The Karyopherin proteins, Crm1 and Karyopherin β1, are overexpressed in cervical cancer and are critical for cancer cell survival and proliferation

Abstract

The Karyopherin proteins are involved in nucleo-cytoplasmic trafficking and are critical for protein and RNA subcellular localization. Recent studies suggest they are important in nuclear envelope component assembly, mitosis and replication. Since these are all critical cellular functions, alterations in the expression of the Karyopherins may have an impact on the biology of cancer cells. In this study, we examined the expression of the Karyopherins, Crm1, Karyopherin β1 (Kpnβ1) and Karyopherin α2 (Kpnα2), in cervical tissue and cell lines. The functional significance of these proteins to cancer cells was investigated using individual siRNAs to inhibit their expression. Microarrays, quantitative RT-PCR and immunofluorescence revealed significantly higher expression of Crm1, Kpnβ1 and Kpnα2 in cervical cancer compared to normal tissue. Expression levels were similarly elevated in cervical cancer cell lines compared to normal cells, and in transformed epithelial and fibroblast cells. Inhibition of Crm1 and Kpnβ1 in cancer cells significantly reduced cell proliferation, while Kpnα2 inhibition had no effect. Noncancer cells were unaffected by the inhibition of Crm1 and Kpnβ1. The reduction in proliferation of cancer cells was associated with an increase in a subG1 population by cell cycle analysis and Caspase-3/7 assays revealed increased apoptosis. Crm1 and Kpnβ1 siRNA-induced apoptosis was accompanied by an increase in the levels of growth inhibitory proteins, p53, p27, p21 and p18. Our results demonstrate that Crm1, Kpnβ1 and Kpnα2 are overexpressed in cervical cancer and that inhibiting the expression of Crm1 and Kpnβ1, not Kpnα2, induces cancer cell death, making Crm1 and Kpnβ1 promising candidates as both biomarkers and potential anticancer therapeutic targets.

Cervical cancer is the second most common cancer among women worldwide,¹ with nearly 80% of cases occurring in developing countries.² The primary risk factor in the development of the disease is infection with the Human Papillomavirus (HPV),^3,4 and more than 90% of cervical cancers carry high-risk HPV DNA.^4,5 The HPV E6 and E7 oncoproteins are responsible for cancer development, and evidence has shown that they alone are sufficient to immortalize human foreskin keratinocytes.⁶ Their continued expression is essential for maintaining the transformed state.⁶ HPV E6 and E7 promote cellular transformation by binding to and blocking the functions of the cell cycle regulatory proteins, p53 and pRb, respectively.^7,8 Prophylactic vaccines against low-risk (HPV6, 11) and high-risk (HPV16 and 18) HPV types have recently been developed.⁹ They rely on the vaccination of women before exposure to the virus; hence, their benefit to women already infected with HPV is still unclear, as well as their benefit to women infected with HPV types other than HPV6, 11, 16 and 18.

Although the high-risk HPV proteins are the causative agents behind cervical cancer development, they initiate a cascade of oncogenic events within the cell, which ultimately drive its transformation. For example, the cell develops an increased telomerase activity, which results in the deregulation of many important cellular genes.^10,11 The progression of cervical cancer is thus a complex biological process, accompanied by many genetic alterations. We used cDNA microarray analysis to profile normal and cancer tissue of the cervix, to identify potential cancer biomarkers and genes functionally relevant to cervical tumorigenesis. Amongst the genes that were found to be differentially expressed in the cervical cancer tissues were members of the Karyopherin superfamily of nuclear transport proteins. Specifically, we identified increased expression of Crm1 (Exportin1), Karyopherin β1 (Importin β1) and Karyopherin α2 (Importin α1) in cervical cancer tissue compared to normal cervical epithelium.

The Karyopherin proteins are involved in the shuttling of cargo proteins, and certain RNAs, across the nuclear pore complex into and out of the cell nucleus. They have also been shown to have a key role in transit through the cell cycle, mitosis and replication.¹² Crm1 is a Karyopherin protein involved in nuclear export. It recognizes the leucine-rich nuclear export signal (NES) of cargo proteins that require shuttling out of the nucleus, and carries them into the cytoplasm.¹³ Nuclear import of proteins is mediated by the recognition of a nuclear localization signal (NLS) on the cargo protein by Karyopherin β1 (Kpnβ1), which subsequently binds the cargo and transports it into the cell nucleus.¹⁴ Kpnβ1 can import proteins directly, or, in the case of the classical nuclear import pathway, an adaptor molecule, of the Karyopherin α family, is required to bridge the interaction between Kpnβ1 and its cargo. The concerted action of Kpnβ1 and Karyopherin α2 (Kpnα2) is necessary for the nuclear import of proteins containing a classical NLS.¹⁴ Nuclear transport also requires Ran, a small Ras-like GTPase involved in both import and export processes.¹⁵ In the case of cervical cancer cells, HPV proteins utilize the nuclear transport proteins for transport into and out of the host cell nucleus, and the nuclear transport proteins hence play a role in promoting viral pathogenesis. HPV capsid proteins, L1 and L2, enter the nucleus via interaction with Karyopherins β1 and α2, amongst other importin proteins, and it is suggested that the L2 protein facilitates nuclear import of the viral DNA.^16-18 Once in the nucleus the viral DNA and capsid proteins can be assembled into virion particles. HPV E6 is also known to enter the nucleus via interaction with various Karyopherin proteins, whereafter it can interact with nuclear transcription factors.¹⁹ While several studies have focused on understanding the import strategies of HPV proteins, less is known regarding their interaction with the export receptor, Crm1. However, it has been reported that the HPV E2 protein uses Crm1 for nuclear export,²⁰ and that high-risk HPV E6 proteins require Crm1 function for their role in the degradation of p53.²¹

In this study, we focus on the analysis of the Karyopherin proteins in cervical cancer patient specimens and cell lines. This study provides evidence for a role of these proteins in contributing to the cancer phenotype. Our data suggests that while Crm1, Kpnβ1 and Kpnα2 are all overexpressed in cancer, Crm1 and Kpnβ1 in particular have functional significance to cancer biology.

