Gene silencing of MIR22 in acute lymphoblastic leukaemia involves histone modifications independent of promoter DNA methylation


Shiang Huang, Institute of Hematology, Union Hospital, Wuhan, China. E-mail:;
Xian-Ming Chen, Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE 68178, USA. E-mail:


Aberrant epigenetic regulation has recently been implicated in the downregulation of tumour suppressor microRNAs (miRNAs). Histone modification and DNA methylation can have different roles in gene silencing in cancer. To investigate whether histone modifications would contribute to the dysregulation of miRNAs in acute lymphoblastic leukaemia (ALL), the effect of a histone deacetylase inhibitor, trichostatin A (TSA), on miRNA expression profile was analysed by microarray assay in a precursor B-cell ALL cell line NALM-6. A total of 10 miRNAs were downregulated and 31 were upregulated significantly following TSA treatment. Among TSA-upregulated miRNAs, MIR22 is an extronic miRNA and resides in the second exon of the non-coding transcript MGC14376. Upregulation of MIR22 transcription was found in both NALM-6 cells and primary human ALL malignant cells treated with TSA. Whereas a CpG island was identified within the promoter element of MIR22, no promoter DNA methylation was detected in these cells. In contrast, accumulation of the repressive histone marker H3K27 trimethylation (H3K27triM) was indentified around the transcriptional start point of the gene, which was reduced by TSA treatment. Thus, accumulation of H3K27triM independent of promoter DNA methylation may be a novel epigenetic mechanism for MIR22 silencing in ALL.

MicroRNAs (miRNAs) are a class of recently identified endogenous small non-coding RNAs of c. 22 nucleotides that can induce direct mRNA degradation or translational inhibition of their target genes (He & Hannon, 2004). These single-strand RNAs are actively involved in many crucial cellular processes, such as proliferation, development, differentiation and apoptotic cell death (Bartel, 2004; Miska, 2005; He et al, 2007). There is now ample evidence that miRNA expression profile is altered in tumour cells (Calin & Croce, 2006; Esquela-Kerscher & Slack, 2006; Mi et al, 2007) and dysregulation of miRNA expression has been implicated in the carcinogenesis process. In general, the majority of miRNAs are downregulated in cancer specimens. Several miRNAs that are downregulated miRNAs in cancer specimens can inhibit the translation of proto-oncogenes in normal cells, and are, therefore, considered as ‘tumour suppressor miRNAs’ (Calin et al, 2002; Takamizawa et al, 2004; Cimmino et al, 2005; Iorio et al, 2005; Johnson et al, 2005). Certain miRNAs are upregulated in tumour cells and may act as ‘oncomiRNAs’ (Hayashita et al, 2005; He et al, 2005; Tam & Dahlberg, 2006; Volinia et al, 2006; Voorhoeve et al, 2006). Nevertheless, the mechanisms by which miRNA expression is altered in cancer cells are still unclear.

