Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Antisense oligonucleotides containing locked nucleic acid improve potency but cause significant hepatotoxicity in animals

Antisense oligonucleotides containing locked nucleic acid improve potency but cause significant... Published online 19 December 2006 Nucleic Acids Research, 2007, Vol. 35, No. 2 687–700 doi:10.1093/nar/gkl1071 Antisense oligonucleotides containing locked nucleic acid improve potency but cause significant hepatotoxicity in animals Eric E. Swayze*, Andrew M. Siwkowski, Edward V. Wancewicz, Michael T. Migawa, Tadeusz K. Wyrzykiewicz, Gene Hung, Brett P. Monia and and C. Frank Bennett Isis Pharmaceuticals, Inc., 1896 Rutherford Road, Carlsbad, CA 92008,USA Received August 18, 2006; Revised October 27, 2006; Accepted November 9, 2006 include recruitment of RNase H, which cleaves the RNA ABSTRACT strand of a DNA:RNA duplexes, and activation of the RNA A series of antisense oligonucleotides (ASOs) interference (RNAi) pathway utilizing short RNA duplexes containing either 2 -O-methoxyethylribose (MOE) or (siRNA) or hairpins (shRNA). In addition to the use of locked nucleic acid (LNA) modifications were de- oligonucleotides as research reagents there is increasing inter- signed to investigate whether LNA antisense oligo- est in exploiting oligonucleotides as therapeutic agents. Cur- nucleotides (ASOs) have the potential to improve rently there is one approved antisense product (Vitravene, upon MOE based ASO therapeutics. Some, but not fomivirsen) (1) and >30 products in active development. The first generation of ASO therapeutics were 2 -DNA all, LNA containing oligonucleotides increased oligomers uniformly modified with the phosphorothioate potency for reducing target mRNA in mouse liver up (PS) backbone substitution and work predominantly through to 5-fold relative to the corresponding MOE con- an RNase H-dependent mechanism. The substitution of sulfur taining ASOs. However, they also showed profound for oxygen in the phosphate ester confers several properties hepatotoxicity as measured by serum transamina- onto ASOs which are crucial for their use as systemic ses, organ weights and body weights. This toxicity drugs (2). Foremost, the PS linkage greatly increases stability was evident for multiple sequences targeting three to nucleolytic degradation (3), such that they possess suffi- different biological targets, as well as in mismatch cient stability in plasma, tissues and cells to avoid meta- control sequences having no known mRNA targets. bolism prior to reaching the target RNA after systemic Histopathological evaluation of tissues from LNA administration to an animal. Additionally, the PS modifica- tion confers a substantial pharmacokinetic benefit by increas- treated animals confirmed the hepatocellular ing the binding to plasma proteins, which prevents rapid renal involvement. Toxicity was observed as early as excretion (4). While greatly increasing the stability of ASOs, 4 days after a single administration. In contrast, the PS modified drugs are still subject to metabolism, and have corresponding MOE ASOs showed no evidence for tissue half-lives of 1–3 days (4), which is sub-optimal for a toxicity while maintaining the ability to reduce target parenterally administered drug. Furthermore, the PS modifi- mRNA. These studies suggest that while LNA ASOs cation reduces affinity for the target RNA (the ultimate bio- have the potential to improve potency, they impose logical receptor for ASO drugs) (3), which adversely effects a significant risk of hepatotoxicity. potency. In order to improve upon the first generation ASO drugs, many different modifications to the core nucleoside monomer INTRODUCTION unit of the ASO have been evaluated for their effects on affin- Synthetic oligonucleotides and their analogs are commonly ity for complementary RNA, nuclease resistance and ASO used as research reagents to modulate gene expression in potency. Most of the modifications which enhance affinity cell culture and in animal models. The most broadly utilized and nuclease resistance, in particular the 2 -substituted nucleo- mechanism by which oligonucleotides are exploited for sides, also limit the ability of the ASO to support RNase modulation of gene expression is through binding of the anti- H-mediated cleavage of the targeted RNA (5). Efforts to sense oligonucleotide (ASO) to a specific mRNA or pre- optimize the design of ASOs to retain the beneficial proper- ties of 2 -modifications, yet maintain RNase H activity have mRNA by Watson–Crick base pairing. Upon binding, the led to the development of chimeric oligonucleotide designs oligonucleotide can modulate RNA processing, inhibit trans- which employ higher affinity 2 -substituted nucleosides lation or promote degradation. Mechanisms of degradation *To whom correspondence should be addressed. Tel: +1 760 603 3825; Fax: +1 760 603 4653; Email: [email protected] 2006 The Author(s). This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. 688 Nucleic Acids Research, 2007, Vol. 35, No. 2 culture and in rodent liver after systemic administration. Our results indicate that although LNA modified ASOs have the potential to improve potency, they impose a significant risk of hepatotoxicity which must be considered when design- ing LNA containing antisense therapeutics. MATERIALS AND METHODS Oligonucleotide design and synthesis ASOs 1, 3 and 4 (sequence 5 -GCTCATACTCGTAGGCCA- Figure 1. (a) Gap design of current generation of ASO therapeutics having 0 0 0 0 0 0 2 -modified ‘wings’ at the 3 and 5 ends flanking a central 2 -deoxy gap 3 , position 791–808) and 2 (sequence 5 -CTCATACTCGT- region. (b) Structures of MOE and LNA nucleosides. AGGCC-3 , position 792–807) are complementary to Mus musculus TNFRSF1A-associated via death domain (TRADD) mRNA (Genbank accession no. NM_001033161). The ASO lead 1a is the murine homolog (a G to A base change at posi- combined with DNA regions which support RNase H activity. tion 5) of the human TRADD lead reported previously (28). There are several designs for such chimeric oligonucleotides Control oligonucleotides 5 (5 -GCCCAATCTCGTTAGC- with ‘gapmer’ designs being most common in which a GA-3 ) were designed with six mismatches to 4, such that 0 0 central DNA region of 7–14 nt is flanked on the 5 and 3 they contained >4 mismatches to all known mouse sequence. ends by 2–6 2 -modifications (Figure 1a) (6). ASOs 6 and 7 (sequence TCTGGTACATGGAAGTCTGG, The most advanced second generation antisense designs position 8232–8251) and 8 (sequence AAGTTGCCACCCA- are 2 -O-methoxyethylribose (7) (MOE) gapmer oligo- CATTCAG, position 5586–5605) are complementary to Mus nucleotides (Figure 1). MOE modified ASOs show increased musculus apolipoprotein B (ApoB) mRNA (Genbank acces- affinity toward a complementary RNA, and are highly resis- sion no. XM_137955.5). The sequences were identified by tant toward degradation by nucleases (8). These improve- a screen of 5-10-5 MOE 20mer ASOs as described previously ments result in a substantial (>20-fold) increase in (29–31). ASOs 9, 10 and 11 (sequence 5 -CTGCTAGCCTC- oligonucleotide potency in cell culture, relative to first gene- TGGATTTGA-3 , position 1931–1950) are complementary ration ASOs (9,10). In animals, MOE ASOs have been shown to M.musculus phosphatase and tensin homolog (PTEN), to possess both excellent pharmacokinetic properties (11–13) mRNA (Genbank accession no. NM_008960). ASO 9 (18) and robust pharmacological activity (14,15). Furthermore, the and control oligonucleotide 12 (19) have been described selective inhibition of gene expression with MOE ASOs previously. elicits positive pharmacological activity in several animal MOE phosphoramidites were prepared as described models of human disease when given systemically with clini- previously (7,32,33). LNA and 2 -deoxyribonucleoside phos- cally relevant routes and schedules of administration (15–19). phoramidites were purchased from commercial suppliers. These beneficial properties have translated to human thera- Oligonucleotides were prepared similar to that described pre- peutics. For example, a MOE ASO targeting ApoB has viously (34) on either an Amersham AKTA 10 or AKTA 100 shown a dose-dependent reduction of target protein con- oligonucleotide synthesizer. Modifications from the reported current with lowering of LDL cholesterol. Doses as low as procedure include: a decrease in the detritylation time to 100 mg per week produced statistically significant reductions 1 min, as this step was closely monitored by UV analysis in ApoB protein, and a dose of 200 mg per week reduced for complete release of the trityl group; phosphoramidite con- serum ApoB protein by 50% (20). Furthermore, to date centration was 0.1 M; 4,5-dicyanoimidazole catalyst was MOE ASOs have an excellent safety record in human clinical used at 0.7 M in the coupling step; 3-picoline was used trials (20–22). instead of pyridine for the sulfurization step, and the time The improvement in potency of MOE ASOs has, in part, decreased from 3 to 2 min. The oligonucleotides were then been attributed to the increased affinity for target mRNA con- purified by ion-exchange chromatography on an AKTA ferred by the MOE modification. Although MOE provides a Explorer and desalted by reverse phase HPLC to yield modi- substantial improvement in affinity, bicyclic nucleoside fied oligonucleotides in 30–40% isolated yield, based on 0 0 modifications such as 2 ,4 -methylene bridged nucleic acids the loading of the 3 -base onto the solid support. Oligo- (23,24) commonly called ‘locked nucleic acid’ (LNA, nucleotides were characterized by ion-pair-HPLC-MS analysis Figure 1b) (25,26) have been shown to provide a further (IP-HPLC-MS) with an Agilent 1100 MSD system. The increase in affinity. LNA containing chimeric ASOs are just purity of the oligonucleotides was >90% (Supplementary entering human clinical trials, and have recently been shown Table S1). to inhibit growth in human tumor xenograft models (27). However, studies of the effects of LNA ASOs versus endoge- Cell culture assays nous targets in liver, a tissue where antisense effects have been extensively characterized, have not been reported to For determining potency in cell culture, mouse brain endo- date. In order to investigate whether LNA ASOs have the thelial (bEND) cells (American Type Culture Collection) potential to improve upon MOE based ASO therapeutics, were transfected with the indicated concentration of oligo- we directly compared the potency and therapeutic index of nucleotide for 4 h using 3 mg/ml Lipofectin in OptiMEM. several designs of MOE and LNA containing ASOs in cell Transfection mixes were then replaced with normal growth Nucleic Acids Research, 2007, Vol. 35, No. 2 689 0 0 media [DMEM-high glucose, 10% fetal bovine serum (FBS) primer (RP) 5 -TGAAGAGTCAGTGGCCGGTT-3 and probe containing penicillin-streptomycin]. Cells were harvested (PR) 5 -6FAM-TTTCTGTTCCACGGGCAGCTCGTAGT- 24 h later and RNA was purified using Qiagen 96-well TAMRA-3 . Primers used for determination of PTEN RNA RNeasy plates. RNA was analyzed for TRADD and level are as follows: FP 5 -ATGACAATCATGTTGCAGCA- 0 0 0 cyclophilin A RNA levels. TRADD RNA levels, normalized ATTC-3 ,RP5 CGATGCAATAAATATGCACAAATCA 3 to those of cyclophilin A, are expressed as percent untreated and PR 5 -6FAM-CTGTAAAGCTGGAAAGGGACGGACT- control (% UTC). Each treatment was performed in triplicate. GGT-TAMRA-3 . Primers used for determination of ApoB IC values were determined using GraphPad Prism software RNA level are as follows: forward primer (FP) 5 -GAAAAT- by fitting the data to a sigmoidal dose–response curve (vari- AGACTTCCTGAATAACTATGCATT-3 , reverse primer (RP) 0 0 0 able slope) using a defined top of 100% and bottom of 0%. 5 -ACTCGCTTGCCAGCTTGC-3 and probe (PR) 5 -6FAM- For caspase induction, A549 cells (American Type Culture TTTCTGAGTCCCCGTGCCCAACA-TAMRA-3 . Primers Collection), a human lung carcinoma cell line, were seeded used for determination of cyclophilin A RNA level are as 0 0 0 into 96-well plates and transfected the following day with follows: FP 5 -TCGCCGCTTGCTGCA-3 ,RP5 -ATCGGCC- 0 0 300 nm ASO for 4 h using Lipofectin (Invitrogen). Transfec- GTGATGTCGA-3 and PR 5 -6FAM-CCATGGTCAACCCC- tion mixes were subsequently replaced with normal growth ACCGTGTTC-TAMRA-3 . media (Ham’s F12K media containing 10% FBS). Cells were assayed 44 h later for caspase 3 activity and cell number Western blot analysis (CyQuant) as described previously (35). Frozen tissue samples were homogenized in RIPA buffer (PBS containing 1% NP-40, 0.5% sodium deoxycholate, Animal treatment 0.1% SDS) containing Complete protease inhibitors (Roche), and protein concentrations were determined by Bio- All animal experiments were conducted according to the Rad protein assay. Protein samples were separated on a 10% American Association for the Accreditation of Laboratory PAGE gel (Invitrogen) and subsequently transferred to a Animal Care guidelines and were approved by the Animal PVDF membrane (Invitrogen). Membranes were incubated Welfare Committee. Male Balb/c mice, aged 6–8 weeks, at room temperature in blocking buffer consisting of 5% were obtained from Charles River Laboratories. Compounds non-fat dry milk in TBS-T for 1 h. Rabbit polyclonal antibod- were suspended in phosphate-buffered saline (PBS), filter ies were obtained from commercial sources, and used at sterilized and administered by intraperitoneal (i.p.) injection 1:1000 dilution. Phospho-eIF2alpha (Ser51) antibody was according to the indicated dosing schedules in a volume cor- obtained from Cell Signaling Technology (Catalog no. responding to 10 ml/g animal weight. Animals were main- 9721). HRP conjugated anti-rabbit secondary antibodies tained at a constant temperature of 23 C and were allowed were obtained from Jackson ImmunoResearch and were standard lab diet and water ad libitum and animal weights used at 1:2500 dilution. Protein bands were visualized using were monitored prior to dosing throughout the live phase of ECL-plus reagent (Amersham). the study. Immediately prior to sacrifice, mice were anes- thetized with isoflurane and terminal bleed was performed by cardiac puncture. Plasma or serum was isolated from Capillary gel electrophoresis whole blood and analyzed for clinical chemistries. Alanine Immediately following removal, organs were frozen in liquid aminotransferase (ALT) and aspartate aminotransferase nitrogen and stored at 80 C until ASO extraction. Tissues (AST) levels were determined using an Olympus AU400e were weighed and ASO was extracted as described previously bioanalyzer. Immediately following terminal blood draw, (12,36). Briefly, tissue samples were homogenized in a mice were sacrificed by cervical dislocation while under BioSavant (Bio 101), and subjected to solid phase extraction anesthesia. In conjunction with necropsy, liver and spleen using a phenyl bonded SPE column (Isolute). Concentrations weights were determined. Effects of compounds on organ of ASO in tissue as well as metabolite profiles were deter- weights, normalized to body weight, are expressed relative mined by capillary gel electrophoresis (CGE) using a to those of the saline treated group. Beckman P/ACE model 5010 capillary electrophoresis unit. RNA analysis Hematoxylin and eosin (H&E) staining Tissues were homogenized in 4 M guanidine isothiocyanate, Tissue samples were fixed in formalin for a minimum of 25 mM EDTA, 50 mM Tris–HCl, pH 6, containing 1 M 1 day followed by incubation in 70% ethanol for a minimum b-mercaptoethanol immediately following sacrifice and of 1 day. Tissue samples were further dehydrated and pro- homogenized. RNA was extracted using RNeasy columns cessed using a Leica ASP300 tissue processor. Tissues were (Qiagen) according to manufacturer’s