[CANCER RESEARCH 63, 8173–8180, December 1, 2003]

Comparative Analysis of Cervical Cancer in Women and in a HumanPapillomavirus-Transgenic Mouse Model: Identification ofMinichromosome Maintenance Protein 7 as an InformativeBiomarker for Human Cervical Cancer

Tiffany Brake,1 Joseph P. Connor,2 Daniel G. Petereit,3 and Paul F. Lambert1

1McArdle Laboratory for Cancer Research and 2Department of Obstetrics and Gynecology, University of Wisconsin Medical School, Madison, Wisconsin, and 3Department ofRadiation Oncology, John T. Vucurevich Cancer Care Institute, Rapid City, South Dakota

ABSTRACT

Human papillomaviruses (HPVs), such as HPV-16, are associated with>99% of cervical cancers in women. Two viral oncogenes, E6 and E7, areselectively expressed in these cancers. K14E6 and K14E7 transgenicmouse strains, which express the HPV16 E6 or E7 gene in stratifiedsquamous epithelia, display many acute and long-term phenotypes indic-ative of the oncogenic potential of E6 and E7 including epithelial hyper-plasia, abrogation of normal DNA damage responses, and spontaneousskin tumors. When treated with estrogen, these HPV-16 transgenic micedevelop a progressive disease leading to cervical cancer that shows manyhistopathological parallels to cervical cancer in women. In this study, weevaluated the cervical lesions that arise in these transgenic mice for theexpression of biomarkers induced in human cervical cancer. These anal-yses, which showed close parallels in the timing and pattern of expressionof cyclin E and Ki-67 in the mouse cervical disease compared with that inhumans, provided further validation of this HPV-16 transgenic mousemodel for human cervical cancer. We then used our mouse model toidentify minichromosome maintenance protein 7 (MCM7), an E2F-inducedcellular DNA replication factor, as a novel biomarker for cervical cancer.In both the mouse and human disease, strong, full thickness staining forMCM7 was seen selectively in the epithelium of high-grade intraepitheliallesions and in frank cancer. The uniform staining pattern and strongsignal for MCM7 suggest that MCM7 may be a highly informative biomar-ker for cervical cancer.

INTRODUCTION

A subset of HPVs4 is associated with at least 99% of cervicalcancers, a leading cause of cancer deaths in women worldwide (1). Acommon feature among cervical cancers is the selective up-regulationof two HPV genes, E6 and E7, which encode multifunctional proteinsbest known for their ability to inactivate the cellular tumor suppres-sors, p53 and pRb, respectively (reviewed in Ref. 2). Tissue cultureand animal model studies have clearly established the oncogenicpotential of E6 and E7. They contribute to cell immortalization intissue culture (3), induce epithelial cell hyperplasia in vivo (4, 5),abrogate DNA damage responses (6, 7), and induce genomic insta-bility (8, 9). Using transgenic mice expressing either E6 or E7 fromHPV-16, the papillomavirus genotype most commonly associatedwith human cervical cancer, behind the human keratin 14 promoter,we have previously dissected the contributions of E6 and E7 in cancerboth in the skin (4, 5, 10) and the cervix (11). In the cervix, E7 was

found to cooperate with estrogen to induce high grade dysplasia(HSIL), carcinoma in situ, and microinvasive cancers (11). AlthoughE6, in the absence of E7, had little effect in the cervix beyond thatseen with estrogen treatment of nontransgenic mice [i.e., hyperplasiaand low-grade dysplasia (LSIL) only], the two viral oncogenes to-gether gave rise to large, frank cervical cancers in �25% of theestrogen-treated doubly transgenic mice, indicating a role of E6 inmalignant progression (11). The histopathology of the progressivedisease that arises in these HPV transgenic mice mimics closely thatseen in women with cervical dysplasia and invasive carcinoma (11).In the current study, we assess the expression pattern of knownbiomarkers for human cervical cancer in the cervix of HPV transgenicmice and, using this animal model, identify a novel biomarker forhuman cervical cancer.

Ki-67 is a general proliferation marker used widely to characterizemalignant lesions, including those of the cervix. The degree of Ki-67positivity is correlated to the severity of cervical lesions: the numberof cells staining positive increases with an increasing degree ofdysplasia, as does the penetration of the staining into the upper layersof the epithelium (12). Cyclin E is another reported cervical cancermarker. Cyclin E stains both cervical dysplasia and cervical squamouscarcinoma, and its presence is strongly associated with HPV presence(13–15). Tissue culture studies suggest that E7 is responsible for theincrease in cyclin E expression and that this increase is specific tohigh-risk rather than low-risk E7, likely because of the heightenedability to inactivate pRb by high-risk E7 (16, 17). We examined theexpression of both of these markers in our mouse model and foundsimilarities in their timing and expression in comparison to humancancer. These studies provide further validation of this HPV-16 trans-genic mouse model for human cervical cancer.

We then used this mouse model to examine new potential candidatebiomarkers for human cervical cancer. Because our prior studies usingthe HPV transgenic mice indicated E7 to contribute to the progressivecervical disease leading up to the development of cancer, we focusedour search for informative markers of cellular proteins likely to beaffected in their expression by E7. E7, by inactivating pRb and therelated proteins p107 and p130, activates the E2F family of transcrip-tion factors and thereby induces expression of E2F-responsive genes.MCM7, a component of a cellular DNA helicase (reviewed inRef. 18), is an E2F-responsive cellular gene (19, 20). We foundMCM7 expression to be induced during the course of the progressivedisease that arises in the HPV transgenic mice. MCM7 staining wasrestricted to basal and immediate parabasal layers in both normal andhyperplastic tissue. Full thickness staining was observed in high-gradedysplasia and invasive cancers. We found a similar pattern of induc-tion in MCM7 expression in the progressive human disease, withstrong, full-thickness staining for MCM7 seen in the epithelium ofHSILs as well as in frank cancer. The uniform staining pattern andstrong signal for MCM7 in these human lesions suggests that MCM7may be a highly informative biomarker for cervical cancer.

