Discovery of a highly potent and orally available importin-β1 inhibitor that overcomes enzalutamide-resistance in advanced prostate cancer

Abstract

Nuclear transporter importin-β1 is emerging as an attractive target by virtue of its prevalence in many cancers. However, the lack of druggable inhibitors restricts its therapeutic proof of concept. In the present work, we optimized a natural importin-β1 inhibitor DD1 to afford an improved analog DD1-Br with better tolerability (>25 folds) and oral bioavailability. DD1-Br inhibited the survival of castration-resistant prostate cancer (CRPC) cells with sub-nanomolar potency and completely prevented tumor growth in resistant CRPC models both in monotherapy (0.5 mg/kg) and in enzalutamide-combination therapy. Mechanistic study revealed that by targeting importin-β1, DD1-Br markedly inhibited the nuclear accumulation of multiple CRPC drivers, particularly AR-V7, a main contributor to enzalutamide resistance, leading to the integral suppression of downstream oncogenic signaling. This study provides a promising lead for CRPC and demonstrates the potential of overcoming drug resistance in advanced CRPC via targeting importin-β1.

Graphical abstract

Structural optimization of a natural importin-β1 inhibitor DD1 led to a potent, orally available, and tolerable analogue DD1-Br that overcomes enzalutamide-resistance in advanced prostate cancer. This study provides an invaluable lead for CRPC, and demonstrates the possibility of overcoming drug resistance in advanced CRPC via targeting importin-β1.

1. Introduction

Prostate cancer (PC) is the most frequent male cancer and a leading cause of malignancy-related mortality for males^1,2. Since PC initially depends on androgen for survival and growth, androgen deprivation therapy (ADT) with chemical or surgical castration represents the standard treatment for early-stage PC³. However, this disease progresses into a lethal stage known as castration-resistant prostate cancer (CRPC), which could grow under low androgen conditions with restored androgen receptor (AR) signaling^4,5. Although patients with CRPC could gain certain benefits from AR signaling targeting therapies, such as AR antagonist enzalutamide (ENZ) and androgen synthesis inhibitor abiraterone (ABI), durable responses are limited, largely due to acquired resistance^6,7. Therefore, the need for novel therapeutics with unique mechanisms to overcome drug resistance in CRPC remains acute.

Karyopherins, known as importins and exportins, are members of nuclear transport receptors involved in the translocation of macromolecules through the nuclear pore complex (NPC)8, 9, 10. In tumors, specific importins and/or exportins are often overexpressed to meet the extra transportation demand for maintaining the abnormally high rate of cell growth^11,12. Thus, inhibiting tumor-suppressor proteins being exported out of or oncoproteins being imported into the nucleus via regulating karyopherins is considered as a novel approach for cancer therapeutics. With the successful launch of exportin drug, selinexor, a selective inhibitor of exportin-1 (XPO1) to treat multiple myeloma (MM) patients and B-cell lymphoma (BCL) patients, karyopherin members are revealing as appealing therapeutic targets in cancer drug discovery^13,14. However, the development of importin drugs has been stagnating, owing to the lack of druggable inhibitors for proof-of-concept study.

Importin-β1 (KPNB1) is a key nuclear transporter overexpressed in CRPC¹⁵. Its expression level is positively correlated with patient's deterioration degree, but the regulatory role remains unclear. In our previous study, we discovered a very potent anti-CRPC agent DD1 and identified that importin-β1 as the direct target (binding sites: Gln560, Phe577, and Leu580), and further confirmed for the first time that importin-β1 is a druggable vulnerability in CRPC¹⁵. However, DD1 had a narrow therapeutic window (LD₅₀ ≈ 4 mg/kg for i.p.) and poor oral bioavailability. Besides, the therapeutic potential of such inhibitors in resistant CRPC still remains unclear. Therefore, an improved analogue with better druggability is in urgent need to clarify the prospect of importin-β1-based CRPC therapy. In the current study, we optimized DD1 to generate an improved importin-β1 inhibitor DD1-Br with better tolerability and oral bioavailability. DD1-Br displayed potent anti-CRPC efficacy in vitro and in vivo, and could overcome the enzalutamide resistance. Mechanistic study showed that by targeting importin-β1, DD1-Br significantly inhibited the nuclear accumulation of multiple CRPC drivers, such as AR, E2F1, and MYC, integrally suppressing downstream oncogenic signaling. In particular, the down-regulation of AR-V7 signaling may account for the main mechanism in overcoming enzalutamide resistance. This study opens a new perspective for CRPC drug development, demonstrating the possibility of overcoming drug resistance in advanced prostate cancer via importin-β1.

