Telomerase is expressed in the majority (>85%) of tumors, but has restricted expression in normal tissues. Long-term telomerase inhibition in malignant cells results in progressive telomere shortening and reduction in cell proliferation. Here we report the synthesis and characterization of radiolabeled oligonucleotides that target the RNA subunit of telomerase, hTR, simultaneously inhibiting enzymatic activity and delivering radiation intracellularly. Oligonucleotides complementary (Match) and noncomplementary (Scramble or Mismatch) to hTR were conjugated to diethylenetriaminepentaacetic dianhydride (DTPA), allowing radiolabeling with the Auger electron-emitting radionuclide indium-111 (111In). Match oligonucleotides inhibited telomerase activity with high potency, which was not observed with Scramble or Mismatch oligonucleotides. DTPA-conjugation and 111In-labeling did not change telomerase inhibition. In telomerase-positive cancer cells, unlabeled Match oligonucleotides had no effect on survival, however, 111In-labeled Match oligonucleotides significantly reduced clonogenic survival and upregulated the DNA damage marker γH2AX. Minimal radiotoxicity and DNA damage was observed in telomerase-negative cells exposed to 111In-Match oligonucleotides. Match oligonucleotides localized in close proximity to nuclear Cajal bodies in telomerase-positive cells. In comparison with Match oligonucleotides, 111In-Scramble or 111In-Mismatch oligonucleotides demonstrated reduced retention and negligible impact on cell survival. This study indicates the therapeutic activity of radiolabeled oligonucleotides that specifically target hTR through potent telomerase inhibition and DNA damage induction in telomerase-expressing cancer cells and paves the way for the development of novel oligonucleotide radiotherapeutics targeting telomerase-positive cancers.
These findings present a novel radiolabeled oligonucleotide for targeting telomerase-positive cancer cells that exhibits dual activity by simultaneously inhibiting telomerase and promoting radiation-induced genomic DNA damage.
Human telomeres consist of repetitive TTAGGG nucleotide sequences at the end of chromosomal DNA. These structures protect chromosomes from fusion, deterioration, and recognition by DNA damage response proteins. In normal somatic cells, telomeres erode during each cell division as a consequence of the inability of unidirectional DNA synthesis to completely replicate the linear chromosomes in the so-called end replication problem (1, 2). When telomeres shorten to a critical length, cells undergo senescence or apoptosis, limiting the number of possible cell divisions (1, 2). In the case of cancer, cellular processes are activated to maintain telomere length, providing the cell with unrestricted cell division potential (3). In the majority of malignancies (>85%), telomere maintenance is the result of the upregulation of telomerase, a protein that has a restricted expression profile in normal tissue (4). Human telomerase is a ribonucleoprotein minimally composed of human telomerase reverse transcriptase (hTERT), a catalytic subunit, and a human telomerase RNA component (hTR), a template sequence-containing subunit encoding the telomeric repeats (5). The enzyme acts by catalyzing the de novo addition of hexanucleotide telomeric repeats to the chromosome-ends, resulting in cell immortality (6). In addition, extra-telomeric properties of telomerase have also been uncovered. hTERT can influence tumor development, oncogenesis, and inflammation via NF-κB and Wnt/β-catenin pathways (7). hTR has been shown to protect against apoptosis and oxidative stress (8, 9), and to activate DNA-dependent kinase (DNA-PKcs) to phosphorylate hnRNPA1, which is critical for capping telomeres (10, 11).
As a consequence of its differential expression in cancer versus normal tissue and its role in the maintenance of the malignant phenotype, telomerase is regarded as an attractive cancer therapeutic target and several telomerase-inhibition strategies have been developed. In particular, oligonucleotides complementary to the hTR template sequence act as catalytic inhibitors of telomerase and have shown promise in preclinical investigations (12–14). The use of oligonucleotides as therapeutic agents has been potentiated by the modulation of nuclease susceptibility, cell permeability, and target affinity, while oligonucleotide-based inhibitors offer exquisite specificity of action and high binding affinity, owing to the necessary formation of specific base-pairing during hybridization (15). One such inhibitor (GRN163L, Imetelstat) is undergoing clinical trials in patients with a variety of cancer indications and positive outcomes have been reported in patients with essential thrombocytopenia and myelofibrosis (16, 17). Interestingly, it has been demonstrated that hTR inhibition results in chemo- and radiosensitization (18–22), which is attributed to both telomere length-dependent and -independent mechanisms (23).
