Silencing of E7 oncogene restores functional E

Carcinogenesis vol.29 no.7 pp.1441–1447, 2008 doi:10.1093/carcin/bgn145 Advance Access publication June 19, 2008

Silencing of E7 oncogene restores functional E-cadherin expression in human papillomavirus 16-transformed keratinocytes Jean-Hubert D.Caberg,y, Pascale M.Huberty, Dominique Y.Begon, Michael F.Herfs, Patrick J.Roncarati, Jacques J.Boniver and Philippe O.Delvenne Department of Pathology, Groupe Interdisciplinaire de Geńoproteómique Appliqueé-Cancer, B23, University of Liege, Centre Hospitalier Universitaire Sart Tilman, 4000 Liege, Belgium To whom correspondence should be addressed. Tel: þ32 43664282; Fax: þ32 43662919; Email: [email protected]

Introduction Chronic infection of the uterine cervix mucosa by human papillomavirus (HPV), most notably HPV16, is associated with the malignant transformation of keratinocytes (KCs) in metaplastic areas of the transformation zone (1). Despite the evidence that HPV is strongly implicated as the causative agent of cervical cancer and its precursors [squamous intra-epithelial lesion (SIL)], HPV infection alone is not sufficient for tumour development (2). It has been suggested that the intrinsic sensitivity of the cervical transformation zone to HPVinduced SIL and cancer development is linked to quantitative and qualitative alterations of dendritic cells (DCs)/Langerhans cells (LCs) (3). There is accumulating evidence that tumour- and virus-infected cells are able to inhibit protective immune reactions. This negative effect might be mediated by a decreased production of soluble molecules, which normally activate the migration and function of DC/LC (4). DC constitutes a network of sentinels considered to be the most important professional antigen-presenting cells in the immune system. LC constitutes a subfamily of DC localized in the suprabasal layers of the epidermis and in mucosal squamous epithelia. After antigen exposure, activated LCs subsequently leave the epithelium and migrate via afferent lymphatics to regional lymph nodes, where they localize in T cell areas as interdigitating DC, able to activate naive T cells to mount immune responses to antigen encountered in the epithelium (5). Abbreviations: AP, activating protein; DC, dendritic cell; HPV, human papillomavirus; IFN, interferon; KC, keratinocyte; LC, Langerhans cell; mRNA, messenger RNA; PCR, polymerase chain reaction; Rb, retinoblastoma; RT, reverse transcription; siC, control siRNA; siE7, E7 siRNA; SIL, squamous intra-epithelial lesion; siRNA, small interfering RNA. y

These authors contributed equally to this work.

Materials and methods Cervical biopsy specimens Fifty formalin-fixed, paraffin-embedded cervical biopsy specimens including 19 normal HPV-negative exocervical tissues and 31 HPV16-positive cervical lesions (1 low-grade SIL, 13 high-grade SILs and 17 squamous cell carcinomas) were retrieved from the archives of the Pathology Department at the University Hospital of Liege. DNA was prepared from all biopsies with the QIAamp DNA Mini kit (Qiagen, Valencia, CA) and tested, by polymerase chain reaction (PCR), for the presence of HPV16 DNA sequences using previously described methods and appropriate controls (20). The protocol was approved by the Liege University Hospital Ethics Committee. Immunohistochemistry Serial sections of cervical biopsy specimens underwent staining using monoclonal antibodies directed against E-cadherin (HECD-1, Zymed Laboratories, San Francisco, California) and CD1a (MTB1, Novocastra, Newcastle, UK). Immunoperoxidase staining was performed with the use of the Vectastain ABC kit (Vector Laboratories, Burlingame, CA) or the Envision kit (Dako, Glostrup, Denmark) according to the supplier’s recommendations. Positive cells were visualized by a 3,3#-diaminobenzidin substrate and the sections were counterstained with haematoxylin. The E-cadherin immunostaining and the density of CD1aþ LC were evaluated, in cervical biopsies, by using a semi-quantitative score and analysed as described previously (8). Briefly, for E-cadherin expression, scoring of the intensity and extent of the staining were performed according to an arbitrary scale. For staining intensity, 0 represented samples in which membrane E-cadherin expression was undetectable, whereas 1þ, 2þ and 3þ denoted samples with, respectively, a low, moderate and strong staining. For staining extent, 1þ represented samples in which E-cadherin expression was detectable in up to 33% of the epithelium, 2þ denoted samples in which 33–66% of the epithelium presented a detectable E-cadherin expression and 3þ represented those in which .66% of the epithelial cells were stained. In order to provide a global score for each case, the results obtained

Ó The Author 2008. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

1441

Downloaded from http://carcin.oxfordjournals.org/ by guest on June 10, 2013

Human papillomavirus (HPV) infection, particularly type 16, is causally associated with cancer of the uterine cervix. The persistence or progression of cervical lesions suggests that viral antigens are not adequately presented to the immune system. This hypothesis is reinforced by the observation that most squamous intraepithelial lesions show quantitative and functional alterations of Langerhans cells (LCs). Moreover, E-cadherin-dependent adhesion of LC to keratinocytes (KCs) is defective in cervical HPV16-associated (pre)neoplastic lesions. The possible role of viral oncoprotein E7 in the reduced levels of cell surface E-cadherin was investigated by silencing HPV16 E7 by RNA interference (siRNA). This treatment induced an increased cell surface Ecadherin expression in HPV16-positive KC and a significant adhesion of LC to these squamous cells. The E-cadherin re-expression following HPV16 E7 silencing was associated with increased detection levels of retinoblastoma protein and the activating protein (AP)-2a transcription factor. These data suggest that HPV16 E7induced alterations of LC/KC adhesion may play a role in the defective immune response during cervical carcinogenesis.

