Targeting the Nuclear Import Receptor Kpnβ1 as an Anticancer Therapeutic Free

Karyopherin β1 (Kpnβ1) is a member of the Karyopherin β superfamily of nuclear transport proteins. There are over 20 members of the Karyopherin β protein family, which can function as either import or export receptors, mediating either the nuclear entry or exit of proteins (1). Kpnβ1, also known as Importin β, is a major nuclear import receptor in the cell that transports proteins containing a nuclear localization signal (NLS) through the nuclear pore complex (NPC) into the nucleus.

The classical nuclear import pathway is characterized by the recognition of the NLS on the cargo protein by the Kpnβ1 adaptor protein, Karyopherin α (Kpnα), also known as Importin α. After cargo recognition, Kpnα binds Kpnβ1 and the trimeric complex translocates into the nucleus, via Kpnβ1-interactions with the nucleoporins (Nups) that comprise the NPC. Once on the nucleoplasmic side of the NPC, the complex is dissociated by the binding of RanGTP to Kpnβ1. The binding of RanGTP triggers the dissociation of the complex due to the fact that RanGTP and Kpnα share an overlapping binding site on Kpnβ1. RanGTP binds with higher affinity to this site, displacing Kpnα, and causing the cargo protein to be released (2). Upon cargo release, RanGTP-Kpnβ1 translocates back into the cytoplasm where RanGTP is hydrolyzed to RanGDP and Kpnβ1 is released for another round of nuclear transport. While Kpnβ1 classically transports its cargo proteins in concert with an adaptor, it can function on its own, for example, in the transport of the sterol regulatory element-binding protein 2 (SREBP-2; ref. 3).

Kpnβ1 not only plays an important role in the nuclear import of cargoes, but also in the regulation of mitotic progression when the nuclear envelope breaks down. This is highlighted by the fact that ectopic overexpression of Kpnβ1 results in distinct mitotic defects (4,5). The right balance of Kpnβ1 is thus necessary for correct cell functioning. Interestingly, recent work has shown that Kpnβ1 expression is upregulated in certain cancers. We previously showed that Kpnβ1 mRNA and protein are expressed at elevated levels in cervical tumors and cervical cancer cell lines (6), and that the promoter is more highly active in cervical cancer cells due to its activation by the cell-cycle regulator, E2F (7). Smith and colleagues (2010) found that Kpnβ1 mRNA was elevated in ovarian cancer cell lines and transformed ovarian cells (8) and Kuusisto and colleagues (2011) also described increased levels of Kpnβ1 in several transformed cell lines (9). These lines of evidence suggest that the Kpnβ1 protein is associated with cellular transformation and cancer. Indeed, inhibition of Kpnβ1 protein expression in cancer cells leads to apoptosis (6), highlighting the role of this protein in cancer development. Inhibition of Kpnβ1 protein expression in noncancer cells, however, has only a minor effect on cell viability (6), pointing to a potential use for Kpnβ1 as an anticancer therapeutic target (10).

Although evidence points to its potential as an anticancer target, currently Crm1 (Exportin 1) is the only nuclear transport protein that has been investigated as a therapeutic target for cancer treatment. The well-known natural Crm1 inhibitor, leptomycin B (LMB), was shown to possess strong antitumor activity in vitro but was poorly tolerated in vivo. This was proposed to be due to its off-target effects; hence several more recent selective inhibitors of nuclear export (SINE) were developed and have shown great potential in preclinical studies. One such compound, Selinexor, is currently being evaluated in clinical trials (11). The success in targeting Crm1 highlights the potential effectiveness of targeting nuclear transport proteins for cancer therapy. Recently, a few inhibitors of nuclear import have been described but none have been tested for specificity, nor have they been tested for their anticancer effects. Small molecule peptidomimetic inhibitors of Kpnα/β-mediated transport were identified in an in vitro screen; however, these inhibitors have low potency and are not cell permeable, hence no inhibition of Kpnα/β-mediated nuclear import could be observed in vivo (12). Peptide inhibitors that bind Kpnα with a strong affinity have also been described, yet these do not inhibit Kpnβ1 directly (13). Ivermectin is a broad-spectrum antiparasitic reagent recently found to inhibit Kpnα/β, but it does not appear to block import mediated by Kpnβ1 alone (14). Karyostatin 1A (15) and importazole (16) are the first novel small molecule inhibitors of Kpnβ1 to be identified, which inhibit RanGTP binding to Kpnβ1; however, their off-target effects have not yet been examined, nor have their anticancer effects. There thus remains a need for the identification of novel effective Kpnβ1 inhibitors and for these to be tested for their effects on cancer cells.

