Background
Cervical cancer is the second most common cancer in women worldwide and is caused by a persistent infection with high-risk types of the human papillomavirus (hrHPV) [
1‐
3]. The development of cervical squamous cell carcinomas (SCCs, representing about 80% of cases) occurs via well-recognisable premalignant precursor lesions (cervical intraepithelial neoplasia (CIN), graded 1-3), whereas less is known about the different precursor stages preceding cervical adenocarcinomas (AdCAs, accounting for 10-20% of cases).
Even taking the promising results of recently introduced prophylactic HPV vaccines into account, cervical screening will remain necessary in the foreseeable future [
4‐
7]. Recent studies have shown that the sensitivity of hrHPV testing is superior to that of cytology as a screening tool [
8‐
10]. However, hrHPV testing also results in the identification of a considerable number of hrHPV-positive women without (pre)cancerous lesions, necessitating the development of proper triage tests for hrHPV-positive women. Assays detecting (epi)genetic changes that besides hrHPV are crucial for malignant progression will likely contribute to the discrimination of hrHPV positive women with cervical (pre)cancer.
DNA methylation-mediated silencing of an increasing number of protein-coding tumour suppressor genes is known to be involved in cervical cancer. Therefore, the present study aimed to investigate whether epigenetic changes relevant in hrHPV-mediated cervical carcinogenesis may affect the expression microRNAs (miRNAs) as well [
11‐
16]. miRNAs, ~23 nucleotide long, non-coding RNAs, regulate expression of protein-coding genes at the posttranscriptional level by sequence specific base pairing in the 3' untranslated region (UTR) of the target mRNAs. Recent proteomic studies have shown that a single miRNA can regulate expression of hundreds of targets [
17,
18]. The potential importance of miRNAs in cervical carcinogenesis in general, is underlined by a number of studies. miRNA loci are significantly associated with fragile sites, which are known insertion sites of HPV in cervical cancers. In addition, even though no HPV-encoded miRNAs have been identified so far, HPV-encoded genes were shown to influence the miRNA expression of its host cell [
19,
20]. Altered miRNA expression was found in cervical cancer cell lines and/or cervical carcinomas compared to normal controls and a number of these differentially expressed miRNAs were shown to influence proliferation rates of cervical cancer cell lines SiHa or HeLa [
20‐
24].
Interestingly, similar to protein-coding tumour suppressor genes, expression of a substantial number of miRNAs was shown to be under epigenetic regulation [
25‐
31]. A well-known epigenetically silenced miRNA in human carcinogenesis is
hsa-miR-124. DNA methylation of
hsa-miR-124 was first shown by Lujambio
et al in colon, breast and lung cancer, as well as in leukaemia and lymphoma [
29]. Subsequent studies confirmed frequent
hsa-miR-124 methylation in leukaemia affecting clinical outcome and additionally showed frequent
hsa-miR-124 methylation in gastric cancer and hepatocellular carcinoma [
25,
26,
30,
32]. At present, no studies have been performed to investigate the role of epigenetic silencing of miRNAs in cervical cancer.
In this study we evaluated the potential role of DNA methylation-based silencing of
hsa-miR-124 during cervical carcinogenesis. The mature
hsa-miR-124 sequence is processed from 3 separate premature sequences, located at chromosomes 8p23.1 (miR-124-1), 8q12.3 (miR-124-2) and 20q13.33 (miR-124-3), all of which contain CpG islands in their promoter regions. We investigated the methylation status of all 3 genomic loci encoding the mature
hsa-miR-124 in cervical cancer cell lines, a longitudinal
in vitro model system of hrHPV-induced carcinogenesis [
33], cervical tissue specimens (n = 139), and hrHPV-positive cervical scrapes (n = 43) of women with and without a CIN3 diagnosis in follow-up. In addition, effects of (ectopic)
hsa-miR-124 expression on cellular proliferation, migration and mRNA expression of
IGFBP7, a potential target gene, were studied.
