Sotirios Papageorgiou, MD, PhD, Second Department of Internal Medicine Propaedeutic, Haematology Unit, University General Hospital “Atttikon”, 1 Rimini str., 124 62 Haidari, Greece. Tel: +30 210 58 31 663; Fax: +30 210 53 26 454; e-mail: email@example.com
MicroRNAs are small non-coding RNAs that act at the post-transcriptional level, regulating protein expression by repressing translation or destabilizing mRNA target. Because of their discovery, microRNAs have been associated with almost every normal cell function, including proliferation, differentiation and apoptosis. Several lines of evidence suggest that they have an important role in normal hematopoiesis as exemplified by the role of mir-155 and mir-150 in the differentiation of B and T lymphocytes, the suppressive role of mir-221 and mir-222 in erythroid differentiation, the inhibitory effect of mir-181 on hematopoietic differentiation and the induction of myeloid differentiation by mir-223. Moreover, they play a role both as oncogenes, probably by a variety of mechanisms, namely through elimination of tumor suppressor proteins, or as tumor suppressor genes by targeting oncogenic mRNAs. Their aberrant expression has been associated with solid tumors and hematopoietic malignancies as suggested by the frequent deletion of mir-15a and mir-16-1 in chronic lymphocytic leukemia, the increased levels of mir-155 in diffuse large B-cell lymphomas and the increased levels of mir-181 in acute myeloid leukemia M1 and M2. The purpose of this review is to summarize current knowledge on the role of microRNAs in normal hematopoiesis and hematopoietic malignancies and, moreover, to highlight their role as potential therapeutic tools.
MicroRNAs are small (19–24 nucleotide), non-protein-coding RNAs that regulate post-transcriptional gene expression by inhibiting protein translation or destabilizing target transcripts. The first microRNA, lin-4, was described in Caenorhabditis elegans in 1993 (1, 2). Since 2001, when the term microRNA was used for the first time (3), a great progress has been made in understanding the role of microRNAs in normal and malignant hematopoiesis, and these recently discovered small molecules have been found to play critical roles in many different cell procedures such as differentiation, proliferation and apoptosis (programmed cell death).
Most human microRNAs are encoded within protein-coding genes and in both introns and exons of mRNA-like non-coding RNAs (4). The biosynthesis of microRNAs, which begins in the nucleus and is completed in the cytoplasm (5–18), is illustrated in Fig. 1. MicroRNAs recognize target sites in the 3′-untranslated regions (UTRs) of mRNAs and probably in the 5′-UTRs through perfect (in plants) or imperfect (in mammals) base-pairing (19). They bind to the target transcript by their nucleotides 2–8, counted from the 5′ end. These nucleotides constitute the ‘seed’ region of the microRNA (18, 20).
MicroRNA genes constitute about 1–5% of the predicted genes in worms, mice and humans. At present, miRBase (microrna.sanger.ac.uk) contains 706 microRNAs that have been cloned in human, but their number is predicted to be up to 2000 or even more. Every type of cell seems to have a specific microRNA profile. Each microRNA can have one or more target transcripts while each transcript may be regulated by one or more microRNAs. According to miRBase, every human microRNA is predicted to regulate the expression of approximately 1000 mRNAs or even more. In addition, the recent study of Calin et al. (21) seems to confirm these computational predictions, because it demonstrates that a single cluster (mir-15a/16-1 in this case) up- or down-regulates directly or indirectly 14% of total genes in human genome. These impressive data imply that microRNAs participate in a great number of overlapping pathways in human protein expression.
MicroRNAs play a very important role in normal hematopoiesis because they regulate hematopoietic differentiation in almost every stage, while their aberrant expression has been associated with many diseases, including hematological malignancies. MicroRNAs can act as oncogenes and tumor suppressors by mechanisms not completely clarified. However, there are some clues that allow us to understand their role in tumorigenesis. As previously stated, microRNAs down-regulate the expression of their target transcript or induce its degradation resulting in reduced amounts of the protein encoded by the target mRNA. Oncogenic microRNAs target and eliminate the expression of tumor suppressing mRNAs. Increased levels of oncogenic microRNAs eliminate the expression of tumor suppressing proteins, increasing cell proliferation and/or inhibiting cell death. Tumor suppressing microRNAs, on the other hand, target oncoprotein encoding mRNAs. Reduced levels of these microRNAs cause an increase in the levels of oncoproteins favoring cell survival or proliferation (reviewed by Waldman et al.) (22). There is evidence that microRNAs may contribute to tumorigenesis even in a direct way. Costinean et al. (23), for example, demonstrated that transgenic (TG) mice, which over-express mir-155, developed a B-cell malignancy and that only the over-expression of the aforementioned microRNA is sufficient for tumorigenesis. These data suggest that microRNAs will become a useful tool in understanding the complicated pathways that are involved in oncogenesis.
This review summarizes the most important studies on microRNA expression in normal and malignant hematopoiesis.
MicroRNAs in normal hematopoiesis
MicroRNA expression in normal hematopoiesis has been studied in murine and human hematopoietic tissue by several groups. It has been shown that these non-coding RNAs play critical roles in almost every stage of hematopoiesis (Table 1). However, there are many differences in microRNA expression between human and mouse hematopoietic cells (Fig. 2) (24). This fact should be taken into consideration when these results are evaluated.
Table 1. MicroRNA expression in normal hematopoiesis
The first study demonstrating the relation between microRNAs and hematopoiesis was published in 2004. In this study, Chen et al. cloned 150 microRNAs from murine bone marrow and found that three microRNAs (mir-181, mir-223 and mir-142) were preferentially expressed in hematopoietic cells. Mir-181was highly expressed in B-lymphoid cells of murine bone marrow. Its ectopic expression in hematopoietic stem cells resulted in an increase in the proportion of B-lineage cells in vivo and in vitro. In addition, mir-223 and mir-142 were highly expressed in myeloid lineage (mir-142 in B-lymphoid cells too). However, ectopic expression of these two microRNAs resulted in an increase in the proportion of T cells, but not of B or myeloid cells as it might be expected (25). On the other hand, Ramkinssoon et al. (24) who studied expression profiles of mir-142, mir-181, mir-223 and mir-155 found differences in expression patterns between human and mouse hematopoietic cells. Thus, it is very important to extrapolate results from studies in mice to humans with caution.