The demonstration that LCs express high levels of E-cadherin and that E-cadherin mediates LC binding to KC in vitro (6) suggests that this adhesion molecule promotes LC retention in epithelial tissues. E-cadherin expression of LC is attenuated as a consequence of cell activation and maturation and the loss of E-cadherin-mediated adhesion of LC to KC is among the first events in the multistep cascade, leading to LC migration from the epithelium to regional lymph nodes. In squamous epithelia, E-cadherin is one of the major cell adhesion molecules defining the tissue architecture and differentiation (7). The cell membrane expression of E-cadherin is low in epithelial cells of SIL and squamous cell carcinoma biopsy specimens as compared with normal exocervical epithelium (8). High-risk types of HPV encode for three oncoproteins, namely E5, E6 and E7 (9). Of these, E6 and E7 have been the more actively investigated. The most well-characterized target of E7 is the retinoblastoma (Rb) tumour suppressor, which plays a major role in the maintenance of the epithelial phenotype. Accordingly, Rb activates the E-cadherin gene by directly binding to the activating protein (AP)-2 transcription factor in vivo and acting synergistically with it (10). Interaction between E7 and Rb disrupts the ability of Rb to repress E2F-responsive genes (11,12) and leads to the proteasomal degradation of Rb (13–15). In the present work, we tested the hypothesis that E7 is involved in the decreased levels of cell surface Ecadherin in HPV-transformed KC via the Rb–AP-2 pathway, thereby affecting the number of LC resident in the infected tissue. We also investigated whether HPV16 E7 regulates the activity of Slug or Snail transcription factors described previously as strong repressors of E-cadherin expression (16–19). By using cell lines expressing HPV16 E7 and a heterotypic adhesion assay between HPV16-positive KC and LC, we finally demonstrated that silencing of E7 oncogene by RNA interference restores a functional E-cadherin expression in HPV-transformed KC.

J.-H.D.Caberg et al.

with the two scales were multiplied, yielding a single scale with steps of 0–9. A similar scoring system was used for CD1a evaluation, with the modification that the staining intensity was replaced by the density of positive cells [low (1þ), moderate (2þ) and high (3þ)]. Cell cultures ‘SiHa’ is a cervical carcinoma cell line containing one copy of integrated HPV16 DNA. ‘KE7’ is a primary skin KC cell line stably transfected with E7 gene of HPV16 (gift of F.Ro¨sl, Heidelberg) (21). ‘C33A’ is a cervical carcinoma cell line negative for HPV. These cell lines were grown and maintained in a mixture of Ham’s F-12 and Dulbecco’s modified Eagle’s medium (22). Human exocervical epithelial cells (KN) were obtained from women who underwent total hysterectomy for non-cervical benign uterine disease. Cell cultures were established and maintained following a previously reported method (23). Briefly, tissue fragments were incubated with 0.25% trypsin (Gibco BRL, Gaithersburg, MD) for 24 h at 4°C. The epithelial cells were detached and cultured with irradiated 3T3 mouse fibroblasts as a feeder in a one-third mixture of HAM F12 (Gibco BRL)/Dulbecco’s modified Eagle’s medium (Gibco BRL), as described previously (24). LCs were generated by culturing CD34þ cord blood mononuclear cells, with previously optimized concentrations of cytokines and haematopoietic growth factors (24). These study protocols were approved by the Ethics Committee of the University Hospital of Liege.