protocol. RNA was embedded in paraffin and 4 m sections were mounted on eluted from the columns with water. RNA samples were ana- positive charged glass slides. Deparaffinized and rehydrated lyzed by fluorescence-based quantitative RT–PCR using an samples were stained for hematoxylin and eosin (H&E), Applied Biosystems 7700 sequence detector. Levels of target using a Leica Autostainer XL. RNAs as well as those of cyclophilin A, a housekeeping gene, were determined. Target RNA levels were normalized Immunohistochemical staining to cyclophilin levels for each RNA sample. Primers used for determination of TRADD RNA level are as follows: Immunohistochemical studies were performed to detect the 0 0 forward primer (FP) 5 -GGCCGCCTGCCAGAC-3 , reverse cleaved form of caspase 3, as well as Bcl2-associated X 690 Nucleic Acids Research, 2007, Vol. 35, No. 2 Table 1. Oligonucleotide targets, sequences and chemistry a m 0 0 Structure code is nucleotide units 5 to 3 . Capital letter is base code: G ¼ guanine, A ¼ adenine, C ¼ cytosine, C ¼ 5-methylcytosine, T ¼ thymine. Small subscript are sugar and linkage codes: ¼ MOE sugar, ¼ LNA sugar, ¼ deoxyribose sugar, ¼ PS linkage, ¼ PO linkage. e l d s o protein (Bax) and growth arrest and DNA-damage-inducible RESULTS beta (GADD45b) protein. Formalin-fixed, paraffin-embedded Effect of LNA modification in cell culture tissue sections were mounted on positive charged glass slides. To directly compare the effects of LNA and MOE modifi- Deparaffinized and rehydrated samples were heated for cations, we designed a series of oligonucleotides targeting 20 min at 95 C in citrate buffer solution. The slides were mouse TRADD mRNA. We have previously published the cooled for 20 min and endogenous peroxidase was blocked with 3% hydrogen peroxide (H O ) in methanol for 10 min identification and characterization of MOE modified oligo- 2 2 at room temperature, followed by rinsing in distilled water. nucleotides targeting human TRADD, a death domain adapter To detect cleaved caspase 3, Bax and GADD45b, the sections protein that interacts with TNF receptor family members (28). The lead ASO from this study, a 4-10-4 MOE design, were incubated 1 h at RT with anti-cleaved caspase 3 poly- contains a single mismatch to murine TRADD mRNA, clonal antibody, (1:50 dilution; Cell Signaling Technology, which was corrected to provide the mouse TRADD ASO Danvers, MA), anti-Bax monoclonal rabbit antibody, E63 1a. Transfection of 1a (Table 1) into bEND cells reduces (1:200 dilution; Epitomics, Burlingame CA) and anti- TRADD mRNA levels in a concentration-dependent manner GADD45b polyclonal antibody, C18 (0.5 mg/ml; Santa as measured by quantitative RT–PCR (Figure 2 and Table 2), Cruz, CA), respectively. The HRP conjugated secondary to give an IC of 8 nM. The 4-10-4 LNA version of this antibodies were obtained from Jackson ImmunoResearch sequence (1b) reduced potency by 4-fold relative to the cor- Laboratories (1:200 dilution; West Grove, PA). The antigen responding MOE ASO. Because LNA has an increased affin- was visualized with DAB (Dako Cytomation—Cat3K3466) ity relative to MOE, we hypothesized that the shorter ASOs for 5 min. For negative controls, primary antibodies were with reduced LNA content may maintain or improve potency, replaced with isotype matched normal IgG. The slides were and thus prepared the 3-10-3 MOE and LNA ASOs 2a and then counterstained with hematoxylin. Nucleic Acids Research, 2007, Vol. 35, No. 2 691 Table 2. Activity and liver levels of MOE (a series) and LNA (b series) ASOs ASO IC (nM) Est. ED (mg/kg) Liver conc. 50 50 a b c in vitro in vivo (mg/g tissue) ± SD 1a 8.3 13 116 ± 20 2a >60 >25 54 ± 5.9 3a 8.5 11 92 ± 11 4a 27 9 96 ± 14 1b 35 13 2b 15 6 64 ± 6.1 3b 1.6 4 69 ± 4.5 4b 8.4 2 48 ± 8.9 5a 60 25 5b 60 25 IC values for reduction of TRADD mRNA in bEND cells after transfection with Lipofectin. ED values for reduction of TRADD mRNA in mouse liver after dosing twice weekly for three weeks estimated by interpolation. Determined by capillary gel electrophoresis for the 4.5 mmol/kg dose groups at end of study. which maintained potency relative to 1a. Further substitution of two additional MOE residues gave the 2-14-2 ASO 4a, which reduced potency of 3–4-fold. In contrast, replacing two LNA units of 1b with deoxy nucleosides resulted in a marked improvement, with 3b having an IC of 1–2 nM. As with the MOE series, further substitution to afford 4b reduced the potency, such that the LNA 2-14-2 4b was approximately equipotent relative to the parent 4-10-4 MOE 1a.As expected, both MOE (5a) and LNA (5b) four base mismatch control ASOs were inactive. In vivo effects of 2 -MOE and LNA modified oligonucleotides To determine if the behavior of LNA modified ASOs was similar in animals, mice were treated with ASOs at several dosage levels two times per week for 3 weeks. As expected, Figure 2. Reduction of TRADD mRNA in bEND cells after transfection with MOE ASO 1a or LNA ASO 1b (top panel) or MOE ASO 4a or LNA ASO 4b treatment of mice with the 4-10-4 MOE ASO 1a twice per (bottom panel). week reduced TRADD mRNA in liver in a dose-dependent manner, producing a 77% reduction at the 4.5 mmol/kg 2b. Interestingly, the MOE 16mer 2a had greatly reduced dose (30 mg/kg, Figure 3). Doses of mmol per kg were potency, whereas the potency of the 16mer LNA 2b was employed as the compounds were of slightly different molecular weight, and we wished to directly compare improved relative to the 18mer LNA ASO 1b. These results potency on a molar basis. Consistent with cell culture results, are consistent with a previous study of LNA ASOs (37), in the corresponding LNA analog 1b was less efficacious at this which the high affinity of LNA facilitated good target dose, resulting in only a 65% inhibition of mRNA. The reduction with 16mer LNA ASO designs. Since RNase H activity has been found to be an important potency of this ASO is difficult to assess from the limited factor in antisense potency both in vitro and in vivo (38), dose–response data, but it appears to have a reduced efficacy, optimization of the gap region is of crucial importance. at least at the doses evaluated. Also consistent with cell cul- Because LNA has been shown to alter the conformation of ture results, the MOE 16mer ASO 2a was weakly active, adjacent DNA nucleotides for several residues (39), and while the corresponding LNA 2b showed 80 and 60% duplex conformation is critical for RNase H activity (5), we reductions in TRADD mRNA at the 4.5 and 1.5 mmol/kg included ASOs having increased gap sizes (3 and 4) in our doses, respectively. SAR set. This replacement of high affinity modifications by The 3-12-3 and 2-14-2 MOE ASOs 3a and 4a produced a additional 2 -deoxy units could increase RNase H activity small increase in potency relative to the 4-10-4 1a, as well as by either adding potential cleavage sites, or by reducing the an increase in efficacy at the 4.5 mmol/kg dose, with 2-14-2 conformational transmission effects of LNA. We reasoned design giving an 89% reduction in mRNA. The LNA ASO that the increased affinity of LNA relative to MOE would com- 3b (3-12-3 design) showed an increase in potency, achieving pensate for the reduced number of 2 -modifications, and may 80% reduction of mRNA at the 1.5 mmol/kg dose. In con- increase potency. Replacing two internal MOE modifications trast to cell culture results, the 2-14-2 LNA ASO 4b further of 1a with deoxy nucleosides provided the 3-12-3 ASO 3a, improved activity, giving a 75% inhibition of TRADD 692 Nucleic Acids Research, 2007, Vol. 35, No. 2 Figure 3. Reduction of TRADD mRNA for MOE (a series) and LNA (b series) ASOs. *Not Analyzed: due to severe toxicity at 8 days, the 4.5 mmol/kg dose group of 4b was terminated early. expression at the lowest dose of 0.5 mmol/kg (3 mg/kg). showed severe hepatotoxicity by histopathological analysis, Interpolation of the TRADD mRNA reduction data allowed discussed further below. estimation of a rough ED , which could be converted to The ALT and AST levels for all the remaining groups mg/kg dosing for a comparison of in vivo potency at the termination of study on Day 21 indicate a striking (Table 2). For 3b, the data suggest an 3-fold increase in difference in the liver function profile of the MOE and LNA potency over the corresponding MOE 3a and its 4-10-4 par- ASOs (Figure 4). All LNA ASOs studied showed at least ent 1a. The 2-14-2 LNA gapmer design 4b provided an even a 10-fold increase in transaminases for the 4.5 mmol/kg larger increase in potency, the magnitude of which could not dose groups, and some showed increases of >100-fold such be estimated from this experiment because the maximal effect as 3b, 4b (despite receiving only 3/6 doses) and 5b. In con- was achieved at the lowest dose tested. The efficacy of inhibi- trast, the transaminase levels in mice treated with MOE ASOs tion of TRADD expression for the high-dose group could not were within the normal range. The observed hepatotoxicity be obtained, as the high-dose group of animals treated with in the LNA ASO series was both compound and dose- 4b was sacrificed early due to severe toxicity observed as dependent, with ASOs having a larger DNA gap increasing discussed below. Mismatch control MOE and LNA ASOs hepatotoxicity. Further evidence for the toxicity of the LNA 5a and 5b showed no reduction in mRNA, indicating a ASOs was evident from the organ weights (Supplementary specific in vivo antisense effect on the target mRNA. Figure S2). LNA ASOs produced large (45–62%, relative to saline) liver weight increases in a dose-dependent manner, except for 5b, which had a smaller (25%), non-dose- LNA ASOs induce profound hepatotoxicity dependent increase. In contrast, the MOE ASOs showed As part of the routine monitoring of animals during the study, much smaller (0–17%) liver weight increases. Increase in plasma bilirubin and transaminase levels for all the high-dose spleen weight was more variable, with 2b and 3b producing groups were examined at Day 8 of the study in order to assess large (>100%, relative to saline) increases. The increase in liver function. Bilirubin, ALT and AST levels were within the spleen weights could be a direct proinflammatory effect of normal range for all animals, except for the 2-14-2 LNA the oligonucleotide or secondary to the severe hepatotoxicity. ASOs 4b and 5b. LNA 4b showed elevations of 186-, 75- and 18-fold for ALT, AST and bilirubin, respectively, relative LNA ASOs induce apoptosis in vitro and in vivo to saline treated animals. Notably, even the control LNA 5b, which showed no TRADD mRNA reduction, showed 46- and Examination of H&E stained liver sections from mice treated 25-fold increases in ALT and AST, respectively. Prior to with LNA ASOs 1b and 4b, as well as control oligo- receiving the fourth dose of ASO on Day 11 of the study, nucleotide 5b, confirmed hepatotoxic events. Histopathological all the animals in the 4.5 mmol/kg group receiving LNA observations included signs of apoptosis, profound eosinophilic ASO 4b experienced significant weight loss, losing 25% cytoplasmic degeneration with glycogen depletion and of their body weight (Supplementary Figure S1). Because hyperchromatic nuclei, as well as centrilobular coagulative of this severe weight loss, coupled with the very large necrosis surrounded with inflammatory infiltrate containing transaminase increases seen with 4b, the study was termi- neutrophils, monocytes and lymphocytes (Figure 5). Lesions of nated early for this dose group. Upon necropsy, these animals intracytoplasmic microvesicular changes were also visualized Nucleic Acids Research, 2007, Vol. 35, No. 2 693 Figure 4. Plasma transaminase levels for MOE (a series) and LNA (b series) ASOs. **Data from 8 days. AST 4020 ± 850, ALT 6470 ± 1450, and severe weight loss led to early termination of 4.5 mmol/kg dose group of 4b. in liver sections of 4b and 5b treated animals, suggesting early neoantigen formed by caspase cleavage of cytokeratin 18. signs of steatosis. An increased number of mitotic hepatocytes M30 immunoreactive cells appear at an early stage of apoptosis were also observed in liver sections from LNA treated animals, in epithelial cells, and are not detectable in vital or necrotic likely indicating regeneration of damaged tissue. epithelial cells (40). Animals treated with LNA ASOs 1b, 4b To further characterize the hepatotoxicity associated with or 5b all showed greatly increased M30 staining relative to LNA ASO treatment, we conducted immunohistopathological saline treated animals (Figure 5), providing further evidence evaluation to characterize the type of toxicity observed for apoptosis induced by treatment with the LNA ASOs. (Figure 5 and Supplementary Table S2). Liver sections To examine whether LNA ASOs could induce apoptosis in from animals treated with LNA ASOs were stained for the cell culture, we examined representative TRADD ASOs 2a, expression of GADD45b, the activated form of caspase 3, 4a at 300 nM in A549 cells for their ability to induce caspase peroxisome membrane protein 70 and Bax. Increased expres- 3 activity, a common marker of apoptosis, in cell culture sion of activated caspase 3 and Bax suggested increased after transfection (35). The LNA ASOs 2b and 4b resulted apoptosis in livers of LNA, but not MOE, oligonucleotide in a 3.1- and 6.2-fold induction of caspase 3 activity relative treated mice. GADD45b is a p53 and NF-kB regulated to control ASO 12, respectively, whereas the corresponding gene induced in response to cell stress. The LNA but not MOE ASOs gave no change relative to the control ASO MOE modified oligonucleotides increased GADD45b expres- (Supplementary Figure S3). sion dramatically. Evidence of peroxisome proliferation was also suggested by increased expression of peroxisome mem- LNA ASO improves potency but also increases toxicity brane protein 70 by IHC. Increased staining of each of these markers was dose-dependent and correlated well with ALT To verify the activity and toxicity observed with 4b,it increases. To further confirm apoptosis involvement of liver was tested in a repeat set of experiments along with 1a as a injury mediated by LNA, liver sections of saline and LNA comparator (Figure 6). In addition to the previously utilized treated animals were stained with the monoclonal antibody doses of 4.5, 1.5 and 0.5 mmol/kg, we also examined 0.9, M30, which is an apoptosis marker that monitors the 0.3 and 0.1 mmol/kg doses of 4b in order to assess to what 694 Nucleic Acids Research, 2007, Vol. 35, No. 2 Figure 5. Histopathology of liver sections from mice treated MOE (a series) and LNA (b series) ASOs at 4.5 mmol/kg twice weekly for 3 weeks. Livers from animals treated with LNA ASOs 1b, 4b and 5b present with significant hepatotoxicities as demonstrated in (A), a routing H&E stain showing profound swollen eosinophilic degeneration, cell death (right arrow) and hyperchromatic nuclei (up arrow) of hepatocytes. The immunohistochemistry reveals that the injured hepatocytes appear with cytoplasmic staining (brown stain cells, up arrows) of cleavage caspase 3 (B), pro-apoptotic protein BAX (C) and M30 (D), a neo- epitope generated in epithelial cells as a result of caspase activation (cleavage). In addition a DNA damage and repair associated protein GADD45b (E) and peroxisome membrane protein PMP70 (Supporting Supplementary Figure S2) were both found to have increased cytoplasmic expression in LNA treated livers. IgG control slides of all four IHC markers were negative (data not shown). Nucleic Acids Research, 2007, Vol. 35, No. 2 695 Figure 6. Transaminases (bar graph, left scale) and reduction of TRADD mRNA (points on line graph, right scale) after treatment with the 4-10-4 MOE gapmer 1a or the 2-14-2 LNA gapmer 4b. extent potency was increased, as well as whether there was an