Expression of the p53 Homologue p63 in Early Cervical Neoplasia

Gynecologic Oncology 80, 24 –29 (2001) doi:10.1006/gyno.2000.5953, available online at http://www.idealibrary.com on

Expression of the p53 Homologue p63 in Early Cervical Neoplasia 1,2 Bradley J. Quade, M.D., Ph.D.,* Annie Yang, B.A.,† Yunmei Wang, Ph.D.,† Deqin Sun, B.S.,* Jeong-ja Park, M.D.,‡ Ellen E. Sheets, M.D.,§ Aida Cviko, M.D.,* Jerome M. Federschneider, M.D.,§ Roxana Peters, ¶ Frank D. McKeon, Ph.D.,† and Christopher P. Crum, M.D.* *Division of Women’s and Perinatal Pathology, Department of Pathology, and §Division of Oncology, Department of Obstetrics, Gynecology, and Reproductive Biology, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts 02115; †Department of Cell Biology, Harvard Medical School, Boston, Massachusetts; ‡Dongsan Hospital, Taegu, South Korea; and ¶Department of Biology, Simmons College, Boston, Massachusetts Received April 7, 2000; published online December 4, 2000

INTRODUCTION

Background. p63, a homologue of the tumor suppressor gene p53, is expressed in embryonic, adult murine, and human basal squamous epithelium and encodes both transactivating and dominant negative transcript isoforms. Mouse embryos functionally deficient in p63 fail to replenish basal squamous epithelial cells, resulting in multiple defects that include absent genital squamous epithelium. This study investigated the expression of p63 in the human cervical transformation zone and early cervical neoplasia. Methods. Tissue localization of p63 was determined by immunohistochemistry in a wide range of epithelia. A correlation was also made between p63 expression and squamous basal cell (keratin 14), endocervical columnar cell (mucicarmine), and cell-cycle specific (Ki-67) markers. Results. p63 expression by immunostaining delineated basal and parabasal cells of maturing ectocervical squamous mucosa, squamous metaplasia in the cervix, and basal and subcolumnar cells of the cervical transformation zone. In atrophic epithelia immunostaining for p63 was present in all cell strata. In early cervical neoplasia, p63 expression was inversely correlated with both squamous cell maturation and nonsquamous differentiation in CIN. This biomarker also identified basal cells in a subset of preinvasive cervical neoplasms with endocervical cell differentiation that were bcl-2 and keratin 14 negative. Conclusions. In the lower female genital tract, p63 is preferentially expressed in immature cells of squamous lineage and is not linked to cell proliferation. The broader range of p63 expression relevant to keratin 14 and bcl-2 indicates that p63 may identify additional subsets of benign and neoplastic epithelial basal cells in the cervical transformation zone and may be useful in studying cell differentiation in the early stages of neoplastic change in this region. © 2001 Academic Press Key Words: p63; p53 homologue; cervical neoplasia; squamous intraepithelial lesion (SIL); human papilloma virus (HPV); reserve cell.

During reproductive life, the uterine cervical mucosa is a dynamic of diverse epithelial phenotypes, evolving in association with hormonal changes, inflammatory stimuli, and human papillomavirus (HPV)-mediated neoplastic transformation. This process takes place at the transition from ectocervical squamous mucosa to endocervical columnar mucosa (transformation zone). The initial phase of transformation zone development is characterized by proliferation of immature epithelial cells at the squamocolumnar junction that, over time, mature into squamous epithelium identical to the ectocervical epithelium (epidermidalization). Cephalad to this process, subcolumnar reserve cells are situated beneath the endocervical columnar epithelium and possess the capacity to undergo squamous differentiation (metaplasia) [1–3]. The squamocolumnar junction of the transformation zone is vulnerable to HPV infections that can result in preinvasive and invasive squamous neoplasia [4]. In addition to squamous cell neoplasms, nonsquamous cervical tumors (glandular and neuroendocrine) also develop in this region, consistent with either an aberrant pathway of neoplastic cell fate or a multiplicity of target cells for viral infection. This alternate pathway of tumorigenesis is distinguished by its frequent association with a particular HPV subtype, HPV 18 [4 – 6]. Two of us previously described the cloning of a new gene bearing a high homology to the tumor suppressor gene, p53, designated p63 [7]. p63 encodes two distinct transcript isoforms, including one full-length (p63 TA) form with transactivating properties and one truncated (p63 ⌬N) form with dominant-negative activity [7]. Immunohistochemical studies detected p63 expression in proliferating cells of mouse and human tissues, including epidermis, urothelium, prostate, and cervix. Recently, we have shown that functional inactivation of murine p63 results in complex deformities in the late mouse embryo, including severe defects in craniofacial and limb development [8]. The abnormalities are accompanied by a nonregenerative differentiation of the epidermis and squamous mucosa including that of the genital tract, indicating that this

