Transcriptional regulation of telomeric repeat-containing RNA by acridine derivatives

ABSTRACT

Telomere is a specialized DNA–protein complex that plays an important role in maintaining chromosomal integrity. Shelterin is a protein complex formed by six different proteins, with telomeric repeat factors 1 (TRF1) and 2 (TRF2) binding to double-strand telomeric DNA. Telomeric DNA consists of complementary G-rich and C-rich repeats, which could form G-quadruplex and intercalated motif (i-motif), respectively, during cell cycle. Its G-rich transcription product, telomeric repeat-containing RNA (TERRA), is essential for telomere stability and heterochromatin formation. After extensive screening, we found that acridine derivative 2c and acridine dimer DI26 could selectively interact with TRF1 and telomeric i-motif, respectively. Compound 2c blocked the binding of TRF1 with telomeric duplex DNA, resulting in up-regulation of TERRA. Accumulated TERRA could bind with TRF1 at its allosteric site and further destabilize its binding with telomeric DNA. In contrast, DI26 could destabilize telomeric i-motif, resulting in down-regulation of TERRA. Both compounds exhibited anti-tumour activity for A549 cells, but induced different DNA damage pathways. Compound 2c significantly suppressed tumour growth in A549 xenograft mouse model. The function of telomeric i-motif structure was first studied with a selective binding ligand, which could play an important role in regulating TERRA transcription. Our results showed that appropriate level of TERRA transcript could be important for stability of telomere, and acridine derivatives could be further developed as anti-cancer agents targeting telomere. This research increased understanding for biological roles of telomeric i-motif, TRF1 and TERRA, as potential anti-cancer drug targets.

INTRODUCTION

Telomere is a specialized DNA–protein complex that caps the end of eukaryotic chromosome and plays important roles in maintaining genomic integrity and chromosome stability. Shelterin is a protein complex formed by six different proteins, with telomeric repeat factors 1 (TRF1) and 2 (TRF2) binding to double-strand telomeric DNA. Telomeric DNA consists of G-rich and C-rich repeats, which could form G-quadruplex and intercalated motif (i-motif) structure, respectively. Shelterin binding sites, TTAGGG repeats, are maintained by telomerase, and telomere attrition predicts mortality and ageing-related diseases [1]. Human telomeres consist of G-rich and C-rich repeats prone to form a classical Watson–Crick double helix with a single-stranded (ss) 3ʹ-overhang. The telomeric G-rich strand can form a four-stranded G-quadruplex structure which have gained considerable attention due to their significant stability under physiological conditions [2,3]. Its complementary C-rich strand has propensity to adopt i-motif structure, whose formation usually requires acidic condition and is dynamic under physiological condition. Thus, there are few evidences about biological function of telomeric i-motif. Several factors have been found to stabilize i-motif structures at physiological conditions such as molecular crowding [4], single-walled carbon nanotubes (SWNTs) [5] and silver cations [6]. It has been reported that carboxyl-modifiedSWNTs could selectively stabilize human telomeric i-motif, leading to telomere uncapping [7]. Researchers have generated and characterized an antibody fragment that can recognize i-motif structures selectively, which provides direct evidence for existence of i-motif in living cells. It has been demonstrated that in vivo formation of i-motif is cell-cycle and pH-dependent, and i-motif structures play an important role in gene regulation [8]. Therefore, targeting of telomeric i-motif with small molecules could be important for a better understanding of its biological function and potentially significant in medicinal chemistry.

Shelterin protects the DNA ends and regulates the access of processing enzymes to telomeric DNA [9,10]. Shelterin can interact with telomeric DNA, which makes telomere a compact globular structure. Upon disruption of shelterin components, telomere decompact up to 10-fold increase in telomere volume, causing telomere de-protection accompanied with telomere abnormalities [11]. Shelterin is a protein complex with six subunits. Within shelterin, telomeric repeat factor 2 (known as TRF2 or TERF2) and telomeric repeat factor 1 (known as TRF1) are both important components, which bind to double-strand DNA. Components of TRF1 and TRF2 have similar architecture: a C-terminal MYB domain required for binding to telomeres, a flexible hinge domain involved in protein–protein interactions and a TRFH domain required for homodimerization [12,13]. The TRFH domain plays a critical role in the ability of TRF2 to condense DNA and stimulate telomeric invasion [14]. However, TRF1 has inefficiently condensed DNA because of the acidic domain in the full-length protein [15]. Besides, in contrast to TRF2, TRF1 disruption has been indicated to compromise telomere replication and significantly increase the levels of fragile telomeres [16].

Telomeric repeats are transcribed into long non-coding RNAs known as telomeric repeat-containing RNAs (TERRA), which are essential for telomere structural stability and heterochromatin formation [17]. TERRA-containing UUAGGG repeats are transcribed from telomeric C-rich strand driven by RNA polymerase II [18,19]. Previous studies have confirmed that TERRA is associated with TRF2 directly in vivo in multiple cell types and can bind to amino terminal GAR domain of TRF2 by forming an intramolecular G-quadruplex structure [17,20]. Besides, it has been shown that stabilization of TERRA G-quadruplex can result in inhibition of U2SO cell proliferation [21]. TRF2 can negatively regulate TERRA transcription since the levels of TERRA transcripts have been up-regulated upon TRF2 knockdown. A similar effect has been observed upon removal of TRF1, but not upon depletion of other shelterin components [13]. However, expressed full-length wild-type TRF2 but not TRF1 prevents TERRA up-regulation upon depletion of endogenous TRF2, which suggests that TRF2 and TRF1 suppress the levels of TERRA transcripts through distinct mechanisms [13]. The mechanism for TRF1 to suppress the levels of TERRA transcripts is unknown. In addition, TERRA participates in telomere heterochromatin formation [17] and various regulatory functions, including telomerase activity [22,23], which is regarded as a promising therapeutic target.

In our present study, after extensive screening, we found that acridine derivative 2c and acridine dimer DI26 could selectively interact with TRF1 and telomeric i-motif, respectively. Compound 2c could block the binding of TRF1 with telomeric duplex DNA, resulting in up-regulation of TERRA. In contrast, compound DI26 could destabilize telomeric i-motif, resulting in down-regulation of TERRA. Both compounds exhibited anti-tumour activity for A549, but induced different DNA damage pathways. The function of telomeric i-motif structure was first studied with a selective binding ligand, indicating that telomeric i-motif could play an important role in regulating TERRA transcription. Our results showed that appropriate level of TERRA transcript could be important for stability of telomere, and acridine derivatives could be further developed as anti-cancer agents targeting telomere. Our research increased understanding for biological roles of telomeric i-motif, TRF1 and TERRA, and further explored their possibility of becoming anti-cancer drug targets.

RESULTS

Screening for small molecules blocking the binding of TRF1 with telomeric DNA

To investigate the interaction between shelterin and telomeric DNA, TRF1 and TRF2, two components of shelterin, were expressed and purified as described in Supplementary Materials and as shown in Figures S1a-b and S2a-b. The activity of TRF1 was verified by using electrophoretic mobility shift assay (EMSA), enzyme-linked immunosorbent assay (ELISA) and filter-binding assay, and its dissociation constant with telomeric DNA was determined to be 13.9 ± 2.5 nM, as shown in Figure S1c-e. We screened more than 400 compounds by using EMSA for compound to disrupt the interaction between TRF1 and telomeric DNA as shown in Figure S3. We found that acridine derivative 2c (compound 367, Figure 1A) had the most potent activity of breaking the interaction between TRF1 and telomeric DNA. With addition of increasing concentration of compound 2c, the amount of the TRF1-DNA complex was found to be decreased, and the corresponding free DNA was found to be increased (Figure 1B,D). Then we used ELISA to evaluate the activity of compound 2c on inhibiting the binding between TRF1 and telomeric DNA, and its IC₅₀ value was determined to be 4.3 μM, as shown in Figure 1C. In comparison, compound 2c did not affect the binding between TRF2 and telomeric DNA (Figure 1C).

