SNAIL trail in myeloid malignancies

In this issue of Blood, Carmichael et al¹  provide compelling evidence for the pivotal role of SNAI1 (Snail family transcriptional repressor 1, also known as SNAIL) overexpression in the development and maintenance of myeloid malignancies, thus opening a new window for therapeutic intervention.

Transcription factors known to induce the epithelial-to-mesenchymal transition (EMT) (such as ZEB1/2 [zinc finger E-box binding homeobox 1/2], SNAI1/2/3, and TWIST1/2) have been undoubtedly implicated in tumorigenesis, cancer progression, metastasis, and chemoresistance in solid tumors² ; however, their role in normal and malignant hematopoiesis has been underappreciated for many years. To date, the ZEB proteins are probably the most studied EMT regulators, since recurrent genetic alterations or deregulated expression of ZEB1 and ZEB2 genes has been reported in several subtypes of lymphoma, as well as in lymphoid and myeloid leukemia.³  Indeed, recent studies have shown the functional impact of ZEB2 in acute myeloid leukemia (AML) and T-cell acute lymphoblastic leukemia development.^4,5 

In this study, Carmichael et al showed that high expression of SNAI1 in primary samples from AML patients is linked to inferior prognosis of these patients. Next, using a series of human and animal models, they elegantly demonstrated that SNAI1 overexpression impairs myeloid differentiation/maturation and promotes proliferation and self-renewal of myeloid progenitors in vitro. This was correlated with an increased proportion of short-term hematopoietic stem cells and a bias of hematopoietic progenitors toward granulomonocytic progenitors and the myeloid lineage in SNAI1-transgenic mice. Strikingly, aged mice overexpressing SNAI1 developed myeloproliferative neoplasms and were predisposed to myeloid leukemia. Last, they showed that loss of Snai1 prolonged survival in 2 independent murine AML models, further emphasizing the role of SNAI1 in leukemia development, even when driven by various genetic alterations.

Mechanistically, Carmichael et al highlighted that the leukemogenic impact of SNAI1 overexpression is correlated, at least partly, to its direct interaction with the histone demethylase KDM1A/LSD1, and decreased methylation of active H3K4me1/2 marks at genomic regions bound by SNAI1. Briefly, the integration of results obtained from next-generation sequencing experiments (RNA-seq, ChIP-seq, and ATAC-seq) lead to a model where SNAI1 overexpression corrupts LSD1 function through potentially 3 mechanisms (hijacking, replacement, and depletion), to impair myeloid differentiation and promote MPN/AML development (see figure). This molecular interplay is particularly relevant since targeting LSD1 has recently emerged as a new therapeutic option to treat AML, with high efficacy in vitro, but low responses in clinical trials.⁶  Therefore, this study highlights the need for a greater understanding of the mechanisms of action of SNAI1/LSD1 complexes, which will facilitate the development of alternative approaches targeting the SNAI1/LSD1 interaction or clinically relevant downstream effectors. Whether this could replace the use of LSD1 inhibitors in SNAI1-overexpressing AML or, on the contrary, potentiate their efficacy by releasing KDM1A/LSD1 from genomic regions bound by SNAI1, is an attractive area of investigation. Of potential interest also, SNAI1 has been shown to interact with many other epigenetic regulators with known functions in AML development, including HDAC1/2, Sin3A, and the PRC2 complex in other biological systems.⁷  Therefore, full characterization of the epigenetic landscape in SNAI1-overexpressing AML cells will facilitate the development of combination therapies potentially including inhibitors of epigenetic regulators other than LSD1.

Finally, while the consequences of SNAI1 overexpression in the development of myeloid malignancies are now coming to light, the mechanism(s) leading to increased expression of SNAI in AML samples remains elusive. During EMT and cancer progression, SNAI1-overexpression can be mediated by several extracellular signals (including TGFβ, WNT, NOTCH, RTK) and microRNAs (miR-34, miR-29b),²  some of these being genetically altered or deregulated in leukemia samples. Although the present study assessed SNAI1-expression in a small number of AML samples, integration of large datasets may refine this analysis and unravel genetic or phenotypic reasons explaining high level of SNA1 transcripts. SNAI1 protein stability should also be assessed, since posttranscriptional mechanisms have been described.²  Notably, there is growing evidence that several EMT-regulators, SNAI1, ZEB1, and TWIST1, are overexpressed in AML samples,^8,9  and since these factors have now been implicated in leukemia development and maintenance, it will be of interest to investigate whether these regulators cooperate functionally in leukemogenesis, as they do during EMT. Moreover, because EMT-regulators are implicated in chemoresistance in solid tumors, additional experiments exploring this in hematological malignancies are warranted.

In conclusion, this study by Carmichael et al adds a new pillar to support the idea that EMT-regulators are key players in the development of human myeloid leukemia, in addition to other hematological malignancies, thus providing new therapeutic opportunities to improve patient outcomes.