1 Supported in part by National Institutes of Health Grants CA 72594 (B.J.Q.) and CA75340 (F.D.M) and by grants from the Partner’s Dana Farber Cancer Center (C.P.C.) and Harvard Medical School/Brigham and Women’s Hospital (F.M., C.P.C.). 2 Presented in part at the United States Academy of Pathology meeting, San Francisco CA, March 1999, and at the Gordon Conference on Cancer, Salve Regina University, Newport, RI, August 1999.

0090-8258/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

24

25

p63 IN CERVICAL NEOPLASIA

gene is critical to normal mucocutaneous epithelial development and function [8]. We studied p63 expression in adult human cervical epithelium with the goal of determining the distribution of p63 expression in nonneoplastic and preinvasive lower female genital tract epithelium. The relationship of p63 expression to other squamous, columnar, and basal cell biomarkers and the p63 transcripts encoded in these conditions was also examined. MATERIALS AND METHODS Case Selection and Classification Archived cervical specimens representing a wide range of cervical disease processes were selected for analysis from the case files of the Division of Women’s and Perinatal Pathology at Brigham and Women’s Hospital. Seventy-nine different examples of epithelium were studied, including 23 samples of mature or immature squamous epithelium with normal or reactive epithelial changes, 6 cases of atrophy, 26 cervical intraepithelial neoplasms (CIN) of all grades, 12 adenocarcinomas in situ (ACIS), and 12 preinvasive lesions exhibiting both squamous (CIN) and columnar (ACIS) differentiation. A portion of the latter have been studied in a separate report by these authors [9]. Immunohistochemical Analysis p63. Preparation of a monoclonal antibody to p63 was described in detail previously [7]. Briefly, a cDNA fragment containing the N-terminal portion (aa 1–205) of ⌬Np63 was overexpressed as a GST fusion protein and used for subsequent monoclonal antibody production. Immunostaining for p63 was performed on deparaffinized sections using horseradish-peroxidase-conjugated goat anti-mouse IgG antibody as previously described [7]. Other markers. Assessment of cell proliferation was determined by immunostaining for Ki-67 using the Mib-1 antibody, as previously described [10, 11]. Keratin 14, a type I intermediate filament distinguishing stratifying epithelial cells from simple epithelium, was detected using a monoclonal antibody to a conjugated synthetic peptide derived from the carboxy terminus (Research Diagnostics, Flanders, NJ) [12]. Bcl-2, a protein playing a central role in the inhibition of apoptosis and which has been localized to basal squamous and subcolumnar reserve cells, was detected using an antibody to a synthetic peptide composed of amino acids 41–54 of the BCL-2 protein (Dako, Carpinteria, CA) [13]. All immunohistochemical analyses were performed according to the manufacturer’s recommendations using peroxidase and diaminobenzidine. Pretreatments/dilutions were as follows: bcl-2 (microwave/1:25) and keratin 14 (microwave/1:50). Mucicarmine staining was performed according to the manufacturer’s recommendations (Sigma, St. Louis, MO). Controls. Specificity of staining for the antigens under study, including p63, was verified by the following: (1) In each