The value of Kpnβ1 inhibition as an anticancer approach lies in the fact that the targeting of this protein results in the disruption of numerous cellular pathways critical for tumor progression. This is since hosts of proteins, critical for tumorigenesis, are reliant on Kpnβ1. Diverse cancer processes rely on Kpnβ1-mediated nuclear import. For example, a recent review highlights its role in control of the epithelial-to-mesenchymal transition (EMT; ref. 17). Kpnβ1 has been shown to be involved in the nuclear import of a number of EMT-promoting proteins, including Notch (18), Snail (19), and Smad (20). This role for Kpnβ1 in promoting EMT suggests that it is necessary for cancer cell invasion and metastasis. Indeed, the nuclear transport proteins, Ran and Crm1, have both recently been implicated as key proteins in the metastatic progression of cancer (21–23). Another role for Kpnβ1 which has recently been highlighted is its role in inflammation (24). Kpnβ1 has been shown to be a critical mediator of inflammation due to its nuclear import of many inflammatory-associated transcription factors, including NFκB (25), STAT family members (26), AP1 (27), and GATA-3(28). Cancer-related inflammation has been termed the seventh hallmark of cancer (29), and the role for Kpnβ1 in this process highlights its regulation of diverse cellular functions necessary for cancer progression.

The inhibition of Kpnβ1 represents a powerful approach to target the communication route shared by many cancer-sustaining pathways. In this study, we performed a structure-based in silico screen to identify inhibitors of Kpnβ1, and identified INI-43 (3-(1H-benzimidazol-2-yl)-1-(3-dimethylaminopropyl)pyrrolo[5,4-b]quinoxalin-2-amine) to be an effective inhibitor of Kpnβ1-mediated nuclear import with anticancer activity.

A computational screen was performed at the Molecular Modelling Facility of the James Graham Brown Cancer Centre, University of Louisville (Louisville, KY), using a rational structure-based approach, to identify compounds that bind the overlapping Ran- and Kpnα2-binding region of Kpnβ1, based on their crystal structures (PDB codes 1ibr, 2bku, and 1qgk). A library of 12,662,570 unique chemical compounds in the 2010 ZINC drug-like database (30) was screened against the target sites using the 1ibr structure, according to the methodology described in (31). Compounds were ranked according to their predicted Kpnβ1-binding affinity.

Forty-seven of the top-scoring compounds were purchased from MolPort, ChemBridge, and Enamine and screened for their ability to block nuclear import, as well as their effect on normal and cancer cell viability. Compounds were obtained in a powder form and dissolved in DMSO to a stock concentration of 100 mmol/L. Inhibitor of Nuclear Import-43, INI-43 (3-(1H-benzimidazol-2-yl)-1-(3-dimethylaminopropyl)pyrrolo[5,4-b]quinoxalin-2-amine) was obtained from Chembridge (ZINC identification no. 20547783). Cisplatin was obtained from Sigma and dissolved in 150 mmol/L NaCl. Ivermectin and importazole were obtained from Sigma and dissolved in DMSO. leptomycin B was also obtained from Sigma in methanol: water (7:3). AG1478 was obtained from Calbiochem and dissolved in DMSO.

All cell lines were maintained in DMEM containing 10 % FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin, except for MCF12A benign breast cancer cells (obtained from the ATCC), which were maintained in 50 % HAMS F12 and 50 % DMEM, supplemented with 5 % FBS (Gibco), 100 U/mL pencillin, 100 μg/mL streptomycin 20 ng/mL EGF (Gibco), 100 ng/mL Cholera Toxin (Sigma), 10 μg/mL insulin (Gibco) and 500 ng/mL hydrocortisone (Sigma), and the ovarian cancer cell lines, A2780, CP70, and OVCAR3, which were maintained in RPMI, containing 10 % FBS, 100 U/mL penicillin and 100 μg/mL streptomycin. KYSE30 and KYSE150 cells were obtained from DSMZ (32) and WHCO5 and WHCO6 oesophageal carcinoma cell lines were acquired from Dr R. Veale (33). FG0 and DMB normal skin fibroblasts were obtained from the Department of Medicine, UCT. Unless specified above, all other cell lines were obtained from the ATCC (Rockville). HeLa, SiHa, CaSki, MS751, C33A, WHCO5, SVWI38, WI38, CCD-1068SK, FG0, WHCO6, Kyse30, and MDA-MB-231 cell lines were authenticated by DNA profiling using the Cell ID System (Promega).

To assay for NFAT luciferase activity, 30,000 HeLa cells were plated per well in 24-well plates, and transfected with 50 ng GFP-NFAT plasmid (Addgene plasmid # 24219, gift of Jerry Crabtree; ref. 34), 50 ng NFAT-luciferase (Addgene plasmid # 10959, gift of Toren Finkel; ref. 35), and 5 ng pRL-TK, using 0.4 μL Genecellin Transfection Reagent (Celtic Molecular Diagnostics). The NFAT-luciferase reporter plasmid contains three tandem repeats of a 30-bp fragment of the IL2 promoter known to bind NFAT. The following day cells were treated with the various compounds and stimulated with 100 nmol/L PMA (Sigma) and 1.3 μmol/L ionomycin (Santa Cruz Biotechnology) for 1.5 or 3 hours, following which luciferase activity was measured.