Most miRNAs are generated by RNA polymerase II as primary transcripts (pri-miRNAs), which form a stem-loop structure in the nucleus. MicroRNA genes can be located in introns or exons of their host genes and are transcribed together (Baskerville & Bartel, 2005). MicroRNAs that have their own promoters are independently expressed and miRNAs organized in clusters share the same transcriptional regulation (Yu et al, 2006). Like many other human genes, miRNA genes are susceptible to epigenetic regulation. Histones dynamically regulate transcriptional activities through their various post-translational modification patterns and have emerged as an important epigenetic regulator (Wang et al, 2004). The active role of the aberrant histone modifications in transcriptional silencing of genes is becoming increasingly understood and may involve a synergy between the methylation and histone deacetylases (HDAC) activity. This synergy can be mediated directly by HDAC interaction with DNA methylating enzymes and by recruitment through complexes involving methyl-cytosine binding proteins (Baylin et al, 2001). Indeed, recent studies revealed that DNA methylation facilitates the accumulation of repressive histone markers around the promoter regions of tumour suppressor genes resulting in gene silencing in cancer cells (Cameron et al, 1999; Fahrner et al, 2002). H3K27 trimethylation (H3K27triM), one of the two most common repressive histone modification markers (Jenuwein & Allis, 2001), has been linked to de novo DNA methylation (Viréet al, 2006; Ohm et al, 2007; Schlesinger et al, 2007). Promoter hypermethylation has been demonstrated to induce downregulation of tumour suppressor miRNAs, such as MIR127, MIR124A, MIR9-1, MIR137 and MIR193A, in various cancers (Saito et al, 2006; Lujambio et al, 2007; Kozaki et al, 2008; Lehmann et al, 2008). Histone modification and DNA methylation can have different roles in gene silencing during tumorigenesis. Recent evidence indicates that relationships between DNA methylation and histone modification have implications for understanding mechanisms of gene repression in cancer (Cedar & Bergman, 2009). Alterations in miRNA profiles have been identified in acute lymphoblastic leukaemia (ALL) (Mi et al, 2007; Roman-Gomez et al, 2009). Aberrant miRNA methylation has been reported to be associated with the clinical outcome of these patients (Roman-Gomez et al, 2009). Nevertheless, the role for promoter histone modification in dysregulation of miRNA expression in ALL is still obscure.

This study demonstrated that a specific HDAC inhibitor, trichostatin A (TSA), induced significant alterations in the expression profile of mature miRNAs in a precursor B-cell ALL cell line, NALM-6. Of these altered miRNAs, we identified that MIR22 showed an increased transcription in NALM-6 cells following treatment with TSA, but not with a DNA methylation inhibitor, 5-AZA-2′-deoxycytidine (AZA). Interestingly, transcriptional suppression of MIR22 in NALM-6 cells was associated with promoter histone modification, probably mediated by H3K27triM. Whereas a CpG island was identified within the promoter element of MIR22, no significant DNA methylation in the region was detected in NALM-6 cells. Moreover, histone modification-mediated transcriptional suppression of MIR22 was further confirmed in the peripheral blood mononuclear cells (PBMCs) from patients with ALL, but not in PBMCs from normal control individuals or from patients with acute myeloid leukaemia (AML). These data indicate that recently identified accumulation of H3K27triM independent of promoter DNA methylation in other tumour cells (Kondo et al, 2008) may be a novel epigenetic mechanism for MIR22 silencing in ALL.

Materials and methods

Patients and NALM-6 cells

A total of 34 patients were included in this study, and included 27 samples of leukaemia (18 ALL, 9 AML) and seven normal controls. All patient blood samples were collected at the time of diagnosis and with informed consent at the Union hospitals of Wuhan, China. PMBCs were isolated from patient blood samples using Ficoll-Paque (GE Healthcare Bio-sciences AB, Uppsala, Sweden) density centrifugation. The NALM-6 (precursor B-cell ALL) cell line was a gift of Dr. Georgakilas from East Carolina University, Greenville, NC, USA and cells were grown in flasks with RPMI 1640 medium supplemented with 10% fetal bovine serum, l-glutamine and gentamicin.

TSA and AZA treatment

Treatment was conducted during the log phase of cell growth with TSA (Invivogen, San Diego, CA, USA) or AZA (Invivogen), whereas the control cells were not treated. NALM-6 cells were seeded at 5 × 106/ml. TSA was added at a final concentration of 1 μmol/l to the culture medium and incubated for 6 h, whereas AZA was added at a concentration of 1 μmol/l and incubated for 72 h with change of culture medium containing AZA every 24 h.

RNA isolation and miRCURY™ LNA Array analysis

Total RNAs were isolated from the TSA-treated and control NALM-6 cells by using the mirVana miRNA Isolation kit (Ambion, Austin, TX, USA) according to the manufacturer’s instructions. RNAs were then processed for the microarray chip analysis performed and analysed by Exiqon (Vedbaek, Denmark). Briefly, prior to initiating the analysis, samples were subject to RNA quality control in order to assess the integrity of the RNAs, the content and concentration of small RNAs. The miRCURY™ LNA Array Power labelling kit which allows highly efficient and uniform labelling of miRNAs with minimal sequence bias was then applied for labelling. Arrays were scanned with highly sensitive equipment in an ozone-free environment to reduce day-to-day variation to a minimum. Image Analysis was then performed to quantify the signals on the arrays. The data obtained were normalized and presented as log2 (Hy5/Hy3) ratios which passed the filtering criteria as previously reported (Gong et al, 2009).


