A Role for the CRM1 Nuclear Export Receptor in HOXA Gene Deregulation and CALM-AF10 Leukemogenesis

Background: CALM-AF10 translocations are seen in 5-10% of T-cell acute lymphoblastic leukemia (T-ALL) and are characterized by HOXA gene overexpression, a feature shared with leukemias associated with translocations of the Mixed-Lineage Leukemia (MLL) gene. The ability of CALM-AF10 to aberrantly activate HOXA gene transcription is mediated by interaction of the AF10 moiety with the H3K79 methyltransferase DOT1L. We previously demonstrated that a nuclear export signal (NES) within CALM is essential for CALM-AF10's ability to upregulate HOXA genes and cause leukemia in mice. The CRM1 (XPO1) nuclear export receptor specifically interacts with NES motifs, including the CALM NES. Fusing an 18 amino acid NES motif to AF10 recapitulates the oncogenic properties of CALM-AF10. Interfering with the CRM1/CALM-AF10 interaction by either genetic (mutation of the CALM NES motif) or pharmacologic (Leptomycin B - LMB) inhibition abolishes CALM-AF10's ability to bind to HOXA genes and activate their expression. We discovered that CRM1 itself is able to bind to HOXA loci, suggesting that CRM1 may recruit CALM-AF10 to its target genes. The NES binding cleft within CRM1 is maintained in a closed conformation by the C-terminal helix (CRM1 aa 1029-1071), and its truncation greatly enhances the NES-binding affinity of CRM1.

Hypothesis: Since the NES motif of CALM is necessary and sufficient for the leukemogenic properties of CALM-AF10, we hypothesized that a CRM1-AF10 fusion protein would also be leukemogenic. To determine the importance of the NES binding ability of CRM1 in leukemogenesis, we investigated a deletion mutant of CRM1-AF10 that lacks the inhibitory C-terminal 43 amino acids of CRM1 (CRM1Δ-AF10).

Methods: Confocal immunofluorescence microscopy of transfected HEK293 cells was used to study the intracellular localization of CRM1-AF10 and CRM1Δ-AF10 fusion proteins, and their transcriptional activity wasmeasured with a HOXA7 -luciferase reporter. We used LMB to study the impact of blocking the CRM1/NES interaction on HOXA7 transcriptional activation and cellular localization. Retroviral transduction and transplantation of murine bone marrow were performed to determine the ability of CRM1-AF10 and CRM1Δ-AF10 to induce leukemias.

Results: CALM-AF10, CRM1-AF10 and CRM1Δ-AF10 fusion proteins localize to both the cytoplasm and nucleus (in a speckled pattern). While all three fusion proteins induced HOXA gene transcription compared with empty vector, CALM-AF10 (7.3 fold increase) and CRM1Δ-AF10 (7.8 fold) were significantly more potent than CRM1-AF10 (5.1 fold). LMB treatment (4 nM, 25 h) repressed HOXA7 transcription induced by CALM-AF10 (68% repressed), CRM1-AF10 (31% repressed), and CRM1Δ-AF10 (60% repressed). LMB exposure (2 nM, 2 h) induced nuclear retention of all three fusion proteins. In vivo, while all three fusions induced murine myeloid leukemias characterized by overexpression of Hoxa genes, CRM1-AF10 leukemias displayed incomplete penetrance (50%) with prolonged latency (250-300 days) compared with both CALM-AF10 and CRM1Δ-AF10, which demonstrated complete penetrance and similar latencies (100-130 days).

Conclusions: We have shown that CRM1 can substitute for CALM to induce HOXA gene expression and cause leukemias when fused to AF10, suggesting that the contribution of the CALM moiety in the CALM-AF10 oncoprotein is to recruit CRM1. Deletion of the C-terminal inhibitory helix of CRM1 potentiates the oncogenic properties of CRM1-AF10. Moreover, pharmacologic inhibition of the NES/CRM1 interaction with LMB counteracts the ability of CRM1-AF10 and CRM1Δ-AF10 to activate the transcription of HOXA genes. These results suggest that binding to an as yet unidentified, NES-containing partner protein plays an important role in CRM1-AF10 leukemogenesis. Of note, the CRM1 C-terminal helix is missing in a CRM1-AF10 translocation described in a T-ALL patient (Bond, BLOOD, 2014). These studies shed light on the importance of CRM1 in leukemias associated with HOXA deregulation, and warrant further studies of the mechanisms underlying the therapeutic efficacy of CRM1 inhibitors (selective inhibitors of nuclear export - SINEs) that target the CRM1/NES interaction.