Targeting Translation Termination Machinery with Antisense Oligonucleotides for Diseases Caused by Nonsense Mutations

ASOs used in this study were chemically modified with phosphorothioate in the backbone and 2′-methoxyethyl (MOE) modification in the wings with a central 10-nucleotide deoxy gap (5-10-5 gapmer). Oligonucleotides were synthesized using an Applied Biosystems 380B automated DNA synthesizer (PerkinElmer Life and Analytical Sciences) and purified as previously described [27,29]. Lyophilized ASOs were dissolved in sterile Dulbecco's phosphate-buffered saline (DPBS; without calcium or magnesium) and quantified by ultraviolet spectrometry, diluted to the desired concentration, sterilized through a 0.2 μm filter.

All the animals of wild-type (C57BL/6; purchased from JAX) and hemophilia mice [30] (licensed from The Children's Hospital of Philadelphia and maintained in Taconic) were housed under standard condition in a pathogen-free mouse facility. All animal procedures were performed in accordance with National Institution of Health guidelines and approved by the Institutional Animal Care and Use Committee at Ionis Pharmaceuticals. ASOs were administrated subcutaneously at a volume of 10 μL/g. Geneticin (G418; 5.6 mg/mL) was administrated subcutaneously at a volume 5 μL/g.

Blood samples were collected by cardiac puncture at time of sacrifice. Plasma chemistry values were measured on the AU480 Clinical Chemistry Analyzer (Beckman Coulter).

Animal tissues were homogenized in guanidine isothiocyanate solution (Invitrogen) supplemented with 8% 2-mercaptoethanol (Sigma-Aldrich). Total RNA was prepared using the RNeasy Mini Kit (Qiagen). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using Express One-Step Superscript qRT-PCR kits on an ABI StepOne Plus Real-Time PCR system. Taqman primer probe sets: eRF1 forward ATTTCCAGGGAATGGAGTACCA, reverse GGGTATGCTCCTTGGGTTGA, probe CGGCAACCGTGCCTCACCCT; eRF3a forward CGGAACCTGTAGAGTCCTCTCAA, reverse CATTTCTGTCTCTCCATTTTCTACAACA, probe TCGTGTGAAGGTTCAAA; hFIX forward TGAGGAAGAATTCAACAGTGTGTCT, reverse CCTCTGGTCTAGGCAACTTCAAC, probe CAGCAGTGTTCAGAGCCAAGCAAGA; C/EBP homologous protein (CHOP) forward TGAGCCTAACACGTCGATTATATCA, reverse TCTGGAACACTCTCTCCTCAGGTT, probe CAGCGACAGAGCCAGAATAACAGCCG; glyceradehyde 3-phosphate dehydrogenase (Gapdh) forward GGCAAATTCAACGGCACAGT, reverse GGGTCTCGCTCCTGGAAGAT, probe AAGGCCGAGAATGGGAAGCTTGTCATC.

Protein levels were measured using western blot. Animal tissues were homogenized in RIPA buffer (Thermo Fisher Scientific) containing Halt Protease Inhibitor Cocktail (Life Technologies). Protein concentrations were determined using the BioRad DC protein assay, and protein was loaded at 40 μg onto a 4%–15% Criterion™ TGX™ Precast Midi Protein Gel (BioRad). Western blot membranes were probed with primary anti-eRF1 antibody (ab31799; Abcam) or anti-eRF3a antibody (ab49878; Abcam). An antibody against β-actin (A5316; Sigma) was used as a loading control. Membranes were then incubated with IRDye secondary antibodies (Li-COR) and scanned using an Odyssey infrared system (Li-COR). Images were quantified using Image Studio (Li-COR).

hFIX protein level in mouse plasma was measured by enzyme-linked immunosorbent assay (ELISA) using Human Factor IX (FIX) ELISA Kit (ab188393; Abcam) following manufacture instructions.

FIX activity assay was performed at UCSD Murine Hematology and Coagulation Core Laboratory. In brief, clotting times were determined in duplicate with an ST4 semi-automated mechanical coagulation instrument (Diagnostica Stago). Thirty microliters of citrated sample plasma diluted 1/10 in HN/BSA buffer was incubated with 30 μL of APTT reagent and 30 μL of human citrated plasma deficient of FIX at 37°C for 5 min, followed by the addition of 30 μL of 25 mM 37°C CaCl₂ to initiate clotting. Time until clot formation was measured and interpolated on a standard curve of serial dilutions citrated normal (BL/6 pool) mouse plasma tested as described. Final result was reported in % hFIX-WT mice.

