Stable cell lines that inducibly overexpress GFP and GFP-tagged Nup88 in response to doxycycline were previously established from HeLa R19 cells containing a single Flp recombination site in the genome (32,33). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM; FUJIFILM Wako Pure Chemical Corporation) supplemented with 10% fetal bovine serum (HyClone™; Cytiva), penicillin/streptomycin (FUJIFILM Wako Pure Chemical Corporation) and 1 µg/ml doxycycline (Nacalai Tesque, Inc.) in a humidified atmosphere with 5% CO₂ at 37°C.

pEGFP-N2-NUP88, a plasmid that overexpresses GFP-fused Nup88, was constructed previously (32). pEGFP-N2-GLI1, a plasmid that overexpresses GFP-tagged Gli1 was constructed as follows. The GLI1 gene was obtained from pOTB7 coding GLI1 cDNA by PCR using the following pair of oligonucleotide primers (Forward: 5′-CTCAAGCTTCGAATTCCGCCACCATGTTCAACTCGATGACCCCA-3′, Reverse: 5′-GTCGACTGCAGAATTCGGCACTAGAGTTGAGGAATTCTGT-3′). The underlined sequences represent the regions complementary to the GLI1 gene. The amplified product was cloned into the EcoRI site of pEGFP-N2 (Takara Bio, Inc.) using an In-Fusion HD Cloning Kit (Takara Bio, Inc.). The plasmids were transformed into competent cells via heat shock at 42°C for 30 sec. After the addition of LB medium, the cells were incubated at 37°C for 30 min and then cultured overnight on LB agar plate containing appropriate antibiotics. A plasmid that overexpresses Flag-tagged GLI1 was constructed by replacing the eGFP gene in pEGFP-N2-GLI1 with the 3×FLAG gene. pDPNR223 containing the truncated KIF7 gene (ID100072702A), which encodes amino acid residues 514–1343 of full-length Kif7 (1343 aa), was purchased from DNAFORM. mCherry-tagged truncated KIF7 [pmCherry-KIF7(514–1343)] was constructed as follows. The entire truncated KIF7 in pDPNR223 was amplified by PCR using the following primer set (Forward: 5′-gtcgacggtaccgcgggcccgggcgccaccatggagcagtacaaactgca-3′, Reverse: 5′-tccttgtagtcttggatcccgggcagggggtttttccggac-3′) and cloned into pEGFP-N2 using an In-Fusion HD Cloning Kit. The underlined sequences represent the regions complementary to the KIF7 gene. The eGFP gene in the plasmid was then swapped with the mCherry gene. siRNAs for control (cat. no. 4390843), KIF7 (cat. no. 4427037, ID: s51568), GLI1 (cat. no. 4427037, ID: s5816), GLI2 (cat. no. 4427037, ID: s5817), GLI3 (cat. no. 4427037, ID: s533814), and KRAS (cat. no. 4427038, ID: s7939) were purchased from Thermo Fisher Scientific, Inc.

Plasmids and siRNAs were transfected using Lipofectamine^® LTX and Lipofectamine^® RNAiMax reagent, respectively, according to the manufacturer's instructions (Invitrogen; Thermo Fisher Scientific, Inc.). The transfection complexes prepared were added to cells at room temperature, and the cells were then incubated at 37°C in a CO₂ incubator. Subsequent experiments were performed 24 h after transfection for plasmids. For the siRNAs, cells were analyzed 48 h after transfection for KRAS, GLI1, GLI2, and GLI3-targeting siRNAs and 24 h after transfection for KIF7-targeting siRNA.

Total RNA was prepared from ~1×10⁷ cells using an RNeasy Kit (Qiagen GmbH). RT-qPCR was performed using One Step TB^® Green PrimeScript RT-PCR Kit II (Takara Bio, Inc.) and StepOnePlus™ (Applied Biosystems; Thermo Fisher Scientific, Inc.). All experiments using these kits and apparatus were performed according to the manufacturer's instructions or protocols. RT-qPCR conditions were as follows: Initial denaturation at 95°C for 30 sec, followed by 40 cycles of 95°C for 5 sec and 60°C for 30 sec. The expression of genes of interest was assessed using the 2^−ΔΔCq method with human 18S rRNA (RN18S1) as the internal control for normalization (34). Oligonucleotide primers employed in this assay were as follows: RN18S1 Forward: 5′-CGGCTACCACATCCAAGGAA-3′; RN18S1 Reverse: 5′-GCTGGAATTACCGCGGCT-3′; KIF7 Forward: 5′-AGACGGAAGAGATCGCGGCATT-3′; KIF7 Reverse: 5′-CTGCTGTAGCACCTTCTCCATC-3′.

Antibodies used for immunoblotting were as follows: Anti-GFP (cat. no. 598), anti-RFP (cat. no. PM005), anti-α-tubulin (cat. no. PM054), and anti-β-actin (cat. no. PM053-7) from Medical & Biological Laboratories Co., Ltd.; anti-Kif7 (cat. no. ABS1458) and anti-FLAG^® M2 (cat. no. F1804) from MilliporeSigma; anti-DNMT1 (D63A4; cat. no. 5032), anti-Gli1 (C68H3; cat. no. 3538), anti-p44/42 MAPK (ERK1/2; 137F5; cat. no. 4695), anti-phospho-p44/42 MAPK (Thr202/Tyr204; D13.14.4E; cat. no. 4370P), anti-mTOR (cat. no. 2983), anti-p-mTOR (S2448) (cat. no. 2971), anti-AKT (cat. no. 9272), anti-phospho-AKT (Thr308) (cat. no. 9275), anti-phospho-AKT (Ser473) (cat. no. 4060P), anti-Smad2/3 (cat. no. 8685T), and anti-Lamin A/C (cat. no. 4777S) from Cell Signaling Technology, Inc.; anti-Nup88 (cat. no. 611896; BD Biosciences); anti-Gli2 (cat. no. 18989-1-AP), anti-phospho-ERK(1/2; Thr202/Tyr204; cat. no. 80031-1-RR), and anti-KRAS (cat. no. 12063-1-AP) from Proteintech Group, Inc.; anti-human/mouse Gli3 (cat. no. AF3690; R&D Systems); anti-phospho-mTOR (Ser2481) (cat. no. 09-343SP) from Merck Millipore. Peroxidase-conjugated goat anti-rabbit IgG (cat. no. 115-035-003) and anti-mouse IgG (cat. no. 115-035-062) were purchased from Jackson Immuno Research Laboratory, Inc. 5-Aza-2′-deoxycytidine (cat. no. A2232) and U0126 (cat. no. S1102) were purchased from Tokyo Chemical Industry and Selleck Biotechnology, respectively.

Cells were seeded into 24-well plates containing coverslips (3×10⁴ cells/well). After 2 days of culture, cells were washed once with PBS and fixed with 4% paraformaldehyde at 4°C for 10 min. For indirect immunofluorescence imaging, the cells were permeabilized with 0.5% Triton X-100 in PBS for 10 min and then incubated with blocking buffer [3% bovine serum albumin (FUJIFILM Wako Pure Chemical Corporation), 0.05% Triton X-100 in PBS] for 1 h, followed by incubation with primary antibodies diluted 1:1,000 in the same blocking buffer for 1 h at room temperature. Following primary antibody incubation, the cells on coverslips were incubated with wash buffer (1.5% bovine serum albumin, 0.05% Triton X-100 in PBS) three times at 5 min intervals. Secondary antibodies conjugated with Alexa Fluor 568 or 488 (Thermo Fisher Scientific, Inc.) were diluted 1:1,000 in the same blocking buffer and then added for 1 h at room temperature in dark conditions. Following secondary antibody incubation, the cells on coverslips were washed three times in the same wash buffer. For direct fluorescence imaging, or after secondary antibody incubation in indirect immunofluorescence, cells were stained with Hoechst 33342 in PBS for 10 min at room temperature. Coverslips were mounted using ProLong Gold antifade reagent (Thermo Fisher Scientific, Inc.), and fluorescence was detected using a 20× objective of a FSX100 inverted fluorescence microscope (Olympus Corporation).

Whole cell lysates prepared with radioimmunoprecipitation assay (RIPA) buffer (FUJIFILM Wako Pure Chemical Corporation) were quantified using the Bradford assay reagent (Nacalai Tesque, Inc.) and then the lysates (~30 µg of protein per lane) were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using 7, 10, or 12% gels depending on the molecular weight of the target proteins. Proteins separated in the gel were electroblotted onto an Immobilon-P transfer membrane (MilliporeSigma). The membrane was blocked with blocking buffer (5% skimmed milk and 0.05% Tween 20 in PBS) for 30 min at room temperature, followed by incubation overnight with primary antibody diluted 1:1,000 in dilution buffer (1% skimmed milk and 0.05% Tween 20 in PBS). The membrane was then washed three times with wash buffer (0.05% Tween 20 in PBS) for 5 min each and further incubated with peroxidase-conjugated secondary antibodies at 4°C for 1 h. For each immunoblot, loading controls were detected on the same membrane as the corresponding target proteins. Chemiluminescence on the membrane produced by secondary antibodies in the presence of Luminate Crescendo Western HRP substrate (MilliporeSigma) was detected using a LAS2000 imaging analyzer (FUJIFILM Corporation). Densitometric analysis of target protein band intensities was performed using ImageJ version 1.54p (National Institutes of Health).

Conditions used for live-cell single-cell imaging were established previously (22). Briefly, cells were seeded with 0.7×10⁴ cells on 4-well collagen-coated glass-bottom dishes (Matsunami Glass Ind., Ltd.) and incubated at 37°C overnight with a phenol red-free culture medium in a humidified atmosphere with 5% CO₂. The medium was then changed just before setting the dishes on a stage top CO₂ incubator of a confocal microscope. Fluorescence signals from cells were acquired using a C2 confocal microscope system (Nikon, Tokyo, Japan) at 10 min intervals for 24 h. The migration velocity of each cell was analyzed using ImageJ version 1.54p (National Institutes of Health).

ChIP assays were performed using a SimpleChIP^® Enzymatic Chromatin IP Kit (Cell Signaling Technology, Inc.), according to the manufacturer's instructions. Cells (4×10⁶ cells per reaction) were cross-linked, lysed, and digested with Micrococcal Nuclease. Chromatin lysate (20 µg) was incubated with primary antibody (2 µg) overnight at 4°C with rotation and 2% of the lysate was used as the Input control. Immune complexes captured with protein G magnetic beads were washed with low-salt and high-salt reagents. After cross-link reversal, purified DNA was subjected to PCR using Taq DNA polymerase (Takara Bio Inc.) under the following conditions: 95°C for 5 min; 30 cycles of 95°C for 30 sec, 55°C for 30 sec, 72°C for 30 sec and 72°C for 1 min. PCR products were electrophoresed on a 2% agarose gel, visualized with ethidium bromide, and evaluated solely by visual inspection.

GFP-tagged proteins were retrieved from mammalian cell lysates using GFP-Trap A beads (Proteintech Group, Inc.). Cell lysates (500 µg; 1 mg/ml) containing GFP or GFP-tagged Gli1 and FLAG-tagged truncated Kif7 were incubated with the beads at 4°C overnight. Beads were then washed three times with lysis buffer at 4°C for 5 min and incubated with SDS sample buffer at 100°C for 1 min to elute the bound proteins.

The expression data for KIF7 used in this study were obtained from the DNA microarray analysis performed in our previous study (22). The raw dataset is accessible via GSE310626. KIF7 expression data were extracted from this pre-existing dataset.

Statistical significance was evaluated using the paired t-test for most of the immunoblotting data. For the analysis of Fig. 3H, one-way analysis of variance followed by Dunnett's test was applied. Migration assays were analyzed using one-way ANOVA followed by the Tukey-Kramer multiple-comparison test. P<0.05 was considered to indicate a statistically significant difference.

We previously demonstrated by DNA microarray analysis that overexpression of Nup88 alters gene expression, which may influence cell motility (22). The present study, to identify additional factors involved in this Nup88-dependent migration, focused on KIF7, which showed a marked decrease in expression in our previous microarray analysis (22) (Table SI). Kif7 is a negative regulator of the Hh pathway that contributes to cancer development. It has been reported that in prostate cancer cell lines, the expression of KIF7 is epigenetically repressed through hypermethylation of CpGs within its promoter (5′ upstream) region, which promotes formation of the malignant phenotype (29). The present study therefore hypothesized that the downregulation of Kif7 could also contribute to motility in HeLa cells. To verify the DNA microarray results, previously established HeLa cell lines that stably overexpress GFP or GFP-tagged Nup88 were utilized (32). As shown in Fig. 1A and C, both proteins were overexpressed and exhibited the expected subcellular localization described previously (32). Specifically, immunoblotting with both anti-GFP and anti-Nup88 antibodies (Fig. 1C) allowed the present study to identify the Nup88-GFP fusion protein and verify its expression level relative to endogenous Nup88. Using these cell lines, the present study performed RT-PCR. This analysis confirmed that Kif7 expression at the RNA level was markedly decreased in the Nup88-overexpressing cells (Fig. 1B). This decreased expression of Kif7 was subsequently confirmed at the protein level (Fig. 1C). If the expression of Kif7 is epigenetically regulated, demethylation of DNA should restore its expression. Thus, it was examined whether the decreased expression of Kif7 was due to epigenetic regulation. Expression of Kif7 in cells stably overexpressing Nup88-GFP was monitored after treatment with a DNA methylation inhibitor, 5-aza-2′-deoxycytidine, which not only inhibits the activity of DNA methyltransferases (DNMTs) but also induces selective degradation of DNMT1 (35). In the experiment, cells overexpressing GFP or Nup88-GFP were cultured in standard medium containing the inhibitor (0 days) and then cells were collected every 2 days for up to 6 days to estimate the expression of Kif7 by immunoblotting. As shown in Fig. 1D, the expression of DNMT1 in both cell lines was markedly reduced 2 days after treatment with the inhibitor and the effect persisted for at least 6 days, indicating that the inhibitor remained effective throughout this time period. Under the given experimental conditions, expression of Kif7 in control cells was largely unaffected (Fig. 1D; lanes 1–4), while its expression in the Nup88-GFP overexpressing cells markedly recovered by days 4 and 6 (Fig. 1D; lanes 7 and 8). These results suggested that the expression of KIF7 is regulated by DNA methylation in Nup88-overexpressing HeLa cells. It has been reported that Drosophila Nup88 binds to silent chromatin (36). Therefore, it was examined whether Nup88 binds to the KIF7 promoter and contributes to DNA methylation. However, ChIP assays revealed that Nup88 did not specifically bind to the KIF7 promoter (Fig. 1E).

Decreased expression of Kif7 has been reported to induce malignant phenotypes in prostate cancer cell lines, which can be suppressed by the overexpression of a Kif7 fragment (Kif7-CC) comprising a discontinuous coiled-coil and a globular C-terminal tail domain (29). Thus, complementation assays were performed to investigate whether the decreased expression of Kif7 also contributes to the malignant phenotype in Nup88-overexpressing cells. In these experiments, the migration activity of GFP- or Nup88-GFP-overexpressing cells was assessed using co-overexpression of mCherry or mCherry-tagged truncated Kif7 (514–1343 aa). The tagged protein was derived from human KIF7 cDNA (BC112271) encoding a region largely comparable to Kif7-CC but lacking a portion of the Gli2-binding domain (481–541 aa). As shown in Fig. 2A and B, the expression of mCherry and mCherry-tagged truncated Kif7 was confirmed. Immunoblotting using anti-mCherry and anti-Kif7 antibodies confirmed that the transiently expressed proteins were correctly produced as intended fusion proteins at the expected molecular weights. Using these validated cells, live cell imaging analysis was performed, and it was found that, in cells expressing mCherry alone, the migration velocity of Nup88-GFP-overexpressing cells was almost twice that of GFP-overexpressing cells (Fig. 2C and Video S1). By contrast, in cells expressing mCherry-tagged truncated Kif7, the migration velocity of Nup88-GFP-overexpressing cells was suppressed to a level comparable to that of GFP-overexpressing cells (Fig. 2C and Video S1). These results indicate that downregulation of Kif7 contributes to the enhanced cell migration.

Kif7 is a major negative regulator of the mammalian Hh pathway. Therefore, decreased expression of Kif7 was expected to activate the Hh pathway. Activation of the Hh pathway leads to increased expression of Gli1, which can promote malignant phenotypes in certain cancers (23). Thus, it was hypothesized that decreased expression of Kif7 might lead to malignant phenotypes by promoting expression of Gli1 via the Hh pathway in Nup88-overexpressing cells. To examine this possibility, immunoblotting was used to compare the expression of Gli1 between GFP- and Nup88-GFP-overexpressing cells. As expected, the expression of Gli1 was enhanced in Nup88-GFP-overexpressing cells (Fig. 3A and B). However, Gli1 expression was unaffected when both Gli2 and Gli3, which are involved in Gli1 transcription in the Hh pathway, were depleted by RNA interference in either GFP- or Nup88-GFP-overexpressing HeLa cells (Fig. 3C-H). These findings suggest that the Nup88-induced increase in Gli1 expression is independent of the Hh pathway. The expression of Gli1 is known to be regulated by oncogenic signaling pathways, such as phosphoinositide 3-kinase-Protein kinase B-mechanistic target of rapamycin (PI3K-AKT-mTOR), Transforming Growth Factor-β (TGF-β), and Ras-Rapidly accelerated Fibrosarcoma (Raf)-Mitogen-Activated Protein Kinase/ERK Kinase (MEK-ERK) (Ras-Raf-MEK-ERK) pathways (23,24). The present study therefore examined the involvement of these pathways in the expression of Gli1. Although PI3K-AKT-mTOR and TGF-β signaling were not specifically activated by the overexpression of Nup88 in HeLa cells (Fig. S1), the levels of phosphorylated (p) ERK1/2 increased in Nup88-GFP-overexpressing cells (Fig. 4A and B). Furthermore, blocking the phosphorylation of ERK with U0126, an inhibitor of MEK1/2, strongly suppressed the expression of Gli1 (Fig. 4C and D). These results indicated the involvement of the MEK-ERK pathway in Gli1 expression. Ras is a major upstream regulator of the MEK-ERK pathway; the present study therefore examined the involvement of Ras in the expression of Gli1. Among three Ras isoforms, Kirsten's rat sarcoma virus (K-Ras) is the predominant isoform in HeLa cells (37). When K-Ras was depleted by RNA interference, the level of pERK1/2 remained unchanged (Fig. 4E and F), suggesting that Nup88 activates the MEK-ERK pathway to regulate the expression of Gli1 independently of K-Ras.

Aberrant expression of Gli1 induces various malignant phenotypes in cancer (23). Therefore, the present study examined whether increased expression of Gli1 in Nup88-overexpressing HeLa cells contributes to malignant phenotypes. To investigate the effect of Gli1 on motility, the migration velocity of Nup88-overexpressing HeLa cells with or without knockdown of Gli1 was monitored using live-cell single-cell imaging analysis. Knockdown of Gli1 in both cell lines was confirmed by immunoblotting (Fig. 5A). When control siRNA was transfected into both cell lines, the migration velocity of Nup88-GFP-overexpressing cells was nearly double that of GFP-overexpressing cells (Fig. 5B and Video S2). This finding was consistent with the result shown in Fig. 2C. By contrast, when GLI1-specific siRNA was transfected, the migration velocity of cells overexpressing Nup88-GFP was suppressed to a level comparable to that of cells overexpressing GFP (Fig. 5B and Video S2). These results indicate that the increase in Gli1 expression contributes to cell motility.

The present study demonstrated that the overexpression of Nup88 in HeLa cells promoted cell migration through the epigenetic repression of Kif7 (Fig. 2) and the MEK-ERK pathway-dependent increase of Gli1 (Fig. 5). Since both Kif7 and Gli1 are components of the Hh signaling pathway, the possibility that the expression of these two proteins might be interdependently regulated was considered. Therefore, the present study examined the effect of Kif7 knockdown on the expression of Gli1, as well as the effect of the overexpression of Gli1 on the expression of Kif7. However, neither the knockdown of Kif7 (Fig. 6A-C) nor the overexpression of Gli1 (Fig. 6D and E) affected the expression of the other protein. These results suggested that the changes in expression of Kif7 and Gli1 are independent of each other.

The mechanism by which the truncated Kif7 suppresses the migration activity of Nup88-overexpressing cells remains unclear (Fig. 2). The present study considered two plausible mechanisms. One possibility is that co-overexpression of the truncated Kif7 reduces the expression levels of phosphorylated ERK and/or Gli1. The other possibility is that the truncated Kif7 binds to Gli1 and suppresses its activity, as it does for Gli2/3, although the truncated Kif7 lacks a portion of the Gli2-binding domain. To address the former possibility, expression of phosphorylated ERK and Gli1 was examined in cells co-overexpressing Nup88-GFP and truncated Kif7. However, overexpression of the truncated Kif7 did not affect the expression of either phosphorylated ERK or Gli1 (Fig. 7A and B). For the latter possibility, a GFP-pulldown assay using parental HeLa cells co-overexpressing truncated Kif7 and GFP-tagged Gli1 was performed to test their interaction. The pulldown assay showed that the truncated Kif7 specifically co-precipitated with Gli1 (Fig. 7C). Taken together, these results suggested that the truncated Kif7 interacts with Gli1 and potentially sequesters it, thereby inhibiting its function without affecting its expression levels.