AND METHODS
The study was conducted on 63 consecutive ECs cases not submitted to previous treatment registered at the Department of Head and Neck Surgery of the A.C. Camargo Hospital in 1994 and 1995.Tumor and nonn~l adjacent mucosa specimens were collected during surgery and fixed in Carnoy solution.After fixation, the material was embedded in paraffin and 4 ~m thick sections were obtained and stained with HE.The material was then submitted to immunohistochemical reaction for the determination of the expression of anti-protein p53 antibody (anti-human p53 protein clone D07, M700 I DAKO, AIS, Denmark).

25
The material was deparaffinized and immersed in a bath containing decreasing concentrations of xylol and ethyl alcohol.Tne antigen was recovered by treatment in a microwave oven under immersion in a nitric citrate solution, with 4 baths of 5 min each at lnaxilnum potency.Endogenous peroxidase (H 2 0 2 ) was blocked and the material was incubated 1: 100 with primary antibody overnight at 4°C, followed by incubation with I:200 biotinylated secondary antibody for 30 lnin at 37°C.After incubation with 1:800 StrepAB complex:HRP for 30 lnin at 37°C (DAKO StrepAB cOlnplex/GRP Duet, mouse/rabbit kit), the material was ilnmersed in 60 mg percent DAB (2,3 diaminobenzidine tetrahydrochloride, Siglna Chemical Co., D5637) plus 1 percent dilnethyl sulfoxide PA (Labsynth) for 5 min at 37°C and counterstained with helnatoxylin.
When more than 20 percent of the tumor cell nuclei stained dark brown, the sample was considered positive.The same criterion was used for the evaluation of normal adjacent mucosa._ Positive and negative controls were systelnatically performed for each reaction.Genes involved in carcinogenesis may be classified as proto-oncogenes and tumor-suppressor genes according to their action mechanism.23 ,35 Tumor-suppressor genes co de for proteins with functions related to the control and suppression of cell growth and proliferation by transactivation or transrepression of specific genes.These genes actively affect the cell cycle, regulating it in a repressi ve manner.II The role of gene p53 has been the subject of several studies on solid human tumors.5,6,15,16,IX,19,25,26,30 P53 mutations are the genetic alterations most frequently detected in head and neck ECS,2XJ4 specifically in the larynx, and seem to actively contribute to the carcinogenic process of the gene 31 as a tumor suppressor and to its association with tobacco.4 ,IO,17 The accumulation of lnutation of gene p53 associated with head and neck ECs has been reported in the literature, although it has not been correlated with other prognostic factors.1.2,7.13A better understanding of the role and mechanism of action of gene p53 in the various stages of carcinogenesis may be of help in terms of specific therapeutic approaches, and may identify new tumor markers for early diagnosis.
The objective of the present study was to investigate the expression of p53 protein in head anCl neck ECs by ilnmunohistochemistry and to determine its relation to other prognostic factors.

Statistica/ ana/ysis
Double-entry convergence tables were constructed for the analysis of correlations between the clinicalpathologic variables studied and the immunohistochemical expression of mutated p53 protein.Statistical significance was calculated by the chi -square test or by the exact Fisher test when applicable.

RESULTS
Expression of mutated p53 protein was considered to be posi ti ve in 30 of the 63 specimens of primary tumors (48 percent).
The study did not show a significant correlation between protein p53 overexpression and patient demographic variables (family history of cancer, cigarette smoking or alcohol consumption) (Table 1).Cases for whom demographic data were not available were excluded from the table.
Figure 1 shows a grade 11 EC of the tongue with p53 overexpression.
To facilitate data analysis, the tumors were schematically assigned to the three following topographies: pharynx, including the cervical esophagus; oral cavity, including palate, mouth fIoor, tongue and gingiva; and larynx.Table 2 shows the correlation between p53 expression and anatomical location, histologic grade and lymph node status.
Epidermoid carcinoma of the larynx was significantly correlated with p53 expression which did not occur in other locations.
Histological grading was schematically divided into grades 11. 111