A very important microRNA in hematopoiesis is mir-150. It is preferentially expressed in mature, resting T and B cells but not in their progenitors. It is up-regulated during B-cell and T-cell maturation but down-regulated again during further differentiation of naïve T cells into Th1 and Th2 cells of murine hematopoietic tissue (26). Ectopic expression of mir-150 in hematopoietic stem cell progenitors reduces mature B-cell levels in the circulation, spleen and lymph nodes with little effect on T-cell or myeloid cell levels. Further experiments provide evidence that mir-150 blocks the transition from pro-B to pre-B cell during B-cell maturation (27) probably explaining the way that ectopic expression of this microRNA reduces mature B cells. In addition, TG and knockout (KO) mice models have shown that mir-150 induces the apoptosis of pro-B cells. Moreover, mir-150 was found to target c-MYB, a transcription factor that controls lymphocyte development. As previously mentioned, mir-150 is preferentially expressed in mature B lymphocytes but not in their progenitors. C-MYB, on the other hand, is highly expressed in lymphocyte progenitors and down-regulated in mature cells. Xiao et al. (28) demonstrated, in vivo, an interesting relationship between these expression patterns. Indeed, mir-150 directly down-regulates C-MYB, by binding to its mRNA and consequently controls its protein expression in a dose-dependent manner.
Another important microRNA in lymphoid differentiation is mir-155. It is a product of B-cell integration cluster (BIC) transcript that was first described as a frequent site of integration for the avian leucosis virus (29). High levels of mir-155 are present in activated B cells and T cells and in activated monocytes (reviewed by Fabbri et al.) (30). The function of mir-155 has been studied in both TG and KO mice. The first TG mice that specifically over-expressed mir-155 in B cells developed a pre-leukemic pre-B-cell proliferation evident in spleen and bone marrow, followed by a B-cell malignancy (23). Two independent groups developed mir-155 KO mice to study the role of mir-155 in B-cell and T-cell differentiation. Thai et al. (31) showed that mir-155 regulates germinal center reaction and T-helper cell differentiation by affecting cytokine production while Rodriguez et al. (32) showed that bic/mir-155 regulates the function of both lymphocytes and dendritic cells leading to defective immune response.
Finally, Georgantas et al. found 33 microRNAs that were expressed in human stem-progenitor cells (HSPC) from bone marrow and peripheral blood. The authors used bioinformatic techniques to combine these results with HSPC mRNA expression data and with miRNA–mRNA target prediction so as to create a miRNA–mRNA interaction database, the Transcriptome Interaction Database. Then, they formed a model for microRNA control of hematopoiesis. According to this model, mir-181 and mir-128 inhibit differentiation of all hematopoietic lineages while mir-146 inhibits differentiation of multipotent progenitor cell (MPP) into common lymphoid progenitors (CLP) (33).
MicroRNAs in granulocyte and monocyte differentiation
A microRNA that plays an important role in myeloid differentiation is mir-223. Fazi et al. studied myeloid differentiation in acute promyelocytic leukemia (APL) cells and demonstrated that mir-223 up-regulates mouse granulopoiesis in association with the transcription factors negative nuclear factor IA (NFIA) and CCAAT/enhancer binding protein a (CEBPA). Both transcription factors are able to bind to mir-223 promoter. NFIA down-regulates while CEBPA up-regulates mir-223 expression. Treatment of APL cells with retinoic acid (RA) resulted in CEBPA replacement of NFIA, up-regulation of mir-223 and enhanced granulocytic differentiation. Interestingly, mir-223 was found to inhibit translation of NFIA, resulting in a negative-feedback loop that favors granulocytic differentiation (34).
However, two recent studies do not confirm the previous results. In the first study, Fukao et al. found that myeloid expression of mir-223 might be specified by the conserved 5′ proximal cis-regulatory element where transcription factors PU.1 and C/EBP cooperatively act on. In addition, by studying the APL cell model that was used by Fazi et al., they found that RA induced differentiation of APL cells, repressed PU.1 and resulted in down regulation of mir-223 (35). In the second study, Johnnidis et al. found that mir-223 negatively regulates progenitor proliferation and granulocyte differentiation and activation in KO mice. In addition, they showed that mir-223 targets Mef2c, a transcription factor that promotes myeloid progenitor proliferation and that genetic ablation of Mef2c suppresses progenitor expansion and corrects the neutrophilic phenotype in mir-223 null mice (36). Further investigations are required to shed light on the real mechanisms that regulate mir-223 expression.
Fontana et al. investigated the role of mir-17-5p, mir-20a and mir-106a in monocytic differentiation and maturation. In unilineage monocytic cultured cells these microRNAs are down-regulated while acute myeloid leukemia 1 (AML1) transcription factor, which promotes M-CSFR transcription (M-CSF receptor), is up-regulated at the protein but not at the mRNA level. Transfection with mir-17-5p, mir-20a and mir-106a caused the opposite results in AML-1 protein expression followed by enhanced blast proliferation and inhibition of monocytic differentiation and maturation. Further experiments confirmed that these microRNAs target AML-1. In addition, AML1 binds the microRNA 17-5p-92 cluster and 106a-92 cluster promoters and inhibits the expression of mir-17-5p-20a-106a as a negative feedback, indicating that monocytopoiesis is controlled by a circuitry involving mir-17-5p, mir-20a, mir-106a, AML-1 and M-CSF (37).
Finally, according to the earlier mentioned model of Georgantas et al. (33), mir-155, mir-24a and mir-17 may inhibit the differentiation of MPP into common myeloid progenitor (CMP), while mir-16, mir-103 and mir-107 may inhibit the differentiation of CMP into granulocytic-macrophage progenitor (GMP).
