Allele-Selective Inhibition of Mutant Huntingtin with 2-Thio- and C5- Triazolylphenyl-Deoxythymidine-Modified Antisense Oligonucleotides

Abstract

We report the effect of introducing a single incorporation of 2-thio-deoxythymidine (2S-dT) or C5-Triazolylphenyl-deoxythymidine (5-TrPh-dT) at four positions within the gap region of RNase H gapmer antisense oligonucleotides (ASOs) for reducing wild-type and mutant huntingtin mRNA in human patient fibroblasts. We show that these modifications can modulate processing of the ASO/RNA heteroduplexes by recombinant human RNase H1 in a position-dependent manner. We also created a structural model of the catalytic domain of human RNase H bound to ASO/RNA heteroduplexes to rationalize the activity and selectivity observations in cells and in the biochemical assays. Our results highlight the ability of chemical modifications in the gap region to produce profound changes in ASO behavior.

Introduction

Huntington's disease (HD) is an autosomal dominant neurodegenerative disorder caused by expansion of a polyglutamine (CAG) tract in exon 1 of the huntingtin (HTT) gene [1]. Wild-type huntingtin protein (wt HTT) is essential for normal neuronal development in early embryonic stages, but its role in the adult brain remains poorly understood [2]. In contrast, the mutant huntingtin protein (mu HTT) is toxic to cells and causes HD [3]. As a result, allele-selective suppression of mu HTT in the brain is an attractive therapeutic strategy for treatment of HD [4].

Several oligonucleotide-based approaches have been investigated as potential allele-selective and allele-nonselective HD therapeutics [5]. We recently described second-generation antisense oligonucleotides (ASOs) targeting SNP rs7685686_A in intron 42 of the huntingtin (HTT) gene for allele-selective suppression of mu HTT [6]. SNP rs7685686_A is tightly linked to the CAG expansion, and ASOs targeting this SNP can provide an allele-selective therapeutic option for roughly half of the Caucasian HD population. Second-generation gapmer ASOs are typically fully phosphorothioate (PS) [7]-modified chimeric oligonucleotides with a central DNA gap region flanked on either end with affinity-enhancing nucleotides [8]. Gapmer ASOs bind to their cognate mRNA in cells and promote RNA degradation through RNase H1-mediated hydrolysis [9]. RNase H1 is a ubiquitously expressed enzyme, which selectively cleaves the RNA strand of a RNA/DNA heteroduplex [10,11].

The starting point for our studies was the previously identified ASO A1, which is fully complementary to the mu HTT mRNA, but has a TG mismatch to the wt HTT mRNA at position 8 corresponding to the SNP site (Fig. 1) [12]. This ASO has a 9-base DNA gap flanked on either end with 2′,4′-constrained-2′-O-ethyl bridged nucleic acid (cEt) [13] and 2′-O-methoxyethyl RNA (MOE) nucleotides [14]. We showed that replacing deoxythymidine (dT) at position 5 of ASO A1 with 2-thio-deoxythymidine (2S-dT) produced a profound enhancement in allele selectivity, while analogous substitution at position 6 did not [12]. Interestingly, ASOs modified with 2S-dT differed from A1 only in the replacement of one oxygen atom with sulfur at the C2 position on the pyrimidine nucleobase. This suggested that very small structural changes on the nucleobases in the ASO can have a profound impact on the RNase H activity.

FIG. 1.

Sequence of parent ASO A1 and of the complementary RNA strands representing the wild-type and mutant HTT mRNA. The black letters indicate DNA, orange MOE, blue cEt, and red letters show site of replacement of 2′-deoxythymidine (dT) nucleotides in the gap of parent ASO A1 with 2S-dT or 5-TrPh-dT. Underlined T indicates nucleotide across from the SNP site. All ASOs are fully phosphorothioate backbone modified. ASO, antisense oligonucleotide. Color images available online at www.liebertpub.com/nat

The C5-position on the pyrimidine nucleobases is another position amenable for substitution, and C5-methyl-substituted pyrimidines represent the most widely used nucleobase modification in antisense therapeutics [15]. The 5-methyl groups enhance nuclease stability and increase duplex thermal stability (T_m) by stacking between the heterocyclic nucleobases in oligonucleotide duplexes [16]. In addition, 5-methyl substitution on deoxycytidine nucleotides greatly suppresses the immunostimulatory profile of CpG oligonucleotides [17].

Replacing the 5-methyl group on the pyrimidine nucleobase with other substituents such as 5-alkynyl [18 –22], 5-thiazolyl [23], or 5-heteroaryl [23] further enhanced duplex thermal stability and nuclease stability. Elaboration of these designs has recently led to the synthesis of 5-triazole-substituted pyrimidine DNA analogs, which can provide dramatic enhancements in duplex thermal stability when positioned sequentially within an oligonucleotide as a result of stacking of the aromatic moieties in the major groove [24 –26]. In these designs, 2′-deoxythymidine with bulky 5-triazolylphenyl bearing a distal sulfonamide moiety substituted (5-TrPh-dT, Fig. 1), produced the highest increase in duplex thermal stability. However, the biological evaluation of oligonucleotides modified with 5-TrPh-dT monomers has never been reported. Furthermore, 5-TrPh-dT presented an opportunity to compare the effects of introducing a large substituent in the major groove versus substitution on the Watson–Crick base pairing face (2S-dT) of DNA/RNA heteroduplexes, on the activity and selectivity profile of SNP targeting ASOs.