Total RNAs were isolated from cells with the mirVana™ miRNA Isolation kit (Ambion). An amount of 0·05 μg total RNAs was reverse-transcribed by using the Taqman MicroRNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA). For the quantification of mature MIR22, a set of primer/probes specific to the mature MIR22 sequence (Applied Biosystems) was used. For the quantification of pri-MIR22 transcripts, specific primers for pri-MIR22 were utilized (listed in Table SI). Comparative real-time PCR reactions were performed in triplicate using Taqman Universal PCR Master Mix (Applied Biosystems) on the Applied Biosystems 7500 FAST real-time PCR System. Normalisation was performed by using RNU6B primers and probes. Relative expression was calculated by using the comparative threshold cycle (CT) method (Chen et al, 2005).

For conventional reverse transcription (RT)-PCR analysis of Deleted in Liver Cancer 1 (DLC1) mRNA expression, total RNAs were isolated from cells with the TRIzol reagent (Ambion). Total RNA (1 μg) was reverse transcribed to cDNA by using a Moloney Murine Leukaemia Virus (M-MLV) Reverse Transcriptase kit (Invitrogen, Carlsbad, CA, USA). The primers used are described in Table SI. PCR products were detected by electrophoresis. GAPDH was used as a loading control.

Promoter luciferase reporter assay

Various fragments representing different regions within the promoter element for human MIR22 were amplified by PCR from human genomic DNA. PCR primers are described in Table SI. The PCR products were separated by agarose gel electrophoresis, and DNA fragments were then isolated and cloned in the MluI- and HindIII- digested pGL3 Basic Vector (Promega, Madison, WI, USA) using T4 DNA ligase (Fisher scientific, Pittsburgh, PA, USA). All constructs were confirmed by sequencing. Cells were transfected with each reporter construct followed by assessment of luciferase activity 24 h after transfection. Luciferase activities were measured and normalized to the control β-gal level. The luciferase activity of each construct was compared with that of the promoterless pGL3 basic vector.

Methylation analysis

Genomic DNA was isolated from treated and untreated cells using proteinase K digestion and organic extractions according to standard procedures (Laird et al, 1991). A total of 0·5 μg genomic DNA was treated with sodium bisulfite according to the manufacturer’s recommendations (EZ DNA Methylation-Gold™ kit; Zymo Research, Orange, CA, USA). Bisulfite-treated DNAs were then used as a template for PCR with primers designed using MethPrimer that were specific to the CpG island region. The combined bisulfite restriction analysis (COBRA) primers for the CpG island region of the MIR22 promoter are described in Table SI. PCR products were digested with BstUI (New England Biolabs, Ipswich, MA, USA), which recognizes sequences unique to methylated alleles but cannot recognize unmethylated alleles (Xiong & Laird, 1997). For the bisulfite sequencing analysis, the PCR products were subcloned by using the TA cloning kit (Invitrogen). Plasmid DNAs from 5 insert-positive clones were isolated using the QIAquick Plasmid Mini Prep kit (Qiagen, Valencia, CA, USA) and then sequenced. The methylation-specific PCR (MSP) primers described in Table SI were used for the PCR reactions to differentiate between methylated and unmethylated sequences. The PCR products were run in 1·5% agarose gels and were then stained with ethidium bromide to visualize COBRA and MSP products.

