Molecular dysfunctions in acute myeloid leukemia revealed by integrated analysis of microRNA and transcription factor
- Authors:
- Published online on: April 15, 2016 https://doi.org/10.3892/ijo.2016.3489
- Pages: 2367-2380
Abstract
Introduction
Acute myeloid leukemia (AML) is an aggressive hematological malignancy accounting for ~80% of all adult leukemia cases, and encompasses 15–20% of leukemia cases in children (1,2). It is estimated that nearly 15% of mortality among patients with hematologic malignancy were attributed to AML (3), and more than 13,000 new cases were diagnosed each year in the USA (4), thus, making AML a major cause of hematologic disorders. AML is usually characterized by malignant clonal proliferation of hematopoietic stem cells and progenitors (blasts) in the bone marrow with the complete or partial blockage at different stages of myeloid differentiation (1,2). Despite advances in chemotherapy and hematopoietic stem cell transplantation, the outcome of AML patients have not improved substantially in the last four decades, with an overall 5-year survival rate of ~25% (1,2,5). Relapse occurs frequently, as ~50–70% of AML patients who achieve complete remission within three years after frontline therapy (2). Among these relapsed patients, only one-third are salvageable with current treatment regimens (5). Therefore, it is urgently needed to search for new therapeutic targets for AML. Securing further advances in therapy is dependent on our increasing understanding of the factors and molecular mechanisms of leukemogenesis in AML. Currently, insight into the pathogenesis of AML has largely come from the investigations of cytogenetic abnormalities and molecular genetic mutations. Many chromosomal structural aberrations and somatic mutations have been identified, which include rearrangements of RUNX1-RNNX1T1, CBFB-MYH11, PML-RARA and MLL gene (6), mutations of CEBPA, NPM1, FLT3, DNMT3A, TET2, IDH1/2 and ASXL1 gene (1,7,8). Despite the pathological role and prognostic implications of these genes have been well elucidated, understanding of the multistep pathogenesis of AML remains limited. Specifically, the regulatory networks of AML gene expression are still unclear.
miRNAs are small non-protein coding RNA molecules of ~19–25 nucleotides that mediate post-transcriptional regulation of gene expression level (9). Pri-miRNAs are transcribed by RNA polymerase II and subsequently processed by the Drosha-DGCR8 enzyme complex in the nucleus to form ~70 nt hairpin pre-miRNAs. Finally, pre-miRNAs were cleaved in the cytoplasm by the endoribonuclease Dicer to yield mature miRNAs. The mature miRNA may bind to the 3′-UTR of target mRNAs through its seed region, which leads to mRNA degradation or inhibition of translation (10). miRNAs play important roles in a variety of biological processes and are implicated in the initiation and progression of many types of human cancers, including AML (11,12). It has been reported that ectopic expression of miR-29a would induce hematopoietic stem cells/progenitors progresses to AML in a mouse model (13). High expression of miR-100 and miR-375 was found in pediatric AML patients, which were correlated with poorer relapse-free and overall survival (14,15). miR-193a expression is downregulated in AML1/ETO-positive leukemia cells, suppression of miR-193a expands the oncogenic activity of the fusion protein AML1-ETO involving a feedback circuitry in miR-193a and AML1-ETO/DNMTs/HDACs (16). Recently, increased miR-181a expression was shown to be associated with improved prognosis in cytogenetically normal AML. In xenograft mouse models of AML, ectopic miR-181a expression inhibits tumor growth (17). Although the aberrant expression and molecular function of AML-related miRNAs were reported in these investigations, the global miRNA regulatory network in AML remains unknown.
TFs are trans-acting protein factors that control the transcription of target genes through specifically binding to the TFBS located within promoter region of target genes (18). As two major types of regulators of gene expression, TFs and miRNAs are able to tightly coordinate to ensure precise and accurate gene expression. TF and miRNA may reciprocally regulate one another and both can regulate the expression of target genes in a combinatorial manner (18,19). It has been reported that miR-223 and transcription factor E2F1 regulate each other to form negative feedback loop in AML (20). Furthermore, transcription factor C/EBPα can exert its effects by inducing miR-30c inactivating the Notch1 protein and enhancing granulocytic differentiation in AML (21). Nevertheless, global regulation of gene expression involved in the TFs and miRNAs in AML are still poorly understood. Therefore, integrated analysis correlating changes in the expression patterns of miRNAs and TFs, as it relates to AML pathogenesis, requires examination.
In the present study, we analyzed miRNA and TF expression profiles in bone marrow samples of AML patients and control groups, utilizing miRNA sequencing and TF array technology, respectively. Based on a combined strategy of the target prediction of miRNAs and TFs, we integrated differentially expressed miRNAs, AML candidate genes and differentially expressed TFs to construct a comprehensive miRNA-TF mediated regulatory network specifically for AML. GO and pathway analysis was performed to determine associated functions and signaling pathway of the network nodes. After calculating the number of the node degree, we found some hub miRNAs and TFs, and further investigated their regulation in the subnetwork. Our results demonstrate that the altered expression levels of miRNAs and TFs may have implications in AML pathogenesis, and that integrative analysis of miRNAs and TFs may provide a new foundation for molecularly targeted therapy of AML in the future.
Materials and methods
Patient samples and ethics statement
A total of 25 bone marrow biopsy samples, including 15 patients with newly diagnosed and untreated AML and 10 age-matched controls, were enrolled in the study. The AML samples were collected at Shenzhen People's Hospital (Shenzhen, China) before any chemo- or radiotherapeutic treatment. Diagnosis and classification of AML were determined according to the French-American-British classification system (22,23). To serve as controls, the samples with normal bone marrow morphology were provided from 10 patients with unexplained anemia or fever by the Affiliated Hospital of Guangdong Medical University (Zhanjiang, China). All the clinical characteristics of AML patients and controls are shown in Table I. The study was carried out according to the Helsinki convention criteria and approved by the ethics committee of the Shenzhen People's Hospital and the Affiliated Hospital of Guangdong Medical University. Written informed consent to use biological samples and clinical data was obtained from patients or their parents.