transaminase levels, organ weights, body weights or increase in therapeutic index. In this experiment, mice treated histopathological analysis of liver tissue samples. Thus, with three doses of 4b at 4.5 mmol/kg (at 11 days) showed though LNA ASO 4b was more potent than the comparator minimal weight loss, and the study was continued for the MOE ASOs, it was not more efficacious, and furthermore course of six administrations. At termination of the study, did not produce maximal efficacy in the absence of observ- the animals treated with 4b had lost 10% of their body able toxicity. These results demonstrate that despite the weight, as opposed to a 10% gain in the control group (Sup- increase in potency observed with the LNA ASOs, the thera- plementary Figures S4 and S5). Furthermore, organ weights, peutic index is not increased, and is probably decreased relat- transaminase increases and histopathological observations ive to the corresponding MOE ASOs. upon necropsy were consistent with the previous results, It is possible that both the improved potency and increased with >50-fold increases in AST and ALT at the 1.5 and toxicity could be due to increased distribution of LNA modi- 4.5 mmol/kg dose levels. The target reduction for 4b observed fied oligonucleotides to liver. To determine if this was the was similar to the first experiment, with 70 ± 8, 76 ± 6 and case, the concentration of ASO in the liver was measured 66 ± 8% reduction of TRADD mRNA at the 4.5, 1.5 and at the conclusion of the study for ASOs 1–4 (Table 2). The 0.5 mmol/kg doses, respectively. The 0.9 mmol/kg dose repeat group of 4b treated animals was used for this analysis, (5 mg/kg) produced the maximal effect of 4b, providing as the mice received the same number of total doses. The an 81 ± 4% reduction of TRADD mRNA, but also appeared liver concentrations ranged from a high of 116 mg/g for 1a to be above the maximum non-toxic dose, as transaminases to a low of 48 mg/g for 4b. These results are consistent were increased 5-fold. with expectations of higher metabolic stability for ASOs The 2-14-2 LNA 4b clearly demonstrated a dose- having greater numbers of MOE and LNA modifications, dependent reduction of TRADD mRNA, with an ED of and also with increased distribution to liver for ASOs having 0.37 mmol/kg (corresponding to 2 mg/kg, Table 2). This more PS linkages. As the concentration trend of ASO in liver suggests an improvement in potency of 4–5-fold relative does not correlate with potency or toxicity, the increased to the corresponding 2-14-2 MOE 4a, and 6–7-fold relative potency and/or toxicity of LNA oligonucleotides is not due to the comparator 4-10-4 MOE gapmer 1a. At dosage levels to more accumulation in the liver. used for the previous study, the activity of 1a was essentially LNA ASO 4b shows hepatotoxicity with a single identical to that observed previously. A higher dose of administration 6.2 mmol/kg (40 mg/kg) of 1a produced a larger reduction (86 ± 3%) of TRADD mRNA. At all doses of 1a employed, To better characterize the nature of the hepatotoxicity, we there was no evidence for toxicity as measured by administered a single dose of either MOE ASO 1a or LNA 696 Nucleic Acids Research, 2007, Vol. 35, No. 2 Figure 7. Transaminases (bar graph, left scale), and reduction of ApoB mRNA (points on line graph, right scale) after treatment with MOE (a series) and LNA (b series) ASOs. ASO 4b to mice, and examined effects at 2 and 4 days post- mRNA in mice treated with 20mer MOE ASOs. Three active dosing. In addition to plasma transaminases and histopatho- ASO sequences (6, 8 and 9, Table 1) targeting either mouse logy, levels of cleaved caspase 3 and phosphorylated eIF2a ApoB or mouse PTEN were identified in cell culture assays (p-eIF2a) were examined by western blot. Phosphorylation using methods previously described for these targets. Because of eIF2a has been shown to mediate apoptosis, presumably the LNA ASO design having two LNA nucleosides flanking a via the inhibition of translation (41). Additionally, levels of large deoxy gap region appeared to exhibit the greatest Bax, GADD45b, PUMA, p53, TNFa and MDM2 mRNA increase in potency and hepatotoxicity, we utilized a 2-16-2 were evaluated by RT–PCR. For all groups except for the design, and applied it to the previously identified 5-10-5 12 mmol/kg dosage level of 4b, there was no discernable dif- MOE sequences. ference from saline for any measured endpoint. In contrast, For the ApoB target, ASOs were dosed at 2.5 and either administration of 12 mmol/kg of the LNA ASO 4b gave a 0.5 or 0.4 mmol/kg twice weekly for 3 weeks (Figure 7). 7- and 2-fold elevation in ALT and AST, respectively, at The 5-10-5 MOE ASO 6 produced a modest 24% inhibition the day 4, but not at the day 2 time-point. Concurrent with of ApoB mRNA at the 0.4 mmol/kg dose and an 84% inhibi- this increase in transaminases at day 4, an increase in tion at the 2.5 mmol/kg dose. The 2-16-2 MOE ASOs 7a p-eIF2a was observed by western blot (Supplementary and 8a gave no inhibition at the low dose, and produced Figure S6). No increases in other genes studied were evident 50% reduction in ApoB-100 mRNA at the 2.5 mmol/kg by RT–PCR or western blot. However, weak staining of acti- (18 mg/kg) dose. The potency increase of the LNA ASOs vated caspase 3, Bax and the M30 neoantigen were observed was variable, with LNA 7b providing a 5-fold estimated by immunohistochemistry (Supplementary Table S3). These increase in potency over the corresponding 2-16-2 MOE 7a, results indicate that the hepatotoxicity induced by LNA 4b but not the parent 5-10-5 MOE ASO 6. LNA 8b was only can occur in as little as 4 days, and are consistent with marginally more active than MOE 8a, and substantially acute liver injury caused by induction of apoptosis. less active than the MOE 6. None of the MOE ASOs showed evidence for toxicity as measured by organ weights and serum transaminase increases. In contrast the LNA ASOs LNA effects on potency and hepatotoxicity are 7b and 8b both resulted in >20-fold increases in AST and independent of target ALT, along with increased organ weights at the 2.5 mmol/ To help rule out a potential target related contribution to the kg dose. hepatotoxicity and to verify the improved potency for other ASOs targeting murine PTEN were dosed at 0.083– targets, we tested additional LNA ASOs targeting other 2.25 mmol/kg twice weekly for 3 weeks (Figure 8). The mouse genes. We have previously published data demonstrat- parent MOE 5-10-5 ASO 9 as well as its gap widened coun- ing specific reduction of mouse ApoB (19) and PTEN (18) terpart 10a reduce target mRNA in a dose-dependent manner Nucleic Acids Research, 2007, Vol. 35, No. 2 697 Figure 8. Transaminases (bar graph, left scale), and reduction of PTEN mRNA (points on line graph, right scale) after treatment with MOE (a series) and LNA (b series) ASOs. LNA ASO shows hepatotoxicity in rats without evidence for toxicity as measured by transaminase As the sequence of the PTEN ASOs was homologous to levels, organ weights and body weights. The LNA ASO rat, we were able to examine if LNA 10b was more potent 10b was significantly more potent (estimated 5–10-fold) in rat, and importantly if the toxicity observed in mouse trans- than either MOE ASO. However, once again, this increase lated to another species. Accordingly, rats were treated with in potency correlated with an increase in hepatotoxicity. ASOs 9, 10b, or control 12 at dose of 0.83, 2.5 or The 0.75 mmol/kg dose group started to show mild transami- 7.5 mmol/kg twice weekly for 3 weeks (Supplementary nase elevations, while the higher dose group resulted in large Figures S7–S9). The MOE ASO 9 showed a dose-dependent (>50-fold) increases in both AST and ALT, increases in liver reduction of PTEN mRNA in rat liver, with an ED of and spleen weights and caused significant weight loss in trea- 2.5 mmol/kg. In contrast to the mouse data, the LNA ted animals. A non-targeted control 5-10-5 MOE ASO 12 had ASO 10b was only slightly more potent in rat liver, with no effect on either target or toxicity measures. an estimated ED of 1.5 mmol/kg. However, the LNA ASO was also hepatotoxic in rats. At the 7.5 mmol/kg dose, LNA ASO hepatotoxicity is not likely due to body weights were decreased 25% relative to control ASO LNA degradation products 12, AST (but not ALT) was increased 5- and 10-fold, respec- To help determine if the increased toxicity of the LNA ASOs tively in two of the four animals, and bilirubin was increased was due to the oligonucleotide or to degradation products, we dramatically in the same two animals. Histopathological prepared mixed backbone versions of the PTEN 2-16-2 ASOs evaluation of H&E stained liver sections confirmed hepato- containing phosphodiester (PO) linkages between LNA toxicity, showing moderate eosinophilic cytoplasmic degene- ration with focal single cell apoptosis and mild mononuclear nucleosides as well as at the LNA/DNA junction (Table 1). cell infiltration (Supplementary Table S4). No hepatotoxicity These ASOs are much more rapidly metabolized in vivo, was observed by histopathological evaluation, ALT, AST, and presumably will release either free MOE or LNA nucleo- bilirubin, organ weights or body weights for either of the sides or nucleotides. If these nucleosides or nucleotides were MOE ASOs. This data suggest that the potency increase of the source of the observed hepatotoxicity, these ASOs should LNA ASOs relative to MOE ASOs is more pronounced in be more toxic, while if the intact ASO was causing the mouse than in rat, though the observed hepatotoxicity is hepatotoxicity, the mixed backbone versions should be less still present in rat. toxic. Neither the MOE (11a) nor LNA (11b) ASO caused significant target reduction at the doses tested. Importantly, the LNA containing ASO showed no evidence of hepatotoxi- city, and analysis of drug levels in liver confirmed near com- DISCUSSION plete metabolism of the intact drug 11b. These results suggest that the intact LNA oligonucleotides are responsible for the The main goal of our study was to determine if LNA contain- observed hepatotoxicity. ing ASOs would improve potency and therapeutic index 698 Nucleic Acids Research, 2007, Vol. 35, No. 2 relative to the current generation of MOE ASOs. Our assump- increased Bax expression and increased expression of the tion entering the work was that an improvement in potency M30 neo-epitope. The upregulation of the pro-apoptotic pro- would yield an improved therapeutic index, since it has tein Bax suggests involvement of the p53-mediated apoptosis been generally believed that many of the toxicities of ASOs pathway, as Bax is a key response gene to p53 activation are due to class effects as a result of the PS backbone. How- (42,43). Furthermore, GADD45b, a key downstream target ever, this proved not to be the case with the LNA ASOs gene of p53 during DNA damage and repair process studied. (44), was highly up-regulated in the injured hepatocytes. Our results clearly demonstrate the ability to improve GADD45b appears to help protect cells against programmed potency with some, but not all, LNA containing ASO designs, cell death through blocking the c-jun N-terminal kinase cas- particularly for the TRADD and PTEN targets. This improve- cade, and is probably induced in response to cellular damage. ment was occasionally fairly large, as much as 5–10-fold, and It is unclear why LNA oligonucleotides cause this level of was most pronounced for LNA ASOs of length 18–20 nt hepatotoxicity though the corresponding MOE oligo- which contained 2–3 LNA residues at each end. As little as nucleotides do not. One possibility is that antisense effects 0.75–1 mmol/kg (5–6 mg/kg) of these ASOs given twice on genes partially complementary to the hepatotoxic ASOs weekly for 3 weeks reduced target mRNA by 80%. The opti- are playing a role in the toxicity, as LNA ASOs have been mal LNA ASO design in vivo appeared to be different than shown to decrease the selectivity for a perfectly complemen- that observed in cell culture, where we found that two LNA tary target relative to the corresponding MOE ASOs (45). nucleotides on each end of the ASO provided the largest However, this seems unlikely because multiple unrelated potency increase. This is evident from a comparison of 3b LNA sequences cause similar toxicities. All oligonucleotides and 4b, where 3b was 5-fold more potent in cell culture, were prepared and purified in the same laboratory using the but less potent in vivo. It is unlikely that the improved identical methods. The impurity profiles of LNA and MOE potency is due solely to increased affinity of the ASO for tar- ASOs were nearly identical, and contained only the expected get RNA, as adding more LNA to the ASO actually decreased impurities resulting from PS oligonucleotide synthesis (PO, potency both in cell culture and in vivo (compare 2b and 3b N-1, etc.). This makes it extremely unlikely that the toxicity with 1b). Because of these trends, combined with the lack of is due to impurities resulting from LNA, but not MOE increased distribution of LNA ASOs to liver, it is likely that oligonucleotide synthesis. Additionally, since the metaboli- other factors are contributing to the increased potency of cally unstable PO containing ASO 11b (which should be LNA modified ASOs observed in our studies. Additional metabolized in vivo to LNA nucleosides and nucleotides) investigations will be required to further characterize the was non-toxic, it is not likely that the toxicity is due to the nature of this potency improvement. LNA monomers. This suggests that the intact LNA contain- Unfortunately, the increased activity of LNA containing ing PS oligonucleotides are responsible for the observed toxi- ASOs was also accompanied by the observation of severe city. There are distinct structural differences between MOE hepatotoxicity, such that there was little or no separation and LNA which may allow LNA containing oligonucleotides between toxic doses and those that produced significant to selectively effect hepatotoxicity. Perhaps importantly, the levels of mRNA reduction. Hepatotoxicity was chemistry-, rigid acyclic 2 -methoxyethyl side chain of MOE protects sequence- and design-dependent, as it was only observed the corresponding 3 -phosphorothioate linkage from inter- with LNA containing ASOs, and the onset occurred at actions via increased steric bulk and hydration (8), relative slightly different dose levels for different compounds. The to the compact and more hydrophobic cyclic structure of fact that the MOE ASOs in some cases (compare 2a with LNA. This could cause selective binding affinity differences 2b, and 6 with 7b and 8b) produced similar reductions in tar- between MOE and LNA oligonucleotides for as yet unknown get RNA without producing observable toxicity suggest that macromolecular binding partners, and/or result in differential the toxicity is not secondary to reduction in target gene compartmentalization of the two classes of oligonucleotides expression. This is further supported by the observation of within liver tissue. severe hepatotoxicity with control 5b, which has >3 mis- The mild hepatotoxicity induced 4 days after a single matches to all known mouse sequences. Hepatotoxicity also administration of LNA ASO 4b occurred concurrently with seemed to be the most severe for the more potent LNA apoptosis and activation of Bax and caspase 3 in hepatocytes ASO designs regardless of target (see 4b, mismatch 5b, 7b as evidenced by histopathological evaluation. Furthermore, and 10b). Thus, therapeutic index was not improved, and an increase in p-eIF2a was observed to coincide with the was likely decreased relative to the MOE ASOs. onset of toxicity. Phosphorylated eIF2a inhibits translation The hepatotoxicity was evident from the observation of initiation, and has been shown to mediate apoptosis, multiple parameters, including histopathological evaluation possibly by preventing the synthesis of short lived anti- of liver tissue upon necropsy as well as large increases in apoptotic factors (41,46). There are four known kinases plasma levels of aminotransferases (ALT and AST). Further- which phosphorylate eIF2a: PKR, which is activated by more, the toxicity was commonly accompanied by large binding of double stranded RNA (dsRNA); GCN2, which is increases in liver and/or spleen weights, likely as a con- activated by amino acid deprivation; HRI, which is activated sequence arising from a response to hepatic injury induced by low heme levels; and PERK, which responds to stress by the LNA ASOs. In several cases, the toxicity was severe in the endoplasmic reticulum. It is unclear from our data enough to cause extensive weight loss in the animals. Histo- how treatment with hepatotoxic LNA oligonucleotides logy data clearly demonstrated both LNA oligonucleotide- results in increased phosphorylation of eIF2a; however, it indiced liver necrosis and activation of apoptosis pathways, is tempting to speculate that PKR could be involved. PKR as evidenced by H&E staining, as well as cleaved caspase 3, is activated by binding of dsRNA to distinct dsRNA binding Nucleic Acids Research, 2007, Vol. 35, No. 2 699 domains which serve as allosteric inhibitors of the kinase has reactivated or is persistently active despite other therapies in patients with AIDS. Am. J. Ophthalmol., 133, 475–483. domain (47,48). It remains to be determined if LNA oligo- 2. Eckstein,F. (2000) Phosphorothioate oligodeoxynucleotides: what is nucleotides interact with PKR, and more extensive mechanis- their origin and what is unique about them? Antisense Nucleic Acid tic work is required to confirm both p-eIF2a mediation of an Drug Dev., 10, 117–121. acute apoptotic response in hepatocytes, and elucidate the 3. Stein,C.A., Subasinghe,C., Shinozuka,K. and Cohen,J.S. (1988) Physicochemical properties of phosphorothioate precise biochemical mechanism responsible for the observed oligodeoxynucleotides. Nucleic Acids Res., 16, 3209–3221. hepatotoxicity. 