2. Results and discussion

2.1. Structural optimization of DD1 afforded the improved analogue DD1-Br

DD1 is a structurally complex daphnane diterpenoid (DD, 5/7/6 ring system) containing an orthoester functionality. It was selected from a series of active DDs (DD1−DD14, Fig. 1A) in anti-CRPC screening campaign towards a natural diterpenoid library¹⁵. As the activity is a major factor prior to other druggability parameters such as toxicity and bioavailability, we firstly tried to obtain the DD analogues with decent anti-CRPC activity. To expand the sample size of this category, simple chemical modifications were conducted on the major isolates DD1 and DD4. As shown in Fig. 1A, the acylation of 5-OH or 20-OH in DD1 and DD4 with acetic anhydride or 2-thiophenecarbonyl/benzoyl chlorides afforded a series of acylated derivatives, DD1a−DD1d and DD4a−DD4d. Alkaline hydrolysis of DD4 obtained DD4e. As halogenation is known to increase the lipophilicity, membrane permeability, and oral absorption in late-stage lead modification¹⁶, 6,7-epoxy ring was opened by a series halogenides. The ring-opening of 6,7-epoxy in DD1 with PBr₃ afforded 6- or 7-brominated derivatives DD1-Br and DD1g. The ring-opening of 6,7-epoxy in DD1 and DD4 with 36% HCl at room temperature afforded four 6- or 7-chlorinated derivatives DD1e, DD1f, DD4f, and DD4g. Besides, the hydrogenation of DD1 afforded DD1h and DD1i. Thus, a small library containing 31 natural or synthetic DDs was constructed. The anti-proliferative activity of these DDs on CRPC cell line C4-2B revealed brief structure–activity relationships (SARs): 1) 9,13,14-orthoester group was essential for the activity, as the absence of this functionality resulted in a dramatic decrease of the activity (e.g., DD11, DD12, and DD14); 2) The double bond between C-1 and C-2 is beneficial to the activity (DD1 vs DD8). 3) The free OH at C-5 and C-20 was favorable to the activity (DD4 vs DD4a−4d; DD1 vs DD1a−1d), while it was unfavorable at C-12 (DD4 vs DD4e). 4) The chlorinative and hydroxylated ring-opening of 6,7-epoxy were harmful to the activity (DD1 vs DD10, DD1 vs DD1e and DD1f), but the bromine at C-6 and C-7 restored the activity of the ring-opening products (DD1-Br and DD1g vs DD10). The above modifications showed most of the functionalities in DD1 were untouchable, with only the bromine-substituted analogues DD1-Br and DD1g and Δ¹⁵-hydrogenated analogue DD1h maintaining comparable activity to DD1. Then, the acute toxicity assay was carried out on these compounds. The results showed that DD1-Br had the most improved tolerability (LD₅₀ > 100 mg/kg for i.p.), being 25-fold greater than that of DD1 (Fig. 1B and C and Supporting Information Fig. S1A). Furthermore, the pharmacokinetic study also showed that DD1-Br had an improved profile in both intraperitoneal (i.p.) and oral (p.o.) administrations, with the C_max values being 5-fold increase in i.p. and oral bioavailability being 32-fold increase (DD1-Br F = 64.1%; DD1 F = 2.21%, Fig. 1D and Fig. S1B and S1C). Taken together, DD1-Br was turned out to be an improved analogue of DD1 with comparable activity but better druggability.

Figure 1.