On the basis of these observations, we hypothesized that targeting hTR with a radiolabeled complementary oligonucleotide would provide a largely unexplored and potent approach for the targeted radiotherapy of telomerase-positive cancer. The localization of telomerase in the nucleus creates an opportunity for targeting with short-range (nm scale) Auger electron-emitting radionuclides, such as Indium-111 (111In). High linear energy transfer Auger electrons are highly toxic when emitted in close proximity to DNA, but have little effect outside this range, thereby providing a potentially highly selective therapy with a low side-effect profile (24). In this study, we describe the design and biological characterization of 111In-labeled oligonucleotides that target hTR, thereby presenting a novel strategy for targeting telomerase-positive tumors by combining selective telomerase inhibition and molecular radiotherapy in a single agent.
Materials and Methods
All chemicals were purchased from Sigma-Aldrich, unless otherwise stated.
MDA-MB-435 (human melanoma), SKBR3 (human breast cancer), U2OS (osteosarcoma), and WI38 (human fibroblasts) cells were obtained from ATCC. The MDA-MB-231/H2N human breast cancer cell line was a gift from Dr. Robert Kerbel (University of Toronto, Toronto, Canada). Cells were maintained in DMEM supplemented with 10% FBS (Gibco) and 1% penicillin/streptomycin/glutamate. Cell line authentication (short tandem repeat profiling) confirmed the identity of the cells used in this work. Cells were regularly checked for Mycoplasma infection and were not used beyond passage 20.
Oligonucleotide synthesis and labeling
2′ O-methyl RNA (2′OMeRNA) oligonucleotides, with or without 5′-amino linkers, were synthesized and HPLC purified by Sigma-Aldrich. Sequences: Match 5′-CAGUUAGGGUUAG-3′; Scramble 5′-GCAGUGUGAUGAU-3′; Mismatch 1 5′-CAGUUA CGCUUAG-3′; and Mismatch 2 5′-CAGAUACGCUUAG-3′. As a control, phosphorothioate (PS) DNA oligonucleotides were similarly obtained. Sequences: PS Match 5′-CAGTTAGGGTTAG-3′; PS Mismatch 5′-CAGTTAGAATTAG-3′ (mismatched bases are underlined).
For diethylenetriaminepentaacetic dianhydride (DTPA) conjugation, oligonucleotides with 5′-amino linkers were reconstituted in sodium bicarbonate buffer (0.1 mol/L, pH 8.3). DTPA was combined with oligonucleotides in 20-fold molar excess and incubated at room temperature for 60 minutes. Fractions were separated by size exclusion chromatography (SEC) on a P4-column (Bio-Rad) with sodium citrate buffer (0.1 mol/L, pH 5). Oligonucleotide concentration was determined by measuring absorbance at 260 nm using a NanoDrop Spectrometer (Thermo Fisher Scientific). Radiolabeling was performed in sodium citrate buffer by the addition of 0.2 MBq of [111In]InCl3 (subsequently referred to as 111InCl3) followed by incubation at room temperature for 60 minutes. Fractions were separated on a P4-column and gamma-counted for 60 seconds using a Wizard Scintillation Counter (PerkinElmer). Oligonucleotides were radiolabeled to the desired molar activity and the radiolabeling efficiency determined by instant thin-layer chromatography (ITLC) in sodium citrate buffer (0.1 mol/L, pH 5).
Oligonucleotides were linked to Cy3 for confocal microscopy studies. Oligonucleotides with 5′-amino linkers were reconstituted in sodium bicarbonate buffer (0.1 mol/L, pH 8.3) to a final concentration of 220 μmol/L. Monoreactive functional Cy3 dye (GE Healthcare) was reconstituted in DMSO and added to the oligonucleotide, before incubation for 60 minutes at room temperature. Reaction products were separated into 50 μL PBS fractions using P4 SEC, as before. Fractions were analyzed using a NanoDrop Spectrometer (Thermo Fisher Scientific) at 260 and 550 nm.
Telomeric repeat amplification protocol
The telomeric repeat amplification protocol (TRAP) assay (EMD-Millipore) was setup according to the manufacturer's instructions, with an optimized PCR cycle (30°C, 30 minutes; 36 cycles of 94°C/30 seconds; 53.5°C/30 seconds; 72°C/60 seconds; and a final extension step of 72°C/3 minutes), using 200 ng of protein from MDA-MB-435 cell lysate, unless otherwise stated. For inhibition experiments, inhibitor or control was added immediately prior to the telomerase extension step (30°C, 30 minutes). Samples were analyzed using a fluorescence plate reader (Tecan). The determined telomerase activity was normalized to untreated control.