4. Geary,R.S., Yu,R.Z. and Levin,A.A. (2001) Pharmacokinetics of In conclusion, we have shown that optimization of size phosphorothioate antisense oligodeoxynucleotides. Curr. Opin. and gap configuration employing LNA containing ASOs Investig. Drugs, 2, 562–573. can improve potency in mouse liver. However, this potency 5. Lima,W.F., Nichols,J.G., Wu,H., Prakash,T.P., Migawa,M.T., Wyrzykiewicz,T.K., Bhat,B. and Crooke,S.T. (2004) Structural increase is strongly correlated with the onset of a severe requirements at the catalytic site of the heteroduplex substrate for hepatotoxicity not seen with the corresponding MOE ASOs, human RNase H1 catalysis. J. Biol. Chem., 279, 36317–36326. and therapeutic index is not improved. These results suggest 6. Monia,B.P., Lesnik,E.A., Gonzalez,C., Lima,W.F., McGee,D., that while LNA ASOs have the potential to improve potency Guinosso,C.J., Kawasaki,A.M., Cook,P.D. and Freier,S.M. (1993) 0 0 Evaluation of 2 -modified oligonucleotides containing 2 -deoxy gaps as of antisense therapeutics, they impose a significant risk of antisense inhibitors of gene expression. J. Biol. Chem., 268, hepatotoxicity. The toxicities observed herein also pose chal- 14514–14522. lenges for interpreting target validation and pharmacology 7. Martin,P. (1995) Ein neuer Zugang zu 2-O-Alkylribonucleosiden und experiments using LNA ASOs in rodent models. It is possible Eigenschaften deren Oligonucleotide. Helv. Chim. Acta, 78, 486–504. more extensive screening efforts may identify less toxic LNA 8. Teplova,M., Minasov,G., Tereshko,V., Inamati,G.B., Cook,P.D., Manoharan,M. and Egli,M. (1999) Crystal structure and improved sequences, or that other ASO designs utilizing LNA may not antisense properties of 2 -O-(2-methoxyethyl)-RNA. Nature Struct. elicit hepatotoxicity. Additionally, bicyclic nucleosides simi- Biol., 6, 535–539. lar to LNA such as amino-LNA, thio-LNA (27) and ENA 9. McKay,R.A., Miraglia,L.J., Cummins,L.L., Owens,S.R., Sasmor,H. (which contains one extra methylene unit in the bicyclic and Dean,N.M. (1999) Characterization of a potent and specific class of antisense oligonucleotide inhibitor of human protein kinase C-alpha bridge) (49) have been reported to show antisense effects expression. J. Biol. Chem., 274, 1715–1722. similar to LNA and MOE ASOs. Although optimized designs 10. Altmann,K.H., Martin,P., Dean,N.M. and Monia,B.P. (1997) Second of one or more of these modifications may ultimately improve generation antisense oligonucleotides—inhibition of pkc-alpha and potency while maintaining the safety profile of MOE ASOs, c-raf kinase expression by chimeric oligonucleotides incorporating 0 0 no detailed characterization of their toxicological properties 6 -substituted carbocyclic nucleosides and 2 -O-ethylene glycol substituted ribonucleosides. Nucleosides Nucleotides, 16, 917–926. has yet been reported. It is clear that additional structure– 11. Geary,R.S., Watanabe,T.A., Truong,L., Freier,S., Lesnik,E.A., activity and structure–toxicity relationship studies will be Sioufi,N.B., Sasmor,H., Manoharan,M. and Levin,A.A. (2001) required to more fully assess the therapeutic potential of 0 Pharmacokinetic properties of 2 -O-(2-methoxyethyl)-modified LNA and other bicyclic nucleoside modified oligo- oligonucleotide analogs in rats. J. Pharmacol. Exp. Ther., 296, 890–897. nucleotides. Ultimately, the value of LNA oligonucleotide 12. Yu,R.Z., Geary,R.S., Monteith,D.K., Matson,J., Truong,L., Fitchett,J. drugs will be determined in human clinical trials, where and Levin,A.A. (2004) Tissue disposition of 2 -O-(2-methoxy) ethyl they are currently being developed for the treatment of modified antisense oligonucleotides in monkeys. J. Pharm. Sci., 93, cancer. Our results are based only on studies in rodents; how- 48–59. ever, caution should be exercised in the clinical development 13. Geary,R.S., Yu,R.Z., Watanabe,T., Henry,S.P., Hardee,G.E., Chappell,A., Matson,J., Sasmor,H., Cummins,L. and Levin,A.A. (2003) of LNA modified oligonucleotides, especially for chronic, Pharmacokinetics of a tumor necrosis factor-alpha phosphorothioate non-life threatening indications. 2 -O-(2-methoxyethyl) modified antisense oligonucleotide: comparison across species. Drug Metab. Dispos., 31, 1419–1428. 14. Zhang,H., Cook,J., Nickel,J., Yu,R., Stecker,K., Myers,K. and SUPPLEMENTARY DATA Dean,N.M. (2000) Reduction of liver Fas expression by an antisense oligonucleotide protects mice from fulminant hepatitis. Supplementary Data are available at NAR Online. Nat. Biotechnol., 18, 862–867. 15. Yu,R.Z., Zhang,H., Geary,R.S., Graham,M., Masarjian,L., Lemonidis,K., Crooke,R., Dean,N.M. and Levin,A.A. (2001) ACKNOWLEDGEMENTS Pharmacokinetics and pharmacodynamics of an antisense phosphorothioate oligonucleotide targeting Fas mRNA in mice. The authors thank Charles Allerson, Amy Dan and Rick Carty J. Pharmacol. Exp. Ther., 296, 388–395. for synthesis of oligonucleotides, Megan Pearce for determi- 16. Liang,Y., Osborne,M.C., Monia,B.P., Bhanot,S., Watts,L.M., She,P., Decarlo,S.O., Chen,X. and Demarest,K. (2005) Antisense nation of ASO levels in liver tissue, Denis Drygin for the oligonucleotides targeted against glucocorticoid receptor reduce caspase 3 activation data, Richard H. Griffey for numerous hepatic glucose production and ameliorate hyperglycemia in diabetic thought provoking discussions and Stanley T. Crooke and mice. Metabolism, 54, 848–855. Arthur A. Levin for a critical read of the manuscript. Funding 17. Zinker,B.A., Rondinone,C.M., Trevillyan,J.M., Gum,R.J., Clampit,J.E., to pay the Open Access publication charges for this article Waring,J.F., Xie,N., Wilcox,D., Jacobson,P., Frost,L. et al. (2002) PTP1B antisense oligonucleotide lowers PTP1B protein, normalizes was provided by Isis Pharmaceuticals. blood glucose, and improves insulin sensitivity in diabetic mice. Proc. Natl Acad. Sci. USA, 99, 11357–11362. Conflict of interest statement. None declared. 18. Butler,M., McKay,R.A., Popoff,I.J., Gaarde,W.A., Witchell,D., Murray,S.F., Dean,N.M., Bhanot,S. and Monia,B.P. (2002) Specific inhibition of PTEN expression reverses hyperglycemia in diabetic REFERENCES mice. Diabetes, 51, 1028–1034. 19. Crooke,R.M., Graham,M.J., Lemonidis,K.M., Whipple,C.P., Koo,S. 1. Vitravene Study Group (2002) Randomized dose-comparison studies of and Perera,R.J. (2005) An apolipoprotein B antisense oligonucleotide intravenous fomiversen for treatment of cytomegalovirus retinitis that 700 Nucleic Acids Research, 2007, Vol. 35, No. 2 lowers LDL cholesterol in hyperlipidemic mice without causing phosphorothioate RNA utilizing phenylacetyl disulfide (PADS). hepatic steatosis. J. Lipid Res., 46, 872–884. Bioorg. Med. Chem. Lett., 16, 2513–2517. 20. Kastelein,J.J.P., Wedel,M.K., Baker,B.F., Su,J., Bradley,J.D., Yu,R.Z., 35. Drygin,D., Barone,S. and Bennett,C.F. (2004) Sequence-dependent Chuang,E., Graham,M.J. and Crooke,R.M. (2006) Potent reduction of cytotoxicity of second-generation oligonucleotides. Nucluic Acids Res., apolipoprotein B and LDL cholesterol by short-term administration of 32, 6585–6594. an antisense inhibitor of apolipoprotein B. Circulation, 114, 36. Leeds,J.M., Graham,M.J., Truong,L. and Cummins,L.L. (1996) 1729–1735. Quantitation of phosphorothioate oligonucleotides in human plasma. 21. Chi,K.N., Eisenhauer,E., Fazli,L., Jones,E.C., Goldenberg,S.L., Anal. Biochem., 235, 36–43. Powers,J., Tu,D. and Gleave,M.E. (2005) A phase I pharmacokinetic 37. Frieden,M., Christensen,S.M., Mikkelsen,N.D., Rosenbohm,C., and pharmacodynamic study of OGX-011, a 2 -methoxyethyl antisense Thrue,C.A., Westergaard,M., Hansen,H.F., Orum,H. and Koch,T. oligonucleotide to clusterin, in patients with localized prostate cancer. (2003) Expanding the design horizon of antisense oligonucleotides with J. Natl Cancer Inst., 97, 1287–1296. alpha-l-LNA. Nucleic Acids Res., 31, 6365–6372. 22. Sewell,K.L., Geary,R.S., Baker,B.F., Glover,J.M., Mant,T.G., Yu,R.Z., 38. Wu,H., Lima,W.F., Zhang,H., Fan,A., Sun,H. and Crooke,S.T. (2004) Tami,J.A. and Dorr,F.A. (2002) Phase I trial of ISIS 104838, a Determination of the role of the human RNase H1 in the pharmacology 2 -methoxyethyl modified antisense oligonucleotide targeting tumor of DNA-like antisense drugs. J. Biol. Chem., 279, 17181–17189. necrosis factor-alpha. J. Pharmacol. Exp. Ther., 303, 1334–1343. 39. Bondensgaard,K., Petersen,M., Singh,S.K., Rajwanshi,V.K., Kumar,R., 23. Obika,S., Nanbu,D., Hari,Y., Morio,K.-I., In,Y., Ishida,T. and Wengel,J. and Jacobsen,J.P. (2000) Structural studies of LNA:RNA 0 0 Imanishi,T. (1997) Synthesis of 2 -O-4 -C-methyleneuridine duplexes by NMR: conformations and implications for RNase H and -cytidine. Novel bicyclic nucleosides having a fixed C3 -endo activity. Chem.-Eur. J., 6, 2687–2695. sugar puckering. Tetrahedron Lett., 38, 8735–8738. 40. Grassi,A., Susca,M., Ferri,S., Gabusi,E., D’Errico,A., Farina,G., 24. Obika,S., Nanbu,D., Hari,Y., Andoh,J.-I., Morio,K.-I., Doi,T. and Maccariello,S., Zauli,D., Bianchi,F.B. and Ballardini,G. (2004) Imanishi,T. (1998) Stability and structural features of the duplexes Detection of the M30 neoepitope as a new tool to quantify liver containing nucleoside analogs with a fixed N-type conformation, apoptosis: timing and patterns of positivity on frozen and 0 0 2 -O-4 -C-methyleneribonucleosides. Tetrahedron Lett., 39, paraffin-embedded sections. Am. J. Clin. Pathol., 121, 5401–5404. 211–219. 0 0 25. Wengel,J. (1999) Synthesis of 3 -C- and 4 -C-branched 41. Scheuner,D., Patel,R., Wang,F., Lee,K., Kumar,K., Wu,J., Nilsson,A., oligodeoxynucleotides and the development of locked nucleic acid Karin,M. and Kaufman,R.J. (2006) Double-stranded RNA-dependent (LNA). Acc. Chem. Res., 32, 301–310. protein kinase phosphorylation of the alpha-subunit of eukaryotic 26. Koshkin,A.A., Nielsen,P., Meldgaard,M., Rajwanshi,V.K., Singh,S.K. translation initiation factor 2 mediates apoptosis. J. Biol. Chem., 281, and Wengel,J. (1998) LNA (locked nucleic acid): an RNA mimic 21458–21468. forming exceedingly stable LNA:LNA duplexes. J. Am. Chem. Soc., 42. Chipuk,J.E., Kuwana,T., Bouchier-Hayes,L., Droin,N.M., 120, 13252–13253. Newmeyer,D.D., Schuler,M. and Green,D.R. (2004) Direct activation 27. Fluiter,K., Frieden,M., Vreijling,J., Rosenbohm,C., De Wissel,M.B., of Bax by p53 mediates mitochondrial membrane permeabilization and Christensen,S.M., Koch,T., Orum,H. and Baas,F. (2005) On the in vitro apoptosis. Science, 303, 1010–1014. and in vivo properties of four locked nucleic acid nucleotides 43. Mihara,M., Erster,S., Zaika,A., Petrenko,O., Chittenden,T., Pancoska,P. incorporated into an anti-H-Ras antisense oligonucleotide. Chem. and Moll,U.M. (2003) p53 has a direct apoptogenic role at the Biochem, 6, 1104–1109. mitochondria. Mol. Cell, 11, 577–590. 28. Siwkowski,A.M., Madge,L.A., Koo,S., McMillan,E.L., Monia,B.P., 44. Xiao,G., Chicas,A., Olivier,M., Taya,Y., Tyagi,S., Kramer,F.R. and Pober,J.S. and Baker,B.F. (2004) Effects of antisense Bargonetti,J. (2000) A DNA damage signal is required for p53 to oligonucleotide-mediated depletion of tumor necrosis factor (TNF) activate gadd45. Cancer Res., 60, 1711–1719. receptor 1-associated death domain protein on TNF-induced gene 45. Simoes-Wust,A.P., Hopkins-Donaldson,S., Sigrist,B., Belyanskaya,L., expression. Mol. Pharmacol., 66, 572–579. Stahel,R.A. and Zangemeister-Wittke,U. (2004) A functionally 29. Chiang,M.Y., Chan,H., Zounes,M.A., Freier,S.M., Lima,W.F. and improved locked nucleic acid antisense oligonucleotide inhibits Bcl-2 Bennett,C.F. (1991) Antisense oligonucleotides inhibit intercellular and Bcl-xL expression and facilitates tumor cell apoptosis. adhesion molecule 1 expression by two distinct mechanisms. Oligonucleotides, 14, 199–209. J. Biol. Chem., 266, 18162–18171. 46. Saelens,X., Kalai,M. and Vandenabeele,P. (2001) Translation 30. Dean,N.M. and McKay,R. (1994) Inhibition of protein kinase C-alpha inhibition in apoptosis: caspase-dependent PKR activation and expression in mice after systemic administration of phosphorothioate eIF2-alpha phosphorylation. J. Biol. Chem., 276, 41620–41628. antisense oligodeoxynucleotides. Proc. Natl Acad. Sci. USA, 91, 47. Nanduri,S., Rahman,F., Williams,B.R. and Qin,J. (2000) A 11762–11766. dynamically tuned double-stranded RNA binding mechanism for the 31. Vickers,T.A., Koo,S., Bennett,C.F., Crooke,S.T., Dean,N.M. and activation of antiviral kinase PKR. EMBO J., 19, 5567–5574. Baker,B.F. (2003) efficient reduction of target RNAs by small 48. Wu,S. and Kaufman,R.J. (1997) A model for the double-stranded RNA interfering RNA and RNase H-dependent antisense agents. (dsRNA)-dependent dimerization and activation of the J. Biol. Chem., 278, 7108–7118. dsRNA-activated protein kinase PKR. J. Biol. Chem., 272, 32. Ross,B. and Song,Q. (2004) US U.S. Pat. Appl. Publ. 20040082775. 1291–1296. 33. Ross,B.S., Song,Q. and Han,M. (2005) Kilo-scale synthesis process for 49. Koizumi,M., Takagi-Sato,M., Okuyama,R., Araki,K., Sun,W., 2 -O-(2-methoxyethyl)-pyrimidine derivatives. Nucleosides Nakai,D., Tsutsumi,S. and Kawai,K. (2006) Direct comparison of Nucleotides Nucleic Acids, 24, 815–818. in vivo antisense activity of ENA oligonucleotides targeting PTP1B 34. Ravikumar,V.T., Andrade,M., Carty,R.L., Dan,A. and Barone,S. (2006) mRNA with that of 2 -O-(2-methoxy)ethyl-modified oligonucleotides. Development of siRNA for therapeutics: efficient synthesis of Oligonucleotides, 16, 253–262. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

Antisense oligonucleotides containing locked nucleic acid improve potency but cause significant hepatotoxicity in animals

Download PDF
14 pages

Loading next page...