Total RNA extraction
Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and a RNeasy kit (Qiagen, Hilden, Germany) following manufacturer's instructions, to include a DNase digestion step. RNA purity and concentration were determined using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) measuring absorbance at 230, 260 and 280 nm.
miRNA sequencing and data analysis
Total RNA was ligated sequentially to 3′ and 5′ RNA adapters using T4 RNA ligase (Promega, Madison, WI, USA). Ligation products were then reverse-transcribed and PCR amplified using Illumina's proprietary RT primers and amplification primers (Illumina, San Diego, CA, USA). Subsequently, ~135–155 bp PCR amplified fragments (correspond to ~15–35 nt small RNAs) were isolated and purified from the PAGE gels. After quantified with Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA), the completed libraries were diluted to a final concentration of 8 pM and loaded on the Illumina cBot for cluster generation using TruSeq Rapid SR cluster kit (Illumina) according to the manufacturer's protocol. Finally, the sequencing was carried out on Illumina HiSeq 2000 using TruSeq Rapid SBS kit (Illumina) as recommended by the manufacturer.
After sequencing, image analysis and base calling were carried out using Off-Line Basecaller software (version V1.8.0; Illumina). The low quality reads were removed from raw sequencing reads by Solexa CHASTITY quality control filter. Subsequently, 3′ adapter sequences were deleted from clean reads and the reads that were shorter than 15 nt were excluded from further analysis. To identify miRNAs, the 3′-adapter-trimmed-reads were aligned to known human pre-miRNAs in the miRBase database (Release 19.0, http://www.mirbase.org/) using Novoalign software (Version v2.07.11, http://www.novocraft.com/) with at most one mismatch. In order to correct for the difference in read counts between samples, the read counts per miRNA in each sample were normalized to tpm (the clone number of transcripts per million based on the sum number of reads aligned to known human pre-miRNAs in miRBase 19.0). Differentially expressed miRNAs may be identified by a fold-change filtering. The threshold is fold-change ≥2.0 (Based on the normalized most abundant tag counts).
Transcription factor/DNA array and computational analysis
The transcription factor/DNA array analysis was carried out using TranSignal™ Protein/DNA Combo Arrays (Panomics, Redwood City, CA, USA) following the protocols provided by the manufacturer, which included 345 major transcription factors. Briefly, nuclear proteins were extracted using the Panomics nuclear extract kit (Panomics) and protein concentrations were determined by BCA protein assay kit (KangChen Bio-Tech Inc., Shanghai, China). Nuclear extract (10 μg) were incubated with 10 μl of biotin-labelled DNA binding oligonucleotides (TranSignal™ Probe Mix) for 30 min at 15°C to allow the formation of transcription factor/DNA complexes. Such complexes were isolated from the free probes by spin column separation system (Panomics). The bound DNA probes were extracted, denatured and hybridized to oligonucleotides (representing 345 consensus binding sequences for TFs) on TransSignal array membrane at 42°C overnight. Subsequently, the blots were washed and incubated with HRP and HRP substrate working solution (Millipore, Billerica, MA, USA), and exposed to Hyperfilm ECL (Amersham Pharmacia Biotech, Uppsala, Sweden). Hybridization signals were detected using GBoX Imaging System (Syngene, Cambridge, UK), and quantitative analysis of the resulting spots were performed using the ScanAlyze software (version 1.0.3, http://graphics.stanford.edu/software/scanalyze/). Any spot showing at least 2-fold increase or decrease is considered significant.
Quantitative PCR
Total RNA from each sample was reverse transcribed to generate cDNA with gene-specific primer (for miRNAs) and oligo dT primer (for TFs) using a Superscript™ III reverse transcriptase kit (Invitrogen, Carlsbad, CA, USA) in accordance with the manufacturer's instructions. QPCR was carried out in a total reaction volume of 10 μl, including 2 μl of template cDNA, 5 μl of 2X SYBR-Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA), 0.5 μl of PCR forward primer (10 μM), 0.5 μl of PCR reverse primer (10 μM) and 2 μl of double-distilled water. The reactions were incubated at 95°C for 10 min, followed by 40 cycles at 95°C for 10 sec, 60°C for 60 sec. All reactions were run in triplicate. After PCR amplification, melt curve analysis was carried out to determine the reaction specificity. Human U6 snRNA and 18S rRNA was used to normalize the expression levels of miRNAs and TFs, respectively. Expression fold-changes were calculated using the 2−ΔΔCt method (24). The differences in gene expression levels between AML samples and controls were compared using the Student's t-test. Statistical significance was set at P<0.05. All primers used in cDNA synthesis and qPCR are shown in Tables II, III and IV.
Association analysis of the different expression of miRNAs and TFs
Target gene prediction for different expression miRNAs. Differentially expressed miRNAs (fold change, ≥2.0) were subjected to bioinformatic analysis for target gene prediction mainly by combinatorial utilization of four different online databases, including miRanda (version v5, http://www.ebi.ac.uk/enright-srv/microcosm/htdocs/targets/v5/), TargetScan (Release 6.2, http://www.targetscan.org/), PicTar (Release 2007, http://pictar.mdc-berlin.de/cgi-bin/PicTar_vertebrate.cgi) and miRTarBase (Release 4.5, http://mirtarbase.mbc.nctu.edu.tw/) (25,26). To decrease the number of false-positive results, we integrated the predicted targets from miRanda, TargetScan and PicTar based on sequence complementarity, evolutionary conservation and free energy of RNA duplexes, which predicted by at least two of three databases were accepted as positive (25). (TFs were treated as genes when predicting miRNA→TF regulatory relations). Subsequently, we merged the results with experimentally validated miRNA targets from miRTarBase.
Combined AML candidate genes with miRNA target prediction
To avoid the redundancy, AML-associated genes were obtained from the MalaCards database (version 1.08.564, http://www.malacards.org/, MalaCards ID: LKM061) (27) and overlapped with the predicted miRNA targets above to form a miRNA→gene regulatory relations. The extracted miRNA and target gene pairs were then subjected to transcription factor binding site (TFBS) analysis.