For analysis of 111In-labeled oligonucleotides, the TRAP assay was modified for the accommodation of radioactive samples. In this system, DTPA-conjugated oligonucleotides were radiolabeled to a molar activity of 27 MBq/nmol and preincubated with cell lysate for 24 hours. Radiolabeled inhibitors were combined with MDA-MB-435 lysate at 27 ng/μL protein concentration in a total volume of 60 μL. Following incubation for 24 hours to allow for radioactive decay, inhibitor–enzyme complexes (7.5 μL) were added to 48 μL of TRAP PCR mix to give 200 ng of lysate protein and the final concentration of inhibitor as indicated. Data were fitted using fixed-slope nonlinear regression and compared using an exact-sum-of-squares F-test and subject to one-way ANOVA with post hoc Tukey test.
Internalization of radiolabeled oligonucleotides
Cells were seeded in a 24-well plate (2 × 104 cells/well). Oligonucleotides were combined with Transfast Transfection Reagent (Tfx; Promega) according to the manufacturer's protocol in antibiotic-free medium to a concentration of 220 nmol/L and incubated for 15 minutes at room temperature. Cell medium was removed, replaced with transfection complexes, and incubated for the indicated time. Following incubation, medium was aspirated and retained. Cells were washed twice with PBS, and the washes combined with the medium to constitute the free fraction. Cell membranes were washed using glycine buffer (0.1 mol/L, pH 2.5) at 4°C for 6 minutes and rewashed with PBS. To lyse the cells, sodium hydroxide (0.1 mol/L) was added and plates incubated for 20 minutes at room temperature. The internalized fraction was collected and combined with two PBS washes. Fractions were counted using a Wizard Gamma-Counter (PerkinElmer).
Cells were seeded in a 24-well plate (2 × 104 cells/well). Cells were transfected with 250 μL of 220 nmol/L oligonucleotide for 2.5 hours before addition of 250 μL medium and incubated for a further 24 hours. Cells were harvested with trypsin (Gibco) following washing in PBS. Harvested cells were counted before plating in 6-well plates at a density sufficient to give >50 colonies for counting. Untreated cells were typically seeded at 750 cells/well. Colonies were grown for >7 days, washed in PBS, and stained with 1 % methylene blue (Alfa Aesar) in 50% methanol (Thermo Fisher Scientific). Colonies containing >50 cells were counted. The surviving fraction (SF) was calculated using the plating efficiency of untreated cells. Data were fitted using the linear-quadratic function and curves were compared using an exact-sum-of-squares F-test. Differences between treatment groups were analyzed with two-way ANOVA.
Cell viability assay
WI38 cells were seeded in a 24-well plate (1 × 104 cells/well). Cells were transfected with 250 μL of oligonucleotide (220 nmol/L) for 2.5 hours before addition of 250 μL medium and incubation for a further 24 hours. At 24 hours, the solution was removed and replaced with fresh medium. After 72 hours, 0.5 mg/mL MTT (thiazolyl blue tetrazolium bromide) dissolved in serum-free medium was added for 120 minutes and the formazan product eluted using DMSO. Absorbance at 540 nm was quantified by plate reader (reference wavelength 650 nm) and the relative cell viability of cell populations determined as a fraction of mock-treated control.
Cells were seeded in 8-well chamber slides (Thermo Fisher Scientific; 2 × 104 cells/well). Cells were transfected for the indicated time with 220 nmol/L oligonucleotide in 100 μL, with an additional 100 μL of medium added after 2.5 hours. A 137Cs-irradiator was used to deliver 4 Gy of γ-radiation to induce γH2AX foci, as a positive control for staining. Irradiated cells were incubated at 37°C for 1 hour to allow for induction of foci. Immunostaining (anti-γH2AX 50-636-KC, Millipore and anti-coilin ab87913, Abcam) and confocal image acquisitions were performed as described previously (25). Intranuclear DNA damage foci were counted manually in >100 cells per treatment. Nuclear Cy3 mean intensity was quantified from >50 cells per treatment. Groups were compared with one-way ANOVA with a post hoc Tukey test. Images represent several fields of view acquired using a 40×/1.2 or 63×/1.4 objective lens, with 4× digital zoom.
Data were plotted as the mean ± SD of the stated number of replicates, unless otherwise indicated. Statistical significance was determined using GraphPad Prism 6.03 software and is indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001 (two-sided).
