Chemical modification patterns conferring high efficiency

Abstract

The present invention relates to a method of identifying a chemical modification pattern which confers an increased splice-modulation efficiency to a single-stranded RNA oligonucleotide with a predetermined nucleobase sequence, as compared to a single-stranded RNA oligonucleotide with the same predetermined nucleobase sequence having every nucleotide chemically modified with a 2' ribose modification, the method comprising: (a) applying a plurality of chemical modification patterns (Patterns A to G) to a plurality of single-stranded RNA oligonucleotides respectively, wherein each of the plurality of single-stranded RNA oligonucleotides has one of Patterns A to G applied to it, (i) wherein each of the plurality of chemical modification patterns comprises: (A) at least one nucleotide chemically modified with the 2' ribose modification; and (B) one to eight nucleotides chemically modified with a constrained nucleotide (CN); and (ii) wherein the plurality of chemical modification patterns comprises Patterns A to G of CN-modified nucleotides; and (b) measuring the splice- modulation efficiencies of the chemically-modified single-stranded RNA oligonucleotides and comparing these with the splice-modulation efficiency of the single-stranded RNA oligonucleotide having every nucleotide chemically modified with the 2' ribose modification to determine the chemical modification pattern which confers the increased splice-modulation efficiency.

Description

DESCRIPTIONTITLE OF THE INVENTION: CHEMICAL MODIFICATION PATTERNS CONFERRINGHIGH EFFICIENCYFIELD OF THE INVENTION

[0001] The present invention relates generally to the field of molecular biology. In particular, the invention relates to a method of identifying a chemical modification pattern which confers the highest efficiency to a single-stranded RNA oligonucleotide.BACKGROUND OF THE INVENTION

[0002] The broad plethora of compounds embraced under the label of nucleic-acid therapeutics (NATs) have helped overcome multiple medical challenges in the recent decade, especially improving the lives of patients affected by debilitating genetic diseases. This group of therapeutic modalities encompasses formulations that contain a modified or natural nucleic- acid payload, such as siRNA, CRISPR systems (with their guide RNAs), and antisense oligonucleotides (ASO), which among others include GAPmers and steric blocking spliceswitching oligonucleotides (SSOs). Unfortunately, despite the initial success in the clinic, poor bioavailability and / or potency of NATs besets their relevance for most clinical indications.

[0003] The incorporation of numerous chemical modifications on both the sugar moiety and inter-nucleotide linkage of an antisense oligonucleotide have bestowed resistance to nucleases and enhanced its therapeutic efficacy. However, despite extensive research, there is no predictable rule for where specific chemical modifications should be placed along the sequence of the antisense oligonucleotide to achieve the highest efficacy. This necessitates extensive empirical testing and optimisation for each NAT, significantly increasing the complexity, time, and cost of drug development.

[0004] There is thus a need for a method of identifying chemical modification pattern that confers the highest efficiency that overcomes the drawbacks of the prior art. Furthermore, other desirable features and characteristics will become apparent from the subsequent detailed description and the appended claims, taken in conjunction with the accompanying drawings and this background of the disclosure.SUMMARY

[0005] In one aspect, the present invention provides a method of identifying a chemical modification pattern which confers an increased splice-modulation efficiency to a singlestranded RNA oligonucleotide with a predetermined nucleobase sequence, as compared to a single-stranded RNA oligonucleotide with the same predetermined nucleobase sequence having every nucleotide chemically modified with a 2’ ribose modification, the method comprising: (a) applying a plurality of chemical modification patterns (Patterns A to G) to a plurality of single-stranded RNA oligonucleotides respectively to result in a plurality of chemically-modified single-stranded RNA oligonucleotides, wherein each of the plurality of single-stranded RNA oligonucleotides has one of Patterns A to G applied to it, and wherein the plurality of single-stranded RNA oligonucleotides have the same predetermined nucleobase sequence, (i) wherein each of the plurality of chemical modification patterns comprises: (A) at least one nucleotide chemically modified with the 2’ ribose modification; and (B) one to eight nucleotides chemically modified with a constrained nucleotide (CN); and (ii) wherein the plurality of chemical modification patterns comprises the following positional configurations of CN-modified nucleotides: (A) Pattern A comprising a CN cluster comprising one or more consecutive CNs, beginning at position +1 , +2 or +3 of the single-stranded RNA oligonucleotide; (B) Pattern B comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the single-stranded RNA oligonucleotide, and a second CN cluster comprising one or more consecutive CNs, ending at position -2 or - 1 of the single-stranded RNA oligonucleotide; (C) Pattern C comprising a CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to the CN cluster and that is not at position -2 or -1 of the single-stranded RNA oligonucleotide; (D) Pattern D comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the single-stranded RNA oligonucleotide, and a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to both the first and second CN clusters; (E) Pattern E comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 , and a second CN cluster comprising one or more consecutive CNs, ending at position -1 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters; (F) Pattern F comprising a first CN at position +3 and a second CN at position -2 of the single-stranded RNA oligonucleotide; and (G) Pattern G comprising a first CN at position +3 and a second CN atposition -1 of the single-stranded RNA oligonucleotide; and (b) measuring the splicemodulation efficiencies of the chemically-modified single-stranded RNA oligonucleotides and comparing these with the splice-modulation efficiency of the single-stranded RNA oligonucleotide having every nucleotide chemically modified with the 2’ ribose modification to determine the chemical modification pattern which confers the increased splice-modulation efficiency.

