Introduction
Mammalian microRNAs (miRNA, miR) are short non-coding RNAs that regulate preferentially gene expression by inhibiting translation of specific target mRNAs. MiRNA-mRNA matching is based on imperfect sequence base-pairing with the required complementarity centered over positions 2 - 8 of mRNA's seed sequence [
1]. Depending on specific target genes, miRNAs regulate many cellular functions such as developmental timing, signal transduction, apoptosis, cell proliferation and tumorigenesis [
2‐
5]. Thus, gene expression and role of miRNAs are currently being largely studied in human malignancies and chemical compounds that regulate miRNA levels are potentially very important for developing new treatment strategies in chronic myeloid leukemia (CML). The first miRNA molecules that have been associated with human leukemia pathogenesis were found in chronic lymphocytic leukemia (CLL) [
6]. MiR-15 and miR-16 are located in a genomic region that is frequently deleted in CLL, thus the expression of these two miRNAs is downregulated. Other works brought the evidence that many miRNAs are indeed found at chromosomal breakpoints and genomic regions associated with cancer [
7,
8].
In CML the following miRNAs were associated with the disease pathogenesis. For instance, the miR-203 was found to be epigenetically silenced in human leukemic Philadelphia chromosome-positive (Ph+) cell lines; this is in line with the observation that BCR-ABL and ABL kinases are miR-203 putative targets [
9]. Derivative 9q+ chromosome deletions carrying miR-199b that occurred in some CML patients were associated with miR-199b decrease [
10]. Venturini et al. [
11] showed miR-17-92 cluster (onkomir-1) to be aberrantly expressed in CD34+ cells of CML patients. Agirre at al. [
12] analyzed the expression of 157 miRNAs in mononuclear and CD34+ cells separated from bone marrow of 6 CML patients at diagnosis and found 11 miRNAs (e.g. miR-150, miR-151, miR-221, miR-127, miR-16) aberrantly expressed in CD34+ cells and 53 miRNAs differentially expressed in mononuclear cells (e.g. miR-150, miR-126, miR-221, miR-222, miR-21). Two recent works contributed to the knowledge about expression change in specific microRNAs associated with resistance to imatinib or responsiveness to imatinib after the treatment initiation in CML patients [
13,
14]. A group of 19 miRNAs (e.g. miR-191, miR-29a, miR-422b, miR-100, miR-326, miR-26a) were identified as possible predictors for clinical resistance to imatinib in patients with newly diagnosed CML [
13]. A relatively rapid increase in the expression of miR-150 and miR-146a and decrease of miR-142-3p and miR-199b-5p in peripheral blood mononuclear cells (PBMCs) of patients newly diagnosed with CML was found two weeks after imatinib initiation [
14].
In this study, we used an array platform to characterize differentially expressed miRNAs in peripheral blood total leukocytes of patients at different stages of CML including diagnosis, major molecular response, therapy failure, hematological relapse, accelerated phase and blast crisis with the aim to identify microRNAs associated with pathogenesis of CML. To the best of our knowledge, such integrated microRNA profiling during the course of CML has not yet been performed. Hierarchical clustering analysis based on expression profiles of 49 miRNAs clearly separated patients at diagnosis, hematological relapse and blast crisis from those in major molecular response and therapy failure. We used
in silico analyses to better understand the targets of 17 selected miRNAs whose deregulation was confirmed by real-time quantitative PCR (RT-qPCR). Based on our previous results demonstrating that miR-150 downregulation is associated with CML [
15], we further validated miR-150 expression in a larger number of patients (n = 70). As
MYB represents functionally validated target of miR-150 [
16], its gene expression analysis was performed on the same patient cohort. Our data provide significant inverse correlations between miR-150 and
MYB expression and
BCR-ABL transcript level and indicate that this relationship is potentially important for pathogenesis in CML.
Discussion
Specific microRNAs regulate hematopoietic cell differentiation and development [
26]. The main interest is in whether there exists a link between levels of miRNAs and leukemia pathogenesis. The first work dealing with miRNA expression in CML demonstrated enhanced expression of the miR-17-92 cluster in CML CD34+ cells [
11]. Other works that reported miRNA aberrant expression in CML appeared very recently. For example, it demonstrated that several miRNAs dysregulated in CML (miR-150, miR-146a, miR-142-3p, miR-199b-5p) were rapidly restored under imatinib treatment [
14]. Several miRNAs (e.g. miR-191, miR-29a, miR-422b, miR-100, miR-326, miR-26a) are promising predictors of imatinib resistance in newly diagnosed CML patients [
13].
This study investigates microRNA differential expression profiles that were initially analyzed at different stages of CML using microarrays. Pooling of patient samples was applied for microarray analysis to reduce individual variability and to find common features of the disease.