Structural optimization of DD1 afforded an improved analogue DD1-Br with better tolerability and oral bioavailability. (A) Structures of DD1−DD14 and the synthesis of daphnane derivatives. (a) 36% HCl, THF, rt, 0.5 h. (b) PBr₃, CH₂Cl₂, −60 °C. (c) 1% NaOH in MeOH (m/v), rt, 0.5 h. (d) H₂, 10% Pd/C, MeOH, 0.5 h. (e) Acetic anhydride/benzoyl chlorides/2-thiophenecarbonyl, Pyr, rt/50 °C. IC₅₀s for DDs in C4-2B cells treated for 48 h. (B) Survival curve showing the survival of the mice (KM) after the indicated single dose treatments. The estimated LD₅₀ value was indicated. (C) SARs of DDs on anti-proliferative activity and toxicity. (D) ICR mice were treated p.o. or i.p. with 1 mg/kg of DD1 or DD1-Br for pharmacokinetic (PK) analysis. Plasma concentration of DD1 or DD1-Br was measured within 24 h after a single dose treatment. Results are shown as mean ± SD. n = 3 mice at each time point.

2.2. DD1-Br is a potent inhibitor of CRPC cell growth and survival

To systematically evaluate the activity of DD1-Br, we tested it on a subset of AR-positive (AR⁺) CRPC cell lines, AR-negative (AR⁻) PC cell lines, and normal cells. The results showed that DD1-Br could selectively inhibit the cell growth of AR⁺ CRPC cell lines (such as C4-2B, VCaP-CRPC, and 22Rv1) at sub-nanomolar range, with SI value being ∼1000 to AR⁻ cell lines (DU145 and PC-3) and prostatic normal cells (RWPE-1) (Fig. 2A and B and Supporting Information Fig. S2A). Further studies showed that DD1-Br could induce cell apoptosis, as proved by the appearance of the cleaved caspase-7 and poly polymerase 1 (PARP1), and suppress the expression of oncoproteins involved in oncogenesis, survival, and proliferation, such as MYC and cyclin B2 (Fig. 2C and Fig. S2B). Consistent with these effects, DD1-Br potently inhibited CRPC cell colony formation (Fig. 2D) and significantly inhibited the growth of the 3D organoid derived from prostate cancer patient-derived xenograft (PDX) (Fig. 2E). Overall, these results indicated that DD1-Br could potently inhibit CRPC cell growth and survival.

Figure 2.

DD1-Br significantly inhibits the growth and survival of CRPC cells. (A) IC₅₀s of DD1, DD1-Br, ENZ, and DOX in AR-positive or AR-negative cell lines treated for 48 h were shown in visual heat map. (B) Dose-dependent inhibition of CRPC cell viability in DD1-Br treatment for 48 h. (C) Immunoblotting analysis of DD1-Br on expression levels of the designated proteins in C4-2B and VCaP-CRPC cells. Cells were incubated with different concentrations of DD1-Br for 24 h. (D) 22Rv1 and C4-2B cells were incubated with different concentrations of DD1-Br for 12 days, and then colony formation was assessed. (E) Patient-derived xenograft derived organoids (PDO) were incubated with various concentrations of DD1-Br for 6 days. Typical photographs were recorded by microscope and PDO cell viability was assessed. Scale bar = 200 μm. Data in (D) and (E) are presented as mean ± SD. of three independent experiments; ∗P < 0.05, ∗∗P < 0.01 vs control; Student's t test.

2.3. DD1-Br targets importin-β1 and its anti-CRPC activity is importin-β1 dependent

To verify if DD1-Br still targets importin-β1, the surface plasmon resonance (SPR) experiments and cellular thermal shift assay (CETSA) were performed to investigate the interactions between DD1-Br and importin-β1. As shown in Fig. 3A, DD1-Br strong bound to importin-β1 (K_D = 0.219 μmol/L, being similar to that of DD1 (K_D = 0.229 μmol/L)¹⁵, and the ability of DD1-Br (20 nmol/L) to enhance the thermostabilization of importin-β1 protein in C4-2B cells indicates target engagement (Fig. 3B). Furthermore, in endogenous affinity pull-down assay, the importin-β1 precipitated by DD1-photoaffinity probe (PT2)¹⁵ could be competitively offset by DD1-Br (Fig. 3C), suggesting that DD1-Br shared the same binding modes with DD1. This was further confirmed by the high binding affinity between DD1-Br and C-terminal 550–876 aa domain of importin-β1 (K_D = 0.201 μmol/L, Fig. S2C) as well as the nice fitting of DD1-Br in hydrophobic pocket formed by the importin-β1 residues Gln560, Phe577, and Leu580 in molecular modeling (Fig. 3D).