Figure 1.

Effect of compound 2c on the binding of TRF1 with telomeric DNA. (A) Structure of 2c. (B) Electrophoretic mobility shift (EMSA) assay for effect of 2c on disrupting the binding between TRF1 and oligonucleotide T-21GC. Lane 1, 1 μM T-21GC. Lane 2, 3 μM TRF1. Lanes 3–10, DMSO or 2c (0.39, 0.78, 1.56, 3.125, 6.25, 12.5 and 25.0 μM) was added into the mixture of TRF1 with T-21GC. TRF1-DNA complex is the binding combination of TRF1 and T-21GC, while free DNA is T-21GC only. (C) Enzyme-linked immunosorbent assay (ELISA) for the effect of 2c on the binding of TRF1 or TRF2 to oligonucleotide T-21GC. The data were derived from three experiments and were shown as the means ± S.E.M. (D) Filter-binding assay result for the effect of 2c on the binding of TRF1 to oligonucleotide T-21GC

Compound 2c could selectively bind to TRF1 without significant interaction to TRF2

In order to know how compound 2c disrupted the binding between TRF1 and telomeric DNA, we performed microscale thermophoresis experiment (MST) to study its interaction with TRF1 and telomeric DNA, respectively. Our data showed that compound 2c bound with TRF1 tightly with its K_D value determined to be 4.49 ± 0.18 μM (Figure 2A), without significant interaction with telomeric DNA (Figure.2B). Since TRF1 mainly binds to telomeric DNA with its C-terminal Myb-like DNA binding domain [24], it is possible that compound 2c also bound to TRF1 on its C-terminal Myb-like DNA binding domain, resulting in disruption of the interaction between TRF1 and telomeric DNA. In comparison, compound 2c had no significant effect on protein TRF2 and TRF2-DNA complex (Figure 2C), which indicated that binding conformation of C-terminal Myb-like DNA binding domain for TRF1 could be different from that for TRF2. Then, we carried out chromatin immunoprecipitation (ChIP) assay to study whether compound 2c could inhibit the binding of TRF1 to telomeric DNA in cells. As shown in Figure 2D, incubation of A549 cells with 1.5 and 3 μM compound 2c for 48 h caused significant dissociation of TRF1 from telomeric DNA. Amplification of the immunoprecipitated samples containing the oligonucleotide of telomere showed a decrease of TRF1-bound DNA with increasing 2c concentration in a dose-dependent manner, indicating that 2c indeed disrupted the binding of TRF1 with telomeric DNA inside A549 cells. These results showed that 2c bound well with TRF1 in vitro and in cells, without significant binding interaction with TRF2.

Figure 2.

Further studies for effect and selectivity of compound 2c. (A) MST experimental result for the binding of fluorescently labelled TRF1 to 2c. The data of MST were evaluated using NT Analysis 1.4.23. (B) MST experimental result for the binding of fluorescently labelled T-21GC to 2c. (C) MST experimental result for the binding of fluorescently labelled TRF2 to 2c. (D) ChIP experiment was used to evaluate the effect of 2c on disruption of the binding between TRF1 and telomeric DNA inside A549 cells. Normal rabbit IgG was used as a negative control for mock immunoprecipitation. Immunoprecipitated DNA samples were PCR-amplified to show TRF1 occupancy of the oligonucleotide of telomere, and the amplified products were separated on 2.0% agarose gel

Screening for small molecules binding to telomeric i-motif

Telomere is a specialized DNA–protein complex, and double-strand telomeric DNA consists of G-rich and C-rich repeats, which could form G-quadruplex and i-motif structure, respectively, during replication and transcription. Although various telomeric G-quadruplex binding ligands have been well established, specific telomeric i-motif binding ligand has not been investigated. In order to study the role of telomeric i-motif, we screened various compounds for telomeric i-motif binding ligand through surface plasmon resonance (SPR) experiment. Our screening results are shown in Table S1, which indicated that acridine dimer DI26 (Figure 3A) had the strongest binding with telomeric i-motif, with dissociation constant K_D value determined to be 0.46 μM, as shown in Figure 3B. We also carried out MST experiment, which also showed that DI26 could bind with telomeric i-motif tightly with dissociation constant K_D value of 0.13 μΜ (Figure 3C). In contrast, DI26 did not show significant binding interaction to telomeric G-quadruplex and duplex DNA (Figure S4a,b).

Figure 3.

Acridine dimer DI26 could bind to and interact with telomeric i-motif in vitro. (A) Structure of DI26. Their binding affinity constant K_D value was determined by using SPR experiment (B) and MST experiment (C). CD titration experiment (D) and fluorometric titration experiment (E) were performed to study the interaction of DI26 with telomeric i-motif. (F) CD spectra for telomeric i-motif (Tel-C) under different temperatures at pH 5.0. (G) CD spectra for Tel-C with five molar equivalents of DI26 under different temperatures at pH 5.0. (H) The stability of telomeric i-motif in the presence or absence of DI26 was studied through measurement of transitional temperatures with molar ellipticity at 288 nm of CD spectra versus temperature

Compound DI26 could bind to and destabilize telomeric i-motif with good selectivity

To further investigate the interaction of DI26 with telomeric i-motif, we carried out circular dichroism (CD) titration and fluorescence spectroscopic experiments. The transitional pH was determined to be 6.11 for further physicochemical properties study (Figure S4c). Telomeric i-motif showed a positive peak at 288 nm and a negative peak near 260 nm at pH 5.5 (Figure 3D), and the peak shifted towards infrared wavelength in a concentration-dependent manner upon addition of DI26, which indicated that the compound interacted strongly with telomeric i-motif and induced a conformational change. Our CD titration experiment also showed that CD spectra of telomeric G-quadruplex and telomeric duplex DNA had no obvious change upon addition of DI26 (Figure S4d,e), which indicated that DI26 had no significant interaction with telomeric G-quadruplex and telomeric duplex DNA. Compound DI26 had a maximum absorption around 400 nm and displays fluorescence emission upon excitation at 390 nm. Upon addition of increasing concentration of telomeric i-motif to compound DI26, significant fluorescence enhancement was observed before reaching a platform as shown in Figure 3E. In contrast, upon addition of increasing concentration of telomeric G-quadruplex to compound DI26, no obvious fluorescence change was observed (Figure S4f), which was probably due to lack of specific interaction between DI26 and telomeric G-quadruplex [25]. To demonstrate the interaction between telomeric i-motif and DI26, NMR titration of telomeric i-motif with different concentrations of DI26 was performed. As shown in Figure S5, NMR signal of telomeric i-motif reduced with increasing concentrations of DI26. These results showed that DI26 interacted with telomeric i-motif selectively.

To evaluate whether DI26 could affect stability of telomeric i-motif, we carried out CD-melting experiment. In the presence of 5 eq DI26, melting temperature (T_m) of telomeric i-motif decreased for 8.65°C (Figure 3F–H) at pH 5.0 and 3.39°C (Figure S6a–c) at pH 5.5, which suggested that DI26 could destabilize telomeric i-motif. In contrast, DI26 had less significant effect on stability of telomeric G-quadruplex with its T_m value increased for only 0.74°C (Figure S6d–f). Our above results suggested that DI26 could selectively bind to and destabilize telomeric i-motif.

