Targeted CRM1-inhibition perturbs leukemogenic NUP214 fusion proteins and exerts anti-cancer effects in leukemia cell lines with NUP214 rearrangements

Chromosomal translocations fusing the locus of nucleoporin NUP214 each with the proto-oncogenes SET and DEK are recurrent in, largely intractable, acute leukemias. The molecular basis underlying the pathogenesis of SET-NUP214 and DEK-NUP214 are still poorly understood, but both chimeras inhibit protein nuclear export mediated by the ß-karyopherin CRM1. In this report, we show that SET-NUP214 and DEK-NUP214 both disturb the localization of proteins essential for nucleocytoplasmic transport, in particular for CRM1-mediated protein export. Endogenous and exogenous SET-NUP214 and DEK-NUP214 form nuclear bodies. These nuclear bodies disperse upon targeted inhibition of CRM1 and the two fusion proteins re-localize throughout the nucleoplasm. Moreover, SET-NUP214 and DEK-NUP214 nuclear bodies reestablish shortly after removal of CRM1 inhibitors. Likewise, cell viability, metabolism, and proliferation of leukemia cell lines harboring SET-NUP214 and DEK-NUP214 are compromised by CRM1 inhibition, which is even sustained after clearance from CRM1 antagonists. Our results indicate CRM1 as a possible therapeutic target in NUP214-related leukemia. This is especially important, since no specific or targeted treatment options for NUP214 driven leukemia are available yet.


Introduction
Chromosomal translocations involving the nucleoporin NUP214 have been described in de novo and therapy-related acute myeloid leukemia (AML) as well as acute lymphoblastic leukemia (ALL). NUP214-related malignancies are frequently associated with poor treatment response and poor prognosis 1-6 . The fusion proteins SET-NUP214 [del(9)(q34.11q34. 13)] and DEK-NUP214 [t(6;9)(p23;q34)] result from the fusion of the almost entire SET and DEK proteins with the C-terminal part of NUP214 1, 7,8 . NUP214 is an integral part of the nuclear pore complex (NPC) and it plays important roles in nuclear export mediated by chromosomal region maintenance 1 (CRM1, or exportin 1/XPO1) 9,10 . CRM1 is the major nuclear export receptor for proteins and ribonucleoprotein (RNP) complexes carrying a characteristic nuclear export signal (NES) 11,12 . NUP214 functions as a terminal docking site for CRM1 nuclear export complexes on the cytoplasmic side of NPCs and depletion of NUP214 results in nuclear accumulation of NEScontaining cargoes [13][14][15][16] .
The C-terminal phenylalanine-glycine (FG) repeat domain of NUP214 exhibits multiple CRM1-binding sites, which are preserved in SET-NUP214 and DEK-NUP214 [16][17][18][19] . In fact, both fusion proteins can bind CRM1 and its co-factor, the small GTPase Ran, and inhibit the nuclear export of NES-containing proteins and RNPs 17,18 . Targeted CRM1 inhibition by small molecule antagonists has become an appealing anti-cancer strategy, for both solid and hematologic malignancies [20][21][22][23][24] . Leptomycin B (LMB), a fungal metabolite from Streptomyces spp, was the first identified small molecule inhibitor specifically targeting CRM1 25 . LMB has potent anticancer activity, but its application in patients was withdrawn after a single phase I clinical trial because of its low efficiency and high toxicity 26 . Selective inhibitors of nuclear export (SINEs) comprise a novel class of CRM1 antagonists with anti-cancer properties both in vitro and in 4 vivo [20][21][22][23][27][28][29] . Indeed, the SINE compound KPT-330 is currently tested in phase 2/3 clinical trials for a wide variety of cancers, including leukemia and other hematologic malignancies 30 . The anti-cancer effects of CRM1 inhibitors are based on the induction of cell death by apoptosis and on cell cycle arrest due to activation of the transcriptional programs of tumor suppressor genes, such as TP53, RB1 and FOXO-related tumor suppressors 20,21,31 . Despite the functional proximity between NUP214 and CRM1-mediated nuclear export and the functional relevance of CRM1 in some leukemia, the impact of CRM1 inhibition in the context of NUP214-related leukemia has not been studied yet.
Here we used patient-derived leukemia cell lines expressing endogenous SET-NUP214 and DEK-NUP214 to address the anti-cancer potential of CRM1 inhibition in NUP214-rearranged leukemia. We report that CRM1 inhibition by LMB or the SINE compound KPT-185 is sufficient to disturb the localization of endogenous SET-NUP214 and DEK-NUP214, coinciding with reduced cell viability and proliferation. CRM1 inhibition reduces cell viability and metabolic activity in a sustained manner after drug withdrawal. However, after drug removal, proliferation and SET-NUP214 and DEK-NUP214 nuclear bodies are restored. Our data suggest that CRM1 constraint is an interesting candidate for the development of an anti-cancer therapeutic approach in NUP214-related leukemia, for which no efficient targeted therapy has been developed so far.