[0006] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern H) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern H comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide, and a second CN cluster comprising two or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide.

[0007] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern I) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern I comprising the following positional configuration of CN-modified nucleotides: a CN cluster comprising (i) one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide; and (ii) an additional CN that is not adjacent to the CN cluster and that is not at position -2 or -1 of the singlestranded RNA oligonucleotide.

[0008] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern J) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern J comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to both the first and second CN clusters.

[0009] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern K) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern K comprising the following positional configuration of CN-modified nucleotides: a CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to the CN cluster.

[0010] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern L) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern L comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +1 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters.

[0011] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern M) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern M comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +2 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters.

[0012] In one embodiment, the 2’ ribose modification is selected from the group consisting of 2’-O-methyl (2’-OMe) and 2’-O-methoxyethyl (2’-MOE).

[0013] In one embodiment, the constrained nucleotide (CN) is selected from the group consisting of locked nucleic acid (LNA), S-constrained ethyl ((S)-cET), ethylene-bridged nucleic acid (ENA), 2’-O-N-methylacetamido (2’-O-NMA), threose nucleic acid (TNA), hexitol nucleic acid (HNA), methyleneoxy nucleic acid (MNA), 2’-Fluoro and 2’-Fluoroarabino nucleic acid (2’-FANA).

[0014] In one embodiment, all nucleotides that are not CN-modified comprise the 2’ ribose modification.

[0015] In one embodiment, the single-stranded RNA oligonucleotide is between 15 and 22 nucleotides in length.

[0016] In one embodiment, the nucleobase sequence of the single-stranded RNA oligonucleotide has a melting temperature of 70°C or below.BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The invention will be better understood with reference to the detailed description when considered in conjunction with the non-limiting examples and the accompanying drawings, in which:

[0018] Fig. 1 shows that the positional configurations of SSO mixmer increase potency of AVA1. (A) AAUC plot for MOE, OMe, MOL, and OML. Cumulative difference of the Area Under the Curve in the dose responses between AVA1A1 SSO and untreated (UTR) cells. For each SSO exon skipping efficiency (FS pPSI) and cellular toxicity (by nuclei count) over four concentrations (6, 15, 37.5, and 93.8nM) were determined by high content confocal imaging at 24 hours after transfection into CT26 NATU A-expressing cells. The area under the curve (AUG) was calculated from the dose response curve and subtracted with those from untreated cells to obtain the AAUC in pPSI (vertical axis) and cell viability (horizontal axis). Both AAUC (pPSI) and AAUC (cell viability) correlate with higher efficiency and higher cell viability respectively (Material and Methods). The plot shows the average and SD of two biological replicates. A minimum of 6 fields and 250 cells were imaged for each replicate. Unpaired t test was performed between OML and MOL (Welch's correction, unequal SDs, one-tailed P-value). (B) LNA placements in AVA1A1 MXs by Lateral and Central PCs (positional configurations). Lateral and Central PCs are further defined in the “Linear regression model” under “Material and Methods” below. (C) AAUC plot for AVA1A1 MXs (OML), clustered by Lateral and Central PCs, and AVA1A1 MC (2'-OMe). Unpaired t test (Welch's correction for unequal SD) was performed between Lateral and Central PC. Refer to Fig. 1A legend. (D) pPSI dose response curves of AVA1A1.12 (Lateral PC, OML), AVA1A1.6 (Central PC, OML) and AVA1A1 (2'-OMe, MC) obtained in CT26 NATURA cells at 24 hours post-transfection. AAUC of UTR versus AVA1A1 MC is shaded for illustration purpose. (E) Single-cell flow-cytometry derived pPSI of ASGR+ dissociated cells from NATURA mice's livers. Samples were collected and processed 48 hours post-subcutaneous injection of SSOs (35mg / kg). Kruskal-Wallis test with Dunn's multiple comparisons (assuming non-Gaussian distribution). (F) Fold-change in pPSI change of OML MiXmers compared to OMe. Each datapoint represents an individual mouse.

[0019] Fig. 2 shows that the mixmers positional rule applies to a large set of SSO sequences targeting the NATURA gene. (A) AAUC plot for AVA1A1 , AVA1 B1 , AVA1 B2 and AVA1C2 MXs (OML), clustered by (refined) Lateral and Central PCs, and their respective MCs (2'-OMe) from the secondary screen in CT26 NATURA cells. Controls used include a scrambled sequence with mono-chemistry (SCR.MC) and with three LNAs in a lateral PC (SCR. MX), complementary sequences to AVA1A1 (cAVA1A1 , and CAVA1A1.12 with identical lateral PC as AVA1A1.12), AVA1B1 (CAVA1B1) and AVA1C2 (cAVA1C2). pPSI and cell count was obtained by flow cytometry with at least 500 cells for each replicate. Acquisition was performed 24 hours post-transfection. Each datapoint represents the average and SD of two independent biological replicates. Unpaired, nonparametric Mann-Whitney test (one-tailed) was performed between Lateral and Central PC. (B) Relative skipping potency of AVA1s OMLMXs in free uptake with calcium-enriched medium (CEM). Huh7 NATURA cells were incubated with 118nM of SSOs and analyzed by flow cytometry at 48h. The potency (fold change) of OML MX over OMe MC was obtained by normalizing the pPSI of MX relative to UTR with the pPSI of MC relative to UTR. Each dot represents the average of a biological duplicate. Unpaired t test (Welch's correction for unequal SD, one-tailed) was performed between Lateral and Central PC. (C) CEM dose-responses (pPSI) of representative OML MXs and their respective MCs in Huh7 NATURA cells at 48h post-incubation. AVA1A1.12, AVA1 B2.2 and AVA1B2.3 were modified with Lateral PCs. Analysis performed through flow cytometry, each dot represents the median of at least 500 cells. (D) Luminescence of Huh7 cells at48h post-incubation with AVA2 OML MXs and MC under CEM condition. The luciferase luminescence was first normalized to cell count which was detected using MTS assay and then normalized to the baseline intensity from untreated CEM condition. Mean and SEM of technical duplicates.

