Ku80-Targeted pH-Sensitive Peptide–PNA Conjugates Are Tumor Selective and Sensitize Cancer Cells to Ionizing Radiation

As most current cancer therapies are toxic to healthy tissue, they often result in undesirable side effects that hinder their clinical use. For example, the dosing of radiotherapy, which is estimated by the American Society of Radiation Oncology to treat approximately two thirds of all patients with cancer, is limited by toxicities including inflammation, vascular damage, and fibrosis. The development of radiation sensitizers that specifically target cancer cells while sparing healthy tissue could be used to enhance the efficacy of this widely used treatment without increasing its toxicity, thus providing an important advance in the field.

Because DNA repair is crucial for tumor cell survival after exposure to DNA-damaging agents, such as ionizing radiation (IR), it represents a promising target for cancer therapy. Furthermore, upregulation of DNA repair occurs after exposure to DNA-damaging agents, and so may underlie resistance to these therapies (1–3). Thus, inhibiting DNA repair renders cells more sensitive to IR, reducing cell survival after IR exposure (radiosensitization). As radiation-induced DNA double-strand breaks (DSB) are lethal threats to genomic integrity, repair of DSBs by homology-directed repair (HDR) or nonhomologous end-joining (NHEJ) is particularly important for cellular survival. The NHEJ pathway, which is responsible for repairing 90% of DSBs, requires the DNA-dependent protein kinase (DNA-PK) complex, of which KU80 is an indispensible factor (4, 5). In mice, KU80 deficiency is associated with impaired DSB repair and radiosensitization (6, 7). Although KU80 lacks a known enzyme activity and therefore is not considered a classically “druggable” protein, inhibition of KU80 using antisense has been shown to radiosensitize cells in culture (8, 9). However, this approach has not yet been tested in vivo, likely due to the limitations of traditional antisense delivery methods.

We sought to overcome these limitations in targeting KU80 for use as a radiosensitizer by using peptide nucleic acids (PNA; ref. 10) with (R)-diethylene glycol at the γ position (^MPγPNA; ref. 11) to develop a molecule that potently reduces KU80 expression via an antisense mechanism [αKu80(γ)]. The γ position substitution confers helical preorganization of the PNA oligomer that substantially improves binding affinity to complementary DNAs or RNAs (11). We further improved the efficacy of αKu80(γ) using a tumor-specific delivery system, pH-low insertion peptides (pHLIP), which take advantage of the acidity of the tumor microenvironment to deliver cargo specifically to tumors in vivo. Our group has previously used pHLIPs to deliver PNA antisense agents against an oncogenic miRNA in a mouse lymphoma model, causing substantial antitumor effects without systemic toxicity (12). Here, we utilized similar technologies to develop a tumor-specific radiosensitizer by conjugating an ^MPγPNA antisense targeting KU80 to pHLIP [pHLIP-αKu80(γ)]. In cell culture, pHLIP-αKu80(γ) delivers to cancer cells and suppresses Ku80 expression with pH specificity. When delivered systemically to mouse models in vivo, pHLIP-αKu80(γ) targets to tumors and causes knockdown of KU80 expression in tumors but not in nonmalignant tissues. Furthermore, systemic treatment with pHLIP-αKu80(γ) causes tumor radiosensitization in vivo but does not appear to cause normal tissue toxicity. These results establish pHLIP-αKu80(γ) as a tumor-selective radiosensitizing agent.

