The XPO1 Inhibitor KPT-8602 Synergizes with Dexamethasone in Acute Lymphoblastic Leukemia

Acute lymphoblastic leukemia (ALL) is the most common pediatric cancer and entails B-cell ALL (85% of cases) and T-cell ALL (15%; refs. 1, 2). ALL is caused by a stepwise accumulation of mutations in lymphoid precursor cells, leading to malignant B- or T-cell transformation associated with uncontrolled proliferation, survival, and a differentiation block (3, 4). Intense chemotherapy regimens remain the first-line treatment for ALL and long-term survival is now achieved in almost 90% of pediatric and up to 70% of adult cases (5). Despite the increased overall survival, relapse and failure to achieve clinical remission remain major challenges. Further intensification of current treatment regimens is not desirable due to the high toxicity, which leads to severe short-term and long-term side effects, including life-threatening infections, osteonecrosis, neurobehavioral side effects, and growth defects (5, 6).

The glucocorticoids dexamethasone and prednisone are widely used in chemotherapy regimens for ALL. Whereas prednisone was commonly used in the past, dexamethasone is now more often used as it proved to be more potent against central nervous system–infiltrating leukemia (7). Response to dexamethasone during induction phase is important to evaluate further treatment options and is commonly used to predict patient outcome (7). Glucocorticoids act by binding to the cytoplasmic glucocorticoid receptor NR3C1, which then undergoes a conformational change and dissociates from some of its chaperone proteins (8). The activated receptor translocates to the nucleus with the help of chaperone and transporter proteins, where it binds to glucocorticoid receptor elements (GRE) in the genome, mostly in a homodimerized form (9). This results in the activation or repression of many target genes including NR3C1 itself, BCL2, KLF13, GILZ, PER1, and NFKBIA (10–12). Although multiple glucocorticoid-responsive genes have been reported over the past years, there is no consensus about the exact signal transduction pathways that lead to glucocorticoid-induced apoptosis in ALL cells (11). Activated NR3C1 can also remain monomeric and bind to other transcription factors such as activating protein-1 (AP-1) or nuclear factor-κB (NFκB), thereby repressing their activity (8).

Exportin 1 (XPO1, also known as CRM1), one of the nuclear export factors for rRNA and nuclear export signal (NES) bearing proteins, has recently emerged as a novel therapeutic target in cancer (13). Selinexor (KPT-330) is the first generation of selective inhibitors of nuclear export (SINE) compounds, which inhibit the XPO1 export function by a slowly reversible, covalent modification of the active XPO1 site at Cys528 (14). Selinexor has potent anticancer activity against solid tumors and leukemias and is currently being evaluated in various clinical trials. Selinexor was recently approved for the treatment of adult patients with relapsed or refractory multiple myeloma (15–19). KPT-8602 (Eltanexor) is a second-generation XPO1 inhibitor with increased reversibility in XPO1 binding and reduced brain penetration, leading to improved drug tolerance compared with KPT-330 (20). Twice-a-week dosage of KPT-330 resulted in dose-limiting gastrointestinal and constitutional toxicities, while a phase I/II study reported much lower toxicities when dosing KPT-8602 even five times a week. This allowed for better dosage and less therapy discontinuation (21). KPT-8602 monotherapy showed promising activity in ALL and acute myeloid leukemia (AML) cell lines and in patient-derived ALL xenograft models and is currently being evaluated in clinical trials (NCT02649790; refs. 21–23). Recently, KPT-8602 showed its potential as a combination therapy, by reducing the leukemia burden in primary AML and DLBCL patient cells when combined with the BCL2 inhibitor venetoclax (24).

In this study, we aimed to determine whether KPT-8602 synergizes with chemotherapy drugs that are currently used for ALL treatment. We evaluated the synergistic combinations with the aim to improve dexamethasone response in patients, to reduce the dose of this toxic agent in chemotherapy regimens.