case, stromal cells, which do not exhibit p63 positivity in animal models, were used as an internal control in each case; (2) The use of multiple antibodies in most of the cases permitted the exclusion of nonspecific staining linked to other immunohistochemical reagents; (3) A portion of p63 immunoreactive cases were reacted with a monoclonal antibody generated to the TA-isoform of p63, which is expressed in very low levels in benign and neoplastic lower genital tract epithelium (A. Wang, C. P. Crum, and F. D. McKeon, unpublished). Staining with this antibody was consistently negative. RESULTS Tissue Localization of p63 Expression Normal mucosa. p63 was consistently expressed in all normal tissues in several distinct sites. Staining intensities varied somewhat between individual cases, but the consistent expression in certain cell types suggested that differences in intensity reflected technical or fixation variables influencing antigen exposure. Expression of p63 was always nuclear and consistently localized to basal and parabasal cells of the ectocervix and maturing transformation zone. Staining above the basal cell layers varied slightly between epithelia but was typically reduced in intensity and cell number as a function of epithelial maturation (Figs. 1A and 1B). In the transformation zone and endocervix, p63 consistently stained subcolumnar (reserve) cell nuclei (Fig. 1C). In the surface epithelial cells distal to the mature transformation zone mucosa, p63 highlighted basal cells in stratified epithelia exhibiting two pathways of differentiation. The most common consisted of expanded groups of subcolumnar cells undergoing squamous metaplasia (Figs. 1D and 1E). The other and less common pathway consisted of stratified epithelia undergoing gradual transitions from basal cells to mature columnar cells (Figs. 1F and 1G, compare to Fig. 1C). In this epithelium, intervening layers of squamous metaplasia were not observed, and loss of p63 staining coincided with the transition from basal cell to columnar cell differentiation (Fig. 1H). The association of p63 staining with immature squamous cells was preserved in atrophic mucosa of postmenopausal women, where the proportion of epithelial cells staining expanded, including full-thickness staining in cases of marked atrophy (Figs. 2D and 2L). Cervical neoplasia. In cervical intraepithelial neoplasia, p63 was tightly linked to cell maturation similar to normal squamous epithelium. In CIN1, p63 was typically localized to the basal and parabasal cells, extending into the middle and upper layers in less well differentiated lesions (CIN2-CIN3) (Figs. 2A–2C and 2I–2K). In contrast, p63 immunostaining was absent in the neoplastic glands of adenocarcinomas in situ. In lesions consisting of both CIN and ACIS, p63 staining disappeared with transitions from squamous to glandular differentiation. Two exceptions were seen, one of which was the presence of p63-positive basal cells in areas of poorly differentiated adenocarcinoma in situ (see Fig. 3O). In these foci, the

26

QUADE ET AL.

FIG. 1. In the normal ectocervix or mature transformation zone (A, H&E), expression of p63 is concentrated in basal and parabasal cells (B). In endocervical crypts (e), p63-positive subcolumnar cells undermine mature columnar epithelium (C). Areas of immature squamous metaplasia (in D, H&E) stain strongly for p63 (E). In areas of transformation zone undergoing columnar differentiation (F, H&E) with mucicarmine positivity (G, arrows), p63 expression is confined to the basal cells (H, arrows). FIG. 2. Expression of p63 and a cell proliferation marker (Ki-67) in CIN1, CIN2, CIN3, and atrophy (A–D, H&E). Ki-67 activity extends into suprabasal strata or intermediate cells undergoing viral cytopathic effect in CIN1 (E, arrowheads) and superficial cells of CIN2 and CIN3 (F and G), but is absent or sharply reduced in atrophy (H). Note that p63 expression spares the superficial cells of CIN1 and CIN2 (I and J) but stains virtually all cell nuclei of CIN3 and atrophy (K and L), underscoring the lack of relationship between cell cycle activity and p63 expression.

p63 IN CERVICAL NEOPLASIA

27

activity was concentrated in the first suprabasal cell layer (data not shown). In neoplastic epithelium (CIN), Ki-67 and p63 typically coexpressed in immature epithelium (Fig. 2). Discordant expression occurred in CIN1, in which Ki-67-positive mature cells showing viral cytopathic effect were p63 negative (Figs. 2E and 2I), and CIN2, where Ki-67 staining often occurred in cell layers above the epithelial layers where p63 immunostaining ceased (Figs. 2F and 2J, arrow). In immature epithelia exhibiting both high (CIN3, Figs. 2G and 2K) and low (atrophy, Figs. 2H and 2I) levels of Ki-67 expression, p63 expression was consistently high. Comparison of p63, bcl-2, and Keratin 14 Expression