Effects of compound DI26 and 2c on levels of TERRA transcripts

Human telomeric DNA is transcribed into using telomeric C-rich DNA strand as a template [18,26,27]. TERRA is partially located to telomeres and plays important roles in maintaining telomere structures. Based on our result that DI26 could bind to and destabilize the telomeric i-motif, we further studied its effect on the levels of TERRA transcripts by using a practical qPCR approach [28]. A549 cells treated with 10 and 20 μΜ DI26 for 24 h were used for measuring TERRA levels by using reverse transcription-quantitative polymerase chain reaction (RT-qPCR). Previous research has shown that in some cancer cells telomere transcription could occur at many but not all chromosomes, and TERRA transcribed mainly from the 20q and Xp loci [29,30]. Thus, six chromosomal regions were analysed by using subtelomere-specific primers including 20q and XpYp (Supplementary Table S3). As shown in Figure 4A, DI26 could significantly down-regulate the levels of TERRA transcripts in a dose-dependent manner, as much as 34.5% (2q) and in average 45.6%. It was possible that the levels of TERRA transcripts were reduced due to the binding of DI26 to TERRA transcriptional templates and blocked the binding of transcription factors.

Figure 4.

In-depth study for regulation and effect of TERRA. (A) Compound DI26 down-regulated TERRA transcription in A549 cells upon 24 h incubation. (B) Compound 2c up-regulated TERRA transcription in A549 cells upon 24 h incubation. (C) TRF1 could tightly bind to TERRA, which was analysed by using ELISA. (D) Addition of compound 2c had no effect on binding affinity between TRF1 and TERRA, indicating that TRF1 bound to TERRA possibly with its allosteric site. (E) TRF1 could tightly bind to telomeric duplex DNA, and the binding affinity was significantly reduced upon addition of TERRA, indicating that TERRA could be an allosteric inhibitor of TRF1

Based on our result that compound 2c could bind to TRF1 and block its binding to telomeric duplex DNA, we also studied its effect on TERRA transcription. In contrast to compound DI26, our result showed that compound 2c up-regulated the levels of TERRA transcripts, as shown in Figure 4B). It has been shown that TRF1 could negatively regulate TERRA transcription since the levels of TERRA transcripts have been up-regulated upon TRF1 knockdown. A similar effect has been observed upon removal of TRF2, but not upon depletion of other shelterin components [13]. Our present data for up-regulation of TERRA caused by delocalization of TRF1 from telomeric DNA by compound 2c is consistent with the previous TRF1 knockdown result. It is possible that the binding of TRF1 to telomeric DNA prevent transcription factor such as RNA polymerase from accessing telomeric DNA for transcription.

Both TERRA and telomeric duplex DNA have been shown to be capable of binding to TRF1 [31]. TERRA can be recruited to telomeres through direct interactions with shelterin, with TRF1 and TRF2 as the most enriched components [17]. Previous study has confirmed that both TRF1 and TRF2 could directly interact with TERRA in many cell lines. In order to better understand the binding of TRF1 to TERRA and telomeric duplex DNA, we performed ELISA to evaluate the affinity between TRF1 and TERRA [20,21] at first. In this assay, 5ʹ biotinylated DNA was immobilized on streptavidin-coated 96-well plates, and freshly purified His-tagged TRF1 was used for antigen-antibody reaction followed by quantification with anti-His antibody. Our result as shown in Figure 4C indicated that TRF1 could bind to TERRA tightly with K_D value determined to be 39.42 ± 3.49 nM. This binding affinity was only slightly changed to K_D value of 31.36 ± 3.52 nM upon addition of different concentrations of compound 2c. Our data suggested that compound 2c had no significant effect on the binding of TRF1 with TERRA, although compound 2c could bind to TRF1 and block its binding with telomeric duplex DNA. This might indicate that TRF1 interacted with TERRA and telomeric duplex DNA through two distinct binding sites.

We further investigated effect of TERRA on binding affinity of TRF1 with telomeric duplex DNA. In the absence of TERRA, TRF1 could bind to telomeric duplex DNA tightly with K_D value determined to be 19.09 ± 4.19 nM. Then, 500 nM unmodified TERRA was incubated with TRF1 (0–500 nM) at room temperature for 1.5 h before applying to immobilized telomeric duplex DNA. As shown in Figure 4E, our result indicated that the binding of TERRA with TRF1 significantly decreased its binding affinity with telomeric duplex DNA, with the K_D value determined to be 166.73 ± 4.32 nM. These results indicated that TERRA could bind to TRF1 at its allosteric site, resulting in delocalization of TRF1 from telomeric DNA, which could further up-regulate the levels of TERRA transcripts.

Effects of compound 2c and DI26 on DNA damage response

Telomeres serve to cap the ends of eukaryotic chromosomes and play crucial roles in end protection. Shelterin complex prevent telomeres from being recognized as damaged DNA causing DNA damage response (DDR). TRF1 plays an important role in compacting telomeric chromatin into globular structures and preventing DDR, which can maintain correct telomere structure [11]. Since our compounds interfered with telomere through binding with TRF1 or telomeric i-motif, we further studied whether the compounds could induce DDR. We studied protein expression levels of DNA damage-related pathways in A549 cells by using western blot. As shown in Figure 5A, phosphorylated ATM (p-ATM), phosphorylated checkpoint kinase-2 (p-Chk2), phosphorylated p53 (p-p53) and phosphorylated H2AX (γH2AX) were all up-regulated in a dose-dependent manner upon incubation of A549 cells with compound 2c for 48 h. It has been shown that ATM can sense and respond to double-strand DNA breaks (DSBs) primarily [32]. After formation of DSBs, ATM directly phosphorylates checkpoint kinase-2 (Chk2) at Thr68 [33] and p53 at Ser15 [34]. DNA damage could induce cells apoptosis, and up-regulation of p-BRCA1 indicated the happening of DNA repair. For comparison, expression levels of TRF1 were found to be constant with increasing concentration of compound 2c. Our above results indicated that compound 2c could bind to TRF1, resulting in delocalization of TRF1 from telomeric DNA. Without strong binding of TRF1 as an important shelterin component for protection of telomere, DSBs could occur resulting in DDR.

Figure 5.

Analyses of DNA damage response triggered by compound 2c and DI26. Western blot was performed to analyse expression levels of related proteins in DNA damage response-related pathways in A549 cells induced by compounds 2c (A) and DI26 (B)

Since compound DI26 could bind to telomeric i-motif, we studied whether DI26 could also induce DDR. We studied protein expression levels of DNA damage-related pathways in A549 cells by using western blot. As shown in Figure 5B, phosphorylation of γ-H2AX, as a hallmark of DDR, was increased in a dose-dependent manner after incubation with DI26 for 48 h. We found that cells treated with DI26 activated ATR-related DDR [35], while the level of p-ATM remained unchanged. Phosphorylated ATR (p-ATR) was up-regulated in a dose-dependent manner, and p-ATR possibly phosphorylated checkpoint kinase 1 (p-CHK1) at Ser345 to arrest cell-cycle progression. Then, the p53 tumour suppressor was phosphorylated at Ser15 to affect DNA repair, cell-cycle arrest and apoptosis [36,37]. Next, we analysed DNA repair factor BRCA1, one of the substrates for ATR, and found that phosphorylated BRCA1 was down-regulated significantly, which suggested that DNA repair was possibly blocked, thus inducing cells apoptosis. The above results indicated that compound DI26 could induce ATR-related DDR. For comparison, the levels of the same proteins in their non-phosphorylated forms were found to be unchanged upon incubation with compound 2c or DI26, as shown in Figure 5. Our results indicated that compound 2c could induce DSBs, while DI26 might induce single-strand DNA damage.

