Childhood leukemia

AML is the product of a limited number of driver mutations (e.g. fusion genes, point mutations or deletions) in chromatin modifiers, transcription factors, nucleophosmin, regulators of the splicing machinery, tumor suppressors, members of the cohesion complex, and regulators of DNA methylation. Alterations in these pathways often occur in combination with mutated signaling mediators of which the constitutively activated FLT3-ITD receptor tyrosine kinase is among the most prevalent. Our lab is particularly interested in leukemogenic alterations of chromatin modifiers such as mixed lineage leukemia 1 (MLL1), nucleoporin 98 (NUP98) and the nuclear receptor interacting SET domain protein 1 (NSD1).

Alterations of transcriptional regulators of normal hematopoiesis are molecular hallmarks of AML. In over 50% of the patients their leukemic cells harbor chromosomal translocations that frequently lead to the expression of fusion genes with leukemogenic activity. Fusions involving genes such as mixed lineage leukemia (MLL) or nucleoporin 98 (NUP98) are recurrently found in aggressive AML. MLL-AF9, MLL-ENL or NUP98-NSD1 are among the most prevalent, and particularly the latter two are are predominantely found in aggressive pediatric acute leukemia. Rather divergent outcomes of patients with the same MLL fusion indicates that other factors such as the cell of origin in which the driver mutations was initially formed may influence the biology and the clinical outcome of the disease.

To study the cellular origin in AML, we established several transgenic mouse lines in which we can activate a leukemia-associated fusion in a particular cell type and to follow the development of the disease (in collaboration with Antoine Peters, FMI, Basel). Using “iMLL-AF9”, an inducible transgenic mouse model for MLL-AF9-driven leukemia we found that expression of the fusion in long-term hematopoietic stem cells induced in 10-20% of the mice a particularly aggressive with extensive tissue infiltration, chemoresistance, and expression of genes related to epithelial-mesenchymal transition (EMT) a process that characterizes progressing solid cancers. By classifying mouse and human leukemias according to genes that reflect aggressiveness and cell of origin, and by comparative transcriptomics, we identified several EMT-related genes that were significantly associated with poor overall survival of AML patients (Figure 1; Starvopoulou et al., Cancer Cell, 2016)

Induction of iMLL-AF9 in long-term hematopoietic stem cells (LT-HSC) resulted in a more aggressive disease than activation in granulocyte-macrophage progenitors (GMP). Invasive LT-HSC-derived AML was characterized by high expression levels of the Evi1 and Erg transcription factors, and by an EMT-like gene expression signature with direct induction of Zeb1. Interestingly, high expression levels of EVI1 and ERG also characterized AML patients with poor outcome suggesting a similar impact of the cellular origin on the biology of the disease in mouse and man.

The nucleoporin 98 (NUP98) gene is target of recurrent chromosomal translocations leading to expression of NUP98 fusions to >30 different partners The role of NUP98-fusions in AML is poorly understood. In the past, we have cloned several NUP98-fusion cDNAs from patient’s cells and demonstrated the leukemogenic activity of a NUP98-HHEX fusion in vitro and in a mouse model (Jankovic et al., Blood, 2008).

The NUP98-NSD1 fusion is a hallmark of paediatric AML with poor outcome. In majority of these patients the NUP98-NSD1+ tumor cells also carry an activating mutation of a receptor tyrosine kinase called FLT3-ITD. We were able to show that, in contrast to a previous study, cooperation of NUP98-NSD1 with FLT3-ITD is essential for the leukemogenic activity in vivo (Thanasopoulou et al., Haematologica, 2014).

NUP98 acts as a nuclear pore protein but can also regulate gene expression. By comparing the subnuclear localization of several NUP98 fusions we found that despite fusion partner-related differences in localization, NUP98 fusions provoked unusual morphological alterations in the nuclear envelope not only in transfected HeLa cancer cells but also in mouse bone marrow cells immortalized by NUP98 fusions, and in cells from NUP98-fusion+ cells from leukemia patients. Although the functional significance remains to be elucidated, our observations suggest that the transforming activity of NUP98-fusions may be linked to altered nuclear stability (in collaboration with Birthe Fahrenkrog, Brussels; Fahrenkrog et al., PlosOne, 2015).

For a long time it was thought that genetically altered transcriptional regulators could not be pharmacologically targeted. However, the immense improvement of structural molecular analysis resulted in the identification of several small molecules that selectively interfere with gene transcription. Active transcription becomes possible through local loosening of the chromatin by reversible modification of the histone proteins. These markers are then recognized through interaction by distinct structural domains of the transcriptional regulators including the so-called bromodomains.

The Structural Genomics Consortium (SGC, Oxford, S. Knapp, P. Fillpakopoulos and coworkers) has solved the structure more most existing bromodomain that allowed identification of several selectively interacting small molecules. As the activity of MLL-fusions seems to be mediated by multiple bromodomain interactions we have tested the anti-leukemic activity of several scaffolds including Pfi-1 and JQ1 targeting bromodomains of the BET protein family such as Brd4 (Picaud et al., Cancer Res, 2013).

CBP/EP300 are two highly homologous bromodomain-containing proteins acting as transcriptional co-activators that are involved in recurrent leukemia-associated chromosomal translocations. The Knapp laboratory developed a specific and potent acetyl-lysine competitive protein-protein interaction inhibitor I-CBP112 that targets the CBP/EP300 bromodomains. We found that I-CBP112 substantially impaired colony formation and induced cellular differentiation without significant cytotoxicity of several human and mouse leukemic cell lines as well as primary AML cells. I-CBP112 also significantly reduced the leukemia-initiating potential of MLL-AF9+ AML cells in vivo. The synergistic effects of I-CBP112 with current standard therapy (doxorubicin) as well as emerging treatment strategies (BET inhibition) provide new opportunities for combinatorial treatment of leukemia and potentially other cancers (Picaud et al. Cancer Res, 2015; in collaboration with Stefan Knapp, Oxford & Frankfurt).