To assay for AP-1 and p65 luciferase activities, HeLa cells were plated in 24-well plates, and transfected with 50 ng of an AP1 luciferase reporter (containing four copies of the AP-1–binding site; ref. 36) or 50 ng of a p65 luciferase reporter (containing five copies of the p65-binding site; Promega) and 5 ng pRL-TK (encoding Renilla luciferase, Promega), using 0.16 μL TransFectin lipid reagent (Bio-Rad). The following day cells were treated with 0.5 or 1× the IC₅₀ concentration of INI-43, importazole, or ivermectin for 24 hours. Three hours before reading luciferase activity, cells were stimulated using 500 nmol/L PMA. Luciferase activity was assayed using the Dual-Luciferase Reporter assay system (Promega), according to the manufacturer's instructions, and luciferase readings were measured using the Veritas microplate luminometer (Promega) and normalized to Renilla luciferase in the same extract.

HeLa cells were grown on glass coverslips and treated with 10 μmol/L INI-43 for 1.5 or 3 hours. For controls, equivalent volumes of DMSO were used. For p65 experiments, cells were treated with 100 nmol/L PMA for 1 hour and subsequently fixed with 4 % paraformaldehyde. Immunofluorescence analysis was performed using 1:200 p65 antibody (sc-7151X, Santa Cruz Biotechnology) and 1:300 Cy3 goat anti-rabbit secondary antibody (Jackson ImmunoResearch). For Kpnβ1 immunofluorence, a 1:100 Kpnβ1 antibody (sc-11367, Santa Cruz Biotechnology) was used. Cell nuclei were stained with 0.5 μg/mL DAPI. Images were captured using a Zeiss inverted fluorescence microscope under 100× oil immersion.

HeLa-GFP and HeLa-Kpnβ1-GFP cells were grown on glass coverslips and treated with 10 μmol/L INI-43 for 45 minutes, after which cells were fixed using 4 % paraformaldehyde. Because of the overlapping emission spectra of INI-43 and GFP, the INI-43 and GFP signals were separated using spectral unmixing, using a Zeiss LSM 510 Meta confocal microscope.

Nuclear and cytoplasmic protein fractionation was performed using the NE-PER Nuclear Protein Extraction Kit (ThermoScientific), according to the manufacturer's instructions. HeLa and CaSki cells were either treated with different concentrations of INI-43 for 3 hours, or transfected with 20 nmol/L control siRNA (sc-37007, Santa Cruz Biotechnology) or Kpnβ1 siRNA (sc-35736, Santa Cruz Biotechnology) for 96 hours, before protein extraction. Western blot analysis was performed using a rabbit anti-NFYA (H-209) antibody (sc-10779, Santa Cruz Biotechnology). An anti-β-tubulin (H-235) antibody (sc-9104, Santa Cruz Biotechnology) was used to control for cytoplasmic protein loading and a rabbit anti-TATA-box–binding protein (TBP; N-12) antibody (sc-204, Santa Cruz Biotechnology) to control for nuclear protein loading. Densitometry was performed using ImageJ.

Cells were plated in 96-well plates and treated with varying concentrations of drug for 48 hours, after which MTT (Sigma) was added, and crystals solubilized 4 hours later using Solubilisation Reagent (10 % SLS in 0.01 mol/L HCl). Absorbencies were measured at 595 nm the following day using a BioTek microplate spectrophotometer and IC₅₀ curves were generated using GraphPad Prism.

The plasmid for overexpression of Kpnβ1 (pEFIRES-Kpnβ1-GFP) was generated by inserting SacI- and NotI-digested human Kpnβ1-GFP, released from pEGFP-Kpnβ1 (kind gift from Patrizia Lavia, Institute of Molecular Biology and Pathology, CNR National Research Council, Rome, Italy; ref. 4), into the pEFIRES plasmid (kind gift from Yosef Shaul, Weizmann Institute of Science, Israel; refs. 37,38). pEFIRES-expressing GFP only was used as a control. HeLa cells were transfected with Genecellin Transfection Reagent (Celtic Molecular Diagnostics) and transfected cells were selected using Puromycin (Calbiochem, Merck). Pools of stably transfected cells were maintained in 1.5 μg/mL Puromycin.

To assay for cell proliferation after INI-43 treatment, cells were plated in 96-well plates and treated with 5 or 10 μmol/L INI-43, after which cell proliferation was monitored every 24 hours for 4 days, using MTT.

Cells were plated in 60 mm dishes and treated with the appropriate concentration of INI-43 for 3, 6, or 24 hours. Cells were harvested and fixed in 100 % ethanol, after which they were stained with propidium iodide and the cell-cycle profiles analyzed using a BD Accuri C6 flow cytometer (BD Biosciences). Data analysis was performed using ModFit 3.3 software.

HeLa cells were treated with either the IC₅₀ or 1.5 × IC₅₀ concentration of INI-43, importazole, ivermectin, or leptomycin B for varying time points, following which cells and cell floaters were harvested using RIPA buffer (10 mmol/L Tris-Cl, pH 7.4, 150 mmol/L NaCl, 1 % sodium deoxycholate, 0.1 % SDS, 1 % Triton X-100, 1× complete protease inhibitor cocktail (Roche) and Western blot analysis performed, using a rabbit anti-PARP1/2 (H-250) antibody (sc-7150).

