XPO1 inhibitor combination therapy with bortezomib or carfilzomib induces nuclear localization of IκBα and overcomes acquired proteasome inhibitor resistance in human multiple myeloma

Acquired proteasome-inhibitor (PI) resistance is a major obstacle in the treatment of multiple myeloma (MM). We investigated whether the clinical XPO1-inhibitor selinexor, when combined with bortezomib or carfilzomib, could overcome acquired resistance in MM. PI-resistant myeloma cell lines both in vitro and in vivo and refractory myeloma patient biopsies were treated with selinexor/bortezomib or carfilzomib and assayed for apoptosis. Mechanistic studies included NFκB pathway protein expression assays, immunofluorescence microscopy, ImageStream flow-cytometry, and proximity-ligation assays. IκBα knockdown and NFκB activity were measured in selinexor/bortezomib-treated MM cells. We found that selinexor restored sensitivity of PI-resistant MM to bortezomib and carfilzomib. Selinexor/bortezomib treatment inhibited PI-resistant MM tumor growth and increased survival in mice. Myeloma cells from PI-refractory MM patients were sensitized by selinexor to bortezomib and carfilzomib without affecting non-myeloma cells. Immunofluorescence microscopy, Western blot, and ImageStream analyses of MM cells showed increases in total and nuclear IκBα by selinexor/bortezomib. Proximity ligation found increased IκBα-NFκB complexes in treated MM cells. IκBα knockdown abrogated selinexor/bortezomib-induced cytotoxicity in MM cells. Selinexor/bortezomib treatment decreased NFκB transcriptional activity. Selinexor, when used with bortezomib or carfilzomib, has the potential to overcome PI drug resistance in MM. Sensitization may be due to inactivation of the NFκB pathway by IκBα.


INTRODUCTION
Cancer cells utilize the process of nuclearcytoplasmic transport through the nuclear pore complex to effectively evade anti-cancer mechanisms [1][2][3][4][5]. We have shown that knockdown of exportin 1 (XPO1/CRM1) protein by siRNA or with an XPO1 inhibitor will sensitize drug-resistant myeloma cells to the topoisomerase II (TOP2) inhibitor doxorubicin [3,5]. In addition, we found that XPO1 inhibitors are able to prevent nuclear export and promote nuclear accumulation of the tumor suppressor protein p53 [3,5]. XPO1 inhibitors, when used in combination with the proteasome inhibitors (PI) bortezomib and carfilzomib, were found to synergistically kill multiple myeloma (MM) cells and when co-cultured with bone marrow Research Paper stromal cells [5,6]. Our studies have shown that MM patient bone marrow mononuclear cells, when cotreated with an XPO1 inhibitor and PIs, synergistically induced apoptosis in MM cell populations but not in non-myeloma bone marrow cells, indicating that XPO1 inhibition may specifically inhibit cancer cells in MM patients [4,5]. These studies were the first to report cancer cell-specific apoptosis by combinations of XPO1 with PI. However, acquired drug-resistance results in cell lines, in vitro and in vivo, and ex vivo in PI-refractory patients have not been investigated in MM.
What has not been addressed in previous studies is whether XPO1 inhibitors are effective in overcoming acquired drug-resistant MM phenotypes, which develop in patients during treatment with PIs. In patients with MM, drug resistance is the primary limitation to successful treatment. Myeloma is still considered incurable despite significant advances afforded by immunomodulatory drugs (thalidomide, lenalidomide, pomalidomide), PIs (bortezomib, carfilzomib, ixazomib), antibodies targeting SLAMF7 protein (elotuzumab) and CD38 (daratumumab), histone deacetylase inhibitors (panobinostat), and high-dose melphalan with autologous stem cell rescue.
In the present study, we show that XPO1 inhibition sensitized acquired PI-resistant MM cells to bortezomib and carfilzomib in both in vitro and in vivo models and ex vivo in PI-refractory patient CD138+/light chain+ MM cells, thus showing that this combination may provide a means to overcoming acquired drug resistance in MM.