DND41, SUP-T1, 697, and BV-173 cell lines (www.DSMZ.de) were cultured in RPMI1640 (Invitrogen) supplemented with 20% FBS (Invitrogen) in 5% CO₂ at 37°C. Cell line authentication was performed by short-tandem repeat analysis, Mycoplasma testing was performed by the MycoAlert Detection Kit and assay control set (Westburg, catalog No. LO LT07–518 and LO LT07–118). Cells were used for experiments within 3 months after thawing, after which a new batch was used. For drug treatment, cells were seeded at 3.5 × 10⁵ cells/mL in 96-well plates (100 μL/well). KPT-8602 (Selleckchem catalog No. S8397), dexamethasone (Selleckchem catalog No. S1322), doxorubicin (Selleckchem catalog No. S1208), vincristine (Selleckchem catalog No. S1241) or DMSO (Sigma Aldrich) were dispensed at the desired concentrations using a Tecan D300e Digital Dispenser (Tecan). The DMSO concentration was normalized according to the highest DMSO volume used. Experiments were performed as biological triplicates. Cell viability was measured after 48 hours of drug treatment with the ATPlite Luminescence Assay System Kit (PerkinElmer) on a VICTOR Multilabel Plate Reader (PerkinElmer). Quantitation of synergy was performed via the Chou–Talalay method, using CompuSyn software (25). Additive effect is defined as a combination index (CI) of 1, synergism as CI <1, strong synergy as CI <0.2, and antagonism as CI >1.

Apoptosis was measured after 48 hours of drug treatment with the FITC Annexin V Apoptosis Detection Kit with PI (BioLegend). Cells were analyzed on a MACSQuant Vyb (Miltenyi Biotec). Data analysis was performed using FlowJo software (BD Biosciences). For Histone flow cytometry, cells were stained with Fixable Viability Dye eFluor 450 (eBioscience, catalog No. 65–0863–18) prior to fixation (eBioscience, catalog No. 00–5523–00). After 1-hour incubation in the dark, cells were washed, and resuspended in permeabilization buffer with primary histone 3 antibody (Cell Signaling Technology; catalog No. 14269). Overnight incubation was performed at 4°C. Cells were washed and secondary staining (Alexa Fluor 647, catalog No. ab150115) was performed for 2 hours at room temperature. Cells were washed and resuspended in PBS, followed by measurement on the MACSQuant Vyb.

All in vivo experiments were approved and supervised by the ethical committee of the University of KU Leuven and conducted according to EU legislation (Directive 2010/63/EU). Experiments on human samples were approved and supervised by the UZ Leuven ethical committee and informed written consent was obtained from all patients or their parents, according to the Declaration of Helsinki. A total of 1 × 10⁶ Ficoll isolated human leukemic mononuclear cells, obtained at diagnostics from the blood of patients, were injected in the tail vein of 6- to 12-week-old NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG) mice (The Jackson Laboratory). Mutational analysis of patient samples used in this study is provided in Supplementary Table S1. Expansion of the human leukemic cells was monitored in the mice by peripheral blood withdrawal and staining with human anti-CD45 antibody (hCD45, eBioscience, catalog No. 17–9459–42). Mice were closely monitored and sacrificed once hCD45 levels reached 50% in the blood or if the humane endpoint was reached.

To prepare patient-derived xenograft (PDX) samples for bioluminescent imaging, freshly harvested PDX cells from the spleen were transduced ex vivo with a pCH-SFFV-eGFP-P2A-fLuc lentiviral vector (kindly provided by Rik Gijsbers, KU Leuven, Leuven, Belgium). After overnight incubation, the transduced cells were washed with PBS, and cells were reinjected via tail vein injection (1 × 10⁶ cells per NSG mouse). When leukemia development was observed (hCD45 > 50%), the mice were sacrificed, human PDX cells were harvested from the spleen, and GFP-positive cells were sorted using the S3 Sorter (Bio-Rad). A total of 600,000 sorted GFP-positive cells were retransplanted into fresh NSG mice. After multiple engraftment/sorting rounds, we established PDX samples with >95% fLuc/GFP–positive human cells. These samples were injected into a larger cohort of 32 NSG mice (1 × 10⁶ leukemic cells per mouse) for in vivo treatment studies.