Synthesis and stability of 111In-labeled oligonucleotides
2′OMeRNA oligonucleotides were selected on the basis of their resistance to nucleases and serum stability (26). Amino-modified 2′OMeRNA oligonucleotides containing a sequence complementary to the template region of hTR were conjugated to DTPA to enable subsequent labeling with 111In (Fig. 1). The region of complementarity between the Match oligonucleotide and hTR is depicted in Table 1, which despite some overlap is distinct from the GRN163L site. To ensure sequence specificity, several noncomplementary control oligonucleotides (Mismatch 1, Mismatch 2, and Scramble) were generated and labeled. Oligonucleotides were purified (or confirmed pure) by HPLC and DTPA conjugation was confirmed by mass spectrometry (Supplementary Fig. S1), and the conjugation reaction products were diagnostically radiolabeled (0.2 MBq) before separation by SEC. Analysis of the reaction products by gamma-counting demonstrated the elution of two distinct radiopeaks, at 700 μL (peak 1) and 1,400 μL (peak 2; Fig. 2A). Unreacted oligonucleotide was found to elute at 700 μL by absorbance measurement at 260 nm and radiolabeled, unreacted DTPA at 1,400 μL by gamma-counting. Reinjection of the fractions constituting peak 1 on a second SEC column verified the presence of a single radiospecies (Fig. 2B). As expected, both the amino modification and conjugation of DTPA were found to be necessary for oligonucleotide radiolabeling (Supplementary Fig. S2).
|Oligonucleotide .||Sequence 5′–3′ .||Modification .||5′-end .|
|hTR template (3′–5′)||GUCAAUCCCAAUCUGUU||—||—|
|Mismatch 1||CAGUUACGCUUAG||2′OMeRNA||Amine (AmC6)|
|Mismatch 2||CAGAUACGCUUAG||2′OMeRNA||Amine (AmC6)|
|Oligonucleotide .||Sequence 5′–3′ .||Modification .||5′-end .|
|hTR template (3′–5′)||GUCAAUCCCAAUCUGUU||—||—|
|Mismatch 1||CAGUUACGCUUAG||2′OMeRNA||Amine (AmC6)|
|Mismatch 2||CAGAUACGCUUAG||2′OMeRNA||Amine (AmC6)|
NOTE: The template region of hTR is shown in bold and mismatched bases are underlined. For reference, the sequence of GRN163L (Imetelstat) is also shown.
To test the stability of the radiolabeled construct, the SEC profile was analyzed following incubation for 0 or 24 hours. A single radiopeak was observed for both timepoints (Fig. 2C). This peak accounted for 92% and 94% of the radioactivity at 0 and 24 hours, respectively, determined by AUC analysis. These results indicated the short-range Auger electrons do not induce significant autoradiolysis of oligonucleotide constructs over this time period. High radiolabeling efficiency (>90%) was confirmed by ITLC and molar activities up to 27 MBq/nmol were employed for subsequent in vitro experiments.
hTR-targeted 111In-labeled oligonucleotides inhibit telomerase activity
To assess telomerase inhibition by 111In-labeled oligonucleotides, relative telomerase activity was determined by a PCR-based TRAP assay in a panel of cancer cell lines. MDA-MB-435 melanoma cells were found to have the highest telomerase activity, with MDA-MB-231/H2N and SKBR3 breast cancer cells expressing 86.3% ± 0.5% and 42.6% ± 0.4% of that activity, respectively (Fig. 3A). The osteosarcoma cell line, U2OS, was found to lack detectable telomerase activity in the TRAP assay, as described previously (27).
In each cell line, the effect of nonradiolabeled 2′OMeRNA Match oligonucleotide on telomerase activity was tested immediately following addition of the construct. Following normalization to untreated control, inhibition of 75%, 74%, and 72% was recorded in MDA-MB-435, MDA-MB-231/H2N, and SKBR3 lysates incubated with Match oligonucleotide at a concentration of 300 nmol/L, respectively (Fig. 3A). In contrast, Scramble showed no inhibitory activity. The amplification of a preelongated control substrate, TSR8, was not affected by Match (Supplementary Fig. S3), confirming inhibition was occurring during telomerase-dependent substrate elongation (28). Further characterization showed that Match inhibited the telomerase activity of MDA-MB-435 lysate in a dose-dependent manner with maximum inhibition of 91% at a concentration of 1 μmol/L and an IC50 of 147.4 nmol/L [95% confidence interval (CI), 91.1–238.3 nmol/L; Fig. 3B]. DTPA conjugation did not significantly alter the IC50 (101.4 nmol/L; 95% CI, 75.9–135.6 nmol/L) of Match (P > 0.05; Fig. 3C). Furthermore, in a TRAP assay modified for inclusion of radioactive samples, no difference in inhibitory potency was observed following radiolabeling with 111In (Fig. 3D). In this modified assay where, as reported previously, preincubation with inhibitor leads to a perceived increase in potency (29), the IC50 values were found to be 19.6 nmol/L (95% CI, 11.5–33.3 nmol/L) for 111In-DTPA-Match (subsequently referred to as 111In-Match) and 11.5 nmol/L (95% CI, 7.7–17.2 nmol/L) for DTPA-Match (P > 0.05). Inhibition of >96% was achieved with 111In-Match at a concentration of 1 μmol/L. The IC50 of Match recorded in the modified TRAP assay was in agreement with previously published data, particularly where preincubation was explicitly employed (30). In all cases, disruption of oligonucleotide complementarity to hTR (Scramble, Mismatch 1, and Mismatch 2) abrogated inhibitory activity, irrespective of modification with DTPA (Fig. 3B and C; Supplementary Fig. S4). Importantly, radiolabeling did not confer inhibitory capability on control oligonucleotides, with 111In-DTPA-Scramble (subsequently referred to as 111In-Scramble) lacking inhibitory activity (Fig. 3D), confirming the preservation of sequence-specific telomerase inhibition.