In this article, we report the effect of a single incorporation of 5-TrPh-dT or 2S-dT at four positions in the gap region of RNase H ASOs targeting a SNP in the HTT mRNA. We show that single incorporation of these modifications is well tolerated and can provide meaningful improvements in selectivity of SNP targeting ASOs for allele selective suppression of gene expression. In addition, we present a model that provides a structural rationale for the biological results.

Materials and Methods

Oligonucleotide synthesis

Oligonucleotides on a 2 μmol scale were made on an ABI 394 DNA/RNA synthesizer using MOE 5-methyl-C Primer support (loading 215 μmol/g). Fully protected nucleoside phosphoramidites were incorporated using standard solid-phase oligonucleotide synthesis, that is, 3% dichloroacetic acid in dichloromethane for deblocking, 1 M 4,5-dicyanoimidazole 0.1 M N-methylimidazole in acetonitrile as an activator for amidite couplings, acetic acid in THF and 10% 1-methylimidazole in THF/pyridine for capping, and 0.2 M phenylacetyl disulfide in pyridine:acetonitrile 1:1 (v:v) for thiolation. DNA building blocks were dissolved in acetonitrile (0.1 M) and incorporated using two times the 4-min coupling time, while other building blocks were dissolved in acetonitrile:toluene 1:1 (v:v) and coupled for 2 times 6 min. After conclusion of the synthesis, the 5′ DMT group was removed and cyanoethyl protecting groups cleaved using triethylamine:acetonitrile 1:1 (v:v). The remaining protecting groups were removed in concentrated aqueous ammonia at room temperature for 12 h. ASOs were purified by ion-exchange-HPLC using a linear gradient of buffer A and B. Buffer A: 50 mM NaHCO₃, buffer B: 1.5 M NaBr, 50 mM NaHCO₃—both buffers in acetonitrile:water 3:7 (v:v). Purified ASOs were desalted using C18 reverse-phase cartridges and analyzed by UV and mass spectrometry (Table 1).

Table 1.

Analytical and Purity Data for Synthesized Antisense Oligonucleotides

ASO	UV purity (%)	Mass (calc.)	Mass (found)
A1	99.5	5108.4	5107.2
B1	96.3	5124.5	5123.4
B2	96.5	5124.5	5123.4
B3	98.6	5124.5	5123.5
B4	96.0	5124.5	5123.1
C1	98.4	5316.6	5315.2
C2	99.0	5316.6	5315.2
C3	98.6	5316.6	5315.2
C4	98.7	5316.6	5315.2
D1	94.2	7157.2	7157.0
E1	92.9	7152.3	7150.7

ASO, antisense oligonucleotide.

Thermal denaturation studies

ASO and RNA were mixed in the 1:1 ratio (4 μM duplex) in a buffer containing 10 mM phosphate, 100 mM NaCl, and 10 μM EDTA at pH 7.0. Duplexes were denatured at 85°C and slowly cooled to the starting temperature of the experiment (15°C). Thermal denaturation temperatures (T_m values) were measured in quartz cuvettes (path length 1.0 cm) on a Cary 100 UV/VIS spectrophotometer equipped with a Peltier temperature controller. Absorbance at 260 nm was measured as a function of temperature using a temperature ramp of 0.5°C per min. T_m values were determined using the hyperchromicity method incorporated into the instrument software. RNAs used were as follows: RNA^mu: 5′-CUGGUGAUGACAAUUUAUU and RNA^wt: 5′-CUGGUGAUGGCAAUUUAUU. Underlined nucleotides indicate position of the SNP site.

Human RNase H1 cleavage pattern

RNA was 5′-end labeled with ³²P using 20 U of T4 polynucleotide kinase, 120 pmol (7,000 Ci/mmol) of [γ-³²P]ATP, 40 pmol RNA, 70 mM Tris-HCl, 10 mM MgCl₂, and 50 mM DTT, at pH 7.6. The labeling reaction was incubated at 37°C for 30 min followed by heating at 90°C for 1 min. Labeled RNA was purified using 12% denaturing polyacrylamide gel. ASO (200 nM), unlabeled 19mer RNA (100 nM; same RNA as used for T_m experiments), and a small amount of ³²P-labeled RNA were mixed in the hybridization buffer (20 mM Tris, 20 mM KCl, pH 7.5) and heated to 90°C for 2 min. To the hybridization mixture were added 0.1 mM tris(2-carboxyethyl)phosphine hydrochloride (TCEP), 1 mM MgCl₂, and 40 U of RNaseOUT and incubated at 37°C for 1 h. The human RNase H1 enzyme was incubated in the dilution buffer (20 mM Tris, 50 mM KCl, and 2 mM TCEP, pH 7.5) for 1 h at room temperature. The enzyme solution (1 μL/25 μL duplex solution) was added to duplex solution and incubated at 37°C. After 45 min, reactions were quenched in the loading buffer and snap-frozen on dry ice. Cleavage products were separated by 12% denaturing polyacrylamide gel and products were quantitated on a phosphorimager.