Cells were grown in 10 cm plates and treated with 10 μmol/L INI-43 for various time points. Cells were lysed in subcellular fractionation buffer (250 mmol/L sucrose, 20 mmol/L HEPES, 10 mmol/L KCl, 1.5 mmol/L MgCl2, 1 mmol/L EDTA, 1 mmol/L EGTA, 1 mmol/L DTT and 1× complete protease inhibitor cocktail; Roche), passed through a 25-G needle and incubated on ice for 20 minutes. The nuclear pellet was obtained after centrifugation at 720 g for 5 minutes. The supernatant was further centrifuged at 10,000 g and the resulting supernatant used as the cytoplasmic fraction. The pellet containing the mitochondrial fraction was washed in subcellular fractionation buffer and passed through a 25-G needle. The mitochondrial fraction was pelleted at 10,000 g and resuspended in RIPA buffer. Western blot analysis was performed using a mouse anti-Cytochrome C antibody (BD Pharmingen) and an anti-Pereroxiredoxin-3 (PRDX-3; P1247, Sigma) antibody was used as mitochondrial protein loading control.

HeLa cells were plated in 96-well plates and treated with 10 μmol/L INI-43. Caspase-3/7 activity was monitored at various time points using the Promega Caspase-Glo 3/7 assay, according to the manufacturer's instructions. Luminescence was measured using the Veritas microplate luminometer (Promega) and normalized to OD595 readings of MTT experiments performed in parallel.

A toxicology study was performed and revealed that Athymic nude mice (n = 6 mice/group) tolerated various doses of INI-43 (1 mg/kg, 10 mg/kg, and 50 mg/kg). No side effects were experience at all the doses examined and 50 mg/kg was chosen as the dose for further experiments. The mice livers were examined at the end of the treatment and no histologic pathology was reported. This dose is also in line with that used for other chemotherapeutic drugs according to the literature. The treatment schedule is shown in Supplementary Fig. S3C and S3D. WHCO6 and CaSki cells were harvested and resuspended in PBS. For tumor inoculation, WHCO6 or CaSki cancer cells (5×10⁶ per mouse) were s.c. implanted into the hind flanks of female nude mice. Once tumors had reached a palpable size (for CaSki the mice were separated into two groups as the tumors took varying lengths of time to reach the appropriate size), drug treatment was initiated, where tumor-bearing mice were randomized and dosed i.p. with either vehicle (DMSO) or INI-43 (50 mg/kg), every 2 to 3 days for 3 (for WHCO6) or 4 (for CaSki) weeks. CaSki tumors grew slower hence the longer treatment schedule. Of note, 50 mg/kg had been chosen as the MTD in a toxicology study. Tumors were measured on day 0 (first treatment) and every 3 to 4 days thereafter, using calipers, and tumor volume estimated using the following formula: volume = (length × width × width)/2. Mice were sacrificed 3 (for WHCO6) or 4 (for CaSki) weeks after the commencement of drug treatment. Day 21 measurements were excluded from the WHCO6 mouse study due to one mouse having to be sacrificed early on day 20. Animal ethics approval was obtained from the Faculty of Health Sciences Animal Ethics Committee, University of Cape Town, Cape Town, South Africa (reference number 012/009).

For all data comparisons, the Student t test was performed using Microsoft Excel. A P value of < 0.05 was considered statistically significant.

Earlier studies performed in our laboratory showed that the inhibition of Kpnβ1 expression in cervical cancer cell lines using siRNA resulted in cell death via apoptosis (6). This effect appeared more specific to the cervical cancer cells compared with noncancer cells. To expand these findings, cancer, transformed and noncancer cell lines of different tissue origins were transfected with Kpnβ1 siRNA and cell viability monitored 5 days after transfection. Figure 1 shows that Kpnβ1 inhibition results in a significant reduction in cell number in cervical cancer, esophageal cancer, and transformed cell lines, while noncancer cells were relatively unaffected (Fig. 1A). Kpnβ1 knockdown in the cancer and noncancer cells was confirmed by Western blot analysis (Fig. 1B). These results suggest that Kpnβ1 expression and activity are necessary for the survival of cancer cells.

The killing of cancer cells after Kpnβ1 inhibition with siRNA presented a unique opportunity to identify Kpnβ1 small molecule inhibitors with potential anticancer activity. To identify Kpnβ1 small molecule inhibitors, a structure-based virtual screen was performed. Virtual screening has yielded many new lead compounds for different protein targets and is a process whereby libraries of small molecules are screened to identify those structures which are most likely to bind a drug target. This screen aimed to identify compounds that can bind the overlapping Ran- and Kpnα2-binding region of Kpnβ1 (corresponding to amino acids 331–363). This region of Kpnβ1 was previously identified to be critical for its function, as its deletion resulted in an inability of Kpnβ1 to transport an NLS-HSA cargo into the nucleus (2). Seventy-four potential “hit” compounds were identified and numbered according to their predicted Kpnβ1-binding affinity, and 47 of the available top-scoring compounds were purchased for further analysis. As the intention was to find an inhibitor of nuclear import with anticancer activity, the compounds were first tested for their effect on cancer cell viability, using the cervical cancer cell line CaSki as a model cell line. Cell viability as measured using the MTT assay revealed that 16 compounds displayed antiproliferative IC₅₀ values of less than 50 μmol/L (Supplementary Table S1), which is within the range of the IC₅₀ obtained using current chemotherapeutic drugs, for example, cisplatin (Supplementary Table S1).