Boc-protected PNA monomers, including ^MPγ monomers, were obtained from ASM Research Chemicals. All oligomers were synthesized on a solid-support resin using standard Boc chemistry procedures. To facilitate linkage of pHLIP to PNAs, cysteine was conjugated to the C-terminus of PNAs using a Boc-miniPEG-3 linker (11-Amino-3,6,9-Trioxaundecanoic Acid, DCHA, denoted in the sequences by -ooo-) obtained from Peptide International. Single isomer 5-Carboxytetramethylrhodamine (TAMRA, Thermo Fisher) was also conjugated to the N-terminus of PNAs using a Boc-miniPEG-3 linker. PNA oligomers were cleaved from the resin using a cocktail solution consisting of m-cresol:thioanisole:TFMSA:TFA (1:1:2:6) and were precipitated with ether. PNAs were then purified by high performance liquid chromatography (HPLC) using a C18 column (Waters) and a mobile phase gradient of water and acetonitrile with 0.1% trifluoroacetic acid. Purified PNAs were further characterized using matrix-assisted laser desorption/ionization– time of flight (MALDI–TOF). PNA stock solutions were prepared using nanopure water, and the concentrations were determined on a Nanodrop Spectrophotometer (Thermo Scientific) at 260 nm using the following extinction coefficients: 13,700 M⁻¹ cm⁻¹ (A), 6,600 M⁻¹ cm⁻¹ (C), 11,700 M⁻¹ cm⁻¹ (G), and 8,600 M⁻¹ cm⁻¹ (T).

The following PNA sequences were used (underline indicates ^MPγ monomer):

• αKu80: CCCATGAAGAATCTTCTCTG-ooo-Cys

• αKu80-TAMRA: TAMRA-ooo-CCCATGAAGAATCTTCTCTG- ooo-Cys

• αKu80(γ): CCCATGAAGAATCTTCTCTG-ooo-Cys

• αKu80(γ)-TAMRA: TAMRA-ooo-CCCATGAAGAATCTTCTCTG-ooo-Cys

• scr(γ): ATATCGTTCACGCACGTATC-ooo-Cys

To facilitate conjugation of pHLIP to thiolated PNAs, a cysteine group derivatized with 3-nitro-2-pyridinesulphenyl (NPys) was incorporated to the C-terminus of pHLIP [AAEQNPIYWARYADWLFTTPLLLLDLALLVDADEGT(CNPys)G, New England Peptide]. To generate pHLIP–PNA conjugates, pHLIP-Cys(NPys) and PNAs were reacted overnight at a 1:1 ratio in the dark in a mixture of DMSO:DMF:0.1 mmol/L KH₂PO₄ pH 4.5 (3:1:1). pHLIP–PNA conjugates were then purified and characterized using HPLC and MALDI-TOF, respectively, and concentrations were determined as described above. For all in vitro and in vivo studies, pHLIP–PNAs were heated at 55°C for 5 minutes to prevent aggregation.

pHLIP–PNAs were incubated at a 2:1 ratio with a DNA sequence corresponding to the predicted binding site of PNAs to KU80 mRNA and allowed to anneal at 37°C overnight. Note that 125 mmol/L TCEP was added to a subset of samples to assess for disulfide reduction. Samples were run on a polyacrylamide gel, and SYBR Gold (Life Technologies) was used to visualize KU80 (pHLIP and free PNA are not detected by this stain).

The following DNA sequences were used:

• Human KU80: 5′-AGGTTCAGAGAAGATTCTTCATGGGAAATC-3′

• Mouse Ku80: 5′-AGGTTCATAGAAGATTCTTCATGGGACATC3′

A549 cells were obtained from the ATCC. DLD1-BRCA2KO cells were obtained from Horizon Discovery. EMT6 cells were provided by Dr. Sara Rockwell. Human cells were authenticated using short tandem repeat profiling (ATCC). All cell lines were regularly tested for mycoplasma contamination (MycoAlert, Lonza). Cells were grown in F12K (A549), RPMI (DLD1-BRCA2KO), or DMEM (EMT6) media containing 10% FBS without antibiotics. For pH-specificity experiments, Leibovitz L-15 media were titrated to pH = 6.2 using HCl. Cells were treated with pHLIP–PNAs in pH-titrated media for 24 hours, followed by a series of several washes with PBS. Cells were then immediately processed for flow cytometry or collected 24 to 48 hours later for Western blot analysis.