Allele-selective polymerase chain reaction conditions

GM04022 fibroblasts heterozygous at SNP rs7685686_A were used to measure the in vitro potency and selectivity of the modified ASOs. The cells were trypsinized and resuspended at a density of 260,000 cells per mL in growth medium before transfecting varying concentrations of ASOs with electroporation (Harvard Apparatus ECM830, 130 V, 6 ms). Treated cells were maintained at 37°C and 5% CO₂ in a minimal essential medium containing 15% fetal bovine serum, nonessential amino acids, and penicillin/streptomycin. Approximately 24 h post-transfection, the cells were washed with Dulbecco's phosphate-buffered saline buffer and lysed. RNA was extracted using the Qiagen RNeasy96 kit, and levels of the human HTT mRNA alleles were determined using the qPCR assay C_2229297_10 at SNP rs362303 from Life Technologies. The mu and wtHTT mRNA levels were measured simultaneously using two different fluorophores, FAM for the mutant allele and VIC for the wild-type allele. Quantitative real-time polymerase chain reaction (RT-PCRs) were run on the ABI 7900HT instrument using the QuantiTect Probe RT-PCR Kit following the manufacturer's instructions. The HTT mRNA levels were normalized to total RNA content, as measured by RiboGreen. Data are expressed as mean±standard deviation (n=3). Sequence of the allele nonselective ASO was TCTCTATTGCACATTCCAAG and of the negative control ASO was CCTTCCCTGAAG GTTCCTCC. Both ASOs were fully PS modified, the underlined letters represent MOE nucleotides, and the central gap region is DNA. The dose–response curves were analyzed using nonlinear regression with normalized response and variable slope, and IC₅₀ values were calculated with GraphPad Prism. Allele selectivity was calculated by dividing the IC₅₀ for inhibiting wtHTT by the IC₅₀ for inhibiting muHTT. If the IC₅₀ for reducing wtHTT was greater than the highest ASO concentration tested, then the allele selectivity was calculated by dividing the highest ASO concentration tested by the IC₅₀ for inhibiting muHTT and expressed as >X fold.

Structural model

The structural model was created from the published X-ray crystal structure (pdb code 2QK9) of the catalytic domain of human RNase H1 bound to a DNA/RNA heteroduplex [27] using Discovery Studio 4.0 visualizer software. Oxygen atoms at the C2-position of pyrimidine nucleobases at relevant locations from the published models were changed to sulfur and distances to relevant amino acid side chains were mapped. The nucleobase of 5-TrPh-dT was created in ChemBio3D Ultra14.0 and used in the structural model.

Results

Effect of 2S-dT and 5-TrPh-dT on thermal discrimination of GT mismatch

The ASOs were first tested in Tm assays versus the matched RNA complement RNA^mu and the mismatched complement RNA^wt to assess the effect of introducing 2S-dT and 5-TrPh-dT at different positions in the gap region of ASO A1 on discrimination of the central GT mismatch (Table 2). In general, incorporation of a single 2S-dT monomer resulted in a 0.3°C–2.0°C increase in T_m, while incorporation of 5-TrPh-dT resulted in T_m changes of −2.2°C to +0.7°C. Both modifications did not enhance discrimination of the GT mismatch, except when placed directly across from the mismatch (A1 ΔT_m +2.2°C, B3 ΔT_m +5.6°C, and C3 ΔT_m +4.0°C).

Table 2.

Antisense Oligonucleotide Sequence, Modification Patterns, Tm, % RNA Cleaved at 45 min, Activity Versus mu HTT mRNA, and Fold Selectivity Versus wt HTT mRNA

ASO	Mod.	Sequence (5′-3′)	T_m °C RNA^mu	T_m °C RNA^wt	ΔT_m °C	% RNA^mu cleaved	% RNA^wt cleaved	muHTT IC₅₀ (μM)	Fold select
A1	NA	TAAATTGTCATCACC	54.6	52.4	+2.2	69	18	0.29	17.6
B1	2S-dT	TAAATTGTCATCACC	55.6	53.8	+1.8	44	7	0.49	>>20
B2	2S-dT	TAAATTGTCATCACC	56.4	54.5	+1.9	44	24	0.13	5.4
B3	2S-dT	TAAATTGTCATCACC	54.9	49.3	+5.6	69	52	0.15	23.4
B4	2S-dT	TAAATTGTCATCACC	55.0	53.0	+2.0	70	19	0.16	17.6
C1	5-TrPh-dT	TAAATTGTCATCACC	52.4	50.0	+2.4	60	21	0.24	13.3
C2	5-TrPh-dT	TAAATTGTCATCACC	55.3	52.9	+2.4	68	17	0.19	16.8
C3	5-TrPh-dT	TAAATTGTCATCACC	53.4	49.4	+4.0	67	22	0.52	30.4
C4	5-TrPh-dT	TAAATTGTCATCACC	53.2	51.1	+2.1	47	19	1.14	15.8
D1	NA	Nonselective ASO	NA	NA	NA	NA	NA	0.4	1.1
E1	NA	Neg. control ASO	NA	NA	NA	NA	NA	>10	NA

The black letters indicate DNA, orange MOE, blue cEt, and red and green letters show site of replacement of 2′-deoxythymidine (dT) nucleotides in the gap of parent ASO A1 with 2S-dT or 5-TrPh-dT, respectively. Underlined letters indicate position across from the SNP site. All ASOs are fully phosphorothioate backbone modified. Nonselective ASO is a 5–10–5 MOE gapmer, which targets HTT mRNA at site distinct from SNP rs7685686_A. Negative control ASO is a 5–10–5 MOE gapmer, which does not target HTT. Color images available online at www.liebertpub.com/nat

NA, not applicable.