Cell lines
LOUCY (T-cell

Plasmids and Transfections
To generate SET-NUP214-GFP, total RNA was extracted from LOUCY cells using the High Pure RNA Isolation Kit (Roche Life Sciences, Basel, Switzerland) according to the manufacturer's instructions. cDNA synthesis was performed by reverse transcription-PCR.
Cloning of SET-NUP214 is described in the Online Supplementary Methods.
The pENTR1-DEK-NUP214 plasmid was a gift from Dr. Martin Ruthardt (Cardiff University, UK) and was subcloned into the peZY-EGFP destination vector using the Gateway™ LR clonase™ enzyme mix (Invitrogen, Merelbeke, Belgium), as described in the Online Supplementary Methods.

Immunofluorescence of suspension cells
Leukemia cells were seeded at 0.8x10 6 cells/ml and grown for 24 h -48 h. Cells were fixed and processed for immunofluorescence as detailed in the Online Supplementary Methods.

Immunofluorescence of adherent cells
HCT-116 cells were seeded on polylysine-coated glass coverslips 24 h prior to transfection.
After transfection, cells were fixed in 2% formaldehyde and processed for immunofluorescence as detailed in the Online Supplementary Methods.
After treatment, cells were washed and processed for immunofluorescence as detailed in the Online Supplementary Methods.

Inhibition of CRM1-mediated nuclear export
Cells were seeded at 0.

Statistics
Experiments were performed at least three times and the results represent the mean ± SEM for three independent biological replicates. Plots were generated and statistical analysis was performed using GraphPad Prism (Version 5.01; GraphPad Software Inc., CA, USA). Statistical differences were calculated by one-way analysis of variance (ANOVA). During evaluation of the results a confidence interval α of 95% and p values lower than 0.05 were considered as statistically significant. Significance levels are represented as *(p<0.05), **(p<0.01) or **(p<0.001).

Leukemogenic NUP214 fusion proteins locate to nuclear bodies in patient-derived cells
We first determined the localization of NUP214 fusion proteins in different patient-derived leukemia cell lines with anti-NUP214 antibodies and immunofluorescence microscopy. LOUCY and MEGAL cells express SET-NUP214 and in both cell lines SET-NUP214 located to the nuclear rim and to nuclear bodies ( Figure 1A), consistent with previous results 17,32 . FKH-1 cells harbor DEK-NUP214, which localized to smaller nuclear bodies as compared to SET-NUP214 ( Figure 1A) 33 . Similar localizations for GFP-tagged versions of SET-NUP214 and DEK-NUP214 were observed in transiently transfected HCT-116 cells ( Figure 1B). In FKH-1 cells, NUP214 antibodies were also detected at the nuclear rim, which likely corresponds to endogenous NUP214 rather than to the fusion protein, as DEK-NUP214-GFP in HCT-116 was not detected at NPCs (Figure 1 A-B). In OCI-AML1 and MOLM-13 cells, which do not express NUP214 fusion proteins, NUP214 staining displayed the typical punctate pattern of nucleoporins at the nuclear rim ( Figure 1A).
The FG domains of nucleoporins are, due to their amino acid composition, intrinsically disordered and exhibit variable degrees of cohesiveness, which is important for the maintenance of the NPC permeability barrier [34][35][36] . 1,6-hexanediol (HD) is a mild alcohol that interferes with hydrophobic interactions established between FG repeats 34 , thereby disrupting the NPC permeability barrier 37,38 . To address the potential role of FG repeat cohesion in the formation of NUP214 nuclear bodies, we treated LOUCY cells, which have the most prominent nuclear bodies, with 5% HD and monitored SET-NUP214 nuclear bodies over time by immunofluorescence microscopy. As shown in Figure 1C, untreated cells represented with prominent SET-NUP214 nuclear bodies of different size and frequency. These SET-NUP214 nuclear bodies started to dissolve as early as 2 min after HD treatment and vanished after 10 min HD exposure, as indicated by a largely spread and homogeneous distribution of SET-NUP214 throughout the nucleus ( Figure 1C). Hence, we concluded that the formation of SET-NUP214 nuclear bodies depends, at least in part, on cohesive interactions between the FG repeats of NUP214.