FIG. 3. Comparisons of p63, keratin 14, and bcl-2 expression. In maturing squamous epithelium (A, H&E), p63 stains both basal and immature suprabasal cells (F), whereas bcl-2 and keratin 14 are limited to the basal cell layer (K and P, respectively). Single rows of subcolumnar cells (B, H&E) are consistently stained by p63 and bcl-2 (G and L, respectively), but are frequently keratin 14 negative (Q). In atrophic mucosa (C, H&E), p63 staining extends into the upper layers (H), in contrast to bcl-2 and keratin 14, which are more basal located (M and R). In CIN3 (D, H&E), p63 immunostaining is extensive (I), but bcl-2 and keratin 14 staining are variable or absent (N and S, respectively). In adenosquamous carcinoma in situ (E, H&E) with mucicarmine-positive cells (J, arrow), p63 highlights foci of intensely staining basal cells with loss of staining in areas of columnar differentiation (O). In contrast, keratin 14 staining is negative (not shown) and bcl-2 reveals rare weakly staining cells (T).

epithelium resembled CIN, but mucin production was identified in the suprabasal cell layers. The latter were p63 negative. A second exception was the presence of p63-positive nuclei beneath the basal epithelium of ACIS. The distribution of these cells was characteristic of normal reserve cells and they were interpreted as residual normal reserve cells not replaced by the neoplastic process (not shown). Correlation of p63 and Cell Cycle Activity in Benign and Neoplastic Epithelium To determine whether the relationship between p63 expression and differentiation was related to cell proliferation, 27 epithelia were also stained for Ki-67 with the Mib-1 antibody. In normal mature squamous mucosa p63 expression was present in both basal and suprabasal epithelium, whereas Ki-67

Thirty-six and 37 epithelia, respectively, were stained for bcl-2 and cytokeratin 14. In mature squamous epithelium (Fig. 3A) and mature squamous metaplasia, bcl-2 and keratin 14 immunostaining was cytoplasmic and typically confined to basal cells or, less commonly, basal and parabasal cells (Figs. 3K and 3P). In contrast, p63 immunostaining extended into the lower to middle third of some squamous epithelia (Fig. 3F). In some squamous mucosa, keratin 14 staining was patchy and involved the upper layers. In contrast, p63 staining was more uniformly distributed and invariably linked to maturation. Subcolumnar cell staining (Fig. 3B) by the three antibodies varied. p63 and bcl-2 consistently stained single layers of subcolumnar cells (Figs. 3G and 3L), but keratin 14 staining tended to be weak or negative (Fig. 3Q), except in areas of subcolumnar cell expansion with conspicuous squamous differentiation (not shown). In atrophic mucosa (Fig. 3C), bcl-2 and keratin 14 staining remained concentrated in the basal layers (Figs. 3M and 3R) while the distribution of p63 typically expanded to the middle and upper thirds (Fig. 3H). In CIN, immunostaining for both bcl-2 and keratin 14 varied. In some cases, staining with both was intense (not shown). In others, particularly poorly differentiated (CIN3) lesions (Fig. 3D), staining for bcl-2 and keratin 14 (Figs. 3N and 3S) was focal or absent, in contrast to p63 which was more intense (Fig. 3I). In summary, bcl-2 staining targeted a smaller, basal situated subset of benign and neoplastic squamous/reserve cells. Keratin 14 staining typically targeted immature cells undergoing squamous differentiation, but varied, often sparing subcolumnar cells and occasionally expanding into maturing squamous cells. In contrast, p63 was consistently expressed in both squamous and subcolumnar cells and maintained a broad distribution of expression in high-grade CIN. A striking difference in localization between p63 and the other two biomarkers was observed in “adenosquamous” carcinomas in situ, which were stratified intraepithelial lesions resembling CIN3 (Fig. 3E) but exhibiting columnar cell differentiation with mucin production (Fig. 3J). In these lesions, discrete foci of intense basal/parabasal p63 immunostaining were identified (Fig. 3O). These foci did not stain for keratin 14 (not shown) and stained only weakly or focally for bcl-2 (Fig. 3T). In mixed lesions containing both adenosquamous carci-