MicroRNAs in erythroid and megakaryocytic differentiation
There is an increasing interest in the role of microRNAs in erythroid and megakaryocytic differentiation too. Mir-221 and mir-222 are down-regulated during erythroid differentiation and maturation. The first eight nucleotides of the 5′ sites of these microRNAs (seed region) are identical. This fact indicates that mir-221 and mir-222 can bind to the same target, which is KIT receptor, a key factor for the proliferation control of hematopoietic cells. Down-regulation of mir-221 and mir-222 probably unblocks KIT expression causing the expansion of erythroblasts (38). In addition, both erythroid and myeloid colony formation are reduced by mir-155 (33).
Additionally, Bruchova et al. studied microRNA expression in normal and polycythemia vera erythropoiesis. They evaluated 12 microRNAs by quantitative real-time PCR (qRT-PCR) in a large number of samples from peripheral blood mononuclear cells, cultured in a three phase liquid system. During normal erythropoiesis, mir-150, mir-155, mir-221 and mir-222 were progressively down-regulated (in agreement with the study mentioned earlier), mir-451 (which was found to be erythroid specific) and mir-16 (in reticulocytes) were up-regulated while mir-339 and mir-378 showed a biphasic expression pattern (39). Georgantas et al. had also showed that mir-155 blocks both myeloid and erythroid differentiation (33), while the study by O’ Connel et al. demonstrates the role of mir-155 in erythropoiesis too. In this study, mouse transplanted with MPP cells that over-express mir-155 developed a myeloproliferative disorder with decreased erythroid/ megakaryocytic lineage in bone marrow (40).
Moreover, Dore et al. showed that the transcription factor GATA-1 activates in vivo the transcription of a single pri-microRNA, encoding mir-144 and mir-451. Zebrafish embryos depleted of mir-451 form erythroid precursors, but their development into mature circulating red blood cells is strongly and specifically impaired. That was not true when mir-144 was silenced, suggesting that mir-451 has a specific role in erythropoiesis (41).
Recently, mir-223 was also found to control erythroid differentiation too. Its down-regulation in progenitor and precursor cells of erythroid lineage seems to be necessary for erythroid maturation while the expression levels of mir-223 are inversely related to the levels of LIM—only protein 2 (LMO2, RBNT2), a highly conserved protein that plays in critical role in hematopoietic differentiation. Further experiments demonstrated that mir-233 down-regulates LMO2 protein, by binding to its 3′-UTR, and inhibits differentiation and maturation of erythroid cells (42).
MicroRNAs seem to be important in megakaryocytic maturation. Garzon et al. performed microRNA expression profiling of in vitro differentiated megakaryocytes derived from CD34+ hematopoietic progenitors. Twenty microRNAs (including mir-10a, mir-126, mir-106, mir-10b, mir-17 and mir-20) were found to be down-regulated. It was also confirmed that mir-130a targets the transcription factor MAFB, which is involved in the activation of the GPIIB promoter (a key protein for platelet physiology) and that mir-10a directly targets HOXA1 in megakaryopoiesis. These two genes are over-expressed during megakaryopoiesis suggesting that down-regulation of microRNAs unblocks their expression (43).
Moreover, Labbaye et al. studied leukemic cells lines and showed that megakaryopoiesis is controlled by a pathway, in which PZLF transcription factor suppresses mir-146 transcription and thereby activates the SDF-1 receptor CXCR4 (44).
Finally, the transcription factor RUNX1 (also known as AML1 protein) is up-regulated during early hematopoiesis and up-regulates mir-27a. During differentiation of megakaryocytic lineage, mir-27a negatively regulates RUNX1 expression forming a feedback loop (45).
MicroRNAs expression in hematological malignancies
MicroRNAs expression in hematological malignancies has been extensively studied. There is an increasing number of reports demonstrating a key role of these non-coding RNAs in the pathogenesis and prognosis of hematological malignancies, the most important of which are summarized in Table 2.
Chronic lymphocytic leukemia (CLL) is the most common leukemia in the Western world, accounting for 30% of all leukemia cases in the US (46).
The del13(q14) is the most common cytogenetic abnormality in CLL and is usually associated with good prognosis (47). Hemizygous and/or homozygous loss at 13(q14) occur in more than half CLL cases. Deletions of 13(q14) loci are frequently found in other types of cancers, such as mantle cell lymphoma, multiple myeloma (MM), prostate cancers and pituitary tumors (reviewed by Nicoloso et al.) (48), suggesting that one or more tumor suppressor genes are located in this region. However, until recently, several studies had failed to identify the involved gene located in the deleted region.
Calin et al. showed that a cluster of two microRNA genes, mir-15a and mir-16-1 (with < 200 bp distance between them), is located exactly in 13(q14) region. These genes are highly expressed in normal CD5+ B-lymphocytes with mir-16-1 being expressed at higher levels than mir-15a. It was also found that both genes are down-regulated or deleted in the majority (68%) of the CLL samples and that down-regulation of mir-15a and mir-16-1 expression correlates with allelic loss at 13q14, suggesting that these two microRNAs may act as tumor suppressors and that their loss contributes to the pathogenesis of CLL (49).
However, in a more recent report, down-regulation of mir-15a and mir-16-1 was found only in a small subset of patients. Fulci et al. evaluated 56 patients with the diagnosis of CLL for microRNA profiling using microRNA cloning and qRT-PCR. Low expression of these microRNAs was found in 11.7% of patients by qRT-PCR and in 11.1% by cloning (50).
These discrepancies may be attributed to the methodological differences between the two studies, namely the northern blot procedure in the first and the microRNA cloning and qRT-PCR in the recent study, the different controls used (CD5+ cells from tonsils in the first and CD5+ cells from cord blood in the latter study), as well as in the different population of patients. More specifically 68% of informative samples displayed loss of heterozygosity (LOH) at the 13q locus in the study by Calin et al., while 26/67 (38.8%) analyzed patients revealed del(13q) by FISH, in the study by Fulci et al. Thus a significantly lower percentage of patients in the latter study had deletions at the 13q locus accounting perhaps for the lower frequency of mir-15a /mir-16-1 deletions.
