MicroRNAs (miRNAs) are small non-coding RNAs involved in the regulation of critical cell processes such as apoptosis, cell proliferation and differentiation. A small set of miRNAs is differentially expressed in hematopoietic cells and seemingly has an important role in granulopoiesis and lineage differentiation. In this study, we analysed, using a quantitative real-time PCR approach, the expression of 12 granulocytic differentiation signature miRNAs in a cohort of acute promyelocytic leukemia (APL) patients. We found nine miRNAs overexpressed and three miRNAs (miR-107, -342 and let-7c) downregulated in APL blasts as compared with normal promyelocytes differentiated in vitro from CD34+ progenitors. Patients successfully treated with all-trans-retinoic acid (ATRA) and chemotherapy showed downregulation of miR-181b and upregulation of miR-15b, -16, -107, -223, -342 and let-7c. We further investigated whether the APL-associated oncogene, promyelocytic leukemia gene (PML)/retinoic acid receptor α (RARα), might be involved in the transcriptional repression of miR-107, -342 and let-7c. We found that PML/RARα binds the regulatory sequences of the intragenic miR-342 and let-7c. In addition, we observed, in response to ATRA, the release of PML/RARα paralleled by their transcriptional activation, together with their host genes, EVL and C21orf34α. In conclusion, we show that a small subset of miRNAs is differentially expressed in APL and modulated by ATRA-based treatment.
Acute promyelocytic leukemia (APL) is a specific acute myelogenous leukemia subtype characterized by maturation arrest at the promyelocytic stage of development and caused by a novel fusion protein resulting from the reciprocal translocation involving the retinoic acid receptor α (RARα) on chromosome 17 with the promyelocytic leukemia gene (PML) on chromosome 15. The resulting APL-associated PML/RAR oncogene (PML/RARα) contributes to the maturation arrest of hematopoietic precursors and leukemogenesis by disrupting PML function and silencing the retinoic acid signaling pathway (Melnick and Licht, 1999; Tallman et al., 2002).
MicroRNAs (miRNAs) are small non-protein-coding RNAs that regulate gene expression at the post-transcriptional level and influence many aspects of cell biology (Bartel, 2004; Meltzer, 2005; Pasquinelli et al., 2005; Mallardo et al., 2008). A number of miRNAs have been shown to be differentially expressed in hematopoietic tissues and to have an important role both in lineage differentiation and in human hematological malignancies (Chen et al., 2004; Fazi et al., 2005; Fontana et al., 2007; Fabbri et al., 2008). Among these, miRNA-223 is a key member of a regulatory circuitry involving C/EBPα and NFI-A that controls granulocytic differentiation in all-trans-retinoic acid (ATRA)-treated APL cells (Fazi et al., 2005). More recently, it has been reported that treatment in vitro of the APL cell line NB4 with the differentiating agent ATRA leads to downregulation of miR-181 and upregulation of nine miRNAs, revealing a potential tumor suppressor function of these miRNAs in APL (Garzon et al., 2007). In this study, we analysed the expression of 12 selected granulocytic signature miRNAs and the impact of ATRA-based therapy in a cohort of APL patients. We found that miR-15a, -15b, -16, 142-3p, -142-5p, -181b, -223, let-7a and let-7d are increased, whereas miR-107, -342 and let-7c are decreased in APL blasts obtained at the time of diagnosis as compared with human normal promyelocytes differentiated in vitro for 7 days from umbilical cord blood (UCB)-derived CD34+ hematopoietic progenitors (Figure 1).