Chromatin Immunoprecipitation (ChIP) assay

ChIP analysis was performed with a commercially available ChIP Assay kit (Upstate Biotechnologies, Lake Placid, NY, USA) in accordance with the manufacturer’s instructions. In brief, 1 × 107 NALM-6 cells were cultured in 15-cm culture dishes and then either treated with 1 μmol/l TSA for 6 h or left untreated. The chromatin fraction was immunoprecipitated overnight at 4°C using anti-trimethyl-K27 histone H3 (Upstate Biotechnologies) or anti-acetyl-histone H3 (Upstate Biotechnologies). PCR amplification was done in 25 μl with specific primers for the analysed promoter regions. The primers used are described in Table SI.

Databases and GenBank accession number

The gene sequences were analysed using the University of California at Santa Cruz Human Genome Browser (, CpG Island Searcher Program ( (Takai & Jones, 2002), and MethPrimer ( The miRNA sequences were analysed using miRBase ( The GenBank accession number of MGC14376 mRNA is NM_032895.

Statistical analysis

Data were compared using the analysis of variance (anova) test. < 0·05 was considered to represent statistical significance.


TSA treatment alters miRNA expression profiles in NALM-6 cells

To test the role of histone modifications in miRNA expression in ALL, we treated NALM-6 cells for 6 h with TSA (1 μmol/l) followed by microarray analysis for mature miRNAs. Using the mercury LNA array analysis service with the Exiqon Human MicroRNA Biochips, covering all the currently known (over 800) human mature miRNAs as provided and performed by Exiqon (Vedbaek), we detected expression of a total of 376 mature miRNAs in NALM-6 cells. Forty-one miRNAs showed significant change in their cellular levels (10 down- and 31 up-regulated) in NALM-6 cells following TSA treatment (Fig 1). All microarray data reported in this article were described in accordance with MIAME [minimum information about a microarray experiment; Microarray and Gene Expression Data (MGED) Society] guidelines (Brazma et al, 2001) and were deposited at ArrayExpress under accession no. E-MEXP-2047 ( Among the 31 upregulated miRNAs, MIRLET7A, MIR15A, MIR181 and MIR22 have previously been shown to possess anti-tumour activities (Pekarsky et al, 2006; Sampson et al, 2007; Calin et al, 2008; Chang et al, 2008). MIRLET7A and MIR22 have also been reported to be frequently downregulated in ALL patients (Mi et al, 2007).

Figure 1.

 TSA treatment alters miRNA expression profile in NALM-6 cells. NALM-6 cells were exposed to TSA (1 μmol/l) for 6 h and the miRNA expression profile was measured using the miRCURY™ LNA Array provided and performed by EXIQON. Data are presented as the log2 (Hy5/Hy3) ratios, which passed the filtering criteria variation across samples from non-stimulated (= 3) and TSA-stimulated (n = 3) samples. TSA-induced alterations in miRNA expression in NALM-6 cells represented as a heat map in the right panel. The red and blue indicate an increase and decrease in miRNA expression, respectively.

TSA, but not AZA, increases expression of MIR22 in NALM-6 cells

To test how histone modifications alter miRNA expression in NALM-6 cells, we focused on the transcriptional regulation of MIR22, a TSA-induced miRNA in NALM-6 cells by our miRNA array analysis. Because the aberrant methylation of the promoter CpG island region is a common epigenetic mechanism for transcriptional silencing of tumour suppressor genes, we also measured the effects of AZA, a DNA demethylating agent, on the transcriptional regulation of MIR22 in NALM-6 cells to test whether DNA methylation is involved. Real-time PCR reactions were performed to quantify mature MIR22 levels in NALM-6 cells following treatment with TSA or AZA. We detected a significant increase of mature MIR22 in cells after TSA treatment (Fig 2A), consistent with our array results. In contrast, no significant change of mature MIR22 expression was detected in cells following AZA treatment (Fig 2A).

Figure 2.