Material and methods

Tissue specimens and histology of cervical carcinomas

Twenty-six cervical specimens, including sixteen cervical cancer biopsies and 10 normal cervical specimens, were collected from patients at Groote Schuur Hospital (South Africa). Tissue specimens were obtained with patient consent and the Research Ethics Committee of the University of Cape Town approved all aspects of the project (REC REF153/2004). Normal specimens were collected from women undergoing hysterectomies for reasons other than cervical cancer. Approximately 2 mm of the epithelial cell layer at the transformation zone of the normal specimens was dissected by a pathologist. Patients’ ages ranged from 29 to 76 years. All tissue samples were snap-frozen in liquid nitrogen and stored at −70°C. Histological examination of the cancers showed 15 squamous cell carcinomas and 1 small cell carcinoma (CD56 and synaptophysin-positive). Fourteen of the squamous cell carcinomas were moderately differentiated, with the balance being poorly differentiated. There was keratinization in only 5 cases, which was focal in nature. The HPV status of the patient specimens was determined by PCR amplification of the HPV gene L1, and hybridizing the PCR product to an HPV chip (LCD-Array HPV2.5, Chipron, Germany) to determine the HPV type present in each sample (according to the manufacturer’s protocols). The albumin gene was also amplified, to determine the integrity of the DNA. All normal specimens positive for the albumin gene were negative for HPV L1, while all cancer specimens positive for albumin were positive for HPV L1. This correlates with published data that suggests that >90% of cervical cancer biopsies carry HPV DNA. HPV16 was the most common type with 70% of the samples positive for HPV16, of which 2 patients were positive for both HPV16 and HPV45. HPV45 and HPV33 were detected in 20 and 10% of the samples, respectively.

Cell lines and cell culture

The following cell lines were obtained from the American Type Culture Collection (ATCC): cervical carcinoma cell lines, HeLa (HPV18), SiHa (HPV16), CaSki (HPV16 and 18), ME180 (similar to HPV39), MS751 (HPV18) and C33A (HPV-negative). Cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with penicillin (100 U/ml), streptomycin (100 μg/ml) and 10% fetal calf serum (Gibco, Paisley, Scotland). Primary cervical epithelial cells (HCX) and an HPV16 E6/E7-transformed cervical epithelial cell line (HPV16E6/E7-HCX) were obtained from Dr. C. Baker, NIH, USA.²² HCX and HPV16E6/E7-HCX cells were maintained in keratinocyte growth medium supplemented with EGF (10 ng/ml) and pituitary extract (50 μg/ml). E6/E7-transformed HCX were grown in the presence of 50 μg/ml G418. Because of difficulties encountered with the culturing of primary HCX cells, immortalized cell lines of fibroblast origin were included in the study. Normal lung fibroblasts, WI38, its transformed counterpart, SVWI38, and normal breast skin fibroblasts, CCD-1068SK, were also obtained from ATCC, while FG_o normal skin fibroblast were obtained from the Department of Medicine, UCT, and transformed fibroblasts, CT1, as referred to in Ref. 23. Normal hTERT-immortalized human oesophageal keratinocytes, EPC2-hTERT, were obtained as a gift from Prof. A. K. Rustgi (University of Pennsylvania, Philadelphia, USA). These cells were maintained in keratinocyte growth medium supplemented with 1 ng/ml EGF and 50 μg/ml pituitary extract. All cells were cultured in 95% air and 5% CO₂ at 37°C.

RNA isolation, amplification and microarray analysis

RNA was isolated from the patient tissues using Trizol reagent (Invitrogen, Rockville, MD, USA), according to the manufacturer’s protocols. RNA was amplified using the Eberwine RNA amplification procedure^23,24 and used to prepare cDNA probes as described previously.²⁵ Human reference RNA (Stratagene, La Jolla, Ca, USA) served as a control on all of the arrays. Amplified human reference RNA was labeled with Cy3-dUTP and amplified RNA from tissue specimens was labeled with Cy5-dUTP. Labeled reference cDNA was combined with labeled patient cDNA in a 40 μl reaction containing 2 μg human COT1 DNA (Gibco), 16 μg polyA (Sigma, St. Louis, MO, USA) and 8 μg yeast tRNA in 1× hybridization solution (5× SSC, 25% formamide and 0.2% SDS). After denaturation at 95°C for 1 min and snap cooling on ice, the probes were hybridized to cDNA microarray slides containing elements produced by the Microarray Core Facility, National Cancer Institute, using Incyte Genomics UniGEM clones. Slides had been prehybridized (5× SSC, 0.1% SDS, 10 mg/ml BSA) for 1 hr and washed in water followed by isopropanol, before hybridization with the probe. Slides were incubated with labeled probes in humidified slide chambers at 42°C for 16 hr. Arrays were washed in 2× SSC, 0.1% SDS solution, and then subjected to more stringent washes in 1×, 0.2× and 0.05× SSC, respectively. Arrays were scanned using a 10 μm resolution GenePix 4000 scanner (Axon Instruments, Foster City, Ca, USA). The resulting TIFF images were analyzed with GenePix software v4.0 (Axon Instruments) and exported into the NCI microarray database (MaDB) developed by Dr. John Powel. Unsupervised and supervised analyses were performed using cluster and Treeview software.²⁶ Normal and cancer group comparisons were performed in MaDB using a multivariate permutation test. A total of permutations were performed, identifying a list of genes containing <10 false positives using a p < 0.0005. The methodology used for the microarray analysis follows the MIAME (Minimum Information about a Microarray Experiment) guidelines.