In order to know whether DNA damage was located at telomeres, double immunofluorescence experiments were performed on A549 cells upon incubation with compounds. Since the binding between TRF1 and telomere could be disturbed by compound 2c, TRF2 was used to locate the telomere. The representative immunofluorescence images of γH2AX, 53BP1 and TRF2 foci in A549 cells after incubation with compound 2c (1.5 μM) for 48 h are shown in Figure 6A and Figure S7a. As shown in Figure 6B, a significant increase in γH2AX foci with a mean of 89 foci per nucleus was observed. Then, we quantitatively analysed the co-localization of γH2AX with TRF2, so-called telomere dysfunction-induced foci (TIFs). Upon incubation of A549 cells with 2c, the number of TIFs was significantly increased to 53 TIFs per nucleus (Figure 6C). Thus, about 60% of γH2AX foci were co-localized to TRF2. In comparison, upon incubation of A549 cells with compound 2c, the number of TRF2 foci had no significant change. We also analysed γH2AX foci induced by DI26, as shown in Figure S7b. γH2AX foci and the co-localization with TRF1 obviously increased upon treatment of the cells with DI26, which indicated that DNA damage also occurred mainly at telomere region.

Figure 6.

Representative immunofluorescence images of γH2AX (green) and TRF2 (red) foci in A549 cells. (A) The nuclei were stained with DAPI (blue), and typical co-localization foci are displayed in yellow in the enlargement picture merged with DAPI. (B) Quantification of γH2AX foci number per nucleus. (C) Quantification of TIF numbers per nucleus. In all experiments, >50 nuclei were counted in each group and the s.d. was calculated from three replicates. *** P < 0.001 compared with DMSO

Effects of compound 2c and DI26 on cell-cycle arrest and apoptosis

Next, we studied whether these two compounds could cause cell-cycle arrest and apoptosis. We used a flow cytometry to analyse percentage of A549 cells in each phase of cell cycle and apoptosis cells. The cells were incubated with various concentrations of compounds for 48 h, and then dyed with propidium iodide (PI) for cell-cycle detection or double staining with annexin V-FITC and PI for apoptosis. As shown in Figure 7A, B, A549 cells treated with compound 2c showed a remarkable increase from 24.83% to 46.46% in G2/M phase accompanied by a noteworthy decrease of the cells in G0/G1 phase from 54.60% to 29.84% all in dose-dependent manners, which indicated that compound 2c induced a G2/M phase arrest. In contrast, DI26 induced a significant accumulation of cells in G0/G1 phase accompanied by a decrease of the cells in G2/M phase (Figure 7E, F), which indicated that compound DI26 induced a G0/G1 phase arrest. The above results might be related with their different DNA damage pathways, which were revealed in our western blotting experiments.

Figure 7.

Effect of compound 2c and DI26 on A549 cell cycle and apoptosis. (A) Cell-cycle analysis after propidium iodide (PI) staining of A549 cells incubated with various concentrations of 2c for 48 h. (B) The histogram of percentage of cells in different phases of the cell cycle was analysed by using EXPO32 ADC software. (C) Representative images of apoptosis analysis after annexin V-FITC and PI staining of A549 cells incubated with 2c for 48 h. (D) The histogram of percentage of cells in different phases of the apoptosis was analysed by using EXPO32 ADC software. (E) Cell-cycle analysis after propidium iodide (PI) staining of A549 cells incubated with DI26 for 48 h. (F) The histogram of percentage of cells in different phases of the cell cycle was analysed by using FlowJo 10.0 software. (G) Representative images of apoptosis analysis after annexin V-FITC and PI staining of A549 cells incubated with DI26 for 48 h. (H) The histogram of percentage of cells in different phases of the apoptosis was analysed by using FlowJo 10.0 software. The data were derived from three experiments and were shown as the means ± S.E.M

Our results of apoptosis experiment (Figure 7C, D) indicated that compound 2c caused a significant increase of cells in early apoptosis phase (A4) from 2.84% to 42.66% together with an obvious dissipation of normal cells (A3) from 96.98% to 56.12% in comparison to dimethyl sulfoxide (DMSO)-treated cells. In contrast, cells treated with DI26 mainly showed a late apoptosis as shown in Figure 7G, H (percentages of late apoptotic cells were 4.54%, 14.9% and 37.5%) accompanied by a fraction of early apoptosis (9.25%, 15.0% and 18.8%). For comparison, percentage of apoptotic cells for non-treated cells was only 0.53%. Our above results showed that compound 2c and DI26 had different effects on A549 cells at both molecular level and cellular level, which were possibly due to their different binding affinity with TRF1 and telomeric i-motif, respectively. Our results indicated that TRF1 and telomeric i-motif could be targeted for different purposes at molecular and cellular levels.

Cells cytotoxicity and anti-cancer activity of compound 2c and DI26

In order to know whether compound 2c and DI26 could impair cell viability and suppress tumour cell proliferation, acute cytotoxic effects of these two compounds were assessed by using methylthiazolyl tetrazolium (MTT) assays on various types of cancer cell lines with normal cells as control. In order to know whether 2c could be taken into A549 cells, we performed cells uptake assay. Compound 2c or its HCl acid form was incubated with A549 cells for 48 h, and percentages for cells uptake were determined to be over 80%, as shown in Figure S8c, with standard curves of UV absorption as shown in Figure S8a, b. Our results showed that compound 2c had a strong inhibition on A549 cells in a 48-h cell viability assay with its IC₅₀ value determined to be 3.3 μM. Compound 2c had a relatively weak effect on normal cells, such as H9c2 (rat myocardial cell line) cells with IC₅₀ value of 11.1 μM, and NRK-52E (rat kidney cell line) cells with IC₅₀ value of 20.5 μM.

Compound DI26 had a relatively weak effect on viability of A549 cells with IC₅₀ value determined to be 14.7 μM, and no significant inhibition on Siha, U2OS and HCT116 cells (IC₅₀ > 50 μM). However, compound DI26 had a mild inhibition effect on normal cells such as LX2 cells with IC₅₀ value determined to be 21.4 μM. In order to know whether amine chain linkage and large molecular weight of DI26 could influence membrane permeability, we also performed cells uptake assay. Our results showed that only ~50% DI26 was taken into A549 cells upon incubation with the cells for 18 h. If incubation time was increased to 48 h, DI26 could be taken into A549 cells at ~80% as shown in Figure S8e, with standard curve for UV absorption as shown in Figure S8d. Our results showed that A549 cells were more sensitive to compound 2c rather than DI26, which indicated that delocalization of TRF1 could be a more detrimental cell event than destabilization of telomeric i-motif [38].

To further assess the anti-proliferation activity, real-time cellular analysis (RTCA) assays were carried out for real-time monitoring A549 cells upon treatment with various concentrations of compounds. RTCA plots were generated using RTCA software 1.1.2. The results showed that these two compounds both had remarkable arrest of cell growth in a dose-dependent manner, as shown in Figures S9b and S10c.

Colony formation assays were also employed to provide more visualized results for their effects of growth suppression on A549 cells. As shown in Figures S9a and S10a, dose-dependent inhibitions of cell growth were observed upon treatment of A549 cells with increasing concentrations of compound 2c (DMSO as a control) and DI26 for 9 days. We also carried out cell scrape assay to evaluate their effects on A549 cell migration. As shown in Figures S9c and S10b, these two compounds could both inhibit A549 cell migration.

