Altered microRNA expression associated with chromosomal ... - Nature

Oncogene (2013) 32, 106 - 116 & 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13 www.nature.com/onc

ORIGINAL ARTICLE

Altered microRNA expression associated with chromosomal changes contributes to cervical carcinogenesis SM Wilting1, PJF Snijders1, W Verlaat1, A Jaspers1, MA van de Wiel2,3, WN van Wieringen2,3, GA Meijer1, GG Kenter4, Y Yi5, C le Sage6,7, R Agami6, CJLM Meijer1 and RDM Steenbergen1 Little is known about the alterations in microRNA (miRNA) expression patterns during the consecutive stages of cervical cancer development and their association with chromosomal instability. In this study, miRNA expression in normal cervical squamous epithelium, high-grade precancerous lesions (cervical intraepithelial neoplasia (CIN2 -- 3)), squamous cell carcinomas (SCCs) and adenocarcinomas (AdCAs) was integrated with previously generated chromosomal profiles of the same samples. Significantly differential expression during the consecutive stages of cervical SCC development was observed for 106 miRNAs. Of these differentially expressed miRNAs, 27 showed early transiently altered expression in CIN2 -- 3 lesions only, 46 miRNAs showed late altered expression in SCCs only and 33 showed continuously altered expression in both CIN2 -- 3 and SCCs. Altered expression of five significantly differentially expressed miRNAs, hsa-miR-9 (1q23.2), hsa-miR-15b (3q25.32), hsa-miR-28-5p (3q27.3), hsa-miR100 and hsa-miR-125b (both 11q24.1), was directly linked to frequent chromosomal alterations. Functional analyses were performed for hsa-miR-9, representing a potential oncogene with increased expression linked to a chromosomal gain of 1q. Hsa-miR-9 overexpression was found to increase cell viability, anchorage-independent growth and migration in vitro. Upon organic raft culturing, hsa-miR-9 hampered differentiation and induced proliferation in all strata of the epithelial layer. These findings support a potential oncogenic function of hsa-miR-9 in cervical cancer. In summary, differential expression of 106 miRNAs, partly associated with chromosomal alterations, was observed during cervical SCC development. Altered expression of hsa-miR-9 associated with a chromosomal gain of chromosome 1q was shown to be functionally relevant, underlining the importance of deregulated miRNA expression in cervical carcinogenesis. Oncogene (2013) 32, 106 -- 116; doi:10.1038/onc.2012.20; published online 13 February 2012 Keywords: microRNA; squamous cell carcinoma; adenocarcinoma; CIN lesion; HPV; hsa-miR-9

INTRODUCTION Cervical cancer is caused by a persistent infection with high-risk types of the human papillomavirus (hrHPV).1,2 Histologically, cervical cancer can be divided into different subtypes, including squamous cell carcinomas (SCCs; about 80% of cases) and adenocarcinomas (AdCAs; about 5 -- 20% of cases).3,4 Whereas SCCs develop via well-defined precursor stages, called cervical intraepithelial neoplasia (CIN, graded 1 -- 3), precursor stages for AdCAs are less well characterised. To reflect their relative risk of progression to cervical cancer, CIN1 lesions are nowadays often referred to as low-grade CIN whereas CIN2 and 3 lesions together are considered high-grade CIN. Whereas high-grade CIN can arise rather fast following hrHPV infection (2 -- 3 years), subsequent development of an invasive carcinoma may take at least one or more decades,5 - 8 indicating that CIN2-3 lesions represent a heterogeneous disease. Nearly all high-grade CIN lesions (CIN2 -- 3) and virtually all carcinomas are associated with a so-called transforming hrHPV infection, which is characterised by deregulated expression of the viral oncogenes E6 and E7 in the proliferating (supra)basal cells of the epithelium. Genetic instability due to the presence of a transforming infection can ultimately result in the accumulation of 1

specific (epi)genetic changes in the host-cell genome driving progression to a malignant phenotype. Although previous studies have identified chromosomal alterations, gene expression changes and aberrant promoter methylation associated with cervical cancer, little is known about the role of microRNAs (miRNAs). Whereas the importance of miRNAs in human carcinogenesis is becoming increasingly recognised, the potential role of miRNAs in cervical cancer development has been demonstrated only in few studies. HPV-encoded genes have been shown to influence the miRNA expression of its host cell.9 - 14 Muralidhar et al. 15,16 showed that a frequently occurring gain of chromosome 5p in cervical cancer is associated with functionally relevant elevated expression of the miRNA-processing enzyme Drosha, and altered miRNA expression profiles. Furthermore, a panel of miRNAs able to predict cervical cancer survival was recently described.17 A number of studies have shown altered miRNA expression patterns in cervical-cancer cell lines and cervical-cancer tissues.18 - 21 So far, only limited data on genomewide miRNA expression patterns in CIN lesions is available.22,23 As human miRNAs are frequently located at fragile sites and chromosomal regions affected in cancer,24 chromosomal alterations are thought to represent a major mechanism underlying

Department of Pathology, VU University Medical Center, Amsterdam, The Netherlands; 2Department of Epidemiology & Biostatistics, VU University Medical Center, Amsterdam, The Netherlands; 3Department of Mathematics, VU University, Amsterdam, The Netherlands; 4Department of Obstetrics & Gynaecology, VU University Medical Center Amsterdam, Amsterdam, The Netherlands; 5Division of Genetic Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, Nashville, TN, USA and 6Division of Gene Regulation, the Netherlands Cancer Institute, Amsterdam, The Netherlands; 7Current Address: The Wellcome Trust/Cancer Research UK Gurdon Institute, University of Cambridge, Cambridge, United Kingdom. Correspondence: Dr RDM Steenbergen, Department of Pathology, Unit of Molecular Pathology, VU University Medical Center, PO box 7057, 1007 MB Amsterdam, The Netherlands. E-mail: [email protected] Received 28 August 2011; revised 13 December 2011; accepted 6 January 2012; published online 13 February 2012

microRNA expression profiling of cervical cancer SM Wilting et al

107

altered miRNA expression in cancer. Indeed, an association between chromosomal alterations and differential miRNA expression has been described for ovarian cancer, breast cancer, melanoma, neuroblastoma and myeloma cell lines.25 - 28 In cervical cancer we previously showed that particularly gains on chromosomes 1, 3q and 20q may be involved in malignant progression.29 Integration of chromosomal alterations with miRNA expression may further clarify the functional role of frequent chromosomal alterations and aid in the identification of novel potential tumour suppressors and oncogenes in cervical carcinogenesis. In this study, we aimed to obtain a comprehensive overview of the alterations in miRNA expression that occur during the progression from normal cervical epithelium, via high-grade CIN lesions, to SCCs and to relate the observed changes in miRNA expression to chromosomal alterations. Genome-wide miRNA expression profiles of micro-dissected specimens of normal squamous epithelium, CIN2 -- 3, SCCs and AdCAs were obtained and directly integrated with previously generated chromosomal profiles of the same samples.29,30 Functional effects of altered expression of hsa-miR-9, which was associated with a frequent chromosomal gain, were further investigated in vitro. RESULTS Distinct miRNA expression profiles related to histology In this study genome-wide miRNA expression profiles were determined in microdissected SCCs (n ¼ 10), AdCAs (n ¼ 9), CIN2 -- 3 (n ¼ 18) and normal cervical squamous epithelial (n ¼ 10) samples. To obtain an overview of the similarities between the overall miRNA expression profiles, unsupervised hierarchical clustering was performed, resulting in two major clusters (Figure 1). Cluster 1 contained 90% of normal samples, 22% of CIN2 -- 3 and 10% of SCCs. Cluster 2 contained only one normal sample (10%) and the vast majority of all cervical (pre)malignant lesions (78% of CIN2 -- 3, 90% of SCCs and 100% of AdCAs). Within cluster 2, separate sub-clusters could be observed. Although one of them (Figure 1 2a) comprises of a mixture of four CIN2 -- 3, 2 AdCAs and one normal sample, the other sub-clusters contained only AdCAs (Figure 1 2b), CIN2 -- 3 (Figure 1 2c) or SCCs (Figure 1 2d). The fact that CIN2 -- 3 lesions are most scattered across the different clusters may reflect the heterogeneity of high-grade CIN disease in general. Our clustering results suggest that miRNA expression in a subset of the CIN lesions is still similar to hrHPV-positive normal epithelium, supporting the observation that CIN2 -- 3 lesions may take one or two decades to progress into an invasive carcinoma.5 On the other hand, cluster 2a suggests that another subset of the CIN 2 -- 3 lesions may resemble AdCAs with respect to their miRNA expression profile. Interestingly, CIN 2 -- 3 lesions are frequently found adjacent to AdCAs, which may be a result of bidirectional differentiation of hrHPV-infected metaplastic epithelial cells of the transformation zone.31 Differential miRNA expression related to histology miRNA expression was compared between normal epithelium, CIN2 -- 3, SCCs and AdCAs. In total, 106 miRNAs were differentially expressed in CIN2 -- 3 and/or SCCs compared with normal epithelium (Table 1a), whereas 18 miRNAs were differentially expressed between SCCs and AdCAs (Table 1b). To verify the reliability of the miRNA arrays we re-analysed the expression of hsa-miR-9/15b/21/28-5p/100/125b/203/375 in 39 of 47 samples analysed by array using quantitative reverse transcription--PCR (qRT--PCR). For all miRNAs, significant correlations between array and qRT--PCR results were found (range 0.67 -- 0.95; Supplementary Figure 1; Po0.01 for all miRNAs). Fold changes between sample groups detected by qRT--PCR were in general larger than those detected by microarray (Supplementary Figure 2). & 2013 Macmillan Publishers Limited

