Introduction
Hepatocellular carcinoma (HCC) is the fifth most common malignancy in men and the eighth most common in women worldwide. It is estimated to cause approximately half a million deaths annually (1). Several risk factors for HCC have been reported, including infection with hepatitis B and C viruses, dietary intake of afratoxin and alcohol consumption. However, the molecular pathogenesis of HCC remains poorly understood.
MicroRNAs (miRNAs) are ∼22 nucleotide non-coding RNAs that function as endogenous silencers of target genes. Currently, more than 1,000 miRNAs have been identified in the human genome (the miRBase database) and each miRNA is predicted to control hundreds of gene targets. miRNAs are expressed in a tissue-specific manner and play important roles in development, cell proliferation, apoptosis and oncogenesis (2–5). Dysregulation of miRNAs in cancer has been repeatedly described (6–8). HCC is no exception and various HCC-specific miRNA signatures have been described (9).
DNA methylation of CpG islands within the promoter regions of tumor suppressor genes is known to inhibit transcriptional initiation and thereby silence these genes. Growing evidence indicates that some tumor-suppressive miRNAs are also epigenetically silenced by promoter DNA methylation in cancer (10), suggesting the diagnostic and therapeutic potential of these miRNAs.
In the present study, we aimed to identify miRNA genes that are silenced by DNA hypermethylation in HCC. We screened for genes with promoter DNA hypermethylation using a genome-wide methylation microarray analysis and found that miR-335, which is harbored within an intron of its protein-coding host gene MEST, is downregulated by aberrant promoter hypermethylation in HCC.
Materials and methods
Cell lines and primary tumors
The following 21 HCC cell lines were examined: HLE, HLF, PLC/PRF/5, Li7, Huh7, Hep3B, SNU354, SNU368, SNU387, SNU398, SNU423, SNU449, SNU475, JHH-1, JHH-2, JHH-4, JHH-5, JHH-6, JHH-7, Huh1 and HepG2 (11).
Paired tumor and non-tumor tissues were obtained from 32 HCC patients who underwent surgery. All specimens were immediately frozen in liquid nitrogen and were stored at −80°C until further use. Genomic DNA and total RNA were isolated using the Puregene DNA isolation kit (Gentra, Minneapolis, MN) and TRIzol reagent (Invitrogen, Carlsbad, CA), respectively. Twenty tumor samples were available for DNA methylation analyses and 32 paired tumor and non-tumor samples were available for microRNA and mRNA analyses. This study was approved by the ethics committees and conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from each patient.
Methylation array analysis
We performed a genome-wide DNA methylation analysis called microarray-based integrated analysis of methylation by isoschizomers (MIAMI), as previously described (12–14). The complete experimental procedure can be obtained at http://grc.dept.med.gunma-u.ac.jp/~gene/image/MIAMI%20Protocol%20V4.pdf. Changes in methylation were judged by assessing the differences in methylation-sensitive HpaII cleavage and methylation-insensitive MspI cleavage between samples. We used a custom microarray, which contains ∼38,000 probes chosen from the Agilent promoter array, on an eArray system (http://earray.chem.agilent.com/earray/).
TaqMan miRNA assay
Reverse transcription (RT) reactions and real-time quantitative polymerase chain reactions (PCR) were performed using the TaqMan MicroRNA RT kit (Applied Biosystems, Darmstadt, Germany), TaqMan MicroRNA Assays (Applied BioSystems) and ABI PRISM 7300 Fast Real-time PCR system (Applied Biosystems), according to the manufacturer’s instructions. RNU6B was used as an endogenous control for miRNA levels.
Drug treatment
Cells were treated with 1 or 5 μM of 5-aza-2′-deoxycytidine (5-aza-dCyd; Sigma-Aldrich, St. Louis, MO) for 4 days or 50 ng/ml of trichostatin A (TSA; Wako, Osaka, Japan) for 1 day. In assessing drug synergy, cells were cultured in the presence of 1 or 5 μM of 5-aza-dCyd for 4 days and were then treated for an additional 24 h with 50 ng/ml of TSA.
Methylation analysis
Methylation status was examined by methylation-specific PCR (MSP), bisulfite PCR followed by restriction enzyme digestion [combined bisulfite and restriction analysis (COBRA)] (15) and bisulfite sequencing analysis, as previously described (16). The primers used are listed in Table I. Briefly, for MSP, genomic DNA was treated with sodium bisulfite using an EZ DNA Methylation kit (Zymo Research, Orange, CA) and subjected to PCR using specific primer sets. For COBRA, genomic DNA was treated with sodium bisulfite and subjected to PCR. The PCR products were digested with BstUI, which recognizes sequences unique to the methylated alleles, but cannot recognize unmethylated alleles and the digested products were electrophoresed on 3% agarose gels and stained with ethidium bromide. Methylation levels were calculated as the ratio of the gray scale value of the methylated band to that of the combined methylated and unmethylated bands. The gray scale value was obtained by scanning the gel with Adobe Photoshop CS3 Extended software (Adobe Systems Inc., San Jose, CA, USA). For bisulfite-sequencing, the PCR products were cloned and then sequenced. CpGenome universal unmethylated and methylated DNA (Chemicon, Billerica, MA) served as controls for unmethylated and methylated DNA, respectively.