The effect of DI26 on long-term A549 cell viability was analysed through population doublings [7,39], in comparison with 10 μM telomerase inhibitor BIBR1532. Compound DI26 showed obviously better inhibition of cell proliferation than BIBR1532 upon incubation for 1 week (Figure S11a), followed by apparent growth arrest in the next 2 weeks, which were accompanied by massive cell death. In order to know whether long-term treatment with DI26 could induce cell senescence, we performed Senescence-associated β-Galactosidase Staining assay. After exposing to 20 μΜ DI26 for 15 days, SA-β-Gal activity was increased with A549 cells senescence reached ~60% (Figure S11b, c).

Drug combination studies with telomerase inhibitor BIBR1532

It has been shown that BIBR1532 can selectively inhibit telomerase activity through non-competitive binding to the active site of hTERT, with IC₅₀ value determined to be 100 nM in a cell-free assay. However, IC₅₀ value of its anti-proliferative activity has been determined to be more than 50 μM [40,41]. In order to know whether BIBR1532 could have synergistic effect with our compounds on anti-neoplastic activity, we studied their drug combination effects. We first assessed effects for combined usage of compound 2c and BIBR1532 on A549 cells. After 48 h incubation of A549 cells with compound 2c and varying concentrations of BIBR1532, the inhibition ratio and IC₅₀ values were determined based on MTT assays. In our study, IC₅₀ value of BIBR1532 alone on A549 cells was determined to be 46.77 μM, and addition of compound 2c or DI26 apparently had further impact on cells viability, as shown in Figure S12a, b. Synergistic effects for combined usage of compounds were calculated based on Chou–Talalay equation with CalcuSyn software [42]. Combination index CI < 1 indicates synergism, CI = 1 indicates additive effect and CI > 1 indicates antagonism. For compound 2c, with dosage of closing to ED50, when ratio of BIBR1532:2c was 20:1, such as 16.4 μM BIBR1532 and 0.82 μM 2c, CI values were determined to be around 1 or more than 1, indicating the possibility of additive effect or antagonism. When dosages were above ED75, CI values were determined to be 0.57, indicating the possibility of synergism. For compound DI26, remarkable synergistic effect was observed when the ratio of BIBR1532 and DI26 was 5:3. The CI values for dosage of closing to ED50 and ED75 were determined to be 0.83 and 0.29, respectively. This might indicate that we could target both telomere and telomerase through drug combination for enhanced effect of anti-cancer therapy.

Compound 2c suppressed tumour growth and induced DNA damage in vivo

Since compound 2c showed strong inhibition on A549 cells with its IC₅₀ value determined to be 3.3 μM, we further investigated its anti-tumour activity in vivo to evaluate whether compound 2c could become a lead compound for cancer treatment. We studied the effect of compound 2c in an A549 xenograft mouse model of human lung cancer. Compound 2c was first mixed with hydrochloric acid in methanol to obtain its hydrochloride form for the purpose of increasing its water solubility for enhanced absorption and bioavailability. Our EMSA, MST and cellular experiments showed that compound 2c and its hydrochloride form had no difference in binding with TRF1 (Figure S13a–d), and inducing the G2/M phase arrest and apoptosis of A549 cells (Figure S14a–d). After A549 cells were incubated with compound 2c hydrochloride for 48 h, the uptake of the compound was determined to be 87.0% (Figure S8c). Therefore, the hydrochloride form of compound 2c was used to study its in vivo anti-tumour activity.

A549 xenografted tumours were established through subcutaneous injection of A549 cells into the right armpit of immunocompromised mice. When the volume of the tumour reached 300 mm³, the tumour was divided into 3 mm³ and subcutaneously transplanted to mice. The mice were randomly divided into five groups (the vehicle-treated group, the compound of 6.25 , 12.5 and 25.0 mg/kg treated group and the cisplatin 0.5 mg/kg treated group, n = 7/group) when the tumour size reached approximately 50 mm³. The mice were treated with compound 2c hydrochloride and vehicle via intraperitoneal (ip) injection daily, while with cisplatin every other day, and the tumours were removed after 20 days of treatment and analysed. Compared with the vehicle group (mean value of 1334.5 mg as shown in Figure 8C), the 2c hydrochloride-treated groups at 6.25, 12.5 and 25.0 mg/kg resulted in a statistically significant reduction in tumour weight with a tumour growth inhibition (TGI) of 51.2%, 55.8% and 64.9%, respectively (mean values of 511.6, 557.7 and 467.9 mg, P < 0.005, P < 0.003 and P < 0.0004, respectively, as shown in Figure 8A, C), and the wide-spectrum anti-tumour antibiotic cisplatin was used as the reference molecule (TGI = 51.4%, mean of 648.4 mg, P < 0.005, as shown in Figure 8A, C). Besides, treatment with 2c hydrochloride at 6.25, 12.5 and 25.0 mg/kg resulted in a noteworthy decrease in the final tumour volume (means of 593.0, 574.7 and 396.4 mm³; P < 0.02, P < 0.02 and P < 0.0002, respectively, as shown in Figure 8B) in comparison with the vehicle group (mean of 1577.6 mm³, as shown in Figure 8B). In the period of the whole experiment, every mouse looked healthy without visible signs of discomfort distress or pain. Compared with the vehicle group, the 2c hydrochloride-treated groups showed no significant difference in body weight and visceral organ weight, which indicated that the compound was tolerated well at these doses (Figure 8D and Figure S15). In conclusion, 2c hydrochloride exhibited a good anti-cancer activity for suppressing lung tumour growth in BALB/C-nu/nu mice with A549 xenografts via the suppression of A549 cell proliferation.

Figure 8.

Compound 2c hydrochloride inhibited tumour growth in A549 xenograft model in vivo. After treatment with 2c hydrochloride at 6.25 mg/kg or 12.5 mg/kg or 25.0 mg/kg or cisplatin at 0.5 mg/kg for 3 weeks, the mice were sacrificed, and the tumours were weighed. (A) Images of excised tumours from each group. (B) Tumour volume of the mice in each group during the observation period. (C) Weights of the excised tumours from each group. (D) Body weight of the mice in each group at the end of the observation period. The data are presented as the mean ± SEM. * P < 0.05, ** P < 0.01, *** P < 0.001, significantly different compared with the vehicle group by t-test, n = 7

In order to know the possible mechanism for the anti-cancer effect of compound 2c, we used western blot and immunohistochemical assay to analyse the amount of DNA damage marker γH2AX in post-mortem tumour. As shown in Figure 9, 2c hydrochloride-treated groups showed an increase of the DNA damage marker γH2AX in a concentration-dependent manner. Our result indicated that the above anti-tumour activity of compound 2c could be related with its DNA damage effect, and compound 2c suppressed tumour growth possibly through inducing DNA damage mainly at telomere region.

Figure 9.

Analysis of DNA damage marker γH2AX for possible mechanism of compound 2c on suppressing tumour growth. (A) Western blot image of γH2AX protein levels in 2c hydrochloride- or cisplatin- or vehicle-treated A549 xenograft tumours. (B) Representative images of γH2AX-positive cells per field in 2c- or cisplatin- or vehicle-treated A549 xenograft tumours

DISCUSSION

Telomere has repetitive telomeric DNA at chromosome ends and telomere-associated proteins protecting telomeric DNA from unwanted repair. Cancer cells survive cellular crisis through telomere maintenance mechanism. Maintenance of telomere is undertaken by telomerase, which is a key factor that is activated in more than 80% of cancer cells but is absent in most normal cells. Telomere function and telomerase level are strongly associated with the development and progression of a wide range of cancers, such as prostate cancer, cervical cancer, gliomas and breast cancer [43,44]. Therefore, molecules targeting telomere maintenance could be developed as potential anti-tumour lead compounds.