Figure 1. Unsupervised hierarchical clustering results. Normal cervical squamous epithelial samples (n ¼ 10) are labelled black, CIN2 -- 3 (n ¼ 18) are labelled red, SCCs (n ¼ 10) are labelled blue and AdCAs (n ¼ 9) are labelled green. Expression levels of all miRNAs are on a green (low expression) to red (high expression) scale.

Results of the comparisons involving normal epithelium, CIN2 -3 and SCCs were subsequently combined to determine patterns of differential miRNA expression during cervical SCC development. MiRNAs with significant differential expression in CIN2 -- 3 compared with normal (false discovery rate (FDR)o0.05) that were unchanged in SCCs compared with normal (FDR40.1) were designated ‘early transient’ (Figure 2a; 18 upregulated and 9 downregulated miRNAs). MiRNAs showing significant differential expression in SCCs compared with normal and/or compared with CIN2 -- 3 (FDRo0.05), but not in CIN2 -- 3 compared with normal (FDR40.1) were classified as ‘late’ (Figure 2b; 27 upregulated and 19 downregulated miRNAs). Finally, ‘early continuous’ differentially expressed miRNAs (Figure 2c; 18 upregulated miRNAs and 15 downregulated miRNAs) showed concordant differential expression in SCCs and CIN2 -- 3 compared with normal (one FDRo0.05 and the other FDRo0.1). Differential miRNA expression related to chromosomal alterations Differential gene locus mapping analysis (DIGMAP) was applied to compare the expression profiles of all miRNAs present on the array grouped by their chromosomal location between CIN2 -- 3/SCCs and normal epithelium. This yielded four loci, located at chromosomes 9, 13, 14 and X, exhibiting differential miRNA Oncogene (2013) 106 - 116


108 Table 1a.

Differentially expressed miRNAs in CIN2 - 3 and/or SCCs compared with normal squamous epithelium

Name

Class

CIN vs normal

Upregulated miRs hsa-miR-192 hsa-miR-135b hsa-miR-101 hsa-miR-191 hsa-miR-34c-5p hsa-miR-150 hsa-miR-125a-5p hsa-miR-30a hsa-miR-143 hsa-miR-146b-5p hsa-miR-181b hsa-let-7g hsa-miR-26a hsa-miR-29c hsa-miR-29b hsa-miR-10a hsa-miR-29a hsa-miR-145 hsa-miR-30c hsa-miR-425 hsa-miR-24 hsa-miR-331-3p hsa-miR-151-3p hsa-miR-107 hsa-miR-652 hsa-miR-17* hsa-miR-9 hsa-miR-185 hsa-miR-339-5p hsa-miR-18a hsa-let-7d hsa-miR-17 hsa-miR-30d hsa-miR-130b hsa-miR-15a hsa-miR-106a hsa-miR-19a hsa-miR-200c hsa-miR-20b hsa-miR-363 hsa-miR-155 hsa-miR-141 hsa-miR-93 hsa-miR-15b hsa-miR-16 hsa-miR-28-5p hsa-miR-338-5p hsa-miR-206 hsa-miR-200a* hsa-miR-92b hsa-let-7i hsa-miR-181d hsa-miR-92a hsa-miR-30e hsa-miR-34b* hsa-miR-592 hsa-miR-19b hsa-miR-106b hsa-miR-595 hsa-miR-34a hsa-miR-25 hsa-miR-146a hsa-miR-21

Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early

Downregulated miRs hsa-miR-205 hsa-miR-27a

Early transient Early transient

Oncogene (2013) 106 - 116

transient transient transient transient transient transient transient transient transient transient transient transient transient transient transient transient transient transient

continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous

SCC vs normal

SCC vs CIN

FDR

FC

FDR

FC

FDR

FC

0.008 0 0.034 0.011 0.002 0.046 0.01 0.007 0.034 0.002 0 0.004 0 0.002 0.029 0 0 0.014 NS NS 0.115 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS 0.046 0.008 0.042 0.037 0 0.034 0 0.047 0.02 0 0 0.02 0.046 0.027 0.037 0.002 0 0.037

1.11 1.12 1.20 1.26 1.26 1.35 1.35 1.38 1.38 1.39 1.46 1.47 1.52 1.53 1.57 1.78 1.87 1.92 1.04 1.10 0.81 0.96 1.02 0.97 0.98 1.03 0.98 1.07 1.02 1.03 1.06 1.00 0.95 1.05 0.98 1.03 1.19 0.89 1.13 1.15 1.20 0.91 1.26 1.19 1.19 1.06 1.10 1.11 1.13 1.13 1.17 1.23 1.25 1.25 1.28 1.28 1.33 1.35 1.35 1.40 1.52 1.73 2.13

NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS 0.027 0.026 NS 0.043 0.02 0.052 0.009 0.024 0.046 0.013 0.04 0.044 0.014 0.026 NS 0.033 0.074 0.02 0.017 0.016 0.035 0.032 0.021 0.011 0 0.027 0.017 0.068 0 0.075 0.004 0 0.029 0 0.009 0.098 0.015 0.004 0.009 0 0.076 0.052 0 0.065 0.007

1.07 1.11 1.17 1.16 1.15 1.06 1.05 1.30 1.05 1.25 1.24 1.09 1.01 1.12 1.18 1.08 0.98 0.87 1.12 1.13 1.16 1.17 1.18 1.19 1.25 1.26 1.26 1.26 1.27 1.28 1.29 1.29 1.29 1.31 1.34 1.35 1.37 1.37 1.37 1.38 1.44 1.59 1.68 2.06 2.17 1.22 1.26 1.15 1.11 1.21 1.32 1.17 1.39 1.22 1.18 1.09 1.66 1.98 1.38 1.55 1.73 1.67 3.68

NS NS NS NS NS NS 0.179 NS NS NS NS NS 0.018 NS NS 0.04 0.007 0.005 NS NS 0.009 0.018 0.041 0.009 0 0.036 0.02 0.013 0.031 0.021 0.011 0.042 0.036 NS 0.045 NS NS 0.024 NS NS NS 0.005 NS 0.02 0.02 NS NS NS NS NS NS NS NS NS NS NS NS 0.031 NS NS NS NS NS

0.96 0.98 0.98 0.92 0.91 0.79 0.78 0.94 0.76 0.90 0.85 0.74 0.66 0.73 0.75 0.61 0.53 0.45 1.08 1.02 1.43 1.23 1.15 1.23 1.28 1.22 1.28 1.19 1.24 1.24 1.21 1.29 1.36 1.25 1.37 1.30 1.15 1.54 1.22 1.19 1.20 1.74 1.34 1.74 1.83 1.15 1.15 1.04 0.99 1.07 1.13 0.96 1.12 0.98 0.93 0.85 1.25 1.47 1.02 1.11 1.14 0.97 1.73