Figure 3.

DD1-Br directly targets importin-β1 and its anti-CRPC activity is importin-β1 dependent. (A) Direct interaction between DD1-Br and importin-β1 in vitro was analyzed by SPR assay. (B) Melt curves of importin-β1 protein in DD1-Br (20 nmol/L for 1 h) treated C4-2B cells were measured by CETSA. (C) Endogenous competition assay. 22Rv1 cells were incubated with different concentrations of DD1-Br (0, 5 or 10 μmol/L), and meanwhile treated with PT2 (5 μmol/L) for 6 h. (D) The binding mode of DD1-Br and importin-β1 (ID: 2P8Q) according to docking simulation. (E) DD1-Br blocks importin-β1-mediated nuclear import in living cells. HEK-293T cells stably expressing NFAT-GFP were treated with DD1-Br (40 nmol/L) for 1 h before treatment with ionomycin for 30 min to induce NFAT-GFP nuclear accumulation. Representative images and quantification of the percentage of cells with nuclear NFAT-GFP. Scale bars = 40 μm. (F) Anti-proliferation assay. CRPC cells were transfected with scrambled siCON or KPNB1 siRNA for 2 days, followed by incubation with different concentrations of DD1-Br for 2 days (in C4-2B) or 3 days (in 22Rv1). (G) 22Rv1-EV (empty vector), 22Rv1-KPNB1^wt and 22Rv1-KPNB1^m1 cells were incubated with various concentrations of DD1-Br for 4 day, viable cells were counted. Data in (F) and (G) are represented as mean ± SD. of three independent experiments; ∗P < 0.05, ∗∗P < 0.01, n.s., not significant vs control; Student's t test.

Given the known role of importin-β1 in nuclear import, we next investigated the influence of DD1-Br on importin-β1-mediated nuclear import in cells by using a reporter system¹⁷. As shown in Fig. 3E, under the stimulation of calcium ionophore ionomycin, the importin-β1-dependent nuclear accumulation of NFAT-GFP could be significantly decreased by DD1-Br, which was consistent with the observed effects of DD1 and importazole (IPZ, a known importin-β1 inhibitor)¹⁵. We also detected the expression of importin-β1 after DD1-Br treatment, and the results showed it had no significant effect (Fig. S2D). Next, we investigated if the anti-proliferative efficacy of DD1-Br is dependent on importin-β1. As shown in Fig. 3F, in C4-2B or 22Rv1 cells the knockdown of KPNB1 significantly attenuated the efficacy of DD1-Br as compared to control cells. Importantly, contrary to what was surveyed when expressing KPNB1^wt in 22Rv1 cells, exogenic expression of KPNB1^m1 (mutation the binding sites Gln560, Phe577, and Leu580) significantly weakened the anti-proliferative effects of DD1-Br (Fig. 3G). Collectively, DD1-Br exhibited these anti-proliferative effects in CRPC by targeting importin-β1.

2.4. DD1-Br reprograms the nuclear localization of key CRPC survival proteins and shuts down their downstream oncogenic signaling

Given importin-β1 could mediate nuclear import of a cluster of oncoproteins to promote prostate cancer aggressiveness^15,18, we next examined the effect of DD1-Br on the nuclear translocation of key oncogenic factors in CRPC. As shown in Fig. 4A and B, DD1-Br could significantly down-regulated the nuclear accumulation of multiple CRPC drivers, including some transcription factors (e.g., MYC, E2F1, AR, and FOXM1) and coactivators (e.g., PCNA, CREBBP, and PARK7). To further dissect the influence of DD1-Br on gene transcription programs, we analyzed RNA-seq transcriptomes of C4-2B cells with the treatment of DD1-Br. The results revealed that 4567 genes were markedly down-regulated, while 3916 genes were up-regulated. Among them, the majority of the down-regulated genes were associated with AR, MYC, and E2F pathways using hallmark gene sets analysis, which were subject to the regulation of aforementioned key CRPC transcription factors (Fig. 4C). This was further corroborated by the gene-set enrichment analysis (GSEA) and RT-PCR verification of target genes (Fig. 4D and E).