Our work aims to understand the molecular mechanisms underlying childhood acute myeloid leukemia (AML), a rare but aggressive disease often associated with poor outcome. In the past we have identified and functionally characterized several AML-associated genetic alterations, identified and characterized aberrantly activated pathways and defined potential novel therapeutic strategies. During the last year, we followed three topics:

1) We investigated the impact of the cellular origin for AML biology and clinical outcome. Hereby we have established novel conditional transgenic mouse models for acute leukemia induced by fusion oncogenes like MLL-AF9 and MLL-ENL. We found that activation of the MLLAF9 fusion in hematopoietic stem cells leads to significantly more aggressive disease than activation in more committed progenitor cells. Importantly, we observed significant overlaps of gene expression signatures of aggressive stem cell-derived mouse MLL-AF9 AML with signatures from patients with MLL-fusions. Interestingly, we identified several genes previously known for their role in progression of solid cancers to be drivers of aggressive AML, that might not only be prognostic markers but also potential origin-related therapeutic targets (Stavropoulou et al., Cancer Cell, 2016).

2) We addressed the role of the NUP98-NSD1 fusion, a molecular hallmark of very aggressive childhood AML. By using a retroviral expression transplantation approach we found, in contrast to previous reports, that the leukemogenic activity of NUP98-NSD1 depends on cooperation with an activating mutation of the FLT3 receptor tyrosine kinase (Thanasopoulou et al., Haematologica 2014). To overcome the limitations of viral integration and overexpression we now study the leukemogenic effects of NUP98-NSD1 in a novel inducible mouse model (in preparation).

3). We studied the sub-nuclear localization of 10 different NUP98 fusions (in collaboration with Birthe Fahrenkrog, Brussels). Despite the difference in localization, expression of all tested NUP98-fusions provoked unusual morphological alterations in the nuclear envelope (NE). Importantly, such aberrations were not only observed in transiently transfected adherent cancer cells (HeLa) but also in mouse bone marrow cells immortalized by NUP98 fusions, and in cells derived from leukemia patients harboring NUP98 fusions. Although the functional significance of these alterations remains to be investigated, our observations suggest that NUP98 fusion-associated NE alterations that may contribute to leukemogenesis (Fahrenkrog et al., Plos One, 2015).

4) We explored the anti-leukemic potential of a novel small molecule that selectively binds to the bromodomain of the CBP/EP300 transcriptional co-regulators (established by the group of Stefan Knapp, SGC Oxford). Exposure of human and mouse leukemic cell lines to I-CBP112 resulted in substantially impaired colony formation and induced cellular differentiation without significant cytotoxicity. I-CBP112 also significantly reduced the leukemia-initiating potential of MLL-AF9⁺ acute myeloid leukemia cells in a dose-dependent manner in vitro and in vivo. Interestingly, I-CBP112 increased the cytotoxic activity of BET bromodomain inhibitor JQ1 as well as doxorubicin. The synergistic effects of I-CBP112 with current standard therapy (doxorubicin) as well as emerging treatment strategies (JQ1 BET inhibitor) provide new opportunities for combinatorial treatment of leukemia and potentially other cancers (Picaud et al., Cancer Res, 2015).

Critical role of retinoid/rexinoid signaling in mediating transformation and therapeutic response of NUP98-RARG leukemia. Qiu JJ, Zeisig BB, Li S, Liu W, Chu H, Song Y, Giordano A, Schwaller J, Gronemeyer H, Dong S, So CW. Leukemia. 2015 May;29(5):1153-62.

Generation of a Selective Small Molecule Inhibitor of the CBP/p300 Bromodomain for Leukemia Therapy. **Picaud S, **Fedorov O, **Thanasopoulou A, **Leonards K, Jones K, Meier J, Olzscha H, Monteiro O, Martin S, Philpott M, Tumber A, Filippakopoulos P, Yapp C, Wells C, Che KH, Bannister A, Robson S, Kumar U, Parr N, Lee K, Lugo D, Jeffrey P, Taylor S, Vecellio ML, Bountra C, Brennan PE, O'Mahony A, Velichko S, Müller S, Hay D, Daniels DL, Urh M, La Thangue NB, Kouzarides T, Prinjha R, Schwaller J*, Knapp S*. Cancer Res. 2015 Dec 1;75(23):5106-19 (*shared senior authorship; **shared 1st authorship).

Expression of Leukemia-Associated Nup98 Fusion Proteins Generates an Aberrant Nuclear Envelope Phenotype. Fahrenkrog B, Martinelli V, Nilles N, Fruhmann G, Chatel G, Juge S, Sauder U, Di Giacomo D, Mecucci C, Schwaller J. PLoS One. 2016 Mar 31;11(3):e0152321.

MLL-AF9 Expression in Hematopoietic Stem Cells Drives a Highly Invasive AML Expressing EMT-Related Genes Linked to Poor Outcome. Stavropoulou V, Kaspar S, Brault L, Sanders MA, Juge S, Morettini S, Tzankov A, Iacovino M, Lau IJ, Milne TA, Royo H, Kyba M, Valk PJ, Peters AH*, Schwaller J*. Cancer Cell. 2016 Jul 11;30(1):43-58 (*shared senior authorship).