Received 8/6/03; revised 9/19/03; accepted 9/29/03.Grant support: NIH Grants CA22443, CA84227, CA09135, and CA 07175.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: Paul F. Lambert, McArdle Laboratory for Cancer Research,1400 University Avenue, Madison, WI 53706. Phone: (608) 262-8533; Fax: (608) 262-2824; E-mail: lambert@oncology.wisc.edu.

4 The abbreviations used are: HPV, human papillomavirus; HSIL, high-grade squa-mous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion; MCM7,minichromosome maintenance protein 7; CIN, cervical intraepithelial neoplasia; TUNEL,terminal deoxynucleotidyl transferase-mediated nick end labeling.

8173

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

MATERIALS AND METHODS

Histological Specimens. Histological sections were obtained from the re-productive tracts of nontransgenic K14E6 (line 5737; Ref. 4), K14E7 (line2304; Ref. 5), and K14E6/K14E7 double transgenic female mice that had orhad not been treated chronically with 17�-estradiol (0.05 mg/60-day period)for 6 months. The treatment of these mice and the histopathological charac-terization of these tissues have been described previously (11). All mice wereof the FVB/N inbred genetic background. Histological sections from archivalparaffin-embedded patient biopsies with no personal identifiers were obtainedfrom the University of Wisconsin Hospital and Clinics.

Immunohistochemistry. Paraffin-embedded sections were deparaffinizedin two changes of xylene, hydrated though a graded series of ethanol:watermixes, rinsed in PBS, and quenched for endogenous peroxidase activity bytreatment with 3% H2O2 in methanol for 20 min. After washing with PBS,sections were microwaved in antigen retrieval buffer (10 mM sodium citrate,pH 6.0) for 20 min and allowed to cool at room temperature for an additional30 min. Sections were rinsed in PBS and blocked in 1.5–5% horse serum inPBS for 30 min at room temperature. Primary antibodies (Ki-67, Dako,Carpinteria, CA, 1:25; Cyclin E, Santa Cruz Biotechnology clone M-20, SantaCruz, CA, 1:400; Mcm7, Neomarkers, Fremont, CA, 1:200; and p16, Neo-markers clone JC8, Fremont, CA, 1:200) were diluted in blocking solutionbefore being applied to the sections and incubated for at least 3 h at roomtemperature in a humidified chamber. Sections were washed in PBS, followedby a 30-min incubation with Vectastain Universal Elite secondary antibody(1:200) or biotinylated donkey anti-rat IgG (PharMingen, San Diego, CA,1:100). After washing in PBS, the biotin signal was amplified using theVectastain Universal Elite substrate and developed using 3,3�-diaminobenzi-dine substrate, according to the manufacturer’s instructions (Vector, Burl-ingame, CA). Sections were counterstained with hematoxylin (Gill’s formu-lation; Vector, Burlington, CA), dehydrated through ethanols and cleared inxylene, then mounted using Cytoseal XYL reagent (VWR, West Chester, PA).

Sections from raft cultures of the immortalized human foreskin keratinocytecell line (NIKS, also known as BC1-Ep/SL) harboring no viral DNA, wild-typeHPV-16 genomes, or E7 null HPV-16 genomes (also referred to as E7TTLmutant genomes), which were generated previously (21), were stained forMCM7 as described above, and four �20 microscopic fields from two or moreindependent rafts were quantified for the frequency of MCM7-positive cells.

For detection of apoptosis, sections of cervical tumors were stained with theApoptag Plus Fluorescein In Situ Detection Kit (Intergen #S7111, Purchase,NY) according to the manufacturer’s instructions, after being deparaffinizedand rehydrated as described above. After extensive washes in PBS, sections

were mounted in Vectashield mounting medium containing propidium iodidecounterstain (Vector).

RESULTS

Expression of Established Biomarkers for Human CervicalCancer in the Mouse Model for HPV-associated Cervical Cancer.A number of cellular factors are induced in human cervical cancersand their precursor lesions, and several are now being tested or usedas biomarkers for diagnosing and/or staging cervical cancers and theirprecursor lesions in the clinic, with varying degrees of success.Histopathologically, the progressive disease that arises in the cervixesof HPV-16 transgenic mice (K14E6 and K14E7), when treated withestrogen, compares favorably to that seen in patients who developHPV-associated human cervical cancer (11). To evaluate further therelevance of our mouse model for HPV-associated cervical cancer, weinvestigated the pattern of expression of several known biomarkers forhuman cervical cancer. These included Ki-67 and cyclin E.