Materials and methods
Cell lines and cell culture
Establishment and culture of the HPV16 (FK16A/FK16B) and HPV18 (FK18A/FK18B) immortalised cell lines have been described previously [
33]. Primary human keratinocytes, referred to as EK cells, were isolated from foreskin and cultured as described previously [
33]. The human cervical carcinoma cell lines SiHa, CaSki and HeLa were obtained from the American Type Culture Collection (Manassas, VA, USA). SiHa cells were treated with 5000 nM 5-aza-2'-deoxycytidine (DAC, Sigma Chemical Co., St. Louis, MO, USA) dissolved in PBS to analyse the effect of global methylation inhibition on
hsa-miR-124 expression.
Clinical tissue specimens
We used frozen specimens of normal cervix (n = 5), CIN2/3 (n = 7), SCC (n = 9) and AdCA (n = 5) as well as formalin-fixed, paraffin-embedded (FFPE) biopsy specimens of normal cervix (n = 18), CIN1 (n = 36), CIN3 (n = 41), SCC (n = 29) and AdCA (n = 15). All specimens were collected during the course of routine clinical practice and stored at the Department of Pathology at the VU University Medical Center (Amsterdam, the Netherlands). Normal specimens were obtained from non-cancer patients undergoing hysterectomy. The mean age of all women included in this study was 40.3 years (range 18-79). Per histological subgroup the women had the following mean ages: 49.1 years (range 34-70) in the normal group; 35.9 (range 22-52) years in the CIN1 group; 35.9 years (range 27-74) in the CIN3 group; 53.0 years (range 35-61) in the SCC group; 45.6 years (range 28-79) in the AdCA group. The mean age in any of the groups of women with cervical (pre)malignant disease was not significantly higher than that of the women with normal histology.
Cervical scrapings were obtained from the population-based cervical screening trial POBASCAM, registered as an International Standard Randomized Controlled Trial under number ISRCTN20781131 [
8,
34]. For this study, we selected 22 cervical scrapes from hrHPV-positive women who had normal cytology without evidence of CIN disease up to the next screening round (i.e. 5 years) and 21 scrapings classified as severe dyskaryosis or worse from hrHPV-positive women who had a CIN3 diagnosis within 18 months of follow-up. The mean age of the women with normal cytology without CIN disease was 33.2 years (range 18-53), and that of women with abnormal cytology with CIN3 was 35.14 years (range 25-55). This study followed the ethical guidelines of the Institutional Review Board of the VU University Medical Center.
For methylation analysis, DNA was extracted from FFPE specimens by proteinase K digestion and purified using the High Pure PCR Template Preparation Kit (Roche Diagnostics, Almere, The Netherlands) following the manufacturer's recommendations. Genomic DNA from cell lines and frozen specimens was extracted by proteinase K digestion followed by standard phenol-chloroform extraction as described previously [
35].
For
hsa-miR-124 expression analysis, frozen specimens of normal cervix, CIN2/3 lesions, SCCs, and AdCAs were first enriched for epithelial cells by means of laser capture microdissection using a Leica ASLMD microscope (Leica, Heidelberg, Germany) as described before [
36]. Subsequently, total RNA was isolated from these samples and cell lines using TRIzol reagent (Life Technologies, Breda, The Netherlands), according to the manufacturer's instructions.
HPV typing of clinical specimens was performed using the general primer GP5+/6+ PCR, followed by reverse line blot, as described previously [
37,
38] (Additional file
1).