All studies were approved by the Yale University Institutional Animal Care and Use Committee. For in vivo xenograft studies, 1 × 10⁶ human DLD1-BRCA2KO cells were implanted subcutaneously in 100 μL of media bilaterally in the flanks of athymic nu/nu mice (Harlan). For in vivo mouse tumor studies, 2 × 10⁵ mouse EMT6 or EMT6-GFP cells were implanted subcutaneously in 100 μL of media in the flanks of BALBc/Rw mice. Three times a week, tumors were measured using calipers to determine tumor volume by the following formula: V = 1/2(4π/3)(length/2)(width/2)(height). Mice were treated when tumors reached an average size of 100 mm³, and irradiation (7 or 10 Gy) was applied locally to tumors using a Seimen's Stalipan.

Mice were euthanized, subcutaneous EMT6 tumors and selected organs were dissected, and ex vivo fluorescence imaging was immediately performed on an IVIS Spectrum System (Caliper) using TAMRA filter sets.

Tumors were dissected and dissociated to single-cell suspension using dispase/collagenase followed by trypsinization. Red blood cells were removed by lysis in ammonium chloride solution. A mouse cell depletion kit (Miltenyi Biotech) was used to isolate human tumor cells from DLD1-BRCA2KO tumors. EMT6-GFP tumors were sorted into GFP⁺ and GFP⁻ fractions using a FACSAria cell sorter (BD Biosciences).

Organs collected from treated mice were flash-frozen in liquid nitrogen, homogenized, and lysed with AZ lysis buffer (50 mmol/L Tris, 250 mmol/L NaCl, 1% Igepal, 0.1% SDS, 5 mmol/L EDTA, 10 mmol/L Na₂P₂O₇, and 10 mmol/L NaF) supplemented with Protease Inhibitor Cocktail (Roche) and Phospho-STOP (Sigma). Cells were also lysed in protease and phosphatase inhibitor supplemented AZ lysis buffer. KU80 expression was detected using antibody #611360 (BD). Expression of the DNA damage marker γH2Ax was measured using antibody #9718 (Cell Signaling Technologies, CST). Apoptosis was evaluated using antibodies targeting cleaved caspase-3 (CST, #9664) and cleaved PARP (CST, #9544). GAPDH (Santa Cruz Biotechnology, sc-365062) was used as an endogenous loading control. Where shown, band intensities were quantified using ImageJ64 software.

Cells were treated with siRNA for KU80 (Dharmacon SmartPOOL) at final concentration of 0.001 to 10 nmol/L for 72 hours. A nontargeting scrambled siRNA (Dharmacon SmartPOOL) was used as a control. Extent of knockdown was confirmed using Western blot.

A549 or DLD1-BRCA2KO cells were seeded at a density of 1 × 10⁵ cells per well in 6-well plates. For siRNA experiments, cells were treated with control nontargeting or KU80 siRNA (0.01 or 1 nmol/L) for 72 hours. For pHLIP–PNA experiments, cells were treated with pHLIP-scr(γ) or pHLIP-αKu80(γ) (1 μmol/L) 72 and 24 hours prior to irradiation. Cells were then reseeded at a density of 500 cells per well in a 6-well plate and irradiated with 2 or 4 Gy using an X-RAD 320 X-Ray Biological Irradiator (Precision X-Ray Inc). Nonirradiated controls were handled in parallel but kept outside of the irradiator during treatment. Approximately 1.5 weeks later, cells were permeabilized in 0.9% saline solution and stained with crystal violet in 80% methanol. Colonies were counted manually.