To gain further insights into the specificity of the miRNAs signature of APL blasts as compared with different stages of granulocytic differentiation, we compared the expression of our set of miRNAs in UCB-derived CD34+ unilineage granulocytic cultures at different stages of normal granulocytopoiesis. This serum-free ex vivo culture systems of granulocytic lineage differentiation allows the analysis of ‘pure’ unilineage precursors at discrete sequential stages of development, starting from initial UCB-derived CD34+ progenitor cells through terminal granulocyte maturation in a 2-week period, as indicated by cell morphology at days 0, 7 and 15 (Figures 2a and b), and by increased expression of granulocytic-specific membrane antigens (data not showed). As shown in Figure 2c, we found a strong upregulation (fold>5) of miR-15a, -15b, -107, -181b, -223, let-7a, let-7d and to a lesser extent (2<fold<5) of miR-16, -142-3p, -142-5p and miR-342 in promyelocytes as compared with that in UCB-derived CD34+ cells. We also found that high levels of miR-15b, -223, -142-5p and -342 (fold>2) and reduced expression of miR-181b and let-7c distinguished fully differentiated granulocytes from normal promyelocytes. Altogether these findings define a specific miRNA signature of APL blasts at diagnosis that differs from normal promyelocytes.
Next, we evaluated the impact of a standard ATRA-containing regimen on selected miRNAs in leukocyte APL patients obtained at time of remission (after third consolidation). We found that most successfully treated patients displayed a striking decrease of miR-181b expression (91.7% of patients), whereas in the majority of patients miR-15b (83.3%), -16 (70.8%), -107 (87.5%), -223 (91.7%), -342 (83.3%) and let-7c (62.5%) were significantly increased after therapy (Figures 3a and b). Interestingly, whereas upregulation of miR-15b and -342 in treated patients mimics the behavior of these miRNAs in the differentiation of normal promyelocytes toward mature granulocytes, upregulation of miR-16 and -107 (that are not upregulated in granulocytes as compared with those in promyelocytes) and let-7c (that decreases all along granulocytic differentiation) is likely related to a specific effect of ATRA. This conclusion is further supported by the observation that, in contrast to the parental NB4 cells, the ATRA-resistant NB4 sub-clone MR2 cells (Rosenauer et al., 1996), showed no significant modulation in the selected miRNA expression in response to ATRA, with the exception of miR-16, that was downregulated, and miR-223 that was moderately upregulated (Figure 3c). We also analysed the expression levels of the same set of differentiation-related miRNAs in sequential samples obtained from three patients (#46, #48, #50) who underwent disease relapse during follow-up. All patients showed a strong upregulation of miR-15b, -107, -223, -342, let-7c and, to a lesser extent, of miR-15a and miR-16 in samples collected during remission. In patients #46 and #50, who underwent hematological relapse, all these miRNA were sharply downregulated at time of disease recurrence, returning to levels comparable to or lower than those observed at diagnosis (Figure 3d). In patient #48, who had a molecular relapse at the time of its second testing, we observed a sharp downregulation of miR-15b and -223, but not of miR-107, -342 and let-7c (Figure 3d). Whether the difference observed between patients with overt relapse and early molecular relapse are due to true biological differences, or are simply related to the timing of analysis or reflect the dilution of the re-emerging leukemic clone, will require further investigations.
Altogether, these results expand and confirm in vivo previous observations obtained in vitro (Garzon et al., 2007; Saumet et al., 2009) and suggest a role of these miRNAs in therapeutic response in APL patients.
We then focused on the regulation of miR-107, -342 and let-7c. These miRNAs were all decreased in APL blasts as compared with those in normal promyelocytes (Figure 1), and strikingly increased in APL patients after ATRA-based therapy (Figures 3a and b).
miR-107, -342 and let-7c are intragenic; they are located in the introns of protein-encoding genes and are supposedly transcribed coincidentally with their host genes (Rodriguez et al., 2004; Kim and Kim, 2007; Grady et al., 2008).
First, we confirmed that miR-107, -342 and let-7c are co-regulated with their respective host genes PANK1, EVL and C21orf34α in promyelocytic NB4 cells exposed for 120 h to ATRA (Figure 4a, left panel). Moreover, in silico analysis of the putative PANK1/miR-107, EVL/miR-342 and C21orf34α/let-7c promoters revealed the presence of multiple RARE sites (Figure 4c). In the case of let-7c an additional RARE site was found in the C21orf34a intron 6, upstream of the let-7c initiation site (Figure 4c, lower panel). To investigate the role of PML/RARα in ATRA-mediated regulation of miR-107, -342 and let-7c expression, we performed experiments in PR9 cells, a U937-derived cell line in which exogenous PML/RARα expression is Zinc-inducible (Grignani et al., 1993). As shown in Figure 4b (right panel), zinc-induced expression of PML/RARα led to a clear-cut decrease of miR-342 and let-7c transcripts, whereas ATRA treatment strongly induced their expression. MiR-107 expression was not affected by PML/RARα induction, suggesting that in the case of miR-107, induction by ATRA might be due to an indirect mechanism.