Quantitative real-time reverse transcription-PCR

Quantitative-RT-PCR was performed using the StepOne Real-time PCR System (Applied Biosystems, Foster City, Ca, USA). The comparative threshold cycle (C_T) method²⁷ was used for the calculation of expression fold change between samples. For first-strand cDNA synthesis, 2 μg of total RNA was reverse-transcribed using ImProm-II Reverse Transcriptase (Promega, Madison, WI, USA). Two microliters of the cDNA (~200 ng) was subsequently amplified using the KAPA SYBR qPCR Master Mix (Kapa Biosystems, Cape Town, South Africa), containing one of the following primer pairs: Crm1 (F 5′-GCA CCT CTT GGA CTG AAT CG-3′; R 5′-AAG CGA CAG CAC ACA CAC AC-3′), Kpnβ1 (F 5′-CCA GTG CCG AGT GGA ATG-3′; R 5′-AAA TCC CTG ACC CTC CTT C-3′), Kpnα2 (F 5′-AGC ACT CGC CGT CTT TC-3′; R 5′-AGA GAA CAC TGA CAA GGA ATG G-3′), Cyclophilin (F 5′-TGA GAC AGC AGA TAG AGC CAA GC-3′; R 5′-TCC CTG CCA ATT TGA CAT CTT C-3′) and GusB (F 5′-CTC ATT TGG AAT TTT GCC GAT T-3′; R 5′-CCG AGT GAA GAT CCC CTT TTT A-3′). All experiments were performed on a selection of patient samples and standardized to the housekeeping genes, GusB and cyclophilin. C_T values of GusB and cyclophilin were averaged for a single reference value.

Western blot analysis

Cells in culture were grown to 80% confluency and lysed in ‘RIPA buffer’ [10 mM Tris–Cl (pH 7.4), 150 mM NaCl, 1% sodium deoxycholate, 0.1% SDS, 1% Triton X-100, 1× complete protease inhibitor cocktail (Roche, Mannheim, Germany)]. Protein concentrations were quantified using a BCA protein assay kit (Pierce, Rockford, IL, USA) and separated on polyacrylamide gels, and transferred to Hybond™-ECL™ nitrocellulose membranes (Amersham, Buckinghamshire, UK). Western blot analyses were performed using the polyclonal rabbit anti-Crm1 (H-300) [sc-5595], rabbit anti-Karyopherin β1 (H-300) [sc-11367], goat anti-Karyopherin α2 (C-20) [sc-6917], mouse anti-p53 (M7001, DakoCytomation, Glostrup, Denmark), rabbit anti-p21 (H-164) [sc-756], rabbit anti-p27 (C-19) [sc-528], rabbit anti-p18 (N-20) [sc-865], rabbit anti-p16 (H-156) [sc-759] and rabbit anti-β-Tubulin (H-235) [sc-9104] antibodies (Santa Cruz Biotechnology, Santa Cruz, Ca, USA).

Immunofluorescent analysis of patient specimens

Paraffin blocks of the 16 cervical tumors and 10 normal cervical biopsies were collected from the archives of Groote Schuur Hospital. Briefly, tissue sections of ~5 μm were cut from the paraffin blocks, deparaffinized and rehydrated in a graded ethanol series. Antigen retrieval was carried out by pressure cooking in EDTA, pH 8.0, for 2 min. Slides were preincubated with 0.2% gelatin in PBS for 30 min to prevent nonspecific immunoreaction. Tissue sections were incubated with respective primary antibodies, rabbit anti-Crm1 (1:100; sc-5595, Santa Cruz Biotechnology), rabbit anti-Karyopherin β1 (1:100; sc-11367) and goat anti-Karyopherin α2 (1:100; sc-6917), for 45 min in a humidified chamber. Slides were given three 5-min washes in PBS containing 0.05% Tween 20, and Sudan Black (0.3% in 70% ethanol) applied to prevent autofluorescence. Sections were then incubated with secondary antibodies: Cy3-conjugated goat antirabbit antibody (1:300; Jackson Immunoresearch, West Grove, PA, USA) and Alexa488-conjugated donkey antigoat antibody (1:150, Jackson ImmunoResearch) for 45 min in a humidified chamber, and then washed. Cell nuclei were stained with DAPI (at 100 ng/ml) and slides were mounted and coverslipped. Fluorescence was visualized using standard fluorescence microscopy and confocal laser scanning microscopy. Standard optical filters were used for the separate imaging of the Cy3, Alexa488 and DAPI signals and all images were analyzed by defining regions of approximately equal size and cell number, using an interactive graphics screen (Axiovision, Gottingen, Germany, Release 4.5 software). A total of 6 views per slide were used to calculate the average fluorescence intensity per slide.

Proliferation assays and Leptomycin B treatment

Leptomycin B (LMB; Sigma) was stored as a 10.2 μM stock in methanol. CaSki, HCX and HCX-E6/E7 sensitivity to LMB was determined using the MTT assay. Briefly, 2 × 10⁴ cells were suspended into 96-well tissue culture plates. After a preincubation at 37°C overnight, cells were treated with various concentrations of LMB: 0, 0.1, 0.5, 1, 10, 20 and 100 nM. After incubation for 48 hr, viable cells were measured using the MTT reagent (Cell Proliferation Kit I, Roche) and absorbance at 595 nm measured. Experiments were performed in quadruplicate and dose-response curves generated.

