Participation of hMLH1, p63, and MDM2 proteins in the pathogenesis of syndromic and nonsyndromic keratocystic odontogenic tumors

(1)

Participation of hMLH1, p63, and MDM2 proteins in the

pathogenesis of syndromic and nonsyndromic keratocystic

odontogenic tumors

Bárbara Vanessa de Brito Monteiro, DDS, PhD,aRoberta Barroso Cavalcante, DDS, PhD,b Renato Luiz Maia Nogueira, DDS, PhD,cMárcia Cristina da Costa Miguel, DDS, PhD,a

Cassiano Francisco Weege Nonaka, DDS, PhD,dand Éricka Janine Dantas da Silveira, DDS, PhDa

Objectives. To evaluate the expression of hMLH1, p63, and MDM2 in Gorlin syndromeeassociated keratocystic odontogenic tumors (SKOTs) and nonsyndromic keratocystic odontogenic tumors (NSKOTs).

Study Design. Seventeen primary NSKOTs, 17 SKOTs, and 8 recurrent NSKOTs were analyzed by using immunohistochemistry.

Results. No significant differences in the hMLH1, p63, or MDM2 labeling indices were observed between groups (P ¼ .398; P ¼ .232; P ¼ .426, respectively). Higher hMLH1 immunoexpression was found in the basal layer of primary NSKOTs. Most KOTs exhibited p63 immunoexpression in the upper layers of the epithelium. MDM2 immunoexpression was observed in the upper epithelial layers of SKOTs and recurrent NSKOTs.

Conclusion. It was not possible to correlate the immunoexpression of hMLH1, p63, and MDM2 in SKOTs and primary and recurrent NSKOTs, suggesting that these proteins exert independent effects on the development of these groups of tumors. (Oral Surg Oral Med Oral Pathol Oral Radiol 2015;120:52-57)

Keratocystic odontogenic tumors (KOTs) are character-ized by high rates of recurrence, inﬁltrative growth, and an occasional association with nevoid basal cell carcinoma syndrome (NBCCS).1-4 There are uncertainties as to whether differences exist in the biologic behavior of Gorlin syndromeeassociated KOTs (SKOTs) and nonsyndromic KOTs (NSKOTs). In this respect, studies evaluating the proteins involved in cell cycle and apoptosis,5-7as well as extracellular matrix components8and proteases,7-11 sup-port the existence of a different biologic behavior in SKOTs and NSKOTs.

Protein p63 plays an important role in epithelial devel-opment and in the proliferation of craniofacial structures. In the human oral mucosa, p63 is restricted mainly to the basal and parabasal layers of normal epithelium. Since cystic odontogenic lesions are of epithelial origin, p63 may play an important role in the growth and progression of these lesions.3

The mismatch repair (MMR) system plays an important role in the maintenance of genetic stability.12,13This sys-tem is involved in the removal of errors in DNA bases that

arise during replication or as a result of DNA damage. Mutations in four genes of the MMR system (MSH2, hMLH1, MSH6, and PMS2) can inﬂuence the expression of p63 and MDM-2 and predispose to a variety of tumors.14-16

According to Castrili et al.,15hMSH2 and hMLH1 are expressed in human cells that are undergoing rapid renewal; their reduced expression has been reported in several tumors. MDM2 protein, together with p53, participates in a complex regulatory network in response to genotoxic stress. Thus, defects in this process are related to the development of several types of cancer. Other genes in the TP53 gene family, such as TP63 and TP73, exert similar functions in tumor suppression, apoptosis induction, and cell cycle control.3,5

Several studies have investigated the role of growth factors, tumor suppressor genes, and oncogenes in the onset and progression of KOT using immunohistochem-istry.17-22 However, studies focusing on DNA repair proteins are sparse, and there are no studies explaining the participation of hMLH1 and its association with cell cycle

a

Postgraduate Program of Oral Pathology, Department of Oral Pathology, Federal University of Rio Grande do Norte, Natal, Rio Grande do Norte, Brazil.

b

Department of Oral Pathology, University of Fortaleza, Fortaleza, Ceara, Brazil.

c

Department of Oral Surgery, Federal University of Ceara, Fortaleza, Ceara, Brazil.

d

Postgraduate Program of Dentistry, Department of Dentistry, State University of Paraiba, Campina Grande, Paraiba, Brazil.