TFBS prediction for different expression miRNAs and its pair target genes
To retrieve predicted TFBS information, We utilized the data underlying the TFBS Conserved Track (http://genome.ucsc.edu/cgi-bin/hgTables?hgsid=350051003&hgta_doSchemaDb=hg19&hgta_doSchemaTable=tfbsConsFactors) at the UCSC Genome Browser (28). These binding sites are conserved across the human/mouse/rat alignment. To further reduce the false-positive prediction, P-value of 0.05 was used as a cut-off for high-quality TFBSs. A gene is identified as the target of a TF if at least one TFBS is located in the gene's promoter region (5,000 bp upstream and 1,000 bp downstream of the TSS) and its P-value was <0.05. miRNA overlaps with a known host transcript (mRNA/lncRNA) and serve as a part of the same transcription unit. Therefore, the promoter for this specific transcript is used as the miRNA promoter. We identified the target miRNAs of the TFs by using the same approach as for protein-coding genes and lncRNAs. A miRNA is regarded as the target of a TF if at least one TFBS falls within the TSS region (from 5,000 bp upstream to 1,000 bp downstream) of the host gene or lncRNA and its P-value was <0.05.
Combined differentially expressed TFs with TFBS prediction
To further increase the accuracy of TFBS prediction, differentially expressed TFs (fold-change, ≥2.0) were overlapped with the predicted TFs above. Subsequently, we extracted TFs and miRNA pairs, TFs and miRNA target genes pairs to form TF→miRNA and TF→gene regulatory relations, respectively.
miRNA-TF regulatory network construction and network node analysis
The TF→miRNA and TF→gene interactions were incorporated into miRNA→gene interactions to construct a comprehensive miRNA-TF mediated regulatory network, which was visualized using Gephi software (version 0.8.1-beta, http://gephi.github.io/). For further network functional evaluation, GO analysis and KEGG pathway analysis were performed using the Functional Annotation Tool of DAVID database (version v6.7, http://david.abcc.ncifcrf.gov/) (29) to identify the functional categories enriched and pathways for the network nodes. GO term enrichment and the KEGG pathways utilized significant P-values (<0.05) relating to the nodes of miRNA-TF regulatory network. To assess network characteristics, the node degree is measured by the number of direct links of the node in the network. Nodes having a total degree ≥20 were defined as hub nodes inside the network. The subnetworks were then constructed by combining all the directly linked nodes for the hubs.
Statistical analysis
Statistical analyses were performed are presented as means ± standard deviation (SD). The using SPSS for Windows (version 13.0; SPSS, Inc.). All values statistical difference between means were analyzed by the Student's t-test and P-value <0.05 was considered significant.
Results
Aberrant miRNA and TF expression in AML
To obtain AML-related miRNAs and TFs, we analyzed miRNA and TF expression profiles in bone marrow samples of AML patients and control groups, using miRNA sequencing and TF array technology, respectively. Among the miRNA transcripts examined, we found 308 miRNAs were differentially expressed (fold-change ≥2.0) in AML samples relative to their controls (Fig. 1A), with 233 being upregulated, while 75 were downregulated. miR-206 (fold-change, 1144) was the most significantly upregulated miRNA, while miR-941 (fold-change, 184.2) was the most significantly downregulated miRNA.
The TF expression profiling data showed 84 TFs to be differentially expressed (fold-change ≥2.0) in AML samples relative to their controls (Fig. 1B), with 76 upregulated, while 8 were downregulated. Among these TFs, PREB (fold-change, 25.9) showed the highest degree of upregulation, while HiNF-B (fold-change, 2.8) was the most downregulated TF.
Quantitative PCR validation
To validate the miRNA sequencing and TF array results, 6 differentially expressed miRNAs and 6 differentially expressed TFs were selected and analyzed via qPCR. For the miRNAs, the results demonstrated that miR-155-5p, miR-9-5p and miR-100-5p were upregulated and that miR-223-3p, miR-16-5p and miR-106b-5p were down-regulated in the AML samples compared with control samples (P<0.05 for each miRNAs; Fig. 2A). For the TFs, the expression of MYC, NFKB1, NR2F1, FOXO1, FOXL1 and TFAP2A showed statistically significant differences between the two sets of bone marrow samples (all P<0.05; Fig. 2B). These qPCR results were consistent with the miRNA sequencing and TF array data.
miRNA and TF regulatory network in AML
The basic integrated regulatory network for miRNA and TF consists of three types of interactions: miRNA→gene (including miRNA→TF), TF→miRNA and TF→gene regulatory relations. After removing a few isolated nodes, we merged the three regulatory relations and constructed a miRNA-TF regulatory network for AML (Fig. 3). The numbers of nodes and pairs in the network are listed in Table V. Among the AML-related miRNAs in the network, most of (117/173) them are upregulated. Moreover, there were three pairs of miRNA→gene regulatory relationships (miR-126-5p and miR-126-3p represses PLK2 and miR-17-5p represses RUNX1) that have been experimentally confirmed in AML (30,31). Two TFs (NFKB1 and MYC) in this network have been reportedly associated with the development of AML (32,33). NFKB1 and MYC transcriptionally suppressed miR-29b-3p expression by binding to its promoter, and MYC was confirmed as a transcriptional repressor of miR-15a-5p, which were in accord with the same TF→miRNA regulatory relationships in our network (34,35). As key regulators of gene expression, miRNAs and TFs may reciprocally regulate each other to form FBLs, or co-regulate the expression of the same targets to form FFLs (18). In our miRNA-TF regulatory network, we identified 13 FBL and 1156 FFL motifs, reflecting the tight relationships between miRNAs and TFs in the network.