XPO1 inhibition sensitizes PI-resistant MM cell lines to bortezomib and carfilzomib
Apoptosis results (flow cytometry using activated caspase 3) from human PI-resistant and parental MM cells after 20-hour concurrent treatment with selinexor (300 nM) or KOS-2464 (10 nM) ± bortezomib (10 nM) or carfilzomib (20 nM) are shown in Figure 1. Both U266 and 8226 parental cell lines were highly sensitive to single-drug treatment with bortezomib or carfilzomib at log-phase growth densities (5 x 10 5 cells/mL). PIresistant U266PSR and 8226B25 MM cell lines [16,17] were resistant to single-agent bortezomib (up to 10-fold) or carfilzomib (up to 9-fold) when compared to parental cells ( Figure 1). When the XPO1 inhibitor selinexor was added, both U266PSR and 8226B25 PI-resistant cells were highly sensitized to bortezomib (P = 0.00055 and P = 0.0054, respectively) or carfilzomib (P = 0.0017 and P = 0.0033, respectively) treatment compared with singleagent treatment ( Figure 1). Equivalent results were found when PIs were used with the XPO1 inhibitor KOS-2464 [18] (Figure 1).

Ex vivo treatment of newly diagnosed, relapsed, and PI-refractory patient MM cells with selinexor and KOS-2464 sensitizes cells to bortezomib and carfilzomib
Using flow cytometry, we gated on the CD138/ light-chain immunoglobulin double-positive myeloma cell population in patient bone marrow aspirates. Apoptosis, as measured by activated caspase 3 expression showed that newly diagnosed (n=8), relapsed (n=5), and bortezomib (n=8)/carfilzomib (n=6) refractory MM patient samples were sensitized by selinexor and KOS-2464 to both bortezomib (P = 0.043 to 0.002) and carfilzomib (P = 0.044 to 0.001) ( Figure 3A, 3C, and 3E). When gating on the CD138/light-chain double-negative non-myeloma cells, we found that they were not sensitized to apoptosis by XPO1 inhibitors ( Figure 3B, 3D, and 3F). These data indicate that myeloma cells were targeted by the XPO1 inhibitor/PI drug combination and that non-myeloma cells were relatively unaffected.

Immunofluorescence microscopy of IκBα in patient MM treated with selinexor
Bone marrow aspirates from newly diagnosed, relapsed, and refractory MM patients were treated ex vivo with selinexor ( Figure 6). Kappa/lambda light chain antigen-positive MM cells had increased IκBα protein staining, especially in the cell nuclei with ex vivo selinexor treatment compared with untreated controls from the same patient.

ImageStream flow cytometry shows an increase in nuclear IκBα with drug treatment
8226B25 human MM cells treated for 20 hours with bortezomib, selinexor, or selinexor/bortezomib were    (Figure 7). A histogram of nuclear and cytoplasmic IκBα showed that the percentage of total cellular IκBα in untreated cells was 29.7% nuclear and 70.4% cytoplasmic. Bortezomib treatment increased nuclear IκBα to 43.4% (P = 0.006) and selinexor treatment to 76.6% (P = 0.000067), with the largest nuclear shift found in the selinexor/bortezomib combination (81.3%) (P = 0.000057). Five thousand cells per treatment group were analyzed ( Figure 7A) (n = 3). Images of representative cells in real time show increased IκBα protein in the cell nuclei with selinexor/bortezomib treatment ( Figure 7B).

Bortezomib-selinexor synergy may be linked to increased IκBα expression and subsequent down-regulation of NFκB transcriptional activity
SiRNA knockdown of IκBα in IM-9 and 8226 MM cells produced a 9.5-fold (P = 0.023) and 25.4-fold (P = 0.0062) increase in selinexor IC 50 values, making these cells less sensitive to selinexor ( Figure 8A). IκBα knockdown was > 60% at 24 hours following transfection. IκBα knockdown also highly reduced apoptosis (activated caspase 3) in selinexor/bortezomib-treated cells (P = 0.0086) compared with control siRNA ( Figure 8B). MM.1S cells treated with selinexor/bortezomib combined and as single agents were assayed for transcriptional activity. Single-agent selinexor or bortezomib treatment lowered NFκB transcriptional activity 2-and 5-fold, respectively (P = 0.0000023 and 0.00024); however, when selinexor and bortezomib were used together, transcriptional activity was reduced 12-fold compared with TNFα controls (P = 0.000019) and 3-fold below the baseline (P = 0.00013) (no TNFα) activity, as measured by chemiluminescent transcription assay ( Figure 8C). In addition, Western blots to measure NFκB transcriptional activity showed a decrease in the anti-apoptotic proteins IAP-1 (84%) and IAP-2 (72.8%) and cell cycle Figure 6: Immunofluorescence microscopy of IκBα in patient MM treated with selinexor. Bone marrow aspirates from newly diagnosed, relapsed, and refractory MM patients were treated ex vivo with 100-300 nM selinexor for 20 hours. MM cells were identified by kappa/lambda light chain antigen staining (green), nuclei identified by DAPI (blue) staining, and IκBα protein (red). All patient MM samples showed an increase in IκBα protein, especially in the cell nuclei with ex vivo selinexor treatment compared with untreated controls from the same patient. www.impactjournals.com/oncotarget (proliferation) proteins c-myc (62%) and cyclin D2 (42%) in selinexor/bortezomib-treated cells compared with untreated controls (inset in Figure 8C). Protein loading was confirmed by GAPDH expression.