To generalize these findings using an alternative oligonucleotide chemistry, we employed PS DNA oligonucleotides as an additional control (29, 31). However, PS oligonucleotides were found to inhibit telomerase in a sequence-independent fashion (Supplementary Fig. S5) and so were unsuitable for therapeutic development. This finding is in agreement with published work (29, 32).
Delivery of 111In-labeled oligonucleotides to cancer cells
To facilitate in vitro cellular characterization 111In-Match and 111In-Scramble were introduced into cancer cells using liposomal transfection reagents. For each cell line, the transfection conditions were optimized to approximately equalize the peak delivery of 111In-Match and 111In-Scramble. Examining the uptake of oligonucleotides over a 24-hour time period, maximum internalization occurred after approximately 4 hours of treatment in MDA-MB-435, U2OS, SKBR3, and MDA-MB-231/H2N cells (Fig. 4A–D). The peak uptake of 111In-Match achieved was 8.9% ± 1.6%, 6.1% ± 0.1%, 10.8% ± 1.1%, and 6.8% ± 0.3% of total administered radioactivity in MDA-MB-435, U2OS, SKBR3, and MDA-MB-231/H2N cells, respectively. Interestingly, while the maximum uptake for cell lines was comparable, 111In-Match was retained to a far greater extent than 111In-Scramble in the telomerase-positive cell lines (Fig. 4A, C, and D); an effect not apparent in telomerase-negative U2OS cells where both sequences demonstrated comparable temporal uptake profiles (Fig. 4B).
hTR-targeted oligonucleotides enter the nucleus and associate with Cajal bodies
As telomerase is predominantly localized to the nucleus (1) and Auger electron-emitting radionuclides exert their therapeutic effects most effectively when in close proximity to DNA, the subcellular localization of the oligonucleotides was investigated using fluorophore-labeled constructs in MDA-MB-435 cells. Cy3-labeled oligonucleotides were synthesized and purified by SEC (Supplementary Fig. S6). Following transfection for 2.5 hours, a robust Cy3 signal was observed primarily in the cytoplasm of cells treated with Cy3-Match or Cy3-Scramble oligonucleotides (Fig. 4E). In contrast, 24 hours after treatment, Cy3-Match oligonucleotides were retained, with a focal nuclear distribution apparent in approximately 30% of cells, while Cy3-Scramble–treated cells showed a substantial decrease in signal intensity, which remained predominantly cytoplasmic (Fig. 4E). Consistent with the uptake of radiolabeled oligonucleotides, a reduced signal was recorded in cells treated with label alone (Cy3), transfected-unlabeled oligonucleotides, or nontransfected Cy3-oligonucleotides (Supplementary Fig. S7). Quantification of the oligonucleotide–Cy3 signal confirmed the preferential nuclear retention of Match over Scramble constructs after 24 hours (Supplementary Fig. S8).
Previous studies have reported that hTR localizes to subnuclear Cajal bodies during telomerase biosynthesis (33). Importantly, the punctate nuclear oligonucleotide signal that was observed in cells incubated with Cy3-Match for 24 hours was found to associate with the periphery of coilin, a protein marker of Cajal bodies (Fig. 4E) (34). Such an association was absent in telomerase-negative U2OS cells (Supplementary Fig. S9). These findings suggest telomerase-targeted oligonucleotides localize to and are retained at the nuclear sites of telomerase assembly.
hTR-targeted 111In-labeled oligonucleotides reduce the survival of telomerase-positive cancer cells
Next, the potential of 111In-Match to selectively reduce the survival of telomerase-positive cancer cells was assessed by clonogenic assay. MDA-MB-435 and U2OS cell lines were employed as telomerase-positive and telomerase-negative cells, respectively, while treatment with 111In-labeled noncomplementary Scramble or Mismatch oligonucleotides were included to provide an indication of sequence specificity. Transfection conditions were optimized to provide comparable levels of cellular uptake for each individual oligonucleotide; approximately 5%–10% internalized radioactivity (Supplementary Fig. S10).