NSG mice injected with PDX cells were treated after engraftment was established (luminescent flux >10⁶ as assessed by bioluminescence imaging (BLI), or the presence of >1% hCD45⁺ cells in the peripheral blood). Six to 12-week-old age-matched female NSG mice were randomized into groups of eight to equally distribute the leukemic burden (as assessed by BLI or hCD45 staining) and weight over the different treatment groups. KPT-8602 was dissolved in DMSO, followed by formulation in 0.5% methylcellulose (Sigma Aldrich, catalog No. M7027–100G) plus 1% Tween-80 (Sigma Aldrich, catalog No. P1754–500ML). KPT-8602 was administered by oral gavage (200 μL per mouse) at a concentration of 5 mg/kg in cycles of 5 days on/2 days off during the 2 weeks of treatment. The mice in the placebo and dexamethasone only groups were treated with vehicle as second treatment. Dexamethasone (Selleckchem, catalog No. S4028) was dissolved in the drinking water at a dose of 4 mg/L following a cycle of 3 days on/2 days off during the 2 weeks of treatment. Previous research has shown that discontinuous and continuous treatment reached equal efficacy (26). Endpoint for leukemic-free survival was reached when mice had 50% of hCD45⁺ cells in the blood. The investigators were not blinded to the treatment groups. Cells from blood, spleen, and bone marrow were harvested and stained with hCD45 antibody (eBioscience, catalog No. 17–9459–42) for flow cytometry.

NSG mice were imaged with the IVIS Spectrum In Vivo Imaging System (PerkinElmer). Mice were sedated with 2% isoflurane (Iso-Vet 1000 mg/g, Dechra), followed by subcutaneous luciferin injection and subsequent imaging. Quantification and image processing were performed using the Living Imaging Software (PerkinElmer).

We and others have previously shown that the XPO1 inhibitor KPT-8602 inhibits the proliferation of ALL cell lines and reduces leukemic burden in in vivo PDX models for B-ALL and T-ALL (22, 27). These results with KPT-8602 monotherapy prompted us to investigate synergistic interactions between KPT-8602 and the cytotoxic agents dexamethasone, doxorubicin, and vincristine, currently used to treat ALL.

We treated B-ALL (697 and BV-173) and T-ALL (DND41 and SUP-T1) cell lines with various concentrations of KPT-8602, dexamethasone, doxorubicin, or vincristine, or a combination of KPT-8602 with one of the other drugs and measured cell proliferation (mutational profile of the cell lines is provided in Supplementary Table S2). Synergy scores were calculated using the Chou–Talalay method, where a combination index (CI) value below 1 indicates synergy, and very strong synergistic combinations have a CI value below 0.2 (25). Strong synergy was observed between KPT-8602 and dexamethasone in all tested cell lines (Fig. 1A). In contrast, combination of KPT-8602 with doxorubicin or vincristine led to moderate synergistic effects in three cell lines and antagonism in one cell line (Supplementary Fig. S1A–S1D).

Treating 697 cells with dexamethasone alone led to a 70% reduction of cell proliferation, which was further reduced by KPT-8602 addition (Fig. 1B). BV-173 proliferation was almost completely blocked with 200 nmol/L of dexamethasone. Addition of KPT-8602 led to a faster decrease and eventual block of proliferation at 20 nmol/L dexamethasone. Whereas dexamethasone as a single agent could only partially inhibit DND41 and SUP-T1 cells, adding KPT-8602 to the same dexamethasone concentrations led to nearly complete inhibition of cell proliferation (Fig. 1B).

We next investigated whether the synergistic effect of the dexamethasone and KPT-8602 combination was due to apoptosis. To this end, the different cell lines were treated with KPT-8602, dexamethasone, or a combination, using for each cell line the concentrations that led to the highest synergy (Supplementary Table S3). We found that three of four cell lines showed significantly more cell death (determined as the sum of AnnexinV⁺/PI⁻ and AnnexinV⁺/PI⁺ cells) after combination treatment compared with single treatments (Fig. 1C). Even upon 10-fold reduction of the dose of dexamethasone in the combination treatment, we achieved equal or higher levels of apoptosis compared with single dexamethasone treatment (Fig. 1C).

To validate the findings from ALL cell lines in a relevant preclinical setting, we treated B- and T-ALL PDX models, selected for their different mutational background, with dexamethasone, KPT-8602, or the combination of both drugs. When engraftment was established, a 2-week treatment with either vehicle, KPT-8602, dexamethasone, or the combination of both drugs was started. To determine whether this combination was synergistic, and to minimize toxicity of dexamethasone, we used relatively low concentrations of dexamethasone (4 mg/L) and KPT-8602 (5 mg/kg).