[0020] Fig. 3 shows that the mixmer improved positional configuration applies to SSO sequences targeting therapeutically relevant genes. (A) Schematics of the MXs modifications with either LNA or cET of the GLDC targeting SSO. The residues modified with constrained nucleic acids have been highlighted in dark grey. (B) Relative change in PSI of exon 7 of GLDC transcripts upon treatment with SSOs at 100 nM in CEM for 72 hours in A549 cells. Each dot represents a biological replicate. (C) PSI of SMN2 exon 7 upon transfection with Spinraza (20nM) and its MX permutations (in lateral positions) for 24 hours in SMN1 KO motoneurons. Percentage of improved inclusion of MXs in comparison to MC is shown above the bars. Each dot represents a biological duplicate with a technical duplicate each replicate. One-way ANOVA statistical analysis was carried out. (D) PSI of DMD exon 51 upon transfection with Drisapersen and its MX Lateral permutations in RH30 cells at 24 hours. Each dot represents a biological duplicate with a technical duplicate for each replicate. Two-way ANOVA with main effect only, Tukey's multiple comparison test with single pooled variance is shown for MC vs MX only.

[0021] Fig. 4 shows (A) AAUC plot for AVA1A1 MXs (MOL), clustered by Lateral and Central PCs, and AVA1A1 MC (2'-MOE). Transfection in CT26-NATURA cells. Refer to Fig. 1A legend. (B) pPSI of AVA1A1 MXs transfected at 37.5nM in CT26 cells, clustered by Lateral and Central PCs, plotted against their experimental melting temperature (Tm). (C) AAUC of pPSI of AVA1A1 OML MXs and OMe MC in CT26 cells, clustered by Lateral and Central PCs, plotted against their experimental melting temperature (Tm). (D) AAUC of pPSI of AVA1A1 MOL MXs and MOE MC in CT26 cells, clustered by Lateral and Central PCs, plotted against their experimental melting temperature (Tm). (E) TaqMan-derived PSI (tPSI) of AVA1A1.12(Lateral PC, OML), AVA1A1.6 (Central PC, OML) and AVA1A1 MC (OMe) obtained from 3T3 and CT26 NATURA cells at 24hour post-transfection at 37.5nM. Each bar represents a technical triplicate. (F) Manual cell count, normalized for Untreated (UTR), of the corresponding cells analyzed in panel E. Each dot represents an replicate measurement. (G) AAUC plot for AVA1A1 MXs (OML and MOL) and MCs (2'-0Me and 2'-M0E) obtained from B16F10, 3T3, HepG2, Huh7, and HeLa NATURA-expressing cell lines. Left panel: refer to Fig. 1A legend. Middle and right panels: AAUC plots clustered by Lateral and Central PCs for AVA1A1 MXs (OML) and AVA1A1 MXs (MOL) respectively. Refer to panel 2C for the statistical analysis. (H) Biological replicate of the experiment described in figure 1 E. Left panel: Each box represents a single mouse. Right panel: 3 mice receiving the same treatment (PBS, AVA1A1 OMe, AVA1A1.12 OML) are merged into a single box plot. Ordinary one-way ANOVA with Turkey's correction applied. (I) pPSI derived from flow cytometry of ASGR+ cells as described in figure 1 E, Each box plot represents an individual mouse, while each dot represents a single cell. Nested ANOVA. (J) Relative abundance of transcripts coding for EGFP (Ex2-3A), tRFP (Ex1-3A), and Luciferase (Ex2-3B) obtained from whole liver cDNA and subsequent TaqMan PCR. (K) Correlation between tPSI (obtained from the values in figure 4J) and pPSI (obtained from values in figure 4I). (L) Acute liver toxicity markers (ALT, ALP, AST) from blood of mice used in all the in vivo experiments described previously. Blood was collected at 48 hours post injections. Each dot represents a single mouse.