Effect of 2S-dT and 5-TrPh-dT on activity and allele selectivity in human fibroblasts

ASOs B1–B4 and C1–C4 were tested in GM04022 human fibroblasts for their ability to reduce HTT mRNA, and the reduction of the individual alleles was measured using allele-selective qRT-PCR (Fig. 2 and Table 2) [12]. In addition, we also included a nonallele-selective ASO D1, which does not bind at the targeted SNP site on the HTT mRNA and ASO E1, which does not target the HTT mRNA, as controls for the assay. As seen in our previous study, introducing 2S-dT at position 5 resulted in a profound increase in allele selectivity while maintaining potency versus the mutant allele (<20% reduction in wt HTT at 10 μM; IC₅₀ for reducing mu HTT 0.49 μM). In contrast, introducing 2S-dT at position 6 (B2) resulted in similar potency, but reduced selectivity compared to A1. Introducing 2S-dT at positions 8 (B3) and 11 (B4) showed similar potency and selectivity profile as the parent ASO A1.

FIG. 2.

Reduction of (A) mutant HTT mRNA and (C) wild-type HTT mRNA in GM4022 human fibroblasts using ASOs A1, B1–B4, C1–C4, D1, and E1 delivered by electroporation. Dose–response curves for reducing (B) mutant HTT mRNA and (D) wild-type HTT mRNA calculated using GraphPad Prism 5. (E) All ASOs tested had no effect on total cellular RNA as measured by RiboGreen dye. Color images available online at www.liebertpub.com/nat

In comparison, introducing 5-TrPh-dT at positions 5, 6, and 11 maintained or reduced potency versus the mutant allele, but did not enhance selectivity compared to parent ASO A1. Interestingly, placing 5-TrPh-dT across from the SNP site (position 8) reduced potency 2-fold versus the mutant allele, but appeared to enhance selectivity versus the wild-type allele relative to parent ASO A1. The slightly higher selectivity observed with ASOs B3 and C3 relative to parent ASO A1 could be attributed to the enhanced binding discrimination for the TG mismatch (Table 2), as a result of introducing the chemical modifications directly across from the SNP site. All ASOs were well tolerated and showed no reductions in total RNA as a measure of nonspecific cellular toxicity (Fig. 2E).

Effect of 2S-dT and 5-TrPh-dT on RNase H cleavage versus wt and mu HTT RNA

We mapped the cleavage sites produced by recombinant human RNase H1 on ASO A1/RNA^wt and A1/RNA^mu heteroduplexes in a biochemical assay [12]. Five cleavage sites (a, b, c, d, and e) were observed on the fully matched A1/RNA^mu heteroduplex with the major cleavage a occurring adjacent to the SNP site (Fig. 3). In case of A1/RNA^wt, the presence of the TG mismatch in the center of the gap ablates cleavage sites a, b, and c (black crosses), and the cleavage was instead shifted to sites d and e.

FIG. 3.

Mapping cleavage sites (a, b, c, d, and e) of recombinant human RNase H1 on ASO/RNA heteroduplexes with matched RNA^mu representing mutant HTT and RNA^wt representing wild-type HTT alleles, respectively. Introducing 2S-dT at position 5 (red nucleotide) enhances allele selectivity, while introducing 2S-dT at position 6 does not. The black, red, and green crosses indicate cleavage sites ablated by the TG mismatch, 2S-dT, and 5-TrPh-dT, respectively. The black letters indicate DNA, orange MOE, blue cEt, and red and green letters show site of replacement of 2′-deoxythymidine (dT) nucleotides in the gap of parent ASO A1 with 2S-dT or 5-TrPh-dT, respectively. Underlined letters indicate position across from the SNP site. Color images available online at www.liebertpub.com/nat

Introducing 2S-dT at position 5 of the ASO B1 ablated cleavage sites d and e in both duplexes (red crosses), while the TG mismatch ablated sites a, b, and c in the B1/RNA^wt heteroduplex. The ablation of RNase H activity versus RNA^wt provides a rationale for the greatly improved allele selectivity in cell culture with this ASO. In contrast, introducing 2S-dT at position 6 ablated cleavage sites c and d, but not site e, in both heteroduplexes. Incorporation of 2S-dT at position 8 across from the SNP site ablated cleavage sites a and b on RNA^wt and RNA^mu. Interestingly, this ASO had an identical cleavage pattern for both heteroduplexes and greatly enhanced cleavage site c on the B3/RNA^wt, which is typically ablated by the TG mismatch, highlighting the ability of positional incorporation of 2S-dT to modify RNase H cleavage patterns. Introducing 2S-dT at position 11 (ASO B4) showed identical cleavage patterns as parent ASO A1 on both heteroduplexes. The inability of ASOs B2, B3, and B4 to suppress RNase H cleavage on RNA^wt provides a likely explanation of the lack of improvement in allele selectivity relative to control ASO A1.