Moreover, an alternative mechanism of mir-15a and mir-16-1 reduction is germ line mutations. A germ-line mutation in the mir-16-1-mir-15a primary precursor, in association with deletion of the normal allele was detected in two CLL patients. Interestingly, one patient had a family history of CLL and breast cancer (51).
In addition, Mir-15a and mir-16-1 have been associated with protein expression in CLL. Mostly non-dividing B-CLL cells over-express BCL-2. No mechanism for this aberrant expression has been identified, although the overexpression of bcl-2 may be attributed in the juxtaposition of BCL-2 to immunoglobulin loci, this mechanism refers to < 5% of the CLL cases. Mir-15a and mir-16-1 expression is inversely correlated with BCL-2 protein levels. Although in normal CD5+ lymphoid cells the levels of both microRNAs were high and the BCL-2 protein was expressed at low levels, in the majority of leukemic cells both mir-15 and mir-16 were expressed at low levels and BCL-2 was over-expressed. Further research showed that mir-15a as well as mir-16-1 down-regulate BCL-2 at the post-transcriptional stage of expression. The MEG-01 cell line (a leukemia-derived cell line with deletion of one allele and alteration of the other mir-15a and mir-16-1 loci and no expression of mir-15a and mir-16-1) was selected to be transfected by a vector that contains the full mir-15a/16-1 cluster. BCL-2 levels were highly reduced in comparison to WT MEG-01 cells (not transfected) or cells transfected with an empty vector. It was also shown that both microRNAs can affect BCL-2 protein expression separately and that mir-15 and mir-16 regulate BCL-2 expression at the post-transcriptional level (52). These data first and foremost indicate a possible mechanism for BCL-2 aberrant expression and secondly explain how mir-15a and mir-16-1 exert their tumor-suppressing function in CLL.
However, in contrast to the previous results, Fulci et al. reported that down-regulation of mir-15 and mir-16 was not followed by significant increase in BCL-2 expression levels. These data do not support the hypothesis that in CLL, the high levels of BCL-2 expression are secondary to a down-regulation of mir-15 and mir-16 (50).
Two other microRNAs, mir-29b and mir-181b, act as tumor-suppressors in aggressive CLL by targeting the Τ-cell leukemia/lymphoma 1 (TCL-1) oncogene. The levels of both microRNAs inversely correlated with levels of TCL-1 in CLL patients (53). TCL-1 is a coactivator of AKT, an oncogene that inhibits apoptotic processes and has a critical role in the regulation of many pathways involved in cell survival, proliferation and death. Its high expression is associated with expression of unmutated IgVH and ZAP-70, factors that indicate aggressive CLL (48).
Several groups have investigated microRNA expression profiling of CLL. Calin et al., compared microRNA expression in CLL cells vs. normal CD5+ B cells with a microarray chip and confirmed their results with Northern blot and qRT-PCR. Two groups of microRNAs, the first composed of 55 genes and the second of 29 genes, were found to have statistically significant differences in expression levels between normal and malignant cells. Some of them are located in fragile sites, such as mir-183 at FRA7H, mir-190 at FRA12A, miR-24-1 at FRA9D and miR-213 (54).
To investigate whether the expression of microRNAs genes is associated with factors that predict the clinical course of CLL (ZAP-70 expression and mutated/unmutated status of IgVH), Calin et al., performed a genome wide microRNA microchip in a large number of CLL patients. By studying 94 samples of CLL, the group demonstrated that there is a signature of 13 microRNA genes (mir-15a, mir-195, mir-221, mir-23b, mir-155, mir-223, mir-29a-2, mir-24-1, mir-29b-2, mir-146, mir-16-1, mir-16-2, mir-29c) that has prognostic value, because the expression levels of the members of the signature correlate with CLL prognostic factors such as ZAP-70 expression and IgVH mutation status (51). Finally, miR-150, miR-223, miR-29b and miR-29c were over-expressed in patients with IgVH mutated cells supporting previous results (50).
MicroRNA expression in Hodgkin lymphoma
There are a few studies referring to microRNAs alterations in Hodgkin Lymphoma (HL). Kluiver et al., analyzed HL cell lines and tissue samples and detected high levels of BIC and mir-155 in HL (55). Three years later, Navarro et al., analyzed the expression of 157 microRNAs in 49 lymph nodes from classical HL and ten reactive lymph nodes (RLN) by qRT-PCR and revealed three well-defined groups: nodular sclerosis cHL, mixed cellularity cHL and RLNs. They also found a distinctive signature of 25 microRNAs which differentiated cHL from RLNs, and 36 microRNAs that were differentially expressed between nodular sclerosis and mixed cellularity subtypes. Finally, mir-96, mir-128a and mir-128b were found to be selectively down-regulated in cHL with Ebstein Barr virus (EBV) infection, while mir-138 was related with Ann Arbor stage I-II cHL (56). These findings indicate different microRNAs expression patterns in different histological subtypes as well as in early stage disease.
Recently, Nie et al. showed that PRDM1/Blimp-1, an important regulator in terminal B-cell differentiation, is targeted by mir-9 and let-7a in cultured Hodgkin/Reed-Sternberg (HRS) cells. MicroRNA expression profile demonstrated that both microRNAs are highly expressed in Hodgkin/Reed-Sternberg cells, suggesting that mir-9 and let-7a act as oncogenes in HL cells (57).
MicroRNA expression in non-HLs
Diffuse large B-cell lymphoma (DLBCL) is the most common type of lymphoma representing nearly 40% of lymphoid malignancies. There are at least two types of DLBCL that have different prognostic value, the germinal center type (GC) with a better outcome and activated B-cell type (ABC) with worse prognosis (58).