0.009 0.002

0.49 0.60

NS NS

0.75 0.91

NS 0.016

1.52 1.52

& 2013 Macmillan Publishers Limited


109 Table 1a (Continued ) Name

hsa-miR-27b hsa-miR-221 hsa-miR-193a-3p hsa-miR-212 hsa-miR-770-5p hsa-miR-484 hsa-miR-636 hsa-miR-494 hsa-miR-125b hsa-miR-375 hsa-miR-99a hsa-miR-188-5p hsa-miR-148a hsa-miR-671-5p hsa-miR-199b-3p hsa-miR-513b hsa-miR-378 hsa-miR-195 hsa-miR-486-5p hsa-miR-26b hsa-miR-376a hsa-miR-199a-5p hsa-miR-497 hsa-miR-100 hsa-miR-660 hsa-miR-218 hsa-miR-203 hsa-miR-638 hsa-miR-370 hsa-miR-575 hsa-miR-193b hsa-miR-572 hsa-miR-149 hsa-miR-210 hsa-miR-622 hsa-miR-23b hsa-miR-493 hsa-miR-296-5p hsa-miR-617 hsa-miR-134 hsa-miR-365

Class

Early Early Early Early Early Early Early Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Late Early Early Early Early Early Early Early Early Early Early Early Early Early Early Early

CIN vs normal

transient transient transient transient transient transient transient

continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous continuous

SCC vs normal

SCC vs CIN

FDR

FC

FDR

FC

FDR

FC

0.017 0.002 0.004 0.015 0.048 0.005 0.046 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS 0.002 0.007 0.042 0 0.005 0.009 0.002 0.007 0.027 0.02 0.034 0 0.016 0.046 0.004

0.66 0.76 0.77 0.78 0.88 0.92 0.94 1.01 0.95 0.63 0.87 0.65 1.12 0.69 1.11 0.85 0.80 1.08 0.89 1.19 1.00 1.42 1.09 1.07 1.14 1.09 0.19 0.30 0.30 0.38 0.41 0.53 0.54 0.56 0.70 0.72 0.80 0.81 0.84 0.86 0.88

NS NS NS NS NS NS NS 0.013 0.013 0.004 0.043 0.016 0.013 0.027 0.007 0.015 0.041 0.037 0.027 NS 0.013 NS NS 0.159 NS NS 0 0.014 0.014 0.004 0.004 0.02 0.041 0.078 0.029 0.074 0.08 0.074 0.009 0.074 0.007

0.70 0.89 1.17 0.89 0.92 0.99 0.94 0.32 0.34 0.47 0.53 0.56 0.64 0.66 0.67 0.69 0.73 0.80 0.82 0.83 0.83 0.89 0.92 0.93 0.93 0.94 0.24 0.24 0.20 0.36 0.50 0.46 0.56 0.80 0.68 0.72 0.81 0.91 0.79 0.86 0.86

NS NS 0.007 NS NS 0.013 NS 0.007 0.013 0.02 NS NS 0 NS 0 0 NS 0.016 NS 0 0.011 0 0 0.041 0 0.036 NS NS 0.031 NS NS NS NS NS NS NS NS 0.074 NS NS NS

1.06 1.16 1.52 1.14 1.04 1.08 1.00 0.32 0.35 0.75 0.61 0.86 0.57 0.95 0.60 0.82 0.91 0.74 0.92 0.70 0.83 0.63 0.84 0.87 0.82 0.86 1.24 0.79 0.68 0.94 1.22 0.86 1.04 1.43 0.97 1.00 1.02 1.12 0.95 1.01 0.98

Abbreviations: CIN, cervical intraepithelial neoplasia; FC, fold change; FDR, false discovery rate; miRNAs, microRNAs; NS, not significant; SCC, squamous cell carcinoma. Differential expression of miRNAs printed in bold is associated with a chromosomal alteration.

expression between CIN2 -- 3/SCCs and normal epithelium (Table 2). The loci on chromosomes 9 and 14 were not positioned in a frequently altered chromosomal region; however, these loci are known to contain clusters of miRNAs located within 10 kb of each other. For chromosome X no chromosomal information was available, but miRNAs in the identified locus on chromosome X are also clustered. Finally, the locus on chromosome 13 showed increased expression of a well-known cluster of onco-miRs (17 -- 92 cluster) located there, despite the presence of a chromosomal loss. Previously generated genome-wide chromosomal profiles of all CIN2 -- 3 and SCCs used in this study also enabled us to directly investigate the association between copy number and miRNA expression. Expression of miRNAs was compared between CIN2 -- 3 and SCCs with a gain or loss of the respective miRNA gene locus and samples without this chromosomal alteration using IntCNGEAn.32 Only miRNAs located at chromosomal alterations occurring in at least 30% of samples were included (n ¼ 81). IntCNGEAn analysis identified five miRNAs that also showed significantly differential expression in CIN2 -- 3/SCC compared with normal epithelium (printed in bold in Table 1a). For hsa-miR-9


(1q23.2), hsa-miR-15b (3q25.32) and hsa-miR-28-5p (3q27.3) the significantly increased expression found was associated with the presence of a chromosomal gain, whereas for hsa-miR-100 and hsa-miR-125b (both 11q24.1) the significantly decreased expression was related to a chromosomal loss. To validate these results we re-analysed the expression of the five identified miRNAs in 30 of 47 samples analysed by array using qRT--PCR. Increased expression of hsa-miR-9, hsa-miR-15b and hsa-miR-28-5p was found in samples with a gain of 1q (hsa-miR-9; P ¼ 0.09) or 3q (hsa-miR-15b; P ¼ 0.08 and hsa-miR-28-5p; P ¼ 0.03) (Supplementary Figures 3A -- C). Decreased expression of hsa-miR-100 and hsamiR-125b in samples with a loss of 11q was also observed by qRT-PCR, but was not significant (hsa-miR-100; P ¼ 0.27 and hsa-miR125b; P ¼ 0.30), which might be partly due to the small number of samples with a loss of 11q tested (n ¼ 9) (Supplementary Figures 3D and E). Functional involvement of hsa-miR-9 in vitro We found increased expression of hsa-miR-9 in cervical cancer to be associated with a gain of chromosome 1q, suggesting that hsaOncogene (2013) 106 - 116


110 Table 1b.

Differentially expressed miRNAs in SCCs compared with

AdCAs Name

SCC vs AdCA FDR

FC

Upregulated miRNAs hsa-miR-33a hsa-miR-23a hsa-miR-221 hsa-miR-224 hsa-miR-27a hsa-miR-210 hsa-miR-222 hsa-miR-205

0.032 0.011 0 0.038 0.006 0 0 0

1.21 1.37 1.40 1.53 1.78 2.53 3.33 15.60

Downregulated miRNAs hsa-miR-192 hsa-miR-10a hsa-miR-200b hsa-miR-375 hsa-miR-194 hsa-miR-145 hsa-miR-215 hsa-miR-199b-5p hsa-let-7g

0.006 0.013 0.016 0 0 0.032 0.033 0.013 0.013

0.21 0.34 0.37 0.46 0.49 0.52 0.57 0.62 0.65

Abbreviations: AdCA, adenocarcinoma; FC, fold change; FDR, false discovery rate; miRNAs, microRNAs; NS, not significant; SCC, squamous cell carcinoma.