Received for publication Aug 18, 2014; returned for revision Mar 10, 2015; accepted for publication Mar 12, 2015.

http://dx.doi.org/10.1016/j.oooo.2015.03.011

Statement of Clinical Relevance

This study contributes to the understanding of the biologic behavior of keratocystic odontogenic tumor (KOT). The association of hMLH1, MDM2, and p63 has been identiﬁed in colon cancer and head and neck carcinoma, and may provide additional insight into understanding the mechanisms related to the development and differences in behavior of KOT.

(2)

proteins in KOT. Based on the fact that microsatellite instability (MSI) is mainly caused by methylation-induced silencing of the hMLH1 gene, we conducted the present study to evaluate whether differences in the behavior of KOTs are associated with changes in the expression of the hMLH1 protein, in addition to p63 and MDM2, which can be inﬂuenced by defects in the production of this protein.

MATERIALS AND METHODS

The study was approved by the Ethics Committee of the Federal University of Rio Grande do Norte (UFRN), Natal, Brazil (No. 246.800). Seventeen cases of primary NSKOTs, 17 cases of SKOTs, and 8 cases of recurrent NSKOTs were retrieved from the archives of the Patho-logical Anatomy Services of the Departments of Oral Pathology of UFRN and University of Fortaleza (UNI-FOR) for the period between January 2005 and December 2012. The number of available institutional archival cases deﬁned the size of the sample. The histologic diagnosis was based on the third WHO classiﬁcation in all cases.23 All patients with Gorlin syndrome had been diagnosed according to the criteria proposed by Evans et al.4 and had multiple KOTs. The patients with sporadic KOTs had single lesions and had been submitted to clinical and radiographic assessment to exclude the presence of other manifestations of Gorlin syndrome. All recurrent lesions were treated with surgical enucleation and did not correspond to recurrences from any of the primary NSKOTs in this study. Serial 3-

m

m-thick sections were cut from the tissue blocks and processed for immunohistochemical analysis.

Immunohistochemistry

The tissue sections were deparafﬁnized and immersed in 3% hydrogen peroxide to block endogenous peroxidase activity. The sections were then washed in phosphate-buffered saline (PBS). Antigen retrieval, antibody dilu-tion, and clone type for p63, MDM2, and hMLH1 are shown inTable I. After treatment with normal serum, the sections were incubated with the primary antibodies in a

moist chamber. Next, the sections were washed twice in PBS and treated with a polymer-based complex (Envisionþ Dual Link System HRP; Dako, Carpin-teria, CA) at room temperature to bind the primary anti-bodies. Peroxidase activity was visualized by immersing the tissue sections in diaminobenzidine (Liquid DABþ; Dako, Carpinteria, CA), which resulted in a brown reac-tion product. Finally, the secreac-tions were counterstained with Mayer’s hematoxylin and coverslipped. Sections of colon carcinoma served as positive controls for all anti-bodies. Samples treated as described above were used as negative controls; however, the primary antibody was replaced with a solution of bovine serum albumin in PBS. Analysis of immunostaining

The tissue sections were examined in a blind fashion under an Olympus CX31 light microscope (Olympus Japan Co., Tokyo, Japan). The immunoexpression of p63, MDM2, and hMLH1 was evaluated only in the epithelial lining of KOTs by establishing the labeling index. Briefly, tissue sections were examined at 100 magnification to identify the area of the epithelial lining containing the largest number of immunostained cells. Next, 1000 epithelial cells were counted at 400 magnification in consecutive fields throughout all epithelial layers. The p63, MDM2, and hMLH1 label-ing indices are expressed as the percentage of immu-nostained nuclei in relation to the total number of nuclei counted. The intensity of the immunopositivity was not considered in the samples of the present study. Statistical analysis

The results were analyzed using the Statistical Package for the Social Sciences (version 17.0; SPSS, Inc., Chicago, IL). The nonparametric Kruskal-Wallis test was applied to compare the percentage of p63-, MDM2-, and hMLH1-immunopositive cells in the epithelial lining among SKOTs, primary NSKOTs, and recurrent NSKOTs. The Spearman correlation test was used to determine possible correlations among the percentages of p63-, MDM2-, and hMLH1-immunopositive cells. A level of signiﬁcance of 5% (P< .05) was adopted for all tests.