MicroRNA expression in B-cell lymphomas has been studied by several groups. Eis et al. analyzed lymphoma samples and cell lines and found increased levels of mir-155, in DLBCL. Both BIC RNA and mir-155 were quantified in 23 DLBCL samples. BIC RNA levels were elevated from two to tenfold in DLBCL in comparison with normal CD19+ B cells, which were used as normal control. Mir-155 levels were increased to even greater extents. Significantly higher levels of mir-155 were detected in the ABC type compared to the GC phenotype. It was also found that the levels of mir-15a were reduced in DLBCL, which shows that the reduction of this microRNA is not specific to CLL. On the other hand, the levels of mir-16 showed no specific pattern of increase or decrease relative to the control B cells (59). In addition, Kluiver et al. (55) also demonstrated that BIC and mir-155 are highly expressed in DLBCL and primary mediastinal lymphomas, but not in other non-HLs.
Roehle et al., who analyzed the expression of microRNAs in DLBCL, follicular lymphoma (FL) and non-neoplastic lymph nodes, found that mir-155 was up-regulated in DLBCL as well. Moreover, this group demonstrated that mir-330, mir-17-5p, mir-106a and mir-210 discriminate DLBCL, FL and RLNs (60).
In agreement with the previous results, Lawrie et al. showed that ABC and GC subtypes of DLBCL have distinct microRNA expression profile. Mir-155, mir-21 and mir-221 are highly expressed in ABC in comparison with GC DLBCL. These microRNAs were over-expressed in clinical cases of DLBCL and FL as well as in DLBCL that had undergone transformation from FL. Univariate and multivariate analysis suggested that mir-21 expression is an independent prognostic indicator in de novo DLBCL and that high expression of this microRNA was associated with a better outcome (61).
The aforementioned microRNAs (mir-155, mir-21, and mir-221) were compared in serum samples from DLBCL patients to healthy controls. The levels of these microRNAs were found to be up-regulated in patients’ serum. In addition, high levels of mir-21 in DLBCL patient serum was associated with improved relapse-free survival (62).
A possible target of mir-155 is the transcription factor PU.1 that is a member of Ets domain-transcription factor family and is required for late differentiation of B cells (63). In addition, the mRNA of the transcription factor C/EBPβ has a potential target site for mir-155 in its 3′-UTR (59). In addition, a correlation between mir-155 and NF-kB expression has been found in DLBCL cell lines and patient samples (64). Recently, mir-155 was found to directly target Src homology-2 domain-containing inositol 5-phosphatase 1 (SHIP1), which acts in the PTEN/AKT pathway. Mir-155 binds to 3′-UTR of SHIP 1 and down-regulates SHIP 1 in vivo and in vitro. Down-regulation of SHIP 1by mir-155 resulted in both myeloproliferative and lymphoproliferative disorder indicating a role of this pathway in myeloid and lymphoid cells (65, 66).
Primary effusion lymphoma (PEL) is a subtype of DLBCL associated with Kaposi sarcoma-associated herpes virus (KSHV) while some of PELs are coinfected by EBV. PEL cells contain two types of microRNAs encoded by either cellular genes or by KSHV or EBV. A 68-microRNA-specific signature was identified containing KSHV and EBV microRNAs. Mir-155 was found to be down-regulated in contrast to ABC DLBCL cases (67).
A recent study associates microRNA expression in DLBCL with immunophenotype, survival and transformation from FL. Lawrie et al. studied 80 samples of DLBCL and 18 samples from FL with microarrays and found distinct expression patterns between these two lymphomas. Furthermore, microRNAs expression was associated with GC-like and non-GC-like immunophenotypes (supporting previous results), international prognostic index status and event-free survival in cyclophosphamide, doxorubicin, oncovin, prednisone (CHOP) and rituximab-CHOP treated patients (68). These data indicate that microRNAs may be used in the future as prognostic and diagnostic tools in DLBCL.
High expression of precursor mir-155/BIC RNA was described in pediatric Burkitt’s lymphoma (69), but not in adult primary cases of the disease (70). Expression of BIC and mir-155 in three latency type III EBV-positive Burkitt’s lymphoma cell lines and in all primary post-transplantation lymphoproliferative disorder cases suggests a possible role for EBV latency type III-specific proteins, in the induction of BIC expression (70).
Mantle cell lymphoma (MCL) represents 5–10% of all non-HLs and has the worst prognosis among B-cell lymphomas. The t(11;14)(q13;q32) is characteristic of MCL and displays the cyclin D1 (CCND1) gene on chromosome 11 downstream to the enhancer region of the IgH gene on chromosome 14 resulting in CCND1 over-expression. Point mutations and genomic deletions in CCND1 create a truncated mRNA and result in increased proliferation and shorter survival (71). Cell lines, with increased ratio of truncated CCND1 have increased total mRNA levels, increased CCND1 protein expression. Mir-16 is able to down-regulate CCND1 protein expression (72). However, the truncation of CCND1 mRNA lacks mir-16-1 binding sites within the CCND1 mRNA 3′-UTR and alters microRNA regulation in MCL (72) suggesting a possible role of this mir-16-1 in CCND1 repression.
The genetic locus 13q31, which is commonly found amplified in many B-cell lymphomas, such as DLBCL, MCL and solid tumors (reviewed by Lawrie) (58), encodes the microRNA cluster mir-17-92, which includes mir-17-5p, mir-18a, mir-19a, mir-20a, mir-19-b-1 and mir-92 (73). He et al. found that expression of the mir-17-92 cluster components in mice also over-expressing MYC, enhanced lymphomagenesis while each microRNA of the cluster did not (73). In addition, O’ Donnell et al. showed that the c-MYC oncogene activates the transcription of the mir-17-92 cluster (74). C-MYC also induces the expression of the transcription factor E2F1 gene, which in turn can induce c-MYC by means of a positive feedback. Mir-17-5p and mir-20a (members of the mir-17-92 cluster) directly down-regulate the transcription factor E2F1 (74). By targeting E2F1, members of the mir-17-92 cluster can reduce the reciprocal activation MYC/E2F1 and control c-MYC-mediated cellular proliferation.
MicroRNA expression in T-cell lymphomas
The role of microRNAs in T-cell lymphomas remains unclear. However, the available data suggest a role of microRNAs in the pathogenesis of T-cell lymphomas.