Figure 4.

DD1-Br alters the key CRPC oncoproteins nuclear accumulation and shuts down their downstream oncogenic signaling. (A, B) Immunoblot and confocal microscopy images detection of the designated proteins in cytoplasm and nucleoplasm in C4-2B cells. Cells were incubated with DD1-Br (10 nmol/L) for 6 h. Scale bars = 20 μm. (C) The enrichment analysis of down-regulated genes by DD1-Br was displayed by bubble chart. Down-regulated genes were obtained from RNA-seq data of C4-2B cells incubated with DD1-Br (20 nmol/L) for 24 h. (D) GSEA profiles for AR, E2F, and MYC pathway signature gene sets from RNA-seq data. (E) qRT-PCR analysis of AR, E2F, and MYC target genes. C4-2B cells were incubated with various concentrations of DD1-Br for 24 h. Data are presented as mean ± SD. of three independent experiments; ∗∗P < 0.01 vs control; Student's t test. (F and G) Venn diagram (F) showed genes that were regulated by DD1 (20 nmol/L, 24 h) and DD1-Br (20 nmol/L, 24 h) treatment of C4-2B cells from RNA-seq. GO analysis (G) on DD1-Br uniquely regulated genes in C4-2B cells.

Interestingly, RNA-seq analysis revealed that among the DD1-Br-affected genes 61.4% were common to those regulated by DD1 (Fig. 4F). Gene ontology analysis and GSEA suggested that the genes uniquely regulated by DD1-Br were associated with biological processes such as cell death, survival, metabolism process, oxygen and migration (Fig. 4G), while those uniquely regulated by DD1 were related to xenobiotics metabolism, nervous system dysfunction, and dilated cardiomyopathy (Supporting Information Fig. S3A and S3B). These differences may partly explain the improved tolerability and bioavailability of DD1-Br as compared with DD1.

Taken together, DD1-Br could inhibit the nuclear accumulation of vital CRPC survival transcription factors and robustly interrupt their target gene transcription programs.

2.5. DD1-Br is a potent, orally available, and tolerable anti-CRPC agent in vivo

Next, we evaluated the efficacy of DD1-Br in vivo by C4-2B xenograft and PDX models. The results showed that i.p. administration of DD1-Br at 0.1 or 0.5 mg/kg markedly inhibited the increase of tumor volume in C4-2B xenografts (Fig. 5A), and i.p. injected with DD1-Br at 0.1 mg/kg could totally prevent the tumor growth in PDX model (Fig. 5B). Immunohistochemical studies supported that the efficacy of DD1-Br was paralleled by a decrease of Ki67, an increase of caspase-3, as well as the suppressed nuclear localization of AR, E2F1, MYC, and PARK7 (Fig. 5C and D and Supporting Information Fig. S4A). As the pharmacokinetic profiles of DD1-Br showed an improved oral bioavailability, we measured its oral efficacy in 22Rv1 xenograft model for providing data more relevant in a clinical setting. The results indicated that oral dosing of DD1-Br at 2 or 0.5 mg/kg still maintained a good efficacy (Fig. 5E).

Figure 5.