As a first model, we used a NOTCH1-mutant T-ALL sample (X10), which was engineered to express luciferase and could be followed by bioluminescent in vivo imaging. After 2 weeks of treatment, we observed a significant reduction in leukemic burden in the mice treated with the combination of KPT-8602 + dexamethasone (Fig. 2A and B). At the end of the treatment, four mice of each treatment group were sacrificed and analyzed (Fig. 2A). Because of a high dexamethasone response, there was no significant difference in spleen weight and leukemic infiltration in peripheral blood after dexamethasone versus combination treatment (Fig. 2C and D). However, the combination treatment significantly improved clearance of leukemia burden from the spleen and bone marrow compared with the single treatments (Fig. 2E and F). We stopped treatment for the remaining four mice of each group and followed the leukemia development by BLI. The leukemic evolution was significantly slower for the mice that received combination treatment compared with single-drug treatment or placebo (Fig. 2G). The median survival was almost twofold higher (median survival of 33 days) compared with KPT-8602 (17.5 days) or dexamethasone (15 days; Fig. 2H).

In a similar approach, we used a second T-ALL sample (XC65), which had a different mutational background (JAK3 and NOTCH1 mutation). We again observed the lowest leukemic infiltration in the assessed organs and on whole-body BLI after 2 weeks of combined dexamethasone + KPT-8602 treatment (Fig. 3A–F). In a third PDX experiment, we engrafted NSG mice with a TCF3-PBX1–positive B-ALL sample (XC56), and these mice underwent the same treatment scheme. After 2 weeks of treatment, spleen and bone marrow infiltration was significantly higher in the single dexamethasone or KPT-8602 groups compared with the mice that received combination treatment (Fig. 3G–I). Already, at the first measurement, we detected almost no leukemic cells in the peripheral blood of mice treated with a combination of KPT-8602 and dexamethasone, while in the mice treated with only KPT-8602, dexamethasone, or vehicle, the leukemic infiltration in the blood continued to increase significantly during treatment (Fig. 3J).

As we noticed different responses to dexamethasone, we checked the NR3C1 levels (Supplementary Fig. S2). PDX X10, which responded best, had the highest amount of NR3C1, while NR3C1 expression was the lowest in the poor dexamethasone responder XC65. Of note, XC65 has a JAK3 mutation, which was previously shown to cause resistance to dexamethasone (28, 29). XC56 had intermediate-to-high levels of NR3C1, showing that that NR3C1 expression levels are not the sole explanation for the observed dexamethasone response. Altogether, the combination of dexamethasone with KPT-8602 was synergistic to treat PDX samples with different genetic backgrounds in vivo, regardless of their response to dexamethasone.

Dexamethasone binds and activates the glucocorticoid receptor (NR3C1), which is a transcription factor that activates or suppresses a variety of target genes, including the NR3C1 gene itself (30). To determine whether KPT-8602 treatment directly influenced NR3C1 transcriptional activity, we studied the DNA binding of NR3C1 by chromatin immunoprecipitation sequencing (ChIP-seq) and the associated gene expression changes by transcriptome sequencing. We treated both 697 and SUP-T1 cells for 24 hours with vehicle and synergy concentrations of dexamethasone, KPT-8602, or dexamethasone + KPT-8602 (Supplementary Table S3) and performed ChIP-seq and RNA-seq on these cells. We used NR3C1 antibodies as well as H3K27ac and H3K4me3 antibodies to determine active regulatory enhancer or promoter regions via ChIP-seq. NR3C1 did not bind DNA in the absence of dexamethasone (Fig. 4A; Supplementary Fig. S3A). Upon dexamethasone treatment, we detected 362 NR3C1 peaks in 697 cells and 710 NR3C1 peaks in SUP-T1 cells, related to 266 and 408 genes, respectively (Supplementary Table S4). NR3C1-binding sites were mainly located in enhancer regions, defined as H3K27Ac⁺/H3K4me3⁻ regions (Fig. 4B; Supplementary Fig. S3B). Analysis by i-CisTarget and RSAT confirmed enrichment of the NR3C1-binding motif (GRE motif) in the NR3C1-bound regions (31–33). Gene-set enrichment analysis (GSEA) illustrated that dexamethasone treatment indeed led to higher expression of the NR3C1-bound target genes for both 697 and SUP-T1 cell lines (Fig. 4C; Supplementary Fig. S3C).