Lum et al. found a high density of integrations upstream of the mir-106a microRNA cistron. In tumors containing integration, the primary transcript encoding the mir-106a cistron was over-expressed 5–20-fold compared with control tumors. The mature mir-106a and mir-363 and mir-17-92 were over-expressed too (75). Landais et al. confirmed pri-mir-106-363 over-expression in 46% of human T-cell leukemias tested (76).
MicroRNA expression in acute leukemia
There is an increasing number of publications investigating the role of microRNAs in acute leukemia. Debernardi et al. analyzed the expression of microRNAs in 30 samples from AML patients with normal karyotype. The levels of mir-181a (mentioned earlier as potential inhibitor of differentiation of all human hematopoietic lineages) (33) were elevated in M1 or M2 AML, compared with M4 or M5 AML. The elevated expression of mir-181a observed in AML implies its involvement in the leukemic process. Moreover, mir-10a, mir-10b and mir-196a-1 were found to correlate with HOX gene expression (77).
Garzon et al. studied the microRNA expression in APL. Cell lines treated with the differentiating agent all-trans-RA (ATRA) showed up-regulation of some microRNAs, including few known tumor suppressor microRNAs, such as mir-15a and mir-16-1, members of the let-7 family and mir-223, mir-342 and mir-107. Mir-181 was found to be down-regulated after treatment with ATRA (78).
Isken et al. analyzed 50 patient samples with AML and 12 control samples for the expression of 154 human microRNAs. The expression level of mir-26a, mir-26b and mir-29b was intermediate between CD34+ cells and normal bone marrow (NBM) samples. There was a significant difference in the expression pattern of mir-23b, mir-34a and the mir-221/mir-222 cluster between AML and controls. Except for mir-23b, all these were up-regulated in AML samples compared to CD34+ and NBM samples. High levels of expression of mir-221/mir-222 cluster promote growth of several cancer cells lines, by inhibiting the expression of CDKN1B, an inhibitor of cell cycle progression and known tumor suppressor. In addition, they secondary affect KIT expression. Finally, mir-181a expression was reduced in samples with French-American-British (FAB) classification M4 and M5 compared to FAB M1 and M2, supporting the results of the aforementioned study by Debernardi et al. (77).
More recently, Marcucci et al. studied samples of leukemia cells from adults under the age of 60 yr who had cytogenetically normal AML and internal tandem duplication in the fms-related tyrosine kinase 3 gene (FLT3-ITD), a wild type nucleophosmin (NPM1) or both using microRNA expression profiling. A signature of 12 microRNAs was associated with clinical outcome. Among them, members of mir-181 family are down-regulated in high risk AML, with their expression levels being inversely correlated with levels of proteins involved in pathways controlled by toll-like receptors and interleukin-1β (79).
Moreover, there is evidence of a unique microRNA signature that distinguishes NPMc + AML (AML with NPM1 mutations and cytoplasmic NPM) from unmutated cases. This signature contains mir-10a, mir-10b, let-7 family and mir-29 family. Some of the microRNAs that are down-regulated, are associated with HOX genes indicating that HOX up-regulation in NPMc + AML is attributed, at least in part, to the loss of microRNA regulation (80).
In 240 AML patient samples with predominantly intermediate and poor cytogenetics, Garzon et al. found miRNA signatures, associated with balanced 11q23 translocation, isolated trisomy 8 and FLT3-ITD mutations. In addition, mir-199a and mir-191 were found to have a prognostic value, because they were associated with reduced overall and disease-free survival (81).
A screening of deregulated microRNAs in acute lymphocytic leukemia (ALL) has been published, providing a list of microRNAs involved in leukemogenesis. The five highly expressed microRNAs in ALL were mir-128b, mir-204, mir-218, mir-331 and mir-181b-1. The most commonly represented microRNA in ALL is mir-128b with a 436, fivefold difference compared to normal CD19+ B cells. The second most highly expressed microRNA in ALL is mir-204. On the contrary, the four microRNAs with the lowest expression levels are mir-135b, mir-132, mir-199s, mir-139 and mir-150. The mir-17-92 cluster was found to be up-regulated in ALL (82). Taking into consideration that the mir-17-92 antagonizes the expression of the pro-apoptotic protein BIM and favors the survival of B-cell progenitors (83), the involvement of mir-17-92 in ALL is not surprising.
Finally, Mi et al. performed a large-scale genome wide microRNA expression profile assay and identified 27 microRNAs that are differentially expressed between AML and ALL. Among them, mir-128a and mir-128b are significantly over-expressed, whereas let-7b and mir-223 are significantly down-regulated in ALL compared to AML. In addition, over-expression of mir-128 in ALL was at least partially associated with promoter hypomethylation and not with an amplification of its genomic locus (84).
MicroRNA expression in chronic myeloid leukemia (CML)
There are few reports correlating microRNAs with BCR-ABL expression. ABL1, which is activated in CML, as a BCR-ABL1 fusion protein, is a possible target of mir-203. Mir-203 is silenced in some hematopoietic malignancies including CML. However, re-expression of this microRNA reduces ABL1 and BCR-ABL1 fusion protein levels and inhibits tumor cell proliferation (85).
Venturini et al. screened for BCR-ABL and c-MYC-dependent microRNA expression using microarrays and quantitative real-time reverse transcriptase PCR. Treatment of K562 cells with imatinib (to inhibit BCR-ABL kinase activity), specific anti-BCR-ABL RNA interference (to knock-down BCR-ABL gene expression) or anti-c-MYC RNA interference resulted in the down-regulation of mir-17-92 cluster. The results were confirmed by qRT-PCR. Finally, it was demonstrated by the same group that mir-17-92 cluster is over-expressed in primary CML CD34+ cells in chronic phase but not in blast crisis, compared with normal CD34+ cells (86). These results suggest that microRNA expression chronic phase of CML is under the control of a BCR-ABL–c-MYC pathway.