 
/lp/oxford-university-press/antisense-oligonucleotides-containing-locked-nucleic-acid-improve-aZ324tVBev

References (54)

Publisher
Oxford University Press
Copyright
© Published by Oxford University Press.
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/gkl1071
pmid
17182632
Publisher site
See Article on Publisher Site

Abstract

Published online 19 December 2006 Nucleic Acids Research, 2007, Vol. 35, No. 2 687–700 doi:10.1093/nar/gkl1071 Antisense oligonucleotides containing locked nucleic acid improve potency but cause significant hepatotoxicity in animals Eric E. Swayze*, Andrew M. Siwkowski, Edward V. Wancewicz, Michael T. Migawa, Tadeusz K. Wyrzykiewicz, Gene Hung, Brett P. Monia and and C. Frank Bennett Isis Pharmaceuticals, Inc., 1896 Rutherford Road, Carlsbad, CA 92008,USA Received August 18, 2006; Revised October 27, 2006; Accepted November 9, 2006 include recruitment of RNase H, which cleaves the RNA ABSTRACT strand of a DNA:RNA duplexes, and activation of the RNA A series of antisense oligonucleotides (ASOs) interference (RNAi) pathway utilizing short RNA duplexes containing either 2 -O-methoxyethylribose (MOE) or (siRNA) or hairpins (shRNA). In addition to the use of locked nucleic acid (LNA) modifications were de- oligonucleotides as research reagents there is increasing inter- signed to investigate whether LNA antisense oligo- est in exploiting oligonucleotides as therapeutic agents. Cur- nucleotides (ASOs) have the potential to improve rently there is one approved antisense product (Vitravene, upon MOE based ASO therapeutics. Some, but not fomivirsen) (1) and >30 products in active development. The first generation of ASO therapeutics were 2 -DNA all, LNA containing oligonucleotides increased oligomers uniformly modified with the phosphorothioate potency for reducing target mRNA in mouse liver up (PS) backbone substitution and work predominantly through to 5-fold relative to the corresponding MOE con- an RNase H-dependent mechanism. The substitution of sulfur taining ASOs. However, they also showed profound for oxygen in the phosphate ester confers several properties hepatotoxicity as measured by serum transamina- onto ASOs which are crucial for their use as systemic ses, organ weights and body weights. This toxicity drugs (2). Foremost, the PS linkage greatly increases stability was evident for multiple sequences targeting three to nucleolytic degradation (3), such that they possess suffi- different biological targets, as well as in mismatch cient stability in plasma, tissues and cells to avoid meta- control sequences having no known mRNA targets. bolism prior to reaching the target RNA after systemic Histopathological evaluation of tissues from LNA administration to an animal. Additionally, the PS modifica- tion confers a substantial pharmacokinetic benefit by increas- treated animals confirmed the hepatocellular ing the binding to plasma proteins, which prevents rapid renal involvement. Toxicity was observed as early as excretion (4). While greatly increasing the stability of ASOs, 4 days after a single administration. In contrast, the PS modified drugs are still subject to metabolism, and have corresponding MOE ASOs showed no evidence for tissue half-lives of 1–3 days (4), which is sub-optimal for a toxicity while maintaining the ability to reduce target parenterally administered drug. Furthermore, the PS modifi- mRNA. These studies suggest that while LNA ASOs cation reduces affinity for the target RNA (the ultimate bio- have the potential to improve potency, they impose logical receptor for ASO drugs) (3), which adversely effects a significant risk of hepatotoxicity. potency. In order to improve upon the first generation ASO drugs, many different modifications to the core nucleoside monomer INTRODUCTION unit of the ASO have been evaluated for their effects on affin- Synthetic oligonucleotides and their analogs are commonly ity for complementary RNA, nuclease resistance and ASO used as research reagents to modulate gene expression in potency. Most of the modifications which enhance affinity cell culture and in animal models. The most broadly utilized and nuclease resistance, in particular the 2 -substituted nucleo- mechanism by which oligonucleotides are exploited for sides, also limit the ability of the ASO to support RNase modulation of gene expression is through binding of the anti- H-mediated cleavage of the targeted RNA (5). Efforts to sense oligonucleotide (ASO) to a specific mRNA or pre- optimize the design of ASOs to retain the beneficial proper- ties of 2 -modifications, yet maintain RNase H activity have mRNA by Watson–Crick base pairing. Upon binding, the led to the development of chimeric oligonucleotide designs oligonucleotide can modulate RNA processing, inhibit trans- which employ higher affinity 2 -substituted nucleosides lation or promote degradation. Mechanisms of degradation *To whom correspondence should be addressed. Tel: +1 760 603 3825; Fax: +1 760 603 4653; Email: [email protected] 2006 The Author(s). This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. 688 Nucleic Acids Research, 2007, Vol. 35, No. 2 culture and in rodent liver after systemic administration. Our results indicate that although LNA modified ASOs have the potential to improve potency, they impose a significant risk of hepatotoxicity which must be considered when design- ing LNA containing antisense therapeutics. MATERIALS AND METHODS Oligonucleotide design and synthesis ASOs 1, 3 and 4 (sequence 5 -GCTCATACTCGTAGGCCA- Figure 1. (a) Gap design of current generation of ASO therapeutics having 0 0 0 0 0 0 2 -modified ‘wings’ at the 3 and 5 ends flanking a central 2 -deoxy gap 3 , position 791–808) and 2 (sequence 5 -CTCATACTCGT- region. (b) Structures of MOE and LNA nucleosides. AGGCC-3 , position 792–807) are complementary to Mus musculus TNFRSF1A-associated via death domain (TRADD) mRNA (Genbank accession no. NM_001033161). The ASO lead 1a is the murine homolog (a G to A base change at posi- combined with DNA regions which support RNase H activity. tion 5) of the human TRADD lead reported previously (28). There are several designs for such chimeric oligonucleotides Control oligonucleotides 5 (5 -GCCCAATCTCGTTAGC- with ‘gapmer’ designs being most common in which a GA-3 ) were designed with six mismatches to 4, such that 0 0 central DNA region of 7–14 nt is flanked on the 5 and 3 they contained >4 mismatches to all known mouse sequence. ends by 2–6 2 -modifications (Figure 1a) (6). ASOs 6 and 7 (sequence TCTGGTACATGGAAGTCTGG, The most advanced second generation antisense designs position 8232–8251) and 8 (sequence AAGTTGCCACCCA- are 2 -O-methoxyethylribose (7) (MOE) gapmer oligo- CATTCAG, position 5586–5605) are complementary to Mus nucleotides (Figure 1). MOE modified ASOs show increased musculus apolipoprotein B (ApoB) mRNA (Genbank acces- affinity toward a complementary RNA, and are highly resis- sion no. XM_137955.5). The sequences were identified by tant toward degradation by nucleases (8). These improve- a screen of 5-10-5 MOE 20mer ASOs as described previously ments result in a substantial (>20-fold) increase in (29–31). ASOs 9, 10 and 11 (sequence 5 -CTGCTAGCCTC- oligonucleotide potency in cell culture, relative to first gene- TGGATTTGA-3 , position 1931–1950) are complementary ration ASOs (9,10). In animals, MOE ASOs have been shown to M.musculus phosphatase and tensin homolog (PTEN), to possess both excellent pharmacokinetic properties (11–13) mRNA (Genbank accession no. NM_008960). ASO 9 (18) and robust pharmacological activity (14,15). Furthermore, the and control oligonucleotide 12 (19) have been described selective inhibition of gene expression with MOE ASOs previously. elicits positive pharmacological activity in several animal MOE phosphoramidites were prepared as described models of human disease when given systemically with clini- previously (7,32,33). LNA and 2 -deoxyribonucleoside phos- cally relevant routes and schedules of administration (15–19). phoramidites were purchased from commercial suppliers. These beneficial properties have translated to human thera- Oligonucleotides were prepared similar to that described pre- peutics. For example, a MOE ASO targeting ApoB has viously (34) on either an Amersham AKTA 10 or AKTA 100 shown a dose-dependent reduction of target protein con- oligonucleotide synthesizer. Modifications from the reported current with lowering of LDL cholesterol. Doses as low as procedure include: a decrease in the detritylation time to 100 mg per week produced statistically significant reductions 1 min, as this step was closely monitored by UV analysis in ApoB protein, and a dose of 200 mg per week reduced for complete release of the trityl group; phosphoramidite con- serum ApoB protein by 50% (20). Furthermore, to date centration was 0.1 M; 4,5-dicyanoimidazole catalyst was MOE ASOs have an excellent safety record in human clinical used at 0.7 M in the coupling step; 3-picoline was used trials (20–22). instead of pyridine for the sulfurization step, and the time The improvement in potency of MOE ASOs has, in part, decreased from 3 to 2 min. The oligonucleotides were then been attributed to the increased affinity for target mRNA con- purified by ion-exchange chromatography on an AKTA ferred by the MOE modification. Although MOE provides a Explorer and desalted by reverse phase HPLC to yield modi- substantial improvement in affinity, bicyclic nucleoside fied oligonucleotides in 30–40% isolated yield, based on 0 0 modifications such as 2 ,4 -methylene bridged nucleic acids the loading of the 3 -base onto the solid support. Oligo- (23,24) commonly called ‘locked nucleic acid’ (LNA, nucleotides were characterized by ion-pair-HPLC-MS analysis Figure 1b) (25,26) have been shown to provide a further (IP-HPLC-MS) with an Agilent 1100 MSD system. The increase in affinity. LNA containing chimeric ASOs are just purity of the oligonucleotides was >90% (Supplementary entering human clinical trials, and have recently been shown Table S1). to inhibit growth in human tumor xenograft models (27). However, studies of the effects of LNA ASOs versus endoge- Cell culture assays nous targets in liver, a tissue where antisense effects have been extensively characterized, have not been reported to For determining potency in cell culture, mouse brain endo- date. In order to investigate whether LNA ASOs have the thelial (bEND) cells (American Type Culture Collection) potential to improve upon MOE based ASO therapeutics, were transfected with the indicated concentration of oligo- we directly compared the potency and therapeutic index of nucleotide for 4 h using 3 mg/ml Lipofectin in OptiMEM. several designs of MOE and LNA containing ASOs in cell Transfection mixes were then replaced with normal growth Nucleic Acids Research, 2007, Vol. 35, No. 2 689 0 0 media [DMEM-high glucose, 10% fetal bovine serum (FBS) primer (RP) 5 -TGAAGAGTCAGTGGCCGGTT-3 and probe containing penicillin-streptomycin]. Cells were harvested (PR) 5 -6FAM-TTTCTGTTCCACGGGCAGCTCGTAGT- 24 h later and RNA was purified using Qiagen 96-well TAMRA-3 . Primers used for determination of PTEN RNA RNeasy plates. RNA was analyzed for TRADD and level are as follows: FP 5 -ATGACAATCATGTTGCAGCA- 0 0 0 cyclophilin A RNA levels. TRADD RNA levels, normalized ATTC-3 ,RP5 CGATGCAATAAATATGCACAAATCA 3 to those of cyclophilin A, are expressed as percent untreated and PR 5 -6FAM-CTGTAAAGCTGGAAAGGGACGGACT- control (% UTC). Each treatment was performed in triplicate. GGT-TAMRA-3 . Primers used for determination of ApoB IC values were determined using GraphPad Prism software RNA level are as follows: forward primer (FP) 5 -GAAAAT- by fitting the data to a sigmoidal dose–response curve (vari- AGACTTCCTGAATAACTATGCATT-3 , reverse primer (RP) 0 0 0 able slope) using a defined top of 100% and bottom of 0%. 5 -ACTCGCTTGCCAGCTTGC-3 and probe (PR) 5 -6FAM- For caspase induction, A549 cells (American Type Culture TTTCTGAGTCCCCGTGCCCAACA-TAMRA-3 . Primers Collection), a human lung carcinoma cell line, were seeded used for determination of cyclophilin A RNA level are as 0 0 0 into 96-well plates and transfected the following day with follows: FP 5 -TCGCCGCTTGCTGCA-3 ,RP5 -ATCGGCC- 0 0 300 nm ASO for 4 h using Lipofectin (Invitrogen). Transfec- GTGATGTCGA-3 and PR 5 -6FAM-CCATGGTCAACCCC- tion mixes were subsequently replaced with normal growth ACCGTGTTC-TAMRA-3 . media (Ham’s F12K media containing 10% FBS). Cells were assayed 44 h later for caspase 3 activity and cell number Western blot analysis (CyQuant) as described previously (35). Frozen tissue samples were homogenized in RIPA buffer (PBS containing 1% NP-40, 0.5% sodium deoxycholate, Animal treatment 0.1% SDS) containing Complete protease inhibitors (Roche), and protein concentrations were determined by Bio- All animal experiments were conducted according to the Rad protein assay. Protein samples were separated on a 10% American Association for the Accreditation of Laboratory PAGE gel (Invitrogen) and subsequently transferred to a Animal Care guidelines and were approved by the Animal PVDF membrane (Invitrogen). Membranes were incubated Welfare Committee. Male Balb/c mice, aged 6–8 weeks, at room temperature in blocking buffer consisting of 5% were obtained from Charles River Laboratories. Compounds non-fat dry milk in TBS-T for 1 h. Rabbit polyclonal antibod- were suspended in phosphate-buffered saline (PBS), filter ies were obtained from commercial sources, and used at sterilized and administered by intraperitoneal (i.p.) injection 1:1000 dilution. Phospho-eIF2alpha (Ser51) antibody was according to the indicated dosing schedules in a volume cor- obtained from Cell Signaling Technology (Catalog no. responding to 10 ml/g animal weight. Animals were main- 9721). HRP conjugated anti-rabbit secondary antibodies tained at a constant temperature of 23 C and were allowed were obtained from Jackson ImmunoResearch and were standard lab diet and water ad libitum and animal weights used at 1:2500 dilution. Protein bands were visualized using were monitored prior to dosing throughout the live phase of ECL-plus reagent (Amersham). the study. Immediately prior to sacrifice, mice were anes- thetized with isoflurane and terminal bleed was performed by cardiac puncture. Plasma or serum was isolated from Capillary gel electrophoresis whole blood and analyzed for clinical chemistries. Alanine Immediately following removal, organs were frozen in liquid aminotransferase (ALT) and aspartate aminotransferase nitrogen and stored at 80 C until ASO extraction. Tissues (AST) levels were determined using an Olympus AU400e were weighed and ASO was extracted as described previously bioanalyzer. Immediately following terminal blood draw, (12,36). Briefly, tissue samples were homogenized in a mice were sacrificed by cervical dislocation while under BioSavant (Bio 101), and subjected to solid phase extraction anesthesia. In conjunction with necropsy, liver and spleen using a phenyl bonded SPE column (Isolute). Concentrations weights were determined. Effects of compounds on organ of ASO in tissue as well as metabolite profiles were deter- weights, normalized to body weight, are expressed relative mined by capillary gel electrophoresis (CGE) using a to those of the saline treated group. Beckman P/ACE model 5010 capillary electrophoresis unit. RNA analysis Hematoxylin and eosin (H&E) staining Tissues were homogenized in 4 M guanidine isothiocyanate, Tissue samples were fixed in formalin for a minimum of 25 mM EDTA, 50 mM Tris–HCl, pH 6, containing 1 M 1 day followed by incubation in 70% ethanol for a minimum b-mercaptoethanol immediately following sacrifice and of 1 day. Tissue samples were further dehydrated and pro- homogenized. RNA was extracted using RNeasy columns cessed using a Leica