For peripheral blood collection, all mice were anesthetized using open-drop 30% w/v isoflurane in propylene glycol. Blood was collected retro-orbitally using heparinized micro-hematocrit capillary tubes (Fisher Scientific). For complete blood counts, 50 μL blood was evacuated into heparinized coated tubes containing 10 μL 0.5 mol/L EDTA acid, and analysis was performed using a Hemavet 950FS (Drew Scientific) according to the manufacturer's protocol. For blood chemistry analysis, blood from vehicle or pHLIP-αKu80(γ) (5 mg/kg, i.v.) treated mice was collected 72 hours after treatment, serum separated using lithium heparin, and sent to Antech Diagnostics for analysis. For peripheral cytokine analysis, serum was isolated by centrifugation (1,500 rcf for 10 minutes) from blood collected from vehicle or pHLIP-αKu80(γ) (5 mg/kg, i.v.) treated mice 72 hours after treatment. Serum was submitted to the CytoPlex Core Facility at Yale University for luminex-based cytokine detection and quantification using the Bio-Plex Pro Mouse Cytokine 23-Plex (Bio-Rad).

Bone marrow cells were harvested by flushing femurs and tibias from mice 72 hours after treatment with vehicle or pHLIP-αKu80(γ) (5 mg/kg, i.v.). Cells were automatically counted using a Cellometer Auto T4 (Nexcelom). For relative quantification of cell types, cells were stained with fluorescently tagged antibodies to different lineage markers (KIT: #12-1171-82, SCA1: #45-5981-82, TER-119 #11-5921-82, GR1 #11-5931-82, CD45 #11-0452-82; Thermo Scientific), washed with PBS, and analyzed via flow cytometry.

Statistical analysis was performed in GraphPad Prism. Unless otherwise stated, data are presented as mean ± SEM, with n = 3 replicates.

We sought to overcome traditional limitations in targeting KU80 for use as a radiosensitizer by using PNAs, DNA analogs in which the sugar phosphodiester backbone is replaced with homomorphous achiral N-(2 aminoethyl) glycine units (10). PNAs are a particularly useful tool as they hybridize with RNA and DNA with high affinity in a sequence-specific fashion, and their charge-neutral structure confers resistance to both protease and nuclease degradation (13). PNA-DNA and PNA-RNA complexes also have enhanced stability from a reduction of the backbone charge repulsion that destabilizes natural double-stranded complexes. PNAs have been used to decrease gene expression via an antisense mechanism in numerous targets (14–16), and their efficacy can be enhanced by structural modifications and by novel delivery systems (17–19). Using PNA monomers with (R)-diethylene glycol at the γ position (^MPγPNA) to increase solubility and nucleic acid binding (11), we synthesized a ^MPγPNA that we hypothesized would reduce KU80 expression via an antisense mechanism [αKu80(γ)]. We first identified a 20-nucleotide sequence with known antisense activity against KU80 (8). Using solid-phase synthesis of Boc-protected monomers, we prepared both an unmodified PNA and a ^MPγPNA corresponding to this antisense sequence (αKu80 and αKu80(γ), respectively) as well as a scrambled control ^MPγPNA sequence [scr(γ); Fig. 1A]. In some cases, the fluorescent dye TAMRA was conjugated to these PNAs to allow visualization in uptake experiments.

To deliver PNAs in cell culture and in vivo, we utilized a tumor-specific delivery system, pHLIPs, which takes advantage of the acidity at the surfaces of tumor cells. Although extracellular pH is tightly regulated to approximately 7.4 in healthy tissue, tumor cell and tumor macrophage surfaces are largely acidic (20), resulting from high rates of glycolytic lactic acid production, surface carbonic anhydrase expression, and the electrochemical gradient (21). Thus, a delivery system with specificity for acidic conditions at cell surfaces facilitates the selective targeting of cargoes to the tumor microenvironment in vivo. pHLIPs are small peptides that insert their C-terminus across cell membranes at acidic extracellular pH (22, 23). Cargo, including small molecules, PNAs, and dyes, can be linked to the C-terminus of a pHLIP using a disulfide bond that is cleaved in the intracellular reducing environment (ref. 24; Fig. 1B).