Analysis of all NR3C1 ChIP peaks throughout the genome indicated that NR3C1 binding in the combination treatment was maintained (SUP-T1) or even slightly increased (697) compared with dexamethasone alone (Fig. 4D). Moreover, addition of KPT-8602 to dexamethasone further increased the expression of directly bound NR3C1 target genes, including NFKBIA, STAG3, PER1, and TSC22D3 in the four cell lines (Fig. 4E and F; Supplementary Fig. S3D; refs. 34, 35). While KPT-8602 had no or only a minor effect on mRNA expression of these genes, dexamethasone and KPT-8602 typically increased the expression by >2-fold compared with dexamethasone alone, illustrating a clear synergy on the direct target genes (Fig. 4F; Supplementary Fig. S3D). In agreement with this, we could reduce the concentration of dexamethasone four- to 10-fold when combining it with KPT-8602 to achieve a similar level of induction of NR3C1 and NFKBIA genes compared with dexamethasone alone (Fig. 4G).

It was previously suggested that XPO1 inhibition could increase the protein levels and/or the nuclear retention of NR3C1 and IκBα (encoded by NFKBIA; refs. 36, 37). To determine whether this could be part of the synergy mechanism in ALL, we analyzed the presence of total NR3C1 and its active phosphorylated form (ser211) in the nucleus (38). As expected, dexamethasone treatment induced the presence and phosphorylation of NR3C1 in the nucleus. However, addition of KPT-8602 did not lead to a further nuclear increase in phosphorylated or total NR3C1 (Fig. 4H and I).

IκBα is known to be important for glucocorticoid-induced antiproliferative effects (39). In addition, several studies have shown that that nuclear retention of its protein IκBα is important for the selinexor mechanism and for the synergy mechanism between selinexor and dexamethasone in multiple myeloma (36, 37). Although we also see increased NFKBIA transcription in our ALL models (Fig. 4F; Supplementary Fig. S3D), we did not observe increased protein levels or increased nuclear retention of IκBα upon KPT-8602 treatment (Supplementary Fig. S4A–S4C; ref. 37). Moreover, GSEA did not show enrichment of NFκB target genes in the differentially expressed genes upon KPT-8602 treatment alone or in combination with dexamethasone, further suggesting that the NFκB pathway is not changed (Supplementary Fig. S4D). Finally, CRISPR/Cas9-mediated deletion of the GRE in the NFKBIA gene resulted in reduced sensitivity to dexamethasone, but did not alter the sensitivity to KPT-8602, indicating that IκBα is not important for KPT-8602 sensitivity in ALL (Supplementary Fig. S4E–S4G).

Together, our data show that the combination of dexamethasone with KPT-8602 results in a stronger activation of the NR3C1 transcriptional complex, which leads to a significantly higher induction of target gene expression than dexamethasone alone.

Combined ChIP-seq and RNA-seq data analysis revealed that only a small fraction of gene expression changes was mediated by direct NR3C1 binding, suggesting that other transcriptional complexes also contributed to dexamethasone response. Indeed, dexamethasone treatment in 697 cells led to a significant upregulation of 315 genes and a significant downregulation of 275 genes (P_adj < 0.05), of which only 7% were directly bound by NR3C1 (Supplementary Table S4). In SUP-T1, we found 403 significantly upregulated and 745 downregulated genes after dexamethasone treatment (5% bound by NR3C1; Supplementary Table S4). These results indicate that 24 hours of dexamethasone treatment has a broad effect on gene expression with only a fraction of these changes caused by direct NR3C1 binding to its target genes. Interestingly, despite the fact that KPT-8602 had only limited effects on gene expression by itself, the addition of KPT-8602 to dexamethasone led to a significant further up- or downregulation of differentially expressed genes (Fig. 5A; Supplementary Table S5).