miR-9 could serve as an oncogene in hrHPV-mediated carcinogenesis. The oncogenic potential of hsa-miR-9 was determined in the HPV16-immortalised keratinocyte cell line FK16A, as endogenous hsa-miR-9 expression was found to increase from early to latepassages of this cell line (data not shown). Ectopic expression of hsa-miR-9 in early-passage FK16A cells (Figure 3a) increased cell viability (Figure 3b; Po0.01), anchorageindependent growth (Figures 3c and d) and migration (Figure 3e), whereas no effect on invasion was found (Figure 3f). In addition, ectopic expression of hsa-miR-9 in SiHa cells with already high endogenous hsa-miR-9 expression further increased the cell viability of SiHa cells as well, indicating that hsa-miR-9 expression levels are not yet saturated in SiHa cells and confirming the beneficial effect of high levels of hsa-miR-9 expression in hrHPVtransformed cells (Supplementary Figures 4A and B). Transient knockdown of hsa-miR-9 in late-passage FK16A cells (Figure 4a) decreased cell viability (Figure 4b; Po0.01) and anchorage-independent growth (Figures 4c and d), but did not alter migratory and invasive capacity (Figures 4e and f). Knockdown of the endogenously already low hsa-miR-9 expression in early-passage FK16A did not alter cellular viability, confirming that these cells are not yet dependent on hsa-miR-9 expression (Supplementary Figures 4C and D). Finally, organotypic raft cultures of early-passage FK16A cells stably expressing hsa-miR-9 were performed to investigate the potential effect of hsa-miR-9 on differentiation. As shown in Figure 5, overexpression of hsa-miR-9 in early-passage FK16A hampered epithelial differentiation, as reflected by decreased protein expression of cytokeratin 10 (KRT10) (Figures 5d -- f). This reduction in differentiation was accompanied by an increased number of proliferating cells throughout all strata of the epithelium, including the superficial cell layers, as indicated by positivity for the proliferation-related Ki-67 antigen (MKI67) (Figures 5g -- i). The latter is in line with cell viability results (Figure 3b). Unfortunately, raft culture experiments could not be performed with late-passage FK16A with knockdown of hsa-miR-9 expression, due to the transient nature of the knockdown. Oncogene (2013) 106 - 116

Figure 2. Visualisation of the three distinct expression profiles identified within the group of differentially expressed miRNAs. Average intensities in normal samples, CIN2 -- 3 and SCCs as measured by the array for all genes in the class designated as (a) transient early (significantly differential expression in CIN2 -- 3 compared with normal (FDRo0.05) and unchanged in SCCs compared with normal (FDR40.1)). (b) Late (significantly differential expression in SCCs compared with normal and/or compared with CIN2 -- 3 (FDRo0.05), but not in CIN2 -- 3 compared with normal (FDR40.1)) and (c) early continuous (concordant differential expression in SCCs and CIN2 -- 3 compared with normal (one FDRo0.05 and the other FDRo0.1)).

DISCUSSION In this study, we investigated alterations in miRNA expression during cervical carcinogenesis in relation to chromosomal instability. To avoid the identification of miRNA alterations strictly related to the presence of hrHPV and favour the identification of miRNAs contributing to the transformation process following hrHPV infection, we chose to use hrHPV-positive normal cervical epithelium for our comparisons. For 106 miRNAs, significantly differential expression was observed in the histological categories of specimens representing & 2013 Macmillan Publishers Limited


111 Table 2.

Chromosomal loci identified by DIGMAP analysis showing differential miRNA expression between CIN2-3/SCC and normal cervical epithelium Cytoband

T-score

miRNAs in DFR

Cluster

CIN vs normal

SCC vs normal

SCC vs CIN

(miRbase v14, 10 kb) 9q22.33

8.51

13q31.3

8.91

14q32.31-q32.32

23q26.2-q26.3

8.86

11.55

hsa-let-7a hsa-let-7f hsa-let-7d hsa-miR-23b hsa-miR-27b hsa-miR-24 hsa-miR-622 hsa-miR-17* hsa-miR-17 hsa-miR-18a hsa-miR-19a hsa-miR-20a hsa-miR-19b hsa-miR-770-5p hsa-miR-493 hsa-miR-337 hsa-miR-127-3p hsa-miR-432 hsa-miR-136 hsa-miR-370 hsa-miR-299-5p hsa-miR-494 hsa-miR-376a hsa-miR-363 hsa-miR-19b hsa-miR-20b hsa-miR-106a hsa-miR-424

cluster21 cluster21 cluster21 cluster22 cluster22 cluster22 cluster5 cluster5 cluster5 cluster5 cluster5 cluster5 cluster3 cluster3 cluster3 cluster3 cluster3 cluster2 cluster2 cluster2 cluster6 cluster6 cluster6 cluster6 cluster7

FDR

FC

FDR

FC

FDR

FC

NS NS NS 0.020 0.017 0.115 0.027 NS NS NS NS NS 0.020 NS 0.034 NS NS NS NS 0.042 NS NS NS NS 0.020 NS NS NS

1.19 1.22 1.06 0.72 0.66 0.81 0.70 1.03 1.00 1.03 1.19 1.17 1.33 0.88 0.80 1.07 1.07 1.11 1.02 0.30 0.98 1.01 1.00 1.15 1.33 1.13 1.03 0.94

NS NS 0.014 0.074 NS NS 0.029 0.024 0.026 0.044 0.017 NS 0.009 NS 0.080 NS NS NS NS 0.014 NS 0.013 0.013 0.032 0.009 0.035 0.020 NS

1.03 1.12 1.29 0.72 0.70 1.16 0.68 1.26 1.29 1.28 1.37 1.22 1.66 0.92 0.81 1.05 0.97 1.06 0.94 0.20 0.96 0.32 0.83 1.38 1.66 1.37 1.35 0.82

NS NS 0.011 NS NS 0.009 NS 0.036 0.042 0.021 NS NS NS NS NS NS NS NS NS 0.031 NS 0.007 0.011 NS NS NS NS NS

0.87 0.92 1.21 1.00 1.06 1.43 0.97 1.22 1.29 1.24 1.15 1.04 1.25 1.04 1.02 0.98 0.91 0.95 0.92 0.68 0.98 0.32 0.83 1.19 1.25 1.22 1.30 0.87

Abbreviations: CIN, cervical intraepithelial neoplasia; DFR, differential flag regions; DIGMAP, differential gene locus mapping analysis; FC, fold change; FDR, false discovery rate; miRNAs, microRNAs; NS, not significant; SCC, squamous cell carcinoma. T score: log-transformed reciprocal P-value. T score cutoff of 5.75 is used to identify DFRs (2s.d. from the mean of total T scores for all windows).

the consecutive stages of cervical SCC development. Further validating our results, 30 of these miRNAs showed concordant differential expression in previous studies investigating miRNA expression mainly in cervical carcinomas.18,20 - 23 The inclusion of CIN2 -- 3 lesions allowed us to discern distinct expression patterns in the total group of differentially expressed miRNAs. Twentyseven miRNAs showed differential expression only in CIN2 -- 3 compared with normal and were therefore designated as ‘early transient’. This wavy expression pattern may indicate temporal involvement in cervical cancer development, involvement in the defence mechanisms against this transformation process, or that changed expression of these miRNAs merely represents a consequence of the change in phenotype. Previous studies investigating cervical and bronchial precancerous lesions also encountered this phenomenon.23,33 Of 46 miRNAs, differential expression was restricted to SCCs, suggesting these miRNAs are potentially important for the progression of high-grade CIN lesions towards invasive carcinomas. Finally, 33 miRNAs showed concordant differential expression in both CIN2 -- 3 and SCCs. DIGMAP identified four chromosomal loci displaying differential miRNA expression between CIN2 -- 3/SCCs and normal epithelium (9q22.33, 13q31.3, 14q32.1 -- 32.32 and Xq26.2 -- 26.3). As DIGMAP uses a sliding window approach rendering the analysis more powerful, not all miRNAs located within the identified loci showed significantly differential expression at the single miRNA level. Except for the locus at chromosome 13, these loci are not frequently altered on the chromosomal level. However, it is likely that expression of clustered miRNAs (located within 10 kb of each & 2013 Macmillan Publishers Limited

other) is controlled in a similar manner due to the presence of shared regulatory sequences and the formation of polycistronic primary miRNAs. The identified locus on 13q contains a wellknown cluster of onco-miRs (17 -- 92 cluster). Expression of this cluster was found to be increased by DIGMAP analysis despite a frequent chromosomal loss of this locus in cervical cancer. As expression of this cluster of miRNAs was shown to be regulated by the E2F family of transcription factors,34 overexpression of this cluster in cervical cancer might be due to the presence of the viral oncoprotein E7, which induces continuous activation of E2F. A direct relation between chromosomal alterations and significantly differential miRNA expression was found for five miRNAs (hsa-miR-9 (1q23.2), hsa-miR-15b (3q25.32), hsa-miR-285p (3q27.3), hsa-miR-100 and hsa-miR-125b (both 11q24.1)) using an objective statistical method including multiple testing correction (IntCNGEAn). Decreased expression of hsa-miR-100 and hsamiR-125b related to a chromosomal loss of 11q has also been found in oral cancer and altered expression of these miRNAs was shown to be functionally involved in carcinogenesis in vitro.35 For hsa-miR-9 contradictory results have been reported, which may in part be related to the fact that hsa-miR-9 is present at three chromosomal locations in the human genome (at chromosomes 1, 5, and 15). Whereas all three loci give rise to the same mature miRNA sequence, the primary miRNA transcripts are different. Elevated copy numbers of the hsa-mir-9-1 locus (1q), as we found in cervical (pre)cancer, were also described in ovarian cancer, breast cancer and melanoma.27 In ovarian cancer, despite the frequent presence of elevated copy numbers for hsa-mir-9-1, Oncogene (2013) 106 - 116