Shelterin component TRF1 plays a key role in the regulation of telomere replication and chromosome end protection. Telomere can be transcribed into TERRA which partially co-localizes with telomere and may promote the recruitment of proteins and enzymatic activity at chromosome ends. TERRA transcription is mainly driven by CCCTC-binding factor (CTCF) [45], but its regulation mechanism remains unclear. TERRA association with telomere is crucial for the telomeric integrity of the chromosome, and thus, to the genomic integrity of the cells. One study in yeast has shown that over-expression of TERRA at a single telomere is sufficient to result in senescence [46]. Previous data have suggested that levels of TERRA transcripts are actively repressed in telomerase-positive cancer cells to maintain telomeric integrity, and therefore increasing TERRA levels in such cancers could induce telomere loss and eventual cell death. Our result showed that compound 2c blocked the binding of TRF1 with telomeric duplex DNA, resulting in up-regulation of TERRA, which could cause cancer cell apoptosis. Our present work is well consistent with previous research, which could be further developed for the treatment of telomerase-positive cancers.

In this work, we screened more than 400 compounds of different skeletons through EMSA and CD experiments, and found that compound 2c could selectively bind to TRF1 and disturb its binding to telomeric DNA. The effect of compound 2c on the binding of TRF1 with double-strand telomeric DNA was further confirmed through other experiments including filter-binding assay, ELISA, and ChIP. The binding of compound 2c to TRF1 could block its interaction with telomeric duplex DNA without significant effect on the affinity between TRF1 and TERRA. The binding of TERRA to TRF1 could apparently inhibit the interaction between TRF1 and telomeric DNA. These indicated that TERRA could bind to the allosteric site of TRF1. Delocalization of TRF1 by 2c up-regulated TERRA transcription, which indicated that TRF1 could negatively regulate TERRA transcription.

It has been shown that telomeric C-rich DNA sequence can fold into i-motif in live cells; however, its biological function and the possibility of becoming anti-cancer target remain to be explored [8]. Our data from SPR, MST, CD and fluorescence experiments indicated that acridine derivative DI26 could bind to and destabilize telomeric i-motif. By using this compound as a probe targeting telomeric i-motif, we explored its biological role and potential as drug target through RT-qPCR, ELISA and western blotting. Our result showed that telomeric i-motif could also regulate TERRA transcription and trigger DDR. In telomerase-negative cancer cell lines, it has been shown that high level of TERRA leads to increased R-loop formation at chromosome ends to promote telomere maintenance [47,48]. Our results indicated that DI26 could bind to and destabilize telomeric i-motif, resulting in down-regulation of TERRA transcription and DDR. Thus, DI26 could be effective for inducing telomerase-negative cancer cells apoptosis through down-regulation of TERRA transcription. TERRA has been demonstrated to be an RNA polymerase II-dependent transcript in all organisms analysed so far, and TERRA gets rapidly degraded by the 5ʹ to 3ʹ nuclear exonuclease Xrn2 in humans [49,50]. Our present result is consistent with previous data, which indicated that the levels of TERRA are tightly associated with telomere maintenance and cell senescence.

In summary, after extensive screening, we found two acridine derivatives 2c and DI26 targeting selectively at TRF1 and telomeric i-motif, respectively. Compound 2c could up-regulate TERRA transcription, while compound DI26 could down-regulate TERRA transcription. Compounds 2c and DI26 could both break telomere integrity and induce ATM- and ATR-dependent DDR, respectively. Thus, their corresponding targets TRF1 and telomeric i-motif, as well as telomere transcription product TERRA, could all become potential anti-tumour targets. Compounds 2c and DI26 both evoked cell-cycle arrest and induced G2/M phase and G0/G1 phase arrest, respectively. Compounds 2c and DI26 both inhibited cancer cell metastasis and proliferation. Compound 2c had a strong inhibition on A549 cells with its IC₅₀ value determined to be 3.3 μM, while compound DI26 had a relatively weak effect on A549 cells with its IC₅₀ value determined to be 14.7 μM. Since compound 2c had a strong inhibition on A549 lung cancer cell growth, its anti-tumour activity was further evaluated in BALB/C-nu/nu mice with A549 xenografts, which showed that compound 2c could significantly inhibit tumour growth possibly through inducing xenograft tumour DNA damage. The function of telomeric i-motif structure was first studied with a selective binding ligand DI26, indicating that telomeric i-motif could play an important role in regulating TERRA transcription. Our results showed that appropriate level of TERRA transcript could be important for stability of telomere, and acridine derivatives could be further developed as anti-cancer agents targeting telomere. By using compounds 2c and DI26 as mechanistic probes, our present study increased our understanding of the biological roles of telomeric i-motif, TRF1 and TERRA, and further explored their possibility of becoming anti-cancer drug targets.

MATERIALS AND METHODS

Oligonucleotides and compounds

All oligonucleotides used in this study were purchased from Sangon (China) and dissolved in ddH₂O. Their concentrations were based on single-strand DNA concentrations, which were determined according to their absorbance at 260 nm using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, USA) and calculated based on their respective molar extinction coefficients. Their further dilutions to working concentrations were made with relevant buffers. Compounds were dissolved in DMSO at 10 mM concentration and stored in −20°C freezer.

Cell culture and protein purification

Human lung adenocarcinoma A549 cells were grown in Dulbecco’s Modified Eagle’s medium (DMEM) supplemented with 10% foetal bovine serum, at 37°C under humidified atmosphere with 5% CO₂. Detailed information about protein expression and purification is provided in Supplementary Information.

Electrophoretic mobility shift assay

Compound was incubated with protein TRF1 in 20 μL of binding buffer (20 mM HEPES-KOH, pH 8.0, 200 mM KCl, 1 mM MgCl₂, 0.5 mM EDTA, 5% (v/v) glycerol, 4% Ficoll, 0.5 mg/mL BSA (bovine serum albumin), 0.5 mM DTT, 0.1% (v/v) NP-40) at 4°C for 1 h. Also, 1 μM Tel-G and Tel-C (Table S2) was annealed at 95°C for 5 min in 20 mM NaCl containing buffer to obtain duplex, and then incubated for another 1 h at 4°C with or without incubated protein. The samples were resolved by using 8% Native-PAGE, followed with silver staining. The pictures were taken by using ImageScanner III (GE Healthcare, Northampton, MA, USA; Abcam, Cambridge, UK; SantaCruz, Heidelberg, Germany; Life Technologies, Shanghai, China; Cell Signaling Technology, Shanghai, China; Bioteke, Beijing, China; Roche, Basel, Switzerland; Invitrogen, Carlsbad, CA, USA; Beckman Coulter, Brea, CA, USA; Multi Sciences, Hangzhou, China).

Filter-binding assay

Filter-binding assays were performed as follows. In brief, a nylon membrane was placed directly below a nitrocellulose membrane to trap any DNA not retained on the nitrocellulose. The nitrocellulose membrane was treated with 0.5 M KOH for 10 min at 4°C. Single-stranded biotin-labelled Tel-GC (200 fM, Table S2) was incubated with 10 pM TRF1 protein in 20 μL binding buffer (25 mM HEPES, 100 mM NaCl, 1 mM DTT, 1 mM EDTA, 10% glycerol, pH 7.9) at 25°C for 30 min and then incubated for 1 h at 25°C after addition of compound 2c. All samples were applied to the membrane under vacuum and washed for three times with 1× binding buffer. The cross-linking reaction was carried out under UV irradiation at 265 nm for 10 min. Detection of biotin-labelled DNA was carried out using a Chemiluminescent Nucleic Acid Detection Module Kit (Thermo Fisher Scientific).