[0022] Fig. 5 shows (A) Schematics of the AVA1s target sites and their MXs modifications by lateral and central PCs. LNA substituted ribose are depicted in dark grey. For every number of substituted LNAs, there is at least one MX with a Lateral PC and a Central PC. A maximum 6 substituted LNAs at 5' was allowed for a lateral PC, with the remaining LNAs (if any) placed at 3', to avoid confounding as a central PC. Scrambled sequences are the reversecomplement of the AVA1 counterparts (e.g. AVA1 B1 has a non-targeting reverse-complement called NC / AVA1 B1 or CAVA1 B1). (B) Model summary for mixed-effects model mean AAUC pPSI — pF1*pF2*pF3 + pMid + pL3*pL2*pL1 + (oligo type I cell type), where the explanatory variables were the binary variables per SSO based on the presence or absence of LNA on each of the first and last three positions of the ASO (pF1, pF2, pF3, pL3, pL2, and pL1 ), an additional variable that counts the number of LNAs between the first and last three positions (pMid) and their interactions. The weights derived from this model was used to define a quantitative rule to differentiate SSOs with "lateral" and "central" PC. (C) AAUC plots for AVA1A1 , AVA1 B1 , AVA1 B2 and AVA1C2 MXs (OML), clustered by (refined) Lateral and Central PCs, and their respective MCs (OMe) from the secondary screen in B16F10, Huh7, HepG2 NATURA cells. Refer to Fig. 2A legend. (D) AAUC of cell viability from the screeningrepresented in figure 2A, clustered by LNA residues in each oligo. Each dot represents a single SSO in biological duplicate. Linear regression R2 is reported for Central and Lateral PC. (E) Relative exon-skipping potency of OML compared to their OMe counterpart (fold change). The data was obtained in the experiment described in figure 2A. Each dot represents the average of biological duplicates for a single oligo. Median, SD and maximum are shown for lateral and central PC clusters. Unpaired, nonparametric Mann-Whitney test (one-tailed) was performed between Lateral and Central PC. (F) Relative exon-skipping potency of AVA1A2 OML combinations compared to AVA1A2 OMe. (G) AAUC plot for AVA1A1, AVA1 B1 , AVA1 B2 and AVA1C2 MXs (MOL), clustered by (refined) Lateral and Central PCs, and their respective MCs (MOE) from the secondary screen in Huh7 NATURA cells in GEM conditions (free uptake). Refer to Fig. 2C legend. (H) Absolute potency of MOL MX compared to MOE MC and untreated. Data obtained from panel G. AVA1C2 relative potency cannot be calculated as the AVA1C2 MOE MC did not show any exon skipping. (I) CCL22 and FCLR3 RT-qPCR values of the best performing MX SSOs compared to their MC counterparts. The SSO are incubated at 4uM with BJAB cells for 48h. Each dot represents a distinct biological replicate.

[0023] Fig. 6 shows (A) PSI change of the exon targeted by the GLDC SSO in OMe MC or OML MX upon transfection with 5 nM of SSO in A549 cells. Each dot represents a biological replicate. (B) PSI change of the exon targeted by the GLDC SSO in MOE MC or MOL MX upon transfection at 5nM in A459 cells. Each dot represents a biological replicate. (C) PSI change of the exon targeted by the GLDC SSO in OMe, LNA, or cET MC compared to OML and OMe (OMe+cET) MX upon transfection at 25nM in A459 cells. Each dot represents a biological replicate. (D) Live cell count of SMN1 KO motoneurons after incubation with 100nM (left panel) and 250nM (right panel) of SSOs for 48 hours. (E) Relative live cell count of RH30 cells 24 hours post-transfection with 50nM of SSO. (F) PSI of DMD exon 51 upon incubation with Drisapersen (4pM, OMe MC) and its MX permutations in RH30 cells for 48 hours in CEM. Fold-change of MX chemistries compared to MC skipping is shown above the best performing SSOs. Each dot represents a biological duplicate with a technical duplicate each replicate. One-way ANOVA statistical analysis was carried out. (G) PSI of DMD exon 51 upon incubation with Drisapersen (4pM, MOE MC) and its MX permutations in RH30 cells for 48 hours in CEM. Fold-change of MX chemistries compared to MC skipping is shown above the best performing SSOs. Each dot represents a biological duplicate with a technical duplicate each replicate.

[0024] Fig. 7 shows (A) AVA1 A1 screening from figure 1 A with AAUC of pPSI clustered by percentage of GC bases having LNA over the total SSO bases. Each dot represents a single SSO with the top three most efficient SSOs in each cell line marked in light grey for all the following graphs. (B) AVA1A1 screening from figure 1A with AAUC of pPSI clustered bynumber of LNA substitutions present in the SSO. (C) AVA1A1 screening from figure 1A with AAUC of pPSI clustered by LNA base combination. (D) AVA1 A1 screening from figure 1 A with AAUC of pPSI ciustered by specific bases with LNA substitution. (E) AVA1 A1 screening from figure 1A with AAUC of pPSi clustered by LNA position pattern, PC. Unpaired T test was performed for this and ali the previous graphs. (F) AVA1 screening from figure 2A with AAUC of pPSi clustered by SSO nucleotides (nt) length. (G) AVA1 screening from figure 2A with AAUC of pPSi clustered by predicted Tm of the monochemistry SSO in 2'0Me ribose chemistry. (H)AVA1 screening from figure 2A with AAUC of pPSi clustered by percentage of GC bases with LNA substitution over the total SSO bases. (I) AVA1 screening from figure 2A with AAUC of pPSi clustered by number of LNA substitutions present in the SSO. (J) AVA1 screening from figure 2A with AAUC of pPSi clustered by LNA base combination. (K) AVA1 screening from figure 2A with AAUC of pPSi clustered by overall bases with LNA substitution.(L) AVA1 screening from figure 2A with AAUC of pPSi plotted against their experimental Tm.(M) AVA1 screening from figure 2A with AAUC of pPSi clustered by mixmer PC. Unpaired T test was performed.