Reverse transcription–PCR analysis One microgram of total RNA extracted from cell cultures (RNeasy Mini kit, Qiagen) and quantified with a ND-1000 spectrophotometer (NanoDrop, Wilmington, DE) was reverse transcribed using Superscript II reverse transcriptase (Invitrogen B, Lidingo, Sweden) according to the manufacturer’s instructions. The reactions were performed at 42°C for 50 min, followed by inactivation of the enzyme at 75°C for 15 min. The complementary DNA was stored at 20°C. Reverse transcription (RT)–PCRs were performed using the following primer sequences: E-cadherin, forward: 5#-TATTCCTCCCATCAGCTGCCC-3# and reverse: 5#-CAATGCGTTCTCTATCCAGAGG-3#; Snail, forward: 5#-AATCGGAAGCCTAACTACAGCGAG-3# and reverse: 5#-CCTTCCCACTGTCCTCATC TGACA-3#; Slug, forward: 5#-CCTTCCTGGTCAAGAAGCATTTCA-3# and reverse: 5#-AGGCTCACATATTCCTTGTCACAG-3#; AP-2a, forward: 5#AGCTGAATTTCTCAACCGACAAC-3# and reverse: 5#-TAGCCAGGAGCATGTTTTTTCTT-3# and hypoxanthine phosphoribosyl transferase (HPRT), forward: 5#-TTGGATATAAGCCAGACTTTGTTG-3# and reverse: 5#-AGATG TTTCCAAACTCAACTTGAA-3#. Samples were run on 1.8% agarose gels containing ethidium bromide and visualized with an ultraviolet transilluminator. Heterotypic adhesion assay Twenty-four hours after plating 1 105 KC in four-well chamber slides (Lab-Tek, Nunc), the medium was rinsed and 0.5 105 LCs per well were added in 0.5 ml of RPMI. The LCs used in this assay were pre-labelled (in red) with the lipophilic fluorescent marker CM-DiL (Molecular Probes, Leiden, The Netherlands) according to previously described procedures (28). The cultures were placed in a 37°C incubator (5% CO2) for 24 h. Cells were then fixed in 2% paraformaldehyde for 15 min at room temperature and nuclei were revealed with 4#,6-diamidine-2#-phenylindole dihydrochloride staining (Roche, Mannheim, Germany). A mouse anti-E-cadherin monoclonal blocking antibody (67A4 clone, Chemicon International, Temecula, CA; 5 lg/ml final concentration) was added to the medium of some cultures 4 h before adding the LC in order to inhibit the E-cadherin-mediated adhesion of LC to KC. A mouse IgG1 served as negative control (W3/25, Serotec, Oxford, UK; 5 lg/ml final concentration).

1442

Confocal microscopy LCs labelled with the lipophilic fluorescent marker CM-DiL (Molecular Probes) and KC were incubated with a fluorescein isothiocyanate-conjugated mouse antiE-cadherin antibody (Pharmingen, San Diego, CA; 5 lg/ml). Cell nuclei were counterstained (in blue) with Toto3 (Molecular Probes) and observed with a confocal laser scanning microscope (TCS-SP2, Leica Microsystems AG). Flow cytometry analysis For E-cadherin staining, cells were incubated with a mouse anti-E-cadherin antibody (HECD-1, Zymed; dilution of 1:200) for 30 min at 4°C. Cells were washed in phosphate-buffered saline, followed by incubation with phycoerythrin-conjugated anti-mouse IgG in phosphate-buffered saline for 20 min at 4°C. These cells analysed for surface antigens were sorted by a FACSVantage (Becton Dickinson). Negative controls included isotype-matched irrelevant antibodies or direct staining with secondary antibody in the absence of the first antibody.

Results E-cadherin expression and CD1aþ LC density are reduced in cervical HPV16-positive lesions E-cadherin and CD1a expression was studied in cervical HPVnegative normal tissues (n 5 19) and HPV16-positive lesions (n 5 31). The immunostaining results are shown in Figure 1. The normal exocervical epithelium was found to be generally strongly membrane positive for E-cadherin. Labelled cells were mainly observed in (para)basal and intermediate epithelial layers (Figure 1A). In contrast, (pre)neoplastic lesions showed no or a low anti-E-cadherin cytoplasmic immunoreactivity (Figure 1B). In the normal epithelium, CD1aþ cells were intermingled with KC in the (para)basal and intermediate cell layers (Figure 1C). In contrast, CD1aþ cells were rarely observed in cervical lesions (Figure 1D). Semi-quantitative evaluation of E-cadherin and CD1a intra-epithelial expression is shown in Figure 1E. The E-cadherin score was significantly higher in normal exocervical epithelium than in HPV16-positive (pre)neoplastic lesions. The CD1a score was also lower in HPV16-positive lesions than that measured in the normal exocervical epithelium and related with the E-cadherin staining. The E-cadherin and CD1a staining patterns were well correlated within single lesions, as reflected by the Spearman’s rank correlation test (P , 0.001). E7 silencing induces the expression of E-cadherin in HPV16transformed KC To estimate the level of transfection efficiency, each cell line used in this study was transfected with 3# 6-FAM-modified control siRNAs


siRNAs’ oligonucleotides and transfection A mix with an equal ratio of four RNA oligonucleotides corresponding to HPV16 E6 or E7 and designed as described previously (25–27) was purchased from Eurogentec (Eurogentec SA, Seraing, Belgium) (supplementary Table 1, available at Carcinogenesis Online). For the transfection, the cells were trypsinized and platted into six-well plates (Nunc, Naperville, IL) or four-well chamber slides (Lab-Tek, Nunc). After 24 h, the cells were transfected with siRNAs formulated into liposomes (Transfectin, Bio-Rad Laboratories, Nazareth Eke, Belgium) according to the manufacturer’s instructions. The final siRNA concentration was 20 nM. Cells were harvested for analysis at various times thereafter as indicated in the results. Transfection efficiencies were established by transfecting cells as described previously with 3# 6-carboxyfluorescein (FAM)-modified siRNAs (Eurogentec SA). Cells were harvested 24 h following transfection and analysed by flow cytometry using a FACSVantage SE (Becton Dickinson, San Jose, CA) and by confocal laser scanning microscopy (TCS-SP2, Leica Microsystems AG, Wetzlar, Germany).

