Dual Inhibition of MDM2 and XPO1 Synergizes to Induce Apoptosis in Acute Myeloid Leukemia Progenitor Cells with Wild-Type TP53 through Nuclear Accumulation of p53 and Suppression of c-Myc

Background. MDM2 is frequently overexpressed in acute myeloid leukemias (AML) and suppresses p53-mediated apoptosis while p53 mutations are relatively rare in AML. MDM2 inhibitors as a monotherapy have shown limited efficacy in clinical trials in AML (~25% response rate) (Andreeff, Clin Cancer Res 2015). XPO1 transports around 300 proteins, including p53 and other tumor suppressors, from the nucleus to the cytoplasm. Overexpression of XPO1 is associated with unfavorable outcomes in AML (Kojima, Blood 2013). p53 activation or XPO1 inhibition have been reported to decrease c-Myc protein levels through diverse mechanisms (Porter Mol Cell 2017 and Tabe PLoSOne 2015).

Objective: We investigated anti-leukemia effect of dual MDM2 and XPO1 inhibition, with the intent to maximize the pro-apoptotic functions of p53, using the MDM2 inhibitor milademetan (Daiichi-Sankyo), and selinexor, a recently FDA-approved XPO1 inhibitor or its analog eltanexor (Karyopharm).

Results: Treatment with milademetan and selinexor (1:1 molar ratio) induced synergistic apoptosis in AML cell lines with wild-type p53 (ED₅₀, 89.3 ± 18.6 nM, combination index (CI), 0.60 ± 0.08). Activity in p53 mutant AML required 40-fold higher ED₅₀ (3572 ± 1986 nM), reflected in an antagonistic CI of 6.94 ± 3.06. Knockdown of wild-type p53 by shRNA in OCI-AML3 (OCI-AML3 shp53) cells or presence of TP53 mutation (p.R248W) in MOLM-13 cells eliminated the synergistic effects, suggesting that normal p53 function is a major determinant of sensitivity to combined treatment.

Next, we treated primary AML samples with milademetan and selinexor or eltanexor and observed that effects were mutation-agnostic (e.g. RAS and FLT3) except for TP53. Combined treatment significantly reduced AUC determined by absolute live cell numbers compared to each drug alone, and induced synergistic apoptosis in primary AML samples with wild-type p53 (ED₅₀ values, 27.2 - 937.4 nM, CI, 0.51 ± 0.07), with similar efficacies in complex and non-complex karyotype AMLs (279.6 ± 94.7 vs 256.6 ± 56.4 nM, P = 0.84). In contrast, combined treatment showed antagonistic effects in primary AML samples with loss-of-function TP53 mutations (CI > 1.0). Immature CD34+CD38- AML cells were more susceptible to combined treatment than CD34- AML cells (apoptosis induction, 76.2 ± 6.7% vs 47.5 ± 6.8%, P = 0.0002)

Mechanistically, combined inhibition increased p53 protein levels and accumulated p53 but not MDM2 protein in the nucleus compared to each drug alone. Combined treatment induced more TP53 target genes (MDM2, CDKN1A, BBC3, FAS and Bax) in OCI-AML3 cells with control shRNA compared with OCI-AML3 shp53 cells. Combinatorial inhibition showed much enhanced reduction of c-Myc mRNA and protein levels in OCI-AML3 shC cells compared with OCI-AML3 shp53 cells (82% vs 32%). In confirmation, combined inhibition reduced c-Myc protein levels profoundly in wild-type p53 primary AMLs (ANOVA P < 0.0001). In contrast, c-Myc reduction was not observed in primary AMLs with p53-inactivating mutations. Intriguingly, OCI-AML3 cells overexpressing c-Myc by lentiviral transduction showed greater sensitivity to XPO1 inhibitors and the combination compared to empty-vector controls, and baseline levels of c-Myc protein also negatively correlated with ED₅₀ for combined treatment in primary AML samples (Spearman R = -0.5357, P = 0.0422).

Conclusion: These preclinical data suggest that dual inhibition of MDM2 and XPO1 induces synergistic apoptosis through accumulation of nuclear p53 and suppression of c-Myc in wild-type p53 AMLs. A clinical trial testing this concept in AML is under development.