 Primary transcripts of MIR22 and its host transcript MGC14376 are induced by TSA treatment but not by a DNA methylation inhibitor. (A) Expression of mature MIR22 was measured by real-time PCR in NALM-6 cells after treated by TSA (1 μmol/l) for 6 h or AZA (1 μmol/l) for 72 h. Data are mean ± standard deviations (SD) from three independent experiments. (B) Genomic organisation of the human MIR22/MGC14376 locus. Exons are shown as black rectangles. (C) Expression of pri-MIR22 was measured by real-time PCR in NALM-6 cells after treated by TSA (1 μmol/l) for 6 h or AZA (1 μmol/l) for 72 h. Data are mean ± SD from three independent experiments. (D) DLC1 mRNA expression was measured by RT-PCR after treated by TSA (1 μmol/l) for 6 h or AZA (1 μmol/l) for 72 h. GAPDH was used as a control. (E) Alternations of pri-MIR22 expression in PBMCs from health volunteers and patients with ALL or AML after treatment with TSA (1 μmol/l) for 6 h, as analysed by real-time PCR.

Human MIR22 is located in the second exon of the non-coding transcript MGC14376 (Pekarsky et al, 2006) (Fig 2B). Recent studies have identified the promoter element of the gene and demonstrated the co-expression of MIR22 and MGC14376 in various tissues including brain, liver, heart, lung, kidney and spleen (Pekarsky et al, 2006; Calin et al, 2008). Using a real-time PCR approach employing specific primers for the primary MIR22 transcripts (pri-MIR22), we then analysed expression of MIR22 at the transcription level in NALM-6 cells following treatment with TSA or AZA. TSA treatment induced an apparent upregulation of pri-MIR22 in NALM-6 cells (Fig 2C) whereas cells treated with AZA showed no significant change in pri-MIR22 expression (Fig 2C). As control, AZA treatment induced a significant upregulation of DLC1, a well-known methylation silenced gene in NALM-6 cells (Chang et al, 2008) (Fig 2D). These data suggest that histone modifications, but not DNA methylation, may be involved in the silencing of MIR22 in NALM-6 cells.

TSA enhanced MIR22 transcription is not limited to NALM-6 cells

To clarify whether TAS-induced MIR22 transcription is specific to NALM-6 cells, PMBCs from seven healthy volunteers, 18 ALL patients and 9 AML patients were collected and applied to pri-MIR22 analysis. Consistent with previous results (Mi et al, 2007), real-time PCR detected a lower expression of pri-MIR22 in PMBCs from ALL patients compared with that from the healthy volunteers (Fig 2E). Treatment with TSA significantly increased pri-MIR22 expression in PMBCs from ALL patients, but not in cells from the health volunteers (Fig 2E). Whereas PMBCs from ALL patients and AML patients showed comparable levels of pri-MIR22, TSA treatment had no effect on pri-MIR22 expression in PMBCs from AML patients (Fig 2E), suggesting TSA-mediated upregulation of MIR22 transcription in ALL but not AML malignant cells.

The CpG island in the promoter element of MIR22 is not significantly methylated in NALM-6 cells and PMBCs from ALL patients

To further confirm the DNA methylation status at the promoter element of MIR22 in NALM-6 cells, a DNA methylation assay was performed to determine the methylation status of this CpG island region in NALM-6 cells. Using several computerized CpG island identification programs (Takai & Jones, 2002), we detected a CpG Island covering the promoter element of MIR22 at the transcriptional start region (Fig 3A). Using the COBRA assay, bisulfite-treated PCR products using primers covering the CpG island region could no longer be cut by the enzyme BstUI (recognition site CGCG) (Fig 3B). In addition, MSP assay also revealed no DNA hypermethylation at the MIR22 promoter (Fig 3C), indicating a lack of DNA methylation at the CpG island region within the promoter element of MIR22 in NALM-6 cells. To obtain more comprehensive information on DNA methylation status in this CpG island region, we performed bisulfite genomic sequencing experiment. The DNA methylation level was consistently very low or undetectable (Fig 3D). Moreover, the MPS assay was used to analyse the methylation status at the promoter element of MIR22 in primary human specimens. No DNA hypermethylation was detected in PMBCs from the healthy volunteers and patients with either ALL or AML (Fig 3E). These data provide further evidence that MIR22 silencing in NALM-6 cells may be DNA methylation-independent.