Western blotting analysis The cell pellets were re-suspended in RIPA lysis buffer (Tris, pH 7.2, 10 mM; NaCl 150 mM; ethylenediaminetetraacetic acid 5 mM; Triton X-100 1%; sodium deoxycholate 1% and sodium dodecyl sulphate 0.1%), subsequently stirred on ice for 20 min and centrifugated at 14 000 r.p.m. for 10 min. The supernatants were analysed for protein content by the Pierce BCA Protein Assay (Pierce, Rockford, UK). Samples were diluted in sodium dodecyl sulphate buffer (Invitrogen AB), heated at 70°C for 10 min, loaded on a NuPage 4–12% Bis–Tris gel and then electrophoresed in a 3-(N-morpholino)propanesulfonic acid buffer (Invitrogen AB). After the electrophoresis, the proteins were transferred to a polyvinyldifluoride membrane (Invitrogen AB) and allowed to bind with the following antibodies: mouse anti-HPV16 E6 (6F4, gift from E.Weiss, Strasbourg; dilution of 1:2000), mouse anti-HPV16 E7 (ED17, Santa Cruz Biotechnology, Santa Cruz, CA; dilution of 1:500), mouse anti-p53 (DO-7, Dako; dilution of 1:200), mouse anti-AP-2a monoclonal antibody (3B5, Santa Cruz Biotechnology; dilution of 1:1000), mouse anti-p16INK4A monoclonal antibody (16P04/JC2, NeoMarkers, Westinghouse, CA; dilution of 1:200), mouse anti-E-cadherin monoclonal antibody (HECD-1, Zymed; dilution of 1:2000), rabbit anti-Rb (C-15, Santa Cruz Biotechnology; dilution of 1:500), goat anti-KU-70 (M19, Santa Cruz Biotechnology; dilution of 1:2000) and rabbit anti-b-actin (C11, Sigma, Bornem, Belgium; dilution of 1:3000). In the next step, we used goat anti-mouse (P0477, Dako; dilution of 1:5000), rabbit anti-goat (P0160, Dako; dilution of 1:5000) and donkey anti-rabbit (NA9340, Amersham Biosciences, Uppsala, Sweden; dilution of 1:5000) secondary antibodies. Visualization of bound antibodies was detected by chemiluminescence (ECL Plus, Amersham Biosciences).

E-cadherin and E7 expression in HPV161 keratinocytes

(siCs) (Eurogentec SA) and analysed by flow cytometry 24 h following transfection. The transfection efficiency was between 70 and 95% and siRNAs were present both in the cytoplasm and in the nucleus of transfected cells (supplementary Figure 1, available at Carcinogenesis Online). As high concentrations of siRNA have been shown to induce unspecific cell effects by activating interferon (IFN)-stimulated genes involved in the IFNa/b pathway (29–31), we analysed by using a quantitative real-time PCR assay the relative messenger RNA (mRNA) levels of STAT1, PKR, Bcl-2 and OAS2 in SiHa cells transfected by E7 siRNA (siE7) in comparison with a siRNA validated to induce a positive IFN response (siIFNþ). siE7 transfection in cells did not influence any of the IFN response genes, demonstrating that the observed cellular effects are related to the specific silencing of targeted viral genes (supplementary Figure 2, available at Carcinogenesis Online). We next monitored the E7 protein levels by immunoblotting after specific siRNA treatment. From 24 h after transfection, we observed an important decrease in E7 expression and this suppressive effect was sustained for several days (Figure 2A). In addition, siE7 affected the growth of KE7 and SiHa cells for at least 3 days after transfection and induced the down-regulation of cell cycle regulator p16INK4A (Figure 2B–D). This is in agreement with

previous data showing that HPV16 E7 overcomes G1 cell cycle arrest imposed by over-expression of the cyclin dependent kinase inhibitor p16INK4A (32). We also monitored E6 and cellular p53 protein levels after E7 silencing and observed only a minor effect of the anti-siE7 on these proteins (Figure 2 and supplementary Figure 3, available at Carcinogenesis Online), suggesting that E6 mRNA can also be slightly affected by siE7. We next analysed protein level and localization of E-cadherin in KE7 and SiHa cells treated with siE7s by using western blotting and confocal microscopy. Treatment of KE7 and SiHa cells with siE7 resulted in the increased expression of E-cadherin as detected by immunoblotting (Figure 3A). The cell growth retardation or arrest of siE7-treated KE7 or SiHa cells was associated with the presence of irregularly distributed cohesive clusters of cells showing E-cadherin membrane staining (Figure 3I, J, O and P), whereas cells treated with siC appeared more evenly distributed (without cohesive cell clusters) and did not show E-cadherin membrane staining (Figure 3F, G, L and M). Similar data were obtained by flow cytometry (data not shown). A strong membrane E-cadherin expression was also found in HPV-negative exocervical epithelial cells (Figure 3C and D). No modulation of E-cadherin immunoreactivity was observed