DISCUSSION
The hypothesis of the participation of gene p53 in the process of carcinogenesis of head and neck tumors has been discussed in severaI studies.7-lJ.'3.'4.21.24.27.2lJ,32.34.37After the discovery of anti-protein p53 antibodies, severaI retrospective studies have been published, with variable results, i.e., rates of p53 positivity of 30 to 67 percent.Thirty to 65 percent of the carcinomas have characteristic l11utations in gene p53, with an allele usually having a point Inutation and producing a mis-sense p53 protein that accumulates at high leveIs in cancer cells.The second allele was Iost by conversion and/or genetic deletion.22 Celllines from ECs of the upper aerodigestive tract have revealed increased p53 expression in vitro.1.3ó Data concerning p53 expression in the normal adjacent l11ucosa and in the tumor of the same patient are presented in Table 3.The results show that p53 expression at these two sites was not coincident in 29 cases.The tumor was p53+ and the lnucosa p53-in 17 cases, and 12 cases presented opposite results (a p53-tumor and a p53+ adjacent lnucosa).
Figure 2 shows an area ofthe nOrInal adjacent mucosa with lnoderate epithelial dysplasia and strongly p53 positive.Figure 3 shows an area of non11al adjacent l11ucosa that is negative for p53 overexpression.
Although the follow-up til11eof the present series is still short, three patients who presented recurrences had p53+ tUl11ors, whereas only one case had a recurrence among patients without p53 expression.This sl11allnUlnber of recurrences prevents statisticaI analysis.-------------------------------  In another il11l11unohistochelnical study, Frank et aI.I3 detected p53 overexpression in 37 percent of cases of ECs of the hypopharynx.Shin et £11.32 obtained a 43 percent index in an il11111unohistocheI11ic£11 study of head and neck ECs.The presence of p53+ in 48 percent of the 63 head and neck ECs cases positive for p53 in our study is compatible with literature data.
In vitro studies carried out by Yin et aI.,37S0l11erSet aI.34and Weinberg et aI.3óshowed abnorI11alities of gene p53 in celllines froln head and neck carcinolnas, such as overexpression of protein p53 del110nstrated by il11111unohistochel11istryand mutations and deletions of the gene detected by sequential analysis by the polYlnerase chain reaction.
Maestro et aI.24investigated 58 cases of larynx tunlors and found p53 overexpression in 60 percent using £1n iml11unohistochemicallnethod and l110lecular analysis by the polymerase chain reaction and in situ hybridization.The authors concluded that this is the genetic £1lteration most frequently occurring in these tUI11ors.Of 89 cases of ECs of the larynx studied by NadaI et aI.,27 64 percent presented il11munohistochelnically-d~tecteel p53 overexpression.The specific localization for these case.s of cancer of the larynx was statisticalIy significant in terms of p53 expression, a fact not previously reported.The p53 overexpression observed in 10 of the 12 cases of cancer of the larynx studied here was more frequent than in the study discussed above.Nees et aI. (1993) found p53 overexpression in only 44 percent of their cases of ECs of the larynx and in 80 percent of the cases of cancer of the oropharynx.
We are currently expanding our study to determine p53 expression in the various topographies of the head and neck.
Prospective studies with appropriate material for analysis of p53 have only permitted a simple statistical correlation with some isolated factors such as smoking 4 .IO and hUlTIanpapilIomavirus.3 The absence of a correlation between smoking and p53 gene expression in the present study disagrees with the data reported by others. 2 ,4 With a longer foIlow-up time, few studies were consistent in showing the absence of a correlation between the prognostic factors normalIy used and p53 expression in head and neck tumors.Frank et aI.I3 did not detect a correlation between p53 expression in ECs of the hypopharynx and other data such as incidence of secondary tumors, survival, ploidy or histological grade.NadaI et aI.27 did not find a prognostic significance of p53 expression in neoplastic and dysplastic lesions of the Iarynx in a study of 89 cases, and concluded that the presence of lTIutations of gene p53 detected by the poIYlTIeraSe chain reaction is not correlated with radiosensitivity in these tumors.
Taking into account the theory of field cancerization, Nees et aI.29 identified mutations of gene p53 in tUlTIOrS and in the normal adjacent mucosa and suggested the participation of p53 in the carcinogenesis of lTIultipIe tLllTIOrs.Shin et aI.32 and Gusterson et aI.I4 studied premalignant lesions of the squamous epitheIiulTI, including dysplasia, hyperplasia, Bowen disease, and the normal adjacent mucosa and found 21 to 45 percent and 15 to 16.6 percent p53 positivity by immunochemistry in their respective series.The scarcity of data concerning survival prevented these authors from correlating their results with prognostic factors.However, there is no objecti ve evidence of a correlation between p53 expression and prognostic factors for head and neck carcinomas. 27 The absence of a correlation between the expression of p53 in the normal adjacent lTIUCOSa and in the tumor area of the same patients detected here agrees with literature data.Nees et £11. 29found p53 overexpression in the normal adjacent mucosa and absence of p53 protein in patients with head and neck ECs This apparent discrepancy was observed in 12 of our cases.The presence of p53 expression in the adjacent mucosa may indicate the possible origin of multiple tumors or the local recurrence in this area through different carcinogenic processes, involving or not mutation of the p53 gene.The specific localization in cancer of the larynx was statistically significant in terms of p53 expression, a fact that is not uncommon in literature reports.Our series must be expanded for a better analysis of the topographic distribution of tumors according to p53 expression.

CONCLUSIONS
1.In material obtained from 63 cases of head and neck ECs, we observed 48 percent imlnunohistochelTIicallydetectable p53 overexpression, a rate compatible with literature data.2. We did not observe a relation between patient demographic data and p53 expression in the tUlnor or in the normal adjacent mucosa.3. p53 overexpression was significantly more frequent in material from ECs of the larynx.4. The presence of 12 cases with p53 overexpression in the normal adjacent mucosa and a p53-negative tUlnor agrees with the theory of field cancerization.This finding may be a factor for the early detection of patients at high risk to develop other p53-positive tumors.
A longer follow-up of this series \vill pennit a better analysis of these values in terms of recurrence anel the possible appearance of other tumors, which is aboLIt 6 percent (unpublisheel data).

Figure 1 -
Figure 1 -p53 expression in grade II epidermoid carcinoma of the tongue.
of cases, demonstrated by polYlnorphisll1 of single strand confon11ation.The salne investigators later detecteel mutations in gene p53 by the polYlnerase chain re£1ction in 63 percent of the cell lines studied.Of the 3 I cases of head and neck ECs studied by Chung et £11.,7 2 I presenteei at least one mutation of gene p53, analyzed by the single chain polymorphisl11 confon11ation l11ethod.Field et aI.,1O in a study of73 ECs cases, observed the presence of protein p 5 3 o ver e x p re s s io n i n 67 p e rc e n t o f t h e In b y immunohistochemistry.

Figure 2 -
Figure 2 -Normal adjacent mucosa with moderate epithelial dysplasia and strongly p53 positive.

Table 1 Correlation between p53 expression and demographic data
and IV.Usually, grades 111 and IV have the

Table 2
Association between p53 expression and tumor topography, histological grade