Ki-67, a marker for proliferating cells (22) and a reported biomar-ker for human cervical cancer (23, 24), was restricted in its expressionto the basal or parabasal compartment of the hyperproliferative strat-ified squamous epithelium in the cervixes of estrogen-treated, non-transgenic mice (Fig. 1A). A similar pattern of staining was seen inlow-grade dysplasia (LSIL) that arose in the estrogen-treated K14E6mice (Fig. 1B). This pattern of expression was indistinguishable fromthat seen in untreated nontransgenic and K14E6 transgenic mice,which also showed staining limited to the basal and parabasal strata(data not shown). These results indicate that hyperplasia and LSIL, thetwo early stages in progressive cervical disease in our mouse model,do not lead to any abnormalities in the distribution of proliferatingcells within the epithelium. In contrast, Ki-67 expression was inducedin the HSIL-grade neoplastic stratified squamous epithelia of theestrogen-treated K14E7 mice with many Ki-67-positive cells ob-served throughout the supraparabasal compartment of the tissue (Fig.1C). In the large frank cancers of the cervix that arise in the estrogen-treated K14E6/K14E7 double transgenic mice, some Ki-67-positivecells were again found throughout the lesions (Fig. 1E). Biomarkerstaining could not be evaluated in low-grade lesions in either the

Fig. 1. Ki-67 expression in the mouse cervix. Shown are histological cross-sections of the mouse cervical tract from nontransgenic and HPV-transgenic mice treated with estrogenfor 6 months. Sections were stained with Ki-67 monoclonal antibody and counterstained with hematoxylin. Ki-67-positive cells show brown staining in the nucleus. Note that detectionof Ki-67 is restricted to the basal and parabasal layers of the cervical squamous epithelium in both the hyperplasia (A) and LSIL (B), whereas scattered staining is evident throughoutmultiple layers of cells in the HSIL (C). Ki-67 is also detected in a subset of cells throughout the frank cancers (E). D, a H&E stain of the tissue shown in E, included for orientation.

8174

MCM7 AS AN INFORMATIVE BIOMARKER FOR HUMAN CERVICAL CANCER

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

K14E7 or K14E6/K14E7 because of the presence of highly dysplasticepithelium throughout the entire cervical squamous epithelium. Inter-estingly, the Ki-67 staining was not uniform in the HSILs and cervicalcancer lesions in the mouse model. This non-uniform pattern of Ki-67staining is similar to that observed in human cervical cancer where thepercentage of cells staining positively for Ki-67, reported to rangefrom 41 to 82% in frank cancer and 41 to 87% in HSILs (23), reflectsa potential limitation in the usefulness of Ki-67 as a biomarker in theclinic.

Cyclin E, an E2F-regulated cell cycle gene (25, 26) and anotherreported biomarker in human cervical cancer (14, 15, 27, 28), was notdetected in the hyperplastic stratified squamous epithelia of estrogen-treated nontransgenic mice (Fig. 2A) nor in the LSILs in estrogen-treated K14E6 transgenic mice (Fig. 2B). In contrast, some cyclin Estaining was seen in all portions of the HSILs in the estrogen-treatedK14E7 mice (Fig. 2C) as well as in the larger frank cancers of thecervix found in the estrogen-treated, K14E6/K14E7 doubly transgenicmice (Fig. 2D). As with Ki-67, not every cell within a lesion of thetreated K14E7 or K14E6/K14E7 mice stained positively for cyclin E;rather the staining was intermittent. This is similar to what has beenseen in human cervical cancers and high-grade dysplastic precursorlesions (11), again reflecting a potential limitation in the usefulness ofthis biomarker. One difference was seen between the staining patternfor cyclin E in LSILs from our mouse model and that of humansamples. In the LSILs arising in the estrogen-treated K14E6 mice, nocyclin E staining was observed. In contrast, cyclin E staining is seenin LSIL samples from women. We believe this difference reflects theabsence of E7 expression in the K14E6 mice. Cyclin E is an E2F-responsive gene (25) that is likely deregulated in its expression byE7’s inactivation of pRb and perhaps the other pocket proteins p107and p130 (26).

On the basis of the data summarized in Figs. 1 and 2, the progres-sive disease in the mouse model for HPV-associated cancers showsclose similarity in the pattern of expression of two established bio-markers of human cervical cancer, Ki-67 and cyclin E, to that seen inthe progressive human disease. These results further validate therelevance of this mouse model for HPV-associated cervical cancer.

Identification of a Novel Marker for Progressive Disease in theMouse Model for HPV-associated Cervical Cancer. A number ofbiomarkers for human cervical cancer have been identified, includingthe ones tested above (Ki-67 and cyclin E) in our mouse model as wellas p16, which could not be evaluated in mice because of a lack ofsuitable antibodies.5 In most cases (e.g., cyclin E and Ki-67), the lackof strong uniform staining within lesions and/or inconsistencies inpositive readouts for the biomarker among different cervical cancershave detracted from their utility in clinical diagnostic screening.Additionally, some markers such as p16, although quite specific andhighly penetrant for HSILs and frank cancer in the stratified squamousepithelium of the human cervix (13, 29, 30), also stain positively innearby normal glandular epithelium, potentially leading to false pos-itives in clinical screening (31). For these reasons, we were interestedin identifying additional biomarkers for cervical cancer using ourmouse model.

The E7 protein is known to bind and inactivate pRb and the relatedp107 and 130 proteins, all of which normally regulate the activity ofthe E2F family of transcription factors. Therefore, E2F-responsivegenes should be up-regulated in the presence of HPV-16 E7. Cyclin Eis one such gene; however, there are many others that could be moresensitive indicators of E2F deregulation. Using a tissue culture systemfor studying the HPV-16 life cycle, we tested this premise. Human

foreskin keratinocytes harboring wild-type HPV-16 genomes or anE7-null HPV-16 genome were grown in organotypic (raft) culture togenerate stratified squamous epithelium in which the HPV-16 lifecycle can be recapitulated and E7 plays a critical role (21, 32).Histological sections of these raft cultures were stained for MCM7, anE2F-responsive gene (19, 33) that is induced at the RNA level in

5 Our unpublished observations and Ron DePinho and Ned Sharpless, personal com-munication.

Fig. 2. Cyclin E expression in the mouse cervix. Shown are histological cross-sectionsof the mouse cervical tract from nontransgenic and HPV-transgenic mice treated withestrogen for 6 months. Sections were stained with cyclin E polyclonal antibody andcounterstained with hematoxylin. Cyclin E-positive cells show brown staining in thenucleus. Note that cyclin E staining is absent in both hyperplastic lesions (A) and LSILs(B), whereas staining is evident throughout HSILs (C) and frank cancers (D). Cyclin Ealso stained the underlying stroma, serving as an internal positive control, and gavebackground staining of infiltrating immune cells, both of which have been observed inhuman cervical specimens (13, 27).