1443


Fig. 1. (A–E) E-cadherin and CD1a immunostaining in cervical biopsy specimens. E-cadherin expression (A) and density of CD1aþ cells (B) in normal squamous epithelium; high-grade SIL demonstrating a lack of anti-E-cadherin immunoreactivity (C) and a low density of CD1aþ cells (D). Inserts are higher magnifications of E-cadherin staining (A and C) or CD1aþ cell density (B and D). Semi-quantitative evaluation of E-cadherin expression and CD1aþ LC density in cervical biopsy specimens. Asterisks indicate statistically significant differences ( P , 0.01, P , 0.001) (E).


when C33A or normal KCs were treated with E7 or siC (data not shown). In order to analyse the impact of the confluence level of transfected cells on the re-expression of E-cadherin, we performed total protein extraction at 24, 72 and 120 h after transfection of HPV16 KCs with siE7 or siC. These different time points corresponded to different levels of confluence (24 h, siC 5 40%/siE7 5 25%; 72 h, siC 5 80%/ siE7 5 50% and 120 h, siC 5 100%/siE7 5 80%). Western blotting results showed slightly E-cadherin levels in siE7-treated KCs with higher cell confluence (supplementary Figure 3, available at Carcinogenesis Online). However, although the confluence of siC-treated cells was higher than that of siE7-transfected cells at the different time points, E-cadherin expression of siE7-treated cells was more intense than that of siC-treated cells, suggesting that the effect of E7 silencing on E-cadherin expression may not be explained only by higher E-cadherin stabilization due to cell confluence. E7-silenced HPV16 KCs, which re-expressed surface E-cadherin protein, were distributed as cohesive clusters, whereas siC-treated cells had a more homogeneous distribution with sparse aggregates, in spite of their higher confluence. E-cadherin re-expression following HPV16 E7 silencing is mediated by AP-2 transcription factor To determine the potential mechanism implicated in the decreased expression of E-cadherin induced by E7, we examined by RT–PCR (Figure 4A) the effects of E7 silencing on the expression of Slug and Snail transcriptions factors. Figure 4 shows that the up-regulation of E-cadherin following E7 silencing is not associated with the downregulation of Snail or Slug transcriptional repressors. As the half-life of Rb is decreased in cells expressing high-risk HPV E7 protein, we analysed nuclear cell extracts from HPV16positive SiHa cells by western blotting with a Rb-specific antibody (Figure 4B) and by RT–PCR (data not shown). The treatment of

1444

SiHa cells with siE7 resulted in higher levels of Rb protein as detected by immunoblotting, but no modulation of Rb mRNA was observed at the transcriptional level. In contrast, cells treated with siC expressed low protein levels of Rb. As Rb has been shown to induce E-cadherin expression via the AP-2 transcription factor, we also analysed AP-2a by RT–PCR and western blotting. Although the transfection of SiHa cells with siE7 did not induce modifications of AP-2a mRNA (data not shown), this treatment resulted in increased levels of AP-2a protein compared with cells treated with siC (Figure 4). HPV16 E7 silencing increases heterotypic E-cadherin adhesion between LC and HPV16-transformed KC An in vitro cell heterotypic adhesion assay was performed with SiHa cells and CM-DiL-labelled LC. The capacity of LC to adhere to KC was measured by counting the number of labelled LC by fluorescence microscopy. LC adhesion to SiHa cells was significantly increased when E7 oncogene was silenced compared with siC-treated cells (Figure 5A). Fluorescence microscopy revealed heterotypic E-cadherin-mediated adhesion between LC and SiHa cells when E7 oncogene was silenced (Figure 5B–D). An E-cadherin-blocking antibody was added in the culture medium before the co-culture with LC and a significant reduction in the percentage of adherent LC was systematically observed in the presence of the blocking antibody compared with an isotype control antibody (Figure 5A). A very low number of adherent CM-DiL-labelled LC was also observed when the adhesion assay was performed with E-cadherin-negative C33A cells (data not shown). Discussion Potentially complementing the action of cytokines or chemokines, E-cadherin has been shown to be necessary to maintain a balanced turnover of LC within epithelial tissues (6).


Fig. 2. (A) Western blot detection of HPV16 E6 and E7 proteins in total cell extracts from SiHa cells at 24 and 48 h. b-Actin served as loading control. (B) Mock, control (siC) and E7 (siE7) siRNA effects on the growth of SiHa cells. Cells were counted at 20, 44, 68 and 92 h after treatment with different siRNAs. (C) Western blot detection of E7 and p16 protein in total cell extracts from siE7-treated E7 cells at various times (24 and 48 h). E7 cells treated with siC served as controls. KU-70 served as loading control. (D) Mock, control (siC) and E7 (siE7) siRNA effects on the growth of E7 cells. Results are expressed as the means ± SDs of three experiments. Asterisks indicate statistically significant differences ( P , 0.05, P , 0.001).