Table VSummary of relationships in the AML-related miRNA and transcription factor (TF) regulatory network. |
Biological functions of the miRNA-TF regulatory network
To investigate the functional groupings of the miRNA-TF regulatory network in AML, we applied the DAVID database to classify the distinct network nodes according to their GO molecular function, cellular component and biological process terms (Fig. 4). In the GO molecular function classification, a large proportion of nodes associated with binding, protein binding and nucleic acid binding (Fig. 4A). The nodes related to cellular components were mainly located in the cell parts, cell and intracellular (Fig. 4B). The GO biological processes classification showed that nodes were mainly involved in cellular process, metabolic process and biological regulation (Fig. 4C). To ascertain which pathways are enriched in the network nodes, KEGG pathway analysis was also performed by using the DAVID database. The results indicate that the nodes were significantly enriched in 33 different pathways, of which the AML pathway was the most significant followed by the pathways in cancer and Jak-STAT signaling pathways (Table VI). Among these pathways, 17 were related to hematopoiesis and leukemia. In addition, AML pathway, apoptosis pathway, Jak-STAT signaling pathway, MAPK signaling pathway, Wnt signaling pathway, p53 signaling pathway, cell cycle pathway and ABC transporter pathway have been previously experimentally verified to be involved in AML (2,36–42).
Table VIPathway analysis for network nodes in AML-related miRNA and transcription factor regulatory network. |
Network hubs and subnetworks for hubs in the miRNA-TF regulatory network
To ascertain which nodes potentially have the most influence on the overall behavior of the networks, we identified the hub nodes based on the highest total degrees inside the network. As shown in Tables VII and VIII, 22 miRNAs and 11 TFs were defined as hub nodes, which having a total degree ≥20. It is of interest that five of the 22 hub miRNAs belonged to the miR-15 family, namely miR-15a-5p/miR-15b-5p/miR-16-5p/miR-424-5p/miR-195-5p (Table VII), further indicating the important role of the miR-15 family in the regulatory network of AML. Notably, more than half of the 11 hub TFs were either notorious leukemia regulators, such as NFE2 (43) and MYC (33), or related to leukemia development and progression, such as NFKB1 (32), TCF3 (44), MAX (45) and FOXO1 (46). These findings are a preliminary reflection on the robustness of our network. Among these hub nodes, two hub miRNAs (miR-15a-5p and miR-125b-5p) and two hub TFs (MYC and NFKB1) have been previously implicated in AML (12,32,33,47). To further investigate the regulation of the four hub nodes, we extracted their subnetworks by including all their directly linked nodes in the miRNA-TF regulatory network (Fig. 5). The subnetwork analysis indicated that many pathways related to AML and other cancers (Table IX), which suggested that miR-15a-5p, miR-125b-5p, MYC and NFKB1 were important in cancer development including AML. In addition, TCF3 stood out as a promising regulator and gene, which links to all the four subnetworks. As a TF, it regulates miR-125b-5p (Fig. 5B) and MYC (Fig. 5C); while in its capacity as a gene, it is regulated by miR-15a-5p (Fig. 5A), MYC (Fig. 5C) and NFKB1 (Fig. 5D). In particular, 17 predicted target genes (CEBPA, FLT3, STAT5A, RELA, STAT5B, PIM1, SPI1, RUNX1T1, LEF1, PIM2, KIT, JUP, NRAS, EIF4EBP1, KRAS, RARA and MYC) of TCF3 in the four subnetworks participated in the AML pathway. These analyses suggest that TCF3 may play an important role in the miRNA-TF regulatory network linked to AML.
Table VIIIHub transcription factors in AML-related miRNA and transcription factor regulatory network. |
Discussion
AML is the most common form of aggressive leukemia in adults (1,2). Despite extensive efforts to elucidate the cytogenetic and molecular genetic mechanisms involved in disease occurrence and development, the pathogenesis of AML is still not fully understood due to the heterogeneity and complexity of this disease. The genomic complexity makes it difficult to identify hub regulators or genes in the pathogenesis of AML. By integrating diverse data sources, systems biology approaches provide a powerful tool for exploring the interactions and searching for key regulators between disease candidate genes on the network level, which may help to explain key aspects of disease pathogenesis and leading to new candidates for the putative therapeutic targets (48).
As major regulators of gene expression, miRNAs and TFs can function as tumor suppressors or oncogenes in a cooperative way to control gene expression, which triggering global alterations of genetic programs are involved in cell proliferation, differentiation, development and apoptosis in multiple human cancers (18). Ye et al (49) constructed a miRNA-TF regulatory network in T-cell acute lymphoblastic leukemia and demonstrated the roles of miR-19 (hub miRNA) and CYLD (hub gene) in the T-cell leukemogenesis. Therefore, understanding and applying regulatory network information for TF-miRNA-target genes could provide clues for key driver miRNAs and genes in human cancers and subsequently suggest novel therapeutic targets (50).
In the present study, we applied miRNA sequencing and TF array technology to quantitatively analyze the differential expression profiles of miRNAs and TFs, respectively, in bone marrow samples of AML patients and their age-matched controls (Fig. 1). The expression profiling data showed that 308 miRNAs and 84 TFs were differentially expressed (fold-change ≥2.0) in AML samples relative to their controls, with a subset of these findings further validated by qPCR (Fig. 2). It is difficult to achieve by only direct experimental methods to systematically infer the comprehensive miRNA-TF regulatory network, which has promoted development of computational approaches. Therefore, we integrated expression profiling data of miRNAs and TFs, computational miRNA and TF target prediction, and AML candidate genes obtained from MalaCards database to construct a miRNA and TF regulatory network specifically for AML (Fig. 3), which might provide important insights into the pathogenesis of AML. Moreover, we found that some miRNA→gene and TF→miRNA regulatory relationships in our network have been experimentally verified in previous studies. For further network functional evaluation, GO category and KEGG pathway annotation were applied to analyze the network node pool (Fig. 4 and Table VI). Among 33 enriched pathways, the AML pathway was the most significant (Table VI), more than half of these pathways were related to hematopoiesis and leukemia development, and eight pathways have been reported to be implicated in AML pathogenesis. These results preliminarily confirmed the reliability of our network.