The 16 compounds with cancer cell-killing ability were next tested for the ability to block nuclear import using a luminescence-based screening assay. This assay was based on the premise that the transcription factor NFAT shuttles between the nucleus and the cytoplasm in a Kpnβ1-dependent (39) and Crm1-dependent manner (40). Its nuclear-cytoplasmic transport is regulated via calcium exposure (41), where NFAT is predominantly cytoplasmic; however, an increase in intracellular calcium levels leads to its nuclear accumulation. Soderholm and colleagues (2011) showed by fluorescence microscopy that inhibiting nuclear import via Kpnβ1 prevented the nuclear accumulation of NFAT-GFP in response to stimulation with the calcium ionophore, ionomycin (16). We assayed for nuclear NFAT following cotreatment with ionomycin and the compounds identified in the in silico screen. This was performed using an NFAT-luciferase reporter that is activated by nuclear NFAT. The phorbol ester PMA was used to costimulate NFAT, as it has been shown previously that PMA and ionomycin elicit synergistic stimulation of NFAT luciferase activity (42). Results showed that stimulation of NFAT-transfected HeLa cells with PMA and ionomycin resulted in a significant increase in activation of the NFAT reporter, due to the nuclear import of NFAT (Fig. 2B). After screening the 16 compounds that showed cell killing effects, 3-(1H-benzimidazol-2-yl)-1-(3-dimethylaminopropyl)pyrrolo[5,4-b]quinoxalin-2-amine (ZINC identification no. 20547783) emerged as a potential inhibitor of nuclear import, hereafter referred to as Inhibitor of Nuclear Import-43 (INI-43). The molecular structure of INI-43 is shown in Fig. 2A. Treatment with INI-43 significantly diminished NFAT activation in a dose-dependent manner (Fig. 2B). INI-43 was taken up within minutes into the cells and inhibition of NFAT activity was achieved after 1.5 hour at a concentration of 10 and 15 μmol/L (approximately 1× and 1.5× its IC₅₀ concentration). In comparison, recently identified inhibitors of Karyopherin α/β-mediated nuclear transport, ivermectin and importazole (14,43), were unable to inhibit NFAT activity at the same time point and concentrations (Fig. 2B), implicating INI-43 as a more potent inhibitor of nuclear import. After treatment for 3 hours, however, importazole and ivermectin, like INI-43, were able to significantly inhibit NFAT activity at their respective IC₅₀ or 1.5 × IC50 concentrations (IC₅₀ values for importazole and ivermectin in HeLa cells were calculated to be 25.3 and 17.8 μmol/L, respectively; higher than the IC₅₀ for INI-43 in HeLa cells of 9.3 μmol/L; Fig. 2C–E). As a negative control, NFAT activation was measured after a 3-hour treatment with the EGFR inhibitor, AG1478, at its IC₅₀ and 1.5 × IC₅₀ concentrations, and no change in NFAT activation was observed (Fig. 2F).

As an additional indicator that INI-43 was able to inhibit nuclear import, it was next investigated whether INI-43 could interfere with the nuclear translocation of p65 in response to PMA treatment, as p65 has previously been shown to rely on Kpnβ1 for its nuclear import (25). HeLa cells were treated with INI-43 and stimulated with PMA, following which p65 localization was determined by immunofluorescence. In untreated cells, p65 showed a predominantly cytoplasmic localization and stimulation with PMA resulted in a clear influx of p65 into the nucleus (Fig. 3A). Treatment with INI-43 in the presence of PMA noticeably blocked the nuclear entry of p65 at 1.5 and 3 hour time points (Fig. 3A). The cells were categorized as having “predominantly cytoplasmic” or “predominantly nuclear” p65 and the ratio of nuclear to cytoplasmic p65 plotted. INI-43 was shown to markedly influence p65 nuclear translocation (Fig. 3B), supporting its potential as an inhibitor of nuclear import via Kpnβ1.

Luciferase assays were next performed to assay for p65 nuclear activity after PMA stimulation and INI-43 treatment, and PMA-induced p65 nuclear activity was markedly reduced in response to overnight INI-43 treatment at 0.5X IC50 and IX IC50 concentrations, as indicated by significantly reduced luciferase activity (Fig. 3C). The inhibition achieved was similar to that obtained using importazole and ivermectin at concentrations relative to their respective IC50s. AP-1 luciferase activity was also measured in response to PMA stimulation and INI-43 treatment, since this transcription factor has also been shown to rely on Kpnβ1 for nuclear import (25,27). AP-1 luciferase activity was also found to be significantly decreased after INI-43, importazole and ivermectin treatment (Fig. 3D).