We first designed and tested ∼350 ASOs targeting mouse eRF1 and eRF3a mRNAs in mouse cell lines (data not shown). The top 15 ASOs that achieved the best target reduction for each of the translation termination factors were then evaluated for in vivo tolerability at a single-dose level in mice (data not shown). The best tolerated ASOs were selected for subsequent experiments. To determine the threshold of translation termination factor reduction that is well tolerated in animals, we performed ASO dose–response experiments in C57Bl/6J mice. ASOs targeting mouse eRF1 or eRF3a were dosed twice a week for 4 weeks at the indicated doses. Weekly body weights and also terminal organ weights and blood chemistry were measured to determine if the ASO treatments were tolerated. As liver is the most sensitive organ for ASO-mediated target knockdown with second-generation ASO chemistry, liver RNA was extracted and measured for target reduction and plasma levels of the liver enzymes; alanine transaminase (ALT) and aspartate transaminase (AST) were measured to monitor liver injury. Our results demonstrated that reduction of eRF1 to ∼40% of normal levels in mouse liver with 200 mg/kg/week ASO treatment is well tolerated. ASO-treated mice showed normal body weight gain, organ weights, and liver transaminase levels compared with the DPBS-treated control group (Fig. 1A–D). It is possible that reduction of eRF1 to levels <40% of normal levels could also be well tolerated; however, we could not achieve further target reduction with ASOs targeting eRF1 (data not shown). Moreover, we observed that reduction of eRF3a to ∼28% of normal levels in mouse liver with 25 mg/kg/week ASO treatment is well tolerated (Fig. 1E–H). Animals exhibited normal body weight gain, organ weights, and liver transaminase levels compared with the DPBS-treated control group (Fig. 1E–H). However, when eRF3a was depleted to ∼22% or 16% of control levels with either 50 or 100 mg/kg/week ASO treatments, respectively, the mice showed less body weight gain and elevated plasma transaminase levels (Fig. 1E–H), indicating that reducing eRF3a to <28% of the normal levels in mouse liver is not well-tolerated. These results were confirmed using two ASOs for each translation termination factor (Fig. 1; data not shown).

It has been reported that siRNA-mediated depletion of eRF3a also reduces the level of eRF1 protein by affecting its stability [25]; therefore, we postulated that eRF3a could be a more effective target to promote translational read-through. To evaluate this hypothesis, we tested ASO-mediated reduction of eRF3a in a hemophilia B mouse model carrying a nonsense mutation. Hemophilia B is an X-linked bleeding disorder that results from a defect in the gene encoding coagulation FIX, a serine protease that is critical for blood clotting [31]. Patients with severe hemophilia B have functional FIX levels that are <1% of normal values and have frequent bleeding episodes that are associated with crippling arthropathy and early death [32]. Increasing circulating FIX to 1% of normal levels can substantially ameliorate the bleeding phenotype [33]. In this hemophilia B mouse model, a human FIX minigene, either wild-type (hFIX-WT as control) or with a nonsense mutation R338X (hFIX-R338X), is expressed in mice that lack the endogenous mouse FIX gene [30]. hFIX-WT mice contain two copies of the transgene and express hFIX mRNA and protein at similar levels to that of FIX expression in human liver and plasma, respectively. hFIX-R338X mice contain six copies of the transgene and express hFIX mRNA approximately four-fold higher than human FIX mRNA level in liver. Truncated protein expressed from the hFIX-R338X transgene is detected in mouse liver but not in plasma. These mice effectively recapitulate the phenotype of patients carrying the R338X mutation (CGA to TGA) demonstrating severe hemophilia with no detectable circulating FIX protein [30]. The hFIX transgenes are driven by the human transthyretin promoter and thus are primarily expressed in mouse hepatocytes, which are the main cell type for endogenous FIX production [30]. Therefore, it is an optimal model for demonstrating proof-of-concept for ASO-mediated translational read-through approach.