In the miRNA-TF regulatory network, we identified 22 hub miRNAs and 11 hub TFs. With a large number of nodes linked to them, these hubs have the most influence on the networks overall behavior, which likely play critical regulatory roles in AML. To decipher this massive and complex network, and mine the key regulatory components, we subsequently constructed the subnetworks (Fig. 5) from two hub miRNAs (miR-15a-5p and miR-125b-5p) and two hub TFs (MYC and NFKB1) which were well-known regulators in AML. miR-15a-5p plays an important role in myeloid and erythroid differentiation, which was reported as a tumor suppressor in multiple myeloma, chronic lymphocytic leukemia and AML cells (51,52). Overexpression of miR-15a-5p significantly inhibited AML cell proliferation by downregulating the WT1 oncogene (51). miR-125b-5p is an oncogenic miRNA, which was highly expressed in AML carrying the t(2;11)(p21;q23) translocation and inhibited myeloid differentiation (12). Mice overexpressing miR-125b-5p causes a highly aggressive myeloid leukemia related to targeting the gene Lin28A (53). MYC is a basic transcription factor of the helix-loop-helix-leucine zipper family, which is found upregulated in AML and has been shown to block the myeloid cell differentiation (33). Retroviral transduction of MYC into murine bone marrow cells results in AML development (33,54). Similarly, sustained expression of a human MYC transgene culminated in the formation of AML in mice (33). NFKB1 is a transcription factor that is constitutively activated in primitive human AML cells, while inhibition of NFKB1 displayed a rapid induction of apoptosis in AML cells (32,55). After analyzing the topology of subnetwork, we found that TCF3 was a potential key regulator in our network due to its direct linking to all the four hubs in the subnetworks (Fig. 5), which may establish cross-talk among these subnetworks. In particular, TCF3 was predicted to regulate 17 target genes which have been involved in AML pathway in the four subnetworks. As a member of the TCF/LEF transcription factor family, TCF3 is essential for development, lymphopoiesis and stem cell function (44,56). Emerging evidence suggests that TCF3 is involved in the pathogenesis of several types of human cancer including colorectal cancer, prostate cancer and breast cancer (57–59). Furthermore, the t(1;19) translocation between TCF3 and PBX1 genes causes expression of a TCF3-PBX1 fusion protein, and has been identified in B-cell acute lymphoblastic leukemia (B-ALL), which established a potential relationship between TCF3 and the etiology of B-ALL (44). However, the exact molecular mechanism of TCF3 as it relates to AML occurrence and development is still unknown.
In the subnetworks, TCF3 was potentially targeted by miR-15a-5p, MYC and NFKB1, and also targeting miR-125b-5p, MYC, ERG and PIM1 (Fig. 5). ERG is a megakaryocytic oncogene, which is required for normal megakaryopoiesis and play crucial role in establishing definitive hematopoiesis (60). A high expression level of ERG in AML is associated with poor prognosis (61). Transgenic expression of the human ERG causes early progenitor myeloid leukemia in mice similar to human AML, which was mediated by inducing the expression of the oncogenic kinase PIM1 through binding to a novel 3′ enhancer (62). PIM1 is a constitutively active serine/threonine protein kinase that normally functions in the proliferation and survival of hematopoietic cells in response to cytokines and growth factors (63). Overexpression of PIM1 has been detected in a range of solid cancers and hematologic malignancies (64). In AML, PIM1 has been suggested to have important roles in cell proliferation, survival, homing and migration, and the prognostic relevance of PIM1 upregulation was also observed in high ERG-expressing AML by the significantly decreased survival of these patients (62). In addition, ERG was the predicted target of MYC, NFKB1 and TCF3, while PIM1 was predicted to be regulated by miR-15a-5p, MYC and TCF3 in our subnetworks (Fig. 5). The above discussion led us to present a model of the involvement of TCF3 and the four hub nodes in subnetworks within AML (Fig. 6). According to the literature surveys and our subnetwork analysis, we propose that TCF3 is a potential key regulator in the miRNA-TF regulatory network linked to AML, which may play its regulatory role via miR-15a-5p, miR-125b-5p, MYC, NFKB1, ERG or PIM1.
In summary, this is the first study that describes the global expression profiling of miRNAs and TFs relating to AML using miRNA sequencing and TF array technology, respectively. Applying systems biology approaches to integrate the experimental data to multiple types of computational prediction on gene regulation, we constructed a miRNA-TF regulatory network specifically for AML, which may provide some hub regulators and clues for exploring the molecular mechanisms of AML. Furthermore, the present study also provides potential therapeutic targets for AML and proposes directions for further experimental research.
Acknowledgements
We sincerely thank all donors who participated in this investigation.
Abbreviations:
AML |
acute myeloid leukemia |
miRNA |
microRNA |
TF |
transcription factor |
pri-miRNA |
primary miRNA |
pre-miRNA |
precursor miRNA |
3′-UTR |
3′-untranslated region |
TFBS |
transcription factor binding site |
TSS |