The ability of INI-43 to block the nuclear import function of Kpnβ1 was tested further by investigating the localization of another Kpnβ1 cargo, NFY-A (44), by subcellular fractionation and Western blot analysis. Cells were treated with INI-43, following which nuclear and cytoplasmic protein fractions were isolated. Western blot analysis was then performed to determine the localization of NFY-A, and NFY-A levels were found to decrease in the nuclear fraction after 3-hour INI-43 treatment (Fig. 3E). To verify that the banding pattern observed after INI-43 treatment was as expected after inhibition of Kpnβ1, Kpnβ1 was silenced using siRNA, and nuclear and cytoplasmic protein fractions isolated for Western blot analysis. Similar results as those obtained with INI-43 treatment were found where the amounts of NFY-A were decreased in the nucleus after Kpnβ1 inhibition using siRNA (Fig. 3F). In both cases of INI-43 treatment and Kpnβ1 siRNA treatment, NFY-A could not be observed in the cytoplasm. It is possible that the cytoplasmic protein fraction was not concentrated enough to detect NFY-A protein.

To address the question of whether the inhibitory effect of INI-43 on nuclear import associated with Kpnβ1, we investigated the localization of Kpnβ1 itself in control and INI-43–treated cells. We found that treatment of HeLa cells with INI-43 resulted in altered subcellular localization of endogenous Kpnβ1. In untreated cells, endogenous Kpnβ1 predominantly localized to the nucleus (Fig. 4A). Treatment with INI-43 resulted in a change in Kpnβ1 subcellular localization to cytoplasmic and perinuclear within 45 minutes of exposure to the compound, suggesting that the compound acts via interfering with Kpnβ1. To further support this result, a HeLa cell line stably overexpressing Kpnβ1-GFP was generated, termed HeLa-Kpnβ1-GFP, as well as a control cell line overexpressing EGFP alone, termed HeLa-GFP. The HeLa-Kpnβ1-GFP cells expressed the Kpnβ1-GFP at equivalent levels to that of endogenous Kpnβ1. Therefore, approximately twice as much Kpnβ1 was expressed in the HeLa-Kpnβ1-GFP cells than the control cells (HeLa and HeLa-GFP; Fig. 4B). A GFP Western blot analysis confirmed the overexpression of EGFP in the control cell line (Fig. 4C). The localization of GFP in these cells (HeLa-GFP and HeLa-Kpnβ1-GFP) was analyzed after INI-43 treatment for 45 minutes. First, the INI-43 fluorescence signal had to be excluded from the GFP signal, due to their overlapping emission spectra. Thereafter, a shift in Kpnβ1-GFP localization was noted upon INI-43 treatment, where reduced nuclear Kpnβ1-GFP fluorescence was observed in response to INI-43, further supporting the fact that INI-43 interferes with Kpnβ1 nuclear entry (Fig. 4D). EGFP localization did not change upon INI-43 treatment, as expected.

We then postulated that if the change in subcellular localization of p65 was as a result of perturbing Kpnβ1 function, overexpression of Kpnβ1 should rescue the inhibitory effects of INI-43. We therefore next performed a rescue experiment using the HeLa-Kpnβ1-GFP cells. Immunofluorescent analysis of the Kpnβ1 cargo protein p65 showed a predominant cytoplasmic localization in untreated HeLa, HeLa-GFP, and HeLa-Kpnβ1-GFP cells (Fig. 4E). Although stimulation of these cells with PMA resulted in the nuclear translocation of p65 (Fig. 4F), a pretreatment with INI-43 blocked the nuclear entry of p65 in HeLa and HeLa-GFP control cells (Fig. 4G). However, in HeLa-Kpnβ1-GFP cells nuclear entry of p65 was restored (Fig. 4G, bottom). Quantification of these results is shown in Fig. 4H and 4I. These results suggest that the inhibitory effects of INI-43 on nuclear import can be rescued by Kpnβ1 overexpression, providing evidence that INI-43 acts via Kpnβ1.

As the focus of this study was to identify nuclear import inhibitor(s) with anticancer activity, the cell killing effect of INI-43 was tested using a panel of cell lines of various tissue origins. INI-43 was found to kill cancer cell lines of cervical, esophageal, ovarian, and breast origin with an IC₅₀ or approximately 10 μmol/L (Supplementary Table S2). All dose–response curves had sharp slopes, quantified by the Hill slope (Supplementary Table S2 and Supplementary Fig. S1). INI-43 also showed elevated toxicity in the cancer cell lines in comparison with the noncancer lines. Specifically, it showed an approximate 2- to 3-fold selectivity toward cancer cell lines over noncancer mesenchymal cells.

To confirm that INI-43 was more effective at killing cancer cells compared with noncancer cells, representative cancer (cervical and oesophageal) and noncancer cell lines, were grown in the presence of INI-43, and cell proliferation measured over a period of 5 days. Proliferation assays showed that INI-43 had a dramatic impact on cancer cell viability in less than 24 hours, and within 48 to 72 hours complete cell death was observed at a concentration of 10 μmol/L INI-43 (Fig. 5A–D). This effect was sustained over the assay time period and there was no population of cells that recovered from the treatment. Noncancer fibroblast cells, however, proliferated relatively normally in the presence of 10 μmol/L INI-43 (Fig. 5E and F). This result suggests a differential sensitivity of cancer and noncancer cells to INI-43, where at 10 μmol/L INI-43 treatment, the cancer cells underwent 100% cell death but the noncancer cells were largely unaffected.