ASP300 tissue processor. Tissues were (Qiagen) according to manufacturer’s protocol. RNA was embedded in paraffin and 4 m sections were mounted on eluted from the columns with water. RNA samples were ana- positive charged glass slides. Deparaffinized and rehydrated lyzed by fluorescence-based quantitative RT–PCR using an samples were stained for hematoxylin and eosin (H&E), Applied Biosystems 7700 sequence detector. Levels of target using a Leica Autostainer XL. RNAs as well as those of cyclophilin A, a housekeeping gene, were determined. Target RNA levels were normalized Immunohistochemical staining to cyclophilin levels for each RNA sample. Primers used for determination of TRADD RNA level are as follows: Immunohistochemical studies were performed to detect the 0 0 forward primer (FP) 5 -GGCCGCCTGCCAGAC-3 , reverse cleaved form of caspase 3, as well as Bcl2-associated X 690 Nucleic Acids Research, 2007, Vol. 35, No. 2 Table 1. Oligonucleotide targets, sequences and chemistry a m 0 0 Structure code is nucleotide units 5 to 3 . Capital letter is base code: G ¼ guanine, A ¼ adenine, C ¼ cytosine, C ¼ 5-methylcytosine, T ¼ thymine. Small subscript are sugar and linkage codes: ¼ MOE sugar, ¼ LNA sugar, ¼ deoxyribose sugar, ¼ PS linkage, ¼ PO linkage. e l d s o protein (Bax) and growth arrest and DNA-damage-inducible RESULTS beta (GADD45b) protein. Formalin-fixed, paraffin-embedded Effect of LNA modification in cell culture tissue sections were mounted on positive charged glass slides. To directly compare the effects of LNA and MOE modifi- Deparaffinized and rehydrated samples were heated for cations, we designed a series of oligonucleotides targeting 20 min at 95 C in citrate buffer solution. The slides were mouse TRADD mRNA. We have previously published the cooled for 20 min and endogenous peroxidase was blocked with 3% hydrogen peroxide (H O ) in methanol for 10 min identification and characterization of MOE modified oligo- 2 2 at room temperature, followed by rinsing in distilled water. nucleotides targeting human TRADD, a death domain adapter To detect cleaved caspase 3, Bax and GADD45b, the sections protein that interacts with TNF receptor family members (28). The lead ASO from this study, a 4-10-4 MOE design, were incubated 1 h at RT with anti-cleaved caspase 3 poly- contains a single mismatch to murine TRADD mRNA, clonal antibody, (1:50 dilution; Cell Signaling Technology, which was corrected to provide the mouse TRADD ASO Danvers, MA), anti-Bax monoclonal rabbit antibody, E63 1a. Transfection of 1a (Table 1) into bEND cells reduces (1:200 dilution; Epitomics, Burlingame CA) and anti- TRADD mRNA levels in a concentration-dependent manner GADD45b polyclonal antibody, C18 (0.5 mg/ml; Santa as measured by quantitative RT–PCR (Figure 2 and Table 2), Cruz, CA), respectively. The HRP conjugated secondary to give an IC of 8 nM. The 4-10-4 LNA version of this antibodies were obtained from Jackson ImmunoResearch sequence (1b) reduced potency by 4-fold relative to the cor- Laboratories (1:200 dilution; West Grove, PA). The antigen responding MOE ASO. Because LNA has an increased affin- was visualized with DAB (Dako Cytomation—Cat3K3466) ity relative to MOE, we hypothesized that the shorter ASOs for 5 min. For negative controls, primary antibodies were with reduced LNA content may maintain or improve potency, replaced with isotype matched normal IgG. The slides were and thus prepared the 3-10-3 MOE and LNA ASOs 2a and then counterstained with hematoxylin. Nucleic Acids Research, 2007, Vol. 35, No. 2 691 Table 2. Activity and liver levels of MOE (a series) and LNA (b series) ASOs ASO IC (nM) Est. ED (mg/kg) Liver conc. 50 50 a b c in vitro in vivo (mg/g tissue) ± SD 1a 8.3 13 116 ± 20 2a >60 >25 54 ± 5.9 3a 8.5 11 92 ± 11 4a 27 9 96 ± 14 1b 35 13 2b 15 6 64 ± 6.1 3b 1.6 4 69 ± 4.5 4b 8.4 2 48 ± 8.9 5a 60 25 5b 60 25 IC values for reduction of TRADD mRNA in bEND cells after transfection with Lipofectin. ED values for reduction of TRADD mRNA in mouse liver after dosing twice weekly for three weeks estimated by interpolation. Determined by capillary gel electrophoresis for the 4.5 mmol/kg dose groups at end of study. which maintained potency relative to 1a. Further substitution of two additional MOE residues gave the 2-14-2 ASO 4a, which reduced potency of 3–4-fold. In contrast, replacing two LNA units of 1b with deoxy nucleosides resulted in a marked improvement, with 3b having an IC of 1–2 nM. As with the MOE series, further substitution to afford 4b reduced the potency, such that the LNA 2-14-2 4b was approximately equipotent relative to the parent 4-10-4 MOE 1a.As expected, both MOE (5a) and LNA (5b) four base mismatch control ASOs were inactive. In vivo effects of 2 -MOE and LNA modified oligonucleotides To determine if the behavior of LNA modified ASOs was similar in animals, mice were treated with ASOs at several dosage levels two times per week for 3 weeks. As expected, Figure 2. Reduction of TRADD mRNA in bEND cells after transfection with MOE ASO 1a or LNA ASO 1b (top panel) or MOE ASO 4a or LNA ASO 4b treatment of mice with the 4-10-4 MOE ASO 1a twice per (bottom panel). week reduced TRADD mRNA in liver in a dose-dependent manner, producing a 77% reduction at the 4.5 mmol/kg 2b. Interestingly, the MOE 16mer 2a had greatly reduced dose (30 mg/kg, Figure 3). Doses of mmol per kg were potency, whereas the potency of the 16mer LNA 2b was employed as the compounds were of slightly different molecular weight, and we wished to directly compare improved relative to the 18mer LNA ASO 1b. These results potency on a molar basis. Consistent with cell culture results, are consistent with a previous study of LNA ASOs (37), in the corresponding LNA analog 1b was less efficacious at this which the high affinity of LNA facilitated good target dose, resulting in only a 65% inhibition of mRNA. The reduction with 16mer LNA ASO designs. Since RNase H activity has been found to be an important potency of this ASO is difficult to assess from the limited factor in antisense potency both in vitro and in vivo (38), dose–response data, but it appears to have a reduced efficacy, optimization of the gap region is of crucial importance. at least at the doses evaluated. Also consistent with cell cul- Because LNA has been shown to alter the conformation of ture results, the MOE 16mer ASO 2a was weakly active, adjacent DNA nucleotides for several residues (39), and while the corresponding LNA 2b showed 80 and 60% duplex conformation is critical for RNase H activity (5), we reductions in TRADD mRNA at the 4.5 and 1.5 mmol/kg included ASOs having increased gap sizes (3 and 4) in our doses, respectively. SAR set. This replacement of high affinity modifications by The 3-12-3 and 2-14-2 MOE ASOs 3a and 4a produced a additional 2 -deoxy units could increase RNase H activity small increase in potency relative to the 4-10-4 1a, as well as by either adding potential cleavage sites, or by reducing the an increase in efficacy at the 4.5 mmol/kg dose, with 2-14-2 conformational transmission effects of LNA. We reasoned design giving an 89% reduction in mRNA. The LNA ASO that the increased affinity of LNA relative to MOE would com- 3b (3-12-3 design) showed an increase in potency, achieving pensate for the reduced number of 2 -modifications, and may 80% reduction of mRNA at the 1.5 mmol/kg dose. In con- increase potency. Replacing two internal MOE modifications trast to cell culture results, the 2-14-2 LNA ASO 4b further of 1a with deoxy nucleosides provided the 3-12-3 ASO 3a, improved activity, giving a 75% inhibition of TRADD 692 Nucleic Acids Research, 2007, Vol. 35, No. 2 Figure 3. Reduction of TRADD mRNA for MOE (a series) and LNA (b series) ASOs. *Not Analyzed: due to severe toxicity at 8 days, the 4.5 mmol/kg dose group of 4b was terminated early. expression at the lowest dose of 0.5 mmol/kg (3 mg/kg). showed severe hepatotoxicity by histopathological analysis, Interpolation of the TRADD mRNA reduction data allowed discussed further below. estimation of a rough ED , which could be converted to The ALT and AST levels for all the remaining groups mg/kg dosing for a comparison of in vivo potency at the termination of study on Day 21 indicate a striking (Table 2). For 3b, the data suggest an 3-fold increase in difference in the liver function profile of the MOE and LNA potency over the corresponding MOE 3a and its 4-10-4 par- ASOs (Figure 4). All LNA ASOs studied showed at least ent 1a. The 2-14-2 LNA gapmer design 4b provided an even a 10-fold increase in transaminases for the 4.5 mmol/kg larger increase in potency, the magnitude of which could not dose groups, and some showed increases of >100-fold such be estimated from this experiment because the maximal effect as 3b, 4b (despite receiving only 3/6 doses) and 5b. In con- was achieved at the lowest dose tested. The efficacy of inhibi- trast, the transaminase levels in mice treated with MOE ASOs tion of TRADD expression for the high-dose group could not were within the normal range. The observed hepatotoxicity be obtained, as the high-dose group of animals treated with in the LNA ASO series was both compound and dose- 4b was sacrificed early due to severe toxicity observed as dependent, with ASOs having a larger DNA gap increasing discussed below. Mismatch control MOE and LNA ASOs hepatotoxicity. Further evidence for the toxicity of the LNA 5a and 5b showed no reduction in mRNA, indicating a ASOs was evident from the organ weights (Supplementary specific in vivo antisense effect on the target mRNA. Figure S2). LNA ASOs produced large (45–62%, relative to saline) liver weight increases in a dose-dependent manner, except for 5b, which had a smaller (25%), non-dose- LNA ASOs induce profound hepatotoxicity dependent increase. In contrast, the MOE ASOs showed As part of the routine monitoring of animals during the study, much smaller (0–17%) liver weight increases. Increase in plasma bilirubin and transaminase levels for all the high-dose spleen weight was more variable, with 2b and 3b producing groups were examined at Day 8 of the study in order to assess large (>100%, relative to saline) increases. The increase in liver function. Bilirubin, ALT and AST levels were within the spleen weights could be a direct proinflammatory effect of normal range for all animals, except for the 2-14-2 LNA the oligonucleotide or secondary to the severe hepatotoxicity. ASOs 4b and 5b. LNA 4b showed elevations of 186-, 75- and 18-fold for ALT, AST and bilirubin, respectively, relative LNA ASOs induce apoptosis in vitro and in vivo to saline treated animals. Notably, even the control LNA 5b, which showed no TRADD mRNA reduction, showed 46- and Examination of H&E stained liver sections from mice treated 25-fold increases in ALT and AST, respectively. Prior to with LNA ASOs 1b and 4b, as well as control oligo- receiving the fourth dose of ASO on Day 11 of the study, nucleotide 5b, confirmed hepatotoxic events. Histopathological all the animals in the 4.5 mmol/kg group receiving LNA observations included signs of apoptosis, profound eosinophilic ASO 4b experienced significant weight loss, losing 25% cytoplasmic degeneration with glycogen depletion and of their body weight (Supplementary Figure S1). Because hyperchromatic nuclei, as well as centrilobular coagulative of this severe weight loss, coupled with the very large necrosis surrounded with inflammatory infiltrate containing transaminase increases seen with 4b, the study was termi- neutrophils, monocytes and lymphocytes (Figure 5). Lesions of nated early for this dose group. Upon necropsy, these animals intracytoplasmic microvesicular changes were also visualized Nucleic Acids Research, 2007, Vol. 35, No. 2 693 Figure 4. Plasma transaminase levels for MOE (a series) and LNA (b series) ASOs. **Data from 8 days. AST 4020 ± 850, ALT 6470 ± 1450, and severe weight loss led to early termination of 4.5 mmol/kg dose group of 4b. in liver sections of 4b and 5b treated animals, suggesting early neoantigen formed by caspase cleavage of cytokeratin 18. signs of steatosis. An increased number of mitotic hepatocytes M30 immunoreactive cells appear at an early stage of apoptosis were also observed in liver sections from LNA treated animals, in epithelial cells, and are not detectable in vital or necrotic likely indicating regeneration of damaged tissue. epithelial cells (40). Animals treated with LNA ASOs 1b, 4b To further characterize the hepatotoxicity associated with or 5b all showed greatly increased M30 staining relative to LNA ASO treatment, we conducted immunohistopathological saline treated animals (Figure 5), providing further evidence evaluation to characterize the type of toxicity observed for apoptosis induced by treatment with the LNA ASOs. (Figure 5 and Supplementary Table S2). Liver sections To examine whether LNA ASOs could induce apoptosis in from animals treated with LNA ASOs were stained for the cell culture, we examined representative TRADD ASOs 2a, expression of GADD45b, the activated form of caspase 3, 4a at 300 nM in A549 cells for their ability to induce caspase peroxisome membrane protein 70 and Bax. Increased expres- 3 activity, a common marker of apoptosis, in cell culture sion of activated caspase 3 and Bax suggested increased after transfection (35). The LNA ASOs 2b and 4b resulted apoptosis in livers of LNA, but not MOE, oligonucleotide in a 3.1- and 6.2-fold induction of caspase 3 activity relative treated mice. GADD45b is a p53 and NF-kB regulated to control ASO 12, respectively, whereas the corresponding gene induced in response to cell stress. The LNA but not MOE ASOs gave no change relative to the control ASO MOE modified oligonucleotides increased GADD45b expres- (Supplementary Figure S3). sion dramatically. Evidence of peroxisome proliferation was also suggested by increased expression of peroxisome mem- LNA ASO improves potency but also increases toxicity brane protein 70 by IHC. Increased staining of each of these markers was dose-dependent and correlated well with ALT To verify the activity and toxicity observed with 4b,it increases. To further confirm apoptosis involvement of liver was tested in a repeat set of experiments along with 1a as a injury mediated by LNA, liver sections of saline and LNA comparator (Figure 6). In addition to the previously utilized treated animals were stained with the monoclonal antibody doses of 4.5, 1.5 and 0.5 mmol/kg, we also examined 0.9, M30, which is an apoptosis marker that monitors the 0.3 and 0.1 mmol/kg doses of 4b in order to assess to what 694 Nucleic Acids Research, 2007, Vol. 35, No. 2 Figure 5. Histopathology of liver sections from mice treated MOE (a series) and LNA (b series) ASOs at 4.5 mmol/kg twice weekly for 3 weeks. Livers from animals treated with LNA ASOs 1b, 4b and 5b present with significant hepatotoxicities as demonstrated in (A), a routing H&E stain showing profound swollen eosinophilic degeneration, cell death (right arrow) and hyperchromatic nuclei (up arrow) of hepatocytes. The immunohistochemistry reveals that the injured hepatocytes appear with cytoplasmic staining (brown stain cells, up arrows) of cleavage caspase 3 (B), pro-apoptotic protein BAX (C) and M30 (D), a neo- epitope generated in epithelial cells as a result of caspase activation (cleavage). In addition a DNA damage and repair associated protein GADD45b (E) and peroxisome membrane protein PMP70 (Supporting Supplementary Figure S2) were both found to have increased cytoplasmic expression in LNA treated livers. IgG control slides of all four IHC markers were negative (data not shown). Nucleic Acids Research, 2007, Vol. 35, No. 2 695 Figure 6. Transaminases (bar graph, left scale) and reduction of TRADD mRNA (points on line graph, right scale) after treatment with the 4-10-4 MOE gapmer 1a or the 