DNA modification and quantitative methylation-specific PCR (qMSP) analysis
The DNA methylation status of the CpG-island containing promoter regions associated with the three genomic loci encoding
hsa-miR-124 (hsa-miR-124-1, hsa-miR-124-2 and hsa-miR-124-3) was determined by qMSP analysis on sodium bisulfite-treated genomic DNA from cell lines, tissue specimens and scrapings. In brief, genomic DNA was modified using the EZ DNA Methylation kit (Zymo Research, Orange, CA, USA), which induces chemical conversion of unmethylated cytosines into uracils, whereas methylated cytosines are protected from this conversion. Specific primers were designed to amplify the methylated DNA sequence of all 3 promoter regions. Amplicons (hsa-miR-124-1: -191 to -97; hsa-miR-124-2: -301 to -163; hsa-miR-124-3: -106 to -11 relative to the transcription start site, respectively) were detected and quantified using TaqMan probes (Table
1). In addition, the modified, unmethylated sequence of the housekeeping gene β-actin (ACTB) was amplified as a reference [
39]. qMSP reactions were carried out in a 12 μl reaction volume containing 50 ng of bisulfite-treated DNA, 417 nM of each primer, 208 nM probe and 1× QuantiTect Probe PCR Kit master mix (Qiagen, Westburg, Leusden, The Netherlands) using the ABI 7500 Fast Real-Time PCR System (Applied Biosystems, Nieuwerkerk a/d IJssel, The Netherlands).
Table 1
Sequences of qMSP and qRT-PCR primers used in this study
hsa-miR-124-1 | F: CGGCGGGGAGGATGTT | | 58.9 |
| R: ATAAAAAACGACGCGTATACGTACG | 94 | 59.4 |
| P: CGGCGTTTTTTATTTTT-Xsprobe | | 70.0 |
hsa-miR-124-2 | F: GGGTAATTAATTTGGATTTACGTCGTTAT | | 59.9 |
| R: CGTAAAAATATAAACGATACGTATACCTACGT | 138 | 58.8 |
| P: TTTACAACACACGCCTAAA -Xsprobe | | 69.0 |
hsa-miR-124-3 | F: ACGCGGCGAAGACGTTT | | 59.0 |
| R: CGAACGACGAACGTCGAAA | 95 | 59.4 |
| P: AAAATCCTCGCCCGAAAAACGCGA | | 70.4 |
IGFBP7 | F: CCCAGGTGTACTTGAGCTGTGA | 89 | 58.9 |
| R: TGAACTCCATAGTGACCCCTTTTT | | 58.9 |
snRNP U1A | F: TCCTCACCAACCTGCCAGA | 71 | 58.8 |
| R: TGAAGCCAGGGAACTGATTGA | | 59.3 |
All PCR experiments were performed in duplicate (delta Ct ≤1.5 between replicates) and mean values were used for calculations. Methylation values of the 3 target regions were normalised to the reference gene ACTB using the comparative Ct method (2
-ΔCT) [
40]. All methylation negative samples in our study had a Ct for ACTB below 32, indicating sufficient DNA quality and thereby excluding false negative results. To ensure the detection of distinguishable increases in methylation level in (pre)malignant cervical lesions over normal cervical controls, we used the 99% confidence interval of the methylation levels obtained in normal cervical controls as cut-off value. Samples above this threshold were considered positive for methylation. For tissue specimens and scrapings separate cut-off values were determined using the appropriate normal controls.
Retroviral transduction
Retroviral
hsa-miR-124 or empty vector (ctrl) constructs previously described by Voorhoeve
et al[
41] were transfected into the Phoenix A retrovirus producer cell line and supernatants containing the replication-deficient
hsa-miR-124-expressing retrovirus or empty vector retrovirus were harvested 48 hours post-transfection. For the transduction experiments, SiHa and CaSki cells were incubated for 16 hours at 37°C with filtered viral supernatants supplemented with polybrene (15 μg/ml). SiHa/CaSki_miR-124 cells or SiHa/CaSki_ctrl cells were selected by continuous culturing of the transduced cells in the presence of blasticidin (3 μg/ml).
Quantitative Reverse Transcription-PCR (qRT-PCR)
Expression of hsa-miR-124 was measured using TaqMan microRNA assays following the manufacturer's instructions (00046 and 001182; Applied Biosystems) on the ABI 7500 Fast Real-Time PCR System (Applied Biosystems). The small nucleolar RNA transcript RNU43 was included as internal reference for hsa-miR-124 expression (001095; Applied Biosystems). Hsa-miR-124 expression values were normalised to the reference again by using the comparative Ct method as described above.