We also mapped the RNase H cleavage patterns for 5-TrPh-dT modified ASOs C1-C4 heteroduplexes with the matched RNA^mu and mismatched RNA^wt. Introducing 5-TrPh-dT at position 5 and 6 did not change the cleavage patterns relative to the parent ASO A1. Introducing 5-TrPh-dT at position 8 intensified cleavage sites a and b, while reducing cleavage intensity at sites c, d, and e. However, this modification was not able to ablate these sites on the mismatched RNA^wt, resulting only in a slightly enhanced increase in allele selectivity in cell culture. In contrast, introducing 5-TrPh-dT at position 11 shifted cleavage predominantly to sites a and e on RNA^mu and to site e on RNA^wt. The shift in cleavage patterns for ASOs C3 and C4 relative to A1 suggested that the bulky 5-TrPh-dT modification influences RNase H activity when incorporated 5–7 nucleotides away from the site/s of cleavage. This is in contrast to 2S-Dt, which influences RNase H activity when incorporated closer to the cleavage site.

The relationship between magnitude of RNase H1 cleavage of RNA^mu and RNA^wt produced by each ASO after 45 min of incubation in the biochemical assay and the activity/selectivity observations in cell culture was slightly less straightforward (Table 2). In general, all ASOs exhibited higher magnitude of RNase H1-mediated cleavage when paired with the fully matched RNA complement RNA^mu (44%–70% cleaved after 45 min) compared to the singly mismatched RNA complement RNA^wt (7%–52%). ASO B1, which showed the lowest magnitude of cleavage of RNA^wt (7%) after 45 min in the biochemical assay, also showed the best allele selectivity (>>20-fold) in cell culture. In contrast, ASO B3 (2S-dT across from the SNP site) produced 52% cleavage of RNA^wt, but exhibited modestly improved selectivity in cell culture compared to ASO A1 (18% cleavage of RNA^wt). A likely explanation is that the increase in allele selectivity results from enhanced binding discrimination of the TG wobble base pair by placing 2S-dT directly across from the SNP site [28].

Structural model for differential effects of 2S-dT and 5-TrPh-dT on RNase H processing

To better understand these observations, we created a structural model of 2S-dT and 5-TrPh-dT modified DNA/RNA duplexes bound to the catalytic domain of human RNase H1 (Fig. 4A). The human enzyme comprises three domains—the catalytic, linking, and hybrid binding—and differs from the Escherichia coli enzyme, which only has the catalytic domain [11,29]. In the context of gapmer ASOs, the hybrid-binding domain is believed to interact with the 3′-wing region [30], while the catalytic domain has a seven-nucleotide footprint on the ASO/RNA heteroduplex (Fig. 4A) [27]. Crystal structures of the catalytic and hybrid domains bound to DNA/RNA hybrids have been reported previously [27,31]. Since human RNaseH1 will have a partially overlapping but distinct seven-nucleotide footprint for every cleavage site, we mapped the interactions of the 2-thio atom and the bulky 5-TrPh group with the enzyme for one cleavage site on the RNA to simplify the analysis (Fig. 4B).

FIG. 4.

Structural model of the catalytic domain of human RNase H1 bound to a DNA/RNA heteroduplex showing (A) the 5-TrPh-moiety at position 2 (brown carbons) projects into the major groove away from the enzyme, but comes in close proximity to the basic protrusion handle at position 7 (olive green carbons). (B) Pictorial view of the scissile phosphate with RNA nucleotides (light green letters) and the 7-nucleotide footprint of the catalytic domain on the DNA strand. (C) The 2S-moiety (bronze spheres) at position 2 (brown carbons) and position 3 (black carbons), but not position 4 (yellow carbons), makes close contacts with ASN151, ASN182, and GLN183, respectively. Introducing 2S-dT at positions 2 (brown carbons) and/or 3 (black carbons) will ablate RNase H cleavage between the RNA (light green carbons, only two nucleotides in the RNA strand shown for clarity) nucleotides, while 2S-dT at position 4 (yellow carbons) will not ablate this cleavage site. (D) Introducing 5-TrPh-dT (olive green carbons) in the vicinity of position 7 will cause a steric clash with THR232 and SER233 of the basic protrusion handle and potentially modulate RNase H cleavage. Color images available online at www.liebertpub.com/nat

2S-dT at position 2 (brown carbons) in the DNA strand relative to the scissile phosphate (light green carbons) can result in a tight contact with the side chain of ASN151, while 2S-dT at position 3 (black carbons) can result in a tight contact with ASN182 and GLN183 (Fig. 4B, C). In contrast, 2S-dT at position 4 (light yellow carbons) is not in close proximity to the protein backbone or side chains. Presumably, the tight contacts with the sulfur atom in 2S-dT at positions 3 and 4 are alleviated in case of dT because of the smaller van der Waals radius of oxygen. Superimposition of the model onto the cleavage sites observed in the biochemical assay rationalize loss of cleavage site e on RNA^wt when 2S-dT is placed at position 5 in B1 (2S-dT black carbons), as opposed to when 2S-dT is placed at position 6 (2S-dT yellow carbons).

In contrast, the bulky 5-TrPh substituent projects into the major groove away from the enzyme, when introduced at position 2 (brown carbons) in the DNA strand relative to the cleavage site on the RNA (Fig. 4B, D). Introducing this substituent at position 7 (olive green carbons) on the DNA strand brings it in close proximity of SER133 and THR232, which are part of the basic protrusion region of the catalytic domain [27]. These observations explain why this modification did not modify the RNase H cleavage patterns when introduced in the proximity of the cleavage sites but appears to modulate RNase H cleavage when incorporated 5–7 nucleotides in the DNA strand away from the cleavage site on the RNA (Fig. 2).