Figure 3.

 The CpG island in the promoter of MIR22/MGC14376 is not significantly methylated in NALM-6 cells. (A) The percentage of C+G nucleotide (CG%) and the density of CpG dinucleotides are shown for the region of the MIR22/MGC14376 transcript. A CpG island is located in the promoter region. (B) COBRA analysis for methylation statues within the CpG island region of MIR22. (C) MSP analysis for methylation status within the CpG island region of MIR22. (D) Bisulfite sequencing showed no or only slight methylation of CpG island in MGC14376 promoter in NALM-6 cells. (E) Absence of promoter DNA methylation of MIR22 in PBMCs from ALL patients. MSP assay showed no hypermethylation of the MIR22 promoter in normal, ALL and AML PBMCs. U, unmethylated; M, methylated.

TSA reduces accumulation of H3K27triM-associated histone modification at the promoter element of MIR22 to increase transcriptional expression of MIR22 in NALM-6 cells

The DNA methylation-independent nature of MIR22 suppression in ALL malignant cells motivated the analysis of the molecular mechanisms underlying histone modification-associated transcriptional suppression of MIR22 in NALM-6 cells. The histone modification status was tested in the promoter element of MIR22. We first analysed the transcriptional activity of different regions within the MIR22 promoter MIRusing luciferase reporter vectors containing fragments of various promoter regions. Based on previous report on the promoter element of MIR22 (Rodriguez et al, 2004; Chang et al, 2008), we initially cloned a 2 kb fragment (from −1 kb to +1 kb covering the region 1–3) around the transcription start site into the pGL3 Basic Vector to transfect NALM-6 cells (Fig 4A). A sevenfold increase in luciferase activity was observed in transfected NALM-6 cells (Fig 4B), confirming the transcriptional activity of the promoter element. Further deletion mapping analysis showed that the first 500 bp sequence (region 1) may not be necessary for transcriptional regulation, because the removal of this region from the 2 kb fragment did not significantly affect luciferase expression in transfected cells (Fig 4B). Unexpectedly, the last 700 bp sequence (region 3) seemed to possess a strong transcriptional inhibitory ability, because truncated fragments without this region increased the leuciferase activity to 80-fold in transfected H69 cells (Fig 4B).

Figure 4.

 Different regions in the MIR22 promoter and MGC14376 differentially regulate its transcriptional activation. (A) Genomic regions indicated by bars around the MIR22 promoter and MGC14376 were analysed by using the promoter luciferase reporter assay. (B) luciferase reporter analysis of promoter activities. H69 cells were transfected with the indicated fusions of the promoter fragments to the luciferase reporter gene. Cells were analysed by luciferase reporter assay. Data are representative of three independent experiments.

Having demonstrated that different promoter regions differentially regulate MIR22 transcription, we then tested the potential histone modifications in the various promoter regions of MIR22 in NALM-6 cells. Acetylated histone H3 is generally associated with active chromatin and gene transactivation. ChIP assays were performed to detect the presence of acetylated H3 in the MIR22 promoter regions in NALM-6 cells (Fig 5A). The presence of acetylated H3 was found to be localized in all the three regions of the MIR22 promoter (Fig 5B, C). TSA treatment did not significantly affect the acetylated H3 marker in region 1 and 2. However, enhanced accumulation of acetylated H3 was detected in region 3 following TSA treatment (Fig 5B, C).

Figure 5.

 Transcriptional silencing of MIR22 in NALM-6 cells is associated with H3K27triM histone modification. (A) Three regions around the MIR22 promoter for ChIP assay. (B) Alterations of histone modifications around the promoter by TSA treatment. The levels of acetylated H3 and H3-K27triM in the three regions around the promoter were determined by ChIP assay by using the antibody to acetylated H3 and H3-K27triM in NALM-6 cells untreated or treated with TSA. (C) Quantification of representative gels showing the ratio of acetylated H3 (active marker) and H3K27triM (repressive marker) to input DNA in NALM-6 cells untreated or treated with TSA for these three regions. Data are representative of three independent experiments.