MicroRNA expression in multiple myeloma
There are few reports associating microRNAs with the pathogenesis of MM. Masri et al. used microRNA chips to investigate microRNA expression profile in human myeloma cell lines (HMCLs) and MM patient samples. The primary MM cells and HMCLs displayed a distinctive microRNA expression profile compared with normal plasma cells. This profile includes mir-125b, mir-133a, mir-1 and mir-124a. Interestingly, mir-15 and mir-16 are expressed at low levels in some MM patients and HMCLs, but not in normal plasma cells (87).
These results are supported by the recent study by Roccaro et al., who performed microRNA expression profiling of bone marrow derived CD138+ MM cells versus their normal counterparts and validated data with qRT-PCR. A MM-specific microRNA signature characterized by down-regulation of mir-15a, mir-16 and over-expression of mir-222, mir-221, mir-382, mir-181a and mir-181b was identified. It was also reported that mir-15a and mir-16 regulate proliferation and growth of MM cells in vitro and in vivo by inhibiting AKT serine/threonine protein kinase (AKT3), ribosomal-protein S6, MAP-kinases and NF-kappaB activator MAP3KIP3 (88).
Bakkus et al. detected higher expression of let-7a, mir-16 and mir-17-5p and mir-19b (the last two are members of the mir-17-92 cluster) both in MM patients and cell lines. In addition, mir-16 was much higher in the MM plasma cells compared with the monoclonal gammopathy of undetermined significance (MGUS) plasma cells (89).
According to Pichiorri et al., mir-21, mir-106b-25 cluster and mir-181 were over-expressed in MM and MGUS samples with respect to healthy plasma cells. Selective up-regulation of mir-32 and mir-17-92 cluster was identified in MM cells, but not in MGUS or healthy plasma cells. Furthermore, mir-19a and b (members of mir-17-92 cluster) were shown to down-regulate the expression of SOCS-1, a gene frequently silenced in MM that plays a critical role as inhibitor of IL-6 growth signaling. P300-CBP-associated factor, which plays an important role in chromatin remodeling and is involved in p53 regulation, was identified as a target of mir-106b-25 cluster, mir-181a and b and mir-32 (90).
MicroRNAs: from normal to malignant hematopoiesis
There is an increasing number of studies trying to correlate microRNA expression in normal hematopoiesis to that in hematological malignancies. At present, there are very few data indicating a direct role of microRNAs in hematological malignancies. This, at first, is because of the great number of complicated, overlapping pathways that microRNAs participate in, added to the fact that microRNAs have many different targets, which are, at present, mostly computationally predicted. Moreover, most of the studies demonstrate multiple aberrantly expressed microRNAs in hematological malignancies, which complicate the attempt to associate oncogenesis with a unique microRNA.
However, there are some data that allow us, to extrapolate data regarding the role of microRNAs from normal to malignant hematopoiesis. For example, mir-155 regulates GC reaction and is associated with B-cell malignancy in TG mice (23, 31, 32). Indeed, it is up-regulated in DLBCL (ABC type), CLL, HL and primary mediastinal B-cell lymphoma (PMBL) indicate that this microRNA plays a key role in lymphomagenesis (55, 59, 91). Mir-155 is a regulator of myelopoiesis too, because it is regarded to block both myeloid and erythroid differentiation (33). The aberrant expression of mir-155 in TG mice, on the other hand, results in a myeloproliferative disorder and reduction of erythroid/megakaryocytic lineage (40). According to these data, mir-155 acts in a variety of ways while its aberrant expression is associated with both myeloid and lymphoid malignancies.
Mir-15a/16-1, regarded as tumor suppressor agents, are down-regulated or deleted in the majority of CLL cases (54) while mir-16-1 is associated with MCL too (72), suggesting a role of these microRNAs in lymphomagenesis.
In addition, mir-181 is highly expressed in murine B-lymphoid cells, its ectopic expression in progenitor cells causes an increase in the proportion of B cells (25), and it is predicted to inhibit differentiation of all hematopoietic lineages (33). Indeed members of mir-181 family have already been associated with AML because mir-181 a is up-regulated in M1, M2 with normal karyotype (77) and down-regulated in high risk AML (79). Another member of mir-181 family, mir-181b is down-regulated in poor prognosis CLL (53). Nevertheless, the way that mir-181 functions need to be defined.
Another microRNA with a very interesting expression pattern is mir-223. Despite the fact that it is involved in myeloid differentiation, it is found up-regulated in adult T-cell leukemia patients but down-regulated in human T-cell leukemia virus type-I (HTLV-I) cell lines (92). In addition, in AML blasts that carry t(8;21), the fusion protein AML1/ETO directly targets mir-223 and reduces its levels (93).
Finally, mir-17-92 which induces lymphomagenesis in TG mice that over-express MYC (73) is over-expressed in B-cell malignancies such as DLBCL and MCL indicating a role of this mir-17-92 in B-cell malignancies. Apart from E2F1, another target of mir-17-92 is the pro-apoptotic protein BIM. Aberrant expression of mir-17-92 in ALL results in down-regulation of BIM and enhances the survival of B-cell progenitors (83). Thus, aberrant expression of mir-17-92 cluster is associated with several hematological malignancies. However, the exact pathways in which mir-17-92 participates need to be clarified.
MicroRNAs as therapeutic agents
As mentioned earlier, microRNAs can act both as oncogenes and tumor suppressor genes and play key roles to almost every cell function. As a result, the use of microRNAs as therapeutic agents or therapeutic targets seems to be a very appealing scenario. The potential use of anti-microRNA molecules, termed as antagomirs, locked nucleic acid (LNA)-anti-miR oligonucleotides or anti-microRNA oligonucleotides, has been studied by several groups, in vitro and in vivo while the first clinical trial (described below) applying anti-microRNAs agents as drugs has already been launched.