DISCUSSION
Our findings show that XPO1 inhibitors, when used in combination with PIs, are highly effective against acquired PI-resistant human MM in in vitro assays with PI-resistant cell lines, in vivo in mice challenged with PI-resistant MM tumors, and ex vivo in patient MM cells that are refractory to PIs. These data strongly suggest that selinexor, when combined with PIs, may be an effective therapy for PI-resistant MM.
Previous studies have shown that IκBα is a tumor suppressor that dimerizes with NFκB and prevents its transcriptional activity, thus decreasing NFκB-driven proliferation and increased apoptosis in MM cells [20,23]. Proteasome inhibitors have been shown to inhibit NFκB activity through stabilization of IκBα [21]. In addition XPO1 inhibition by leptomycin B has been shown to sequester NFκB-IκBα complexes in the nucleus. These complexes are unable to bind DNA in the nucleus and activate NFκB transcriptional activity [22]. We found that the combination treatment of an XPO1 inhibitor (selinexor) and a PI (bortezomib) synergistically increased IκBα more than single-agent selinexor or bortezomib treatment in both parental and PI-resistant MM cell lines and in patient myeloma cells. This observed synergistic increase in IκBα expression by selinexor and bortezomib resulted in a strong anti-tumor effect. Proximity ligation assays showed that NFκB-IκBα complexes were significantly increased in cells treated with the selinexor and bortezomib combination, further indicating that NFκB is inactivated by PI and XPO1 inhibitor treatment. Therefore, inhibition of NFκB by upregulating IκBα protein and subsequent creation of NFκB-IκBα complexes may be, at least in part, the mechanism behind proteasome inhibition and XPO1 inhibitor anti-tumor synergy. In evidence of this mechanism, we found that siRNA knockdown of IκBα significantly decreased the effect of selinexor, as shown by 10-to 25-fold increase in IC 50 values, and knockdown significantly reduced selinexor/ bortezomib-induced apoptosis. In addition selinexor/ bortezomib treatment was shown to abrogate NFκB transcriptional activity.
In a companion report in this issue, Kashyap et al examined the combination of selinexor with PIs in sarcoma cell lines. They report that the combination treatment of C. Chemiluminescent transcription factor assay. MM.1S cells were pretreated with 1 μM selinexor ± 100 nM BTZ for 2 hours and then exposed to 20 ng/mL of TNFα for 4 hours in serum-free media. TNFα exposure induced NFκB transcriptional activity 6-fold. Single-agent selinexor (P = 0.000023) and BTZ (P = 0.00024) lowered NFκB transcriptional activity, and the combination of selinexor and BTZ further reduced the activity to 3-fold below baseline (P = 0.00013) (no TNFα). (Inset) Western blots of protein from 8226 selinexor/bortezomibtreated cells showed a decrease in NFκB-mediated IAP-1 (84%), IAP-2 (72.8%), c-Myc (62%), and cyclin D2 (42%) protein expression compared with untreated controls. Protein loading was confirmed by GAPDH expression. selinexor with bortezomib sensitizes the sarcoma cells to the cytotoxic effects of PIs. In addition, this report highlights the importance of the NFκB signaling pathway in cancer, especially the role of the inhibition of NFκB or IκBα by selinexor/bortezomib treatment.
Selinexor, an orally active selective inhibitor of XPO1-mediated nuclear export (SINE), is currently undergoing phase I/II studies in a variety of indications, including a combination with carfilzomib, in both relapsed and refractory MM patients (NCT02199665).
The results presented in this study support combinatorial clinical trials in relapsed and refractory MM that utilize PI therapies.

MATERIALS AND METHODS
For human sample acquisition, written informed consent approved by an Institutional Review Board was obtained from all patients, in accordance with the Declaration of Helsinki. Patient samples were deidentified and obtained through the Institutional Review Board-approved Total Cancer Care® protocol at the Moffitt Cancer Center.