DD1-Br is a potent, orally available, and tolerable anti-CRPC agent in vivo. (A) Antitumor activity of the designated treatments (DD1-Br, 0.5 mg/kg or 0.1 mg/kg, intraperitoneally i.p., 5 times a week, n = 6 mice) on the tumor growth of C4-2B xenografts. a (P = 0.0012); b (P = 0.00044). (B) Growth curve of PDX model (TM00298) treated with DD1-Br (0.1 mg/kg intraperitoneally i.p., 5 times a week, n = 7 mice). a (P = 0.013). (C) IHC quantification of AR, E2F1, MYC, and PARK7 nuclear localization in xenograft tumor tissue of mice bearing PDX treated with vehicle or DD1-Br. Data are presented as mean ± SD. ∗∗P < 0.01 vs vehicle; Student's t test. (D) Anti-Ki67 or c-caspase 3 IHC images and quantification of tumor section from PDX were shown. Scale bars = 50 μm. (E) Effects of the indicated treatments (DD1-Br, oral administrations p.o., daily, n = 7 mice) on the growth of 22Rv1 xenografts. a (P = 5.09 × 10⁻⁵); b (P = 1.63 × 10⁻⁸). (F) Tolerance test, expressed as percentage change of mice body weight in previous study¹⁵ (on Days 10 and 22 in 22Rv1 tumor models or on Days 15 and 30 in PDX models) for the indicated DD1 doses. Student's t test. (G) Tolerance test, expressed as percentage change of mice body weight in (B) (on Days 15 and 30) and (E) (on Days 9 and 18) for the indicated DD1-Br doses. Results in (A), (B), and (E) are presented as the mean tumor volume ± SEM.

In a previous study we found that the treatment of DD1 (i.p. 0.5 mg/kg, 5 times per week) resulted in mice body weight loss either in the early stage of administration (weight loss ∼13%) or at the end stage of administration (weight loss ∼9%) in both models (Fig. 5F). In contrast, the treatments of DD1-Br showed no apparent influence on mice body weight in all administration stages in both 22Rv1 xenograft model and PDX model (Fig. 5G and Fig. S4B). Consistent with this, there were no apparent pathologic abnormalities being observed in the appearance and histological study of multiple organs, including lungs, heart, liver, kidneys, spleen, testicles, and epididymis (Fig. S4C). Additional toxicology analysis further suggested the complete blood count with differential, blood chemistry, and kidney/liver function was normal in DD1-Br-treated mice (Fig. S4D). Besides, Discovery Studio software was used to predict and simulate the toxicological properties of DD1 and DD1-Br. As shown in Fig. S4E, consistent with our experimental results, the predicted toxic and side effects of DD1-Br were significantly lower than those of DD1. Altogether, these results suggested that DD1-Br is an improved analogue of DD1 with better in vivo druggability.

2.6. DD1-Br inhibited AR-V7 signaling and sensitized resistant CRPC to ENZ

Enzalutamide (ENZ), the current first-line therapy for CRPC, is encountering the clinical challenge of acquired resistance¹⁹. Mechanisms of ENZ resistance are rather complex, including intratumoral androgen biosynthesis²⁰, AR overexpression and amplification²¹, AR mutations or generation of AR splicing variants^22,23, as well as activation of parallel pathways²⁴. Notably, recent studies have shown that the emergence and continuous activation of AR splicing variants, particularly AR variant 7 (AR-V7), play vital roles in therapeutic resistance²⁵. AR-V7 lacks the functional ligand-binding domain (LBD), but remains recruited onto targeted promoters or coactivators to activate oncogene transcription and drive PC progression even in the presence of ENZ²⁶. Thus, the suppression of AR-V7 signaling in CRPC may bring the benefits to ENZ therapy.

In protein and RNA-seq analysis, we found that DD1-Br could strongly inhibit the nuclear accumulation of AR-V7 and integrally down-regulate its downstream signaling (Fig. 6A‒C and Supporting Information Fig. S5A), suggesting that DD1-Br may synergize with ENZ in killing resistant CRPC cells. To verify this, the anti-CRPC activity of the combination treatment (DD1-Br + ENZ) was evaluated in 22Rv1 and C4-2B/ENZR cells with high AR-V7 expression. As expected, the combination of DD1-Br and ENZ in a relatively low concentration synergistically inhibited the colony formations and proliferation of these resistant CRPC cells (Fig. 6D and Fig. S5B and S5C). Moreover, to further demonstrate the synergistic effects were due to the blockage of AR-V7 signaling, we evaluated the coefficient of drug interaction (CDI) of combination therapy after knockdown of AR-V7. As shown in Fig. S5D, DD1-Br at 15 nmol/L showed a potent synergistic effect with ENZ in normal 22Rv1 cells (CDI = 0.663), while this synergistic effect was significantly diminished after AR-V7 knockdown (CDI = 1.074), indicating that the inhibition of AR-V7 signaling contributed to the main mechanism of DD1-Br + ENZ therapy.