How histone modifications inhibit gene transcription in a DNA methylation-independent manner is still unclear. H3K27triM-associated histone modification has recently been reported to silence genes in cancer cells through promoter methylation-independent mechanisms (Kondo et al, 2008). Therefore, we further examined the H3K27triM status in the MIR22 promoter, particularly in region 3, in NALM-6 cells. While a modest amount of H3K27triM repressive marker was detected in region 2, a significant amount of H3K27triM was identified in region 3 in NALM-6 cells (Fig 5B, C). For both regions 2 and 3, the amount of H3K27triM was decreased by TSA treatment (Fig 5B, C). As for region 1, no H3K27triM was detected in NALM-6 cells in the absence or presence of TSA (Fig 5B, C). Thus, H3K27triM-associated histone modification occurred in the MIR22 promoter element covered by regions 2 and 3 in NALM-6 cells and this process can be released by TSA treatment. Collectively, our data supported that accumulation of H3K27triM at promoter region contributes to DNA methylation-independent epigenetic silencing MIR22 in ALL malignant cells.


There is ample evidence to suggest that abnormal miRNA expression is closely associated with tumour initiation, promotion and progression. It would be of great interest to understand how miRNA expression is regulated in cancer cells. The present study showed that TSA rapidly altered the expression profile of mature miRNAs in NALM-6 cells. Consistent with the previous observation that multiple miRNAs are downregulated in ALL cells (Yu et al, 2006; Mi et al, 2007), we found that the major effect of TSA treatment on NALM-6 cells was to reactivate miRNA expression. A total of 31 miRNAs were upregulated after TSA treatment including those that were previously reported to be downregulated in ALL patients, such as MIRLET7A, MIR27A and MIR22 (Mi et al, 2007). On the other hand, a subset of miRNAs was downregulated by TSA treatment in our array. Obviously, aberrant histone modification is an important epigenetic mechanism for the dysregulation of miRNA expression in NALM-6 cells.

Human miRNA genes are generally transcribed by RNA polymerase II and are susceptible to epigenetic regulation at the transcriptional level (Baskerville & Bartel, 2005). miRNA-22 is an ‘extronic’ transcription miRNA and its gene, MIR22, is found in the second exon of the non-coding transcript MGC14376 (Rodriguez et al, 2004; Chang et al, 2008). Accordingly, we identified an enhanced expression of mature and pri-MIR22, as well as MGC14376, in NALM-6 cells following TSA treatment. TSA-stimulated transactivation of MIR22 was further confirmed in primary leukaemia cells from ALL patients. Notably, a lower expression of pri-MIR22 was detected in PMBCs from ALL patients compared with that in healthy volunteers. Thus, transcriptional suppression of MIR22 expression occurs in ALL cells and this suppression may be regulated by histone modifications. Whether this can happen in other types of leukaemia is unclear. Our analysis of a limited number of AML patients suggests a histone modification-independent suppression of MIR22 in PMBCs from the patients.

As one of the two most common histone modifications associated with epigenetic silencing, H3K27triM serves as a signal for specific chromodomain binding of the polycomb repressor complex (PRC) including BMI1, RING1, EZH2, HPC and HPH5 (Baylin et al, 2001; Chang et al, 2008). Binding of PRC blocks the recruitment of transcriptional activation factors and prevents initiation of transcription by RNA polymerase II (Jenuwein & Allis, 2001; Kondo et al, 2008). We found that transcriptional silencing of MIR22 in ALL cells may involve H3K27triM-associated histone modification in the promoter. Promoter luciferase reporter analysis revealed that the first intron of the promoter, covered by regions 2 and 3, is critical for the transcriptional regulation of the gene. Indeed, TSA treatment enhanced the accumulation of acetylated H3 in region 3 in NALM-6 cells as revealed by ChIP analysis. Moreover, localization of the repressive marker H3K27triM was detected in the region 3 of the promoter and TSA treatment induced a decrease of the H3K27triM marker at this region in NALM-6 cells. Coupled with the induced transactivation of MIR22 in NALM-6 cells following TSA treatment, we speculated that TSA induces MIR22 transactivation via release of H3K27triM-associated transcriptional repression at the first intron (region 3) of MIR22. Histone modification status associated with the intron regions has recently been reported in the transcriptional regulation of human genes, particularly in tumour cells (Roh et al, 2005, 2007; Heintzman et al, 2007). How TSA inhibits H3K27triM and what components of the PRC maybe recruited to the region remain unclear and merit further investigation. A recent report (Fiskus et al, 2006) indicated that TSA could deplete EZH2 in ALL cells.