RNA interference

For the inhibition of gene expression, short-interfering RNA (siRNA) was used [Crm1 siRNA (sc-35116), Kpnβ1 siRNA (sc-35736) and Kpnα2 siRNA (sc-35741), Santa Cruz Biotechnology]. siRNA consisting of a scrambled sequence (sc-37007, Santa Cruz Biotechnology) was used as a nonsilencing control. Cells were transiently transfected with 20 nM siRNA using Transfectin (Bio-Rad, Richmond, Ca, USA). Protein was harvested 24–96 hr after transfection and protein knock-down examined by Western Blot analysis. The effect of the respective siRNAs on cell proliferation was determined using the MTT assay. Briefly, cells were seeded into 96-well tissue culture plates, transfected with 20 nM siRNA and incubated for 24–144 hr. The effect of siRNA on proliferation was assayed using the MTT reagent.

Cell cycle analysis

For analysis of the effect of the siRNAs on the cell cycle, CaSki cells were plated in 60-mm dishes, transfected with 20 nM siRNA and incubated for 96 hr. Cells were harvested and fixed in 95% ethanol. For similar analysis of the effect of LMB on the cell cycle, cells were treated with 10 nM LMB and harvested and fixed 48 hr after treatment. Subsequent to fixing, the DNA content of the cells was stained using propidium iodide, and the cell cycle profiles were analyzed using the Beckman Coulter FC500 Flow Cytometer. Quantification of the percentage of cells at different stages of the cell cycle was performed using FlowJo software.

Apoptosis assays

To assay for Caspase-3/7 activity, the Caspase-Glo™ 3/7 Assay (Promega) was performed, according to the manufacturer’s instructions. Caspase activity was measured 24 hr after LMB treatment and 72 and 96 hr after transfection with siRNA. Luminescence was monitored using the Veritas™ Microplate Luminometer.

Immunofluorescent analysis for p53 and Crm1

For immunofluorescence analysis of p53 and Crm1 in control and Crm1 knock-down cells, cells were plated on round coverslips, transfected with 20 nM siRNA and fixed in 4% paraformaldehyde 120 hr post-transfection. The coverslips were placed in 0.1% Triton X-100 in PBS for 5 min (for cell permeabilization) followed by 50 mM NH₄Cl in PBS for 5 min (for quenching), and blocked in 0.2% gelatin for 30 min. Coverslips were subsequently incubated with α-Crm1 primary antibody (1:100) and α-p53 primary antibody (1:50) for 45 min in a humidified chamber. After 3 washes in PBS, secondary antibodies, Cy3-conjugated goat antirabbit antibody (1:300; Jackson ImmunoResearch) and Alexa488-conjugated goat anti-mouse antibody (1:400; Invitrogen™) were applied for a further 45 min. Cell nuclei were stained with DAPI (100 ng/ml) and coverslips mounted in Mowiol. Fluorescence was visualized using standard fluorescence microscopy.

Statistical analysis

For all data comparisons, the Student’s t test was performed using either Microsoft Excel or Graphpad prism. A P value of <0.05 was considered statistically significant.

Results

Crm1, Kpnβ1 and Kpnα2 expression in normal and cervical cancer tissue specimens

Analysis of the gene expression patterns of normal and cervical cancer tissue specimens using microarray analysis revealed a highly significant overexpression of crm1, kpnβ1, and kpnα2 in cervical cancer biopsies compared to normal cervical tissue (p < 0.0005) (Fig. 1a). As these genes are essential for numerous cellular processes, we hypothesized that their overexpression may have significance to cancer biology. Real-time RT-PCR, on a sample of normal and cervical cancer tissue specimens, confirmed a significant increase in crm1, kpnβ1 and kpnα2 gene expression in cancer tissue compared to normal (at least 2.5-fold, p < 0.0005) (Fig. 1b).

Figure 1 –

Expression of Crm1, Kpnβ1 and Kpnα2 in normal and cervical cancer tissues. (a) Gene expression of crm1, kpnβ1 and kpna2 in normal and cervical cancer biopsies, analyzed by microarray technologies [normal (n = 8), cancer (n = 16), *p < 0.0005]. (b) Real-time RT-PCR analysis confirming upregulation of crm1, kpnβ1 and kpna2 in cervical cancer biopsies compared to normal [normal (n = 9), cancer (n = 13), *p < 0.0005]. Results shown are the mean ± SE. (c–e) Quantification of Crm1 (c), Kpnβ1 (d) and Kpnα2 (e) immunofluorescence in 10 normal and sixteen cervical cancer tissue sections. Fluorescence was quantified using AxioVision, Release 4.5 software. A total of 6 views per slide were used to calculate the average fluorescence intensity per sample. Box-and-whisker plots comparing Crm1 (c), Kpnβ1 (d) and Kpnα2 (e) expression in normal and cancer tissue sections were generated using Graphpad Prism (*p < 0.05). (f) Representative fluorescence images showing hematoxylin and eosin (H&E) staining, Crm1, Kpnβ1 and Kpnα2 expression, and nuclear staining (DAPI) in normal (GSH7) and cervical cancer (GSH28) sections. All images were collected at identical exposure settings and magnification (E: epithelium; S: stroma). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Having established overexpression of the karyopherin genes at the mRNA level, confirmation of increased Crm1, Kpnβ1 and Kpnα2 protein expression in cancer specimens was obtained by immunofluorescent analysis using tissue sections cut from 26 archived patient blocks. Tissue sections from 10 normal patients and 16 patients with cervical cancer were assayed and results showed significantly higher expression of Crm1, Kpnβ1 and Kpnα2 in the cancers compared to the normal specimens (p < 0.05) (Figs. 1c-1e). Fluorescent images of Crm1, Kpnβ1 and Kpnα2 showed predominantly nuclear envelope and cytoplasmic localization, while some expression was also detected in the nucleus (Fig. 1f). The localization patterns of the Karyopherin proteins are in agreement with that described by the manufacturers of the individual antibodies.