Intron-flanking primers for
IGFBP7, a potential target gene of
hsa-miR-124, were selected using Primer Express 3.0 (Applied Biosystems) (Table
1). Total RNA was reverse transcribed using the specific reverse primer and the resulting cDNA was used for real-time PCR. cDNA corresponding to 25 ng of total RNA was amplified in a total reaction volume of 25 μl containing 12.5 μl 2x Sybr Green master mix (Perkin Elmer/Applied Biosystems) and 0.5 μM primers. The house keeping gene U1 small nuclear ribonucleoprotein specific A protein (
snRNP U1A) was included as internal reference (Table
1). Expression values of
IGFBP7 were normalized to this reference by using the comparative Ct method as described above.
Cellular proliferation and migration assays
Cell proliferation was measured using a colorimetric (MTT-tetrazolium) assay (ICN Biomedicals Inc, OH, USA). In this assay the amount of dye conversion, as measured by the optical density at a wavelength of 540 nm, is directly related to the number of viable cells in each well. In brief, 5000 cells (CaSki) or 10000 cells (SiHa) were seeded in triplicate in 96-well plates and assayed for MTT conversion at day 0, day 1, day 2 and day 5 for CaSki and day 0, day 2, day 3 and day 6 for SiHa. The proliferation rate was determined by subtracting the measurement of day 0 from all other time points.
Cellular migration in vitro was determined using a so-called wound-healing assay. Cells were grown to confluency in 24-well plates and a single linear scratch was made in duplicate for all conditions using a sterile tip, resulting in a cell-free zone. Photographs of the scratch were taken immediately post-scratching and following 24 hours of incubation at 37°C. After 48 hours cells were fixed with methanol and stained with crystal violet solution.
Statistical analysis
Proliferation rates between hsa-miR-124-expressing cells and control cells were compared using the Student's t test. The frequency of methylation between normal and low-grade (CIN1) lesions on one hand and high-grade (CIN3) lesions and SCCs on the other hand were compared using χ2-testing. The difference in hsa-miR-124 expression between normal cervical epithelium and CIN2/3 lesions and carcinomas was compared using the Wilcoxon rank test. Linear (Pearson) correlation was determined between hsa-miR-124 methylation levels and hsa-miR-124 expression.
Discussion
In this study we showed that epigenetic silencing of hsa-miR-124 is functionally involved in cervical carcinogenesis and may provide a valuable marker for risk stratification of hrHPV-positive women. Using qMSP analysis, we found methylation of hsa-miR-124-1 and/or hsa-miR-124-2 in none of the normal tissues, 58.5% of CIN3 lesions, 93.1% of SCCs and 93.3% of AdCAs. Increased methylation levels of hsa-miR-124-1 and hsa-miR-124-2 in cervical tissue specimens were significantly correlated with lower hsa-miR-124 expression levels. Analysis of cervical scrapes showed that only 4.5% of hrHPV-positive scrapes without CIN disease in follow-up was positive compared to 71.4% of hrHPV-positive scrapes with CIN3 in follow-up. To the best of our knowledge this study provides the first evidence of DNA methylation-based silencing of a miRNA in cervical cancer.
Methylation of
hsa-miR-124 was found in cervical cancer cell lines SiHa, CaSki and HeLa as well as in late passages of HPV16/18 immortalised keratinocytes, reminiscent of high grade cervical precursor lesions, but not in normal primary keratinocytes. The fact that methylation of
hsa-miR-124 and concomitant reduced
hsa-miR-124 expression was found in late passages of HPV-immortalised keratinocytes but not in early passages, indicates that this event takes place post-immortalisation and is not directly related to the presence of hrHPV. Ectopic expression of
hsa-miR-124 decreased the proliferation rate of both SiHa and CaSki cells and also inhibited the migratory capacity of SiHa cells. Effects on the migratory capacity of CaSki cells were difficult to ascertain due to the specific growth characteristics of this cell line. Consistent with our findings, both Agirre
et al and Furuta
et al observed inhibitory effects on cellular growth upon reintroduction of
hsa-miR-124 expression in acute lymphoblastic leukaemia (ALL)-derived cells and hepatocellular carcinoma cell lines [
26,
32].