112

Figure 3. Functional effects of ectopic hsa-miR-9 expression in early-passage FK16A cells. (a) Expression of hsa-miR-9 in early-passage FK16A cells transduced with either the control vector (FK16A early þ ctrl) or hsa-miR-9 (FK16A early þ miR-9). The ectopic expression level in FK16A early þ miR-9 was 25 times higher than the average expression level observed in SCCs in vivo. In (b) cell viability is shown in FK16A early þ ctrl cells and FK16A early þ miR-9 cells. (c) Representative pictures of colony formation in soft agar of FK16A early þ ctrl cells and FK16A early þ miR-9 cells. (d) Quantification of the number of colonies in soft agar formed by FK16A early þ ctrl cells and FK16A early þ miR-9. In (e) the migratory capacity of FK16A early þ ctrl cells and FK16A early þ miR-9 cells is shown and in (f ) invasion of FK16A early þ ctrl cells and FK16A early þ miR-9 cells.

decreased expression of hsa-miR-9 was found in a small set of ovarian cancers with unknown chromosomal status compared with matched normal and ectopic hsa-miR-9 expression repressed cell growth in vitro.36 In breast cancer, increased expression of the hsa-mir-9-1 primary transcript was found compared with normal breast, whereas a decrease in the expression of primary transcript hsa-mir-9-3 (15q) was found in breast tumours with lymph node metastasis and in tumours with vascular invasion compared with tumours without these characteristics.37 This latter finding is supported by the significant association between methylation at the hsa-mir-9-3 locus and lymph node metastasis in breast cancer, lung cancer and melanoma and increased risk of recurrence in clear-cell renal cell carcinomas.38,39 In addition, it was shown that epigenetic silencing of hsa-mir-9-3 reduced p53-mediated apoptosis in mammosphere-derived breast epithelial cells.40 Methylation of the hsa-mir-9-1 locus has been described in clear-cell renal cell carcinomas, breast cancer and pre-invasive intraductal lesions of the breast as well, but expression levels in these tumours were not related to the expression levels in corresponding normal tissue.38,41 In our study, increased expression of hsa-miR-9 related to a chromosomal gain of the hsa-mir-9-1 locus was found in cervical (pre)cancerous lesions compared with normal epithelium. To confirm the results of our integration analysis, we specifically determined the expression of the primary transcript hsa-mir-9-1 (derived from the locus on 1q) in four SCCs with a 1q gain and Oncogene (2013) 106 - 116

four SCCs without this chromosomal alteration and found a 2.4fold higher expression in the SCCs with a 1q gain (data not shown). Ectopic expression of hsa-miR-9 in early-passage HPV16immortalised cells with low levels of endogenous hsa-miR-9 resulted in an increase in the cell viability, anchorage-independent growth and migration. Vice versa, knockdown of hsa-miR-9 expression in a later passage of the same cell line, showing elevated endogenous hsa-miR-9 levels, decreased cell viability and anchorage-independent growth. Moreover, upon organic raft culturing hsa-miR-9 overexpression reduced epithelial differentiation, accompanied by an increased number of proliferating cells throughout all strata of the epithelium. These findings suggest a putative role for hsa-miR-9 in the switch from differentiation to proliferation during hrHPV-mediated transformation of squamous epithelial cells. Together, our results point to an oncogenic role of hsa-miR-9 in the context of hrHPV-induced malignant transformation. Interestingly, Hu et al.17 showed that relatively high levels of hsa-miR-9 expression in a group of cervical carcinoma patients was associated with a better overall survival rate, potentially indicating better treatment response due to high hsa-miR-9 expression. In this study, we found evidence that chromosomal alterations may be directly involved in the altered expression of part of the miRNAs associated with cervical carcinogenesis. However, recent literature suggests that other mechanisms, including epigenetic alterations, appear highly important as well. It is now estimated & 2013 Macmillan Publishers Limited


113

Figure 4. Functional effects of hsa-miR-9 knockdown in late-passage FK16A cells. (a) Expression of hsa-miR-9 in late-passage FK16A cells transfected with either negative control A (FK16A late þ ctrl) or antagomiR-9 (FK16A late þ anti-miR-9). After knockdown hsa-miR-9 expression levels in FK16A late þ anti-miR-9 were similar to those observed in normal epithelium in vivo. In (b) cell viability is shown in FK16A late þ ctrl cells and FK16A late þ anti-miR-9 cells. (c) Representative pictures of colony formation in soft agar of FK16A late þ ctrl cells and FK16A late þ anti-miR-9 cells. (d) Quantification of the number of colonies in soft agar formed by FK16A late þ ctrl cells and FK16A late þ anti-miR-9. In (e) the migratory capacity of FK16A late þ ctrl cells and FK16A late þ anti-miR-9 cells is shown and in (f ) invasion of FK16A late þ ctrl cells and FK16A late þ anti-miR-9 cells.

that half of all miRNA loci is associated with CpG islands,42 suggesting that DNA methylation may have a prominent role in (ab)normal miRNA expression. In this study nine out of the 43 (21%) significantly downregulated miRNAs were located within a CpG island. We found preliminary evidence of methylation-mediated silencing for three of these nine miRNAs, hsa-miR-203, hsa-miR-572 and hsa-miR-638 in a panel of hrHPV-transformed cell lines (data not shown). Promoter methylation of hsa-miR-203 has already been described in hepatocellular carcinoma, oral cancer, cervical cancer and haematopoietic malignancies,43 - 47 whereas, to the best of our knowledge, methylation-mediated silencing of hsa-miR-572 and hsa-miR-638 has not been previously described in cancer. Based on the foregoing information, further investigations regarding methylation-mediated silencing of miRNAs and its functional relevance during hrHPV-mediated transformation are warranted. In conclusion, this study showed that changes in miRNA expression in cervical (pre)malignant lesions are partly associated with chromosomal alterations. In addition, the data presented here provide valuable insights in the distinct expression profiles of miRNAs during the different stages of cervical cancer development. Integration of miRNA expression profiles with chromosomal alterations resulted in the identification of hsa-miR-9 as a candidate oncogene in cervical cancer, which may represent a promising disease marker, as well as a potential therapeutic target. Together, the findings of this study underline the importance of deregulated miRNA expression in cervical cancer development. & 2013 Macmillan Publishers Limited

MATERIALS AND METHODS Clinical tissue specimens and HPV testing For miRNA microarray analysis we used frozen specimens of 10 SCCs, 9 AdCAs, 18 CIN2 -- 3 and 10 cervical squamous epithelial samples with normal histology (Supplementary Table 1). Only CIN2 - 3 showing strong diffuse p16INK4A staining by immunohistochemistry were included, as p16INK4A overexpression is a marker for CIN2 - 3 lesions that harbour a transforming hrHPV infection and are considered the true precursor lesions of cervical cancer.48 Genome-wide chromosomal profiles of all SCC and CIN2 - 3 samples were previously generated using array-based comparative genomic hybridisation (array CGH).29,30 hrHPV testing on all specimens was performed using GP5 þ /6 þ -PCR with enzyme immunoassay read-out using a probe cocktail of 14 hrHPV types.49 Of the samples analysed by miRNA microarray, all samples except for one normal sample contained hrHPV (Supplementary Table 1). All biopsy samples were collected during the course of routine clinical practice at the Department of Obstetrics and Gynaecology at the VU University Medical Center (Amsterdam). This study followed the ethical guidelines of the Institutional Review Board of the VU University Medical Center.