Fig. 3. (A–J) E-cadherin expression after siE7s transfection. Cells transfected with siC served as controls. (A) Western blot detection of E-cadherin proteins in total cell extracts from SiHa and KE7 cells at 24 and 48 h. b-Actin served as loading control. (B–Q) Confocal laser microscopy images of normal exocervical keratinocytes (KN) (B–D), SiHa cells transfected with siC (E–G), siE7 (H–J) or KE7 cells transfected with siC (K–M), siE7 (O–Q) and stained with Toto3 (B, E, H, K and O), E-cadherin:fluorescein isothiocyanate (FITC) (C, F, I, L and P) or Toto3/ E-cadherin:fluorescein isothiocyanate (D, G, J, M and Q). Bar, 10 lm.

In addition to the homophilic and homotypic interactions of E-cadherin, which are important for the cell adhesion within epithelial tissues, there also exist homophilic and heterotypic interactions of E-cadherin between epithelial cells and LC, which are important for maintaining LC in the epithelium (6). The presence of LC in the epithelium is essential for the immune recognition of foreign antigens on mucosal surfaces. Consequently, the intimate contact between KC and LC is potentially important not only for initiating but also for maintaining the immune response during the malignant transformation of the cervix. In this study, we first confirmed that the intraepithelial expression of E-cadherin is decreased in HPV16-positive

cervical SILs compared with the normal squamous epithelium and is associated with a lower density of CD1aþ LC. A progressive loss of E-cadherin staining was observed with increased severity of cervical lesions. These findings are in agreement with previous studies reporting abnormal cytoplasmic staining or absence of E-cadherin in cervical pre-neoplastic lesions (8,33–35). However, within each histological diagnosis category, no significant correlation was found between the distribution of specific HPV genotypes and the expression of E-cadherin and/or CD1a, suggesting that the decreased density of LC may be just one of the factors required to sustain the persistence of the virus in the KCs. The resulting immunotolerance is probably, however, to be a necessary condition for the expression of the oncogenic potential of high-risk HPVs. Since the major etiological agent of cervical cancer is HPV, we postulated that viral proteins directly interfere with the ability of LC to induce a protective immune response. A previous report already demonstrated that expression of HPV16 E6 in KC reduces the levels of cell surface E-cadherin (36). In the present study, we also found that the E7 oncoprotein affects the expression of cell surface E-cadherin in HPV16-positive KC. HPV16 E7 oncoprotein has been shown to be expressed and to affect a plethora of cellular gene products during the malignant transformation. The most well-characterized target of E7 is the Rb tumour suppressor protein. Interaction between E7 and Rb disrupts the ability of Rb to bind cellular E2F transcription factors, inhibits Rb-mediated repression of E2F-responsive genes (37,38) and leads to the proteasomal degradation of Rb in cultured cells (14,15,39). Interestingly, Rb inactivation was also found to induce a complete mesenchyme-like conversion and a loss of expression of epithelial marker genes including E-cadherin and cytokeratins.

1445


Fig. 4. RT–PCR of E-cadherin, Snail and Slug mRNA (A) and western blot detection of Rb and AP-2a proteins (B) in cell extracts from SiHa cells transfected with siCs or anti-siE7s. HPRT and KU-70 served as loading control.


These changes were concomitant with the appearance of cell invasiveness in vitro (40). One mechanism by which E-cadherin expression is modulated by E7 oncoprotein in cervical (pre)neoplastic lesions may be related to the Rb protein which has been shown to induce E-cadherin expression via the AP-2 transcription factor (10). Interestingly, the silencing of HPV16 E7 in HPV16-transformed KC resulted in Rb re-expression and increased detection levels of AP-2a protein. We next investigated the functional consequences of E-cadherin re-expression by HPV16-infected KC on LC adhesion. E7 gene silencing was shown to induce an increased adherence of LC to HPV16-transformed KC in a heterotypic cell adhesion assay. Accordingly, LCs were more efficient to adhere to E7-silenced HPV16positive KC expressing higher levels of E-cadherin compared with control-transfected HPV16-positive KC. We further demonstrated that the cell surface E-cadherin re-expressed at the membrane of HPV16-positive KC was responsible for the increased LC/KC adhesion, since the recruitment of LC was diminished in the presence of a blocking anti-E-cadherin antibody. However, the observation that the adhesion of LC was not completely suppressed with the blocking antibody suggests the role of other adhesion molecules. In addition, cytokines or chemokines produced by HPV-transformed KC, although at low levels, such as CCL20 (41), defensins (42), GMCSF (43) and MCP-1 (44), could also influence the recruitment of LC/DC. In conclusion, the inability of the local immune system to mount a protective cell-mediated immunity against malignant cells or cells in transformation due to a virus-induced deficiency of adhesion molecules necessary for cell to cell interactions might play an important role in the development of cervical cancer. The altered expression of

1446

E-cadherin induced by E7 oncoprotein could not only be an early indication signalling the behaviour of abnormal neoplastic cells, such as the risk of local invasion and metastasis at a distant site, but might also constitute one of the major determinants for establishing a local immunodeficiency via antigen presentation dysfunctions. Supplementary material Supplementary Table 1 and Figures 1–3 can be found at http://carcin. oxfordjournals.org/ Funding Belgian Fund for Medical Scientific Research;Centre de Recherche Interuniversitaire en Vaccinologie with a grant from the Walloon Region and GlaxoSmithKline (4216); Marshall Programme of the Walloon Region (Neoangio 616476); L.Fredericq Fund and Centre Anti-Cancereux pre`s l’Universite´ de Lie`ge. Acknowledgements P.D. is a Research Director of the Belgian National Fund for Scientific Research. Conflict of Interest Statement: None declared.