Table I. Properties of the primary antibodies

Clone Speciﬁcity Company Dilution Antigen retrieval Incubation

7 JUL P63 Santa Cruz Biotechnology,

Santa Cruz, CA 1:100 Citrate, pH 6.0 Pascal, 25’ at room temperature 60 minutes 1 B10 MDM2 Novocastra Laboratories,

Benton Lane, Newcastle upon Tyne, UK

1:50 Citrate, pH 6.0

Pascal, 15’ in microwave

60 minutes

G168-15 MLH1 Dako, Carpinteria, CA 1:50 Tris-EDTA, pH 9.0

Pascal, and 25’ at room temperature

(3)

RESULTS

Immunoexpression of hMLH1 in the epithelial component of KOT was positive in all cases analyzed. In primary NSKOT, immunostaining for hMLH1 was observed mainly in the basal epithelial layer (58.5%; n ¼ 10). In SKOT and recurrent NSKOT, immuno-staining was found in all layers of the epithelium, with no predominance in a speciﬁc layer. The median hMLH1 labeling index in the lining epithelium was 52.2 (range: 12.1-75.5) in the group of primary NSKOT, 56.9 (range: 23.3-69.7) in SKOT, and 41.2 (range: 10.8-69.0) in recurrent NSKOT. There was no signiﬁcant difference between groups (P ¼ .398, Kruskal-Wallis test) (Table II).

Immunoexpression of p63 in the epithelial component of KOT was positive in all cases analyzed. Immuno-staining predominated in the upper layers of the epithe-lium in 88.2% (n ¼ 15) of primary NSKOT (see

Figure 1B). In SKOT, immunoexpression was detected in the basal and upper layers. Immunostaining predominated in the upper epithelial layers in 75% of recurrent NSKOT. The median p63 labeling index was 57.0 (range: 29.1-74.3) in the group of primary NSKOT, 61.7 (range: 23.6-76.0) in SKOT, and 50.2 (range: 37.8-79.5) in recurrent NSKOT. No signiﬁcant difference was observed between groups (P ¼ .232, Kruskal-Wallis test) (seeTable II).

Immunoexpression of MDM2 in the epithelial component of KOT was positive in all cases analyzed. In primary NSKOT, 76% (n ¼ 13) exhibited immu-nostaining in the basal and upper layers of the epithe-lium. A predominance of immunostaining in the upper epithelial layers was observed in SKOT and recurrent NSKOT (see Figures 1G, 1H, and 1I). The median MDM2 labeling index in the epithelial lining was 61.9 (range: 31.7-91.5) for primary NSKOT, 57.0 (range: 25.9-78.6) for SKOT, and 54.35 (range: 13.8-75.9) for recurrent NSKOT. There was no signiﬁcant

difference between groups (P ¼ .426, Kruskal-Wallis test) (seeTable II).

The morphologic aspects of the lesions and details of the immunostaining are presented inFigure 1(A-L).

Spearman’s correlation test showed no signiﬁcant correlations among the p63, MDM2, and hMLH1 labeling indices in the epithelial component of KOT (P> .05).

DISCUSSION

The MMR repair system is important for maintaining genomic stability, reducing DNA errors, and preventing these errors from becomingﬁxed mutations during cell proliferation. MMR deﬁciencies are associated with a higher risk of development of different types of cancer.25 Immunohistochemical analysis of the expression of the hMSH2 and hMLH1 genes in tumors indicates that the expression of these genes is typically lost in patients with hereditary nonpolyposis colorectal cancer.13,14

Although the identification of mutations in the PTCH gene, among other factors, may support the classification of KOT as a neoplasm,2there are no studies in the literature that define the role of the MMR system in these lesions. In the present study, no significant difference in the hMLH1 labeling indices was observed among primary NSKOTs, recurrent NSKOTs, and SKOTs (P > .05). Therefore, possible differences in the biologic behaviors of these lesions do not occur because of alterations in hMLH1 expression.