[0025] Fig. 8 shows the components of a nucleotide amenable for chemical modifications.

[0026] Fig. 9 shows the classes of modified sugar moieties. Left: 2’-C modified sugars. Right: constrained sugars. Every ribose sugar in an steric blocker antisense oligonucleotide molecule was substituted with either type of moieties, one from each class.

[0027] Fig. 10 shows an embodiment of the CALIMERO gene. Top: An embodiment of the CALIMERO gene, where the number of spliced isoforms is 4. Exon 1 codes for the ATG start codon, Flag-tag, and a nuclear localization signal (NLS). Exon 2 is the FS exon, and is designed to be predominately spliced into the mature mRNA. Exon 3 is composed of the coding sequences of the fluorescent proteins turboRFP and EGFP in tandem, followed by an alternate acceptor splice site demarcating Exon 3B, which is composed of the coding sequences of His-tag and firefly luciferase in tandem. The codon reading frame of EGFP sequence is designed to be inframe with the FS exon, so that the protein is expressed only when FS exon is spliced-in. On the contrary, the turboRFP sequence is inframe when FS exon is spliced-out. The relative protein expressions of turboRFP and EGFP is the internally normalized readout for functional delivery. The codon reading frame of luciferase sequence is inframe with the FS exon while the His-tag sequence is inframe when FS exon is spliced-out. Exon 3A is designed to be preferentially spliced into the mature mRNA over exon 3B. Bottom: schematic representation of the binding sites of the 5 AVA1s on the FS (frame-shift) exon; FS exon exclusion will switch the expression of the CALIMERO gene from the EGFP to tRFP isoforms. SSO#2045 targets an intra-exonic region of FS whereas AVAIbs (SSO#2100 andSSO#2101) each targets the acceptor splice site and AVAIcs (SSO#2102 and SSO#2103) each targets donor splice site of FS.

[0028] Fig. 11 shows representative dose response curves. Dose response curve of OML mixmers are depicted in dark grey and for 2045_2OML in light grey, respectively. The ABC (area between curves) was computed by subtracting the areas, estimated via the trapezium rule, under the dark grey (OML) curve with the light grey (2OML) curve. ABC is thus negative when an OML mixmer is more efficacious than its 2OML counterpart over the range of SSO concentrations (Top Panel), ABC is positive when otherwise (Middle Panel), and is almost zero when there is no difference in potencies (Bottom Panel). Magnitude of ABC is indicative of the potency difference (gain or loss) of an mixmer over the 2OML version. Vertical axis: pPSI is ratio of EGFP protein to total fluorescent proteins (EGFP + tRFP), Figure 10. pPSI decreases with increasing efficiency of SSO#2045. Horizontal axis: SSO concentrations (nM).

[0029] Fig. 12 plotted as described in the figure legend of shows data for AVA2-SSO. Top: Refer to Figure 10 (top) legend. A A2-SSO was designed to skip exon 3A and spliced out the sequences encoding both tRFP and EGFP, which resulted in a transcript an in-frame luciferase sequence. Bottom: Luciferase activity measured in three cell lines (B16F10, CT26 and HepG2). Efficiency of the AVA2-SSO is measured from the level of luciferase induced compared to untreated (UTR). SSO#2087 and SSO#2167 were both modified with 2OMe ribose sugars whereas 2167.2 is an OML mixmer.

[0030] Fig. 13 shows data for GLDC-targeting SSOs. The efficiencies of SSO#2013 mixmers modified with “good”, intermediate” or “bad” positional sugar configurations were denoted by medium grey, dark grey and black circles. Efficiency of the parent SSO#181 (2OM) lead is denoted by a medium-dark grey circle. Efficiency of each SSO was inferred from the relative level of target exon (GLDC exon 7) that was normalized to the scrambled control (NC2). The SSOs were validated at various delivery conditions, namely transfection, calcium- enriched medium (OEM) to stimulate in vivo cell uptake, and without any helper reagent (“Free uptake”), on various cell lines namely A549, HuH7 and TS32 (NSCLC tumor sphere cells that are enriched in tumor stem cells), and at various concentrations. (A) OML mixmers. Ribose was substituted with either 2’-OMe or LNA. Efficiency of the 2’-OMe modified version of SSO#2013 is denoted by a light grey circle. The mean fold change in efficiency of each mixmer over its singly-modified counterpart is depicted below its respective efficiency plot. (B) MOL mixmers. Ribose of was substituted with either 2’-O-MOE or LNA. Efficiency of the 2’-O-MOE modified version of SSO#2013 is denoted by a light grey circle. The mean fold change in efficiency of each mixmer over its singly-modified counterpart is depicted below its respective efficiency plot. (C) OMc mixmers. Ribose was substituted with either 2’-OMe or (S)-cET.Efficiency of the 2’-0Me modified version of SSO#2013 is denoted by a light grey circle. (D) MOc mixmers. Ribose was substituted with either 2’-O-MOE or (S)-cET. Efficiency of the 2’- O-MOE modified version of SSO#2013 is denoted by a light grey circle.