Cell lines
Human MM cell lines RPMI 8226 (8226), IM-9, MM.1S, and U266 were obtained from the American Type Culture Collection (ATCC; Manassas, VA). To establish resistance of human MM cells to bortezomib, U266 and 8226 cells were continuously cultured in gradually increased concentrations of bortezomib. To produce PIresistant U266 cells (U266PSR), bortezomib dosing started at 0.5 nM, increasing in step-wise increments of 0.2 nM to 20 nM [16,19]. To produce PI-resistant 8226 (8226B25) cells, bortezomib dosing started at 1.0 nM, increasing in step-wise increments of 2.5 nM to 25 nM. The U266PSR cell line expressed a modest increase in Mcl-1, resulting in enhanced cell survival by inhibiting apoptosis and markedly lower expression of the apoptosispromoting factor Bim [17]. Both of the bortezomib resistant cell lines, U226PSR and 8226B25, are also highly resistant to carfilzomib (see Figure 1). U266PSR and 8226B25 cell lines were authenticated by the Moffitt Cancer Center Molecular Genomics Core Facility Cell lines using short tandem repeat (STR) DNA typing according to ATCC's "Authentication of Human Cell Lines: Standardization of STR Profiling (2012)." Results were compared with STR databases from ATCC and DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen, GmbH, Braunschweig, Germany) to establish percent identity. Cell lines were considered authenticated when the number of shared alleles across the eight core loci was ≥80% (as described by ATCC).

Drug-resistant cell lines treated with XPO1 inhibitors and bortezomib or carfilzomib
Parental and drug-resistant human 8226, 8226B25, U266, and U266PSR MM cells were grown at lowdensity (log growth phase) conditions (3-4 ×10 5 cells/mL) and cultured for 20 hours with either 300 nM selinexor (Karyopharm Therapeutics) or 10 nM KOS-2464 (Bristol-Myers Squibb) with and without 10 nM bortezomib (LC Labs), or 20 nM carfilzomib (SelleckChem). Optimal drug concentrations were determined by titration experiments for bortezomib and carfilzomib. Cells were fixed and permeabilized, and apoptosis was measured using anti-activated caspase 3/Alexa Fluor 488 (Cell Signaling Technology) staining in accordance with the manufacturer's standard protocol. Percent apoptosis was assayed by flow cytometry on a LSRII (Becton-Dickinson) bench-top analyzer. Data analysis was performed using Flowjo version 9.4 software (Tree Star, Inc) [24].

NOD/SCID-γ mouse studies with selinexor ± bortezomib
All mouse studies were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC), Research Integrity & Compliance -Research & Innovation at the University of South Florida. Bortezomib-resistant U266PSR human myeloma cells (10 6 ) were injected subcutaneously into flanks of female NOD/SCID-γ mice, and tumors were allowed to grow for 14 days before the start of treatment [19]. U266PSR human myeloma tumors were treated twice weekly by intraperitoneal injection with bortezomib (0.5 mg/kg) or twice weekly by oral gavage with selinexor (10 mg/kg) or in combination where selinexor treatment was followed 2 to 3 hours later by bortezomib injection. Five mice were used per experimental group. Tumors were measured by calipers, and tumor volumes (mm 3 ) were calculated by the formula (length x width 2 )/2. Mice were euthanized when a tumor volume exceeded 2000 mm 3 or if the mouse lost >15% of its body weight. Drug toxicity was also assayed by mouse weights with a decrease of ≥10% considered an indication of toxicity by the drug regimen.

