In this work we investigated the molecular mechanisms that control CDC25A transcription and translation in acute myeloid leukemia cells that express the FLT3-ITD mutant tyrosine kinase receptor. This investigation picks up where our previous study leaves off, in which we identified this phosphatase as a master determinant of proliferation and differentiation arrest in this leukemic subtype⁷. CDC25A major function is the activation of CDK complexes, but its direct impact has been also reported on the tyrosine kinase receptor EGRF¹⁷. In consequence, we tested whether in our model, CDC25A could directly regulate FLT3-ITD activity. As can be seen in Supplementary Fig. 3, CDC25 inhibition did not affect tyrosine kinase phosphorylation of FLT3-ITD, nor STAT5 phosphorylation, suggesting that CDC25A most likely regulates AML cells proliferation through its canonical function of CDKs activator.

In general, CDC25A expression is tightly regulated in myeloid malignancies that express tyrosine kinase mutations^7,11,18 in which STAT5 and STAT3 pathways are most often activated. Transcriptional regulations of CDC25A by E2F1¹⁹, myc²⁰ and STAT3²¹ have been described previously in different cellular contexts. However, the question of STAT5-mediated transcriptional regulation of CDC25A remained an open one. In our study, STAT5 is identified at the CDC25A promoter for the first time, and we highlight the importance of STAT5 activity for proper CDC25A transcription downstream of the FLT3-ITD receptor. It has been reported that STAT3 associates with myc or Rb to respectively activate or inhibit CDC25A transcription²¹. Therefore, whether STAT5 associates with transcriptional partners to finely tune CDC25A transcription in AML cells remains an interesting question.

The importance of STAT factors in hematological models has been extensively documented^22,23, and researchers have been taking interest in the clinical relevance of these factors as targets in various cancer models^24,25. Recently, a study in MDS and AML models showed an impairment in leukemic growth in vivo and an increase in differentiation in primary progenitor cells with the use of a clinical antisense nucleotide that targets STAT3²⁶. Our current findings along with those of other studies^27,28 highlight and specify the importance of STAT5 as part of an essential pathway for FLT3-ITD AML proliferation and survival, and efforts to target STAT5 in AML are in progress²⁹.

In addition to this direct transcriptional regulation, we have highlighted another CDC25A regulation pathway downstream of FLT3-ITD involving STAT5 as well. In fact, we identified micro-RNA 16 as a key negative regulator of CDC25A protein expression downstream of FLT3-ITD and STAT5 in AML cells. First, our data support the argument that STAT5 is an actor in miR-16 regulation downstream of FLT3-ITD. STAT5 has been shown to regulate miR-16-2 and miR15-b expression in multiple myeloma³⁰ by inhibiting the transcription of their cluster gene, SMC4. However, in AML cell lines, we found no effect of FLT3 inhibition on miR-15-a, or -b expression, which suggests a more complex regulation of miR-16 that may involve maturation rather than transcriptional regulation of one or both cluster genes (for a review see³¹).

Our data also point to the fact that miR-16 is an important determinant of CDC25A protein level downstream of FLT3-ITD. MiR-16-dependent regulation of CDC25A has been previously reported in response to UV radiation-induced DNA damage, participating in the checkpoint response and proliferation arrest observed in these conditions¹⁴. In this study, we demonstrate that miR-16-dependent regulation of CDC25A is not restricted to DNA damage response, and could participate in the response of leukemic cells to tyrosine kinase inhibitors. There are 2 seed sequences of miR-16 in the 3’UTR region of CDC25A mRNA. This strongly suggests a direct interaction between miR-16 and CDC25A mRNA, as was suggested with a luciferase reporter system in Hela cells in response to UV radiation¹⁴. However, in this study we did not detect a significant impact of miR-16 expression on CDC25A mRNA level in AML cells, which suggests that depending on the cell and the environmental conditions, the functional miR-16 molecular mechanisms may differ.

In a previous work, Kim et al. described a down-regulation of miR-16 by FLT3-ITD¹⁵. Pim1 regulation by miR-16 has been described in a murine cell line model overexpressing FLT3-ITD¹⁵. However our data do not support the argument that Pim1 is a miR-16 target in the human FLT3-ITD AML cell lines MOLM-14 and MV4-11 (see Fig. 3E). It remains to be established whether this discrepancy is due to the distinct AML cell models used in both studies or to other unknown parameters. However, considering that transcriptional regulation of Pim1 by STAT5 is a well-established event downstream of FLT3-ITD^32,33, this may suggest the existence of a complex oncogenic signaling network including STAT5, miR-16, Pim1, and CDC25A that still needs to be clarified.

MiR-16 is considered to be a tumor suppressor because it negatively regulates the expression of pro-oncogenic proteins involved in cell proliferation and cell death. Therefore, some studies aimed at expressing miR-16 as a therapeutic approach have been developed in a murine model of CLL^34,35. However the use of micro RNAs as therapeutic tools remains rather unsuccessful as they have never progressed beyond a phase I trial³⁶. This emphasizes the need for a clearer understanding of the networks of proteins regulated downstream of these oligonucleotides. Although we describe CDC25A as an important target of miR-16 in this study, it is unlikely that all the effects of miR-16 expression in FLT3-ITD cells are only due to CDC25A regulation. For instance, in terms of the effects of miR-16 expression on the differentiation process, it should be noted that this RNA is involved in retinoic acid induced differentiation³⁷, and that PU.1, an important transcriptional regulator of myeloid differentiation, has been described as a miR-16 target³⁸. Whether or not the miR-16-dependent differentiation that we observed in FLT3-ITD AML cells is linked with these pathways remains to be investigated. With regards to its cancer relevance, another interesting miR-16 target that has been established is the anti-apoptotic protein Bcl2³⁹. Bcl2 inhibitors are currently being introduced as potential therapeutic tools in AML⁴⁰, and interestingly, a synergy between venetoclax and the FLT3-inihibitor quizartinib has been described⁴¹. It would be interesting to establish whether miR-16-dependent regulation of Bcl2 is involved in this process.

In conclusion, our data point to the following model, in which FLT3-ITD activity leads to miR16 repression through a STAT5-dependent mechanism that remains to be precised. In consequence, STAT5 activation downstream of FLT3-ITD contributes to CDC25A protein expression by two different ways: STAT5 represses miR16, which is a negative regulator of CDC25A translation in this model, and directly participates to CDC25A gene transcription. This pathways participates to proliferation and differentiation arrest of FLT3-ITD AML cells. This model is summarized in Fig. 6.