Castrili et al.15showed that in ameloblastomas, hMSH2 and hMLH1 proteins are exclusively located in the nuclear compartment of the outer layer of the tumor nests. The restricted expression of hMSH2 and hMLH1 agrees with the probable stem-cell nature of the outer cell compart-ment from which inner stellate reticulum-like and squa-mous cells originate. The authors suggested that the development and progression of these tumors do not depend on defects in the human MMR system.15 Table II. Parameters used for the calculation of the Kruskal-Wallis (KW) test for the evaluation of P63, MDM2, and MLH1 labeling indices in primary nonsyndromic OKCs, recurrent nonsyndromic OKCs, and syndromic OKCs

Variable/group n Median Q25-Q75 Mean of ranks KW P

P63

Primary nonsyndromic OKC 17 57.00 49.35-66.55 21.18 2.924 .232

Recurrent nonsyndromic OKC 8 50.20 41.57-62.50 15.63

Syndromic OKC 17 61.70 56.00-71.45 24.59

MDM2

Syndromic OKC 17 57.00 40.80-69.30 20.24

MLH1

(4)

In primary NSKOT, immunopositive cells predomi-nated in the basal epithelial layer, whereas immunoex-pression was seen in all of the epithelial layers of SKOT and recurrent NSKOT. The presentﬁndings regarding the location of hMLH1 immunoexpression may indicate a more effective and consistent participation of the MMR system in primary KOT, preventing the perpetu-ation of DNA errors to the upper layers of the epithelium. On the other hand, the immunoexpression of hMLH1 in all of the epithelial layers of SKOT and recurrent NSKOT suggests greater genomic instability in these tumors compared with primary NSKOT. Moreover, the slight reduction in the hMLH1 labeling index observed

in cases of recurrent NSKOT might be related to the loss of control of cell proliferation mechanisms, which ex-plains the recurrence of these lesions.

The p63 gene expresses multiple isoforms. The most common isoform is p63

a

, which encodes the p63 protein, particularly

D

Np63

a

. This protein is abundant in the highly proliferative basal cells of adult epithelial tissues, including the oral epithelium. The p63 protein is also highly expressed the embryonic ectoderm, being essential in the differentiation of keratinocytes during embryonic development. Moreover, p63 participation in the prolifer-ation and differentiprolifer-ation of epithelial cells has been sug-gested, since it is a homolog of the tumor suppressor p53.18

Fig. 1. A, Primary KOT (hematoxylin and eosin [H&E], 400). B, Syndrome KOT (H&E, 200). C, Recurrent KOT (H&E,200). D, hMLH1 immunoexpression in basal and suprabasal layers of a primary KOT (labeling index: 57,2); E, hMLH1 immunoexpression in basal and suprabasal layers of a syndromic KOT (labeling index: 41,8); F, hMLH1 immunoexpression in a recurrent KOT observed mainly in superior epithelial layers (labeling index: 48,6) (Envision,400). G, p63 immunoexpression in suprabasal layers of a primary KOT (labeling index: 79,5); H, p63 immunoexpression in basal and suprabasal layers of a syndromic KOT (labeling index: 69,5). I, p63 immunoexpression mainly in epithelial suprabasal layers of a recurrent KOT (labeling index: 49,1). (Envision,400). J, Immunoexpression of MDM2 in all epithelial layers of a primary KOT (labeling index: 46,8); K, Immunoexpression of MDM2 mainly in suprabasal layers of a syndromic KOT (labeling index: 41,6) L, Immunoexpression of MDM2 mainly in epithelial suprabasal layers of a recurrent KOT (labeling index: 52,9) (Envision,400).

(5)

With respect to the presence of p63 in KOT, this protein is thought to contribute to the biologic proﬁle of these tumors, playing a role in the regulation of epithelial differentiation and favoring tumorigenesis.17

Overexpression of p63 has been observed in the epithelial lining of KOT.3,17,18 In the present study, immunoexpression of p63 was observed in all cases studied, but there was no signiﬁcant difference in the p63 labeling indices among primary NSKOT, recurrent NSKOT, and SKOT (P > .05). Similarly, Atarbashi-Moghadam et al.18 found no association between the expression of this protein in KOT and clinical features, such as recurrence and NBCCS.