Chromatin immunoprecipitation

ChIP experiments were performed using a Pierce Agarose ChIP Kit (Thermo Fisher Scientific) following the manufacturer’s protocol with all buffers used included in the kit. Ten per cent of the lysate was removed for using as an input, and 2 μg of antibody against TRF1 (#ab1423, Abcam, Cambridge, UK) was used for ChIP. Normal rabbit IgG (#sc-3888, Santa Cruz, Heidelberg, Germany) was used as a negative control. After extensive washing with a buffer provided in the kit, the DNA was extracted from immunoprecipitated chromatin and amplified by using PCR with Tel1 and Tel2 primers (Table S2).

Microscale thermophoresis experiment

Protein or DNA was fluorescently labelled by using a Monolith NTTM Protein Labeling Kit following the manufacturer’s protocol. The fluorescent molecule was incubated with compounds of different concentrations. The MST analyses were performed by using a Monolith NT.115, and the fitting curve was obtained by using NT analysis 1.4.23. The K_D is the numeric equivalent of the concentration of compound when the response is half of the plateau response (Rmax) on the fitting curve.

Surface plasmon resonance experiment

SPR measurement was performed on a ProteOn XPR36 Protein Interaction Array system (Bio-Rad Laboratories, Hercules, CA) using a Neutravidin-coated GLH sensor chip. For immobilization, Tel-C and duplex DNA (5ʹ-biotin-[TATAGCTATA-HEG-TATAGCTATA]-3ʹ) were biotinylated and attached to a reptavidin-coated sensor chip. Six concentrations were injected simultaneously at a flow rate of 25 mL/min for 200 s of association phase, followed with 300 s of dissociation phase at 25°C. The final graphs were obtained by subtracting blank sensorgrams from i-motif or duplex sensorgrams. Data were analysed with ProteOn manager software.

CD spectroscopy and CD-melting experiment

A final concentration of 2 μM oligomers were re-suspended in 1 × BPES (buffer phosphate saline) (pH 4.5, 5.5 or 6.2) with or without compounds, and then annealed by heating at 95°C for 5 min, gradually cooled to room temperature and stored at 4°C. CD Spectra from 230 to 350 nm were recorded on a Chirascan® circular dichroism spectrophotometer (Applied Photo-physics, UK) with a 10 mm path length quartz cuvette. The buffer blank was subtracted for all spectra, and the final analysis was carried out by using GraphPad Prism5. For CD melting experiments, annealed samples were incubated with 5 eq compound for 6 h, and molar ellipticity at 288 nm was measured over a temperature range of 20–90°C with a heating rate of 2.5°C/min.

Fluorescence spectroscopic assay

Fluorescence spectroscopic assays were conducted on the Fluoromax-4 luminescence spectrophotometer (HORIBA, USA). Individual DNA was prepared in 1 × BPES (pH 6.2) buffer for the designated concentrations and annealed by heating at 95°C for 5 min, and then cooled slowly to 25°C. Before testing, 3 μM DI26 or other compound was added and allowed to equilibrate for 2 min. The excitation wavelength was fixed at 410 nm, and the emission spectra in the range of 430–650 nm were collected at room temperature.

Cytotoxic assay

Cells were plated at a density of 5 × 10³ per well in 96-well microplates with adherent culture overnight. Cells were incubated with various concentrations of compounds (from 0 to 100 μM) for 48 h. Then, 20 μL of 2.5 mg/mL MTT solution was added to each well, and the cells were further incubated for 4 h. Next, medium was discarded and DMSO (100 μL per well) was added to each well to dissolve the formazan dye. The absorbance at 570 nm was measured using a microplate reader (Bio-Tek. USA). All experiments were parallel performed in triplicate, and the IC₅₀ values were determined from the mean OD values of the triplicate tests versus the drug concentration curves. In order to investigate the effect of our compound in combination with telomerase inhibitor BIBR1532, we assessed A549 cell viability upon addition of our individual compound together with BIBR1532. After 48 h incubation, inhibition ratio and IC₅₀ value of the individual compound in combination with BIBR1532 were calculated based on cytotoxic assay data.

Enzyme-linked immunosorbent assay

The streptaWell® High Bind plates (Roche, Basel, Switzerland) were used in the ELISA assays. The 5′-biotin-labelled Tel-C (Table S2) was diluted to 100 nM in 200 μL 20 mM NaCl containing buffer, and then annealed with fivefold excess of non-biotinylated oligomers by heating to 95°C for 5 min and followed by cooling to room temperature. To prepare the protein samples, the concentration of 2c was initially 50 μM and half-diluted six times in blocking buffer (50 mM KH₂PO₄, 100 mM KCl, pH 7.4, 3% BSA). After adding 500 nM TRF1, the samples were incubated at 37°C for 1 h. The wells were rehydrated with 200 μL phosphate buffer saline (PBS), and then incubated with the DNA samples at 4°C overnight. The samples were washed three times with ELISA buffer (50 mM KH₂PO₄, 100 mM KCl, pH 7.4) and blocked with 200 μL of blocking buffer for 1.5 h at 25°C. The samples were washed three times with ELISA buffer, incubated with protein samples at 37°C for 1 h, washed three times with washing buffer (ELISA buffer containing 0.1% Tween-20) and incubated with primary His-probe antibody (H3) (sc-8036, Santa Cruz) at 4°C overnight. Then, the samples were washed three times with washing buffer, incubated with 100 μL secondary HRP-linked Antibody (#7076S, Cell Signaling Technology, Shanghai, China) for 1 h at 37°C and washed three times with washing buffer. Following addition of 100 μL of TMB Chromogen® Solution (Life Technologies, Shanghai, China), plates were incubated in the dark for 5 min, and the reaction was stopped by addition of 50 μL 1 M H₂SO₄. Absorbance was measured at 450 nm, and the data were analysed using GraphPad Prism® 5 via Hill fitting. To evaluate the influence on the affinity between TERRA and TRF1, the 5ʹ-biotin-labelled TERRA (Table S2) was diluted to 100 nM in DEPC water containing 10 mM Tris HCl, pH 7.4, 100 mM KCl. For investigating the binding pattern among TERRA, TRF1 and telomeric duplex DNA, 500 nM TERRA was incubated with TRF1.

Reverse transcription-quantitative polymerase chain reaction

Total RNA was isolated from A549 cells treated with compounds for 24 h using RNAiso Plus (Takara, #39,109), and 3 μg of RNA was used as a template for reverse transcription using a PrimeScript™ RT reagent Kit with gDNA Eraser (#RR047A) following the manufacturer’s protocol. Genomic DNA was removed with gDNA Eraser at 42°C for 5 min and then subjected to cDNA synthesis, which was carried out in 20 μL volume containing 0.1 μM TERRA and glyceraldehyde 3-phosphate dehydrogenase (GAPDH)-specific primers (Table S3). The mixtures were incubated at 50°C for 15 min for reverse transcription and then at 85°C for 15 s. PCR was performed in duplicate using 0.5 μM primers (Table S3) and 2× RealStar Green Fast Mixture with 1 μL cDNA template using the LightCycler®480 system (Roche). PCR was conducted at 95°C for 5 min, and then 40 cycles of 95°C for 15 s and 60°C for 1 min, followed with a dissociation stage at 95°C for 5 s, 65°C for 1 min, and then 40°C for 30 s for melting curve analysis.