2-14-2 LNA gapmer 4b. extent potency was increased, as well as whether there was an transaminase levels, organ weights, body weights or increase in therapeutic index. In this experiment, mice treated histopathological analysis of liver tissue samples. Thus, with three doses of 4b at 4.5 mmol/kg (at 11 days) showed though LNA ASO 4b was more potent than the comparator minimal weight loss, and the study was continued for the MOE ASOs, it was not more efficacious, and furthermore course of six administrations. At termination of the study, did not produce maximal efficacy in the absence of observ- the animals treated with 4b had lost 10% of their body able toxicity. These results demonstrate that despite the weight, as opposed to a 10% gain in the control group (Sup- increase in potency observed with the LNA ASOs, the thera- plementary Figures S4 and S5). Furthermore, organ weights, peutic index is not increased, and is probably decreased relat- transaminase increases and histopathological observations ive to the corresponding MOE ASOs. upon necropsy were consistent with the previous results, It is possible that both the improved potency and increased with >50-fold increases in AST and ALT at the 1.5 and toxicity could be due to increased distribution of LNA modi- 4.5 mmol/kg dose levels. The target reduction for 4b observed fied oligonucleotides to liver. To determine if this was the was similar to the first experiment, with 70 ± 8, 76 ± 6 and case, the concentration of ASO in the liver was measured 66 ± 8% reduction of TRADD mRNA at the 4.5, 1.5 and at the conclusion of the study for ASOs 1–4 (Table 2). The 0.5 mmol/kg doses, respectively. The 0.9 mmol/kg dose repeat group of 4b treated animals was used for this analysis, (5 mg/kg) produced the maximal effect of 4b, providing as the mice received the same number of total doses. The an 81 ± 4% reduction of TRADD mRNA, but also appeared liver concentrations ranged from a high of 116 mg/g for 1a to be above the maximum non-toxic dose, as transaminases to a low of 48 mg/g for 4b. These results are consistent were increased 5-fold. with expectations of higher metabolic stability for ASOs The 2-14-2 LNA 4b clearly demonstrated a dose- having greater numbers of MOE and LNA modifications, dependent reduction of TRADD mRNA, with an ED of and also with increased distribution to liver for ASOs having 0.37 mmol/kg (corresponding to 2 mg/kg, Table 2). This more PS linkages. As the concentration trend of ASO in liver suggests an improvement in potency of 4–5-fold relative does not correlate with potency or toxicity, the increased to the corresponding 2-14-2 MOE 4a, and 6–7-fold relative potency and/or toxicity of LNA oligonucleotides is not due to the comparator 4-10-4 MOE gapmer 1a. At dosage levels to more accumulation in the liver. used for the previous study, the activity of 1a was essentially LNA ASO 4b shows hepatotoxicity with a single identical to that observed previously. A higher dose of administration 6.2 mmol/kg (40 mg/kg) of 1a produced a larger reduction (86 ± 3%) of TRADD mRNA. At all doses of 1a employed, To better characterize the nature of the hepatotoxicity, we there was no evidence for toxicity as measured by administered a single dose of either MOE ASO 1a or LNA 696 Nucleic Acids Research, 2007, Vol. 35, No. 2 Figure 7. Transaminases (bar graph, left scale), and reduction of ApoB mRNA (points on line graph, right scale) after treatment with MOE (a series) and LNA (b series) ASOs. ASO 4b to mice, and examined effects at 2 and 4 days post- mRNA in mice treated with 20mer MOE ASOs. Three active dosing. In addition to plasma transaminases and histopatho- ASO sequences (6, 8 and 9, Table 1) targeting either mouse logy, levels of cleaved caspase 3 and phosphorylated eIF2a ApoB or mouse PTEN were identified in cell culture assays (p-eIF2a) were examined by western blot. Phosphorylation using methods previously described for these targets. Because of eIF2a has been shown to mediate apoptosis, presumably the LNA ASO design having two LNA nucleosides flanking a via the inhibition of translation (41). Additionally, levels of large deoxy gap region appeared to exhibit the greatest Bax, GADD45b, PUMA, p53, TNFa and MDM2 mRNA increase in potency and hepatotoxicity, we utilized a 2-16-2 were evaluated by RT–PCR. For all groups except for the design, and applied it to the previously identified 5-10-5 12 mmol/kg dosage level of 4b, there was no discernable dif- MOE sequences. ference from saline for any measured endpoint. In contrast, For the ApoB target, ASOs were dosed at 2.5 and either administration of 12 mmol/kg of the LNA ASO 4b gave a 0.5 or 0.4 mmol/kg twice weekly for 3 weeks (Figure 7). 7- and 2-fold elevation in ALT and AST, respectively, at The 5-10-5 MOE ASO 6 produced a modest 24% inhibition the day 4, but not at the day 2 time-point. Concurrent with of ApoB mRNA at the 0.4 mmol/kg dose and an 84% inhibi- this increase in transaminases at day 4, an increase in tion at the 2.5 mmol/kg dose. The 2-16-2 MOE ASOs 7a p-eIF2a was observed by western blot (Supplementary and 8a gave no inhibition at the low dose, and produced Figure S6). No increases in other genes studied were evident 50% reduction in ApoB-100 mRNA at the 2.5 mmol/kg by RT–PCR or western blot. However, weak staining of acti- (18 mg/kg) dose. The potency increase of the LNA ASOs vated caspase 3, Bax and the M30 neoantigen were observed was variable, with LNA 7b providing a 5-fold estimated by immunohistochemistry (Supplementary Table S3). These increase in potency over the corresponding 2-16-2 MOE 7a, results indicate that the hepatotoxicity induced by LNA 4b but not the parent 5-10-5 MOE ASO 6. LNA 8b was only can occur in as little as 4 days, and are consistent with marginally more active than MOE 8a, and substantially acute liver injury caused by induction of apoptosis. less active than the MOE 6. None of the MOE ASOs showed evidence for toxicity as measured by organ weights and serum transaminase increases. In contrast the LNA ASOs LNA effects on potency and hepatotoxicity are 7b and 8b both resulted in >20-fold increases in AST and independent of target ALT, along with increased organ weights at the 2.5 mmol/ To help rule out a potential target related contribution to the kg dose. hepatotoxicity and to verify the improved potency for other ASOs targeting murine PTEN were dosed at 0.083– targets, we tested additional LNA ASOs targeting other 2.25 mmol/kg twice weekly for 3 weeks (Figure 8). The mouse genes. We have previously published data demonstrat- parent MOE 5-10-5 ASO 9 as well as its gap widened coun- ing specific reduction of mouse ApoB (19) and PTEN (18) terpart 10a reduce target mRNA in a dose-dependent manner Nucleic Acids Research, 2007, Vol. 35, No. 2 697 Figure 8. Transaminases (bar graph, left scale), and reduction of PTEN mRNA (points on line graph, right scale) after treatment with MOE (a series) and LNA (b series) ASOs. LNA ASO shows hepatotoxicity in rats without evidence for toxicity as measured by transaminase As the sequence of the PTEN ASOs was homologous to levels, organ weights and body weights. The LNA ASO rat, we were able to examine if LNA 10b was more potent 10b was significantly more potent (estimated 5–10-fold) in rat, and importantly if the toxicity observed in mouse trans- than either MOE ASO. However, once again, this increase lated to another species. Accordingly, rats were treated with in potency correlated with an increase in hepatotoxicity. ASOs 9, 10b, or control 12 at dose of 0.83, 2.5 or The 0.75 mmol/kg dose group started to show mild transami- 7.5 mmol/kg twice weekly for 3 weeks (Supplementary nase elevations, while the higher dose group resulted in large Figures S7–S9). The MOE ASO 9 showed a dose-dependent (>50-fold) increases in both AST and ALT, increases in liver reduction of PTEN mRNA in rat liver, with an ED of and spleen weights and caused significant weight loss in trea- 2.5 mmol/kg. In contrast to the mouse data, the LNA ted animals. A non-targeted control 5-10-5 MOE ASO 12 had ASO 10b was only slightly more potent in rat liver, with no effect on either target or toxicity measures. an estimated ED of 1.5 mmol/kg. However, the LNA ASO was also hepatotoxic in rats. At the 7.5 mmol/kg dose, LNA ASO hepatotoxicity is not likely due to body weights were decreased 25% relative to control ASO LNA degradation products 12, AST (but not ALT) was increased 5- and 10-fold, respec- To help determine if the increased toxicity of the LNA ASOs tively in two of the four animals, and bilirubin was increased was due to the oligonucleotide or to degradation products, we dramatically in the same two animals. Histopathological prepared mixed backbone versions of the PTEN 2-16-2 ASOs evaluation of H&E stained liver sections confirmed hepato- containing phosphodiester (PO) linkages between LNA toxicity, showing moderate eosinophilic cytoplasmic degene- ration with focal single cell apoptosis and mild mononuclear nucleosides as well as at the LNA/DNA junction (Table 1). cell infiltration (Supplementary Table S4). No hepatotoxicity These ASOs are much more rapidly metabolized in vivo, was observed by histopathological evaluation, ALT, AST, and presumably will release either free MOE or LNA nucleo- bilirubin, organ weights or body weights for either of the sides or nucleotides. If these nucleosides or nucleotides were MOE ASOs. This data suggest that the potency increase of the source of the observed hepatotoxicity, these ASOs should LNA ASOs relative to MOE ASOs is more pronounced in be more toxic, while if the intact ASO was causing the mouse than in rat, though the observed hepatotoxicity is hepatotoxicity, the mixed backbone versions should be less still present in rat. toxic. Neither the MOE (11a) nor LNA (11b) ASO caused significant target reduction at the doses tested. Importantly, the LNA containing ASO showed no evidence of hepatotoxi- city, and analysis of drug levels in liver confirmed near com- DISCUSSION plete metabolism of the intact drug 11b. These results suggest that the intact LNA oligonucleotides are responsible for the The main goal of our study was to determine if LNA contain- observed hepatotoxicity. ing ASOs would improve potency and therapeutic index 698 Nucleic Acids Research, 2007, Vol. 35, No. 2 relative to the current generation of MOE ASOs. Our assump- increased Bax expression and increased expression of the tion entering the work was that an improvement in potency M30 neo-epitope. The upregulation of the pro-apoptotic pro- would yield an improved therapeutic index, since it has tein Bax suggests involvement of the p53-mediated apoptosis been generally believed that many of the toxicities of ASOs pathway, as Bax is a key response gene to p53 activation are due to class effects as a result of the PS backbone. How- (42,43). Furthermore, GADD45b, a key downstream target ever, this proved not to be the case with the LNA ASOs gene of p53 during DNA damage and repair process studied. (44), was highly up-regulated in the injured hepatocytes. Our results clearly demonstrate the ability to improve GADD45b appears to help protect cells against programmed potency with some, but not all, LNA containing ASO designs, cell death through blocking the c-jun N-terminal kinase cas- particularly for the TRADD and PTEN targets. This improve- cade, and is probably induced in response to cellular damage. ment was occasionally fairly large, as much as 5–10-fold, and It is unclear why LNA oligonucleotides cause this level of was most pronounced for LNA ASOs of length 18–20 nt hepatotoxicity though the corresponding MOE oligo- which contained 2–3 LNA residues at each end. As little as nucleotides do not. One possibility is that antisense effects 0.75–1 mmol/kg (5–6 mg/kg) of these ASOs given twice on genes partially complementary to the hepatotoxic ASOs weekly for 3 weeks reduced target mRNA by 80%. The opti- are playing a role in the toxicity, as LNA ASOs have been mal LNA ASO design in vivo appeared to be different than shown to decrease the selectivity for a perfectly complemen- that observed in cell culture, where we found that two LNA tary target relative to the corresponding MOE ASOs (45). nucleotides on each end of the ASO provided the largest However, this seems unlikely because multiple unrelated potency increase. This is evident from a comparison of 3b LNA sequences cause similar toxicities. All oligonucleotides and 4b, where 3b was 5-fold more potent in cell culture, were prepared and purified in the same laboratory using the but less potent in vivo. It is unlikely that the improved identical methods. The impurity profiles of LNA and MOE potency is due solely to increased affinity of the ASO for tar- ASOs were nearly identical, and contained only the expected get RNA, as adding more LNA to the ASO actually decreased impurities resulting from PS oligonucleotide synthesis (PO, potency both in cell culture and in vivo (compare 2b and 3b N-1, etc.). This makes it extremely unlikely that the toxicity with 1b). Because of these trends, combined with the lack of is due to impurities resulting from LNA, but not MOE increased distribution of LNA ASOs to liver, it is likely that oligonucleotide synthesis. Additionally, since the metaboli- other factors are contributing to the increased potency of cally unstable PO containing ASO 11b (which should be LNA modified ASOs observed in our studies. Additional metabolized in vivo to LNA nucleosides and nucleotides) investigations will be required to further characterize the was non-toxic, it is not likely that the toxicity is due to the nature of this potency improvement. LNA monomers. This suggests that the intact LNA contain- Unfortunately, the increased activity of LNA containing ing PS oligonucleotides are responsible for the observed toxi- ASOs was also accompanied by the observation of severe city. There are distinct structural differences between MOE hepatotoxicity, such that there was little or no separation and LNA which may allow LNA containing oligonucleotides between toxic doses and those that produced significant to selectively effect hepatotoxicity. Perhaps importantly, the levels of mRNA reduction. Hepatotoxicity was chemistry-, rigid acyclic 2 -methoxyethyl side chain of MOE protects sequence- and design-dependent, as it was only observed the corresponding 3 -phosphorothioate linkage from inter- with LNA containing ASOs, and the onset occurred at actions via increased steric bulk and hydration (8), relative slightly different dose levels for different compounds. The to the compact and more hydrophobic cyclic structure of fact that the MOE ASOs in some cases (compare 2a with LNA. This could cause selective binding affinity differences 2b, and 6 with 7b and 8b) produced similar reductions in tar- between MOE and LNA oligonucleotides for as yet unknown get RNA without producing observable toxicity suggest that macromolecular binding partners, and/or result in differential the toxicity is not secondary to reduction in target gene compartmentalization of the two classes of oligonucleotides expression. This is further supported by the observation of within liver tissue. severe hepatotoxicity with control 5b, which has >3 mis- The mild hepatotoxicity induced 4 days after a single matches to all known mouse sequences. Hepatotoxicity also administration of LNA ASO 4b occurred concurrently with seemed to be the most severe for the more potent LNA apoptosis and activation of Bax and caspase 3 in hepatocytes ASO designs regardless of target (see 4b, mismatch 5b, 7b as evidenced by histopathological evaluation. Furthermore, and 10b). Thus, therapeutic index was not improved, and an increase in p-eIF2a was observed to coincide with the was likely decreased relative to the MOE ASOs. onset of toxicity. Phosphorylated eIF2a inhibits translation The hepatotoxicity was evident from the observation of initiation, and has been shown to mediate apoptosis, multiple parameters, including histopathological evaluation possibly by preventing the synthesis of short lived anti- of liver tissue upon necropsy as well as large increases in apoptotic factors (41,46). There are four known kinases plasma levels of aminotransferases (ALT and AST). Further- which phosphorylate eIF2a: PKR, which is activated by more, the toxicity was commonly accompanied by large binding of double stranded RNA (dsRNA); GCN2, which is increases in liver and/or spleen weights, likely as a con- activated by amino acid deprivation; HRI, which is activated sequence arising from a response to hepatic injury induced by low heme levels; and PERK, which responds to stress by the LNA ASOs. In several cases, the toxicity was severe in the endoplasmic reticulum. It is unclear from our data enough to cause extensive weight loss in the animals. Histo- how treatment with hepatotoxic LNA oligonucleotides logy data clearly demonstrated both LNA oligonucleotide- results in increased phosphorylation of eIF2a; however, it indiced liver necrosis and activation of apoptosis pathways, is tempting to speculate that PKR could be involved. PKR as evidenced by H&E staining, as well as cleaved caspase 3, is activated by binding of dsRNA to distinct dsRNA binding Nucleic Acids Research, 2007, Vol. 35, No. 2 699 domains which serve as allosteric inhibitors of the kinase has reactivated or is persistently active despite other therapies in patients with AIDS. Am. J. Ophthalmol., 133, 475–483. domain (47,48). It remains to be determined if LNA oligo- 2. Eckstein,F. (2000) Phosphorothioate oligodeoxynucleotides: what is nucleotides interact with PKR, and more extensive mechanis- their origin and what is unique about them? Antisense Nucleic Acid tic work is required to confirm both p-eIF2a mediation of an Drug Dev., 10, 117–121. acute apoptotic response in hepatocytes, and elucidate the 3. Stein,C.A., Subasinghe,C., Shinozuka,K. and Cohen,J.S. (1988) Physicochemical properties of phosphorothioate precise biochemical mechanism responsible for the observed oligodeoxynucleotides. Nucleic Acids Res., 16, 3209–3221. hepatotoxicity. 