Following normalization to untreated control, the nonradiolabeled Match oligonucleotide (i.e., 0 MBq/nmol) was shown to have no effect on the survival of cancer cells (SF > 0.8; Fig. 5), a result in line with the notion that long-term treatment with existing telomerase inhibitors is necessary to lead to telomere erosion sufficient for therapeutic effect (1). This finding was confirmed in MDA-MB-435 cells treated with different concentrations of radiolabeled and nonlabeled oligonucleotide (Supplementary Fig. S11). A significant effect on clonogenic survival was observed only following treatment with radiolabeled Match (27 MBq/nmol), with nonlabeled Match showing no therapeutic activity over this timeframe.
As shown in Fig. 5A, 111In-Match caused a significant reduction in the clonogenic survival of telomerase-positive MDA-MB-435 cells in an 111In-dependent manner, where treatment with the highest molar activity (27 MBq/nmol) resulted in a SF of 0.14 ± 0.03. In contrast, 111In-Scramble, 111In-Mismatch 1, and 111In-Mismatch 2 sequences had minimal impact on MDA-MB-435 survival even at high molar activities (SF > 0.7 at 27 MBq/nmol). Likewise, 111InCl3 in the presence of transfection agent demonstrated no therapeutic activity. Importantly, 111In-Match did not significantly reduce the survival of telomerase-negative U2OS cells, with SF of 0.73 ± 0.16 at the maximum molar activity of 27 MBq/nmol (Fig. 5B) but exhibited therapeutic activity toward two additional telomerase-positive cancer cell lines, SKBR3 and MDA-MB-231/H2N (Fig. 5C and D). 111In-Match in the absence of liposomal transfection reagent failed to significantly reduce clonogenic survival at the equivalent levels of radioactivity, confirming the requirement for intracellular delivery of 111In for therapeutic activity (Supplementary Fig. S12).
The potential for off-target toxicity of these novel radiopharmaceuticals was further investigated by measurement of the viability of treated normal fibroblasts. These cells cannot be analyzed by clonogenic assay as they do not form colonies. In WI38 cells, shown to lack telomerase activity and exhibiting similar uptake of transfected radiopharmaceuticals (Supplementary Fig. S13), oligonucleotides radiolabeled to 27 MBq/nmol had minimal effect on cell viability (Fig. 5E). Furthermore, the modest reduction in viability occurred in an oligonucleotide sequence-independent manner, indicating a minimal and nonspecific effect on normal cells.
Long-term telomerase inhibition is known to induce downregulation of hTERT and associated downstream cellular effects, for example those leading to altered regulation of the proapoptotic pathway (35). However, in the short timescale necessary for reduction in clonogenic survival following treatment with radiolabeled oligonucleotides, no effect on hTERT expression was observed (Supplementary Fig. S14). These results indicate that hTR-targeted–radiolabeled oligonucleotides induce rapid 111In-mediated radiotoxicity selectively toward telomerase-positive cancer cells.
hTR-targeted 111In-labeled oligonucleotides induce DNA damage specifically in telomerase-positive cancer cells
To further explore the preferential activity of 111In-labeled oligonucleotides toward telomerase-positive cancer cells, the upregulation of γH2AX was investigated. γH2AX is a protein involved in the recruitment of DNA damage repair factors and is a surrogate marker of DNA double-strand breaks (36). In MDA-MB-435 cells, a molar activity–dependent increase in γH2AX foci number was observed following treatment with 111In-Match (Fig. 6A and B). Treatment with Match labeled to 0, 9, 18, and 27 MBq/nmol led to upregulation of γH2AX to 3.1 ± 3.1, 7.4 ± 5.3, 12.5 ± 6.5, and 12.7 ± 6.9 foci per nuclear confocal plane (NCP), respectively. In each case, the number of foci was significantly higher than in cells treated with Tfx alone (3.0 ± 3.3 foci/NCP) and Tfx + 111InCl3 (3.1 ± 2.5 foci/NCP). Notably, the effect of 111In-Scramble labeled to 27 MBq/nmol (3.6 ± 3.4 foci/NCP) was not significantly different from control treatments, confirming the sequence-specific effect of 111In-Match.