The use of 2′-O-methyl antisense nucleotides (antisense RNA molecules, complementary to the microRNA they target) to inhibit siRNA and microRNA function was first applied to Drosophila, C. elegans and mice (94–96). In a fundamental study, for microRNA silencing, antagomirs against mir-16, mir-122, mir-192 and mir-194 were injected into mice and caused reduction of the levels of these microRNAs in several tissues. Mir-122 is preferentially expressed in the liver, and its levels were reduced or even depleted when mice were treated with antagomir-122, a result that was sustained for as long as 23 d. These results indicate that mir-122 plays a key role in cholesterol metabolism is successfully targeted by antisense oligonucleotides and represents an attractive therapeutic target. Apart from mir-122, several other microRNAs have been already silenced by antisense oligonucleotides, such as mir-126, which is expressed in human endothelial cells and was effectively silenced by antagomirs-126 in mice (99).
Aberrant expression of mir-17-92 has been associated with CML, as was previously analyzed. Anti-mir-18a, anti-mir-19-b and anti-mir-20a successfully antagonize mir-18a, mir-19b and mir-20a (members of mir-17-92 cluster), respectively. In addition, the expression levels of the transcription factor E2F1 (recognized as target of mir-20) are increased in a dose-dependent manner following treatment with anti-mir-20 (100). These results suggest that members of mir-17-92 may represent a potential therapeutic target in CML.
A very appealing therapeutic option is the combination of microRNAs with chemotherapy and the effect of chemotherapy in microRNA expression profiling. A study in head and neck squamous cell carcinoma showed that high mobility group AT-hook 2 (HMGA2, a gene that encodes a protein that belongs to the non-histone chromosomal high mobility group protein family), which is associated with sensitivity to doxorubicin, is down-regulated by hypoxia. This situation is accompanied by up-regulation of mir-98. Moreover, it was demonstrated that transfection with pre-mir-98 during normoxia down-regulates HMGA2 and probably contributes to the resistance to doxorubicin. As a result, mir-98 silencing could enhance the sensitivity to doxorubicin, which is a very interesting and promising scenario (101). The role of microRNAs in response to chemotherapy has been studied in human cholangiocarcinoma cells too. Meng et al. (102) showed that inhibition of mir-21 and mir-200b, which are over-expressed in malignant cholangiocytes, increased sensitivity to gemcitabine. Mir-127, which is not expressed in cancer cells, was found up-regulated in T 24 cells (a human bladder carcinoma cell line) that were treated with 5-aza-2′-deoxycytidine and 4-phenylbutiric acid. Equally interesting was the fact that BCL-6, a potential target of mir-127, was down-regulated after treatment, indicating that the reactivation of mir-127 caused the repression in BCL-6 proto-oncogene (103).
The observation that microRNA hypermethylation, which results in microRNA down-regulation, is associated with tumorigenesis raised hope that methylated microRNAs could represent a new target for hypomethylating therapy. Several studies provide evidence that epigenetic regulation of microRNAs by methylation contributes to tumorigenesis or represent adverse prognostic factors. In ALL, for example, DNA methylation results in down-regulation of several microRNAs while treatment with 5-aza-2′-deoxycytidine results in up-regulation of these microRNAs. In addition, methylation of microRNAs seems to represent an independent prognostic factor associated with poor disease-free and overall survival (OS) (104). More specifically, methylation of DNA in ALL patients resulted in down-regulation of mir-124 and up-regulation of its target, cyclin-dependent kinase 6 (CDK6) and proliferation of ALL cells, while patients carrying methylated mir-124 had higher relapse and mortality rate (105). In conclusion, methylated microRNAs seem to represent a very promising therapeutic target of hypomethylating treatment and a useful tool in discriminating patients with poor prognosis.
Finally, the most important step in establishing the role of microRNAs as therapeutic targets has been already made. The first clinical trial worldwide, studying the potential use of LNA-antimir-122 in hepatitis C treatment has been already launched. Mir-122, a microRNA expressed mostly in liver cells (as mentioned earlier) participates in hepatitis C virus (HCV) replication. Indeed, mir-122 enhances HCV translation by binding to two targets in 5′-UTR of viral genome (106–108). The aforementioned phase I, double-blind, randomized clinical trial includes 48 volunteers, divided into six groups each comprised of eight persons. In each group, six persons receive LNA-antimir-122 and 2 receive placebo (Santaris pharma, 2008). This is the first attempt to apply anti-microRNA agents in humans. Phase II studies and studies of more microRNAs probably represent the most important scientific event in the future.
Apart from their potential use as therapeutic agents, microRNAs may represent a useful tool in predicting the clinical outcome of a disease or even identifying subgroups of patients at high risk, who need more intensive therapy or closer follow up. In CLL, for example, the expression of microRNAs is associated with known prognostic factors, such as ZAP-70 expression, IgVH mutational status and time between diagnosis and treatment (51). In addition, mir-29c (previously reported to act as a tumor suppressor factor) and mir-223 are down-regulated during transition from Binet stage A to stage C, as well as in poor prognosis subgroups of CLL. Thus, it is possible that these microRNAs can predict treatment-free survival (TFS) and OS. Stamatopoulos et al. developed a score system to classify patients according to mir-29c and mir-223 expression, ZAP-70 and LPL. This score is inversely correlated with TFS and OS indicating a significant role of microRNAs in discriminating between good and poor prognosis subgroups of patients (109).
Despite these appealing scenarios, there is a long way from our current knowledge to the establishment of microRNAs as therapeutic agents and many studies should be performed to shed light on the complicated pathways in which microRNAs are involved.
There is an increasing number of studies bringing evidence that microRNAs play critical roles in almost every cell function, such as differentiation, proliferation and apoptosis. MicroRNAs, recently identified as post-transcriptional regulators of gene expression, can act both as oncogenes and tumor suppressor genes demonstrating an important role of these small non-coding RNAs in the pathogenesis and prognosis of hematological malignancies.
Their potential use as therapeutic agents or target is of great importance. Using microRNAs with tumor suppressor function as drugs seems to be very tempting. On the other hand, microRNAs that act as oncogenes could represent potential therapeutic targets of antisense microRNAs (antagomirs).
The authors acknowledge grant support from Hellenic Co-operative Oncology Group (HeCOG) MDS-9-2007. We also thank P. Tsiotra and F. Kontsioti (2nd Department of Internal Medicine & Research Institute, Lab. of Molecular Biology) for their valuable assistant.