Figure 6.

DD1-Br inhibited AR-V7 signaling and sensitized resistant CRPC to ENZ. (A) Immunoblotting analysis of AR-V7 in nucleoplasm (Nucle) and cytoplasm (Cyto) of 22Rv1, C4-2B/ENZR, and 22Rv1-KPNB1^m1 cells treated with DD1-Br (10 nmol/L for 6 h in C4-2B/ENZR cells; 25 nmol/L for 24 h in 22Rv1 and 22Rv1-KPNB1^m1 cells). (B) GSEA of AR-V7 up-regulated genes set in C4-2B cells. Cells were incubated with DD1-Br (20 nmol/L) for 24 h. (C) The heatmap presentation of alterations in expression of AR-V7 up-regulated genes in C4-2B cells. Cells were incubated with DD1-Br (20 nmol/L), DD1 (20 nmol/L), or IPZ (20 μmol/L) for 24 h. Gene expression was performed by RNA-seq. (D) C4-2B/ENZR and 22Rv1 cells were incubated with designated concentrations of DD1-Br, or in combination with ENZ for 14 days. Colony formation images were taken (left), and colony numbers were counted (right), Data are presented as mean ± SD. n = 3; ∗∗P < 0.01 vs control; Student's t test. (E, F) Effects of the indicated treatments (DD1-Br, i.p., 0.1 mg/kg; ENZ, p.o., 10 mg/kg; DD1-Br and ENZ combined; 5 times a week) on the tumor growth of 22Rv1 xenografts. a (P = 0.213, ENZ vs vehicle); b (P = 9.71 × 10⁻⁵, DD1-Br vs vehicle, P = 0.0062, DD1-Br vs ENZ); c (P = 4.32 × 10⁻⁷, DD1-Br + ENZ vs vehicle, P = 5.74 × 10⁻⁵, DD1-Br + ENZ vs ENZ, P = 0.0014, DD1-Br + ENZ vsDD1-Br) determined by Student's t test at 18 day. Results are presented as the mean tumor volume ± s.e.m. Volumes (left) and weights (right) are shown. (G) Individual tumor growth trajectories in (E). (H) Effects of the indicated treatments (DD1-Br, i.p., 0.5 mg/kg; ENZ, p.o., 10 mg/kg; DD1-Br and ENZ combined; 5 times a week) on the tumor growth of VCaP-CRPC xenografts in castrated mice. a (P = 0.497, ENZ vs vehicle); b (P = 0.037, DD1-Br vs vehicle, P = 0.012, DD1-Br vs ENZ); c (P = 0.013, DD1-Br + ENZ vs vehicle, P = 0.0021, DD1-Br + ENZ vs ENZ, P = 0.021, DD1-Br + ENZ vsDD1-Br) determined by Student's t test at 30 day. (I) Establishment and treatment of VCaP-CRPC xenograft model. (J) Anti-Ki67 IHC images and quantification of tumor cells from (E) and (H). Scale bars = 50 μm.

Next, we assessed the in vivo combination therapy efficacy in advanced CRPC models. In 22Rv1 xenograft model, administration of DD1-Br alone (0.1 mg/kg, i.p.) substantially inhibited the tumor growth, while dosing ENZ alone (10 mg/kg, p.o.) did not show significant effect. The combination of these two drugs led to complete growth arrest on tumors, suggesting a potent synergistic effect of the combination therapy (Fig. 6E‒G). Moreover, to mimic the human disease more closely, we established a resistant VCaP-CRPC xenograft tumor model based on postcastration relapsed growth of VCaP cells in castrated mice. In this model, the combination of DD1-Br (0.5 mg/kg, i.p.) and ENZ (10 mg/kg, p.o.) elicited an even more potent synergistic effect, resulting in tumor regression (Fig. 6H, I and Fig. S5E‒S5G). Consistent with this, the combination therapy synergistically increased the expression of apoptotic protein cleaved caspase 3 and decreased the expression of proliferative protein Ki67 in tumor tissues of both models (Fig. 6J and Fig. S5H and S5I). It was worth noting that the combination showed no influence on the animal body weight during the treatment, suggesting its potential safety (Fig. S5J).