28

QUADE ET AL.

nomas in situ and conventional adenocarcinomas in situ, the latter were invariably p63 negative. DISCUSSION The findings from this study suggest that in the cervix, p63 is a marker for the immature squamous cell and for basal cells giving rise to both squamous and, in some circumstances, columnar cell differentiation. A consistent relationship between p63 expression and immature cells was maintained in both benign and neoplastic cervical epithelium (Figs. 1–3). This association suggests that p63 may be useful for classifying cervical neoplasia, and a study of p63 expression in cervical carcinomas is in progress (T.-Y. Wang et al., manuscript in preparation). The association between the immature cell phenotype and expression of this gene was also demonstrated in atrophic epithelium, suggesting an inverse relationship between the effects of estrogenic hormones—which stimulate terminal differentiation—and p63 localization. These properties of p63 are consistent with the p63-null mouse model, which infers strong associations between p63 and (1) the perpetuation of squamous (or transitional) epithelial growth and (2) the development of adnexal structures derived from ectoderm, including skin appendages, breast, salivary gland, prostate, and cervix [8]. A common factor linking both is the basal cell, which must be both capable of proliferating to ensure the fate of the squamous epithelium and (in some sites) multipotent, to generate a diversity of cell phenotypes seen in these organs. Because p63 has recently been characterized, the precise mechanisms by which this gene influences—and is regulated in— epithelial differentiation have yet to be elucidated. The dominant-negative isoform (p63 ⌬N) has been shown in earlier studies to be the principal p63 RNA transcribed in epithelial tissues [7]. The functional importance of the dominant-negative isoform is supported by a recent report linking the EEC (ectodermal dysplasia, ectrodactyly, cleft lip/palate) syndrome in humans to a mutation in the p63 gene [14]. The EEC syndrome is characterized by limb and urogenital abnormalities and is linked to heterozygous mutations in the region encoding this transcript [14]. The dominant-negative isoform has also been the predominant transcript isolated in studies of normal and neoplastic squamous epithelium in the cervix and vulva (F. McKeon and C. P. Crum, unpublished). One study suggested that expression of the TA isoform is increased in neoplasia, but this observation remains to be confirmed [15]. Like p53, p63 is not induced during normal cell cycle progression [16]. The absence of either gene does not prevent cell proliferation [8, 16] and, as shown in this report, Ki-67 and p63 are frequently not coexpressed (Fig. 3). Furthermore, absence of p63 immunostaining of koilocytic and other superficial cell nuclei where HPV is commonly expressed indicates that vegetative viral DNA replication does not stimulate p63 expression [10, 11]. This lack of relationship among HPV, cell cycle activity, and p63 expression is supported indirectly by

studies of in vitro interactions between p63 and either its homologue p53 or papillomaviral oncogenes. Consistent with its nuclear localization, p63 is a weak transcriptional activator of p53 targets. Yet, p63 homodimers interact weakly and not at all with p53 and another homologue p73, suggesting that other targets exist and remain to be defined [17, 18]. Of note, viral oncoproteins such as the large T-antigen of SV40 and HPV18 E6 do not inhibit p63 (p51A)-mediated transcription [19]. An alternate hypothesis for p63’s functional role is modulation of transcription. A recent report has found that C-terminal extensions unique to both p63 and a related homologue p73 contain sterile alpha motif (SAM) domains [20]. Because SAM domains contain protein modules involved in protein–protein interactions, the influence of p63 on cell differentiation may be mediated by functional regulation of other protein effector molecules in the nucleus. Moreover, because each p63 isoform has three potential C-termini (␣, ␤, ␥), further transcript modifications in this region conceivably may influence the activity of these proteins [7]. Localization of p63 antigen in cervical epithelium is similar to several other genes, including p53, keratins 14 and 19, bcl-2, CD44, and ␤-integrins [12, 13, 21–24]. Like p53, p63 expression is nuclear and localizes to basal and suprabasal cells of normal mucosa. However, in normal mucosa, p53 expression, as determined by immunohistochemistry, is at a very low intensity relative to p63 [21]. Moreover, the uniform expression of p63 in immature neoplastic squamous mucosa is in sharp contrast to p53, which is detected sporadically [21]. p63, keratin 19, and CD44 define basal and subcolumnar cells in normal transformation zone but, in contrast to p63, the other two proteins are also localized to adenocarcinomas [22, 23]. Proteins most closely resembling p63 in expression pattern are keratin 14, bcl-2, and ␤-integrins [12, 13, 24]. Antibodies to all of these markers react with similar cell populations in squamous mucosa and are reportedly expressed in high-grade squamous precursors [12, 13, 24]. Moreover, expression of both keratin 14 and bcl-2 is generally reduced or absent in cervical adenocarcinomas [12, 13]. In this study, a major difference between p63 and these two biomarkers was the consistent expression of p63 in immature neoplastic squamous epithelial cells. In contrast, bcl-2 tended to concentrate in the basal epithelium, including both atrophic and neoplastic squamous mucosa. This staining pattern was more limited than that of p63 and in keeping with a stronger—and inverse—relationship between bcl-2 and differentiation/apoptosis. Keratin 14, like p63, was usually confined to immature squamous epithelium. However, keratin 14 staining was frequently absent in subcolumnar reserve cells that had not expanded and undergone squamous differentiation. This suggests that keratin 14 is expressed in only the subset of reserve cells undergoing squamous differentiation. In contrast, other subsets express only p63 and bcl-2, raising the possibility that these cells may be multipotential. Although it is presumed that both squamous and glandular neoplasms originate in the vicinity of the cervical transforma-