Western blot

A549 cells treated with compounds for 48 h were collected and lysed in RIPA lysis buffer (Bioteke, Beijing, China), and then the protein concentrations were determined by using BCA protein assay kit (Thermo Fisher Scientific). An equal amount of proteins (20 μg) were resolved by using 10% SDS-PAGE and then transferred to 0.22 μm (PVDF: polyvinylidene fluoride) membrane. The blots were blocked for 1 h with 5% BSA and probed with/primary antibodies (1:1000). After three washes in TBST (Tris buffered saline Tween), the membranes were incubated with corresponding secondary antibodies (1:4,000). Blots were visualized by using chemiluminescence, and blot images were acquired by using Tanon-4200SF gel imaging system (Shanghai, China). Antibodies to GAPDH (#5174, Cell Signaling Technology, Shanghai, China), Phospho-ATR (Thr428) (#2853, Cell Signaling Technology), Phospho-ATM (Ser1981) (#5883, Cell Signaling Technology), Phospho-Chk1 (Ser345) (#2348, Cell Signaling Technology), Phospho-Chk2 (Thr68) (#2197, Cell Signaling Technology), Phospho-p53 (Ser15) (#9286, Cell Signaling Technology), Phospho-BRCA1 (Ser1524) (#9009, Cell Signaling Technology), γH2AX (#9718, Cell Signaling Technology), TRF1 (ab1423, Abcam), β-actin (bsm-33,036 M, Bioss), anti-rabbit IgG-HRP (#7074, Cell Signaling Technology) and anti-mouse IgG-HRP (#7076, Cell Signaling Technology) were used.

Immunofluorescence

A549 cells grown on glass coverslips were fixed with 4% formaldehyde, then permeabilized with 0.1% triton-X100/PBS, and finally blocked with 5% goat serum/PBS. For immunolabeling experiments, cells were incubated with the following primary antibodies: γH2AX antibody (#9718, Cell Signaling Technology), TRF1 (#sc-271,485, Santa Cruz) and TRF2 antibody (#ab23579, Abcam) at 4°C overnight. Then cells were washed with 5% goat serum/PBS and incubated with anti-rabbit Alexa 488-conjugated antibody (#A21206, Life Technology), and anti-mouse Alexa 555-conjugated antibody (#A21427, Life Technology) at 37°C for 3 h. Nuclei were counterstained with DAPI (Invitrogen, Carlsbad, CA, USA). Fluorescence signals were recorded by using a LSM710 microscope (Zeiss, GER) and analysed with ZEN software. For quantitative analysis of γH2AX and TIFs (refer to the co-localization of γH2AX with TRF foci) positivity, 50 nuclei were counted in each group, and the the standard error of the mean (S.E.M.) was calculated from three replicates. Frequency distribution graphs were plotted using GraphPad Prism (GraphPad Software Inc).

Cell-cycle arrest analysis

Cell-cycle arrest analysis was performed by using Cell Cycle Staining Kit (#CCS012, Multi Sciences, Hangzhou, China) following the manufacturer’s protocol in the kit. The cells were analysed by using flow cytometry with an Epics Elite XL flow cytometer (Beckman Coulter, Brea, CA, USA). For each analysis, 10,000 events were collected. The cell-cycle distribution was analysed with EXPO32 ADC software.

FITC annexin V/PI cell apoptosis analysis

FITC annexin V/PI cell apoptosis analysis was performed by using the annexin V-FITC/PI Apoptosis Kit (#AP101-100-kit, Multi Sciences) following the manufacturer’s protocol in the kit. Emitted fluorescence was quantitated by using Epics Elite flow cytometry (Beckman Coulter). For each analysis, 10,000 events were collected.

Real-time cellular analysis

A549 cells were seeded on E-Plate 16 PET and cultured for about 16 h before treatment with compound. Cells were treated with various concentrations of compounds or DMSO as a control. The cells were sampled every minute for 15 min. The data were obtained by using GraphPad Prism 5.

Colony formation assay

Five hundred A549 cells were seeded on cell culture dish (60 × 15 mm) and exposed to compounds at various concentrations and DMSO as a control at 37°C in a humidified atmosphere of 5% CO₂. DMEM was replaced and different concentrations of compounds were added every 3 days. Cells were fixed with methyl alcohol and dyed with crystal violet after cultured for 9 days. The pictures were taken by using cell imager.

Cell scrape assay

A549 cells were subsequently seeded in 6-well culture plates (500,000/well) at 37°C in a humidified atmosphere with 5% CO₂. After 12 h preculture, a cross-shaped scrape was made through the monolayer A549 cells using a plastic pipet’s tip. Cells were washed with PBS and incubated with compounds at different concentrations. Several wounded areas were observed and photographed using microscopy after scratching and then culturing for 0 and 48 h. The edge of the cells was marked with a line for observing easily.

Long-term cell culture

Long-term cell proliferation experiments were conducted using A549 cells. The cells were grown in culture dishes at 2.5 × 10⁵ per dish (60 mm) and were exposed to DI26 at various concentrations and BIBR1532 as a positive control every 4 days. The cells in the control and drug-exposed dishes were counted, and the dishes were reseeded with 2.5 × 10⁵ cells. The process was repeated until a growth platform appeared. Results were expressed as the cumulated population doubling as a function of the time of culture.

Senescence analysis

Cell senescence was analysed with Senescence-associated β-Galactosidase Staining Kit (Beyotime, Shanghai, China) following the manufacturer’s protocol after treatment of cells with compounds for 15 days. The cells were viewed and photographed under an optical microscope.

Evaluation of in vivo anti-tumour activity

Male BALB/c nude mice (5 weeks old) were purchased from and housed at the Experimental Animal Center of Sun Yat-sen University (Guangzhou, China) and maintained in pathogen-free conditions (12 h light−dark cycle at 24 ± 1°C with 60–70% humidity and provided with food and water ad libitum). All procedures were approved by the Animal Care and Use Committee of Sun Yat-sen University and conformed to the legal mandates and national guidelines for the care and maintenance of laboratory animals. Each mouse was injected subcutaneously with a 22-gauge needle containing 1 × 10⁷ A549 cells. When the volume of the tumour reached 300 mm³, the tumour was divided into 3 mm³ and subcutaneously transplanted. The volume of the tumour was measured with an electronic calliper and calculated as 1/2 × length × width² in mm³ once every 3 days after implantation until reached approximately 50 mm³. Then the mice were randomly divided into five groups of seven animals and treated intraperitoneally (ip) with various regimens every day for the entire period of observation (20 days). The mice of control group were treated with an equivalent volume of vehicle, and 2c hydrochloride-treated group was treated with 2c hydrochloride at dosage of 25.0, 12.5 or 6.25 mg/kg body weight, and cisplatin-treated group was treated with cisplatin at dosage of 0.5 mg/kg body weight every other day. The tumour volume and body weight of the mice were measured every 2 days after drug treatment. In the end of the observation period, mice were euthanized by cervical dislocation, and the tumours were removed and weighed. The inhibition ratio (IR) was calculated according to the formula: IR = (1 − mean tumour weight of the experimental group/mean tumour weight of the control group) × 100%.

Funding Statement

We thank National Natural Science Foundation of China (Grant 21977123), Guangdong Basic and Applied Basic Research Foundation (Grant 2019A1515011074), Youth Innovation Talents Project of Colleges and Universities in Guangdong Province (Grant 51340208), Natural Science Foundation of Guangdong Province (Grants 2017A030313089, 2017A030308003), Outstanding Talents of Guangdong Special Plan (2019JC05Y456), and Guangdong Provincial Key Laboratory of Construction Foundation (2017B030314030) for financial support of this study.

Disclosure statement

The authors report no conflict of interest.

Supplementary material

Supplemental data for this article can be accessed here.