Thiol-disulfide interchange reactions were performed to conjugate PNAs to pHLIP via a disulfide linkage [pHLIP-αKu80, pHLIP-αKu80(γ), and pHLIP-scr(γ)]. After purifying and characterizing the pHLIP–PNA conjugates using mass spectrometry and HPLC, we then confirmed the ability of these conjugates to bind the sequence of interest (using a DNA sequence as a proxy for the RNA target) using a gel-shift assay (Fig. 1C). Both pHLIP-αKu80 and pHLIP-αKu80(γ), but not pHLIP-scr(γ), bound the KU80 mRNA sequence as indicated by a shift in migration of the detected band (Fig. 1C), confirming the sequence-specific binding of our pHLIP–PNA conjugates. As expected, pHLIP-αKu80(γ) showed enhanced nucleic acid binding as compared with pHLIP-αKu80, which is indicated by a reduced amount of free, unbound DNA. In some samples, the reducing agent TCEP was added to cleave disulfide linkages between pHLIP and PNAs. The addition of TCEP shifted the size of detected bands, confirming cleavage of conjugates and release of the peptide under reducing conditions. Gel shift assays also confirmed the ability of pHLIP-αKu80(γ) to bind the mouse sequence of Ku80 at the corresponding sequence (Supplementary Fig. S1), which differs from the human sequence by one base pair.

We next confirmed the pH-dependent insertion of pHLIP-αKu80 and pHLIP-αKu80(γ) in cultured cells. To do this, we treated cells with TAMRA-labeled pHLIP–PNA conjugates in acidic (pH = 6.2) or neutral (pH = 7.8) media followed by several washes with PBS. Flow cytometry was used to quantify TAMRA fluorescence. As demonstrated in Fig. 1D and E, these culture conditions led to robust pH-dependent delivery of pHLIP-αKu80 and pHLIP-αKu80(γ).

Using the culture conditions described above, we also confirmed pH-dependent antisense activity of pHLIP-αKu80 conjugates in the human lung adenocarcinoma cancer cell line A549 (Fig. 1F and G). Specifically, after treatment with pHLIP-αKu80(γ) at pH = 6.2, the expression of Ku80 is knocked down to 45±14% of controls in A549 cells. Treatment with pHLIP-αKu80 caused a more modest reduction in KU80 expression (Fig. 1F and G), consistent with its reduced nucleic acid binding. Importantly, KU80 expression is unaffected in cells treated with pHLIP–PNA at pH = 7.8, demonstrating that pHLIP-αKu80(γ) exhibits pH-specific activity against KU80.

To inform timing and dosing for subsequent radiosensitization and synthetic lethality experiments, we established the pharmacokinetic and pharmacodynamic properties of systemic pHLIP-αKu80(γ) treatment in mice bearing EMT6 mouse breast cancer tumors. We used the fluorescently labeled pHLIP-αKu80(γ)-TAMRA conjugate to visualize tumor and organ distribution after i.v. administration (via tail vein injection) using IVIS Spectrum. Intratumoral TAMRA fluorescence was detected and quantified at both 24 and 72 hours after a 5 mg/kg i.v. injection of pHLIP-αKu80(γ)-TAMRA (Fig. 2A and B). As expected, pHLIP-αKu80(γ)-TAMRA was also observed in the kidney (Fig. 2A), likely due to excretion of pHLIP by the renal system as well as uptake of pHLIP-αKu80(γ)-TAMRA by acidic regions in the kidney. We also observed a lower level of fluorescence in the liver (Fig. 2A and B). These results confirm prior work showing relative tumor specificity of pHLIP delivery (22) and also establish successful in vivo delivery of ^MPγPNAs by pHLIP.