NUP214 fusions accumulate the nuclear export factor CRM1 in their nuclei and perturb the localization of endogenous nucleoporins
We have previously shown that SET-NUP214 in transfected cells interacts with the nuclear export factor CRM1 and that this interaction is enhanced in the presence of RanGTP 17 .
Accordingly, we found that SET-NUP214 nuclear bodies accumulate CRM1 in LOUCY ( 39 . To reinforce the notion that NUP214 fusion proteins preferentially bind CRM1-RanGTP nuclear export complexes, we co-expressed SET-NUP214-GFP and DEK-NUP214-GFP in HCT116 cells with RFP-tagged versions of Ran 17 . We employed RanQ69L (RanQ69L-RFP), a non-hydrolyzable mutant of Ran, and RanT24N (RanT24N-RFP), which is resistant to GTP loading and nuclear export complex formation 40 . As expected, both NUP214 fusion proteins sequestered RanQ69L-RFP, but not RanT24N-RFP to nuclear bodies (Supplementary Figure S1 A-B), complementing our previous biochemical data 17 .
Additionally, we observed that the fusion proteins sequestered endogenous NPC components: NUP88 ( Figure 2D) and NUP62 ( Figure 2E Figure 6A) and FKH-1 ( Figure 6C) cells were alive 72 h after exposure to LMB, as compared to 77% of MEGAL cells ( Figure 6B). The effect of KPT-185 on cell viability was slightly weaker in LOUCY ( Figure 6A) and MEGAL ( Figure 6B) cells as compared to LMB, but similar in FKH-1 cells ( Figure 6C). Together these data suggest that CRM1 inhibition affects the viability of cells expressing NUP214 fusion proteins and that MEGAL cells appear to be more resistant to CRM1 inhibition in comparison to LOUCY and FKH-1 cells.

LMB and KPT-185 persistently affect cellular function
As mentioned above, LMB binds CRM1 irreversibly, in contrast to KPT-185 41 . This let us ask how cells respond to treatment withdrawal, i.e. how persistent are the effects of the two CRM1 inhibitors in NUP214-rearranged leukemia cells. We hence exposed the three leukemia cell lines to LMB or KPT-185 as described above and monitored cell viability and metabolic activity at different time points after drug removal. We found that after LMB and KPT-185 removal, cell viability ( Figure 7A) and metabolic activity ( Figure 7B) of MEGAL and FKH-1 cells were significantly reduced for at least 48 h. Similarly, LOUCY cells showed a significant reduction of cell viability for at least 48 h after drug removal, accompanied by a decrease in their metabolic activity, which nevertheless was not statistically significant. Again, the effect of the two CRM1 inhibitors on cell viability was weaker in MEGAL cell as in LOUCY and FKH-1 cells ( Figure 7A; see also Figure 6). The persistent negative effect of CRM1 inhibition on cellular fitness of the leukemia cell lines was similarly observed in colony-forming assays (see Material and Methods) conducted with MEGAL and FKH-1 cells ( Figure 7C). MEGAL cells formed colonies, which were smaller (< 50 µm) than colonies formed by non-treated cells ( Figure 7C, top panel), whereas colonies formed by FKH-1 cells were reduced in size and number compared to non-treated cells ( Figure 7C, bottom panel). These data further support the notion that MEGAL cells are slightly more resistant to CRM1 inhibition, when compared to LOUCY and FKH-1 cell lines. We additionally performed the experiment with LOUCY cells, however due to the small size of the colonies it was not possible to properly evaluate potential differences between treated and non-treated cells.