Cell lines Establishment and culture of FK16A, an HPV16 immortalised keratinocyte cell line, was described previously.50 The human cervical carcinoma cell line SiHa was obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured as described previously.51 Oncogene (2013) 106 - 116


114

Figure 5. Organotypic raft cultures of FK16A early with and without ectopic hsa-miR-9 expression. Representative pictures of H&E staining are shown for the raft cultures of (a) parental early FK16A rafts, (b) FK16A early þ ctrl rafts, and (c) FK16A early þ miR-9 rafts. Immunohistochemical staining patterns for the differentiation marker cytokeratin 10 (KRT10) are shown in (d) parental early FK16A rafts, (e) FK16A early þ ctrl rafts and (f ) FK16A early þ miR-9 rafts. Similarly, protein expression patterns of the proliferation-related Ki-67 antigen (MKI67) are shown in (g) parental early FK16A rafts, (h) FK16A early þ ctrl rafts and (i) FK16A early þ miR-9 rafts.

Laser capture microdissection and RNA isolation Tissue samples were first subjected to laser capture microdissection using a Leica ASLMD microscope (Leica, Heidelberg, Germany) to enrich for neoplastic, dysplastic or normal epithelial cells as described previously.52 Total RNA was isolated from microdissected tissue samples using TRIzol Reagent (Life Technologies, Breda, The Netherlands) according to the manufacturer’s instructions. To ensure comparable RNA integrity between specimens analysed by array, RNA pico chips were performed on an Agilent 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany).

MiRNA microarrays Global miRNA expression profiles were determined using human miRNA microarrays (G4471A-016436, Agilent Technologies, Santa Clara, CA, USA). These arrays contain in situ synthesised 60-mer oligonucleotides representing 472 human miRNAs (MiRbase release 9.1). Labelling and hybridisation was performed according to the manufacturers’ instructions. MiRNA and CGH microarray data are available from the Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) through series accession numbers GSE11573 (CGH data CIN2 - 3 lesions), GSE6473 (CGH data SCCs) and GSE30656 (miRNA expression profiles of all samples).

Microarray analysis Data pre-processing. The obtained single-channel microarray data were log-transformed and normalised using quantile normalisation. Probes with log2 intensities below 6 (corresponding to intensity below 64) in more than 90% of the samples were discarded (n ¼ 44).

Cluster analysis. Unsupervised hierarchical clustering analysis was performed on all probes that met our quality criteria using the Manhattan distance measure and Ward’s linkage. Oncogene (2013) 106 - 116

Differential miRNA expression analysis. Differentially expressed miRNAs were determined using the Wilcoxon two-sample test. The test procedure includes a permutation-based FDR correction for multiple testing, needed to discriminate real differences from chance effects.53 An FDR below 0.05 was considered significant. DIGMAP analysis. Differential gene locus mapping (DIGMAP) analysis was carried out in the CIN2 - 3 and SCCs to investigate aberrant expression patterns of groups of miRNAs based on their chromosomal location as previously described.52,54 With this technique, changes in expression associated with the chromosomal location can be detected and related to chromosomal alterations, as well as other mechanisms, such as shared regulatory sequences. Statistically significant loci were identified using the T-test, which generates confidence (T) scores by log-transformation of the reciprocal P-value. An automated scanning method employing sliding window analysis was used to find so-called differential flag regions. In this analysis, a sliding window size of 10 was used and regions with a T score greater than two standard deviations from the mean of total T scores for all windows were considered as differentially flagged regions (DFR). Integrated copy number and gene expression analysis (IntCNGEAn). MiRNA expression levels were directly linked to the underlying chromosomal copy number of the miRNA as measured by array CGH using integrated copy number and gene expression analysis (IntCNGEAn).32 miRNAs and BAC clones were positioned along the genome using the GRCh 37 freeze. Matching of copy number data and miRNA expression data was performed using the method ‘distance’, in which the midpoint of the copy number and miRNA expression features are calculated and for each miRNA the closest copy number feature (BAC clone) is selected. Only miRNAs that are located on a chromosomal location that was altered in at least 30% of SCCs and CIN2 -- 3 together were tested (81 miRNAs). The & 2013 Macmillan Publishers Limited


115 statistical test performed was a one-tailed weighted Mann -- Whitney test. P-values are calculated via permutation. The P-values are corrected for multiplicity by the Benjamini - Hochberg procedure. miRNAs with an FDR below 0.2 and concordant significant differential expression between CIN2 - 3 and/or SCCs and normal epithelium were considered significant.

Quantitative reverse transcription--PCR Expression of hsa-miR-9, hsa-miR-15b, hsa-miR-21, hsa-miR-28-5p, hsa-miR100, hsa-miR-125b, hsa-miR-203 and hsa-miR-375 was measured using TaqMan miRNA assays following the manufacturer’s instructions (000583, 00390, 000397, 00411, 00437, 00449, 000507, 000564; Applied Biosystems, Nieuwerkerk a/d IJssel, The Netherlands) on the ABI 7500 Fast Real-Time PCR System (Applied Biosystems). The small nucleolar RNA transcript RNU43 was included as a reference gene (001095; Applied Biosystems). miRNA expression values were normalised to the reference using the comparative CT method (2DCT).55

Slides were incubated with 0.3% H2O2 in methanol to inhibit endogenous peroxidase. Antigen retrieval was performed by boiling the slides in a microwave oven either in citrate buffer (10 mM, pH 6.0) for cytokeratin 10 (LHP1, dilution 1:100, Novocastra, Leica Biosystems, Newcastle, UK) or Tris/ EDTA buffer (10 mM, pH9) for MIB1 (MIB-1, dilution 1:40, DakoCytomation, Glostrup, Denmark). Slides were incubated with the monoclonal antibody for 1 h at room temperature. For antibody detection the EnVision horseradish peroxidase system (Dako, Glostrup, Denmark) was used. Sections were counterstained with haematoxylin.

Statistical analysis The non-parametric Spearman’s rank correlation was determined between array and qRT--PCR results. Expression levels of miRNAs as measured by qRT--PCR were compared between samples with and without the associated chromosomal alteration using the non-parametric two-sample Wilcoxon test (one-tailed). Proliferation rates between cell lines were compared using the Student’s t-test.

Retroviral transduction For the transduction experiments, early-passage FK16A cells (p61) and SiHa cells were incubated for 4 and 16 h, respectively, at 37 1C with filtered viral supernatants supplemented with polybrene (15 mg/ml).56 FK16A/SiHa_miR9 and FK16A/SiHa_ctrl cells were selected by continuous culturing of the transduced cells in the presence of blasticidin (2 mg/ml) and ectopic overexpression was verified by qRT--PCR analysis.

AntagomiR transfection Late-passage FK16A cells (Bp200) were transiently transfected with 25 nM mirCURY LNA microRNA power Inhibitor for hsa-miR-9 or negative control A (427460-04 and 199020-04, respectively, Exiqon, Vedbaek, Denmark), using Lipofectamin 2000 Reagent according to the manufacturer’s instructions. Knockdown efficacy was verified by qRT--PCR analysis.

CONFLICT OF INTEREST Dr RDM Steenbergen, Professor Dr PJF Snijders and Professor Dr CJLM Meijer are stockholders of Self-screen BV, The Netherlands. All authors declare no conflict of interest.

ACKNOWLEDGEMENTS This work was supported by the V-ICI institute of the VU University Medical Center, Amsterdam, The Netherlands (grant number CCA20085-04) and the Dutch Cancer Society (KWF, grant number VU2010-4668). We are grateful to Marlon van der Plas, Suzanne Snellenberg, Nour Makazaji and Tim Schutte for excellent technical assistance.