Bone marrow aspirate processing and apoptosis assay of patient myeloma cells
As previously described, bone marrow aspirates were collected from newly diagnosed (n=8), relapsed (n=5), and PI-refractory bortezomib (n=8)/carfilzomib (n=6) patients [5]. PI refractory is defined as patients who progress during treatment or within 60 days after stopping treatment with either bortezomib or carfilzomib. Bone marrow aspirates (20 mL) from patients were isolated by Ficoll-Paque Plus (Amersham Biosciences) [5]. Cytospin slides were used to determine the percent plasma cell population by the microscopic morphology of toluidinestained cells [5].
Isolated bone marrow mononuclear cells from the Ficoll-Paque fraction described above were also incubated at 4 × 10 6 /mL in 200 μL RPMI (Fisher) containing 10% FBS in 96-well plates, treated with either 300 nM selinexor or KOS-2464 with and without 10 nM bortezomib or 20 nM carfilzomib and incubated for 20 hours in a 5% CO 2 humidified incubator. The following day, the cells were fixed and assayed for caspase activation-mediated apoptosis. Bone marrow mononuclear cells were fluorescently labeled with antibodies against activated caspase 3 (ASP175-Alexa 488), CD138 (M115-Alexa 647, BD Bioscience), and light-chain kappa (G20-193 V450, BD Bioscience) or lambda (JDC-12 V450, BD Bioscience). A BD Bioscience LSRII flow cytometer was used to gate MM cell populations, defined as cells that are both CD138 positive and light-chain positive. Nonmyeloma patient bone marrow cells were defined as cells that were both CD138 and light-chain negative. Percent of apoptotic cells from each sample was assayed by activated caspase 3. Data analysis was performed using Flowjo version 9.4 software (Tree Star, Inc).

Proximity ligation assay
Log-phase MM cells, both parental and PI-resistant, were placed at cell densities of 4 × 10 6 cells/mL and treated with 300 nM selinexor for 4-6 hours. Cells were washed with PBS, and cytospins were made with 10 5 cells/slide and fixed with 4% paraformaldehyde. Cells were incubated with primary antibodies to IκBα (E130/ Abcam) and NFκB (L8F6/Cell Signaling). Incubation with the primary antibodies was followed by secondary antibodies conjugated with oligonucleotides provided in the Duolink kit (Olink Bioscience, Uppsala, Sweden) [25]. A red fluorescent signal was generated only when IκBα and NFκB, were in close proximity (<40 nm). 4', 6-Diamidino-2-phenylindole (DAPI) was used to stain the nuclei. Samples were observed with a Leica TCS SP5 AOBS laser scanning confocal microscope through a 63X/1.4NA Plan Apochromat oil immersion objective lens (Leica Microsystems CMS GmbH, Germany). We applied 405 diode, Argon 488, and HeNe 647 laser lines to excite the samples, with tunable emissions used to minimize crosstalk between fluorochromes. Z stack (0.5-μmthick slices) images for each sample were captured with photomultiplier detectors, and maximum projections were prepared with the LAS AF software version 2.6 (Leica Microsystems).
Maximum projection images were analyzed using the Definiens® Developer v2.0 (Definiens AG, Munich, Germany) software suite. First, nuclei were segmented with an auto-threshold segmentation on the DAPI stain. Next, cytoplasms were segmented on the cellular autofluorescence in the green channel by using both size and intensity thresholds. Finally, the total number of foci per nucleus and cytoplasm were analyzed for number and area. This experiment was repeated 3 times. Western blots were made of the treated cells at 4-6 hours for XPO1 expression (according to the methods described below).

Immunofluorescence microscopy and Western blot of IκBα in selinexor/bortezomib-treated drug-resistant MM cells and selinexor-treated patient MM cells
Drug-resistant 8226B25 and U266PSR human MM cells were incubated with selinexor (300 nM), bortezomib (10 nM), or their combination for 20 hours to determine their combined effects on IκBα expression by Western blot and immunofluorescence microscopy.
Immunofluorescence microscopy of cytospins slides was performed as previously described [5] using both PI-resistant and parental cell lines and on patient MM samples (n = 6) after treatment with 300 nM selinexor, 10 nM bortezomib, or in combination. Anti-IκBα rabbit monoclonal antibody (Abcam) was used at 1:100 followed by incubation with anti-rabbit Alexa Fluor 594 (Invitrogen) secondary antibody (1:500). In addition patient bone marrow aspirate cytospins were stained with anti-kappa (Millipore AP505F) or anti-lambda (Millipore AP506F) light-chain FITC-conjugated antibodies to identify MM cells. Slides were washed four times in PBS, air dried, and covered with cover glass and Vectashield mounting media containing antifade/DAPI (Vector Laboratories Inc) to stain nuclei. Images were captured with a high-resolution CCD camera mounted on a Zeiss Automated Upright Fluorescent Microscope.