Discussion

HD is an autosomal dominant genetic disorder caused by expansion of a polyglutamine-encoding CAG tract within exon 1 of the HTT gene [1]. Oligonucleotide-based strategies are ideally suited as potential HD therapeutics because of their ability to target the mRNA directly and suppress production of the disease causing protein. Nonallele-selective RNaseH ASOs [32] and siRNA [33], which target homologous sequence tracts in mu HTT and wt HTT mRNA, partially lower both allelic variants and produce therapeutic benefits in murine HD disease models. Allele selectivity can be achieved by targeting the CAG tract directly [34] or by targeting SNPs tightly linked to the CAG expansion [35].

ASOs targeting SNP rs7685686_A in intron 42 of the HTT gene can selectively suppress mutant HTT mRNA and HTT protein in human patient fibroblasts [12,36]. In addition, these ASOs were well tolerated in wild-type rodents and produced potent and sustained suppression of mu HTT protein in the CNS in transgenic murine HD models. However, ASOs targeting SNP rs7685686_A can only provide an allele-selective therapeutic option for ∼45% of the Caucasian HD population due to heterozygosity of SNPs in the broader population. Therefore, additional ASOs at other SNP sites within the HTT gene are required to provide an allele-selective option for all patients.

Recent studies have shown that ASO structure–activity relationships (SAR) for targeting other SNPs can be somewhat unpredictable because of differences in RNase H processing based on changes in the ASO sequence, position of SNP in the gap, and nature of the mismatch [36,37]. Thus, additional chemical strategies, which can improve allele selectivity by enhancing mismatch discrimination or modulating RNase H cleavage of singly mismatched heteroduplexes, are desirable. In this context, our data contribute to the chemical toolbox and provide design options for enhancing properties of SNP targeting ASOs for treating autosomal dominant disorders.

A few previous studies have characterized the effect of non-DNA-like chemical modifications on activity of human RNase H1 in biochemical experiments [38]. However, translating these observations from biochemical systems to real-life examples in cells and animals is lacking. In this broader context, targeting SNP rs7685686_A provides an experimental system to characterize the effects of introducing non-DNA-like chemical modifications within the gap region on the activity and selectivity profile of RNase H ASOs in cells and in animals [12,39,40].

While we were able to provide a rationale for the changes in allele selectivity by mapping RNase H1 cleavage sites on the fully matched versus singly mismatched RNA/ASO heteroduplexes, gaps in our knowledge still remain. For example, the extent of cleavage of RNA^wt and RNA^mu observed after 45 min of incubation in the biochemical assay matches well with the activity and selectivity trends in cell culture for most ASOs, but not for all. A plausible explanation is that the activity/selectivity observations in cell culture reflect the end result of a complex ensemble of events, including ASO uptake into cells, release from endomembrane compartments, scanning the mRNA space for its cognate site, and recruitment of RNase H to the ASO/RNA heteroduplex. This sequence of events cannot be replicated in the biochemical system using preformed duplexes and fixed concentration of enzyme.

Another confounding variable is that the SNP rs7685686_A is located in intron 42 of the very large HTT pre-mRNA, suggesting that the kinetics of intron splicing could play a role as well. Nonetheless, the most selective ASO in cell culture (B1) produced the slowest (7%) cleavage of RNA^wt in the biochemical assay. In contrast, the slightly higher selectivity observed with ASOs B3 and C3 relative to parent ASO A1 could be attributed to enhanced binding discrimination of the TG mismatch as a result of introducing the chemical modifications directly across from the SNP site.

Our structural analysis showed that the sulfur atom in 2S-dT and the 5-triazolylphenyl group in 5-TrPh-dT occupy very distinct sites (Watson–Crick base pairing face versus major groove) in the RNase H1/heteroduplex complex. The C2-position of the pyrimidine nucleobase is in close proximity to highly conserved amino acids ASN 151, ASN182, and GLN 183 in the catalytic domain of RNase H1 [27], when this modification is inserted in the antisense DNA strand within two nucleotides of the cleavage site on the RNA. Replacing the 2-oxygen atom in pyrimidine nucleobases with sulfur causes tight contacts with these amino acids because of the larger van der Waals radius of sulfur causing a shift in the RNase H1 cleavage pattern toward the 3′-side on the RNA strand.

In contrast, the 5-triazolylphenyl group in 5-TrPh-dT is bigger than the sulfur atom in 2S-dT, but its steric bulk projects into the major groove away from the RNase H1-binding interface, when this modification is introduced in the vicinity of the RNase H cleavage site. However, this modification experiences a steric clash with THR232 and SER233 in the basic protrusion handle of human RNase H1 when it is introduced on the DNA strand 5–7 nucleotides away from the cleavage site on the RNA. Despite this steric clash, this modification has a smaller impact on RNase H1 activity, suggesting that the enzyme can accommodate this chemical modification.