These experiments prompted us to investigate whether the ability of ATRA treatment to induce miR-107, -342 and let-7c expression might be mediated in NB4 cells by a release of PANK1, EVL and C21orf34α promoters from endogenous PML/RARα repression. Using chromatin immunoprecipitation, we found that PML/RARα binds in vivo with variable affinity to the four RARE sites located in the promoter regions of EVL/miR-342 and two out of three RARE sites of the C21orf34α/let-7c promoter, but not the C21orf34α intronic RARE sites and the RARE sites in the PANK1 promoter (Figure 4d). As expected, the lack of PML/RARα recruitment onto the PANK1/miR-107 promoter confirms the results obtained in PR9 cells (Figure 4b, right panel) and strongly suggests that miR-107 modulation by ATRA is PML/RARα-independent and might be due to indirect mechanisms. After ATRA treatment, PML/RARα was released from all the occupied RARE sites (Figure 4d) and this correlated with the recruitment of the acetyltransferase p300 and an increased H3 histone acetylation (PAN-H3ac, H3K9ac and H3K14ac) around the RARE elements of both EVL/miR-342 and C21orf34α/let-7c promoters together with a sharp decrease in lysine 9 trimethyl-histone H3 (H3K9me3) at the RARE site proximal to the C21orf34α transcription start site (Figure 4d). Of note, when similar chromatin immunoprecipitation experiments were performed in the ATRA-resistant MR2 cells, we found that PML/RARα was retained on EVL/miR-342 and C21orf34α/let-7c promoters and that, accordingly, histone H3 acetylation status was not modified after ATRA treatment (Figure 4e).
These results confirm that the EVL/miR-342 and C21orf34α/let-7c promoters are targeted by PML/RARα in APL cells. Although additional mechanisms may be at work (that is, mutations of the PML/RARα fusion protein or deletion-specific miR loci modulated by ATRA), our observations strongly suggest that the ATRA-mediated changes of miR-342 and let-7c transcription are a specific effect of ATRA treatment in APL patients in vivo.
The miRNA target prediction analysis performed using Pictar, miRanda, TargetScan and miRBase Targets software has confirmed a number of putative targets for miR-107, -342 and let-7 family that might have a role in APL pathogenesis and/or ATRA treatment response. miR-107 has been shown to target the transcription factor NFI-A (Garzon et al., 2007), which participates with C/EBPα in the regulation of granulocyte differentiation-associated miR-223 expression (Fazi et al., 2005). miR-342 has been shown to stimulate granulocytic differentiation (De Marchis et al., 2009). Interestingly, the putative target of miR-342 MEIS1, a member of the TALE family of homeodomain genes, has been shown to have a pivotal role in normal hematopoiesis (Abramovich and Humphries, 2005; Diaz-Blanco et al., 2007) and it is frequently upregulated in human AMLs (Kumar et al., 2009). Finally, it is known the involvement of let-7 family members in differentiation and development, as well as in anti-proliferative functions, by targeting the RAS oncogene and the non-histone DNA binding protein HMGA2 (Johnson et al., 2005; Peng et al., 2008).
In conclusion, we identified a small subset of miRNAs that are modulated during normal granulocyte differentiation and represent new markers in the therapeutic response in APL patients.
Conflict of interest
The authors declare no conflict of interest.
We thank Dr WH Miller Jr for kindly providing the NB4 sub-clone MR2 and Dr I Sperduti for his help on statistical analysis. This work was supported by grants from Associazione Italiana per la Ricerca sul Cancro (AIRC), Ministero della Salute-Italy and European Community (Eu Active p53 Consortium).
About this article
Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)
Scientific Reports (2017)