In contrast, in telomerase-negative U2OS cells, upregulation of γH2AX foci number was not oligonucleotide sequence dependent (Fig. 6A and B) and both 111In-Match and 111In-Scramble upregulated γH2AX foci number to a modest extent compared with controls but there was no statistically significant difference between the two constructs, and no apparent dependency on molar activity (111In-Match: 6.1 ± 5.6, 6.1 ± 6, 7.2 ± 6 foci/NCP for 9, 18, and 27 MBq/nmol; 111In-Scramble: 8.2 ± 8.3 foci/NCP for 27 MBq/nmol; and Tfx and Tfx+111InCl3: 2.0 ± 2.9 and 2.7 ± 3.3 foci/NCP, respectively). It was observed that for 111In-Scramble (27 MBq/nmol) and 4Gy γ-radiation there were statistically significantly more foci in U2OS than in MDA-MB-435 cells. We speculate that this may indicate greater DNA repair capacity in MDA-MB-435 compared with U2OS. Despite this, importantly, the induction of γH2AX foci by 111In-Match labeled to specific activities above 9 MBq/nmol was statistically significantly greater in MDA-MB-435 than in U2OS cells (P < 0.001). The modest upregulation in foci number in U2OS cells, coupled to the lack of therapeutic activity of the oligonucleotides in these cells, likely reflects the nonspecific, sublethal induction of DNA damage following internalization of radiolabeled oligonucleotides. In U2OS cells, unlabeled Match and Scramble did not elicit a significant effect (2.6 ± 3.8 and 5.5 ± 5.7 foci/NCP, respectively). Together, these results are consistent with a telomerase- and 111In-dependent mechanism of DNA damage induction in telomerase-positive cancer cells.
Since the original discovery of telomerase and the observation that it is expressed abundantly in malignant compared with normal cells, telomerase inhibition has been intensively investigated as an anticancer strategy, including in combination with external beam radiation or chemotherapy. It has been demonstrated that telomere shortening caused by downregulation of telomerase results in radiosensitization (18–23, 37–39), and, conversely, that increased telomerase expression is linked to apoptosis resistance and enhanced DNA repair (7). Telomerase protects aneuploid cells against replication stress, and promotes genomic stability in cancer cells (40, 41). Moreover, oxidative DNA damage, which is a major determinant of radiation toxicity, stimulates telomerase activity (42). We have shown previously that targeting hTERT with radiolabeled small-molecule inhibitors is a viable anticancer strategy, in which telomerase inhibition is achieved synchronously with delivery of a high radiation dose (43). Together, these findings provide a strong rationale for combining hTR inhibition with molecular radiotherapy for the treatment of neoplastic disease. Although the metal chelator DTPA has been shown to facilitate the therapeutic labeling of a wide range of biomolecules with the Auger electron-emitting radionuclide, 111In (44), the use of radiolabeled oligonucleotides has not been explored in detail. Yet oligonucleotide-based inhibitors offer high binding affinities and specificity, which, if combined with the radiotoxicity of short-range emitters such as 111In, have the potential to generate gene-targeted radiotherapeutics. This concept was proposed by Neumann and colleagues, who examined sequence-specific triplex-forming oligonucleotides to carry 125I for this purpose (45, 46). The use of modified nucleic acids confers improved characteristics on therapeutic oligonucleotides. However, for therapeutic development, the selection of alternative chemistry must be carefully coordinated with the application. This necessitates the inclusion of stringent experimental controls, as exemplified by sequence-independent inhibition of telomerase activity mediated by PS oligonucleotides reported here and previously (32).
We demonstrate that DTPA-conjugation to the 5′-end of 2′OMe-modified oligonucleotides results in site-specific 111In labeling to high molar activity (27 MBq/nmol). An oligonucleotide complementary to the template region of hTR inhibited telomerase activity in a sequence- and concentration-dependent manner, and inhibitory potency was not compromised by DTPA conjugation or 111In labeling. Upon cellular delivery, oligonucleotides targeted to hTR were retained in telomerase-positive cancer cells to a greater extent than the nontargeting control, while clonogenic assay data clearly showed a sequence-specific and dose-dependent effect of radiolabeled oligonucleotides on telomerase-positive cell survival. The rapid radiation-dependent loss of clonogenic survival demonstrated here offers a marked advantage over existing strategies reliant solely on inhibition of telomerase, which require long-term treatment for therapeutic effect (1). Crucially, telomerase-targeted–radiolabeled oligonucleotides had little effect on the survival of telomerase-negative osteosarcoma cells and normal fibroblasts, despite internalization occurring to a similar extent as in telomerase-positive cell lines. The normal fibroblast cell line, WI38, has low expression of hTR (47), and this is a possible explanation for its resistance to 111In-Match. These observations suggest a therapeutic window for the selective targeting of telomerase-positive cancer over normal tissue. At the activities assayed, a small fraction of cells survived treatment with the radiolabeled construct, possibly due to the stochastic nature of radioactive decay and the short range of Auger electrons, the existence of a radioresistant subpopulation of cells, or a failure of delivery. Use of alternative radionuclides and development of an efficient in vivo delivery system will be of future importance to address such questions.