References 1. Delvenne,P. et al. (2004) Epithelial metaplasia: an inadequate environment for antitumour immunity? Trends Immunol., 25, 169–173.


Fig. 5. (A–D) Heterotypic adhesion assay between SiHa cells and CM-DiL-labelled LC. (A) SiHa cells were transfected with the transfecting agent only (Mock), control (siC) and E7 (siE7) (means of triplicates). An E-cadherin-blocking antibody (siRNA þ anti-E-cadherin Ab) or an isotype control antibody (siRNA þ control Ab) was added or not in the culture medium in the presence of LC. (B–D) Confocal microscopy pictures of co-cultures of LC and SiHa cells transfected with siE7s and stained with E-cadherin:fluorescein isothiocyanate (FITC)/Toto3 (B), LC:CM-DiL/Toto3 (C) and E-cadherin:fluorescein isothiocyanate/LC:CMDiL/Toto3 (D). Bar, 10 lm. Asterisks indicate statistically significant differences ( P , 0.001).


23. Hubert,P. et al. (1999) Colonization of in vitro-formed cervical human papillomavirus-associated (pre)neoplastic lesions with dendritic cells—role of granulocyte/macrophage colony-stimulating factor. Am. J. Pathol., 154, 775–784. 24. Hubert,P. et al. (2005) Production of large numbers of Langerhans’ cells with intraepithelial migration ability in vitro. Exp. Dermatol., 14, 469–477. 25. Jiang,M. et al. (2002) Selective silencing of viral gene expression in HPVpositive human cervical carcinoma cells treated with siRNA, a primer of RNA interference. Oncogene, 21, 6041–6048. 26. Tang,S. et al. (2006) The E7 oncoprotein is translated from spliced E6 I transcripts in high-risk human papillomavirus type 16- or type 18-positive cervical cancer cell lines via translation reinitiation. J. Virol., 80, 4249–4263. 27. Tang,S. et al. (2006) Short-term induction and long-term suppression of HPV16 oncogene silencing by RNA interference in cervical cancer cells. Oncogene, 25, 2094–2104. 28. Andrade,W.N. et al. (1996) A method for tracking the migration of blood lymphocytes. Immunol. Invest., 25, 455–467. 29. Pebernard,S. et al. (2004) Determinants of interferon-stimulated gene induction by RNAi vectors. Differentiation, 72, 103–111. 30. Sledz,C.A. et al. (2003) Activation of the interferon system by shortinterfering RNAs. Nat. Cell Biol., 5, 834–839. 31. Sledz,C.A. et al. (2004) RNA interference and double-stranded-RNAactivated pathways. Biochem. Soc. Trans., 32, 952–956. 32. Giarre,M. et al. (2001) Induction of pRb degradation by the human papillomavirus type 16 E7 protein is essential to efficiently overcome p16(INK4a)-imposed G(1) cell cycle arrest. J. Virol., 75, 4705–4712. 33. Bremnes,R.M. et al. (2002) The E-cadherin cell-cell adhesion complex and lung cancer invasion, metastasis, and prognosis. Lung Cancer, 36, 115–124. 34. Fearon,E.R. (2003) Connecting estrogen receptor function, transcriptional repression, and E-cadherin expression in breast cancer. Cancer Cell, 3, 307–310. 35. Kuefer,R. et al. (2003) The role of an 80 kDa fragment of E-cadherin in the metastatic progression of prostate cancer. Clin. Cancer Res., 9, 6447–6452. 36. Matthews,K. et al. (2003) Depletion of Langerhans cells in human papillomavirus type 16-infected skin is associated with E6-mediated down regulation of E-cadherin. J. Virol., 77, 8378–8385. 37. Chellappan,S. et al. (1992) Adenovirus E1A, simian virus 40 tumor antigen, and human papillomavirus E7 protein share the capacity to disrupt the interaction between transcription factor E2F and the retinoblastoma gene product. Proc. Natl Acad. Sci. USA, 89, 4549–4553. 38. Phelps,W.C. et al. (1991) Analysis of trans activation by human papillomavirus type 16 E7 and adenovirus 12S E1A suggests a common mechanism. J. Virol., 65, 6922–6930. 39. Boyer,B. et al. (2000) Induction and regulation of epithelial-mesenchymal transitions. Biochem. Pharmacol., 60, 1091–1099. 40. Martel,C. et al. (1997) Inactivation of retinoblastoma family proteins by SV40 T antigen results in creation of a hepatocyte growth factor/scatter factor autocrine loop associated with an epithelial-fibroblastoid conversion and invasiveness. Cell Growth Differ., 8, 165–178. 41. Herman,L. et al. (2006) MIP3 alpha stimulates the migration of Langerhans cells in models of human papillomavirus (HPV)-associated (pre)neoplastic epithelium. Cancer Immunol. Immunother., 56, 1087–1096. 42. Hubert,P. et al. (2007) Defensins induce the recruitment of dendritic cells in cervical human papillomavirus-associated (pre)neoplastic lesions formed in vitro and transplanted in vivo. FASEB J., 21, 2765–2775. 43. Woodworth,C.D. et al. (1993) Comparative lymphokine secretion by cultured normal human cervical keratinocytes, papillomavirus-immortalized, and carcinoma cell-lines. Am. J. Pathol., 142, 1544–1555. 44. Riethdorf,S. et al. (1998) Expression of the MCP-1 gene and the HPV 16 E6/E7 oncogenes in squamous cell carcinomas of the cervix uteri and metastases. Pathobiology, 66, 260–267. Received November 19, 2007; revised June 11, 2008; accepted June 11, 2008