MiRNA array data showed similar expression pattern of 49 miRNAs in imatinib responders with MMR and patients with failure to achieve complete cytogenetic response. As expected, hierarchical clustering assembled the pools of samples at diagnosis, in hematological relapse and blast crisis, while MMR and TF pools formed a separate cluster (Figure
1). Total leukocytes from blast crisis peripheral blood that consisted of more than 50% blasts of each sample in the pool showed the highest number of strongly deregulated miRNAs.
We applied the functional annotation tool DAVID to look for the biological functions of predicted targets with only conserved sites and high P
CT values of the 17 miRNAs with real-time qPCR-confirmed up-regulation (miR-19a, miR-19b, miR-17, miR-20a, miR-92a, miR-221, miR-222, miR-126, miR-146a, miR-181a, miR-181b, let7c and miR-155) and down-regulation (miR-103, miR-150, miR-451 and miR-144) in blast crisis (Figure
3). Several targets were involved in the processes that were found to be important in CML; endocytosis [
27], mTOR signaling pathway [
28,
29], hedgehog signaling [
30,
31], focal adhesion [
32,
33] and Wnt signaling [
34,
35]. We summarized 19 genes with the probability to be targeted by miR-20a, miR-17, miR-19a, miR-103, miR-144, miR-150, miR-155, miR-181a, miR-221 and miR-222. The encoded proteins were annotated in pathways related to the CML (hsa05220). Out of these, 10 targets are involved in MAPK signaling (BCR, E2F2, E2F3, CBL, RAF1, CRK, CRKL, KRAS, SOS1, MAPK1). Interestingly, inhibition of MAPK signaling in Ph+ cell line K562 induced apoptosis [
36]. Application of MAPK specific inhibitor U0126 showed synergistic effect with imatinib resulting in CD34+ progenitor reduction in CML [
36].
Confirmed increase of miR-19a, miR-19b, miR-17, miR- 20a, miR-92a, miR-106a, miR-221, miR-222, miR-126, miR-146a, miR-181a, miR-181b, let7c and miR-155 was identified in samples of BC pool (Figure
3). This pattern was not found in Dg, Hr, TF or MMR pools. Overexpression of these miRs may be related to the immature character of blasts. Whether the increased level of these miRNAs may contribute to the CML pathogenesis or may simply reflect the stage of the disease is the matter of further investigation. Abnormal expression of onkomir miR-17-92 (miR-17; miR-19a; miR-19b; miR-20a; miR-92a) was described in CML CD34+ cells [
11]. Agirre et al. [
12] found up-regulated miR-221 and miR-222 in mononuclear cells of CML patients in comparison to healthy controls. MiR-155, miR-106a, miR-146a, miR-181 and miR-126 were reported as deregulated miRNAs in CML [
13,
14]. To our knowledge, let7c expression has so far not been described in CML. In this study, our
in silico analyses revealed that miR-221 and miR-103 (P
CT 0.63 and 0.73, respectively) target PIK3R1. PIK3R3 is predicted to be regulated by miR-19a and miR-181a (P
CT 0.92 and 0.81, respectively). PI3K is annotated in ERBB, MAPK and mTOR signaling pathways. KRAS, which is involved in MAPK signaling, is a predicted target of miR-19a (P
CT 0.92). MAPK expression may be regulated by onkomirs miR-17 and miR-19a (P
CT 0.96 and 0.86, respectively). Interestingly, it was reported that RAS/MAPK signaling may contribute to the survival of BCR-ABL positive cells under imatinib selection pressure [
37]. AKT1, a member of the antiapoptotic PI3K pathway, is involved in both, BCR-ABL mediated transformation as well as in response to the BCR-ABL kinase inhibitors. It was shown that the PI3K/AKT/mTOR signaling is activated in imatinib naive cells while under imatinib pressure it may enhance resistance to imatinib [
38]. As shown in our real-time qPCR data (Figure
3), the rather decreased levels of miR-181a, miR-221 and miR-19a in some imatinib treated patients, and miR-103 down-regulation in a number of blast crisis, diagnosis and progressed CML may contribute to the increased level of PI3K and thus may be involved in the previously described PI3K/AKT/mTOR signaling activation and in the resistance development in some CML cases. Though no experimental therapy using miRNA modulation has as yet provided significant and curative approach, the knowledge of deregulation of miRNAs specific for CML may facilitate the development of such therapeutic strategies. Several candidate microRNAs (e.g. miR-181a, miR-221, miR-19a, miR-103) regulating expression in CML target important signaling pathways may represent promising candidate targets for CML therapy.