Anchorage-independent proliferation of cancer cells was next measured in response to INI-43 treatment. Cells were grown on polyheme-coated plates and cell proliferation monitored 9 days after treatment. Both cervical and esophageal cancer cell colony formation was inhibited in the presence of 10 μmol/L INI-43 (Fig. 5G), confirming that INI-43 inhibits both anchorage-dependent and -independent proliferation of cancer cells.

To investigate whether the cell death observed after INI-43 treatment was as a result of Kpnβ1 inhibition, we performed a rescue experiment using the HeLa-Kpnβ1-GFP cell line (Fig. 4B). Although Kpnβ1-GFP overexpression resulted in a modest but significant increase in the IC50 for INI-43 (from 8.33 in HeLa-GFP cells to 9.6 in HeLa- Kpnβ1-GFP cells), cells overexpressing Kpnβ1 were considerably more resistant to INI-43 treatment at concentrations of 5 and 10 μmol/L INI-43, and showed significantly more viable cells compared with controls (Fig. 5H). The rescue experiment therefore confirmed that INI-43, in part, exerted its cell killing effects by targeting Kpnβ1, as overexpression of Kpnβ1 significantly rescued cell viability.

The mechanism of INI-43–induced cell death was next investigated. Cell-cycle analysis showed that INI-43 treatment resulted in an increase in the population of cells in the G₂–M stage of the cell cycle at 6 hours after treatment (Fig. 6A). This correlates with the important role played by Kpnβ1 in the regulation of mitosis. The sub-G₁ population of cells was significantly increased after 10 μmol/L INI-43 treatment for 24 hours and as little as 6 hours with 15 μmol/L INI-43 (Supplementary Fig. S2). The induction of a cell-cycle block and accumulation of a sub-G₁ cell population suggested that apoptosis might be activated; hence the cleavage of PARP-1 was measured by Western blot analysis as an indicator of apoptosis. INI-43 treatment (at 10 μmol/L) resulted in a time-dependent increase in cleaved PARP-1, from as early as 3 hours, but more visibly at 6 hours, after treatment (Fig. 6B). Treatment for 24 hours with 15 μmol/L resulted in almost complete cell death, hence degradation of the loading control, GAPDH. Treatment of HeLa cells with the IC₅₀ and 1.5 × IC₅₀ concentrations of importazole, ivermectin, and the Crm1 inhibitor leptomycin B, similarly resulted in the accumulation of cleaved PARP-1, although this was only visible at later time points (Fig. 6C). This reveals that INI-43 works like other nuclear transport inhibitors, whereby it results in the induction of cancer cell apoptosis. To investigate whether INI-43 treatment resulted in the activation of apoptosis associated with the intrinsic mitochondrial pathway, Cytochrome C subcellular localization was investigated, as Cytochrome C release from the mitochondria is a characteristic feature of apoptosis mediated by the intrinsic mitochondrial pathway. Our results showed that INI-43 treatment resulted in a decrease in mitochondrial Cytochrome C levels, in a time-dependent manner from 6 hours posttreatment (Fig. 6D), suggesting that the intrinsic mitochondrial pathway is activated upon INI-43 treatment and responsible for cancer cell death. Finally, caspase-3/7 activity was measured after INI-43 treatment of HeLa cells. A significant increase in caspase-3/7 activity was observed both 6 and 24 hours after 10 μmol/L INI-43 treatment (Fig. 6E), correlating with the release of Cytochrome C and cleavage of PARP-1 at these time points, and independently confirming the induction of apoptosis after INI-43 treatment.

Because INI-43 displayed cytotoxicity in vitro, we next investigated its antitumor activity in vivo using mouse xenografts models. A liver microsome assay revealed that INI-43 displayed good metabolic stability, with a degradation half-life of over 100 minutes. A toxicology study was performed and revealed that mice tolerated various doses of INI-43, with 50 mg/kg being the MTD tested. A xenograft study was thus conducted where mice were inoculated with cancer cells (WHCO6 esophageal cancer or CaSki cervical cancer cells) and, once tumors had reached a palpable size, were treated i.p. with vehicle (DMSO) or INI-43 (50 mg/kg) every 2 to 3 days and tumor size monitored for 3 to 4 weeks. INI-43 treatment was found to significantly inhibit esophageal and cervical tumor growth (Fig. 6Fa and b). An analysis of the gain in tumor size relative to the starting tumor size is shown in Fig. 6G. The data show that INI-43–treated tumors showed a significantly lower gain in tumor size compared with controls, suggesting that INI-43 treatment decreases tumor growth. Body mass did not change over the INI-43 treatment time courses and mice appeared healthy over the duration of the experiments, suggesting that no significant adverse side-effects were experienced (Supplementary Fig. S3).