One of the previously identified targets mediating the tumour suppressive function of
hsa-miR-124 is
CDK6, via
CDK6 mediated phosphorylation and subsequent inactivation of the tumour suppressor
pRb[
29,
32]. However, in cervical cancer the virally encoded oncoprotein E7 is thought to bind and inactivate
pRb, suggesting
hsa-miR-124 may (partly) function via other targets in this type of cancer. Using data from a recent study by Baek
et al, we identified
IGFBP7 as a promising target gene in cervical cancer [
17].
IGFBP7 mRNA levels were increased in late passages of FK16B and FK18B cells compared to their corresponding early passages, which also showed increased levels of hsa-miR-124-1 and hsa-miR-124-2 methylation and decreased levels of
hsa-miR-124 expression. In addition,
IGFBP7 showed decreased mRNA expression in CaSki cells ectopically expressing
hsa-miR-124 compared to empty vector control cells and parental cells. No effect was seen in SiHa_miR-124 cells, however. These results indicate that
IGFBP7 may be a potential target of
hsa-miR-124 in part of the cervical cancers, but other targets may be relevant for the tumour suppressive function of
hsa-miR-124 in cervical cancer as well.
IGFBP7 is part of the insulin-like growth factor (IGF) axis, which has been implicated in cervical cancer before. It was shown that the IGF-axis may influence the persistence of hrHPV infections and that abnormally balanced co-expression of IGFBP family members is associated with gynaecological malignancy [
42,
43].
IGFBP7 is the only member of the IFGBP family that binds insulin instead of IGF and is relatively unknown compared to its family members. On one hand
IGFBP7 has recently been described as a tumour suppressor gene in colorectal cancer and was shown to induce senescence in cells harbouring oncogenic BRAF [
44‐
46], whereas on the other hand
IGFBP7 was shown to have oncogenic properties in gliomas [
47]. Further functional studies are needed to investigate whether the tumour suppressive function of
hsa-miR-124 in cervical cancer may in part be mediated via
IGFBP7.
Hsa-miR-124 was originally described as a brain-specific miRNA, involved in neuronal differentiation. Lujambio
et al were the first to show methylation-mediated silencing of
hsa-miR-124 in different human cancer types, reaching the highest frequency in colorectal cancer (75%) [
29]. In acute lymphoblastic leukaemia (ALL) methylation of
hsa-miR-124 was shown to negatively affect clinical outcome [
30,
32]. In ALL hsa-miR-124-1 showed the highest frequency of methylation as was also found for cervical cancer in our study. Interestingly, in gastric cancer and hepatocellular carcinoma, hsa-miR-124-3 showed the highest frequency of methylation [
25,
26], whereas in ALL and cervical cancer hsa-miR-124-3 showed the lowest frequencies of methylation positivity. This may indicate that hsa-miR-124-3 methylation is tissue or tumour type dependent. The fact that hsa-miR-124-3 methylation was infrequent in cervical precursor lesions and HPV-immortalised keratinocyte cell lines supports the notion that at least in cervical carcinogenesis hsa- miR-124-3 methylation is a rather late event. Overall, the methylation positivity rates for
hsa-miR-124 found in our study rank among the highest currently reported, although the use of different assays and methods, including non-quantitative MSP and combined bisulfite restriction analysis, makes a direct comparison difficult. Importantly,
hsa-miR-124 methylation in cervical cancer was histotype-independent and could already be detected in CIN3 lesions and scrapes of women with underlying CIN3, underlining its potential value for cervical cancer screening.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SMW and RAAvB performed all experiments and data analysis and drafted the manuscript. RDMS and PJFS participated in the design of the study and the drafting of the manuscript. FEH, RA and ClS greatly contributed to the functional experiments with retroviral constructs. CJLM, BD and GAM contributed to the conception of the study and critically revised the manuscript. All authors read and approved the final manuscript.