Cell viability assay

REFERENCES

Cell viability was measured using a colorimetric (MTT - tetrazolium) assay (MP Biomedicals, Illkrich, France) as described before.57,58 The proliferation rate was determined by subtracting the measurement of day 0 from all other time points.

1 Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 2002; 55: 244 - 265. 2 Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999; 189: 12 - 19. 3 Fu Y, Reagan J. Pathology of the Uterine Cervix, Vagina and Vulva. Major Problems in Pathology. W.B. Saunders and Co.: Philadelphia, 1989, pp 288 - 335. 4 Pisani P, Bray F, Parkin DM. Estimates of the world-wide prevalence of cancer for 25 sites in the adult population. Int J Cancer 2002; 97: 72 - 81. 5 McCredie MR, Sharples KJ, Paul C, Baranyai J, Medley G, Jones RW et al. Natural history of cervical neoplasia and risk of invasive cancer in women with cervical intraepithelial neoplasia 3: a retrospective cohort study. Lancet Oncol 2008; 9: 425 - 434. 6 Wallin KL, Wiklund F, Angstrom T, Bergman F, Stendahl U, Wadell G et al. Typespecific persistence of human papillomavirus DNA before the development of invasive cervical cancer. N Engl J Med 1999; 341: 1633 - 1638. 7 Winer RL, Kiviat NB, Hughes JP, Adam DE, Lee SK, Kuypers JM et al. Development and duration of human papillomavirus lesions, after initial infection. J Infect Dis 2005; 191: 731 - 738. 8 Zielinski GD, Snijders PJ, Rozendaal L, Voorhorst FJ, van der Linden HC, Runsink AP et al. HPV presence precedes abnormal cytology in women developing cervical cancer and signals false negative smears. Br J Cancer 2001; 85: 398 - 404. 9 Li B, Hu Y, Ye F, Li Y, Lv W, Xie X. Reduced miR-34a expression in normal cervical tissues and cervical lesions with high-risk human papillomavirus infection. Int J Gynecol Cancer 2010; 20: 597 - 604. 10 Martinez I, Gardiner AS, Board KF, Monzon FA, Edwards RP, Khan SA. Human papillomavirus type 16 reduces the expression of microRNA-218 in cervical carcinoma cells. Oncogene 2008; 27: 2575 - 2582. 11 McKenna DJ, McDade SS, Patel D, McCance DJ. MicroRNA 203 expression in keratinocytes is dependent on regulation of p53 levels by E6. J Virol 2010; 84: 10644 - 10652. 12 Melar-New M, Laimins LA. Human papillomaviruses modulate expression of microRNA 203 upon epithelial differentiation to control levels of p63 proteins. J Virol 2010; 84: 5212 - 5221. 13 Nuovo GJ, Wu X, Volinia S, Yan F, di LG, Chin N et al. Strong inverse correlation between microRNA-125b and human papillomavirus DNA in productive infection. Diagn Mol Pathol 2010; 19: 135 - 143.

Colony formation in soft agar Anchorage-independent cell growth was analysed as described before.59 Colonies were photographed and counted after 3 weeks of incubation.

Cellular migration and invasion Migration and invasion were determined using transwells containing a fluorescence blocking filter (HTS FluoroBlok; Falcon, BD Biosciences, Breda, The Netherlands). For migration assays the transwell was left uncoated, whereas for invasion assays the transwell was coated with collagen (50 mg/ ml solution in PBS; Collagen Type I rat tail, BD Biosciences). In all cases the bottom of the transwell was coated with fibronectin (2 mg/ml solution in PBS; MP Biomedicals). Cells in medium without supplements were added to the top compartment, whereas complete medium was added to the bottom compartment of the transwell. Migration or invasion was quantified after 48 h by fluorescent labelling of the cells using calceinAM (Molecular Probes, Invitrogen, Leiden, The Netherlands). To determine the relative amount of cells that either migrated or invaded through the transwell, ratios between the fluorescence intensities of the bottom and top compartment were calculated.

Organotypic raft cultures and immunohistochemistry FK16A cells were grown as epithelial raft tissues as described previously.60 For all conditions duplicate rafts were developed. After 8 days the raft tissues were fixed in formalin and embedded in paraffin. For histological examination, 4 mM sections were stained with haematoxylin and eosin. In addition, immunohistochemical staining was performed for differentiation marker cytokeratin 10 and proliferation-related Ki-67 antigen (MKI67). & 2013 Macmillan Publishers Limited

Oncogene (2013) 106 - 116


116 14 Wang X, Wang HK, McCoy JP, Banerjee NS, Rader JS, Broker TR et al. Oncogenic HPV infection interrupts the expression of tumor-suppressive miR-34a through viral oncoprotein E6. RNA 2009; 15: 637 - 647. 15 Muralidhar B, Goldstein LD, Ng G, Winder DM, Palmer RD, Gooding EL et al. Global microRNA profiles in cervical squamous cell carcinoma depend on Drosha expression levels. J Pathol 2007; 212: 368 - 377. 16 Muralidhar B, Winder D, Murray M, Palmer R, Barbosa-Morais N, Saini H et al. Functional evidence that Drosha overexpression in cervical squamous cell carcinoma affects cell phenotype and microRNA profiles. J Pathol 2011; 224: 496 - 507. 17 Hu X, Schwarz JK, Lewis Jr JS, Huettner PC, Rader JS, Deasy JO et al. A microRNA expression signature for cervical cancer prognosis. Cancer Res 2010; 70: 1441 - 1448. 18 Lee JW, Choi CH, Choi JJ, Park YA, Kim SJ, Hwang SY et al. Altered MicroRNA expression in cervical carcinomas. Clin Cancer Res 2008; 14: 2535 - 2542. 19 Lui WO, Pourmand N, Patterson BK, Fire A. Patterns of known and novel small RNAs in human cervical cancer. Cancer Res 2007; 67: 6031 - 6043. 20 Rao Q, Zhou H, Peng Y, Li J, Lin Z. Aberrant microRNA expression in human cervical carcinomas. Med Oncol 2011 (doi:10.1007/s12032-011-9830-2). 21 Wang X, Tang S, Le SY, Lu R, Rader JS, Meyers C et al. Aberrant expression of oncogenic and tumor-suppressive microRNAs in cervical cancer is required for cancer cell growth. PLoS One 2008; 3: e2557. 22 Li Y, Wang F, Xu J, Ye F, Shen Y, Zhou J et al. Progressive miRNA expression profiles in cervical carcinogenesis and identification of HPV-related target genes for miR-29. J Pathol 2011; 224: 484 - 495. 23 Pereira PM, Marques JP, Soares AR, Carreto L, Santos MA. MicroRNA expression variability in human cervical tissues. PLoS One 2010; 5: e11780. 24 Calin GA, Sevignani C, Dumitru CD, Hyslop T, Noch E, Yendamuri S et al. Human microRNA genes are frequently located at fragile sites and genomic regions involved in cancers. Proc Natl Acad Sci U S A 2004; 101: 2999 - 3004. 25 Bray I, Bryan K, Prenter S, Buckley PG, Foley NH, Murphy DM et al. Widespread dysregulation of MiRNAs by MYCN amplification and chromosomal imbalances in neuroblastoma: association of miRNA expression with survival. PLoS One 2009; 4: e7850. 26 Lionetti M, Agnelli L, Mosca L, Fabris S, Andronache A, Todoerti K et al. Integrative high-resolution microarray analysis of human myeloma cell lines reveals deregulated miRNA expression associated with allelic imbalances and gene expression profiles. Genes Chromosomes Cancer 2009; 48: 521 - 531. 27 Zhang L, Huang J, Yang N, Greshock J, Megraw MS, Giannakakis A et al. microRNAs exhibit high frequency genomic alterations in human cancer. Proc Natl Acad Sci U S A 2006; 103: 9136 - 9141. 28 Zhang L, Volinia S, Bonome T, Calin GA, Greshock J, Yang N et al. Genomic and epigenetic alterations deregulate microRNA expression in human epithelial ovarian cancer. Proc Natl Acad Sci U S A 2008; 105: 7004 - 7009. 29 Wilting SM, Steenbergen RD, Tijssen M, van Wieringen WN, Helmerhorst TJ, van Kemenade FJ et al. Chromosomal signatures of a subset of high-grade premalignant cervical lesions closely resemble invasive carcinomas. Cancer Res 2009; 69: 647 - 655. 30 Wilting SM, Snijders PJ, Meijer GA, Ylstra B, van den IJssel PR, Snijders AM et al. Increased gene copy numbers at chromosome 20q are frequent in both squamous cell carcinomas and adenocarcinomas of the cervix. J Pathol 2006; 209: 220 - 230. 31 Tase T, Okagaki T, Clark BA, Twiggs LB, Ostrow RS, Faras AJ. Human papillomavirus DNA in adenocarcinoma in situ, microinvasive adenocarcinoma of the uterine cervix, and coexisting cervical squamous intraepithelial neoplasia. Int J Gynecol Pathol 1989; 8: 8 - 17. 32 van Wieringen WN, van de Wiel MA. Nonparametric testing for DNA copy number induced differential mRNA gene expression. Biometrics 2009; 65: 19 - 29. 33 Mascaux C, Laes JF, Anthoine G, Haller A, Ninane V, Burny A et al. Evolution of microRNA expression during human bronchial squamous carcinogenesis. Eur Respir J 2009; 33: 352 - 359. 34 Woods K, Thomson JM, Hammond SM. Direct regulation of an oncogenic microRNA cluster by E2F transcription factors. J Biol Chem 2007; 282: 2130 - 2134. 35 Henson BJ, Bhattacharjee S, O0 Dee DM, Feingold E, Gollin SM. Decreased expression of miR-125b and miR-100 in oral cancer cells contributes to malignancy. Genes Chromosomes Cancer 2009; 48: 569 - 582. 36 Guo LM, Pu Y, Han Z, Liu T, Li YX, Liu M et al. MicroRNA-9 inhibits ovarian cancer cell growth through regulation of NF-kappaB1. FEBS J 2009; 276: 5537 - 5546. 37 Iorio MV, Ferracin M, Liu CG, Veronese A, Spizzo R, Sabbioni S et al. MicroRNA gene expression deregulation in human breast cancer. Cancer Res 2005; 65: 7065 - 7070.