1447


2. Bosch,F.X. et al. (1995) Prevalence of human papillomavirus in cervicalcancer—a worldwide perspective. J. Natl Cancer Inst., 87, 796–802. 3. Giannini,S.L. et al. (2002) Influence of the mucosal epithelium microenvironment on Langerhans cells: implications for the development of squamous intraepithelial lesions of the cervix. Int. J. Cancer, 97, 654–659. 4. Al-Saleh,W. et al. (1995) Inverse modulation of intraepithelial Langerhans cells and stromal macrophage dendrocyte populations in human papillomavirus-associated squamous intraepithelial lesions of the cervix. Virchows Arch. Int. J. Pathol., 427, 41–48. 5. Kripke,M.L. et al. (1990) Evidence that cutaneous antigen-presenting cells migrate to regional lymph-nodes during contact sensitization. J. Immunol., 145, 2833–2838. 6. Tang,A.M. et al. (1993) Adhesion of epidermal Langerhans cells to keratinocytes mediated by E-cadherin. Nature, 361, 82–85. 7. Hodivala,K.J. et al. (1994) Evidence that Cadherins play a role in the downregulation of integrin expression that occurs during keratinocyte terminal differentiation. J. Cell Biol., 124, 589–600. 8. Hubert,P. et al. (2005) E-cadherin-dependent adhesion of dendritic and Langerhans cells to keratinocytes is defective in cervical human papillomavirus-associated (pre)neoplastic lesions. J. Pathol., 206, 346–355. 9. Zur Hausen,H. (2000) Papillomaviruses causing cancer: evasion from hostcell control in early events in carcinogenesis. J. Natl Cancer Inst., 92, 690–698. 10. Batsche,E. et al. (1998) RB and c-Myc activate expression of the E-cadherin gene in epithelial cells through interaction with transcription factor AP2. Mol. Cell. Biol., 18, 3647–3658. 11. Gage,J.R. et al. (1990) The E7 proteins of the nononcogenic human papillomavirus type 6b (HPV-6b) and of the oncogenic HPV-16 differ in retinoblastoma protein binding and other properties. J. Virol., 64, 723–730. 12. Imai,Y. et al. (1991) Purification and characterization of human papillomavirus type 16 E7 protein with preferential binding capacity to the underphosphorylated form of retinoblastoma gene product. J. Virol., 65, 4966–4972. 13. Boyer,S.N. et al. (1996) E7 protein of human papilloma virus-16 induces degradation of retinoblastoma protein through the ubiquitin-proteasome pathway. Cancer Res., 56, 4620–4624. 14. Gonzalez,S.L. et al. (2001) Degradation of the retinoblastoma tumor suppressor by the human papillomavirus type 16 E7 oncoprotein is important for functional inactivation and is separable from proteasomal degradation of E7. J. Virol., 75, 7583–7591. 15. Jones,D.L. et al. (1997) Analysis of the p53-mediated G(1) growth arrest pathway in cells expressing the human papillomavirus type 16 E7 oncoprotein. J. Virol., 71, 2905–2912. 16. Batlle,E. et al. (2000) The transcription factor snail is a repressor of E-cadherin gene expression in epithelial tumour cells. Nat. Cell Biol., 2, 84–89. 17. Bolos,V. et al. (2003) The transcription factor Slug represses E-cadherin expression and induces epithelial to mesenchymal transitions: a comparison with Snail and E47 repressors. J. Cell Sci., 116, 499–511. 18. Cano,A. et al. (2000) The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression. Nat. Cell Biol., 2, 76–83. 19. Hajra,K.M. et al. (2002) The SLUG zinc-finger protein represses E-cadherin in breast cancer. Cancer Res., 62, 1613–1618. 20. Delvenne,P. et al. (1994) Detection of human papillomaviruses in paraffinembedded biopsies of cervical intraepithelial lesions—analysis by immunohistochemistry, in-situ hybridization, and the polymerase chain-reaction. Mod. Pathol., 7, 113–119. 21. Aguilar-Lemarroy,A. et al. (2002) Restoration of p53 expression sensitizes human papillomavirus type 16 immortalized human keratinocytes to CD95mediated apoptosis. Oncogene, 21, 165–175. 22. Hubert,P. et al. (1998) In vitro propagated dendritic cells from patients with human-papilloma virus-associated preneoplastic lesions of the uterine cervix: use of Flt3 ligand. Cancer Immunol. Immunother., 47, 81–89.