Real-time quantitative RT-PCR
We quantified mRNA using a real-time fluorescence detection method, as previously described (11). Real-time quantitative PCR experiments were performed with the LightCycler system using FastStart DNA Master Plus SYBR Green I (Roche Diagnostics, Penzberg, Germany), according to the manufacturer’s protocol. The primers used are listed in Table I. The endogenous control for mRNA was GAPDH.
Statistical analysis
Spearman’s rank correlation test, Wilcoxon signed-rank test and Mann-Whitney U test were performed using SPSS 15.0 software (SPSS, Inc., Chicago, IL). P-values of <0.05 were considered significant.
Results
Genome-wide DNA methylation profiles in HCC
To identify miRNA genes that are silenced by DNA hypermethylation in HCC, we compared DNA methylation profiles between three HCC cell lines (SNU449, Li-7 and PLC/PRF/5) and one normal liver tissue using the MIAMI method. The microarray covers approximately 38,000 probes (corresponding to promoter regions of about 14,000 genes), which include 411 probes for miRNA (167 miRNA genes). MIAMI analyses revealed that 575 probes (484 genes) were hypermethylated and 350 probes (277 genes) were hypomethylated similarly in the three HCC cell lines compared to normal liver. The hypermethylated genes included eight miRNA genes (miR-let-7b, miR-101-2, miR-122a, miR-146b, miR-149, miR-200b, miR-335 and miR-497). Therefore, further analysis was focused on these eight miRNA genes. The strategy and partial results are shown in Fig. 1.
Expression of candidate miRNAs in HCC cell lines
We analyzed the expression levels of the eight miRNAs in 21 human HCC cell lines and normal liver using TaqMan miRNA PCR. Expression levels of six miRNAs (miR-let-7b, miR-101-2, miR-122a, miR-146b, miR-335 and miR-497), but not two of the miRNAs (miR-149 and miR-200b), were lower in more than half of the 21 cell lines than normal liver (Fig. 2).
Restoration of miRNA expression by the methyltransferase inhibitor
We then assessed the effects of demethylation on the expression of the six candidate miRNAs. Three HCC cell lines (SNU449, Li7 and PLC/PRF/5) were treated with 5-azadCyd, a methyltransferase inhibitor and miRNA expression levels were assayed with TaqMan miRNA PCR. Expression of four miRNAs (miR-101-2, miR-146b, miR-335 and miR-497), but not two of the miRNAs (miR-let-7b and miR-122a), were restored with 5-aza-dCyd treatment in all three HCC cells (Fig. 3), suggesting that aberrant DNA methylation suppressed the expression of these four miRNAs. Additionally, it was observed that treatment with a histone deacetylase inhibitor, TSA, enhanced the expression of these four miRNAs by 5-aza-dCyd in all three cell lines (Fig. 3). These findings suggest that histone deacetylation may also contribute to the transcriptional repression of these four miRNAs.
Methylation of miR-335/MEST in HCC cells
About half of all miRNA genes are encoded in the introns of protein-encoding genes and subsequently excised from a primary transcript in common with protein coding genes, so-called host genes (10,17–19). Thus, these miRNA genes are more likely to be susceptible to transcriptional repression by aberrant DNA methylation of CpG islands located in the host genes. Of the selected four genes (miR-101-2, miR-146b, miR-335 and miR-497), we identified that miR-101-2 and miR-335 are intronic miRNAs using the human genome browser at UCSC (February 2009). miR-101-2 and miR-335 are located within the introns of RNA terminal phosphate cyclase-like 1 gene (RCL1) (Fig. 4A) and mesoderm specific transcript homolog gene (MEST) (Fig. 5A), respectively. We also found CpG islands around the transcription start sites of miR-101-2/RCL1 and miR-335/MEST genes using the genome database of the European Bioinformatics Institute. However, no CpG islands were found around miR-146b or miR-497.
Therefore, we assessed the methylation status of the CpG islands of miR-101-2/RCL1 and miR-335/MEST via MSP in three HCC cells (SNU449, Li7 and PLC/PRF/5) and normal liver. MSP analyses indicated that the CpG island of miR-101-2/RCL1 was not methylated in these HCC cells (Fig. 4B), whereas aberrant DNA methylation within the CpG island of miR-335/MEST was evident in all three HCC cells (Fig. 5B).
