XPO1-dependent nuclear export as a target for cancer therapy

Eukaryotic cells have well-separated nuclear and cytoplasmic compartments. Proper cellular functions require the exchange of large molecules through the nuclear pore complex (NPC). Small molecules can passively diffuse through the NPC, whereas the transport of larger cargoes, including RNAs and proteins, requires various transport receptors of the importin beta superfamily [1, 2]. XPO1 is a major transport receptor, responsible for exporting proteins and multiple RNA species. Originally named as CRM1 (chromosomal region maintenance 1), XPO1 was identified in Schizosaccharomyces pombe as a gene required for maintaining higher-order chromosome structure [3]. Subsequently, it was shown to function as a shuttling protein, mediating the nuclear export of proteins and mRNAs in Saccharomyces cerevisiae and renamed as XPO1 (exportin 1) [4].

XPO1 is a nuclear export receptor with a pleiotropic role in transporting a plethora of proteins and RNA species, including rRNAs, snRNAs, mRNA, microRNAs, and tRNAs [5] (Fig. 1). XPO1 functions together with RAN GTPase, which provides the energy for transport and ensures the directionality of nuclear export [6]. In the nucleus, XPO1 binds to the nuclear export signal (NES) on its target proteins and to RAN in its active GTP-bound form (RAN-GTP). The complex is subsequently docked to NPC and passes through the nuclear membrane into the cytoplasm. Hydrolysis of RAN-GTP to RAN-GDP causes the disassembly of the complex and release of cargoes in the cytoplasm. The directionality of XPO1-mediated export is determined by the concentration gradient of RAN-GTP, which is predominantly confined to the nucleus [7] (Fig. 1). In addition to its role in nuclear-cytoplasmic transport during the interphase of cell cycle, XPO1/RAN regulates mitosis.

XPO1 overexpression is a common feature among many human cancer types, including pancreatic, ovarian, glioma, lung, gastric, prostate, and colorectal cancers, and is associated with poor prognosis [34‐40]. While amplified copy number may explain the XPO1 overexpression in some leukemia and lymphoma subtypes [41], for the majority of human cancers, the mechanism of XPO1 overexpression remains unknown. Genetic studies demonstrate a converse role for MYC and P53 in regulating XPO1 transcription [42]. The MYC overexpression and/or P53 loss of function observed in most human cancers are likely responsible for the widespread XPO1 overexpression. Interestingly, several XPO1 binding partners and/or adaptor proteins involved in nuclear export, namely RAN, HuR, eIF4E, LRPPRC, and NXF3, are also frequently overexpressed in human cancers and correlate with poor prognosis [24, 43‐47], pointing to the aberrations of nuclear export machinery as a major hallmark of tumorigenesis. In addition to overexpression, XPO1 mutations have been found in 0.5–2.9% of solid and hematopoietic tumors, with the highest incidence in lymphomas [48]. A recent large-scale analysis of whole-exome and genome sequencing data from 42,793 patients has identified highly recurrent mutations of XPO1 (E571, R749, and D624) specifically in B cell malignancies. The E571K mutation is enriched in primary mediastinal B cell lymphoma (33%), classic Hodgkin lymphoma (14%), diffuse large B cell lymphoma (DLBCL) (2%), and chronic lymphocytic leukemia (CLL) (3%) [49]. Furthermore, XPO1^E571K mutation has been shown to cooperate with MYC and BCL2 in promoting lymphomagenesis. These observations suggest that gain-of-function mutations convert XPO1 to an oncogenic driver, particularly in B cell malignancies. The exclusive presence of XPO1 mutations in B cell malignancies is largely unknown but is likely related to somatic hypermutations of immunoglobulin genes, where the error-prone polymerase η may introduce a high frequency of A > T and C > G transversions [50‐52]. Due to frequent deregulation in cancers, XPO1 has been identified as a therapeutic target in many tumor types. Recent CRISPR library and RNAi screens have validated XPO1 as a therapeutic target in sarcoma, DLBCL, multiple myeloma, and KRAS-mutant lung cancer [9, 53‐55]. Specific XPO1 inhibitors have been extensively tested and demonstrated efficacy in a broad range of cancer types in preclinical studies.

The development of XPO1/CRM1 inhibitors dates back to early 1980s with the discovery of multiple classes of compounds and has been elegantly summarized in another review [56]. The first characterized XPO1 inhibitor, leptomycin B, was isolated from a strain of Streptomyces as an antifungal agent [57, 58]. Shortly following isolation and purification, its antitumor efficacy was examined in murine transplantation tumor models and demonstrated survival benefits [59]. However, a phase 1 trial of leptomycin B carried out in a small cohort of patients with various advanced cancers demonstrated marked malaise and anorexia, with no partial response [60]. Further clinical development of leptomycin B was therefore halted following the failed trial. XPO1/CRM1 was subsequently identified as the molecular target of leptomycin B with covalent binding at Cys528 [61]. Leptomycin B analogs, including ratjadones, anguinomycins, and KOS2464, also covalently bind to XPO1/CRM1 [62‐64]. In particular, the semisynthetic KOS2464 has demonstrated significant therapeutic efficacy in several solid and hematopoietic tumor types with much reduced toxicity compared to leptomycin B [64].

In recent years, synthetic inhibitors of XPO1/CRM1, including PKF050-638, CBS9106, and selective inhibitors of nuclear export (SINE), have been developed. CBS9106 was shown to bind XPO1/CRM1, suppress its nuclear export activities, and induce XPO1/CRM1 protein degradation [65, 66]. Treatment with CBS9106 induces cell cycle arrest and apoptosis in a broad spectrum of cancer cells. Oral administration of CBS9106 significantly suppresses tumor growth and prolongs survival without a significant loss of body weight in a xenograft mouse model of multiple myeloma [65]. PKF050-638 originally identified as an inhibitor of HIV Rev protein also interferes with XPO1/CRM1 function [67]. Its N-azolylacrylate scaffold was later adopted by the SINE compounds, a series of structurally related small molecule inhibitors, including KPT-185, KPT-276, KPT-335, KPT-330 (selinexor), KPT-8602 (eltanexor), and SL-801 (felezonexor) [68, 69]. SINE compounds specifically bind to Cys528 in the cargo-binding groove of XPO1/CRM1. In contrast to leptomycin B, the covalent binding of SINE compounds to XPO1/CRM1 is slowly reversible, potentially explaining SINEs’ relatively low toxicity [63, 68]. The development of SINEs has enabled the translational application of nuclear export inhibitors as novel therapeutic agents. A number of SINE compounds have been tested extensively in preclinical settings, exhibiting efficacy in solid tumors [14, 70‐74] and hematopoietic malignancies [75‐79]. Among the investigated SINEs, KPT-330, KPT-8602, and SL-80 have been advanced to clinical trials. KPT-330 (selinexor) in particular has been evaluated in the majority of the trials and recently received FDA approval for resistant and relapsed multiple myeloma [80].

An important aspect of normal cell function, nuclear-cytoplasmic export, is often deregulated in cancers, providing a unique therapeutic opportunity. The recent FDA approval of selinexor for the treatment of RR multiple myeloma lends credence to this therapeutic strategy. Future expansion of the clinical indications of selinexor is contingent upon the success of additional clinical trials. More SINE compounds with improved efficacy and reduced adverse events will be tested and potentially approved for clinical use. In this light, extensive genetic and preclinical studies and well-designed clinical trials are the prerequisite to exploiting the full potential of XPO1 inhibitors in cancer treatment.