As the hFIX-R338X transgene is primarily expressed in hepatocytes, we evaluated whether we could further deplete eRF3a in hepatocytes using a triantennary N-acetyl galactosamine (GalNAc)-conjugated eRF3a ASO to improve the read-through efficacy while reducing any unwanted toxic effects in other cell types. GalNAc is a high-affinity ligand for the hepatocyte-specific asialoglycoprotein receptor (ASGPR) [34]. GalNAc conjugation results in preferential ASO delivery to hepatocytes relative to extrahepatic cells, resulting in 10-fold higher ASO potency in mouse liver than unconjugated ASOs of the same sequence [35]. This enables lower doses of ASO to be administered to reduce hepatocyte targets while largely sparing target reduction in extrahepatic tissues.

Dose–response experiments were performed with GalNAc-conjugated eRF3a ASO to determine the optimal level of target reduction to achieve a balance between efficacy and safety in hemophilia mice. Male hemophilia mice between ages 6 and 12 weeks were dosed once a week with ASOs at 0.56, 1.67, 3.3 and 5 mg/kg/week for 4 weeks. eRF3a-GalNAc-ASO treatments led to dose-dependent liver target reduction in hemophilia mice, reducing eRF3a mRNA levels to ∼97%, 65%, 35%, and 19% with respect to the levels of DPBS-treated mice (Fig. 2A). Western blot analysis indicated similar eRF3a target reduction at the protein level (Fig. 2B). In contrast to data previously reported in HEK293 cells whereby siRNA-mediated depletion of eRF3a also significantly reduced eRF1 protein levels [25], reduction of eRF3a in hemophilia mouse livers only caused a minor reduction in eRF1 protein levels (Fig. 2B). Of interest, we observed an upregulation of eRF1 mRNA levels in mouse livers treated with eRF3a ASO (Fig. 2C), indicating that a feedback regulation might exist to increase eRF1 expression in compensation for the loss of eRF3a expression.

Without affecting hFIX-R338X mRNA levels (Fig. 2D), 3.3 and 5 mg/kg/week eRF3a-GalNAc-ASO treatments led to the detection of full-length hFIX protein in mouse plasma, indicating that reduction of eRF3a to <35% of normal levels induces translational read-through of hFIX-R338X mRNAs (Fig. 2E). However, in contrast to other treatment groups, mice that were treated with eRF3a-GalNAc-ASO at 5 mg/kg/week exhibited less body weight gain during the study and slightly elevated liver transaminase levels (data not shown), confirming that reducing eRF3a RNA level to <28% of normal levels in liver is not well tolerated in mice (Fig. 1E–H) even with GalNAc-conjugated ASO that restricts target reduction in the liver primarily to hepatocytes [35]. In line with these observations, an upregulation of CHOP mRNA was also observed in mice treated with 5 mg/kg/week eRF3a-GalNAc-ASO (Fig. 2F), indicating induction of the unfolded protein response (UPR). This effect is likely because of the impairment of global translation accuracy, which is in agreement with a previous report that high concentrations of aminoglycoside treatment cause endoplasmic reticulum stress and activation of the UPR [36].

Therefore, we conclude that a minimum of ∼28% eRF3a is needed to avoid effects on global translation. Although eRF3a-GalNAc-ASO treatment at a well-tolerated dose significantly increased the production of otherwise undetectable full-length hFIX protein, it did not reach the level required (1% of normal levels) to improve the hemophilia disease phenotype [33].

As reduction of eRF3a did not significantly reduce eRF1 protein levels as originally predicted, we evaluated the effects of ASO-mediated reduction of eRF1 on promoting translational read-through in hFIX-R338X hemophilia mice, because it is the translation termination factor that directly recognizes and binds to the stop codon. To deplete eRF1 more efficiently in hepatocytes, we generated a GalNAc-conjugated eRF1 ASO. Male hemophilia mice between ages 6 and 12 weeks were dosed once a week with ASOs at 2.2, 6.6 and 20 mg/kg/week for 4 weeks. eRF1-GalNAc-ASO treatment led to dose-dependent liver target reduction in hemophilia mice and reduced eRF1 mRNA levels to ∼80%, 63%, and 46% of DPBS-treated mice (Fig. 3A). Western blot analysis indicated similar eRF1 target reduction at the protein level (Fig. 3B). All treatment groups demonstrated normal body weight gain, and terminal organ weights and blood chemistry, indicating that all treatments were well tolerated (data not shown). Reduction of eRF1 to these levels did not affect eRF3a expression at the mRNA or protein level (Fig. 3B, C).