8175

MCM7 AS AN INFORMATIVE BIOMARKER FOR HUMAN CERVICAL CANCER

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

HPV-positive cells based on microarray analyses. MCM7 was foundto be highly induced in the raft culture harboring the wild-typeHPV-16 genome (Fig. 3B) compared with the raft culture harboringno viral DNA (Fig. 3A), with the induction evident in both the basaland suprabasal compartments of the epithelium. This induction, par-ticularly in the suprabasal compartment, was nearly completely absentin the raft culture harboring the E7-null mutant HPV-16 genome (Fig.3C). We demonstrated previously that E7 is responsible for repro-gramming suprabasal cells to re-enter the cell cycle and support DNA

synthesis, allowing for the vegetative amplification of the viral ge-nome (21). The current result demonstrates that MCM7 is induced inthe natural host cell in an E7-dependent manner that correlates withE7’s ability to cause cells to re-enter the cell cycle.

Given that MCM7 is a marker for E7’s induction of DNA synthesisin the viral life cycle, we evaluated whether it also is a marker forHPV-associated cancer in the mouse cervix. In the hyperplastic cer-vical epithelium of estrogen-treated nontransgenic mice (Fig. 4A) andthe LSILs in estrogen-treated K14E6 transgenic mice (Fig. 4B),

Fig. 3. MCM7 expression in the viral life cycle. Shown are histological cross-sections of the epithelium present in organotypic (raft) cultures generated with a human keratinocytecell line harboring no HPV (A), wild-type HPV-16 genomes (B), or E7-null HPV-16 genomes (C). These sections were stained with MCM7-specific antibody (staining nuclei brown)and counterstained with hematoxylin. The poorly differentiated basal compartment is located at the bottom of each panel, with terminal differentiation occurring in the top layers ofthe stratified epithelium. Note the frequent (49%) MCM7-positive cells in the suprabasal compartment in the wild-type HPV-16-positive raft (B). That frequency is greatly reduced(to 7%) in the raft harboring the E7-null HPV-16 genome (C).

Fig. 4. MCM7 immunohistochemistry of mouse cervical sections. Shown are histological cross-sections of the mouse cervical tract from nontransgenic and HPV-transgenic micetreated with estrogen for 6 months. Sections were stained with MCM7 antibody and counterstained with hematoxylin. Positive cells show brown nuclear staining. Note that MCM7staining is restricted to the basal and parabasal layers of the cervical squamous epithelium in both the hyperplastic lesions of nontransgenic mice (A) and LSILs arising in K14E6 mice(B), whereas staining is uniformly and strongly positive throughout the full thickness of HSILs arising in K14E7 mice (C), and microinvasive cancers (D), as well as the large, frankcancers (E) arising in the K14E6/K14E7 mice.

8176

MCM7 AS AN INFORMATIVE BIOMARKER FOR HUMAN CERVICAL CANCER

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

MCM7 staining was restricted to the basal and parabasal layers, aswas seen in these mouse strains without estrogen treatment (data notshown). In the estrogen-treated K14E7 and K14E6/K14E7 mice,MCM7 staining was uniformly and strongly positive throughout thefull thickness of the high-grade dysplasias (HSILs; Fig. 4C), micro-invasive cancers (Fig. 4D) ,as well as the large, frank cancers (Fig.4E). This expansion in MCM7 staining correlates with the expansionin the proliferative compartment within these tissues, as reflected inthe Ki-67 immunohistochemistry (Fig. 1) as well as by bromode-oxyuridine incorporation (data not shown). The MCM7 staining pat-tern, however, was much stronger and more uniform than that seenwith either Ki-67 or cyclin E. Thus, in the mouse model, MCM7proves to be specifically and uniformly induced in high-grade dys-plasia and frank cancer of the cervix. As such, MCM7 represents apotentially valuable candidate biomarker not only for detecting frankcancer but also for distinguishing HSILs from LSILs.

MCM7 Is a Highly Specific Biomarker for High-Grade Lesionsand Frank Cancer in the Human Cervix. To investigate the poten-tial utility of MCM7 as a biomarker in human cervical disease, westained archival histological sections of cervixes from patients withprogressive cervical disease as well as normal tissue. The pattern ofMCM7 staining in normal cervixes (Fig. 5A) and in LSILs (Fig. 5B)was similar, with positive cells limited to the basal and parabasallayers of the normal stratified squamous epithelium and expandedwithin the lower third of the epithelium in LSILs. In contrast, inHSILs (Fig. 5C) and in frank cancer (Fig. 5D), �90% of the squa-mous cells within the lesion stained positively for MCM7. The onlyexception to their being uniformly strong MCM7 staining in cancerswere in regions of cancers with high concentrations of necrotic/apoptotic cells (Fig. 6) or inflammatory infiltrates (as judged by CD45staining; data not shown).