With respect to the distribution of p63 immunoex-pression in primary and recurrent NSKOT, higher expression was observed in the upper layers of the epithelium. The same has been reported in the literature for KOT, regardless of the type of tumor (isolated, recurrent, or syndromic).18-20 The immunoexpression of this protein in more superﬁcial epithelial layers may be related to disturbances in cell cycle control, increasing the intrinsic growth potential. This fact would explain the inﬁltrative and aggressive growth of KOTs, as well as their higher recurrence rates, compared with other odontogenic lesions.3In general,

high p63 immunoexpression in KOT may reﬂect the immaturity of epithelial cells, favoring tumorigenesis, and therefore supports the hypothesis of a neoplastic origin of these tumors as well as the existence of a suprabasal proliferative compartment.17

In most human cancers, the tumor suppressor function of p53 is compromised by mutations in the TP53 gene or by partial inhibition of the regulation of this protein. Under normal conditions, p53 levels are precisely regulated by MDM2.21In the present study, MDM2 immunoexpression was observed in all KOT cases, in contrast to the literature, which reports immunoexpression in only 80% of the cases studied,21or even the complete absence of this marker in KOT.24 However, high MDM2 immunoreactivity in KOT cases compared with other odontogenic tumors has also been reported.22 Furthermore, in the present study, MDM2 immunoexpression predominated in the middle and upper compartments. This ﬁnding agrees with the observations of Shariﬁ-Sistani et al.,21who compared the immunoexpression of MDM2 and p53 between KOT and ameloblastoma and found a correlation between these proteins, that is, an increase in one protein would lead to an increase in the other. The authors suggested that the increased immunoexpression of MDM2 in some tumors would not be able to reduce the amount of mutated p53.21

(6)

The tumor suppressor gene functions of p53 have been shown to be mediated by the interaction with MMR genes and proteins.14,26DNA mismatch repair genes also play a role in the development of sporadic human neoplasms. Advances in our understanding of mismatch repair de ﬁ-ciency have important applications. Furthermore, the main function of MDM2 is to negatively regulate p53 by bind-ing directly to the latter. This bindbind-ing results in the for-mation of the MDM2/p53 complex, which inhibits the transcription activity of p53 and induces its degradation.26 However, no correlation among the hMLH1, p63, and MDM2 labeling indices was observed in the present study, suggesting that these proteins exert independent effects on the development of these tumors; inFigure 2, we have presented the pathway involving hMLH1, p63, and MDM-2. Future studies that include different mem-bers of DNA mismatch repair genes (hMSH2, hMSH6, and hPMS2) should be performed to further elucidate their role in the pathogenesis of these lesions.

REFERENCES

1.Nayak MT, Singh A, Singhvi A, Sharma R. Odontogenic kera-tocyst: what is in the name? J Nat Sci Biol Med. 2013;4:282-285. 2.Figueroa A, Correnti M, Avila M, Andea A, DeVilliers P, Rivera H. Keratocystic odontogenic tumor associated with nevoid basal cell carcinoma syndrome: similar behavior to sporadic type? Otolaryngol Head Neck Surg. 2010;142:179-183.

3.Gonçalves CK, Fregnani ER, Leon JE, Silva-Sousa YT, Perez DE. Immunohistochemical expression of p63, epidermal growth factor receptor (EGFR) and notch-1 in radicular cysts, dentigerous cysts and keratocystic odontogenic tumors. Braz Dent J. 2012;23:337-343. 4.Evans DG, Ladusans EJ, Rimmer S, Burnell LD, Thakker N,

Farndon PA. Complications of the naevoid basal cell carcinoma syndrome: results of a population based study. J Med Genet. 1993;30: 460-464.

5.Kimi K, Kumamoto H, Ooya K, Motegi K. Immunohisto-chemical analysis of cell-cycle- and apoptosis-related factors in lining epithelium of odontogenic keratocysts. J Oral Pathol Med. 2001;30:434-442.

6.Kolár Z, Geierová M, Bouchal J, Pazdera J, Zboril V, Tvrdý P. Immunohistochemical analysis of the biological potential of odontogenic keratocysts. J Oral Pathol Med. 2006;35:75-80. 7.Leonardi R, Matthews JB, Loreto C, et al. Beta-catenin and

sur-vivin expression in keratocystic odontogenic tumor (KCOT). A comparative immunohistochemical study in primary, recurrent, and nevoid basal cell carcinoma syndrome (NBCCS)-associated lesions. Histol Histopathol. 2013;28:1175-1184.

8.Amorim RF, Godoy GP, Galvão HC, Souza LB, Freitas RA. Immunohistochemical assessment of extracellular matrix com-ponents in syndrome and non-syndrome odontogenic keratocysts. Oral Dis. 2004;10:265-270.

9.Cavalcante RB, Pereira KM, Nonaka CF, Nogueira RL, de Souza LB. Immunohistochemical expression of MMPs 1, 7, and 26 in syndrome and nonsyndrome odontogenic keratocysts. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2008;106:923-929.