[0031] Fig. 14 shows data for SLC25A13-targeting SSOs. The efficiencies of SSO#2034 mixmers modified with “good”, intermediate” or “bad” positional sugar configurations were denoted by medium grey, dark grey and black circles. Efficiency of the parent SSO#2008(2OM) is denoted by a medium-dark grey circle. Efficiency of each SSO was inferred from percentage of splicing correction attained. The SSOs were validated under transfection and calcium-enriched medium (OEM) to stimulate in vivo cell uptake, on both HuH7 and HepG2 cells, and at various concentrations. (A) OML mixmers. Ribose was substituted with either 2’-OMe or LNA. Efficiency of the 2’-OMe modified version of SSO#2034 is denoted by a light grey circle. The mean fold change in efficiency of each mixmer over its singly-modified counterpart is depicted below its respective efficiency plot. (B) MOL mixmers. Ribose was substituted with either 2’-O-MOE or LNA. Efficiency of the 2’-O-MOE modified version of SSO#2034 is denoted by a light grey circle. The mean fold change in efficiency of each mixmer over its singly-modified counterpart is depicted below its respective efficiency plot.

[0032] Fig. 15 shows data for OTC-targeting SSOs. The efficiencies of SSO#2044 mixmers modified with “good”, intermediate” or “bad” positional sugar configurations were denoted by medium grey, dark grey and black circles. Efficiency of the parent SSO#2021(2OM) lead is denoted by a medium-dark grey circle. Efficiency of each SSO was inferred from percentage of splicing correction attained. The SSOs were validated under transfection and calcium-enriched medium (OEM) to stimulate in vivo cell uptake, on both HuH7 and HepG2 cells, and at various concentrations. The mean fold change in efficiency of each mixmer over its singly-modified counterpart is depicted side-by-side. (A) OML mixmers. Ribose of SSO#2044 was substituted with either 2’-OMe or LNA. Efficiency of the parent SSO#2021(2OM) is denoted by a medium-dark grey circle. The mean fold change in efficiency of each mixmer over its singly-modified counterpart is depicted side-by-side. (B) MOL mixmers. Ribose of SSO#2044 was substituted with either 2’-MOE or LNA. The mean fold change in efficiency of each mixmer over its singly-modified counterpart is depicted side-by- side.

[0033] Fig. 16 shows an embodiment of a lateral positional configuration.

[0034] Fig. 17 shows an extension of more ribose moieties studied. (A) Chemical structure of all the ribose moieties used in the study. (B) Positional configurations of the mixmers tested. The sequence is the 2045 nucleotide sequence (5’-GTCCAGAGCTTTCATTCTGT-3’) (SEQID NO: 1). O indicates a 2’-0Me, while Y is used to show where the other chemistries are introduced.

[0035] Fig. 18 shows the on-target potency of mixmers with more types of ribose moieties. Splice-switching efficiency of lipofectamine-transfected 2045 SSO in 3T3 cells with the various mixmers configurations. Highlighted in grey rectangular boxes the oligos 2045.333 and 2045.444 in “central configuration”. All other oligos are in “lateral configuration”. The spliceswitching efficiency has been normalized for the 2045.101 in OML lateral configuration, which is the best-performing mixmer from the previous study in 2’-Ome+LNA).DETAILED DESCRIPTION OF THE PRESENT INVENTION

[0036] Chemical modification of natural sugars has been a major breakthrough that enabled multiple nucleic acid therapeutics (NATs) to successfully enter the clinic. This implementation evolved by substituting the sugars with multiple ribose chemistries in specific patterns within the same oligo, with improvements in stability, specificity and safety profile of siRNAs.

[0037] More recently, researchers have started to apply a similar chemical modification strategy on another class of NATs called Splice-Switching Oligonucleotides (SSOs), which are designed to alter specific splicing events. The present disclosure uncovers a set of mixed- modification patterns involving 2’-O-methyl (2’-OMe), 2’-O-m ethoxyethyl (2’-MOE), locked nucleic acid (LNA), constrained Ethyl (cET), ethylene-bridged nucleic acid (ENA), 2’-Fluoro, 2’-Fluoroarabino nucleic acid (2’-FANA), 2’-O-N-methylacetamido (2’-O-NMA), threose nucleic acid (TNA), hexitol nucleic acid (HNA) and methyleneoxy nucleic acid (MNA) ribose moieties, that confers significantly higher SSO potency, which the present inventors defined as Lateral PC (positional configuration). Remarkably, these patterns work across different SSO sequences, suggesting a universal strategy to enhance SSO-based treatments and streamline their development by reducing the need for extensive empirical testing of modifications.

[0038] The present inventors deployed the NATURA system (Figure 10, Figure 12) to quantitate the absolute potencies of SSO mixmers whose sugars were substituted with LNA and 2’OMe or 2’O-MOE in high-throughput. Specific rules of PCs that resulted in striking potencies and reduced cell toxicity, as compared to the uniformly-modified counterparts, were discovered. Further investigation showed that they were not only agnostic to SSO sequences but are extensible to the (S)-cET moiety (Figure 9). Advantageously, the existence of suchrules as described in the present disclosure eliminates the need to screen many PC permutations.