To evaluate the potential value of MCM7 as a biomarker forcervical disease in humans, we evaluated the staining pattern inarchival, paraffin-embedded tissue from multiple patients with normalcervical epithelium or progressive cervical disease (Table 1). Wefound MCM7 to be induced in all frank cancers and HSILs andun-induced in all normal and LSIL tissues. Within the HSILs, wefound the expansion of MCM7 staining to correlate closely with CINII versus CIN III subgrades of disease. The two HSILs limited to CINII grade disease had an incomplete expansion (two-thirds of thethickness in the staining pattern for MCM7). All four HSILs with CINIII grade disease characteristics had uniform and complete expansionof MCM7 staining pattern to the full thickness of the stratifiedsquamous epithelium, with one exception, a lesion that showed full-thickness staining in some portions and two-thirds thickness in otherportions. These results indicate that MCM7 is a highly informativebiomarker for human cervical cancer and the progressive disease thatleads to the development of cervical cancer.

Comparison of MCM7 Staining Patterns in Human CervicalDisease to That of p16. p16 is currently considered an informativebiomarker for human cervical disease because it is induced inhigh-grade dysplasias, which are likely to give rise to cancers aswell as to be contained in the cancers themselves (29, 34). Giventhat we likewise see a similar strong and uniform induction ofMCM7 staining in the same subset of lesions, we directly com-pared the MCM7 and p16 staining patterns in the same humanclinical samples. A very similar pattern of stain for MCM7 and p16was found, with strong and uniform induction in the HSILs andinvasive cancers with both markers (Fig. 5). The only differencewas that MCM7 is restricted in its staining to the nucleus, whereasp16 staining is seen in both the nucleus and cytoplasm. Also, p16did not uniformly stain positively the basal/parabasal cells of thenormal cervix (Fig. 5E) or LSILs (Fig. 5F).

DISCUSSION

In this study, we evaluated the expression patterns of known bi-omarkers of human cervical cancers as a means of further validatingour animal model for HPV-associated cervical cancer. The similaritiesin the Ki-67 and cyclin E staining patterns in the progressive cervicaldisease in our mouse model compared to that in humans verifies thatour model is targeting molecules of physiological relevance to thedevelopment of cervical cancer in women. This biomarker analysis,together with the prior knowledge that the histopathology of theprogressive disease arising in our mouse model closely compares withthat in women (11), provides strong validation of this animal model.

In our animal model, as seen in human cancers, the pattern ofinduction of Ki-67 and cyclin E was not robust. Only a fraction ofcells stained positively for these markers even in frank cancer. Theselimitations raise concern over the utility of Ki-67 and cyclin E asinformative biomarkers in diagnosing cervical cancer. It is also inter-esting to note that our analyses of the E6 versus E7 transgenic miceallowed us to correlate induction of expression of Ki-67 and cyclin Especifically with E7 expression. This correlation, at least in the case ofKi-67, may reflect the fact that in our hands, the disease arising in theE6-only mice is limited to LSILs (CIN I). It also could reflect anenhanced ability of E7 to induce these biomarkers. Consistent withthis, E7 can induce cyclin E in differentiating cultures of humankeratinocytes (16). Cyclin E was likewise induced to some degree inthe cervixes of some E7 transgenic mice, even in the absence ofestrogen treatment (data not shown).

Using our animal model we identified MCM7 to be a novel andhighly informative biomarker for human cervical cancer. Our premisefor studying MCM7 arose from the analysis of the role of E7 in theHPV life cycle. We had described previously a role of E7 in repro-gramming terminally differentiating stratified squamous epithelia tosupport DNA synthesis and specifically the vegetative amplificationof the viral DNA genome (21). We have hypothesized that this abilityof E7 to reprogram cells to support DNA synthesis is reflective of itsinactivation of pRb, the cellular tumor suppressor that modulates theactivity of the E2F family of transcription factors. We thereforepredicted that E2F-responsive genes would be up-regulated by E7,and this up-regulation would correlate with the induction of DNAsynthesis in the suprabasal compartment of HPV-infected epitheliawhen grown in organotypic culture. This proved the case as wedemonstrated in Fig. 3. Indeed, the induction of MCM7 was greatestin the suprabasal compartment in which E7 reprograms cells tosupport DNA synthesis, the same cellular compartment in which E7�sown expression is heightened during the life cycle (35). Clearly, theinduction is dependent upon E7, given the absence of induction incells harboring an E7-null genome. These data provide further supportfor the role of E7’s inactivation of pRb in mediating its role in theviral life cycle and provided the impetus for us to monitor MCM7expression in cervical cancer.

In the progressive cervical disease in our mouse model, MCM7 washighly induced in HSILs and frank cancers and arose in mice onlyexpressing E7, consistent with it being a readout for E7 function. Inhuman cervical disease, a very similar pattern of expression of MCM7was seen. In contrast to Ki-67 and cyclin E, MCM7 was uniformlyand highly induced throughout more aggressive HSILs (CIN III) andfrank cancer in the human lesions (Table 1), similar to that seen withp16. In the less aggressive HSILs (CIN II), we noted an intermediatestaining pattern in which the bottom two-thirds of the cervical epi-thelium uniformly stained positively for MCM7, compared with thefull-thickness staining in the more aggressive HSILs (CIN III). Thus,MCM7 provides one the ability to discriminate between the severityof HSIL disease. In addition, we found 100% of the human HSILs and

8177

MCM7 AS AN INFORMATIVE BIOMARKER FOR HUMAN CERVICAL CANCER

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

frank cancers to stain robustly for MCM7, demonstrating this biomar-ker to be highly informative. MCM7 was also detectable withoutunmasking steps in CaSki cells prepared in a mock ThinPrep PAPsmear sample (data not shown), indicating that surveillance for thismarker is compatible with current clinical sampling for early detectionof cervical disease.