10.Leonardi R, Matthews JB, Caltabiano R, et al. MMP-13 expres-sion in keratocyst odontogenic tumor associated with NBCCS and sporadic keratocysts. Oral Dis. 2010;16:795-800.

11.Nonaka CF, Cavalcante RB, Nogueira RL, de Souza LB, Pinto LP. Immunohistochemical analysis of bone resorption regulators (RANKL and OPG), angiogenic index, and myoﬁbroblasts in syn-drome and non-synsyn-drome odontogenic keratocysts. Arch Oral Biol. 2012;57:230-237.

12.Graves B, Thompson P, Xia M, et al. Activation of the p53 pathway by small-molecule induced MDM2 and MDMX dimerization. Proc Natl Acad Sci U S A. 2012;109:11788-11793. 13.Li GM. Mechanisms and functions of DNA mismatch repair. Cell

Res. 2008;18:85-98.

14.Martin SA, Lord CJ, Ashworth A. Therapeutic targeting of the DNA mismatch repair pathway. Clin Cancer Res. 2010;16:5107-5113. 15.Castrilli G, Piantelli M, Artese L, et al. Expression of hMSH2 e

hMLH1 proteins of the human mismatch repair system in ame-loblastoma. J Oral Pathol Med. 2001;30:305-308.

16.Kheirelseid EA, Miller N, Chang KH, et al. Mismatch repair protein expression in colorectal cancer. J Gastrointest Oncol. 2013;4:397-408.

17.Gurgel CA, Ramos EA, Azevedo RA, Sarmento VA, Carvalho AM, Santos JN. Expression of Ki-67, p53 and p63 proteins in keratocyst odontogenic tumours: an immunohisto-chemical study. J Mol Hist. 2008;39:311-316.

18.Atarbashi-Moghadam S, Atarbashi-Moghadam F, Mokhtari S, Eini E. Immunohistochemical analysis of p63 expression in odontogenic lesions. Biomed Res Int. 2013;2013:624176. 19.Foschini MP, Cocchi R, Marucci G, et al. High DeltaN p63 isoform

expression favours recurrences in odontogenic keratocyst e odon-togenic keratocystic tumour. Int J Oral Maxillofac Surg. 2006;35: 673-675.

20.Lo Muzio L, Santarelli A, Caltabiano R, et al. p63 expression in odontogenic cysts. Int J Oral Maxillofac Surg. 2005;34:668-673. 21.Shariﬁ-Sistani N, Zartab H, Babakoohi S, et al.

Immunohisto-chemical comparison of the expression of p53 and MDM2 pro-teins in ameloblastomas and keratocystic odontogenic tumors. J Craniofac Surg. 2011;22:1652-1656.

22.Carvalhais JM, Aguiar MC, Araújo VC, Gomez R. p53 and MDM2 expression in odontogenic in cysts and tumours. Oral Dis. 1999;5:218-222.

23.Philipsen HP. Keratocystic odontogenic tumor. In: Barnes L, Eveson JW, Reichart P, Sidransky D, eds. World Health Orga-nization classiﬁcation of tumors: pathology and genetics of head and neck tumors, IARC Press, Lyon 2005: pp. 284-317. 24.Galvão HC, Gordón-Núnez MA, Amorim RF, Freitas RA,

Souza LB. Immunohistochemical expression of protein 53, mu-rine double minute 2, B-cell lymphoma 2, and proliferating cell nuclear antigen in odontogenic cysts and keratocystic odonto-genic tumor. Indian J Dent Res. 2013;24:369-374.

25.Xiao X, Melton DW, Gourley C. Mismatch repair deﬁciency in

ovarian cancerdmolecular characteristics and clinical implica-tions. Gynecol Oncol. 2014;132:506-512.

26.Marcus VA, Madlensky L, Gryfe R, et al. Immunohistochemistry for hMLH1 and hMSH2: a practical test for DNA mismatch repair-deﬁcient tumors. Am J Surg Pathol. 1999;23:1248-1255.

Reprint requests:

Éricka Janine Dantas da Silveira, DDS, PhD Universidade Federal do Rio Grande do Norte Departamento de Odontologia

Programa de Pós- Graduação em Patologia Oral Av. Senador Salgado Filho

1787, Lagoa Nova, Natal/RN CEP 59056-000, Brazil