Importantly, subcellular localization studies revealed that 24 hours after treatment the non-hTR–targeting oligonucleotides remained mainly cytoplasmic, whereas oligonucleotides complementary to hTR were found within the nucleus of MDA-MB-435 cells in distinct foci in close proximity to Cajal bodies. In agreement with our findings, the nuclear localization of fluorophore-labeled 2′OMeRNA (48) and peptide nucleic acid oligonucleotide telomerase inhibitors, has been reported previously in cancer and immortalized cells (49, 50). Furthermore, unassigned intranuclear foci were observed in a subset of cells (48, 49). This study shows, for the first time, that oligonucleotide telomerase inhibitors accumulate in close association with nuclear Cajal bodies. FISH staining for the RNA component of telomerase, performed postfixation, revealed a cell-cycle–dependent peripheral association between hTR and Cajal bodies (51, 52). This cell-cycle dependency likely explains the presence of oligonucleotide foci in only a proportion, about 30%, of cells in this study. From the perspective of the radiolabeled hTR oligonucleotides, the reported intranuclear accumulation is suggestive of specific target engagement, while also having the potential to localize the constructs in proximity to DNA and so maximize the effects of short-range Auger electron molecular radiotherapy.
Taken together, these findings demonstrate that the enhanced retention and nuclear localization of radiolabeled oligonucleotides targeted to hTR results in a telomerase-dependent cytotoxic induction of DNA double-strand breaks by 111In. In addition, the role of telomerase in the DNA damage response and genomic stability offers the opportunity for simultaneous therapeutic exploitation of the reported interaction between telomerase inhibition and ionizing radiation. It is also possible that perturbation of hTR function by 111In-Match results in inhibition of DNA-PKcs, a phenomenon that has been reported by others (10). DNA-PK plays a central role in nonhomologous end joining, the major pathway for repair of ionizing radiation–induced DNA double-strand breaks in human cells. Therefore, through its inhibition of hTR and, perhaps DNA-PKcs, 111In-Match may be impeding the repair of DNA lesions that it is itself causing.
This work supplements the emerging potential of Auger electron-emitting oligonucleotides as molecular radiotherapeutic agents (45, 53, 54). The intrinsic characteristics of oligonucleotide biotherapeutics, including high affinity and specificity of binding and intermediate molecular weight, offer therapeutic advantages over existing approaches. However, a recognized challenge of therapeutic oligonucleotides is that of effective drug delivery (55) and our current studies are aimed at developing nanoparticulate solutions for cancer cell–specific delivery of telomerase-targeted radio-oligonucleotides.
In conclusion, we report for the first time the development of an hTR subunit–targeting nucleic acid radiopharmaceutical for the selective treatment of cancer. The positive results of 111In-Match in telomerase inhibition and clonogenic survival studies provide a strong incentive for further development of telomerase-targeted molecular radiotherapy.
Disclosure of Potential Conflicts of Interest
K.A. Vallis is a consultant/advisory board member for Blue Earth Diagnostics Ltd. No potential conflicts of interest were disclosed by the other authors.
Conception and design: M.R. Jackson, B.M. Bavelaar, A.H. El-Sagheer, T. Brown, K.A. Vallis
Development of methodology: M.R. Jackson, B.M. Bavelaar, P.A. Waghorn, A.H. El-Sagheer, T. Brown, M. Tarsounas, K.A. Vallis
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.R. Jackson, B.M. Bavelaar, P.A. Waghorn, M.R. Gill, A.H. El-Sagheer
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.R. Jackson, B.M. Bavelaar, A.H. El-Sagheer, K.A. Vallis
Writing, review, and/or revision of the manuscript: M.R. Jackson, B.M. Bavelaar, M.R. Gill, A.H. El-Sagheer, T. Brown, K.A. Vallis
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B.M. Bavelaar
Study supervision: K.A. Vallis
This work was supported by Cancer Research-UK (C5255/A15935); the Medical Research Council (MC_PC_12004); the Engineering and Physical Sciences Research Council (EPSRC) Oxford Centre for Drug Delivery Devices (EP/L024012/1); and the CR-UK/EPSRC Cancer Imaging Centre Oxford (C5255/A16466).
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.