To assess the pharmacodynamic properties of pHLIP-αKu80(γ), we treated DLD1-BRCA2KO human colon cancer xenograft tumor-bearing mice with pHLIP-αKu80(γ) (5 mg/kg, i.v.), and collected tumors and organs for Western blot analysis. In order to better measure the suppression of KU80 expression within tumor cells, we isolated DLD1-BRCA2KO tumor cells from stromal cells by immune depletion of mouse cells. At 48 hours after treatment, we observed an approximately 40% decrease in KU80 expression in isolated DLD1-BRCA2KO tumor cells (Fig. 2C and D). We then confirmed these results in EMT6 tumors, using EMT6 cells expressing GFP (EMT6-GFP) in order to isolate tumor cells from stromal and immune cells on the basis of GFP expression using fluorescence-activated cell sorting (Supplementary Fig. S2). Similar to the DLD1-BRCA2KO tumor cells, in EMT6-GFP cells, systemic treatment with pHLIP-αKu80(γ) caused a partial suppression of KU80 expression (Fig. 2C and D). To determine whether this degree of KU80 knockdown was sufficient to induce radiosensitization, we titrated siRNA-mediated knockdown of KU80 in tumor cells in culture and used a clonogenic survival assay to assess radiosensitization. siRNA doses corresponding to an approximately 40% knockdown were determined to increase cellular sensitivity to irradiation (Supplementary Fig. S3), indicating that the extent of KU80 knockdown induced by pHLIP-αKu80(γ) in vivo would be sufficient to mediate radiosensitization. Furthermore, as compared with controls, tumors treated with pHLIP-αKu80(γ) also had a trend toward higher expression of γH2Ax, a marker of DNA damage, providing further evidence of the functional effects of pHLIP-αKu80(γ) on DNA repair (Fig. 2E and F).

We also assessed for the effects of pHLIP-αKu80(γ) treatment on KU80 expression in nonmalignant tissue. Importantly, we did not observe any change in KU80 expression in any of the organs assessed at this time point, including liver, kidney, heart lung, and spleen (Fig. 2G and H). Together, these results indicate that pHLIP-αKu80(γ) causes a selective suppression of KU80 expression within tumors.

We first assessed for potential tumor radiosensitizing effects of pHLIP-αKu80(γ) in subcutaneous mouse tumor xenografts of the mouse EMT6 breast cancer cell line. In this immune-competent model of breast cancer, pHLIP-αKu80(γ) treatment alone did not affect tumor growth or mouse survival (Fig. 3A and B). However, when administered with local tumor irradiation (1 × 15 Gy), pHLIP-αKu80(γ) significantly delayed tumor growth and prolonged survival (Fig. 3A and B). Other treatment protocols combining pHLIP-αKu80(γ) with local tumor irradiation (2 or 3 i.v. administrations of vehicle alone or pHLIP-αKu80(γ), combined with 1 × 10 Gy or 2 × 7 Gy, had similar effects on tumor growth and mouse survival in the EMT6 tumor model (Supplementary Fig. S4).

We then tested for radiosensitization by pHLIP-αKu80(γ) in tumor xenografts using the human DLD1-BRCA2KO colon cancer cell line in immune-deficient athymic nude mice. These cells are BRCA2-deficient, conferring deficits in the HDR pathway of DNA DSB repair. There is some evidence that HDR deficiency may confer increased susceptibility to NHEJ inhibition (25), and so we first assessed for effects of pHLIP-αKu80(γ) monotherapy on DLD1-BRCA2KO tumor growth and survival (as defined by tumor volume 4X that of initial treatment). We found that pHLIP-αKu80(γ) treatment did not alter tumor growth or survival as compared with vehicle-treated controls (Fig. 3C and D). However, when pHLIP-αKu80(γ) was administered in combination with local tumor irradiation (7Gy), it caused a significant tumor growth delay compared with irradiation alone along with a trend (P = 0.07) toward prolonged survival (Fig. 3C and D), similar to the results observed using the EMT6 model. Importantly, pHLIP-scr(γ) did not affect tumor growth or survival when administered alone or in combination with local tumor irradiation (Fig. 3C and D). Together, these results indicate that systemic administration of pHLIP-αKu80(γ) treatment induces radiosensitization in tumors.