When performing overrepresentation analysis (ORA-analysis) on the significantly downregulated genes in dexamethasone versus DMSO-treated cells, we noticed that DNA replication and cell cycle were among the pathways significantly down in both cell lines, indicating a strong effect of dexamethasone treatment on these pathways (Fig. 5B, dexamethasone vs. DMSO: x-axis). Addition of KPT-8602 to dexamethasone further downregulated the DNA replication and cell-cycle pathways compared to dexamethasone monotherapy in both cell lines (Fig. 5B, Combi vs. dexamethasone: y-axis). These gene expression data indicate that dexamethasone treatment strongly affects cell cycle and DNA replication and that dexamethasone effects were enhanced by addition of KPT-8602.

To identify the transcription factors that could be responsible for these transcriptional changes, we performed an in silico analysis of regulatory sequences using i-CisTarget (31, 32). This analysis revealed a strong enrichment of the E2F-binding motif in the genes that were more up- or downregulated after dexamethasone treatment. In addition, the E2F-binding motif was also enriched in the differentially expressed genes after dexamethasone+KPT-8602 treatment compared with dexamethasone (Fig. 5C). In agreement with this, known E2F target genes were significantly downregulated in dexamethasone-treated cells and were further downregulated in cells treated with dexamethasone + KPT-8602 (Fig. 5D). The E2F transcription factors are important drivers of the cell cycle, and increased activity of E2F members in ALL is mediated by different oncogenic events. The upstream E2F regulator CDKN2A is found deleted in the majority of ALL cases, while activation of the AKT pathway and TAL1 expression lead to upregulation of E2F1, E2F2, and E2F8 (3, 40). In further support for an important role of E2F transcriptional complexes in dexamethasone response, we observed a consistent downregulation of E2F1, E2F2, and E2F8 in the cells treated with dexamethasone and even stronger in cells treated with the combination treatment (Fig. 5E). Overall, this indicates an important role for the E2F transcription factors in the dexamethasone response and in the synergy mechanism of KPT-8602 and dexamethasone.

Finally, we used the NanoString nCounter platform to complement the RNA-seq data, now focusing on cancer-associated pathways and histone genes (our RNA-seq method was biased toward polyadenylated RNA, and thus not suitable for the detection of histone variants that are encoded by non-polyadenylated transcripts). Using this method, we looked at transcriptional changes of 770 genes from 13 cancer-associated canonical pathways. Global analysis of this dataset confirmed the RNA-seq data and showed a weak effect of KPT-8602 on gene expression, a stronger effect of dexamethasone treatment and a more pronounced up- or downregulation in the combination treatment (Fig. 5F). From the 770 genes analyzed, 18 genes were significantly upregulated and 42 genes were significantly downregulated by KPT8602 + dexamethasone compared with dexamethasone alone (FDR < 0.05). The E2F1 transcription factor itself and a large set of E2F transcription factor target genes associated with cell cycle and DNA repair were slightly downregulated by dexamethasone. Again, combination treatment with KPT-8602 + dexamethasone had stronger effect on these genes compared with dexamethasone alone, leading to a further downregulation of E2F1 and E2F transcription factor target genes (Fig. 5F). Remarkably, while KPT-8602 treatment by itself had very limited effect on gene expression, it significantly downregulated all histone 3 variants present in the nCounter gene panel and the combination treatment with dexamethasone further downregulated these histone 3 genes. We confirmed this on protein level by flow cytometry analysis with anti-histone 3 staining in three of four of the tested cell lines, with only BV-173 showing no significant response after treatment with the synergy concentrations (Fig. 5G). These data identify downregulation of histone 3 variants as another mechanism that contributes to inhibition of the cell cycle by combined dexamethasone and KPT-8602 treatment.

The use of glucocorticoids (dexamethasone or prednisone) is one of the pillars of successful ALL treatment. Most ALL cases respond very well to dexamethasone treatment, and a suboptimal response is a poor prognostic marker. However, dexamethasone treatment is associated with adverse side effects and resistance at relapse is still a common problem. In this study, we found that the XPO1 inhibitor KPT-8602, which is currently in clinical trials for other malignancies (21), enhances the effects of dexamethasone on both direct and indirect glucocorticoid receptor target genes, leading to more apoptosis and a further decrease in proliferation of ALL cells.