Expression of Crm1, Kpnβ1 and Kpnα2 in cultured normal, transformed and cancer cells

Western blot analysis was used to determine the expression of Crm1, Kpnβ1 and Kpnα2 in cultured cell lines. Expression levels in the cervical cancer cell lines, HeLa, SiHa, CaSki, ME180, Ms751 and C33A, were compared to that in a primary cervical epithelial culture derived from normal cervix, HCX, as well as in HPV16-E6/E7-transformed HCX cells (HCX-E6/E7). Crm1, Kpnβ1 and Kpnα2 expression was detected in all of the cell lines, with higher expression in the cancer cell lines compared to HCX normal primary cells (Fig. 2a), confirming the patient-derived data. E6/E7-transformed HCX cells similarly showed higher Crm1, Kpnβ1 and Kpnα2 protein expression compared to the untransformed HCX cells.

Figure 2 –

Expression of Crm1, Kpnβ1 and Kpnα2 in cells grown in culture. (a) Western blot analyses of Crm1, Kpnβ1 and Kpnα2 expression in normal primary cervical epithelial cells (HCX), HPV E6/E7-transformed cells (HCX-E6/E7) and cervical cancer cell lines HeLa, SiHa, CaSki, ME180, Ms751 and C33A. β-tubulin was used as a control for protein loading. Results shown are representative of Western blots performed in triplicate. (b) Western blot analyses of Crm1, Kpnβ1 and Kpnα2 expression in normal lung fibroblasts, WI38, compared to transformed fibroblasts, SVWI38 and CT1.

An analysis of Crm1, Kpnβ1 and Kpnα2 protein expression in transformed fibroblast cell lines (SVWI38 and CT-1) compared to normal fibroblasts (WI38) revealed elevated Crm1, Kpnβ1 and Kpnα2 expression in the transformed fibroblasts (Fig. 2b). This suggests that the upregulation of these Karyopherin proteins is likely not confined to cervical cancer, and may be associated with cellular transformation in general.

Effect of Leptomycin B on cell viability

To determine whether elevated levels of the Karyopherin proteins in cancer cells were of functional significance, the effect of inhibiting either their activity or expression was investigated. No drugs are currently available for the inhibition of the importin proteins, Kpnβ1 and Kpnα2; however, a Crm1 inhibitor, Leptomycin B (LMB), was used to inhibit Crm1 function in representative cervical cell lines. The cervical cancer cell line, CaSki, and the E6/E7-transformed cell line, HCX-E6/E7, appeared to be more sensitive to LMB treatment than the normal cervical epithelial cells, HCX (Fig. 3a), suggesting that the viability and survival of cervical cancer and transformed cells are more dependent on Crm1 activity compared to normal cells. To investigate cancer cell death induced by LMB treatment, cell proliferation and cell cycle analyses were performed. MTT assays showed a significant decrease in the proliferation of CaSki cells grown in the presence of LMB compared to controls (Fig. 3b). An analysis of the cell cycle profiles of CaSki cervical cancer cells treated with 10 nM LMB revealed a significant increase in the percentage of cells in the subG1 population, indicative of cell death (Figs. 3c and 3d). Caspase-3/7 assays revealed an ~ 3-fold increase in Caspase-3/7 activity in the LMB-treated cells suggesting that the inhibition of Crm1 activity results in cancer cell death via apoptosis (Fig. 3e).

Figure 3 –

Effect of Leptomycin B (LMB) treatment on cell viability. (a) CaSki cervical cancer cells, HCX-E6/E7-transformed cervical cells and HCX primary normal cervical cells were treated with LMB at increasing nanomolar concentrations and cell viability assayed 48 hr after treatment using MTT reagent. Results shown are the percentage of viable cells after LMB treatment relative to the untreated controls. Experiments were performed in quadruplicate and repeated at least 2 times. (b) CaSki cell proliferation in the absence and presence of 10 nM LMB. Cell proliferation was measured using MTT after a 48-hr incubation in the absence or presence of LMB. Cell proliferation in LMB-treated cells is expressed relative to that in untreated cells (*p < 0.005). (c) Effect of LMB on cell cycle progression. CaSki cells were left untreated or treated with 10 nM LMB for 48 hr, and harvested for cell cycle analysis using flow cytometry. LMB-treated cells showed an increased subG1 population. Experiments were performed in triplicate. (d) Quantitation of the cell cycle data showing the significant increase in the subG1 cell population of cells treated with LMB (*p < 0.01). (e) Caspase-3/7 activity in LMB-treated cells. Cells were treated with 10 nM LMB for 24 hr and Caspase-3/7 activity analyzed as a measure of apoptosis (*p < 0.001).

Inhibition of Crm1 and Kpnβ1, but not Kpnα, interferes with cancer cell proliferation and induces cancer cell death via apoptosis

Considering that no drugs are currently available for the specific inhibition of Kpnβ1 and Kpnα2, siRNA technologies were used to inhibit their expression, and the effect of this inhibition of cellular function determined. Firstly, CaSki cervical cancer cells were transiently transfected with siRNA against Crm1, Kpnβ1 and Kpnα2, respectively, and its effect on protein expression levels was monitored by Western blot analysis. Crm1, Kpnβ1 and Kpnα2 protein levels were significantly reduced on treatment with the different siRNAs (Figs. 4a-4c). A control siRNA was used as a negative control. Time-course analyses of Crm1, Kpnβ1 and Kpnα2 inhibition showed that the gene knock-downs resulted in reduced protein levels from as early as 24 hr post-transfection. To determine whether inhibition of expression had biological relevance, cell proliferation assays were performed in the presence of control, Crm1, Kpnβ1 and Kpnα2 siRNA, respectively. Results showed that CaSki cell proliferation was significantly reduced in cells where Crm1 and Kpnβ1 expression was inhibited (Figs. 4d and 4e). However, cells in which Kpnα2 was inhibited showed no change in cell proliferation, suggesting that the inhibition of this gene does not affect cancer cell biology (Fig. 4f). These findings suggest that Crm1 and Kpnβ1 are important for the proliferation of cervical cancer cells.