To confirm and quantify the methylation status of miR-335/MEST, we assayed DNA methylation levels of the miR-335/MEST CpG island using the COBRA technique, which involves bisulfite PCR followed by restriction enzyme digestion, in 21 HCC cell lines. COBRA analyses (Fig. 5C) revealed that the miR-335/MEST CpG island was hypermethylated in three cell types (JHH7, HLF and PLC/PRF/5) that lack the expression of miR355 (Fig. 2), partly methylated in eight (JHH6, SNU368, SNU398, SNU423, SNU449, SNU475, Huh7 and Li7) with reduced expression of miR355 (Fig. 2) and unmethylated in the remaining 10 cell lines, including HLE. Consistent with the results of COBRA, further analysis of the PCR products with bisulfite-sequencing showed that the CpG island was hypermethylated in HLF cells (methylation rate, 97%) and hypomethylated in HLE cells (methylation rate, 1%) (Fig. 5D). Taken together, these data suggest that the miR-335/MEST CpG island was hypermethylated in some HCC cells. The physical relationship between miR-335, MEST, the CpG island and the primers used for MSP and COBRA are shown in Fig. 5A.
The expression levels of miR-335 significantly correlated with those of MEST in 21 HCC cell lines (Spearman’s rank correlation test, r=0.83; P=0.0001) (Fig. 5E), supporting the notion that the intronic miR-335 is co-expressed with its host gene, MEST, under the control of the host gene promoter.
Methylation and reduced expression of miR-335 in primary HCC tumors
To determine whether the methylation of the miR-335/MEST CpG island observed in HCC cell lines also occurs in primary human HCC, we assessed the methylation status of miR-335/MEST in paired tumor and non-tumor tissues from 20 patients with primary HCC by using COBRA. Methylation of miR-335/MEST was observed in all 20 HCC tumors and in 15 of the 20 non-tumor liver tissues. Although methylation of miR-335/MEST was found in both HCC tumors and non-tumor tissues, the level of miR-335/MEST methylation was significantly higher in 18 (90%) out of 20 tumors, compared to their non-tumor tissue counterparts (Wilcoxon signed-rank test, P<0.001) (Fig. 6A).
To investigate whether the reduced expression of miR-335 observed in HCC cells was relevant in primary HCC tumors, we analyzed the expression of miR-335 in paired tumor and non-tumor tissues from 32 HCC patients via TaqMan miRNA PCR. The expression level of miR-335 was significantly lower in 25 (78%) out of 32 tumors, compared to their non-tumor tissue counterparts (Wilcoxon signed-rank test, P= 0.001) (Fig. 6B). Taken together, these findings suggest that the expression of miR-335 was frequently reduced by aberrant DNA methylation in primary HCCs.
Since miR-335 was identified as a metastasis suppressor miRNA in breast cancer by Tavazoie et al(20), we examined the relationship between the expression levels of miR-335 and the presence of distant metastasis in these 32 primary HCCs. The expression of miR-335 was significantly lower in HCC tumors with distant metastasis than in those without distant metastasis (Mann-Whitney U test, P=0.02) (Fig. 6C), suggesting that a reduced expression of miR-335 may be association with distant metastasis in HCC, as well as in breast cancer.
Discussion
This is the first report that miR-335 is downregulated in HCC via aberrant promoter hypermethylation, which was demonstrated through a number of approaches. First, we screened for genes with promoter DNA methylation in HCC cell lines using MIAMI, a powerful method for genome-wide profiling of promoter methylation in the human genome (12–14) and found eight miRNA genes that were possibly methylated in HCC cells. Further methylation analyses, including MSP, COBRA, bisulfite-sequencing and drug treatment with 5-aza-dCyd and TSA, combined with expression analyses, narrowed down the candidate methylated miRNA genes and confirmed that the miR-335/MEST CpG island was hypermethylated in some HCC cells. In primary HCCs, the level of miR-335/MEST methylation was significantly higher and the expression of miR-335 was significantly lower in tumors compared to their non-tumor tissue counterparts, suggesting that the expression of miR-335 was reduced by aberrant DNA methylation in primary HCCs. Furthermore, our results suggest that a reduced expression of miR-335 may be associated with distant metastasis in HCC.
DNA hypermethylation of CpG islands within promoter regions is known to be an epigenetic aberration leading to the inactivation of tumor-suppressive miRNA in cancer, which is similar to that of many classical tumor-suppressor genes. To date, 19 intergenic miRNA genes, which are located in the non-coding regions between genes and 42 intronic (intragenic) miRNA genes, which are harbored within introns of their protein-coding host genes, have been identified as tumor-suppressive miRNA (10). Of these, miR-335 was reported to suppress metastasis and migration by targeting SOX4 and tenascin C and inhibit tumor initiation in breast cancer (20,21). The transcription of miR-335 was shown to be co-regulated with MEST by promoter hypermethylation in breast cancer cells (21). Furthermore, it was demonstrated that miR-335 regulates Rb1 and controls cell proliferation in a p53-dependent manner (22). Recent studies have shown that miR-335 orchestrates cell proliferation, migration and differentiation in human mesenchymal stem cells (23), as well as inhibits growth and invasion of malignant astrocytoma cells (24). Further work will be aimed at elucidating the role of miR-335 in the carcinogenesis and metastasis of HCC.