Proliferation of LOUCY and FKH-1 cells was restored 48 h after KPT-185 removal.
Proliferation of LMB-treated FKH-1 cells further decreased 48 h after drug removal, but also non-treated cells were less proliferative.

Discussion
The pathologic potential of NUP214 fusion proteins has been widely recognized, but the underlying molecular mechanisms have been only sparsely studied 17,19,[42][43][44] . NUP214 is an important player in nuclear export mediated by the ß-karyopherin CRM1 and CRM1 inhibition in turn has been proven beneficial in clinical trials as anti-cancer strategy. Here we confirm that endogenous and exogenous SET-NUP214 and DEK-NUP214 form nuclear bodies that accumulate CRM1 and its co-factor RanGTP (Figure 1 A,  functions as a scaffold for the assembly of the fusion proteins into nuclear bodies. In any case, the disruption of these structures upon CRM1 inhibition indicates that NUP214 fusion proteins require CRM1 activity for nuclear body formation and possibly for their oncogenic capacities. Consistently, treatment of NUP214-rearranged leukemia cell lines with LMB or KPT-185 affects viability, metabolism, and proliferation, albeit to a different extent (Figure 6-8). This effect is sustained even after drug removal (Figure 7), likely to the stability in drug binding to CRM1.
Binding of LMB and KPT-185 to CRM1 is energetically favored, i.e. both drugs not only bind free CRM1, but they also disrupt existing interactions between CRM1 and its cargoes 41 . Binding of LMB to CRM1 results in a covalent, irreversible interaction, whereas multiple hydrophobic bridges between KPT-185 and CRM1 result in a slowly reversible interaction 41 .
MEGAL cells respond different to CRM1 inhibition compared to LOUCY (and FKH-1) cells, i.e. they appear to be more robust/resistant (Figure 6-8). These discrepancies between MEGAL and LOUCY cells, which both express SET-NUP214, may be explained by differences in CRM1 levels in these two cell lines. Indeed, preliminary data indicate that in particular KPT-185 treatment coincided with reduced CRM1 levels in LOUCY cells and to a lesser extent in MEGAL cells and was also sustained after drug removal although some recovery could be observed (Supplementary Figure S3 A-B). LMB treatment on the contrary appears to not alter CRM1 protein levels. Further detailed analysis are required here, but our results are in accordance with previous reports which have shown that KPT-185 treatment led to reduced CRM1 protein levels, but not so LMB 20,[45][46][47][48] .
Cellular response to CRM1 inhibitors in leukemia has moreover been reported to depend on the mutational status of the tumor suppressor TP53 21 . LOUCY cells harbor a missense mutation p53 V272M (Supplementary Figure S4A), which results in functionally inactive p53 49,50 . MEGAL cells feature a heterozygous frameshift mutation in TP53 due to deletion of cytosine at position 898, resulting in a frameshift after a leucine residue at position 299, which affects p53's 42 Cterminal residues (c.del898C, p53 299fs*42 ; Supplementary Figure S4A). This mutant form of p53 in MEGAL cells is comparatively more abundant than the wild-type protein (Supplementary Figure S4B). The biological implications of p53 299fs*42 remain to be elucidated. However, one may hypothesize that p53 299fs*42 acts as dominant negative and hinders the biological function of the wild type counterpart, which may be stabilized by CRM1 inhibition 51,52 . Such a dominantnegative effect of p53 299fs*42 may contribute to the reduced sensitivity of MEGAL cells to CRM1 inhibition. Such a reduced sensitivity may arise from a reduced nuclear accumulation of tumor suppressor genes, delayed cell cycle, and/or reduced variation in overall protein expression after CRM1 inhibition, as seen in a fibrosarcoma cell line treated with the SINE KPT-330 53 . However, future studies are required to further elucidate this possibility.
Taken together, CRM1 inhibition might be an interesting therapeutic option in NUP214related leukemia, similarly as described for several cancer models, including other forms of leukemia 23,24,29,54,55 . This notion is further supported by a recent study that showed successful disease remission for an AML patient with NUP214 driven leukemia who was treated with KPT-330 as single agent 56,57 . This is particularly important, since no specific or targeted therapy has been developed so far 56 .