Collectively, these results suggested that importin-β1 inhibitor DD1-Br could sensitize resistant CRPC tumors to ENZ via down-regulating AR-V7 signaling, and DD1-Br/ENZ is a promising combination therapy for intractable CRPC.

3. Conclusions

Spurred by the discovery of exportin-1 drug selinexor, importins are emerging as attractive targets in cancer therapies. Several importins, such as importin-β1, importin-α2, and importin-α4, have been demonstrated to be involved in progression in multiple tumors27, 28, 29, 30. However, the biological function of these importins is mainly investigated by RNA interference (RNAi) methods. The lack of druggable inhibitors hinders their therapeutic potential and further clinical investigation. To the best of our knowledge, only two direct importin inhibitors, importazole and goyazensolide, have been reported before^17,31. Importazole is a weak importin-β1 inhibitor commonly used as a tool in importin-β1-related study and goyazensolide is a covalent inhibitor of importin-5 showed in vitro anti-tumor and anti-viral potentials.

In the current study, we developed a highly potent and orally available importin-β1 inhibitor DD1-Br based on previously discovered natural importin-β1 inhibitor DD1. With improved tolerability and pharmacokinetic properties, DD1-Br displayed remarkable therapeutic efficacy in advanced CRPC xenograft models both in monotherapy and combination therapy. Mechanistic study revealed that by targeting importin-β1, DD1-Br significantly blocked the nuclear transport of multiple CRPC drivers, particularly AR-V7, a main contributor to enzalutamide resistance, leading to the integral suppression of downstream oncogenic signaling. This study not only provides an invaluable lead for CRPC treatment, but also demonstrates the possibility of overcoming drug resistance in advanced CRPC via targeting importin-β1.

4. Experimental

4.1. Chemical methods and procedures

Details of general information, isolation and extraction, plant material, structural elucidation of natural new compounds, synthesis of daphnane derivatives, X-ray crystal structure analysis, and NMR data can be found in the Supporting Information.

4.2. Biological experiments

Detailed methods of cell and 3D organoid cultures, cell and 3D organoid viability assays, apoptosis assay, colony formation assay, qPCR, biotin-streptavidin pull-down assay, Western blotting analysis, siRNA transfection, lentivirus preparation, CETSA, confocal immunofluorescence microscopy experiment, molecular docking, expression of recombinant proteins, SPR assay, transcriptome RNA-seq and data analysis, pharmacokinetic studies, acute toxicity experiment, mouse models and treatments, immunohistochemistry, hematology and blood chemistry analysis, and statistics analysis can be found in the Supporting Information.

Acknowledgments

This work was supported by the Natural Science Foundation of China (Nos. 82273804, 81973195, 81872891, and 81973203), the Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai) (No. SML2021SP301, China), Open Program of Shenzhen Bay Laboratory (No. SZBL2021080601007, China), the Guangdong Natural Science Funds for Distinguished Young Scholar (No. 2019B151502016, China), Guangdong-Hong Kong-Macao research team project of the Guangdong Basic and Applied Basic Research Foundation (No. 2022B1515130008, China).

Author contributions

Sheng Yin, and Junjian Wang initiated and supervised the research. Jia-Luo Huang and Xue-Long Yan designed and performed the experiments, interpreted the data, and drafted the manuscript. Lu Gan and Huahua Gao performed partial cellular assays. Dong Huang and Shen Li extracted and purified the tested compounds. Run-Zhu Fan and Xinying Zhu were involved in partial data analysis. Hong-Wu Chen, Fang-Yu Yuan, and Gui-Hua Tang revised the manuscript. All authors read and approved the final manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

Appendix A. Supplementary data

The following are the Supplementary data to this article: