The serine/threonine kinase Polo-like kinase 1 (Plk1) plays a pivotal role in cell proliferation and has been validated as a promising anticancer drug target. However, very limited success has been achieved in clinical applications using existing Plk1 inhibitors, due to lack of sufficient specificity toward Plk1. To develop a novel Plk1 inhibitor with high selectivity and efficacy, we designed and synthesized a pyrrole-imidazole polyamide–Hoechst conjugate, PIP3, targeted to specific DNA sequence in the PLK1 promoter. PIP3 could specifically inhibit the cell cycle–regulated Plk1 expression and consequently retard tumor cell growth. Cancer cells treated with PIP3 exhibited severe mitotic defects and increased apoptosis, whereas normal cells were not affected by PIP3 treatment. Furthermore, subcutaneous injection of PIP3 into mice bearing human cancer xenografts induced significant tumor growth suppression with low host toxicity. Therefore, PIP3 exhibits the potential as an effective agent for targeted cancer therapy. Mol Cancer Ther; 17(5); 988–1002. ©2018 AACR.
Cell-cycle deregulation is a hallmark of human cancer and, as such, targeting the cell cycle in cancer has become a validated therapeutic approach. The Polo-like kinases (Plks), including Plk1, Plk2, Plk3, and Plk4, are a conserved subfamily of serine/threonine protein kinases that play pivotal roles in cell-cycle regulation and genome integrity maintenance (1). Among the four human Plks, Plk1 has been studied most extensively because of its tight association with tumorigenesis (2). Plk1 as a mitotic kinase has been implicated in multiple steps of mitotic progression, such as mitotic entry, centrosome maturation, spindle assembly, chromosome segregation, and cytokinesis (3, 4). Overexpression of Plk1 has been observed in a wide range of human cancers and often correlates with poor prognosis (5, 6). Interference with Plk1 functions by either siRNA or small-molecule inhibitors resulted in mitotic arrest followed by apoptotic cell death in most cancer cells (7–9). Interestingly, specific inhibition of Plk1 had only very minor effect on normal cell growth in vitro and in vivo (10, 11). More importantly, adult mice with inducible knockdown of Plk1 show normal viability, no weight regulation or energy metabolism disturbance, normal histologic morphologies in essential organs, as well as no obviously increased apoptosis (11). On the basis of these studies, Plk1 is suggested to be a specific regulator of tumor cell growth and thus an attractive drug target for cancer therapy (12). To date, a number of chemicals that target ATP-binding site of Plk1 catalytic domain have been developed (13–15). Although some of these chemicals show promising antitumor activity in preclinical experiments, owing to the high degree of similarity of ATP-binding clefts among various kinases, these inhibitors often display low specificity toward Plk1, thus significantly limiting their further clinical use for cancer therapy (13–15).
Pyrrole-imidazole (Py-Im) polyamides (PIP) are a class of synthetic molecules that recognize and bind to DNA with high affinity and specificity (16). Sequence-specific binding of PIPs in the minor groove of double-helical DNA alters the DNA architecture, thus blocking the potential protein–DNA interaction on the specific site (17). Because of this characteristic, designed PIPs can act as gene expression modulators by competing with endogenous transcription factors in binding to their target DNA elements in gene promoters. The DNA recognition by PIPs depends on the linear sequence of pyrrole and imidazole amino acids. Pyrrole prefers binding to A, T, and C bases, but not G base, whereas imidazole selectively targets the G base. In hairpin PIPs, a Py/Py pair recognizes an A/T or T/A base pair, an Im/Py pair binds to a G/C base pair, whereas a Py/Im pair recognizes a C/G base pair (18). Guided by these DNA recognition rules, a number of sequence-specific DNA-binding PIPs have been successfully developed to modulate the expression of disease related genes, such as TGFβ1 (19), NF-κB (20), MMP-9 (21), ABCA1 (22), EVI1 (23), LOX-1 (24), and KRAS mutations (25). However, further development of PIPs for clinical applications has met with little success, presumably due to the low specificity and inadequate nuclear uptake of these molecules (26).
To develop a novel Plk1 inhibitor with high selectivity and efficacy, we designed and synthesized a Py-Im polyamide–Hoechst 33258 conjugate targeting the specific DNA sequence (ggattttaaatcc) in PLK1 promoter. Our results demonstrated that conjugating Hoechst with PIP could recognize longer DNA sequences and also exhibit excellent cellular permeability and nuclear uptake. Furthermore, this agent could specifically inhibit the cell cycle–dependent expression of Plk1 and consequently retarded the tumor cell growth. Consistent with the previously reported phenotypes of Plk1 RNAi, the transcriptional repression of PLK1 significantly delayed the mitotic progression of cancer cells, while normal cells were not affected. In addition, this agent showed effective antitumor activity in various mouse tumor xenograft models. Overall, these results suggest that this PIP–Hoechst conjugate is a promising candidate for Plk1-targeted cancer therapy.
Materials and Methods
All cell lines were obtained from Procell Life Science & Technology Co., Ltd. on April 8, 2015. Cells were maintained at 37°C in a humidified incubator with 5% CO2 and were used within 3 months of resuscitation. No independent authentication was carried out to check the identification, but all cultured cells retained the characteristic phenotype as shown on the ATCC website.
Reagents and solvents were purchased from standard suppliers and used without further purification. Py-Im polyamides were synthesized by following established solid-phase synthesis protocol 1,2. Detailed synthesis and modification procedures are shown in Supplementary Data S1.
The surface plasmon resonance (SPR) assays were performed using a Proteon XPR36 instrument (Bio-Rad). Biotinylated hairpin DNAs were purchased from Invitrogen. The sequence of tested DNA was: 5′- biotin- GGGTTTGGATTTTAAATCCTTTTGGATTTAAAATCCAAACCC - 3′. The biotinylated DNA hairpins were immobilized on streptavidin-coated ProteOn GLH sensor chip to obtain the desired immobilization level (∼1,000 RU). All of the binding operations were performed with the continuous running HBS-EP buffer (10 mmol/L HEPES, pH 7.4, 150 mmol/L NaCl, 3 mmol/L EDTA, and 0.005 % surfactant P20) with 0.1 % DMSO at 25˚C. Samples with various concentrations were prepared in the HBS-EP buffer with 0.1 % DMSO and injected at a flow rate of 20 μL/minute. Sensorgrams of bound polyamides were generated by subtracting the signal of an injection passing over an uncoupled channel from that of immobilized channel. ProteOn Manager software (version 3.1) was used to determine KD values from the association rates (Ka) and dissociation rates (Kd). The proper binding models (models for simultaneous Ka/Kd, 1:1 binding with drifting baseline or 1:1 binding with mass transfer; models for general fitting–steady-state affinity) were used for fitting the sensorgrams to give better fitting (3). All the experiments were performed in triplicates.
Cell culture and synchronization
HeLa, A549, and HUVEC cells were maintained in DMEM (HyClone), supplemented with 10% FBS (HyClone), 100 U/mL of penicillin and 100 μg/mL of streptomycin at 37°C in 5% CO2. To obtain cell synchronization at mitosis, either nocodazole or thymidine block method was used. For nocodazole treatment, cells were incubated in the presence of 100 ng/mL nocodazole for 12 hours. For triple thymidine block and release, cells were treated with 2 mmol/L thymidine for 16 hours. Then, cells were washed in PBS and incubated with fresh medium. Eight hours later, 2 mmol/L thymidine was again added to the medium. This procedure was repeated once more. After the third thymidine block, cells were released into fresh medium and harvested at different time points for cell-cycle analysis.
Plasmids, lentivirus, and stable cell lines
The human PLK1 promoter with a size of 459 bp (−474 to −16 bp upstream of initiation codon ATG) was amplified from genomic DNA of HeLa cells and ligated into PGL4.19 vector. The mutant promoter construct was generated using QuikChange Site-Directed Mutagenesis Kit (Agilent Technologies). The human AURORA KINASE A promoter with a size of 380 bp (−4400 to −4021 bp upstream of initiation codon ATG) was also amplified from genomic DNA of HeLa cells and ligated into PGL4.19 vector, following introducing Flag-Plk1 fragments right after the promoter region to generate the Aurora A:Flag-Plk1 construct for rescue experiments. To generate cell lines stably expressing PLK1 promoter-luciferase, HeLa cells were transiently transfected with constructed wild-type or mutant PLK1 promoter-luciferase plasmids using Megatrans 1.0 (OriGene). Transfection-positive cells were selected by using G418 (800 μg/mL). After 30 days of selection, remaining cells were collected for luciferase activity assays. pHIV-H2B-mRFP expression plasmid from Addgene was used to generate cell lines stably expressing H2B-RFP. In brief, HEK293T cells were cotransfected with pHIV-H2B-mRFP and packaging plasmids (pMD2.G and psPAX2) to produce lentiviruses. Purified lentiviruses were then used to infect HeLa, A549, hTERT-RPE1, or HUVEC cells. Polybrene (4 μg/mL) was added to increase the infection efficiency. Stably transfected cells were sorted by flow cytometer (BD FACSAria III).
Western blot analysis
Briefly, cells were lysed in ice-cold RIPA buffer supplemented with protease inhibitor cocktail-EDTA for 1 hour and then centrifuged at 14,000 × g for 15 minutes at 4°C. Lysed proteins from the supernatant were resolved by SDS-PAGE electrophoresis and electrotransferred onto polyvinylidene difluoride membranes. Blots were probed with primary antibodies for 2 hours at room temperature and then with secondary antibodies coupled to horseradish peroxidase for 1 hour at room temperature. Immunoreactivity was detected by using Luminata Classico Western HRP Substrate (Millipore). The following antibodies were used for Western blot analysis: Plk1 (Santa Cruz Biotechnology; sc-17783), Plk2 (Santa Cruz Biotechnology; sc-25421), Plk3 (BD Biosciences; 556518), Plk4 (Proteintech; 12952-1-AP), and p-Histone H3 (Santa Cruz Biotechnology; sc-8656-R), BubR1 (Santa Cruz Biotechnology; sc-16195), Cdc25c (Santa Cruz Biotechnology; sc-13138), Wee1A (Santa Cruz Biotechnology; sc-5285), PARP (Cell Signaling Technology; 46D11), caspase-3 (Cell Signaling Technology; 5A1E), and GAPDH (Proteintech; 60004-1-Ig). Particularly, antibodies against BubR1, Cdc25c, and Wee1A can distinguish the corresponding phosphorylated target proteins by significant shift in electrophoretic mobility.
Reverse transcription and qRT-PCR
Total RNA was isolated using TransZol (TransGen Biotech) according to the manufacturer's instructions. After total RNA concentration was determined, 1 μg of total RNA was used for cDNA synthesis using the TransScript First-Strand cDNA Synthesis Kit (TransGen Biotech). qRT-PCR was performed in triplicate using TransStart Tip Green qPCR SuperMix (TransGen Biotech) and detected using the CFX96 real-time PCR detection system (Bio-Rad). The oligonucleotides used were as follows: Plk1 (sense: 5′-ACAGTTTCGAGGTGGATGTG-3′; antisense: 5′-GGTTGATGTGCTTGGGAATAC-3′); IL6 (sense: 5′- AGTGAGGAACAAGCCAGAGC-3′; antisense: 5′- GTCAGGGGTGGTTATTGC AT-3′); IL8 (sense: 5′-TCCTGATTTCTGCAGCTCTGT; antisense: 5′-AAATTTGGGGTGGAA AGGTT-3′); GAPDH (sense: 5′-ACTAACATCAAATGGGGTGAGG-3′; antisense: 5′-AAAGTTGTCATGGATGACCTTG-3′). Gene expression was normalized against GAPDH.
Luciferase activity assay
In brief, HeLa cells stably expressing wild-type or mutant PLK1 promoter-luciferase were grown in 12-well plates (30,000 cells/well). Cells were then synchronized by double thymidine block and released into fresh medium in the absence or presence of 10 μmol/L of PIP3. Luciferase activity was measured by using the Dual-Luciferase Reporter Assay System (Promega, E1960).
Cell viability assay
Cell Counting Kit-8 (TransGen Biotech) was used to evaluate the effect of Py-Im polyamide on cell proliferation. Cells were plated on 96-well plates (5,000/well) and treated with different concentrations of PIP3 for 72 hours. Ten microliters of the cell proliferation reagent CCK8 was then added to each well and incubated for additional 2 hours at 37°C. The absorbance at 450 nm (A450 nm) was measured by using Multiskan GO (Thermo Fisher Scientific) and was directly proportional to live cell numbers. IC50 values were calculated by dose–response inhibition module in GraphPad Prism 5. Assays were performed in biological triplicates.
Cells grown on glass coverslips were fixed with 4% paraformaldehyde in PBS for 15 minutes, permeabilized with 0.5% Triton X-100 in PBS for 15 minutes, washed in PBS twice, and blocked with 3% BSA for 1 hour at room temperature. The primary antibodies (mouse anti–α-tubulin at 1:500, rabbit anti–p-histone H3 at 1:200) and the fluorophore-conjugated secondary antibodies were sequentially incubated with cells for 2 and 1 hour, respectively. Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL)–positive apoptotic cells in tumor sections were detected by using In Situ Apoptosis Detection Kit (TransGen Biotech). All immunofluorescence images were taken by using Olympus BX3-CBH fluorescence microscopy or Leica CTR6000 confocal laser microscopy.
Chromatin immunoprecipitation assay
The chromatin immunoprecipitation (ChIP) assay was performed by using Simple ChIP Enzymatic Chromatic IP Kit (#9003, Cell Signaling Technology) according to the manufacturer's instructions. In brief, HeLa cells grown in 100-mm dishes were treated with or without PIP3 for 3 days. Cells were then incubated in 1% formaldehyde for 10 minutes at room temperature for protein–DNA crosslink. The crosslink process was stopped by adding 1× glycine for additional 10 minutes. Then, cells were collected with ice-cold PBS buffer and digested by nuclease enzyme for 20 minutes at 37 °C. The samples were sonicated for 10 minutes to shear genomic DNA to an average fragment size of 200 to 1,000 bp and were centrifuged at 15,000 rpm at 4°C for 10 minutes. After removing the pellets, the supernatant was incubated overnight at 4°C with 5 μg antibodies against Sp1 (17-601, Millipore) or NF-YB (sc-13045, Santa Cruz Biotechnology), followed by another incubation with Protein G Magnetic Beads at 4°C for 2 hours. Magnetic beads were washed and then incubated with ChIP elution buffer for additional 30 minutes at 65°C with gentle vortex (1,200 rpm). The eluates were further incubated with proteinase K for 2 hours at 65°C for reverse cross-links. DNA was recovered by kit purification. Quantitative PCR was performed using the following primers: PLK1-promoter forward: 5′- ATCCACGCCGGGTTTGGTTTC-3′; PLK1-promoter reverse: 5′- ACGCGGCGCTGGGAACGTTAC -3′.
In vivo xenograft experiments
The animal protocol was approved by Institutional Ethical Committee of Animal Experimentation of Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences (Beijing, China). All animals were maintained on a standard light/dark cycle. A549, NCI-H1299, and MDA-MB-231 cells were implanted subcutaneously into the left flank of 4- to 5-week-old female BALB/c nude mice. When the tumor volume (0.5 × L × W2) reached about 50 mm3, vehicle (0.1% DMSO: normal saline) or PIP3 (1 mg/kg) was injected into the peri-region every 3 days for seven injections. Animals were sacrificed 3 days after the final injection. Animal weight and general health care were monitored daily. At least 6 animals were used for each treatment group.
The cryosections of excised tumors were fixed in 4% paraformaldehyde in PBS and permeabilized with 0.1% Triton X-100 in PBS for 15 minutes. The endogenous peroxidase activity was blocked by incubation with 0.3% H2O2 for 30 minutes. Samples were then washed with PBS twice, treated with blocking buffer containing 3% BSA for 30 minutes at room temperature, and stained with primary antibodies overnight at 4°C. Then, the samples were washed with PBS three times and incubated with secondary antibodies conjugated with HRP for 1 hour at room temperature, followed by substrate reaction using diaminobenzidine (DAB) precipitation (Vector Laboratories, Enzo Life Sciences). Reactions were stopped by adding distilled water. Samples were finally stained with hematoxylin and sealed with neutral resin. Images were captured using Olympus BX53 microscope.
The effect of Py-Im polyamide on cell-cycle process was examined by time-lapse microscopy. Cells stably expressing histone H2B-RFP were placed on 35-mm culture dishes. After treatment with PIP3 for 72 hours, cells were imaged under 10× objective at 2-minute intervals for a total time of 6 hours. Point visit was used to follow cells in multiple fields of the same sample. Mitotic stages were defined as follows: prophase, chromosomes are starting to condense; metaphase, condensed chromosomes are well aligned at midplate; anaphase, a pair of condensed sister chromatids are segregating from each other. Time lapse was performed by using Olympus IX81 live cell imaging workstation. For phospho-histone H3 counting, positive cells were manually labeled and counted.
RP-HPLC analysis of PIP stability
HeLa cells treated by 10 μmol/L PIP3 for 72 hours were lysed with RIPA buffer on ice for 1 hour, followed by 1-minute ultrasonication (50% amplitude, sonication 5s + interval 5s, six cycles). The whole-cell lysates were incubated in 100°C water bath for 15 minutes to release the PIP3 from DNA. After centrifugation at 14,000 × g for 10 minutes, the supernatant was collected to monitor the dissolved PIP3 by reverse-phase high-performance liquid chromatography (RP-HPLC) analysis. Pure PIP3 dissolved in water was used as control. Analytic RP-HPLC was performed at room temperature on the Shimadzu LC 20 with UV detector SPD-20A using Inertsil ODS-SP column (4.6 × 250 mm, 5 μm, 100 Å, UV absorbance was measured at 310 nm). The RP-HPLC gradient was started at 10% of B (MeCN), then increased to 100% of B over 30 minutes (A: 0.1% TFA in water).
At least three independent experiments were used for quantification. GraphPad Prism 5 was used for statistical analysis, and all data passed normality tests. All Student t tests were performed assuming Gaussian distribution, and P < 0.05 is considered to be statistically significant.
Design of Plk1-targeting PIPs
Plk1 mRNA expression is low at G1–S transition, increases during the S-phase, and is maximally expressed during G2–M phase (27). Knockdown of Plk1 expression blocks the cells to go through mitotic progression and consequently induces apoptosis (28). Thus, the transcriptional suppression of PLK1 is a promising approach to kill tumor cells. To design appropriate Py-Im polyamides targeting Plk1 for transcriptional block, we first analyzed the PLK1 promoter structure. As previously reported, the activation of Plk1 transcription during mitosis is controlled by at least three activating elements in PLK1 promoter, among which are two transcription factor–binding sites: the SP1 consensus site and the NF-Y–binding CCAAT box (29).
Thus, targeting these two elements by competitive binding of Py-Im polyamide would be an effective way to inhibit Plk1 expression in mitosis. Because the promoter structures of various PLK1 family members are largely different, this strategy allows the Plk1 inhibition to be isoform specific. In a search for the potential targeted DNA sequence between these two elements, we coincidently found a sequence 5′-GGATTTTAAATCC-3′, which possesses a pair of inverted repeats (5′-GGATTT-3′) and is highly A/T rich and thus is a perfect target to a Py-Im hairpin polyamide with a Hoechst 33258 conjugated at the γ-turn position (Fig. 1A). Accordingly, a hairpin Py-Im polyamide–Hoechst conjugate was designed and synthesized. We expected that this agent would have the potential to inhibit the transcriptional activity of both elements. For comparative studies, a series of PIP derivatives were also prepared (Fig. 1B). The results of the SPR experiments showed that the Hoechst conjugation at either γ-turn position or at the C-terminus of PIPs did not alter their binding affinity with targeted DNA sequence significantly (Fig. 1C).
Effect of PIPs on Plk1 expression
We then performed immunoblotting to examine the effect of designed PIPs on Plk1 expression in cultured HeLa cells. Our data showed that among all six designed PIPs, only PIP3, which possesses a Hoechst conjugation at the γ-turn, had the ability to suppress Plk1 expression (Fig. 2A). Unsynchronized cells treated with PIP3 at various concentrations for 3 days showed dose-dependent reduction of Plk1 expression level compared with that of control cells (Fig. 2A). We further showed that PIP3-mediated Plk1 suppression is time dependent, and 1 day of PIP3 treatment could reduce the Plk1 expression effectively (Fig. 2B). Then, we investigated whether PIP3 could suppress the cell cycle–regulated Plk1 expression. By incubation with nocodazole for 12 hours, HeLa cells were largely arrested into mitosis, as evidenced by accumulation of phospho-histone H3. In the absence of PIP3, significantly increased expression of Plk1 was detected in cells with nocodazole treatment. When the cells were preincubated with 10 μmol/L of PIP3 for 24, 48, or 72 hours, however, the increase in Plk1 expression upon nocodazole treatment was nearly abolished (Fig. 2C). Similar results were also obtained by using A549, a human non–small cell lung cancer cell line (Fig. 2D). These results indicate that PIP3 is a potential Plk1 expression inhibitor. However, it is also possible that PIP3 treatment delayed the mitotic entry and thus induced suppression of Plk1 expression as a side effect. To test this possibility, we further examined the Plk1 expression upon PIP3 treatment within pure mitotic cells collected by “shake-off” after nocodazole treatment. As shown, Plk1 expression was also inhibited by PIP3 treatment in these cells (Fig. 2E), which strongly suggested that PIP3 suppresses Plk1 expression directly. Consistent with the abovementioned results, PIP3 treatment also significantly inhibited the Plk1-mediated phosphorylation of BubR1, Cdc25C, and Wee1 in mitosis (Fig. 2F). Together, these results suggest that PIP3 is capable of inhibiting the oscillatory expression of Plk1 during cell-cycle progression.
PIP3 was designed to target PLK1 promoter and was expected to suppress the transcription of PLK1. Therefore, we continued to evaluate the effect of PIP3 on PLK1 transcription. qRT-PCR results revealed that the Plk1 mRNA level was elevated by 7-fold upon 12 hours of nocodazole treatment in HeLa cells. In contrast, pretreatment of cells with 10 μmol/L of PIP3 for 48 hours significantly reduced this increase by approximately 50% (Fig. 2G). To verify these results, we further measured the inhibitory effect of PIP3 on PLK1 promoter activity in HeLa cells that stably express luciferase gene under the PLK1 promoter. Luciferase activity was assessed in cultures synchronized by a thymidine block and release. In line with the previous studies (29, 30), the PLK1 promoter displayed a cell cycle–regulated activity with a relative low level at the S-phase (30-minute release from thymidine block) and an approximately 7-fold of increased level in G2–M phase (10-hour release from thymidine block; Fig. 2H). Upon PIP3 treatment, however, the PLK1 promoter activities at the S-phase and G2–M phase were both inhibited. Similar to the abovementioned qRT-PCR results, cells within G2–M phase showed a more substantial sensitivity to PIP3 treatment (Fig. 2H). Together, these results demonstrate that PIP3 is capable of suppressing PLK1 transcription by modulating its promoter activity.
A previous study showed that the PIP2 was also capable of recognizing NF-κB–binding site (GGAT) and thereby impaired the expression of downstream genes of NF-κB signaling, such as IL6 and IL8 (20). In our study, PIP2 indeed showed strong inhibitory effect on the TNFα-stimulated expression of both IL6 and IL8, whereas PIP3 displayed much less inhibitory activity toward these genes (Supplementary Fig. S1), suggesting that the Hoechst conjugation in PIP3 provides additional sequence selectivity.
Hoechst 33258 was reported to have toxicity in a high concentration (31). We then tested whether Hoechst 33258 in a PIP3-comparable working concentration would affect cell cycle and thus led to Plk1 suppression as a side effect. Our results showed that 10 μmol/L of PIP3 penetrated cells to an extent similar to 1 μmol/L of Hoechst 33258 after 24 hours of incubation (Supplementary Fig. S2A), and Plk1 expression was only inhibited by 10 μmol/L of PIP3 but not by 1 μmol/L of Hoechst 33258 incubation for 72 hours (Supplementary Fig. S2B), demonstrating that Hoechst 33258 on its own cannot suppress Plk1. We also examined the stability of the covalent linkage between PIP and Hoechst 33258 within PIP3 molecules. By applying RP-HPLC, we showed that the PIP3 molecules that were incubated with HeLa cells for up to 3 days still had a robust stability (Supplementary Fig. S2C). Together, these results exclude the possibility that Plk1 suppression by PIP3 is due to deconjugated Hoechst 33258 from PIP3.
Cellular uptake and nuclear localization of PIP3
We next examined the cellular uptake and subcellular distribution of PIP3 in living cells. As PIP3 possesses a Hoechst conjugation, it could be monitored in living cells directly. The experiments were performed in HeLa cells expressing histone H2B-RFP in which the nuclei were fluorescently labeled with RFP. We first investigated the time required for PIP3 uptake into cells. As shown, PIP3 began to enter into cells as early as 1 hour after incubation and gradually accumulated in nuclei from 3 hours of incubation (Supplementary Fig. S3A). At 24 hours of incubation, PIP3 was strongly accumulated in nuclei, although a large amount of PIP3 remained on plasma membrane. These data suggest that PIP3 is permeable to both plasma membrane and nuclear envelope (Supplementary Fig. S3A; Fig. 2I). In particular, the nuclear localization of PIP3 could be detected in both interphase and mitotic cells (Fig. 2I). The same pattern of cellular uptake and nuclear recruitment of PIP3 was also observed in A549 cells (Supplementary Fig. S3B and S3C), indicating that the cellular uptake and nuclear access of PIP3 are cell type independent. Using the same methodology, we further evaluated the cellular uptake and subcellular distribution of other three PIPs with Hoechst conjugation. Under the same condition, cells with PIP4 and PIP5 showed no detectable fluorescence in either interphase nuclei or mitotic nuclei. For PIP6, although some weak fluorescence signal was detected in mitotic nuclei, no signal was observed in interphase nuclei, and the majority of the compound remained in cytoplasm and formed polyamide vesicular bodies (Supplementary Fig. S4). These data clearly showed that only PIP3 has the ability to access the nucleus in living cells, thus providing a good explanation why among all four designed PIP–Hoechst conjugates, only PIP3 showed unique inhibitory activity for PLK1 transcription.
Specificity of Plk1 inhibition by PIP3
To determine the inhibitory specificity of PIP3 on Plk1 transcript, we designed a mismatch polyamide PIP7 that recognizes the sequence 5′-A/TGGCCA/T-3′. The PIP7 also possesses a Hoechst conjugation at the γ-turn position but has decreased binding affinity with PLK1 promoter as compared with PIP3 (Fig. 3A). Consistent with this result, Western blotting revealed that Plk1 expression could not be suppressed by this mismatch PIP7 within the effective concentration range of PIP3 (Fig. 3B). This result was further confirmed by qRT-PCR (Fig. 3C) and luciferase activity assay (Fig. 3D). To validate the DNA-binding specificity of PIP3, we introduced a series of substitution mutations in PIP3-targeting sequences within Plk1-luciferase construct (Fig. 3E). As shown, PIP3 could not inhibit the activity of PLK1 promoter, which possesses mutations in both polyamides recognized sequences (mtL+R; Fig. 3F). However, single mutation of one of the two PIP-targeting DNA sequence (mtL and mtR) or mutation of Hoechst-recognized DNA sequence alone (mtM) all showed partial sensitivity toward PIP3. Together, these data demonstrate that PIP3 suppresses Plk1 in a sequence-specific mode, and both PIP- and Hoechst-recognized DNA sequences contribute to this specificity. To further investigate whether PIP3 suppresses Plk1 expression by competing with the endogenous transcription factor SP1 and NF-Y, we employed a chromatin immunoprecipitation assay. Quantitative PCR results showed that the amount of DNA fragments immunoprecipitated by antibodies against SP1 and NF-Y was obviously reduced in samples treated with PIP3 compared with that of control group (Fig. 3G). These results suggest that PIP3 functions by disrupting the binding of SP1 and NF-Y to the PLK1 promoter. Moreover, we also evaluated the potential inhibitory effect of PIP3 on other Plk family members. Despite that PIP3 showed strong inhibitory ability toward Plk1, it did not reduce the expression level of Plk2, Plk3, and Plk4 (Fig. 3H), demonstrating that the suppression of Plk1 by PIP3 is Plk isoform specific.
Mitotic defects induced by PIP3 treatment in cancer cells
Consistent with the fundamental role of Plk1 in various stages of mitosis, either deletion of Plk1 by RNAi or inhibition of its activity by Plk1 inhibitor significantly delayed the mitotic progression (9, 10). We therefore examined the potential impact of PIP3 on mitotic progression in synchronized HeLa cells by a triple thymidine block and release (Fig. 4A). The mitotic index was examined by staining of mitotic marker phospho-histone H3. As shown, control HeLa cells started to enter mitosis after 6 hours, reached a peak of mitosis after 8 hours, and exited mitosis at 12 hours of release from the thymidine-induced G1–S arrest. Although the kinetics of mitotic progression in PIP3-treated HeLa cells were similar to that of control cells, the mitotic index was significantly increased in these cells (Fig. 4B). We further calculated the percentages of the cells at each stage of mitosis. Our data showed that the number of cells in the metaphase was dramatically increased, while the cells in anaphase/telophase were decreased upon PIP3 treatment (Fig. 4C). Thus, these data strongly suggest that PIP3 treatment induced mitotic arrest. The nature of the mitotic arrest was further determined using time-lapse microscopy in cells stably expressing histone H2B-RFP. In control cells, the progression from chromosome condensation to cytokinesis required about 1 hour. In comparison, when cells were treated with 20 μmol/L of PIP3, the total duration of mitosis was markedly increased to 2 hours on the average (Fig. 4F and G; Supplementary Movies S1 and S2). Separate analysis of the duration of each phase of the mitosis revealed that PIP3 treatment caused a delayed progression from prophase to anaphase (Fig. 4H). Consistent with these mitotic defects, increased frequencies of multipolar spindles and lagging chromatids were detected in PIP3-treated HeLa cells (Fig. 4I). These mitotic defects caused by PIP3 treatment were further confirmed in A549 cells (Fig. 4D–H; Supplementary Movies S3 and S4). Previous studies have demonstrated that Plk1 deletion or inhibition could cause apoptosis in cancer cells (8, 9). In line with this view, PIP3 treatment also induced apoptotic cell death in HeLa cells, as indicated by accumulated PARP cleavage, caspase-3 activation, and increased propidium iodide/Annexin V–positive cells identified by flow cytometry (Fig. 4J; Supplementary Fig. S5). Combined, these findings are consistent with the reported phenotype of Plk1 RNAi in cancer cells (10), further supporting that PIP3 inhibits Plk1 with high selectivity.
We further evaluated the effect of PIP3 on cell proliferation in a panel of human cancer cell lines from diverse organ derivations. The concentrations required to inhibit 50% of cell growth (IC50) were calculated after 3 days of exposure to PIP3. As indicated, the IC50 values of this cell panel were in the range of 1.2 to 13.23 μmol/L (Fig. 5A). Consistent with the previous finding that p53-deficient cells are more sensitive to Plk1 deletion (32), nearly half of the p53-null or mutant cancer cells showed relative lower IC50 values in comparison with the p53 wild-type cells (Fig. 5B), indicative of a close correlation between p53 status and efficacy of PIP3.
We also wonder whether the different sensitivity to PIP3 in various cell lines is correlated to their Plk1 expression levels. For this purpose, we examined the Plk1 expression level in a panel of cancer cell lines. As indicated, we found no obvious correlation between PIP3 sensitivity (as suggested by IC50) and the Plk1 richness implicated as either mRNA level (Supplementary Fig. S6A and S6D) or as protein level (Supplementary Fig. S6B and S6E) in these cell lines.
Cellular impact of PIP3 in nontransformed cells
Previous studies have shown that Plk1 knockdown could induce strong mitotic arrest followed by apoptosis in tumor cells but not in normal cells (10, 11). To evaluate the possible toxicity of PIP3 toward normal cells, we extended the phenotype analysis of PIP3 in two nontransformed cell lines, hTERT-RPE1, a human retinal pigment epithelial cell line that stably expresses human telomerase reverse transcriptase, and HUVEC, a human umbilical vein endothelial cell line. Although PIP3 was also capable of suppressing Plk1 expression in both cell lines at the concentration of 10 μmol/L, it had no obvious effect on either kinetics of mitotic progression (Supplementary Fig. S7A and S7B) or the distribution of mitotic cells over the different mitotic phases (Supplementary Fig. S7C and S7D) in these cells. Time-lapse microscopy revealed that both hTERT-RPE1 and HUVEC cells progressed through mitosis normally regardless that cells were treated with or without PIP3 (Supplementary Fig. S7E–S7G; Supplementary Movies S5–S8). Consistent with these observations, hTERT-RPE1 and HUVEC cells were less sensitive to PIP3, with IC50 values of 35.83 and 41.75 μmol/L, respectively (Fig. 5A). Combined, these results suggest a potentially low toxicity of PIP3 toward normal cells.
Rescue of PIP3-caused cell-cycle delay by exogenous expression of Plk1
To further demonstrate that PIP3-induced cell-cycle delay and cell growth suppression is due to the inhibition of Plk1 expression, we next performed a rescue experiment. For this purpose, we generated a Flag-Plk1 expression construct driven by the promoter of mitotic kinase Aurora A. As Aurora A has a similar cell-cycle expression pattern to Plk1, this construct (Aurora A:Flag-Plk1) can most closely mimic the cell cycle–controlled expression of Plk1 while avoiding the Plk1 overexpression–induced mitotic defect. Because the AURORA A promoter has no PIP3-binding sequence, we expected that the expression of Flag-Plk1 driven by AURORA A promoter is resistant to PIP3 treatment. Indeed, PIP3 treatment could not reduce this exogenous expression of Flag-Plk1 within both unsynchronized and pure mitotic cells (Fig. 6A). Then, we investigated whether transfection of this construct can rescue the mitotic delay induced by PIP3 treatment. By examination of the mitotic progression of synchronized HeLa cells, we found that mitotic index of Aurora A:Flag-Plk1–transfected cells, although higher than that of control group cells during late mitosis, was much less than that of PIP3-treated cells throughout the whole mitotic progression (Fig. 6B), demonstrating that expression of exogenous Flag-Plk1 partially rescued the mitotic defects induced by PIP3. We next examined whether cell growth inhibition by PIP3 can also be rescued by Aurora A:Flag-Plk1 transfection similarly. As shown, in HeLa, A549, H1299, and MDA-MB-231 cell lines, the cell growth inhibition upon PIP3 treatment were all greatly diminished by Aurora A:Flag-Plk1 expression (Fig. 6C–F). Together, these results demonstrate that PIP3 treatment induced cell-cycle delay, and cell growth inhibition is majorly through suppression of endogenous Plk1 expression, thus further confirming the great targeting specificity of PIP3 toward PLK1 promoter.
In vivo antitumor potential of PIP3
We next examined the in vivo activity of PIP3 against human lung cancer A549 xenograft in immunodeficient mice. When the tumor size reached an approximate volume of 50 mm3, the mice were injected subcutaneously with either vehicle (DMSO) or PIP3 (1 mg/kg) once every 3 days for a total of seven injections. Drug administration regimen is referred in previous articles (33, 34). The animals were sacrificed 3 days after the final injection (Fig. 7A). Intense Hoechst signal was observed in the nuclei of xenograft tumor cells (Supplementary Fig. S8C), indicative of the efficient uptake of PIP3 by xenograft tumors. During the observation period of 3 weeks, a significant retardation of tumor growth was observed in mice treated by PIP3 with a T/C (treated vs. the control) value of 35% (Fig. 7B, left; Supplementary Fig. S8A and S8B). As a result, the tumor weight in mice treated with PIP3 reduced to 29% of that in mice injected with vehicle at the experimental endpoint (Fig. 7B, right). The efficacy of PIP3 was further studied in a non–small cell lung cancer model, NCI-H1299, and a breast cancer model, MDA-MB-231. By using the same schedule of injections, we observed excellent tumor growth inhibition with T/C values of 38% and 23%, respectively, and substantial tumor weight reduction with T/C values of 32% and 30%, respectively (Fig. 7C and D; Supplementary Fig. S8A and S8B), in these two models. Notably, all animals examined had no evident reduction in body weight (Fig. 7B–D) and displayed normal behavior, indicative of a good tolerability of PIP3. Together, these results indicate that PIP3 can inhibit the growth of a range of tumor types and may be clinically effective.
To demonstrate that the observed in vivo activity of PIP3 was mediated by Plk1 suppression, we next examined the related biomarkers in tested xenografts by employing immunohistologic methods. Relative to vehicle controls, approximately 50% to 70% reduction of Plk1 expression was detected in PIP3-treated tumors (Fig. 7E and F). Moreover, the number of phospho-histone H3–positive cells (Fig. 7G) was decreased in PIP3-treated tumors, indicative of the reduced proliferation rate in these cells. By TUNEL staining, we further detected a significant increase of cell death in PIP3-treated tumors (Fig. 7H). Collectively, these in vivo data suggest that PIP3 inhibited the tumor growth via suppressing Plk1 expression.
Although small molecules of PIPs have shown great promise in regulation of gene expression, their disadvantages, such as low selectivity attributed by short recognition sequences and poor cellular permeability, really limit their use in clinical practice. Previous studies indicated that the linkage of PIPs to Hoechst molecule could recognize longer DNA sequences and also enhance cellular permeability (26, 35, 36). However, whether this strategy is effective for endogenous gene regulation remains unclear. Here, we demonstrated that a PIP–Hoechst conjugate is effective for suppression of Plk1 expression in tumor cells in vitro and in vivo. Indeed, conjugation of Hoechst significantly improved the cellular uptake and nuclear access of PIPs. Surprisingly, the PIPs without Hoechst or PIPs with Hoechst conjugated at C-terminus could not suppress Plk1 expression, suggesting that the proper Hoechst conjugation increased the sequence selectivity of PIPs. These results are consistent with the previous finding that the DNA-binding affinity of hairpin Hoechst polyamides is driven by both the polyamide interactions within the DNA minor groove and the Hoechst recognizing its target sequence (35). Thus, our study demonstrates a general strategy for improving the DNA-binding specificity of PIPs.
Plk1 is a key player of mitotic progression and cell proliferation and has been widely considered to be a promising target for anticancer therapy (37). However, current Plk1 inhibitors either targeted its kinase domain or substrate interaction domain have their own drawbacks such as off-target effects, poor stability, and less transcellular permeability, thus really limiting their further use for cancer treatment (38–41). In this study, we developed a novel approach to inhibit Plk1 at the transcriptional level using synthetic PIPs. PIPs can bind to the gene promoter in a DNA sequence–specific mode and inhibit gene expression through interfering with the interactions between DNA and transcription factors, thus allowing them to be a new class of transcriptional gene regulating agents with high specificity. Additional properties such as high stability and excellent transcellular permeability in cell cultures (42) and low toxicity in animals (43) further support the PIPs as appropriate agents for gene silencing in disease treatment. Here, we show that PIP3, a PIP–Hoechst conjugate targeting the PLK1 promoter, was capable of suppressing Plk1 expression in vitro and in vivo. Moreover, PIP3 showed high selectivity toward Plk1, but not other Plk family members. Notably, PIP3 could easily permeate cell membranes and localize to nuclei without the need for any delivery systems. Thus, as a promising Plk1 inhibitor candidate, PIP3 has significant advantages over existing Plk1 inhibitors and may be applied for Plk1 inhibition in future molecular biology studies and potentially in cancer therapy.
Multiple lines of evidence have shown that tumors with p53 deficiency may be particularly sensitive to Plk1 inhibition. Knockdown of Plk1 in human cancer cell lines leads to a mitotic arrest by activation of the p53 pathway (8, 44, 45). Accordingly, depletion of Plk1 preferentially reduces the survival of cancer cells with either no p53 or inactivated p53 compared with those harboring wild-type p53 protein. Consistent with these findings, Yang and colleagues (46) reported that transient exposure of a specific Plk1 inhibitor GSK461364A has greater antiproliferative effect in p53-deficient tumors. In this study, we also show that PIP3, as a novel class of Plk1 inhibitor, works better in cancers that have lost p53 function. Although the molecular mechanism behind Plk1 and p53 is not fully illustrated, all these studies support a positive association of loss of p53 function with Plk1-targeted drug response, thus providing a rationale for potential patient selection in future clinic trials using Plk1 inhibitors.
Although Plk1 is overexpressed in many types of cancers, its expression is relatively low in normal tissues (47). In agreement with this view, inhibition of Plk1 by siRNA led to specific killing of cancer cells while normal cells survived (10). Monika and colleagues also demonstrated that inducible knockdown of Plk1 in adult mice did not affect the proliferation of primary cells but significantly retarded the growth of cancer cells (11). Thus, these studies suggest that cancer cells and normal cells have different degrees of dependency on Plk1 activity. It is thought that the growth and survival of cancer cells require high level of Plk1 activity, whereas the residual enzyme activity upon Plk1 knockdown is sufficient for normal cells to survive. Because of this, the development of efficient Plk1 knockdown approach represents an appealing strategy to kill cancer cells while avoiding toxicity toward normal tissues. PIPs are such kind of agents that can inhibit the enhanced expression of Plk1 genes driven by specific transcription factors while preserving the basic expression of Plk1. In the current study, we developed a specific PIP–Hoechst conjugate, PIP3, to suppress Plk1 expression. Our results demonstrated that PIP3 could inhibit the expression of Plk1 and consequently retarded tumor cell growth. However, despite that PIP3 similarly suppressed Plk1 expression in nontransformed cells, the mitotic progression and proliferation of these cells were not affected by PIP3 treatment. This property of PIP3 is different from that of the existing Plk1 inhibitors, most of which kill normal and malignant cells alike. Because PIP3 targets PLK1 gene promoter but not Plk1 proteins, this novel Plk1 inhibitor shows no different sensitivity in cancer cells with different Plk1 expression levels. Thus, PIP3 represents a novel class of Plk1 inhibitor that can kill tumor cells while leaving noncancerous cells undamaged and therefore is a promising candidate for targeted cancer therapy.
In summary, we developed a PIP–Hoechst conjugate that can target PLK1 promoter and suppress Plk1 expression with high specificity. This agent is effective in suppression of cancer cell growth in vitro and in vivo, while nontoxic to normal cells under experimental concentrations. Accordingly, it shows great antitumor efficacy in xenografts with low toxicity. Thus, this PIP–Hoechst conjugate is a promising Plk1 inhibitor candidate for further development in targeted cancer therapy.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: K. Liu, L. Fang, H. Sun, R. Wang, W. Su, Hongchang Li
Development of methodology: K. Liu, L. Fang, H. Sun, Z. Pan, W. Su, Hongchang Li
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Liu, L. Fang, H. Sun, Z. Pan, J. Zhang, Y. Tan, Z. Ding
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Liu, L. Fang, H. Sun, Y. Tan, Z. Ding, Hongchang Li
Writing, review, and/or revision of the manuscript: K. Liu, L. Fang, L. Ao, X. Liu, Huashun Li, R. Wang, W. Su, Hongchang Li
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Liu, L. Fang, J. Zhang, J. Chen, X. Shao, Y. Tan, C. Wu
Study supervision: X. Liu, R. Wang, W. Su
Other (measure the concentration of Py-Im Polyamide and perform SPR assay): W. Wang
Other (chemical synthesis of Py-Im polyamide): C. Wu
This work was supported by a grant from the Ministry of Science and Technology of China no. 2014CB964602 (to H.C. Li); National Natural Science Foundation of China no. 31671397 (to H.C. Li), no. 21402232, and no. 21778068 (to L.J. Fang), no. 21672254 and no. 21432003 (to W. Su), no. 81702952 (to K. Liu), no. 31601174 (to X.M. Shao), no. 81502537 (to J.C. Zhang); Guangdong Natural Science Foundation of Research Team no. 2016A030312006 (to H.C. Li and W. Su); Guangdong Natural Science Foundation no. 2016A030310128 (to J.C. Zhang); Shenzhen Science and Technology Program JCYJ20160229204338907 and JCYJ20150630114942300 (to H.C. Li), JCYJ20150401150223649 (to W. Wang), JCYJ20170307165752275 (to K. Liu), JCYJ20170818153538196 (to W. Su), and KQCX 2015033117354154 (to L.J. Fang).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.