The real-time qPCR validated the down-regulation of miR-150, miR-451, miR-103 and miR-144 overall in individual samples of BC, Hr, Dg pools and in some samples of TF pool (Figure
3). These molecules may be related to the CML pathogenesis and may reflect transformation from chronic to accelerated phases. Agirre et al. found miR-150 downregulation in mononuclear cells and CD34+ cells separated from bone marrow in newly diagnosed CML patients (n = 6) in comparison to healthy donors (n = 6) [
12]. MiR-150 was recently described to be downregulated in untreated CML patients [
14]. Flamant et al. [
14] suggest that miR-150 play a role in leukemic cells and potentially in the more primitive hematopoietic compartment in chronic phase CML patients. This is in line with the knowledge that miR-150 is important in the regulation of hematopoiesis. During normal erythroid differentiation its level is gradually decreased [
39], however; it shows the highest expression in mature lymphocytes [
40]. Others proved that miR-150 expression increases during B-lymphoid differentiation in contrast to myeloid differentiation. It seems likely that miR-150 regulates the development of other two different blood lineages; B lymphocytes and megakaryocytes [
41,
42]. Thus, miR-150 deregulation is found in hematological malignancies; miR-150 is decreased in polycytemia vera reticulocytes [
43] and a marked decrease was recently also detected in MDS-del(5q) [
44] while, in contrast, a twofold increase was found in CLL lymphocytes [
45].
Based on our results [
15] and recent results of others we expanded real-time qPCR assays of miR-150 on the larger cohort of CML patients. Decreased level of miR-150 was confirmed in patients at diagnosis, in the majority of patients with hematological relapse and in accelerated phase and blast crisis. Normal miR-150 level was observed in imatinib treated patients with major molecular response and failure to achieve CCgR. Our observations are consistent with the data of Flamant et al. [
14] showing rapid increase of miR-150 expression after imatinib treatment initiation in patients with newly diagnosed CML. They further found that low miR-150 expression inversely correlated with white blood count and thus speculated that the level reflected the high leukocyte counts in newly diagnosed CML patients. We showed here a significant inverse correlation of miR-150 expression with
BCR-ABL transcript level (p = 0.01; r = -0.501). Non-treated newly diagnosed patients, patients with disease progression and resistant to imatinib showed a high level of
BCR-ABL together with high leukocyte count and decreased amount of miR-150. Normal miR-150 level was detected in imatinib responders (MMR) and patients with failure to achieve CCgR (TF) with normal blood count and low
BCR-ABL transcript level. As imatinib targets Ph+ cells, the normal level of miR-150 in imatinib treated patients in chronic phase with physiological blood count could be the result of the suppression of leukemic cells and the concomitant recovery of normal hematopoiesis under imatinib treatment. Our
in vitro tests showed elevated expression of miR-150 and marked decrease of p-CRKL following imatinib
in vitro treatment of Ph+ cell line MOLM-7. These findings suggest a potential functional relationship between miR-150 and BCR-ABL.
Gene expression of
MYB in our study showed a significant inverse correlation with miR-150 transcript level (p = 0.01; r = -0,409).
MYB is the proven target of miR-150 and encodes a transcriptional factor required for proliferation and survival of normal and leukemic blast cells. A recently published study on a mouse model of blast crisis reported that
c-MYB is required for BCR-ABL dependent leukemogenesis [
46]. Lidonnici et al. [
46] speculated that miR-150 reduction might contribute to the
c-MYB upregulation that is likely induced by BCR-ABL, and may be involved in BCR-ABL driven leukemogenesis in CML. Interestingly, we found a significant correlation between
MYB expression and
BCR-ABL transcript level (p = 0.01; r = 0,782) in CML patients, which is in line with the above described suggestion.
In summary, our data demonstrated that miR-150, miR-20a, miR-17, miR-19a, miR-103, miR-144, miR-155, miR-181a, miR-221 and miR-222 are deregulated in CML. Furthermore, in silico filtering identified targeted genes that are involved in cell cycle, growth inhibition, MAPK, ErBb, transforming growth factor beta and p53 signaling pathways that are reported in CML pathogenesis. MiR-150 expression showed significant negative correlation with its target MYB and with BCR-ABL transcript level. The results of this study outline the mechanisms whereby miRNAs may be implicated in CML pathogenesis. However, if they function in BCR-ABL dependent or independent manner has to be elucidated.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KMP - conception and design, in silico analyses, data evaluation and interpretation, manuscript drafting; TL - contribution to the manuscript drafting, real-time qPCR analyses and evaluation; HK - provision of patient samples, clinical data evaluation, critical revisions; PB - in vitro tests, contribution to data evaluation and interpretation, critical revision; MT - clinical data revision; critical revisions of the manuscript; TS - contribution to data evaluation and interpretation, critical revision; JM - critical revisions. All authors read and approved the final manuscript.