38 Hildebrandt MA, Gu J, Lin J, Ye Y, Tan W, Tamboli P et al. Hsa-miR-9 methylation status is associated with cancer development and metastatic recurrence in patients with clear cell renal cell carcinoma. Oncogene 2010; 29: 5724 - 5728. 39 Lujambio A, Calin GA, Villanueva A, Ropero S, Sanchez-Cespedes M, Blanco D et al. A microRNA DNA methylation signature for human cancer metastasis. Proc Natl Acad Sci U S A 2008; 105: 13556 - 13561. 40 Hsu PY, Deatherage DE, Rodriguez BA, Liyanarachchi S, Weng YI, Zuo T et al. Xenoestrogen-induced epigenetic repression of microRNA-9-3 in breast epithelial cells. Cancer Res 2009; 69: 5936 - 5945. 41 Lehmann U, Hasemeier B, Christgen M, Muller M, Romermann D, Langer F et al. Epigenetic inactivation of microRNA gene hsa-mir-9-1 in human breast cancer. J Pathol 2008; 214: 17 - 24. 42 Weber B, Stresemann C, Brueckner B, Lyko F. Methylation of human microRNA genes in normal and neoplastic cells. Cell Cycle 2007; 6: 1001 - 1005. 43 Bueno MJ, Perez de CI, Gomez de CM, Santos J, Calin GA, Cigudosa JC et al. Genetic and epigenetic silencing of microRNA-203 enhances ABL1 and BCR-ABL1 oncogene expression. Cancer Cell 2008; 13: 496 - 506. 44 Furuta M, Kozaki KI, Tanaka S, Arii S, Imoto I, Inazawa J. miR-124 and miR-203 are epigenetically silenced tumor-suppressive microRNAs in hepatocellular carcinoma. Carcinogenesis 2010; 31: 766 - 776. 45 Kozaki K, Imoto I, Mogi S, Omura K, Inazawa J. Exploration of tumor-suppressive microRNAs silenced by DNA hypermethylation in oral cancer. Cancer Res 2008; 68: 2094 - 2105. 46 Sang CC, Yeung WK, Ying LC, Ping CL, Kwan HP, Yan CS et al. Epigenetic inactivation of the hsa-miR-203 in haematological malignancies. J Cell Mol Med 2011; 15: 2760 - 2767. 47 Botezatu A, Goia-Rusanu CD, Iancu IV, Huica I, Plesa A, Socolov D et al. Quantitative analysis of the relationship between microRNA124a, -34b and -203 gene methylation and cervical oncogenesis. Mol Med Report 2011; 4: 121 - 128. 48 Wentzensen N, von Knebel DM. Biomarkers in cervical cancer screening. Dis Markers 2007; 23: 315 - 330. 49 Jacobs MV, Snijders PJ, van den Brule AJ, Helmerhorst TJ, Meijer CJ, Walboomers JM. A general primer GP5+/GP6(+)-mediated PCR-enzyme immunoassay method for rapid detection of 14 high-risk and 6 low-risk human papillomavirus genotypes in cervical scrapings. J Clin Microbiol 1997; 35: 791 - 795. 50 Steenbergen RD, Walboomers JM, Meijer CJ, van der Raaij-Helmer EM, Parker JN, Chow LT et al. Transition of human papillomavirus type 16 and 18 transfected human foreskin keratinocytes towards immortality: activation of telomerase and allele losses at 3p, 10p, 11q and/or 18q. Oncogene 1996; 13: 1249 - 1257. 51 Steenbergen RD, Kramer D, Meijer CJ, Walboomers JM, Trott DA, Cuthbert AP et al. Telomerase suppression by chromosome 6 in a human papillomavirus type 16immortalized keratinocyte cell line and in a cervical cancer cell line. J Natl Cancer Inst 2001; 93: 865 - 872. 52 Wilting SM, de WJ, Meijer CJ, Berkhof J, Yi Y, van Wieringen WN et al. Integrated genomic and transcriptional profiling identifies chromosomal loci with altered gene expression in cervical cancer. Genes Chromosomes Cancer 2008; 47: 890 - 905. 53 Manduchi E, Grant GR, McKenzie SE, Overton GC, Surrey S, Stoeckert Jr CJ. Generation of patterns from gene expression data by assigning confidence to differentially expressed genes. Bioinformatics 2000; 16: 685 - 698. 54 Yi Y, Mirosevich J, Shyr Y, Matusik R, George Jr AL. Coupled analysis of gene expression and chromosomal location. Genomics 2005; 85: 401 - 412. 55 Schmittgen TD, Livak KJ. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc 2008; 3: 1101 - 1108. 56 Voorhoeve PM, le Sage C, Schrier M, Gillis AJ, Stoop H, Nagel R et al. A genetic screen implicates miRNA-372 and miRNA-373 as oncogenes in testicular germ cell tumors. Cell 2006; 124: 1169 - 1181. 57 Overmeer RM, Henken FE, Bierkens M, Wilting SM, Timmerman I, Meijer CJ et al. Repression of MAL tumor suppressor activity by promoter methylation during cervical carcinogenesis. J Pathol 2009; 219: 327 - 336. 58 Wilting SM, van Boerdonk RA, Henken FE, Meijer CJ, Diosdado B, Meijer GA et al. Methylation-mediated silencing and tumour suppressive function of hsa-miR-124 in cervical cancer. Mol Cancer 2010; 9: 167. 59 Steenbergen RD, Kramer D, Braakhuis BJ, Stern PL, Verheijen RH, Meijer CJ et al. TSLC1 gene silencing in cervical cancer cell lines and cervical neoplasia. J Natl Cancer Inst 2004; 96: 294 - 305. 60 Steenbergen RD, Parker JN, Isern S, Snijders PJ, Walboomers JM, Meijer CJ et al. Viral E6-E7 transcription in the basal layer of organotypic cultures without apparent p21cip1 protein precedes immortalization of human papillomavirus type 16- and 18-transfected human keratinocytes. J Virol 1998; 72: 749 - 757.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene (2013) 106 - 116