We also tested for the effects of in vitro pHLIP-αKu80(γ) treatment on cell survival using clonogenic survival assays in DLD1-BRCA2KO and A549 cells treated with pHLIP-conjugated scr(γ) or αKu80(γ) PNAs under neutral and acidic conditions, with or without irradiation. In DLD1-BRCA2KO cells treated under acidic conditions, pHLIP-αKu80(γ) caused significant radiosensitization as compared with pHLIP-scr(γ) treatment (Supplementary Fig. S5). No radiosensitization was observed in cells treated under neutral conditions (Supplementary Fig. S5). Similar results were observed in A549 cells, with pHLIP-αKu80(γ) treatment causing a trend (P = 0.06) toward radiosensitization under acidic but not neutral conditions (Supplementary Fig. S5).

To better understand the mechanism underlying pHLIP-αKu80(γ)–induced radiosensitization, we analyzed the expression of DNA damage and apoptosis markers in control and pHLIP-αKu80(γ)–treated tumors with or without local tumor irradiation (15 Gy). Tumors treated with pHLIP-αKu80(γ) plus irradiation showed a trend toward increased expression of γH2Ax, a marker of DNA damage, as well as cleaved PARP and cleaved caspase 3, markers of apoptosis (Fig. 3E and F).

After administration of pHLIP-αKu80(γ), no gross toxicity was noted in treated mice, including weight loss (Fig. 4A), skin reactions, or behavioral changes. Furthermore, an evaluation of blood chemistries did not reveal any electrolytes abnormalities (Supplementary Fig. S6) or any signs of renal or hepatic damage (Fig. 4B).

As many cancer therapeutics display bone marrow toxicity, we also evaluated the effects of pHLIP-αKu80(γ) on bone marrow. Analysis of mature blood cells in the periphery did not indicate any bone marrow toxicity, as the number of red blood cells, white blood cells, and platelets were comparable between control and pHLIP-αKu80(γ)–treated mice (Fig. 4C). Furthermore, treatment with pHLIP-αKu80(γ) did not significantly change bone marrow cellularity (Fig. 4D) or the relative percentages of different bone marrow cell types as assessed by flow cytometry (Supplementary Fig. S7).

We also evaluated the immunogenicity of pHLIP-αKu80(γ). To do this, we tested the serum levels of 23 different cytokines in the serum of mice treated with pHLIP-αKu80(γ). pHLIP-αKu80(γ) treatment did not alter the expression of peripheral cytokines (Fig. 4E), consistent with a lack of an immune response. Together, these results do not indicate any normal tissue toxicity with pHLIP-αKu80(γ) treatment.

In this report, we utilize new technologies to target a clinically important pathway for which therapeutic targeting has been historically challenging. Although several small-molecule inhibitors targeting the NHEJ pathway have been identified (26, 27), the clinical potential of these molecules is limited by their lack of specificity and in vivo toxicity (28, 29). In addition, although antisense suppression and siRNA-mediated inhibition of NHEJ factor expression have been demonstrated in cell culture (8, 9, 30–34), they have not been reported in vivo [with the exception of Li and colleagues, in which intratumoral injections of an adenovirus vector, requiring tumoral heat shock (42°C), achieved in vivo radiosensitization]. Here, we report the in vivo use of antisense-based technology via systemic (i.v.) administration to target the nonenzymatic NHEJ factor KU80, which, lacking a known enzyme activity, is not a classically “druggable” target.

Furthermore, we confer tumor selectivity on the antisense PNA molecule using a pH-based delivery system, pHLIP. We demonstrate that in culture, pHLIP-αKu80(γ) delivers the PNA to cancer cells and suppresses KU80 expression with pH specificity. Furthermore, in vivo, pHLIP-αKu80(γ) targets to tumors and causes selective suppression of KU80 expression in tumors. This study builds upon prior work identifying pHLIPs as a potential delivery tool for anticancer therapeutics (12, 35–39) and represents the first delivery of ^MPγPNA using pHLIPs as well as the first usage of pHLIP–PNA technology to reduce protein expression via an antisense mechanism.