The value of MCM7 in discriminating between stages of the pro-gressive cervical disease in women compares well with other biomar-kers. Interestingly, two other MCMs, MCM2 and MCM5, have alsobeen reported by Coleman and colleagues (36, 37) to be informative

biomarkers for distinguishing different stages of cervical disease inwomen. As with our study, MCM2 and MCM5 were highly inducedin HSILs, more so than in LSILs. No description of potential differ-ences between CIN II and CIN III lesions was reported for theseMCMs. These investigators also mentioned that MCM7 labels cervi-cal cancers (37); however, an evaluation of MCM7 staining pattern inprogressive cervical disease (i.e., LSILs and HSILs) was not reported.Thus, multiple MCMs appear to be highly induced in a manner that isinformative in discriminating between different stages of human cer-vical disease. Perhaps this is not surprising because it is now recog-

Fig. 5. MCM7 immunohistochemistry of human cervical samples and comparison to p16 expression. MCM7 staining was performed on human cervical biopsy specimens of normal(A), LSIL (B), HSIL (C), and invasive carcinoma (D). p16 analysis was performed on serial sections of the same tissue used in MCM7 analysis (E–H). All sections were counterstainedwith hematoxylin. MCM7-positive cells show brown staining confined to the nucleus, whereas p16-positive cells show both nuclear and cytoplasmic brown positivity. Note that MCM7staining is restricted to the basal and immediate parabasal layers of the cervical squamous epithelium in the normal cervix (A) and that the thickness of positive cells increases withincreasing lesion grade, staining the lower third of the epithelium in LSIL (B), two-thirds to full thickness of HSIL (C), and throughout the entire thickness of invasive cancer (D). p16fails to stain cells of the normal cervix (E) or of LSIL (F) but displayed the same staining pattern as MCM7 in the HSIL (G) and invasive cancer lesions (H).

8178

MCM7 AS AN INFORMATIVE BIOMARKER FOR HUMAN CERVICAL CANCER

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

nized that many if not all MCM genes are E2F inducible (19, 33).Whether their up-regulation is reflective of increased expression ofHPV E7 that arises in progressive cervical disease, however, is notclear. High-risk HPVs are commonly found integrated in cervicalcancers, leading to the selective up-regulation of expression of twoviral genes, E6 and E7. These integration events, however, are notoften detected in the precursor disease including HSILs. Therefore, itis reasonable to conclude that other alterations in cell physiology priorto the integration of the HPV genome may be contributing to theinduction of MCM expression. Both E6 and E7 induce genomicinstability (8, 9), potentially through the induction of centrosomeabnormalities (38). Such centrosome abnormalities and genetic insta-bility have been noted in the cervical epithelium of our HPV trans-

genic mice (11) and could contribute to changes in cell physiologythat lead to the induction of MCM7.

Note Added in Proof

In the process of preparing this study for publication, we learned from Dr. JohnDoorbar (Cancer Research UK, London, United Kingdom) that his group per-formed a study (K. Middleton et al., J. Virol. 77: 10186–10201, 2003) in whichthey found a similar MCM7 staining pattern in human cervical cancer as seen inour study (personal communication), providing independent demonstration of thevalue of MCM7 as a potentially useful biomarker. This study was recentlypublished (see Middleton et al., J. Virol. 77: 10186–10201, 2003).

ACKNOWLEDGMENTS

We acknowledge Amy Liem for assistance with mouse husbandry, estrogentreatments, and necropsies; Elsa Flores for generation of raft cultures; Anh Dofor MCM7 immunohistochemistry on the raft cultures; and Harlene Edwardsand Jane Weeks of the University of Wisconsin Comprehensive Cancer CenterHistology Core Facility for expert histological preparation of all tissue samplesanalyzed. We are grateful to Josephine Harter for assistance in histopatholog-ical diagnoses.

REFERENCES

1. Walboomers, J. M., Jacobs, M. V., Manos, M. M., Bosch, F. X., Kummer, J. A., Shah,K. V., Snijders, P. J., Peto, J., Meijer, C. J., and Munoz, N. Human papillomavirus isa necessary cause of invasive cervical cancer worldwide. J. Pathol., 189: 12–19, 1999.

Fig. 6. Detection of apoptosis in areas of cancer negative for MCM7.Shown is histological staining of a human invasive squamous cancer. Ashows MCM7 immunohistochemistry, showing brown positive staining ofsquamous cells with a non-staining pocket of tissue in the middle of thelesion. B shows TUNEL staining of a serial section of the same tissue shownin A. TUNEL-positive cells were labeled with fluoroscein and appear yellowwhen layered over the red propidium iodide counterstain. Note that the areaof the lesion which does not stain positive for MCM7 in A is TUNEL-positivein B.

Table 1 Pattern of MCM7 staining in human cervical disease

Grade of diseaseNumber of

samples

MCM7 staining patterna

1/3 or less 2/3 Full thickness

No Disease 13 13 (100%)LSIL 7 7 (100%)HSIL 6 2 (33%) 4 (66%)

Invasive Cancer 8 8 (100%)a MCM7 staining was performed as described in “Materials and Methods.” Indicated in

parentheses are the percentages of samples in each grade of disease with 1/3, 2/3, or fullthickness staining for MCM7 in the stratified squamous epithelium of the cervix in normal(No Disease), LSIL, and HSIL samples or in the cancer itself.