In conclusion, we report the synthesis and biological evaluation of gapmer ASOs modified with bulky C5-triazolylphenyl and 2-thio substituted 2′-deoxythymidine nucleotides. We show that single incorporation of these modifications in the gap region does not reduce activity relative to PS DNA and can provide meaningful improvements in allele selectivity. Furthermore, our analysis illuminates the structural basis for the improvements in allele selectivity observed by introducing non-DNA-like chemical modifications in the gap region of ASOs, which function through the RNase H antisense mechanism.

Footnotes

Author Disclosure Statement

No competing financial interests exist. M.E.O., J.N., A.W., and P.P.S. are employees of Isis Pharmaceuticals, Inc.

References

MacDonald

, Ambrose

, Duyao

, Myers

, Lin

, Srinidhi

, Barnes

, Taylor

, James

, et al. (1993). A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington's disease chromosomes. Cell, 72:971–983.

Nasir

, Floresco

, O'Kusky

, Diewert

, Richman

, Zeisler

, Borowski

, Marth

, Phillips

and Hayden

. (1995). Targeted disruption of the Huntington's disease gene results in embryonic lethality and behavioral and morphological changes in heterozygotes. Cell, 81:811–823.

Walker

. (2007). Huntington's disease. Lancet, 369:218–228.

Zuccato

, Valenza

and Cattaneo

. (2010). Molecular mechanisms and potential therapeutical targets in Huntington's disease. Physiol Rev, 90:905–981.

Southwell

, Skotte

, Bennett

and Hayden

. (2012). Antisense oligonucleotide therapeutics for inherited neurodegenerative diseases. Trends Mol Med, 18:634–643.

Carroll

, Warby

, Southwell

, Doty

, Greenlee

, Skotte

, Hung

, Bennett

, Freier

and Hayden

. (2011). Potent and selective antisense oligonucleotides targeting single-nucleotide polymorphisms in the Huntington disease gene/allele-specific silencing of mutant huntingtin. Mol Ther, 19:2178–2185.

Eckstein

. (2014). Phosphorothioates, essential components of therapeutic oligonucleotides. Nucleic Acid Ther, 24:374–387.

Seth

, Siwkowski

, Allerson

, Vasquez

, Lee

, Prakash

, Wancewicz

, Witchell

and Swayze

. (2009). Short antisense oligonucleotides with novel 2′-4′ conformationaly restricted nucleoside analogues show improved potency without increased toxicity in animals. J Med Chem, 52:10–13.

, Lima

, Zhang

, Fan

, Sun

and Crooke

. (2004). Determination of the role of the human RNase H1 in the pharmacology of DNA-like antisense drugs. J Biol Chem, 279:17181–17189.

10.

Lima

, Wu

and Crooke

. (2001). Human RNases H. Methods Enzymol, 341:430–440.

11.

Cerritelli

and Crouch

. (2009). Ribonuclease H: the enzymes in eukaryotes. FEBS J, 276:1494–1505.

12.

Ostergaard

, Southwell

, Kordasiewicz

, Watt

, Skotte

, Doty

, Vaid

, Villanueva

, Swayze

, et al. (2013). Rational design of antisense oligonucleotides targeting single nucleotide polymorphisms for potent and allele selective suppression of mutant Huntingtin in the CNS. Nucleic Acids Res, 41:9634–9650.

13.

Seth

, Vasquez

, Allerson

, Berdeja

, Gaus

, Kinberger

, Prakash

, Migawa

, Bhat

and Swayze

. (2010). Synthesis and biophysical evaluation of 2′,4′-constrained 2′O-methoxyethyl and 2′,4′-constrained 2′O-ethyl nucleic acid analogues. J Org Chem, 75:1569–1581.

14.

Teplova

, Minasov

, Tereshko

, Inamati

, Cook

, Manoharan

and Egli

. (1999). Crystal structure and improved antisense properties of 2′-O-(2-methoxyethyl)-RNA. Nat Struct Biol, 6:535–539.

15.

Seth

and Swayze

. (2014). Unnatural nucleoside analogs for antisense therapy. In: Natural Products in Medicinal Chemistry. Hanessian

, ed. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, pp 403–440.

16.

Herdewijn

. (2000). Heterocyclic modifications of oligonucleotides and antisense technology. Antisense Nucleic Acid Drug Dev, 10:297–310.

17.

Krieg

. (2002). CPG motifs in bacterial DNA and their immune effects. Annu Rev Immunol, 20:709–760.

18.

Wagner

, Matteucci

, Lewis

, Gutierrez

, Moulds

and Froehler

. (1993). Antisense gene inhibition by oligonucleotides containing C-5 propyne pyrimidines. Science, 260:1510–1513.

19.

Kumar

, Baral

, Anderson

, Guenther

, Ostergaard

, Sharma

and Hrdlicka

. (2014). C5-alkynyl-functionalized alpha-L-LNA: synthesis, thermal denaturation experiments and enzymatic stability. J Org Chem, 79:5062–5073.

20.

Kaura

, Kumar

and Hrdlicka

. (2014). Synthesis, hybridization characteristics, and fluorescence properties of oligonucleotides modified with nucleobase-functionalized locked nucleic acid adenosine and cytidine monomers. J Org Chem, 79:6256–6268.

21.

Kaura

, Guenther

and Hrdlicka

. (2014). Carbohydrate-functionalized locked nucleic acids: oligonucleotides with extraordinary binding affinity, target specificity, and enzymatic stability. Org Lett, 16:3308–3311.

22.