Importantly, we show that pHLIP-αKu80(γ) increases the sensitivity of tumors to radiotherapy, causing a substantial delay in tumor growth and prolonging host survival when administered in combination with local tumor irradiation. Interestingly, although prior work has indicated that inhibition of NHEJ may be toxic in the setting of BRCA deficiency (25), we do not find pHLIP-αKu80(γ) to be effective as a monotherapy in DLD1-BRCA2KO tumor xenografts. This discrepancy may be explained by the fact that pHLIP-αKu80(γ) only induces a partial, and not complete, suppression of KU80 expression, and therefore treated tumors likely retain partial NHEJ capacity. Complete suppression of NHEJ may be required for the previously identified synthetic lethal interactions.

Finally, we provide evidence that pHLIP-αKu80(γ) is safe and nontoxic when administered in vivo, as assessed by several end-points. These indications of safety are in line with previous work that has established PNAs and pHLIPs as potential therapeutic agents with favorable side effect profiles (12, 19, 40). We expect relatively low off-target effects from pHLIP-αKu80(γ) treatment, as uptake of the pHLIP–PNA is primarily limited to tumor cells, and the αKu80(γ) PNA has 100% complementarity to the KU80 mRNA only (based on nucleotide BLAST search). Although PNA–nucleotide binding is sequence specific, and mismatches in PNA sequences have been shown to significantly reduce nucleotide binding (41), it is possible that within cancer cells, the αKu80(γ) PNA may cause off-target effects by binding to sequences of partial complementarity (i.e., PCMTD2, KCNIP4).

Although the suppression of KU80 expression and radiosensitization induced by pHLIP-αKu80(γ) treatment is modest, efficacy could be improved through a number of methods. A more detailed evaluation of dosing and timing protocols could optimize the treatment regimen. An alternative approach could be simultaneous delivery of multiple antisense PNAs targeting KU80. In additional, serine-modified PNAs (^SγPNAs) could be used in place of ^MPγPNAs. Similar to ^MPγPNAs, ^SγPNAs also preorganize into a helical structure and demonstrate increased solubility and nucleic acid binding as compared with unmodified PNAs (42). Furthermore, synthesis of ^SγPNA monomers is simpler than ^MPγPNA monomers. Efforts could also be made to conjugate small-molecule inhibitors of DNA-PK to pHLIPs. However, as these small molecules do not have available thiol groups, this approach would require the development of additional pHLIP-cargo linkage techniques to ensure that the small molecules could be inserted across the cell membrane and cleaved from pHLIP while still retaining activity against DNA-PK.

Together, these results establish pHLIP-αKu80(γ) as a tumor-specific radiosensitizer. The development of compounds like pHLIP-αKu80(γ) may eventually help lead to novel treatment regimens for cancer therapy that have both high specificity and low toxicity.

D.M. Engelman is Scientific Advisor at, and has an ownership interest (including patents) in, pHLIP, Inc. P.M. Glazer is Consultant at, and has an ownership interest (including patents) in, Cybrexa Therapeutics. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A.R. Kaplan, R. Bahal, P.M. Glazer

Development of methodology: A.R. Kaplan, H. Pham, R. Bahal, P.M. Glazer

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.R. Kaplan, H. Pham, Y. Liu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.R. Kaplan, H. Pham, Y. Liu, P.M. Glazer

Writing, review, and/or revision of the manuscript: A.R. Kaplan, H. Pham, R. Bahal, D.M. Engelman, P.M. Glazer

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): H. Pham

Study supervision: P.M. Glazer

Other (provided reagents used to in vivo experiments): S. Oyaghire

This work was supported by grants from the NIH: R35CA197574 to P.M. Glazer, R01GM073857 to D.M. Engelman, and F30CA221065 to A.R. Kaplan. Support for this research was also provided by NIH Medical Scientist Training Program Training Grant T32GM007205. We thank D. Hegan, A. Gupta, and A. Dhawan for assistance.

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.