[0039] In one aspect, the present disclosure provides a method of identifying a chemical modification pattern which confers an increased splice-modulation efficiency to a singlestranded RNA oligonucleotide with a predetermined nucleobase sequence, as compared to a single-stranded RNA oligonucleotide with the same predetermined nucleobase sequence having every nucleotide chemically modified with a 2’ ribose modification, the method comprising: (a) applying a plurality of chemical modification patterns (Patterns A to G) to a plurality of single-stranded RNA oligonucleotides respectively to result in a plurality of chemically-modified single-stranded RNA oligonucleotides, wherein each of the plurality of single-stranded RNA oligonucleotides has one of Patterns A to G applied to it, and wherein the plurality of single-stranded RNA oligonucleotides have the same predetermined nucleobase sequence, (i) wherein each of the plurality of chemical modification patterns comprises: (A) at least one nucleotide chemically modified with the 2’ ribose modification; and (B) one to eight nucleotides chemically modified with a constrained nucleotide (CN); and (ii) wherein the plurality of chemical modification patterns comprises the following positional configurations of CN-modified nucleotides: (A) Pattern A comprising a CN cluster comprising one or more consecutive CNs, beginning at position +1 , +2 or +3 of the single-stranded RNA oligonucleotide; (B) Pattern B comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the single-stranded RNA oligonucleotide, and a second CN cluster comprising one or more consecutive CNs, ending at position -2 or - 1 of the single-stranded RNA oligonucleotide; (C) Pattern C comprising a CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to the CN cluster and that is not at position -2 or -1 of the single-stranded RNA oligonucleotide; (D) Pattern D comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the single-stranded RNA oligonucleotide, and a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to both the first and second CN clusters; (E) Pattern E comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 , and a second CN cluster comprising one or more consecutive CNs, ending at position -1 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters; (F) Pattern F comprising a first CN at position +3 and a second CN at position -2 of the single-stranded RNA oligonucleotide; and (G) Pattern G comprising a first CN at position +3 and a second CN atposition -1 of the single-stranded RNA oligonucleotide; and (b) measuring the splicemodulation efficiencies of the chemically-modified single-stranded RNA oligonucleotides and comparing these with the splice-modulation efficiency of the single-stranded RNA oligonucleotide having every nucleotide chemically modified with the 2’ ribose modification to determine the chemical modification pattern which confers the increased splice-modulation efficiency.

[0040] It would be generally appreciated by the skilled person that the method of identifying a chemical modification pattern which confers an increased splice-modulation efficiency of the present disclosure may be done in a single experiment, in multiple sub-experiments, or in multiple separate experiments. In other words, the method of the present disclosure may be carried out over a series of experiments or multiple stages of an experiment.

[0041] As used herein, the term “efficiency” is meant to include splice-modulation efficiency such as splice-switching efficiency as well as the efficiency in steric blockage of RNA-binding protein to their cognate binding sites. In one embodiment, the term “splice-switching efficiency” is meant to include the capability of an oligonucleotide to modulate RNA splicing. The modulation of RNA splicing includes but is not limited to the induction of exon exclusion, induction of exon inclusion, usage of alternative 5’ or 3’ splice sites, intron retention or restoration, and selective inclusion of one of a pair of mutually exclusive exons. In one embodiment, the baseline of comparison of efficiency (such as splice-switching efficiency) is the RNA oligonucleotide that is fully modified with 2’ ribose (mono-chemistry) (i.e. a singlestranded RNA oligonucleotide having every nucleotide chemically-modified with a 2’ ribose modification). A “higher” or “increased” efficiency can be defined by three non-mutually exclusive outcomes, which can be measured by the abundance, expression or composition of target. First, a mixed-chemistry oligonucleotide (mixmer) has an “increased efficiency” when it is observed to modulate its target in the desired manner at a specified magnitude at a lower dose than the mono-chemistry counterpart having the same base sequence; e g. EC50 (mixmer) < EC50 (mono-chemistry). Second, a mixed-chemistry oligonucleotide (mixmer) has an “increased efficiency” when it is observed to achieve a specified magnitude of target modulation for which the mono-chemistry counterpart cannot achieve (no matter how high the dose is); e.g. the mixmer can achieve EC90 but not its mono-chemistry counterpart. Third, a mixed-chemistry oligonucleotide (mixmer) has an “increased efficiency” when it is observed to achieve a higher average magnitude of target modulation over a specified range of dose than the mono-chemistry counterpart having the same base sequence. These three properties can occur in a mixmer. Accordingly, the “highest efficiency” refers to the chemically-modifiedoligonucleotide that provides the greatest increase in efficiency when compared to its monochemistry counterpart having the same nucleobase sequence.

[0042] As used herein, the term “a first CN cluster” in the phrase “Pattern D comprising a first CN cluster” refers to the first CN cluster in Pattern D, i.e. Pattern D comprising a first CN cluster of Pattern D. In other words, the term “a first CN cluster” in Claim 1 (a)(ii)(D) does not refer to the same CN cluster as the one appearing in Claim 1 (a)(ii)(A), for example. The phrase “Pattern D comprising... a second CN cluster” refers to the second CN cluster in Pattern D. The term “an additional CN” in the phrase “an additional CN that is not adjacent to both the first and second CN clusters” refers to an additional CN in Pattern D that is separate from both the first and second CN clusters of Pattern D.