The interaction between KPT-8602 and dexamethasone was synergistic in ALL cell lines, which was not the case for combinations of KPT-8602 with doxorubicin or vincristine. Moreover, synergy was also confirmed in B-ALL and T-ALL xenografts after in vivo treatments. We did not observe more side effects (weight loss, appetite, general condition, behavior) in the mice treated with the combination compared with the single treatments, which is promising for further clinical testing. We show that it is possible to lower the dose of dexamethasone four to 10 times when combining it with a low dose of KPT-8602, and still achieve the same amount of apoptosis as with higher doses of dexamethasone alone. These findings could translate into improved responses in patients with suboptimal response to dexamethasone or alternatively could allow the reduction of dexamethasone doses to reduce side effects.

Synergy between dexamethasone and the first-generation XPO1 inhibitor KPT-330 (Selinexor) has recently been shown in multiple myeloma (41), sometimes also in combination with proteasome inhibitors (18, 41–43). KPT-330 treatment is associated with various side effects, but these are expected to be reduced when using KPT-8602, as this second generation SINE compound is much better tolerated (20, 23). We have indeed not observed major toxicities in our xenograft mouse models.

Mechanistically, dexamethasone is known to activate the glucocorticoid receptor (NR3C1), resulting in higher expression of proapoptotic proteins and several repressors of various signaling pathways such as the NFκB and the RAS–MAPK pathways (44–46). The exact mechanism by which XPO1 inhibitors cause selective inhibition of cancer cells remains incompletely understood. One proposed mechanism is the nuclear retention of NES-bearing proteins such as tumor suppressors, apoptosis inducers, and cell-cycle regulators (47). Besides its protein export function, XPO1 is important for the nuclear export of mRNA, rRNA, and U snRNA but this has not yet been linked to its anticancer mechanism (48–50). Recently, XPO1 was shown to accumulate at HOX cluster regions, by this recruiting the leukemogenic proteins Nup98-HoxA9, SET-Nup214, and mutant NPM, leading to HOX gene activation in human leukemia cells (51). XPO1 also has a role in the control of mitotic progression and chromosome segregation, making it a complex molecule affecting various pathways in the cell (52).

To increase our insight in the molecular mechanism that could explain the synergy between dexamethasone and KPT-8602, we integrated ChIP-seq and RNA-seq data, which revealed an increased transcriptional activity of NR3C1 in the presence of KPT-8602. Upon dexamethasone treatment, there was an upregulation of the direct NR3C1 target genes such as NFKBIA, STAG3, PER1, and TSC22D3 and these were further upregulated by addition of KPT-8602. This illustrates a strong synergistic mechanism because KPT-8602 itself did not affect expression of these genes. Our Western blot data indicate that NR3C1 is not accumulating in the nucleus upon XPO1 inhibition and that this is not likely to be a mechanism for the enhanced NR3C1 activity. In addition, calreticulin and not XPO1 was found to be the main NR3C1 nuclear export molecule (53–55). We hypothesize that enhanced nuclear localization of NR3C1 cofactors might facilitate the glucocorticoid receptor function, leading to the increased transcriptional activity we observed.

TSC22D3 plays a role in the anti-inflammatory response of glucocorticoids and is thus an important mediator for the antiproliferative effects of glucocorticoids (39, 44). We report here an increased expression after combining KPT-8602 and dexamethasone compared with dexamethasone single treatment, which is interesting for further studies. Another well-known molecule responsible for the anti-inflammatory function of dexamethasone is NFKBIA (IκBα), an inhibitor of NFκB, which gene expression is also further upregulated after dexamethasone+KPT-8602 treatment compared with dexamethasone only (46, 56). Although studies have documented an XPO1-mediated nuclear export of IκBα, we did not observe the previously reported nuclear IκBα localization after addition of KPT-8602 in our ALL models (37, 57). In agreement with this, we did not measure NFκB target gene inhibition after KPT-8602 treatment in ALL, which further reinforces our conclusion that in contrast to studies in multiple myeloma, IκBα nuclear retention is not part of the synergy mechanism in ALL.

