XPO1 Relieves MYC-Induced Replication Stress Limiting the Immunogenicity of DLBCL Cells

The soluble nuclear pore component XPO1 is a therapeutic target in DLBCL and its selective inhibitor selinexor received FDA approval in relapsed/refractory DLBCL. XPO1 inhibitors block the nuclear export of a subset of cargo proteins. Since the spectrum of XPO1 cargos varies among cell types, the mechanism underlying the anti-lymphoma effect of selinexor has not been fully elucidated. Lack of this knowledge hampers the clinical use of XPO1 inhibitors in terms of patient selection and development of rational therapeutic combinations.

We reported that XPO1 is amplified and overexpressed in DLBCL. At the molecular level, we described that XPO1 overexpression functionally remodels the nuclear pore to sustain the expression of DNA damage repair proteins increasing lymphoma cells’ tolerance to genotoxic stress. Here, we investigated whether this XPO1 role is relevant for the anti-lymphoma effect of selinexor. Aberrant MYC expression induces DNA damage resulting from replication stress; thus, we speculated that selinexor kills MYC expressing (MYC+) DLBCL cells by limiting their ability to tolerate replication stress. Using B-cells expressing MYC in an inducible manner, we demonstrated that exposure to selinexor (6 h) caused G2-M cell cycle arrest upon MYC expression (p<0.0001, vs. vehicle-treated cells). Cell cycle arrest was associated with increased levels of DNA damage in S-phase as measured by yH2AX (p<0.0001). Furthermore, although MYC increased the expression of key replication stress repair proteins (i.e. RAD51, WEE and BRCA1), this effect was blunted by selinexor. To determine the clinical relevance of this mechanism, we interrogated publicly available datasets of DLBCLs and found that XPO1 amplification and overexpression was associated with MYC translocation (p<0.005), double-hit lymphoma signature (p<0.005), and elevated levels of chromosomal instability (p<0.005). Overall, these observations suggest that MYC-driven lymphomas sustaining high levels of genomic instability may critically depend on XPO1.

To therapeutically capitalized this mechanism, we evaluated the effect of selinexor on MYC+ DLBCLs. Clinically achievable doses of selinexor induced cell cycle arrest and apoptosis in 83% (5/6) of cell lines investigated in vitro. In a patient-derived xenograft (PDX) MYC+ DLBCL model, a human-equivalent dose of selinexor impaired tumor growth (p< 0.05, vs. vehicle) without significant toxicities. Reduced tumor growth was secondary to apoptosis as determined by TUNEL (p<0.05). To characterize the biological effects of XPO1 inhibition, we compared the transcriptome of PDX from the previous experiment. Pathway analysis of differentially expressed genes revealed that selinexor induced pathways regulating innate immune signaling. The upregulation of key genes from this pathway (i.e. DDX58, DDX60, IRF7 and STAT1) was validated by qRT-PCR. Since unrepaired DNA damage elicits an innate immune response, we hypothesized that these effects could be mechanistically linked. To test this notion, we exposed MYC+ DLBCL cells to sub-lethal doses of the DNA damaging agent etoposide, selinexor, or their combination. Expression of innate immune genes was significantly higher in cells exposed to the combination compared to either drug alone (p<0.001, for both comparisons), indicating that selinexor increases the immunogenicity of MYC+ DLBCL cells and this effect is enhanced by concomitant replication stress. Furthermore, by analyzing primary DLBCLs we found that poorly immunogenic lymphomas, characterized by an immune deserted microenvironment, expressed higher levels of XPO1 and MYC compared to DLBCL with an immune rich microenvironment (P<0.001, for both genes). To determine a potential effect of selinexor in the function of immune effector cells, we performed single cells secretomic analysis (Isolight) of circulating CD8+ T cells obtained from two DLBCL patients before and after selinexor treatment and found that selinexor increases the secretory activity of CD8+ T cells in both patients.

Overall, our study reveals a key oncogenic role for XPO1 in enabling DLBCL cells to tolerate MYC-induced replication stress while limiting their immunogenicity. This suggest that patients with MYC+ DLBCL and elevated genomic instability may benefit from therapeutic approaches based on XPO1 inhibitors in combination with immunomodulatory drugs, an approach we will test prospectively.