Global alterations of DNA methylation have recently been shown to be associated with gene silencing in ALL cells (Taylor et al, 2007). Previous studies have linked H3K27triM to de novo DNA methylation in many cancer cells (Viréet al, 2006; Ohm et al, 2007; Schlesinger et al, 2007). DNA methylation and histone modifications have been demonstrated to downregulate miRNA expression in cancer cells such as MIR127 expression in bladder cancers (Saito et al, 2006). More recently, aberrant miRNA methylation has been demonstrated to be a common phenomenon in ALL that affects the clinical outcome of these patients (Roman-Gomez et al, 2009). However, our data indicate that transcriptional silencing of MIR22 in ALL cells involves H3K27triM-associated histone modification but is independent of promoter DNA methylation. Although a CpG island resides within the promoter element of MIR22, we found no or very slight methylation within the promoter element in both NALM-6 cells and primary leukaemia cells from ALL patients. Consistently, AZA treatment did not reverse MIR22 expression in NALM-6 cells. Our results are consistent with a recent study showing that H3K27triM silencing machinery as a cancer-specific event is not closely associated with DNA methylation in some cancer cells (Kondo et al, 2008).

TSA treatment resulted in significant alterations in expression of multiple mature miRNAs in NALM-6 cells as revealed by microarray analysis. It remains to be tested whether histone modifications alter expression of those miRNA genes at the transcriptional level. In addition, we detected that a subset of miRNAs in their mature form was actually downregulated in NALM-6 cells following TSA treatment, whereas TSA generally activates gene transcription via inhibiting histone deacetylase activities. It is possible that an indirect mechanism may account for this TSA-induced suppression, but the rapid decrease of those miRNAs in TSA-treated cells would argue for a direct transcriptional inhibition.

In conclusion, this study showed that histone modification is involved in miRNA dysregulation in human ALL cells. Specifically, using NALM-6 cells and primary leukaemia cells from ALL patients, we demonstrated that the silencing of MIR22 in ALL cells is associated with the accumulation of histone modification in the promoter element of this gene but independent of DNA methylation. MiRNA-22 is one of the miRNAs frequently downregulated in human ALL cells (Mi et al, 2007) and enforced expression of MIR22 can inhibit the in vivo growth of B-cell lymphoma in mice, indicating the anti-tumour effect of MIR22 (Pekarsky et al, 2006; Sampson et al, 2007; Calin et al, 2008; Chang et al, 2008). Moreover, histone modifications have been implicated in the tumorigenesis of various cancers and HDAC inhibitors possess significant growth inhibitory activity on ALL cells (Scuto et al, 2008). Therefore, it would be of interest to investigate the functional role for MIR22 in histone modification-associated tumour initiation, promotion and progression in ALL cells.


We are grateful to Dr Georgakilas at the East Carolina University for providing the NALM-6 cells. We thank Drs G Hu and A-Y Gong (Department of Medical Microbiology and Immunology, Creighton University Medical Center) for technical assistance and helpful discussion.

Conflict of interest Statement

None declared.


The National Young Scientist’s Program of China (No. 30700330, No. 30671003 and No. 30700429), the Nebraska Tobacco Settlement Biomedical Research Program (LB692) and the Creighton Cancer and Smoking Disease Research Development Program (LB595).