Surprisingly, we found that many gene expression changes upon dexamethasone treatment converged on E2F transcription factor regulation. Many of the genes altered by dexamethasone and dexamethasone+KPT-8602 treatment contained E2F transcription factor motifs and we also observed strong and consistent downregulation of the E2F1, E2F2 and E2F8 transcription factors, from which E2F1 and E2F2 are known to positively regulate the cell cycle (58). Moreover, XPO1 inhibition is well known to result in nuclear retention of E2F4 and E2F7, two negative cell-cycle regulators of the E2F transcription factor family (59, 60). Our drug treatment data are consistent with an important role for E2F target gene inhibition by dexamethasone and the further downregulation of E2F members and E2F target genes after combined KPT-8602 and dexamethasone treatment.

The data presented here offer strategies to increase the effects of dexamethasone, to better treat or prevent relapse. We show that the combination of dexamethasone with KPT-8602 displays strong synergy in inhibiting proliferation and increasing apoptosis in both T-ALL and B-ALL cell lines, through a global increase in NR3C1 action, further potentiating the dexamethasone effect. We particularly noticed a diminished expression of cell cycle and DNA replication/repair genes. The combination is also more effective at diminishing the leukemic burden in a clinically relevant patient derived xenograft model of both B-ALL and T-ALL compared with the monotherapies. We believe that this synergy could lead to a better tolerance of treatment with potentially less side effects, but more importantly an enhanced dexamethasone efficacy.

D. Verbeke reports grants from FWO Vlaanderen and nonfinancial support from NanoString Technologies (nCounter PanCancer pathway panel) during the conduct of the study. J. De Bie reports grants from FWO Research Foundation (PhD fellowship) outside the submitted work. O. Gielen reports grants from VIB KULeuven (Kom Op Tegen Kanker) and VIB KULeuven (KULeuven) during the conduct of the study. K. Jacobs reports grants from KULeuven during the conduct of the study. N. Mentens reports grants from KUL (Kom op tegen Kanker) during the conduct of the study. A. Uyttebroeck reports grants from KULeuven (C1) during the conduct of the study. N. Boeckx reports grants from KULeuven during the conduct of the study. J. Maertens reports grants, personal fees, and nonfinancial support from Gilead and Pfizer; grants and personal fees from MSD; and personal fees and nonfinancial support from Mundipharma, Cidara, and F2G outside the submitted work. H. Segers reports grants from KULeuven (C1 grant) during the conduct of the study. J. Cools reports grants from Kom op Tegen Kanker and KU Leuven; nonfinancial support from NanoString (provided free reagents); and other from FWO (fellowships for D. Verbeke, C. Prieto, and J.D. Bie) during the conduct of the study. No potential conflicts of interest were disclosed by the other authors.

D. Verbeke: Conceptualization, data curation, formal analysis, validation, methodology, writing-original draft, writing-review and editing. S. Demeyer: Data curation, formal analysis, writing-review and editing. C. Prieto: Validation, methodology, writing-original draft, writing-review and editing. C.E. de Bock: Supervision, writing-review and editing. J. De Bie: Data curation, validation, writing-review and editing. O. Gielen: Formal analysis, methodology, writing-review and editing. K. Jacobs: formal analysis, writing-review and editing. N. Mentens: Formal analysis, writing-review and editing. B.M. Verhoeven: Formal analysis. A. Uyttebroeck: Resources, supervision, writing-review and editing. N. Boeckx: Resources, supervision, writing-review and editing. K. De Keersmaecker: Conceptualization, data curation, funding acquisition, writing-review and editing. J. Maertens: Data curation, writing-review and editing. H. Segers: Conceptualization, data curation, supervision, writing-review and editing. J. Cools: Conceptualization, data curation, formal analysis, supervision, funding acquisition, writing-original draft, writing-review and editing.

This work was supported by a grant from Kom op tegen Kanker (to J. Cools) and a C1 grant (C14/18/104) from the KU Leuven (to N. Boeckx, K.De Keersmaecker, J. Maertens, H. Segers, and J. Cools). D. Verbeke, C. Prieto, and J. De Bie were supported by fellowships from FWO Vlaanderen. We thank NanoString Technologies for providing the reagents and data analysis for the nCounter PanCancer pathway panel.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.