Although reduction of eRF1 also did not affect hFIX-R338X mRNA levels (Fig. 3D), full-length hFIX protein was detected in plasma of mice treated with eRF1-GalNAc-ASO at 6.6 and 20 mg/kg/week, indicating that reduction of eRF1 to 63% of normal levels induces translational read-through of hFIX-R338X mRNA (Fig. 3E). Based on these results, it seems that, in contrast to eRF3a, eRF1 is a haploinsufficient gene in its function in translation termination at PTC. Of interest, we did not detect upregulation of CHOP mRNA in any of the treatment groups in this study (Fig. 3F), indicating a lack of UPR induction. These results suggest that ASO-mediated reduction of eRF1 to 46–63% of normal levels promotes translational read-through at PTC without significantly affecting global translation accuracy.

Therefore, eRF1 seems to be a safer and preferable target for promoting translational read-through at PTCs in comparison with eRF3a. Despite this encouraging data, however, the highest average level of full-length hFIX protein detected in eRF1-GalNAc-ASO-treated mice was only ∼0.24% of hFIX-WT mice (Fig. 3E), which would not result in meaningful clinical outcomes for hemophilia disease.

Because GalNAc-ASO-mediated reduction of eRF1 or eRF3a demonstrated limited efficacy in promoting translational read-through and seems less likely to be effective as a single therapy, we wanted to test if ASO-mediated reduction of eRF1 or eRF3a will improve the efficacy of the aminoglycoside geneticin (G418), which was previously shown to promote translational read-through in hFIX-R338X mice [37]. We treated hemophilia mice with the highest tolerable dose of either eRF1-GalNAc-ASO (20 mg/kg/week) or eRF3a-GalNAc-ASO (3.3 mg/kg/week) in combination with G418. Male hFIX-R338X mice between ages 6 and 12 weeks were dosed once a week with either eRF1- or eRF3a-GalNAc-ASO for 4 weeks. Seven days before the end of study, ASO-treated cohorts were divided and half were administered 28 mg/kg G418 daily for 7 days [37]. A separate cohort of mice were treated with G418 alone daily for the last 7 days of the study. Animals were sacrificed 48 h after the last ASO treatment and 9 h after the last dose of G418. All groups of animals demonstrated normal body weight gain throughout the study, as well as normal organ weights and plasma transaminase levels as measured at the time of necropsy (data not shown).

In this study, both eRF1-GalNAc-ASO and eRF3a-GalNAc-ASO demonstrated similar efficacy in promoting translational read-through when administered alone without G418 compared with previous studies (hFIX protein levels in the plasma of eRF1- and eRF3a-GalNAc-ASO treated mice were 0.25% and 0.29% of that of hFIX-WT mice, respectively) (Fig. 4A–C). Of interest, when combined with G418 treatment, both eRF1- and eRF3a-GalNAc-ASO demonstrated synergistic effects with G418 in promoting translational read-through. eRF3a-GalNAc-ASO treatment increased G418 activity by ∼three-fold, increasing plasma hFIX protein levels from ∼4.9% to ∼14.8% compared with untreated hFIX-WT mice (Fig. 4C). eRF1-GalNAc-ASO improved G418 efficacy even further by more than six-fold, reaching ∼32% of untreated hFIX-WT levels (Fig. 4C). Furthermore, both eRF1-ASO/G418 and eRF3a-ASO/G418 treatments significantly improved FIX activity in an activated partial thrombo-plastin time-based coagulation assay (APTT) compared with G418 alone, with eRF1-ASO/G418 demonstrating a trend of greater improvement (Fig. 4D). We also observed an approximately two-fold increase in hFIX-R338X mRNA levels in eRF1-GalNAc-ASO-treated mouse liver, which we did not observe in the previous dose–response study where slightly less eRF1 target reduction was achieved (Fig. 4E). As a measurement of the UPR, CHOP mRNA level did not significantly change in any treatment groups except in the eRF3a-GalNAc-ASO/G418-treated group (Fig. 4F), probably because of additive effects of eRF3a-GalNAc-ASO and G418 on global translation accuracy.

Overall, in hemophilia mice, GalNAc-conjugated ASOs targeting eRF1 and eRF3a worked in synergy with the small molecular read-through drug geneticin, significantly increasing production of full-length hFIX protein, which results in improved plasma FIX activity. GalNAc-ASO-mediated reduction of eRF1 seems to be safer and more effective in promoting the translational read-through of hFIX-R338X mRNA than reduction of eRF3a when combined with G418.