8179

MCM7 AS AN INFORMATIVE BIOMARKER FOR HUMAN CERVICAL CANCER

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

2. zur Hausen, H. Papillomaviruses causing cancer: evasion from host-cell control inearly events in oncogenesis. J. Natl. Cancer Inst., 92: 690–698, 2000.

3. McDougall, J. K. Immortalization and transformation of human cells by humanpapillomavirus. Curr. Top. Microbiol. Immunol., 186: 101–119, 1994.

4. Song, S., Pitot, H. C., and Lambert, P. F. The human papillomavirus type 16 E6 genealone is sufficient to induce carcinomas in transgenic animals. J. Virol., 73: 5887–5893, 1999.

5. Herber, R., Liem, A., Pitot, H., and Lambert, P. F. Squamous epithelial hyperplasiaand carcinoma in mice transgenic for the human papillomavirus type 16 E7 oncogene.J. Virol., 70: 1873–1881, 1996.

6. Demers, G. W., Halbert, C. L., and Galloway, D. A. Elevated wild-type p53 proteinlevels in human epithelial cell lines immortalized by the human papillomavirus type16 E7 gene. Virology, 198: 169–174, 1994.

7. Song, S., Gulliver, G. A., and Lambert, P. F. Human papillomavirus type 16 E6 andE7 oncogenes abrogate radiation-induced DNA damage responses in vivo throughp53-dependent and p53-independent pathways. Proc. Natl. Acad. Sci. USA, 95:2290–2295, 1998.

8. White, A. E., Livanos, E. M., and Tlsty, T. D. Differential disruption of genomicintegrity and cell cycle regulation in normal human fibroblasts by the HPV onco-proteins. Genes Dev., 8: 666–677, 1994.

9. Reznikoff, C. A., Belair, C., Savelieva, E., Zhai, Y., Pfeifer, K., Yeager, T., Thomp-son, K. J., DeVries, S., Bindley, C., Newton, M. A., et al. Long-term genome stabilityand minimal genotypic and phenotypic alterations in HPV16 E7-, but not E6-,immortalized human uroepithelial cells. Genes Dev., 8: 2227–2240, 1994.

10. Song, S., Liem, A., Miller, J. A., and Lambert, P. F. Human papillomavirus types 16E6 and E7 contribute differently to carcinogenesis. Virology, 267: 141–150, 2000.

11. Riley, R. R., Duensing, S., Brake, T., Munger, K., Lambert, P. F., and Arbeit, J. M.Dissection of human papillomavirus E6 and E7 function in transgenic mouse modelsof cervical carcinogenesis. Cancer Res., 63: 4862–4871, 2003.

12. Keating, J. T., Ince, T., and Crum, C. P. Surrogate biomarkers of HPV infection incervical neoplasia screening and diagnosis. Adv. Anat. Pathol., 8: 83–92, 2001.

13. Keating, J. T., Cviko, A., Riethdorf, S., Riethdorf, L., Quade, B. J., Sun, D.,Duensing, S., Sheets, E. E., Munger, K., and Crum, C. P. Ki-67, cyclin E, andp16INK4 are complementary surrogate biomarkers for human papillomavirus-relatedcervical neoplasia. Am. J. Surg. Pathol., 25: 884–891, 2001.

14. Cho, N. H., Kim, Y. T., and Kim, J. W. Correlation between G1 cyclins and HPV inthe uterine cervix. Int. J. Gynecol. Pathol., 16: 339–347, 1997.

15. Kim, Y. T., Choi, E. K., Cho, N. H., Ko, J. H., Yang, W. I., Kim, J. W., and Lee, S. H.Expression of cyclin E and p27KIP1 in cervical carcinoma. Cancer Lett., 153: 41–50,2000.

16. Martin, L. G., Demers, G. W., and Galloway, D. A. Disruption of the G1/S transitionin human papillomavirus type 16 E7-expressing human cells is associated with alteredregulation of cyclin E. J. Virol., 72: 975–985, 1998.

17. Zerfass, K., Schulze, A., Spitkovsky, D., Friedman, V., Henglein, B., and Jansen-Durr, P. Sequential activation of cyclin E and cyclin A gene expression by humanpapillomavirus type 16 E7 through sequences necessary for transformation. J. Virol.,69: 6389–6399, 1995.

18. Blow, J. J., and Hodgson, B. Replication licensing—defining the proliferative state?Trends Cell Biol., 12: 72–78, 2002.

19. Suzuki, S., Adachi, A., Hiraiwa, A., Ohashi, M., Ishibashi, M., and Kiyono, T.Cloning and characterization of human MCM7 promoter. Gene (Amst.), 216: 85–91,1998.

20. Ishida, S., Huang, E., Zuzan, H., Spang, R., Leone, G., West, M., and Nevins, J. R.Role for E2F in control of both DNA replication and mitotic functions as revealedfrom DNA microarray analysis. Mol. Cell. Biol., 21: 4684–4699, 2001.

21. Flores, E. R., Allen-Hoffmann, B. L., Lee, D., and Lambert, P. F. The humanpapillomavirus type 16 E7 oncogene is required for the productive stage of the virallife cycle. J. Virol., 74: 6622–6631, 2000.

22. Gerdes, J., Lemke, H., Baisch, H., Wacker, H. H., Schwab, U., and Stein, H. Cellcycle analysis of a cell proliferation associated human nuclear antigen defined by themonoclonal antibody Ki-67. J. Immunol., 133: 1710–1715, 1984.

23. Isacson, C., Kessis, T. D., Hedrick, L., and Cho, K. R. Both cell proliferation andapoptosis increase with lesion grade in cervical neoplasia but do not correlate withhuman papillomavirus type. Cancer Res., 56: 669–674, 1996.