4. Geary,R.S., Yu,R.Z. and Levin,A.A. (2001) Pharmacokinetics of In conclusion, we have shown that optimization of size phosphorothioate antisense oligodeoxynucleotides. Curr. Opin. and gap configuration employing LNA containing ASOs Investig. Drugs, 2, 562–573. can improve potency in mouse liver. However, this potency 5. Lima,W.F., Nichols,J.G., Wu,H., Prakash,T.P., Migawa,M.T., Wyrzykiewicz,T.K., Bhat,B. and Crooke,S.T. (2004) Structural increase is strongly correlated with the onset of a severe requirements at the catalytic site of the heteroduplex substrate for hepatotoxicity not seen with the corresponding MOE ASOs, human RNase H1 catalysis. J. Biol. Chem., 279, 36317–36326. and therapeutic index is not improved. These results suggest 6. Monia,B.P., Lesnik,E.A., Gonzalez,C., Lima,W.F., McGee,D., that while LNA ASOs have the potential to improve potency Guinosso,C.J., Kawasaki,A.M., Cook,P.D. and Freier,S.M. (1993) 0 0 Evaluation of 2 -modified oligonucleotides containing 2 -deoxy gaps as of antisense therapeutics, they impose a significant risk of antisense inhibitors of gene expression. J. Biol. Chem., 268, hepatotoxicity. The toxicities observed herein also pose chal- 14514–14522. lenges for interpreting target validation and pharmacology 7. Martin,P. (1995) Ein neuer Zugang zu 2-O-Alkylribonucleosiden und experiments using LNA ASOs in rodent models. It is possible Eigenschaften deren Oligonucleotide. Helv. Chim. Acta, 78, 486–504. more extensive screening efforts may identify less toxic LNA 8. Teplova,M., Minasov,G., Tereshko,V., Inamati,G.B., Cook,P.D., Manoharan,M. and Egli,M. (1999) Crystal structure and improved sequences, or that other ASO designs utilizing LNA may not antisense properties of 2 -O-(2-methoxyethyl)-RNA. Nature Struct. elicit hepatotoxicity. Additionally, bicyclic nucleosides simi- Biol., 6, 535–539. lar to LNA such as amino-LNA, thio-LNA (27) and ENA 9. McKay,R.A., Miraglia,L.J., Cummins,L.L., Owens,S.R., Sasmor,H. (which contains one extra methylene unit in the bicyclic and Dean,N.M. (1999) Characterization of a potent and specific class of antisense oligonucleotide inhibitor of human protein kinase C-alpha bridge) (49) have been reported to show antisense effects expression. J. Biol. Chem., 274, 1715–1722. similar to LNA and MOE ASOs. Although optimized designs 10. Altmann,K.H., Martin,P., Dean,N.M. and Monia,B.P. (1997) Second of one or more of these modifications may ultimately improve generation antisense oligonucleotides—inhibition of pkc-alpha and potency while maintaining the safety profile of MOE ASOs, c-raf kinase expression by chimeric oligonucleotides incorporating 0 0 no detailed characterization of their toxicological properties 6 -substituted carbocyclic nucleosides and 2 -O-ethylene glycol substituted ribonucleosides. Nucleosides Nucleotides, 16, 917–926. has yet been reported. It is clear that additional structure– 11. Geary,R.S., Watanabe,T.A., Truong,L., Freier,S., Lesnik,E.A., activity and structure–toxicity relationship studies will be Sioufi,N.B., Sasmor,H., Manoharan,M. and Levin,A.A. (2001) required to more fully assess the therapeutic potential of 0 Pharmacokinetic properties of 2 -O-(2-methoxyethyl)-modified LNA and other bicyclic nucleoside modified oligo- oligonucleotide analogs in rats. J. Pharmacol. Exp. Ther., 296, 890–897. nucleotides. Ultimately, the value of LNA oligonucleotide 12. Yu,R.Z., Geary,R.S., Monteith,D.K., Matson,J., Truong,L., Fitchett,J. drugs will be determined in human clinical trials, where and Levin,A.A. (2004) Tissue disposition of 2 -O-(2-methoxy) ethyl they are currently being developed for the treatment of modified antisense oligonucleotides in monkeys. J. Pharm. Sci., 93, cancer. Our results are based only on studies in rodents; how- 48–59. ever, caution should be exercised in the clinical development 13. Geary,R.S., Yu,R.Z., Watanabe,T., Henry,S.P., Hardee,G.E., Chappell,A., Matson,J., Sasmor,H., Cummins,L. and Levin,A.A. (2003) of LNA modified oligonucleotides, especially for chronic, Pharmacokinetics of a tumor necrosis factor-alpha phosphorothioate non-life threatening indications. 2 -O-(2-methoxyethyl) modified antisense oligonucleotide: comparison across species. Drug Metab. Dispos., 31, 1419–1428. 14. Zhang,H., Cook,J., Nickel,J., Yu,R., Stecker,K., Myers,K. and SUPPLEMENTARY DATA Dean,N.M. (2000) Reduction of liver Fas expression by an antisense oligonucleotide protects mice from fulminant hepatitis. Supplementary Data are available at NAR Online. Nat. Biotechnol., 18, 862–867. 15. Yu,R.Z., Zhang,H., Geary,R.S., Graham,M., Masarjian,L., Lemonidis,K., Crooke,R., Dean,N.M. and Levin,A.A. (2001) ACKNOWLEDGEMENTS Pharmacokinetics and pharmacodynamics of an antisense phosphorothioate oligonucleotide targeting Fas mRNA in mice. The authors thank Charles Allerson, Amy Dan and Rick Carty J. Pharmacol. Exp. Ther., 296, 388–395. for synthesis of oligonucleotides, Megan Pearce for determi- 16. Liang,Y., Osborne,M.C., Monia,B.P., Bhanot,S., Watts,L.M., She,P., Decarlo,S.O., Chen,X. and Demarest,K. (2005) Antisense nation of ASO levels in liver tissue, Denis Drygin for the oligonucleotides targeted against glucocorticoid receptor reduce caspase 3 activation data, Richard H. Griffey for numerous hepatic glucose production and ameliorate hyperglycemia in diabetic thought provoking discussions and Stanley T. Crooke and mice. Metabolism, 54, 848–855. Arthur A. Levin for a critical read of the manuscript. Funding 17. Zinker,B.A., Rondinone,C.M., Trevillyan,J.M., Gum,R.J., Clampit,J.E., to pay the Open Access publication charges for this article Waring,J.F., Xie,N., Wilcox,D., Jacobson,P., Frost,L. et al. (2002) PTP1B antisense oligonucleotide lowers PTP1B protein, normalizes was provided by Isis Pharmaceuticals. blood glucose, and improves insulin sensitivity in diabetic mice. Proc. Natl Acad. Sci. USA, 99, 11357–11362. Conflict of interest statement. None declared. 18. Butler,M., McKay,R.A., Popoff,I.J., Gaarde,W.A., Witchell,D., Murray,S.F., Dean,N.M., Bhanot,S. and Monia,B.P. (2002) Specific inhibition of PTEN expression reverses hyperglycemia in diabetic REFERENCES mice. Diabetes, 51, 1028–1034. 19. Crooke,R.M., Graham,M.J., Lemonidis,K.M., Whipple,C.P., Koo,S. 1. Vitravene Study Group (2002) Randomized dose-comparison studies of and Perera,R.J. (2005) An apolipoprotein B antisense oligonucleotide intravenous fomiversen for treatment of cytomegalovirus retinitis that 700 Nucleic Acids Research, 2007, Vol. 35, No. 2 lowers LDL cholesterol in hyperlipidemic mice without causing phosphorothioate RNA utilizing phenylacetyl disulfide (PADS). hepatic steatosis. J. Lipid Res., 46, 872–884. Bioorg. Med. Chem. Lett., 16, 2513–2517. 20. Kastelein,J.J.P., Wedel,M.K., Baker,B.F., Su,J., Bradley,J.D., Yu,R.Z., 35. Drygin,D., Barone,S. and Bennett,C.F. (2004) Sequence-dependent Chuang,E., Graham,M.J. and Crooke,R.M. (2006) Potent reduction of cytotoxicity of second-generation oligonucleotides. Nucluic Acids Res., apolipoprotein B and LDL cholesterol by short-term administration of 32, 6585–6594. an antisense inhibitor of apolipoprotein B. Circulation, 114, 36. Leeds,J.M., Graham,M.J., Truong,L. and Cummins,L.L. (1996) 1729–1735. Quantitation of phosphorothioate oligonucleotides in human plasma. 21. Chi,K.N., Eisenhauer,E., Fazli,L., Jones,E.C., Goldenberg,S.L., Anal. Biochem., 235, 36–43. Powers,J., Tu,D. and Gleave,M.E. (2005) A phase I pharmacokinetic 37. Frieden,M., Christensen,S.M., Mikkelsen,N.D., Rosenbohm,C., and pharmacodynamic study of OGX-011, a 2 -methoxyethyl antisense Thrue,C.A., Westergaard,M., Hansen,H.F., Orum,H. and Koch,T. oligonucleotide to clusterin, in patients with localized prostate cancer. (2003) Expanding the design horizon of antisense oligonucleotides with J. Natl Cancer Inst., 97, 1287–1296. alpha-l-LNA. Nucleic Acids Res., 31, 6365–6372. 22. Sewell,K.L., Geary,R.S., Baker,B.F., Glover,J.M., Mant,T.G., Yu,R.Z., 38. Wu,H., Lima,W.F., Zhang,H., Fan,A., Sun,H. and Crooke,S.T. (2004) Tami,J.A. and Dorr,F.A. (2002) Phase I trial of ISIS 104838, a Determination of the role of the human RNase H1 in the pharmacology 2 -methoxyethyl modified antisense oligonucleotide targeting tumor of DNA-like antisense drugs. J. Biol. Chem., 279, 17181–17189. necrosis factor-alpha. J. Pharmacol. Exp. Ther., 303, 1334–1343. 39. Bondensgaard,K., Petersen,M., Singh,S.K., Rajwanshi,V.K., Kumar,R., 23. Obika,S., Nanbu,D., Hari,Y., Morio,K.-I., In,Y., Ishida,T. and Wengel,J. and Jacobsen,J.P. (2000) Structural studies of LNA:RNA 0 0 Imanishi,T. (1997) Synthesis of 2 -O-4 -C-methyleneuridine duplexes by NMR: conformations and implications for RNase H and -cytidine. Novel bicyclic nucleosides having a fixed C3 -endo activity. Chem.-Eur. J., 6, 2687–2695. sugar puckering. Tetrahedron Lett., 38, 8735–8738. 40. Grassi,A., Susca,M., Ferri,S., Gabusi,E., D’Errico,A., Farina,G., 24. Obika,S., Nanbu,D., Hari,Y., Andoh,J.-I., Morio,K.-I., Doi,T. and Maccariello,S., Zauli,D., Bianchi,F.B. and Ballardini,G. (2004) Imanishi,T. (1998) Stability and structural features of the duplexes Detection of the M30 neoepitope as a new tool to quantify liver containing nucleoside analogs with a fixed N-type conformation, apoptosis: timing and patterns of positivity on frozen and 0 0 2 -O-4 -C-methyleneribonucleosides. Tetrahedron Lett., 39, paraffin-embedded sections. Am. J. Clin. Pathol., 121, 5401–5404. 211–219. 0 0 25. Wengel,J. (1999) Synthesis of 3 -C- and 4 -C-branched 41. Scheuner,D., Patel,R., Wang,F., Lee,K., Kumar,K., Wu,J., Nilsson,A., oligodeoxynucleotides and the development of locked nucleic acid Karin,M. and Kaufman,R.J. (2006) Double-stranded RNA-dependent (LNA). Acc. Chem. Res., 32, 301–310. protein kinase phosphorylation of the alpha-subunit of eukaryotic 26. Koshkin,A.A., Nielsen,P., Meldgaard,M., Rajwanshi,V.K., Singh,S.K. translation initiation factor 2 mediates apoptosis. J. Biol. Chem., 281, and Wengel,J. (1998) LNA (locked nucleic acid): an RNA mimic 21458–21468. forming exceedingly stable LNA:LNA duplexes. J. Am. Chem. Soc., 42. Chipuk,J.E., Kuwana,T., Bouchier-Hayes,L., Droin,N.M., 120, 13252–13253. Newmeyer,D.D., Schuler,M. and Green,D.R. (2004) Direct activation 27. Fluiter,K., Frieden,M., Vreijling,J., Rosenbohm,C., De Wissel,M.B., of Bax by p53 mediates mitochondrial membrane permeabilization and Christensen,S.M., Koch,T., Orum,H. and Baas,F. (2005) On the in vitro apoptosis. Science, 303, 1010–1014. and in vivo properties of four locked nucleic acid nucleotides 43. Mihara,M., Erster,S., Zaika,A., Petrenko,O., Chittenden,T., Pancoska,P. incorporated into an anti-H-Ras antisense oligonucleotide. Chem. and Moll,U.M. (2003) p53 has a direct apoptogenic role at the Biochem, 6, 1104–1109. mitochondria. Mol. Cell, 11, 577–590. 28. Siwkowski,A.M., Madge,L.A., Koo,S., McMillan,E.L., Monia,B.P., 44. Xiao,G., Chicas,A., Olivier,M., Taya,Y., Tyagi,S., Kramer,F.R. and Pober,J.S. and Baker,B.F. (2004) Effects of antisense Bargonetti,J. (2000) A DNA damage signal is required for p53 to oligonucleotide-mediated depletion of tumor necrosis factor (TNF) activate gadd45. Cancer Res., 60, 1711–1719. receptor 1-associated death domain protein on TNF-induced gene 45. Simoes-Wust,A.P., Hopkins-Donaldson,S., Sigrist,B., Belyanskaya,L., expression. Mol. Pharmacol., 66, 572–579. Stahel,R.A. and Zangemeister-Wittke,U. (2004) A functionally 29. Chiang,M.Y., Chan,H., Zounes,M.A., Freier,S.M., Lima,W.F. and improved locked nucleic acid antisense oligonucleotide inhibits Bcl-2 Bennett,C.F. (1991) Antisense oligonucleotides inhibit intercellular and Bcl-xL expression and facilitates tumor cell apoptosis. adhesion molecule 1 expression by two distinct mechanisms. Oligonucleotides, 14, 199–209. J. Biol. Chem., 266, 18162–18171. 46. Saelens,X., Kalai,M. and Vandenabeele,P. (2001) Translation 30. Dean,N.M. and McKay,R. (1994) Inhibition of protein kinase C-alpha inhibition in apoptosis: caspase-dependent PKR activation and expression in mice after systemic administration of phosphorothioate eIF2-alpha phosphorylation. J. Biol. Chem., 276, 41620–41628. antisense oligodeoxynucleotides. Proc. Natl Acad. Sci. USA, 91, 47. Nanduri,S., Rahman,F., Williams,B.R. and Qin,J. (2000) A 11762–11766. dynamically tuned double-stranded RNA binding mechanism for the 31. Vickers,T.A., Koo,S., Bennett,C.F., Crooke,S.T., Dean,N.M. and activation of antiviral kinase PKR. EMBO J., 19, 5567–5574. Baker,B.F. (2003) efficient reduction of target RNAs by small 48. Wu,S. and Kaufman,R.J. (1997) A model for the double-stranded RNA interfering RNA and RNase H-dependent antisense agents. (dsRNA)-dependent dimerization and activation of the J. Biol. Chem., 278, 7108–7118. dsRNA-activated protein kinase PKR. J. Biol. Chem., 272, 32. Ross,B. and Song,Q. (2004) US U.S. Pat. Appl. Publ. 20040082775. 1291–1296. 33. Ross,B.S., Song,Q. and Han,M. (2005) Kilo-scale synthesis process for 49. Koizumi,M., Takagi-Sato,M., Okuyama,R., Araki,K., Sun,W., 2 -O-(2-methoxyethyl)-pyrimidine derivatives. Nucleosides Nakai,D., Tsutsumi,S. and Kawai,K. (2006) Direct comparison of Nucleotides Nucleic Acids, 24, 815–818. in vivo antisense activity of ENA oligonucleotides targeting PTP1B 34. Ravikumar,V.T., Andrade,M., Carty,R.L., Dan,A. and Barone,S. (2006) mRNA with that of 2 -O-(2-methoxy)ethyl-modified oligonucleotides. Development of siRNA for therapeutics: efficient synthesis of Oligonucleotides, 16, 253–262.

Journal

Nucleic Acids Research – Oxford University Press

Published: Dec 19, 2006

There are no references for this article.