The targeting of nuclear transport presents a novel anticancer approach that has received attention of late. In this study, we have identified a novel inhibitor of nuclear import, 3-(1H-benzimidazol-2-yl)-1-(3-dimethylaminopropyl)pyrrolo[5,4-b]quinoxalin-2-amine, termed INI-43, and shown that this compound displays anticancer activity in vitro and in vivo. Few other inhibitors of Kpnβ1-mediated nuclear import have recently been identified; however, they are not widely available and have not been extensively tested. Moreover, no inhibitor of nuclear import has as yet been tested for its anticancer effects.

INI-43 was identified as a potential inhibitor of nuclear import by analyzing the localization of Kpnβ1 and Kpnβ1-dependent cargo proteins after INI-43 treatment. INI-43 interfered with the nuclear localization of Kpnβ1 as well as that of its cargo transcription factors, NFY, AP-1, p65, and NFAT, all of which displayed diminished nuclear import upon INI-43 treatment. The inhibition of their nuclear import and activity after Kpnβ1 inhibition highlights the extent to which nuclear import inhibition can affect cellular processes crucial for tumorigenesis.

Because INI-43 was found to inhibit nuclear import, its cell-killing effect was analyzed and it was found to kill cancer cells of different origins at a concentration of approximately 10 μmol/L, which is within the in vitro cell killing concentration range of current chemotherapeutic drugs, for example, cisplatin. The overexpression of Kpnβ1 in HeLa cells could, in part, rescue the cell killing effects of INI-43, suggesting that INI-43 acts via Kpnβ1 inhibition. In addition, it was found that cancer cells were more sensitive to INI-43 treatment than the noncancer cells tested. There is evidence in the literature that nuclear export inhibitors targeting Crm1 exhibit potent anticancer activity but only minimal effects on normal cells (45–47), suggesting that inhibition of other Karyopherin β family members might result in similar effects. Moreover, the inhibition of Kpnβ1 using siRNA similarly results in cancer cell apoptosis, while the inhibition of Kpnβ1 in noncancer cells has only a minor effect on cell biology (6,48). As the goal of cancer therapy is to promote the death of cancer cells without causing too much damage to normal cells this selectivity is advantageous and suggests that the development of anti-nuclear import drugs may have therapeutic potential. The selectivity may derive from the increased nuclear transport rates in cancer cells compared with normal cells, and thus increased reliance on the nuclear transport machinery (9). Alternatively, it is also possible that the selectivity toward cancer cells is due to the cancer cells being “primed” to undergo apoptosis (closer in proximity to the apoptotic threshold) compared with normal cells (49).

We found that INI-43 treatment of cervical cancer cells resulted in a G₂–M cell-cycle arrest. Kpnβ1 plays a vital role in mitosis, where it plays a role in spindle regulation and nuclear envelope and pore assembly (50). Both Kpnβ1 inhibition with siRNA (48) and Kpnβ1 overexpression (4,5) have been shown to result in mitotic defects and mitotic arrest, hence the induction of a mitotic block after INI-43 treatment supports INI-43 as an agent that interferes with cell-cycle progression in the G₂–M phase where Kpnβ1 has previously been reported to play a role.

The inhibition of nuclear import was associated with an increase in multiple markers of apoptosis in cancer cells. These included the cleavage of PARP-1, the release of Cytochrome C from the mitochondria, and the induction of caspase-3/7 activity. Cytochrome C release is a feature of the intrinsic mitochondria-mediated apoptosis pathway, suggesting that INI-43 treatment results in activation of the intrinsic apoptotic pathway. The intrinsic apoptosis is similarly induced after Crm1 inhibition in leukemic cells (45) and Kpnβ1 inhibition in cervical cancer cells using siRNA (48). The induction of apoptosis is likely induced by the mislocalization of key proteins after nuclear import inhibition and thus the disruption of cellular homeostasis.

This study has identified INI-43 as a useful tool compound and possibly a good lead for future development of Kpnβ1 inhibitors. In addition, the evidence provided here, showing that Kpnβ1 functioning is critical for cancer cell survival highlights Kpnβ1 as an attractive target for future cancer therapies.

No potential conflicts of interest were disclosed.

Conception and design: J.O. Trent, V.D. Leaner

Development of methodology: P.J. van der Watt, A. Chi, T. Stelma, L. Angus, J.O. Trent

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P.J. van der Watt, A. Chi, T. Stelma, C. Stowell, E. Strydom, S. Carden, D. Lang, W. Wei, V.D. Leaner

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P.J. van der Watt, A. Chi, T. Stelma, C. Stowell, S. Carden, D. Lang

Writing, review, and/or revision of the manuscript: P.J. van der Watt, A. Chi, D. Lang, W. Wei, M.J. Birrer, J.O. Trent, V.D. Leaner

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Hadley, D. Lang

Study supervision: V.D. Leaner

The authors thank Hajira Guzgay and Cecilia Hilmer for technical assistance with proliferation assays, Susan Cooper for assistance with immunofluorescent analysis, and Dr. Mathew Njoroge for assistance with liver microsome assays.

This work was supported by grants obtained by Virna D. Leaner from the South African Medical Research Council, the National Research Foundation, the Cancer Association of South Africa (CANSA), and the University of Cape Town.