Guenther

, Kumar

, Anderson

and Hrdlicka

. (2014). C5-amino acid functionalized LNA: positively poised for antisense applications. Chem Commun (Camb), 50:9007–9009.

23.

Gutierrez

, Terhorst

, Matteucci

and Froehler

. (1994). 5-Heteroaryl-2′-deoxyuridine analogs. Synthesis and incorporation into high-affinity oligonucleotides. J Am Chem Soc, 116:5540–5544.

24.

Kumar

, Hornum

, Nielsen

, Enderlin

, Andersen

, Len

, Herve

, Sartori

and Nielsen

. (2014). High-affinity RNA targeting by oligonucleotides displaying aromatic stacking and amino groups in the major groove. Comparison of triazoles and phenyl substituents. J Org Chem, 79:2854–2863.

25.

Andersen

, Dossing

, Jensen

, Vester

and Nielsen

. (2011). Duplex and triplex formation of mixed pyrimidine oligonucleotides with stacking of phenyl-triazole moieties in the major groove. J Org Chem, 76:6177–6187.

26.

Kumar

, Chandak

, Nielsen

and Sharma

. (2012). Sulfonamide bearing oligonucleotides: simple synthesis and efficient RNA recognition. Bioorg Med Chem, 20:3843–3849.

27.

Nowotny

, Gaidamakov

, Ghirlando

, Cerritelli

, Crouch

and Yang

. (2007). Structure of human RNase H1 complexed with an RNA/DNA hybrid: insight into HIV reverse transcription. Mol Cell, 28:264–276.

28.

Kumar

and Davis

. (1997). Synthesis and studies on the effect of 2-thiouridine and 4-thiouridine on sugar conformation and RNA duplex stability. Nucleic Acids Res, 25:1272–1280.

29.

Nowotny

, Gaidamakov

, Crouch

and Yang

. (2005). Crystal structures of RNase H bound to an RNA/DNA hybrid: substrate specificity and metal-dependent catalysis. Cell, 121:1005–1016.

30.

Lima

, Rose

, Nichols

, Wu

, Migawa

, Wyrzykiewicz

, Vasquez

, Swayze

and Crooke

. (2007). The positional influence of the helical geometry of the heteroduplex substrate on human RNase H1 catalysis. Mol Pharmacol, 71:73–82.

31.

Nowotny

, Cerritelli

, Ghirlando

, Gaidamakov

, Crouch

and Yang

. (2008). Specific recognition of RNA/DNA hybrid and enhancement of human RNase H1 activity by HBD. EMBO J, 27:1172–1181.

32.

Kordasiewicz

, Stanek

, Wancewicz

, Mazur

, McAlonis

, Pytel

, Artates

, Weiss

, Cheng

, et al. (2012). Sustained therapeutic reversal of Huntington's disease by transient repression of huntingtin synthesis. Neuron, 74:1031–1044.

33.

DiFiglia

, Sena-Esteves

, Chase

, Sapp

, Pfister

, Sass

, Yoder

, Reeves

, Pandey

, et al. (2007). Therapeutic silencing of mutant huntingtin with siRNA attenuates striatal and cortical neuropathology and behavioral deficits. Proc Natl Acad Sci U S A, 104:17204–17209.

34.

, Matsui

, Gagnon

, Schwartz

, Gabillet

, Arar

, Wu

, Bezprozvanny

and Corey

. (2009). Allele-specific silencing of mutant huntingtin and ataxin-3 genes by targeting expanded CAG repeats in mRNAs. Nat Biotechnol, 27:478–484.

35.

Pfister

, Kennington

, Straubhaar

, Wagh

, Liu

, DiFiglia

, Landwehrmeyer

, Vonsattel

J-P

, Zamore

and Aronin

. (2009). Five siRNAs targeting three SNPs may provide therapy for three-quarters of Huntington's disease patients. Curr Biol, 19:774–778.

36.

Southwell

, Skotte

, Kordasiewicz

, Ostergaard

, Watt

, Carroll

, Doty

, Villanueva

, Petoukhov

, et al. (2014). In vivo evaluation of candidate allele-specific mutant huntingtin gene silencing antisense oligonucleotides. Mol Ther, 22:2093–2106.

37.

Skotte

, Southwell

, Ostergaard

, Carroll

, Warby

, Doty

, Petoukhov

, Vaid

, Kordasiewicz

, et al. (2014). Allele-specific suppression of mutant huntingtin using antisense oligonucleotides: providing a therapeutic option for all Huntington disease patients. PLoS One, 9:e107434.

38.

Lima

, Nichols

, Wu

, Prakash

, Migawa

, Wyrzykiewicz

, Bhat

and Crooke

. (2004). Structural requirements at the catalytic site of the heteroduplex substrate for human RNase H1 catalysis. J Biol Chem, 279:36317–36326.

39.

Ostergaard

, Gerland

, Escudier

, Swayze

and Seth

. (2014). Differential effects on allele selective silencing of mutant huntingtin by two stereoisomers of alpha,beta-constrained nucleic acid. ACS Chem Biol, 9:1975–1979.

40.

Ostergaard

, Thomas

, Koller

, Southwell

, Hayden

and Seth

. (2015). Biophysical and biological characterization of hairpin and molecular beacon RNase H active antisense oligonucleotides. ACS Chem Biol, 10:1227–1233.