transcription start sites |
GO |
Gene Oncology |
KEGG pathway |
Kyoto Encydopedia of Gene and Genomes |
qPCR |
quantitative PCR |
PAGE |
polyacrylamide gel electrophoresis |
HRP |
streptavidin-conjugated horseradish peroxidase |
lncRNA |
long non-coding RNA |
FBL |
feedback loop |
FFL |
feed-forward loop |
DAVID |
the database for annotation, visualization and integrated discovery |
References
Conway O'Brien E, Prideaux S and Chevassut T: The epigenetic landscape of acute myeloid leukemia. Adv Hematol. 2014:1031752014.PubMed/NCBI | |
Lee HJ, Daver N, Kantarjian HM, Verstovsek S and Ravandi F: The role of JAK pathway dysregulation in the pathogenesis and treatment of acute myeloid leukemia. Clin Cancer Res. 19:327–335. 2013. View Article : Google Scholar | |
Zaidi SK, Trombly DJ, Dowdy CR, Lian JB, Stein JL, van Wijnen AJ and Stein GS: Epigenetic mechanisms in leukemia. Adv Biol Regul. 52:369–376. 2012. View Article : Google Scholar : PubMed/NCBI | |
Jamieson K and Odenike O: Late-phase investigational approaches for the treatment of relapsed/refractory acute myeloid leukemia. Expert Opin Pharmacother. 13:2171–2187. 2012. View Article : Google Scholar : PubMed/NCBI | |
Sabnis H, Bradley HL, Bunting ST, Cooper TM and Bunting KD: Capillary nano-immunoassay for Akt 1/2/3 and 4EBP1 phosphorylation in acute myeloid leukemia. J Transl Med. 12:1662014. View Article : Google Scholar : PubMed/NCBI | |
Naoe T and Kiyoi H: Gene mutations of acute myeloid leukemia in the genome era. Int J Hematol. 97:165–174. 2013. View Article : Google Scholar : PubMed/NCBI | |
Larsson CA, Cote G and Quintás-Cardama A: The changing mutational landscape of acute myeloid leukemia and myelodysplastic syndrome. Mol Cancer Res. 11:815–827. 2013. View Article : Google Scholar : PubMed/NCBI | |
Abdel-Wahab O and Levine RL: Mutations in epigenetic modifiers in the pathogenesis and therapy of acute myeloid leukemia. Blood. 121:3563–3572. 2013. View Article : Google Scholar : PubMed/NCBI | |
Yan S, Yim LY, Lu L, Lau CS and Chan VS: MicroRNA regulation in systemic lupus erythematosus pathogenesis. Immune Netw. 14:138–148. 2014. View Article : Google Scholar : PubMed/NCBI | |
van Rooij E and Kauppinen S: Development of microRNA therapeutics is coming of age. EMBO Mol Med. 6:851–864. 2014. View Article : Google Scholar : PubMed/NCBI | |
Babashah S, Sadeghizadeh M, Tavirani MR, Farivar S and Soleimani M: Aberrant microRNA expression and its implications in the pathogenesis of leukemias. Cell Oncol (Dordr). 35:317–334. 2012. View Article : Google Scholar | |
Marcucci G, Mrózek K, Radmacher MD, Garzon R and Bloomfield CD: The prognostic and functional role of microRNAs in acute myeloid leukemia. Blood. 117:1121–1129. 2011. View Article : Google Scholar : | |
Han YC, Park CY, Bhagat G, Zhang J, Wang Y, Fan JB, Liu M, Zou Y, Weissman IL and Gu H: microRNA-29a induces aberrant self-renewal capacity in hematopoietic progenitors, biased myeloid development, and acute myeloid leukemia. J Exp Med. 207:475–489. 2010. View Article : Google Scholar : PubMed/NCBI | |
Bai J, Guo A, Hong Z and Kuai W: Upregulation of microRNA-100 predicts poor prognosis in patients with pediatric acute myeloid leukemia. Onco Targets Ther. 5:213–219. 2012.PubMed/NCBI | |
Wang Z, Hong Z, Gao F and Feng W: Upregulation of microRNA-375 is associated with poor prognosis in pediatric acute myeloid leukemia. Mol Cell Biochem. 383:59–65. 2013. View Article : Google Scholar : PubMed/NCBI | |
Li Y, Gao L, Luo X, Wang L, Gao X, Wang W, Sun J, Dou L, Li J, Xu C, et al: Epigenetic silencing of microRNA-193a contributes to leukemogenesis in t(8;21) acute myeloid leukemia by activating the PTEN/PI3K signal pathway. Blood. 121:499–509. 2013. View Article : Google Scholar | |
Hickey CJ, Schwind S, Radomska HS, Dorrance AM, Santhanam R, Mishra A, Wu YZ, Alachkar H, Maharry K, Nicolet D, et al: Lenalidomide-mediated enhanced translation of C/EBPα-p30 protein up-regulates expression of the antileukemic microRNA-181a in acute myeloid leukemia. Blood. 121:159–169. 2013. View Article : Google Scholar : | |
Zhang HM, Kuang S, Xiong X, Gao T, Liu C and Guo AY: Transcription factor and microRNA co-regulatory loops: Important regulatory motifs in biological processes and diseases. Brief Bioinform. 16:45–58. 2015. View Article : Google Scholar | |
Cheng C, Yan KK, Hwang W, Qian J, Bhardwaj N, Rozowsky J, Lu ZJ, Niu W, Alves P, Kato M, et al: Construction and analysis of an integrated regulatory network derived from high-throughput sequencing data. PLoS Comput Biol. 7:e10021902011. View Article : Google Scholar : PubMed/NCBI | |
Pulikkan JA, Dengler V, Peramangalam PS, Peer Zada AA, Müller-Tidow C, Bohlander SK, Tenen DG and Behre G: Cell-cycle regulator E2F1 and microRNA-223 comprise an auto-regulatory negative feedback loop in acute myeloid leukemia. Blood. 115:1768–1778. 2010. View Article : Google Scholar : | |
Katzerke C, Madan V, Gerloff D, Bräuer-Hartmann D, Hartmann JU, Wurm AA, Müller-Tidow C, Schnittger S, Tenen DG, Niederwieser D, et al: Transcription factor C/EBPα-induced microRNA-30c inactivates Notch1 during granulopoiesis and is downregulated in acute myeloid leukemia. Blood. 122:2433–2442. 2013. View Article : Google Scholar : PubMed/NCBI | |
Padró T, Ruiz S, Bieker R, Bürger H, Steins M, Kienast J, Büchner T, Berdel WE and Mesters RM: Increased angiogenesis in the bone marrow of patients with acute myeloid leukemia. Blood. 95:2637–2644. 2000. | |