ImageStream flow cytometry
The 8226B25 cell line was treated with 300 nM selinexor, 10 nM bortezomib, or selinexor/bortezomib and incubated for 20 hours. Cells were washed once in cold PBS and fixed for 20 minutes on ice in cytofix/cytoperm (BD Bioscience) solution. Cells were then pelleted and resuspended in perm/wash (BD Bioscience) solution and stored at 4°C. Treated cells (5 x10 5 ) and controls were incubated for 1 hour at room temperature with anti IκBα (E130) (Abcam ab32518) diluted 1:50 in perm/wash (BD Bioscience) solution. Cells were then washed in perm/ wash solution and incubated with anti-rabbit Alexa-488 diluted 1:250 and incubated for 45 minutes in the dark at room temperature. DAPI (Sigma D21490) (5 nM) was added to stain the nucleus immediately before analyses. Flow cytometry was performed on an ImageStreamX MKII high-speed imaging flow cytometer (Amnis Corporation) to analyze intracellular fluorescence. Bright field and fluorescent images were collected at a ×40 magnification. We assayed 10,000 gated cell singlets from each sample. IDEAS Analysis Software (Amnis Corporation) was used to determine nuclear location of IκBα in the treated and untreated cells. The Similarity Feature was used to determine nuclear location. Similarity is the log-transformed Pearson correlation coefficient and is a measure of the degree to which two images are linearly correlated within a designated region. This analysis uses the pixel data in the region that is specified as "the nucleus" (DAPI) and compares the similarity when other fluorochromes (IκBα) occupy the same pixel space. A high positive similarity value indicates that the probes are in the same location, whereas a high negative value indicates they are in different locations. A value near zero indicates there is an equal amount of probe in both the nucleus and the cytoplasm. Values were obtained for percent IκBα in the nucleus and cytoplasm in MM cells in all drug treatment groups.
IκBα siRNA knockdown and its effect on selinexor IC 50 , apoptosis, and NFκB transcriptional activity IM-9 and 8226 cell lines were transfected with 40 nM of IκBα siRNA or 40 nM BLOCK-iT control using Neon Transfection System (#MPK5000, Life Technologies) following manufacturer's instruction. The transfection was performed using antibiotic-free RPMI 1640 media. Twenty-four hours after transfection, the transfected and nontransfected IM-9 and 8226 cells were treated with selinexor starting at 30 μM and diluted 1:3 to a final concentration of 4.5 nM in triplicate for 72 hours. The cell viability was analyzed using CellTiter-Fluor cell viability assay (#G6080, Promega) and the half-maximal inhibitory concentration (IC 50 ) of selinexor for each condition was calculated using XLfit. IκBα knockdown was 60% compared with transfection control. We transfected 8226, 8226B25, and H929 human MM cells with control siRNA or IκBα siRNA followed by treatment 48 hours after transfection with selinexor 300 nM, bortezomib 10 nM, and the combination. After 20hour incubation with drugs, the cells were assayed by flow cytometry for activated caspase 3 apoptosis as above (n=2).
NFκB transcriptional activity was assayed in MM.1S cells pretreated with 1 μM selinexor and/or 100 nM bortezomib for 2 hours and then exposed to 20 ng/mL TNFα for 4 hours in serum-free media. TNFα exposure induced NFκB transcriptional activity 6-fold as measured by a chemiluminescent transcription factor assay kit (Thermo Scientific, catalog no. 89859). TNFα-activated NFκB p65 transcription factor binds to a biotinylated consensus sequence plated on a 96-well plate. The plate is developed using a primary antibody to NFκB p65 followed by a horseradish peroxidase-conjugated secondary antibody and developed with a chemoluminescent substrate.

Statistical analysis
Analysis of variance was used to compare mouse data sets. All other statistical comparisons were made using the t test.

ACKNOWLEDGMENTS
We thank Rasa Hamilton (Moffitt Cancer Center) for editorial assistance. Our study also received valuable assistance from the Flow Cytometry Core, Microscopy Core, and Tissue Core Facilities at the H. Lee Moffitt Cancer Center & Research Institute.

CONFLICTS OF INTEREST
The author and co-authors from the Moffitt Cancer Center and Virginia Commonwealth University have no conflicts of interest to disclose. Co-authors from Karyopharm Therapeutics (Trinayan Kashyap, Yosef Landesmann) provided the XPO1 inhibitor selinexor.

The H. Lee Moffitt Cancer Center & Research
Institute is supported in part by NCI Cancer Center Support Grant P30 CA076292. Additional support was provided by the Bankhead-Coley Team Science Project 2BT03-43424 grant, NIH CA194051 grant, the V Foundation and the Moffitt Cancer Center Foundation. Work done at Virginia Commonwealth University was supported by NIH grants P50-CA142509 and CA167708, CA205607, and the Leukemia and Lymphoma Society of America 6472-15.