24. Resnick, M., Lester, S., Tate, J. E., Sheets, E. E., Sparks, C., and Crum, C. P. Viraland histopathologic correlates of MN and MIB-1 expression in cervical intraepithelialneoplasia [see comments]. Hum. Pathol., 27: 234–239, 1996.

25. Botz, J., Zerfass-Thome, K., Spitkovsky, D., Delius, H., Vogt, B., Eilers, M.,Hatzigeorgiou, A., and Jansen-Durr, P. Cell cycle regulation of the murine cyclin Egene depends on an E2F binding site in the promoter. Mol. Cell. Biol., 16: 3401–3409, 1996.

26. Geng, Y., Eaton, E. N., Picon, M., Roberts, J. M., Lundberg, A. S., Gifford, A.,Sardet, C., and Weinberg, R. A. Regulation of cyclin E transcription by E2Fs andretinoblastoma protein. Oncogene, 12: 1173–1180, 1996.

27. Quade, B. J., Park, J. J., Crum, C. P., Sun, D., and Dutta, A. In vivo cyclin Eexpression as a marker for early cervical neoplasia. Mod. Pathol., 11: 1238–1246,1998.

28. Dellas, A., Schultheiss, E., Leivas, M. R., Moch, H., and Torhorst, J. Association ofp27Kip1, cyclin E and c-myc expression with progression and prognosis in HPV-positive cervical neoplasms. Anticancer Res., 18: 3991–3998, 1998.

29. Klaes, R., Friedrich, T., Spitkovsky, D., Ridder, R., Rudy, W., Petry, U., Dallenbach-Hellweg, G., Schmidt, D., and von Knebel Doeberitz, M. Overexpression ofp16(INK4A) as a specific marker for dysplastic and neoplastic epithelial cells of thecervix uteri. Int. J. Cancer, 92: 276–284, 2001.

30. Sano, T., Oyama, T., Kashiwabara, K., Fukuda, T., and Nakajima, T. Expressionstatus of p16 protein is associated with human papillomavirus oncogenic potential incervical and genital lesions. Am. J. Pathol., 153: 1741–1748, 1998.

31. Riethdorf, L., Riethdorf, S., Lee, K. R., Cviko, A., Loning, T., and Crum, C. P.Human papillomaviruses, expression of p16, and early endocervical glandular neo-plasia. Hum. Pathol., 33: 899–904, 2002.

32. Flores, E. R., Allen-Hoffmann, B. L., Lee, D., Sattler, C. A., and Lambert, P. F.Establishment of the human papillomavirus type 16 (HPV-16) life cycle in animmortalized human foreskin keratinocyte cell line. Virology, 262: 344–354, 1999.

33. Ohtani, K., Iwanaga, R., Nakamura, M., Ikeda, M., Yabuta, N., Tsuruga, H., andNojima, H. Cell growth-regulated expression of mammalian MCM5 and MCM6 genesmediated by the transcription factor E2F. Oncogene, 18: 2299–2309, 1999.

34. Sano, T., Masuda, N., Oyama, T., and Nakajima, T. Overexpression of p16 andp14ARF is associated with human papillomavirus infection in cervical squamous cellcarcinoma and dysplasia. Pathol. Int., 52: 375–383, 2002.

35. Cheng, S., Schmidt-Grimminger, D. C., Murant, T., Broker, T. R., and Chow, L. T.Differentiation-dependent up-regulation of the human papillomavirus E7 gene reac-tivates cellular DNA replication in suprabasal differentiated keratinocytes. GenesDev., 9: 2335–2349, 1995.

36. Williams, G. H., Romanowski, P., Morris, L., Madine, M., Mills, A. D., Stoeber, K.,Marr, J., Laskey, R. A., and Coleman, N. Improved cervical smear assessment usingantibodies against proteins that regulate DNA replication. Proc. Natl. Sci. USA, 95:14932–14937, 1998.

37. Freeman, A., Morris, L. A., Mills, A. D., Stoeber, K., Laskey, R. A., Williams, G. H.,and Coleman, N. Minichromosome maintenance proteins as biological markers ofdysplasia and malignancy. Clin. Cancer Res., 5: 2121–2132, 1999.

38. Duensing, S., Lee, L. Y., Duensing, A., Basile, J., Piboonniyom, S., Gonzalez, S.,Crum, C. P., and Munger, K. The human papillomavirus type 16 E6 and E7oncoproteins cooperate to induce mitotic defects and genomic instability by uncou-pling centrosome duplication from the cell division cycle. Proc. Natl. Acad. Sci. USA,97: 10002–10007, 2000.

8180

MCM7 AS AN INFORMATIVE BIOMARKER FOR HUMAN CERVICAL CANCER

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

2003;63:8173-8180. Cancer Res   Tiffany Brake, Joseph P. Connor, Daniel G. Petereit, et al.   an Informative Biomarker for Human Cervical CancerIdentification of Minichromosome Maintenance Protein 7 asHuman Papillomavirus-Transgenic Mouse Model: Comparative Analysis of Cervical Cancer in Women and in a

  Updated version

  http://cancerres.aacrjournals.org/content/63/23/8173

Access the most recent version of this article at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/63/23/8173.full#ref-list-1

This article cites 38 articles, 17 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/63/23/8173.full#related-urls

This article has been cited by 25 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  SubscriptionsReprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://cancerres.aacrjournals.org/content/63/23/8173To request permission to re-use all or part of this article, use this link

Research. on May 30, 2020. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from