Aurelius J, Martner A, Brune M, Palmqvist L, Hansson M, Hellstrand K and Thoren FB: Remission maintenance in acute myeloid leukemia: Impact of functional histamine H2 receptors expressed by leukemic cells. Haematologica. 97:1904–1908. 2012. View Article : Google Scholar : PubMed/NCBI | |
Pfaffl MW: A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 29:e452001. View Article : Google Scholar : PubMed/NCBI | |
Sui W, Lin H, Peng W, Huang Y, Chen J, Zhang Y and Dai Y: Molecular dysfunctions in acute rejection after renal transplantation revealed by integrated analysis of transcription factor, microRNA and long noncoding RNA. Genomics. 102:310–322. 2013. View Article : Google Scholar : PubMed/NCBI | |
Lin Y, Wu J, Chen H, Mao Y, Liu Y, Mao Q, Yang K, Zheng X and Xie L: Cyclin-dependent kinase 4 is a novel target in micoRNA-195-mediated cell cycle arrest in bladder cancer cells. FEBS Lett. 586:442–447. 2012. View Article : Google Scholar : PubMed/NCBI | |
Rappaport N, Twik M, Nativ N, Stelzer G, Bahir I, Stein TI, Safran M and Lancet D: MalaCards: A comprehensive automatically-mined database of human diseases. Curr Protoc Bioinformatics. 47:1.24.1–1.24.19. 2014. View Article : Google Scholar | |
Felice B, Cattoglio C, Cittaro D, Testa A, Miccio A, Ferrari G, Luzi L, Recchia A and Mavilio F: Transcription factor binding sites are genetic determinants of retroviral integration in the human genome. PLoS One. 4:e45712009. View Article : Google Scholar : PubMed/NCBI | |
Huang DW, Sherman BT, Tan Q, Kir J, Liu D, Bryant D, Guo Y, Stephens R, Baseler MW, Lane HC, et al: DAVID Bioinformatics Resources: Expanded annotation database and novel algorithms to better extract biology from large gene lists. Nucleic Acids Res. 35(Web Server): W169–W175. 2007. View Article : Google Scholar : PubMed/NCBI | |
Li Z, Lu J, Sun M, Mi S, Zhang H, Luo RT, Chen P, Wang Y, Yan M, Qian Z, et al: Distinct microRNA expression profiles in acute myeloid leukemia with common translocations. Proc Natl Acad Sci USA. 105:15535–15540. 2008. View Article : Google Scholar : PubMed/NCBI | |
Fischer J, Rossetti S, Datta A, Eng K, Beghini A and Sacchi N: miR-17 deregulates a core RUNX1-miRNA mechanism of CBF acute myeloid leukemia. Mol Cancer. 14:72015. View Article : Google Scholar : PubMed/NCBI | |
Grosjean-Raillard J, Adès L, Boehrer S, Tailler M, Fabre C, Braun T, De Botton S, Israel A, Fenaux P and Kroemer G: Flt3 receptor inhibition reduces constitutive NFkappaB activation in high-risk myelodysplastic syndrome and acute myeloid leukemia. Apoptosis. 13:1148–1161. 2008. View Article : Google Scholar : PubMed/NCBI | |
Brondfield S, Umesh S, Corella A, Zuber J, Rappaport AR, Gaillard C, Lowe SW, Goga A and Kogan SC: Direct and indirect targeting of MYC to treat acute myeloid leukemia. Cancer Chemother Pharmacol. 76:35–46. 2015. View Article : Google Scholar : PubMed/NCBI | |
Mott JL, Kurita S, Cazanave SC, Bronk SF, Werneburg NW and Fernandez-Zapico ME: Transcriptional suppression of mir-29b-1/mir-29a promoter by c-Myc, hedgehog, and NF-kappaB. J Cell Biochem. 110:1155–1164. 2010. View Article : Google Scholar : PubMed/NCBI | |
Tao J, Zhao X and Tao J: c-MYC-miRNA circuitry: A central regulator of aggressive B-cell malignancies. Cell Cycle. 13:191–198. 2014. View Article : Google Scholar : PubMed/NCBI | |
Steffen B, Müller-Tidow C, Schwäble J, Berdel WE and Serve H: The molecular pathogenesis of acute myeloid leukemia. Crit Rev Oncol Hematol. 56:195–221. 2005. View Article : Google Scholar : PubMed/NCBI | |
Schimmer AD: Novel therapies targeting the apoptosis pathway for the treatment of acute myeloid leukemia. Curr Treat Options Oncol. 8:277–286. 2007. View Article : Google Scholar : PubMed/NCBI | |
Kim HS, Lim GY, Hwang J, Ryoo ZY, Huh TL and Lee S: Induction of apoptosis by obovatol as a novel therapeutic strategy for acute myeloid leukemia. Int J Mol Med. 34:1675–1680. 2014.PubMed/NCBI | |
Mikesch JH, Steffen B, Berdel WE, Serve H and Müller-Tidow C: The emerging role of Wnt signaling in the pathogenesis of acute myeloid leukemia. Leukemia. 21:1638–1647. 2007. View Article : Google Scholar : PubMed/NCBI | |
Li L, Tan Y, Chen X, Xu Z, Yang S, Ren F, Guo H, Wang X, Chen Y, Li G, et al: MDM4 overexpressed in acute myeloid leukemia patients with complex karyotype and wild-type TP53. PLoS One. 9:e1130882014. View Article : Google Scholar : PubMed/NCBI | |
Chim CS, Wong AS and Kwong YL: Epigenetic inactivation of INK4/CDK/RB cell cycle pathway in acute leukemias. Ann Hematol. 82:738–742. 2003. View Article : Google Scholar : PubMed/NCBI | |
Fukuda Y, Lian S and Schuetz JD: Leukemia and ABC transporters. Adv Cancer Res. 125:171–196. 2015. View Article : Google Scholar : PubMed/NCBI | |
Li YJ, Higgins RR, Pak BJ, Shivdasani RA, Ney PA, Archer M and Ben-David Y: p45NFE2 is a negative regulator of erythroid proliferation which contributes to the progression of friend virus-induced erythroleukemias. Mol Cell Biol. 21:73–80. 2001. View Article : Google Scholar | |
Somasundaram R, Prasad MA, Ungerbäck J and Sigvardsson M: Transcription factor networks in B-cell differentiation link development to acute lymphoid leukemia. Blood. 126:144–152. 2015. View Article : Google Scholar : PubMed/NCBI | |