Figure 4 –

Effect of inhibiting Crm1, Kpnβ1 and Kpnα2 expression on CaSki cell proliferation using siRNAs. (a–c) Time-course analyses of Crm1 (a), Kpnβ1 (b) and Kpnα2 (c) inhibition after transfection with siRNA. Protein was harvested at the indicated time points after transfection. Western blot analysis using antibodies specific to Crm1, Kpnβ1 and Kpnα2 was used to analyze inhibition with individual siRNAs. β-tubulin was used as a control for protein loading. (d–f) The effect of Crm1 (d), Kpnβ1 (e) and Kpnα2 (f) siRNA on cell proliferation was determined using the MTT assay. CaSki cervical cancer cells were transfected with 20 nM control siRNA or respective siRNA, and cell growth was monitored using the MTT reagent. Results shown are the mean ± SD of experiments performed in quadruplicate and repeated at least 3 times. Crm1 and Kpnβ1 siRNA resulted in a significant inhibition of cell proliferation (p < 0.00001 at 96 hr).

The growth inhibitory effect mediated by Crm1 and Kpnβ1 siRNA was tested on a panel of cervical cancer cell lines, including CaSki, Hela, SiHa, MS751 and C33A. Cell proliferation monitored 5 days after transfection with the different siRNAs revealed that all of the cervical cancer cell lines tested had a similar sensitivity to the siRNAs, where Crm1 and Kpnβ1 inhibition significantly reduced proliferation, while Kpnα2 inhibition showed no effect (Figs. 5a-5e). Since the inhibition of Crm1 and Kpnβ1 reduced the proliferation of different cancer cell lines, we next determined the effect of inhibition on normal cells. Because of difficulties encountered with the transfection of HCX primary cells, the noncancer fibroblast cell line WI38 that could more readily be transfected with siRNAs was used. Cell proliferation assayed 5 days after siRNA transfection showed that the proliferation of these normal cells was unaffected by Crm1 and Kpnβ1 inhibition (Fig. 6a). This lack of response in WI38 cells was not due to an ineffective protein knock-down, as Western blot analysis revealed effective and specific inhibition of protein expression in cells transfected with either Crm1 or Kpnβ1 siRNA (Fig. 6b). SVWI38, a cell line derived from SV40-transformed WI38 cells, showed a modest decrease in proliferation after inhibition of Crm1 and Kpnβ1 compared to that observed in the CaSki cell line (Figs. 6a and 6b). To substantiate the finding that the proliferation of noncancer cells was unaffected by the inhibition of Crm1 and Kpnβ1, other normal cell lines were tested, including normal skin fibroblasts CCD-1068SK and FG0, as well as normal immortalized epithelial cells, EPC2. Crm1 and Kpnβ1 knock-down in these noncancer cell lines similarly had very little effect on cell proliferation (Fig. 6c).

Figure 5 –

Effect of Crm1, Kpnβ1 and Kpnα2 inhibition on cervical cancer cell proliferation. (a–e) Cells were transfected with siRNA as described in Figure 4 and proliferation monitored 120 hr post-transfection. Proliferation is shown for each cell line relative to that of cells transfected with control siRNA. Cervical cancer cell lines, CaSki (a), HeLa (b), SiHa (c), MS751 (d) and C33A (e), showed similar sensitivities to Crm1 and Kpnβ1 inhibition, and were unaffected by Kpnα2 inhibition. Results shown are the mean ± SD of experiments performed in quadruplicate and repeated at least 2 times (*p < 0.05).

Figure 6 –

Effect of Crm1, Kpnβ1 and Kpnα2 inhibition on the proliferation of normal fibroblasts, WI38, transformed fibroblasts, SVWI38, and CaSki cervical cancer cells. (a) Cells were transfected with siRNA against Crm1 and Kpnβ1 and viable cells assayed using MTT reagent 120 hr post-transfection. Proliferation is shown for each cell line relative to that of cells transfected with the control siRNA. Results shown are the mean ± SD of experiments performed in quadruplicate and repeated at least 2 times. (b) Western blot analysis matched to proliferation assays in (a) showing Crm1 and Kpnβ1 knock-down after transfection with individual siRNAs. (c) Proliferation of normal cells, CCD-1068SK, FG0 and EPC2, 120 hr after Crm1 and Kpnβ1 inhibition. Results shown are the mean ± SD of experiments performed in quadruplicate and repeated at least 2 times (*p < 0.05).

To determine whether inhibition of cancer cell proliferation induced by blocking Crm1 and Kpnβ1 expression was associated with changes in cell cycle progression, cell cycle analysis was performed on cells transfected with control, Crm1 or Kpnβ1 siRNA. A significant increase in a subG1 population in Crm1- and Kpnβ1-inhibited cells was detected (Figs. 7a and 7b). These findings suggested that inhibition of Crm1 and Kpnβ1 expression resulted in an increase in cell death. This was subsequently shown to occur via apoptosis, as measured by Caspase-3/7 activity. Cells assayed 72 and 96 hr after inhibiting Crm1 and Kpnβ1 showed an approximate 2-fold increase in Caspase-3/7 activity compared to controls (p < 0.01) (Fig. 7c). Kpnα2 knock-down cells showed no change, supporting our earlier finding that inhibition of Kpnα2 had no effect on proliferation. These results suggest a role for Crm1 and Kpnβ1 in cancer cell proliferation and survival, while Kpnα2 function appears to be redundant.

Figure 7 –

Crm1 and Kpnβ1 inhibition induces apoptosis in CaSki cells. (a) Effect of Crm1 siRNA on cell cycle progression. CaSki cells were transfected with 20 nM control siRNA or Crm1 siRNA and harvested 96 hr later, stained with propidium iodide and respective cell cycle profiles analyzed using flow cytometry. Crm1 knock-down cells showed an increase in a subG1 population. Experiments were performed in triplicate and repeated at least 2 times. (b) Quantitation of the cell cycle data obtained from cells transfected with control, Crm1 or Kpnβ1 siRNA, showing the significant increase in the subG1 cell population of cells transected with Crm1 and Kpnβ1 siRNA (*p < 0.001) and a minimal but significant decrease in G1 (p < 0.005). (c) Caspase-3/7 activity (y-axis) for CaSki cells transfected with 20 nM control, Crm1, Kpnβ1 or Kpnα2 siRNA was measured 72 and 96 hr post-transfection (x-axis). Crm1 and Kpnβ1 inhibition caused a significant increase in Caspase-3/7 activity (*p < 0.01); however, Kpnα2 inhibition caused no change in Caspase-3/7 activity compared to control siRNA-transfected cells. Results shown are the mean ± SD of experiments performed in quadruplicate and repeated at least 2 times.

Crm1 and Kpnβ1 inhibition results in elevated levels of p53, p21, p27 and p18

p53 stability and degradation has been reported to be dependent on Crm1 activity.²⁸ To determine whether the apoptosis induced by inhibition of Crm1 was associated with changes in p53, p53 levels were analyzed in CaSki cells transiently transfected with either control or Crm1 siRNA. Immunofluorescent analysis for Crm1 expression in control and Crm1 knock-down cells showed significantly lower levels in the Crm1 siRNA-transfected cells (Fig. 8a, 1 and 2). An accompanying increase in p53 nuclear staining was observed in Crm1-inhibited cells compared to controls (Fig. 8a, 3 and 4, marked with arrows). Intense nuclear staining was not, however, observed in every cell, likely due to the nature of transient transfections. DAPI was used to visualize the nuclei of control and Crm1 siRNA-transfected cells and also indicated fewer cells in the Crm1-inhibited group, consistent with earlier proliferation assays (Fig. 8a, 5 and 6).

Figure 8 –

The inhibition of Crm1 and Kpnβ1 results in increased levels of p53, p21, p27 and p18. (a) CaSki cells were transfected with 20 nM control siRNA (1, 3, 5) or Crm1 siRNA (2, 4, 6) and fixed 120 hr post-transfection. Immunofluorescence analysis of Crm1 knock-down (1 vs. 2) and p53 (3 vs. 4) shows elevated nuclear p53 staining (4: marked with arrows) in cells with low Crm1 levels. Results shown are representative of experiments performed in 2 independent analyses. DAPI staining for nuclei are shown (5 and 6). (b) Western blot analysis showing p53, p21, p27, p18 and p16 levels after Crm1, Kpnβ1 and Kpnα2 knock-down. Knock-down of Crm1 and Kpnβ1 resulted in elevated p53, p21, p27 and p18 levels, while p16 remained relatively unchanged. (c) Western blot analysis showing levels of cell cycle inhibitors after Crm1, Kpnβ1 and Kpnα2 inhibition in noncancer WI38 cells. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Nuclear accumulation of p53 in Crm1 knock-down cells suggests that Crm1 inhibition prevents the nuclear export of p53, thereby possibly protecting p53 from degradation as degradation of p53 occurs predominantly in the cytoplasm. We therefore next investigated whether the Crm1 knock-down altered p53 expression levels and that of other growth inhibitory proteins including p27, p21, p18 and p16. p53 levels in Crm1 knock-down cells was found to be elevated compared to control cells (Fig. 8b). Additionally, p21, a known p53 target gene, was also found to have higher levels in Crm1 knock-down cells. Similarly, p27 and p18 levels were elevated in Crm1 knock-down cells; however, p16 levels remained unchanged (Fig. 8b). p53, p21, p27 and p18 expression was also found to be elevated in Kpnβ1 knock-down cells but not in Kpnα2 knock-down cells (Fig. 8b). Taken together, these results suggest that the higher levels p53, p21, p27 and p18 in Crm1 and Kpnβ1 knock-down cells may be part of an apoptotic response when these 2 Karyopherin proteins are inhibited.

Interestingly, in WI38 noncancer cells, where inhibition of Crm1 and Kpn β1 with siRNA had no effect on cell proliferation, Crm1 and Kpnβ1 knock-down resulted in an increase in p53, although to a lesser extent than that observed in the CaSki cervical cancer cells (Fig. 8c). This was accompanied by a marginal induction of p21 expression. p27 levels in the Crm1 and Kpnβ1 knock-down cells was comparable to that in the cancer cells. p18, which had shown elevated expression in response to inhibition of Crm1 and Kpnβ1 in cancer cells, was unchanged and virtually undetectable in the normal cells. p16 levels were similarly very low (Fig. 8c).

Discussion

In this study, we identified altered expression of members of the Karyopherin family of nuclear transport proteins, Crm1, Kpnβ1 and Kpnα2, in cervical tumor tissue and cell lines. We also found that transformed fibroblasts and transformed epithelial cells similarly overexpress Crm1, Kpnβ1 and Kpnα2. To our knowledge, this study is the first to report overexpression of these 3 Karyopherins simultaneously in cancer cells, and suggests that Crm1 and Kpnβ1, not Kpnα2, are essential for cancer cellular function. The Karyopherins have received interest of late as potential targets for cancer therapy.^29-31 It has been suggested that their aberrant functioning can lead to uncontrolled cell growth,^30,32 and thus the directed targeting of these proteins may have potential as an anticancer therapy. The potential of Crm1 as a diagnostic and therapeutic target for aggressive ovarian carcinomas has recently been reported.³³ These authors report that Crm1 overexpression in ovarian cancers correlated significantly with poor patient survival. Overexpression of Kpnα2 has similarly been found to correlate with a shorter disease-free survival in patients with breast cancer^34,35 and melanomas.³⁶

To determine whether Crm1, Kpnβ1 and Kpnα2 have functional relevance to cancer cells, we used 2 approaches to inhibit their activity: drug-mediated inhibition and RNA interference. No Kpnβ1- or Kpnα2-specific drugs are currently available; however, LMB has been shown to be a potent inhibitor of Crm1 activity.^37-39 LMB treatment of cervical cancer cells and transformed cervical epithelial cells was highly cytotoxic, at nanomolar concentrations, while the normal cervical epithelial cells were much less sensitive. This is in agreement with a recent study reporting that keratinocytes expressing HPV oncogenes were considerably more sensitive to LMB treatment compared to normal keratinocytes.⁴⁰ It is possible that the differences in sensitivity to LMB may be related to the differences in the proliferation and metabolic ability of cancer and normal cells. A theory is that for cancer cells to maintain their high rate of proliferation and metabolic activity, higher levels and activity of proteins such as Crm1 may become essential to their biology. It has to be noted that LMB was found to have significant side-effects when tested in clinical trials; however, the question remains whether these side-effects were due to off-target toxicity of the drug or target-specific toxicity. Using the siRNA approach, we found that Crm1 siRNA, like LMB, resulted in cell death, as did Kpnβ1 siRNA, while Kpnα2 siRNA had no visible effect on the cell. Normal fibroblasts and epithelial cells, on the other hand, remained unaffected after knock-down of all 3 Karyopherin proteins. This, together with our finding that normal primary epithelial cells were less sensitive to LMB treatment, suggests that normal and cancer cells respond differently to the inhibition of Crm1 and Kpnβ1, which is promising in terms of the development of anticancer drugs.

We observed the nuclear accumulation of p53 in Crm1 knock-down cells. Crm1 is known to export p53 from the nucleus into the cytoplasm of cervical cancer cells via the C-terminal NES of p53,²⁸ a function that is promoted by HPV E6.²¹ Additionally, nuclear export is required for the efficient degradation of p53 in high-risk HPV-infected cells,²⁸ although nuclear degradation of p53 does occur at low levels.²¹ Accordingly, the inhibition of Crm1 prevents p53 from exiting in the nucleus, resulting in its nuclear accumulation and stabilization. Consistent with our finding, it has been shown that treatment of cervical cancer cells with LMB leads to the accumulation of active p53 in the nucleus and the induction of apoptosis.^41,42 It has also been reported that the introduction of a dominant negative p53 only partially rescued cells from the apoptotic effect induced by LMB.⁴¹ Since p53 only partially mediates the apoptotic effect induced by LMB, it is likely that p53-independent mechanisms are also involved in cell death induced by Crm1 inhibition. Moreover, we observed that C33A cervical cancer cells, which carry mutant p53,⁴³ were susceptible to Crm1 siRNA-induced cell death, suggesting that p53-independent apoptosis occurs within these cells. Furthermore, the loss of Crm1 and Kpnβ1 resulted in elevated levels of not only p53 and p21, but also p27 and p18 in CaSki cervical cancer cells, suggesting that all of these proteins may be contributing to the apoptosis observed when Crm1 and Kpnβ1 are inhibited. p16 levels, however, remained unchanged, possibly due to the fact that p16 is overexpressed in cervical cancer cells.⁴⁴ Crm1 and Kpnβ1 knock-down in noncancer WI38 cells that had showed no effect on cell proliferation surprisingly also resulted in an increase in p53 and p27 levels; however, p21 levels remained largely unchanged. p18 and p16 were unaltered and virtually undetectable in these cells. This suggests that the noncancer cells may in part be protected from the cell death inducing effects of Crm1 and Kpnβ1 knock-down due to a lack of alteration of p21 and p18 levels.

Interestingly, the inhibition of Kpnα2 had no visible effect on the cell lines tested; a finding in line with that of Quensel et al.,⁴⁵ who found that the inhibition of certain karyopherin α proteins blocked proliferation, while the inhibition of others, including Kpnα2, did not. Although Kpnα2 has an important function in the classical nuclear import pathway, wherein it acts as an adaptor molecule linking cargo proteins to Kpnβ1, there are other karyopherin α proteins, which can possibly compensate for loss of Kpnα2. Kohler et al. suggest that while each karyopherin α family member has different substrate specificities, they, however, are not confined to the substrates for which they have high affinity.⁴⁶ In addition, other nonclassical import pathways exist that do not require an adaptor protein, and the proteins are transported by Kpnβ1 alone. A recent study predicted that 57% of nuclear proteins use the classical nuclear import pathway, while about 43% use other mechanisms.⁴⁷ It is interesting that the cancer cells used in our study manage to survive in the absence of Kpnα2, even though they upregulate its expression. Kpnα2 therefore may have potential as a biomarker, but its potential as a therapeutic target remains to be determined.

In conclusion, this study shows that cervical cancer cells overexpress Crm1, Kpnβ1 and Kpnα2 and that these cells become highly dependent on the function of Crm1 and Kpnβ1, but not Kpnα2. The fact that both an importer and an exporter of cargo proteins are upregulated in cancer cells may have an end result of more efficient transport sustaining the faster proliferation rate of cancer cells. Our findings suggest that Crm1 and Kpnβ1 may have potential as both biomarkers and therapeutic targets, especially since normal epithelial and fibroblast cells appear to respond differently to inhibition compared to transformed and cancer cells.