p63 IN CERVICAL NEOPLASIA

tion zone, the cell population from which the latter arises is unclear [1–3]. Possibilities include mature endocervical columnar epithelium, subcolumnar reserve cells, and specialized basal cells dedicated strictly to columnar differentiation. The latter two possibilities are supported indirectly by the strong association between columnar cell neoplasia and the transformation zone and the identification of putative receptors to papillomaviruses in basal epithelial cells [25]. Given the proximity of subcolumnar reserve cells to stratified metaplastic and columnar epithelia, it is likely that the latter represent different differentiation pathways from the former (Figs. 1C–1H). It is significant that this range of differentiation seen in benign transformation zone mucosa parallels neoplasms arising in this site, including squamous, glandular, and adenosquamous lesions. Striking differences in immunostaining among p63, bcl-2, and keratin 14 were observed in preinvasive lesions of the transformation zone, specifically high-grade squamous precursors and adenosquamous carcinomas in situ [9]. The latter lesions include combinations of CIN, ACIS, and intermediate lesions that resemble CIN but are mucicarmine positive. Selective expression of p63 was seen in some of the latter epithelia, suggesting that p63 may identify unique populations of neoplastic basal cells not previously targeted by squamousspecific (keratin 14) biomarkers (Fig. 3). Whether these p63positive cells represent unique subpopulations of neoplastic basal or reserve cells will require additional study. However, p63 may facilitate their identification and the study of additional differentiation pathways in benign and neoplastic cervical transformation zone. REFERENCES 1. Geschickter CF, Fernandez F: Epidermidalization of the cervix. Ann NY Acad Sci 97:638 – 652, 1962 2. Reagan JW, Patten SF Jr: Dysplasia: a basic reaction to injury in the uterine cervix. Ann NY Acad Sci 97:662– 682, 1962 3. Richart RM: Cervical intraepithelial neoplasia. In Sommers SC (ed): Pathology Annual. New York, Appleton, 1973, pp 301–328 4. Alani RM, Munger K: Human papillomaviruses and associated malignancies. J Clin Oncol 16:330 –337, 1998 5. Stoler MH, Mills SE, Gersell DJ, Walker AN: Small-cell neuroendocrine carcinoma of the cervix. A human papillomavirus type 18-associated cancer. Am J Surg Pathol 15:28 –32, 1991 6. Smotkin D, Berek JS, Fu YS, Hacker NF, Major FJ, Wettstein FO: Human papillomavirus deoxyribonucleic acid in adenocarcinoma and adenosquamous carcinoma of the uterine cervix. Obstet Gynecol 68:241–244, 1986 7. Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch V, et al.: p63, a p53 homolog at 3q27–29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Mol Cell 2:305–316, 1998 8. Yang A, Schweitzer R, Sun D, Kaghad M, Walker N, Bronson RT, et al.:

29

p63 is essential for regenerative proliferation in limb, craniofacial and epithelial development. Nature 398:714 –718, 1999 9. Park JJ, Sun D, Quade BJ, Flynn C, Sheets EE, Yang A, McKeon FD, Crum CP: Stratified mucin-producing intraepithelial lesions of the cervix: adenosquamous or columnar cell neoplasia? Am J Surg Pathol, in press 10. Resnick M, Lester S, Tate JE, Sheets EE, Sparks C, Crum CP: Viral and histopathologic correlates of MN and MIB-1 expression in cervical intraepithelial neoplasia. Hum Pathol 27:234 –239, 1996 11. Cattoretti G, Becker MH, Key G, Duchrow M, Schluter C, Galle J, et al.: Monoclonal antibodies against recombinant parts of the Ki-67 antigen (MIB 1 and MIB 3) detect proliferating cells in microwave-processed formalin-fixed paraffin sections. J Pathol 168:357–363, 1992 12. Smedts F, Ramaekers F, Troyanovsky S, Pruszczynski M, Robben H, Lane B, Leight I, Plantema F, Vooijs P: Basal-cell keratins in cervical reserve cells and a comparison to their expression in cervical intraepithelial neoplasia. Am J Pathol 140:601– 612, 1992 13. Ter Harmsel B, Smedts F, Juijpers J, Jeunink M, Trimbos B, Ramaekers F: Bcl-2 immunoreactivity increases with severity of CIN: a study of normal cervical epithelia, CIN, and cervical carcinoma. J Pathol 179:26 – 30, 1996 14. Rollnick BR, Hoo JJ: Genitourinary anomalies are a component manifestation in the ectodermal dysplasia, ectrodactyly, cleft lip/palate (EEC) syndrome. Am J Med Genet 29:131–136, 1988 15. Nishi H, Isaka K, Sagawa Y, Usuda S, Fujito A, Ito H, Senoo M, Kato H, Takayama M: Mutation and transcription analyses of the p63 gene in cervical carcinoma. Int J Oncol 15:1149 –1153, 1999 16. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA Jr, Butel JS, Bradley A: Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature 356:215–221, 1992 17. Shimada A, Kato S, Enjo K, Osada M, Ikawa Y, Kohno K, Obinata M, Kanamaru R, Ikawa S, Ishioka C: The transcriptional activities of p53 and its homologue p51/p63: similarities and differences. Cancer Res 59:2781– 2786, 1999 18. Davison TS, Vagner C, Kaghad M, Ayed A, Caput D, Arrowsmith CH: P73 and p63 are homotetramers capable of weak heterotypic interactions with each other but not with p53. J Biol Chem 274:18709 –18714, 1999 19. Roth J, Dobbelstein M: Failure of viral oncoproteins to target the p53homologue p51A. J Gen Virol 80:3251–3255, 1999 20. Thanos CD, Bowie JU: P53 family members p63 and p73 are SAM domain-containing proteins. Protein Sci 8:1708 –1710, 1999 21. Hunt CR, Hale RJ, Buckley CH, Hunt J: p53 expression in carcinoma of the cervix. J Clin Pathol 49:971–974, 1996 22. Kwaspen FH, Smedts FM, Broos A, Bulten H, Debie WM, Ramaekers FC: Reproducible and highly sensitive detection of the broad spectrum epithelial marker keratin 19 in routine cancer diagnosis. Histopathology 31:503– 516, 1997 23. Shimabukuro K, Toyama-Sorimachi N, Ozaki Y, Goi T, Furukawa K, Miyasaka M, et al.: The expression patterns of standard and variant CD44 molecules in normal uterine cervix and cervical cancer. Gynecol Oncol 64:26 –34, 1997 24. Hughes DE, Rebello G, al-Nafussi A: Integrin expression in squamous neoplasia of the cervix. J Pathol 173:97–104, 1994 25. Evander M, Frazer IH, Payne E, Qi YM, Hengst K, McMillan NA: Identification of the ␣ 6 integrin as a candidate receptor for papillomaviruses. J Virol 71:2449 –2456, 1997