Zada AA, Pulikkan JA, Bararia D, Geletu M, Trivedi AK, Balkhi MY, Hiddemann WD, Tenen DG, Behre HM and Behre G: Proteomic discovery of Max as a novel interacting partner of C/EBPalpha: A Myc/Max/Mad link. Leukemia. 20:2137–2146. 2006. View Article : Google Scholar : PubMed/NCBI | |
Pellicano F, Scott MT, Helgason GV, Hopcroft LE, Allan EK, Aspinall-O'Dea M, Copland M, Pierce A, Huntly BJ, Whetton AD, et al: The antiproliferative activity of kinase inhibitors in chronic myeloid leukemia cells is mediated by FOXO transcription factors. Stem Cells. 32:2324–2337. 2014. View Article : Google Scholar : PubMed/NCBI | |
Gao SM, Xing CY, Chen CQ, Lin SS, Dong PH and Yu FJ: miR-15a and miR-16-1 inhibit the proliferation of leukemic cells by down-regulating WT1 protein level. J Exp Clin Cancer Res. 30:1102011. View Article : Google Scholar : PubMed/NCBI | |
Cho JH, Gelinas R, Wang K, Etheridge A, Piper MG, Batte K, Dakhallah D, Price J, Bornman D, Zhang S, et al: Systems biology of interstitial lung diseases: Integration of mRNA and microRNA expression changes. BMC Med Genomics. 4:82011. View Article : Google Scholar : PubMed/NCBI | |
Ye H, Liu X, Lv M, Wu Y, Kuang S, Gong J, Yuan P, Zhong Z, Li Q, Jia H, et al: MicroRNA and transcription factor co-regulatory network analysis reveals miR-19 inhibits CYLD in T-cell acute lymphoblastic leukemia. Nucleic Acids Res. 40:5201–5214. 2012. View Article : Google Scholar : PubMed/NCBI | |
Yan Z, Shah PK, Amin SB, Samur MK, Huang N, Wang X, Misra V, Ji H, Gabuzda D and Li C: Integrative analysis of gene and miRNA expression profiles with transcription factor-miRNA feed-forward loops identifies regulators in human cancers. Nucleic Acids Res. 40:e1352012. View Article : Google Scholar : PubMed/NCBI | |
Gao SM, Xing CY, Chen CQ, Lin SS, Dong PH and Yu FJ: miR-15a and miR-16-1 inhibit the proliferation of leukemic cells by down-regulating WT1 protein level. J Exp Clin Cancer Res. 30:1102011. View Article : Google Scholar : PubMed/NCBI | |
Aqeilan RI, Calin GA and Croce CM: miR-15a and miR-16-1 in cancer: Discovery, function and future perspectives. Cell Death Differ. 17:215–220. 2010. View Article : Google Scholar | |
Chaudhuri AA, So AY, Mehta A, Minisandram A, Sinha N, Jonsson VD, Rao DS, O'Connell RM and Baltimore D: Oncomir miR-125b regulates hematopoiesis by targeting the gene Lin28A. Proc Natl Acad Sci USA. 109:4233–4238. 2012. View Article : Google Scholar : PubMed/NCBI | |
Luo H, Li Q, O'Neal J, Kreisel F, Le Beau MM and Tomasson MH: c-Myc rapidly induces acute myeloid leukemia in mice without evidence of lymphoma-associated antiapoptotic mutations. Blood. 106:2452–2461. 2005. View Article : Google Scholar : PubMed/NCBI | |
Guzman ML, Neering SJ, Upchurch D, Grimes B, Howard DS, Rizzieri DA, Luger SM and Jordan CT: Nuclear factor-kappaB is constitutively activated in primitive human acute myelogenous leukemia cells. Blood. 98:2301–2307. 2001. View Article : Google Scholar : PubMed/NCBI | |
Howard JM, Nuguid JM, Ngole D and Nguyen H: Tcf3 expression marks both stem and progenitor cells in multiple epithelia. Development. 141:3143–3152. 2014. View Article : Google Scholar : PubMed/NCBI | |
Slyper M, Shahar A, Bar-Ziv A, Granit RZ, Hamburger T, Maly B, Peretz T and Ben-Porath I: Control of breast cancer growth and initiation by the stem cell-associated transcription factor TCF3. Cancer Res. 72:5613–5624. 2012. View Article : Google Scholar : PubMed/NCBI | |
Li C, Cai S, Wang X and Jiang Z: Hypomethylation-associated up-regulation of TCF3 expression and recurrence in stage II and III colorectal cancer. PLoS One. 9:e1120052014. View Article : Google Scholar : PubMed/NCBI | |
Patel D and Chaudhary J: Increased expression of bHLH transcription factor E2A (TCF3) in prostate cancer promotes proliferation and confers resistance to doxorubicin induced apoptosis. Biochem Biophys Res Commun. 422:146–151. 2012. View Article : Google Scholar : PubMed/NCBI | |
Salek-Ardakani S, Smooha G, de Boer J, Sebire NJ, Morrow M, Rainis L, Lee S, Williams O, Izraeli S and Brady HJ: ERG is a megakaryocytic oncogene. Cancer Res. 69:4665–4673. 2009. View Article : Google Scholar : PubMed/NCBI | |
Marcucci G, Maharry K, Whitman SP, Vukosavljevic T, Paschka P, Langer C, Mrózek K, Baldus CD, Carroll AJ, Powell BL, et al; Cancer and Leukemia Group B Study. High expression levels of the ETS-related gene, ERG, predict adverse outcome and improve molecular risk-based classification of cytogenetically normal acute myeloid leukemia: A Cancer and Leukemia Group B Study. J Clin Oncol. 25:3337–3343. 2007. View Article : Google Scholar : PubMed/NCBI | |
Goldberg L, Tijssen MR, Birger Y, Hannah RL, Kinston SJ, Schütte J, Beck D, Knezevic K, Schiby G, Jacob-Hirsch J, et al: Genome-scale expression and transcription factor binding profiles reveal therapeutic targets in transgenic ERG myeloid leukemia. Blood. 122:2694–2703. 2013. View Article : Google Scholar : PubMed/NCBI | |
Burger MT, Han W, Lan J, Nishiguchi G, Bellamacina C, Lindval M, Atallah G, Ding Y, Mathur M, McBride C, et al: Structure guided optimization, in vitro activity, and in vivo activity of Pan-PIM kinase inhibitors. ACS Med Chem Lett. 4:1193–1197. 2013. View Article : Google Scholar | |
Shah N, Pang B, Yeoh KG, Thorn S, Chen CS, Lilly MB and Salto-Tellez M: Potential roles for the PIM1 kinase in human cancer - a molecular and therapeutic appraisal. Eur J Cancer. 44:2144–2151. 2008. View Article : Google Scholar : PubMed/NCBI |