[0043] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern H) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern H comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide, and a second CN cluster comprising two or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide.

[0044] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern I) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern I comprising the following positional configuration of CN-modified nucleotides: a CN cluster comprising (i) one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide; and (ii) an additional CN that is not adjacent to the CN cluster and that is not at position -2 or -1 of the singlestranded RNA oligonucleotide.

[0045] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern J) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern J comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to both the first and second CN clusters.

[0046] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern J) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern J comprising the following positionalconfiguration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to both the first and second CN clusters.

[0047] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern L) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern L comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +1 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters.

[0048] In one embodiment, step (a) further comprises applying a further chemical modification pattern (Pattern M) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern M comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +2 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters.

[0049] In one embodiment, the 2’ ribose modification is selected from the group consisting of 2’-O-methyl (2’-OMe) and 2’-O-methoxyethyl (2’-MOE).

[0050] In one embodiment, the constrained nucleotide (CN) is selected from the group consisting of locked nucleic acid (LNA), S-constrained ethyl ((S)-cET), ethylene-bridged nucleic acid (ENA), 2’-O-N-methylacetamido (2’-O-NMA), threose nucleic acid (TNA), hexitol nucleic acid (HNA), methyleneoxy nucleic acid (MNA), 2’-Fluoro and 2’-Fluoroarabino nucleic acid (2’-FANA).

[0051] In one embodiment, all nucleotides that are not CN-modified comprise the 2’ ribose modification.

[0052] In one embodiment, the single-stranded RNA oligonucleotide is between 15 and 22 nucleotides in length.

[0053] In one embodiment, the nucleobase sequence of the single-stranded RNA oligonucleotide has a melting temperature of 70°C or below.

[0054] In one embodiment, the melting temperature (Tm) of an oligonucleotide refers to the temperature at which 50% of the oligonucleotide is bound to its complement (e.g. temperature at which 50% of a splice-switching oligonucleotide is bound to its target mRNA) and 50% is free in solution (e.g. 50% of the splice-switching oligonucleotide is single-stranded or selfstructured).

[0055] By “oligonucleotide”, it is meant to refer to any polynucleotide. A "polynucleotide" is an oligomer comprised of nucleotides. A polynucleotide may be comprised of DNA, RNA modified forms thereof, or a combination thereof. The term "nucleotide" or its plural as used herein is interchangeable with modified forms as discussed herein and otherwise known in the art. In certain instances, the art uses the term "nucleobase" which embraces naturally occurring nucleotides as well as modifications of nucleotides that can be polymerized. Thus, nucleotide or nucleobase means the naturally occurring nucleobases adenine (A), guanine (G), cytosine (C), thymine (T) and uracil (U) as well as non-naturally occurring nucleobases such as xanthine, diaminopurine, 8-oxo-N6- methyladenine, 7-deazaxanthine, 7- deazaguanine, N4,N4-ethanocytosin, N',N'-ethano-2,6- diaminopurine, 5-methylcytosine (mC), 5-(C[3]- C6)-alkynyl-cytosine, 5-fluorouracil, 5- bromouracil, pseudoisocytosine, 2- hydroxy-5-methyl-4-tr- iazolopyridin, isocytosine, isoguanine, inosine and the "non-naturally occurring" nucleobases described in Benner et ah, U. S. Pat. No. 5,432,272. The term "nucleobase" also includes not only the known purine and pyrimidine heterocycles, but also heterocyclic analogues and tautomers thereof. Further naturally and non-naturally occurring nucleobases include those disclosed in U. S. Pat. No. 3,687,808 (Merigan, et al), which is hereby incorporated by reference in its entirety. In various embodiments, polynucleotides also include one or more "nucleosidic bases" or "base units" which include compounds such as heterocyclic compounds that can serve like nucleobases, including certain "universal bases" that are not nucleosidic bases in the most classical sense but serve as nucleosidic bases. Universal bases include 3-nitropyrrole, optionally substituted indoles {e.g., 5-nitroindole), and optionally substituted hypoxanthine. Other desirable universal bases include pyrrole, and diazole or triazole derivatives, including those universal bases known in the art.

[0056] As used herein, the term “splice switching oligonucleotides” (SSOs) or “splice switching oligomers” is meant to include synthetic antisense nucleic acids that base-pair with a pre-mRNA and disrupt the splicing process by sterically blocking the RNA-RNA base-pairing or protein-RNA binding interactions that occur between components of the splicing machinery and the pre-mRNA. The SSOs may also be known as “antisense nucleotides”, “steric blockers” or “steric hindrance antisense nucleotides" which can modulate splicing. SSOs may modulate splicing via steric blocking. In some embodiments, SSOs may be mixmers. The term “mixmer”includes an oligomer on which different types of chemical modifications are applied on its sugar moieties, or on its backbone linkages, or both. Examples of chemical modifications include phosphorothioate linkages, 2’-O-methyl RNA modifications, 2’-O-methoxyethyl RNA modifications and locked nucleic acid substitutions. The terms “phosphorothioate bond” and “phosphorothioate linkage” are used interchangeably. The chemical modifications may increase the efficacy, selectivity and stability while manifesting superior toxicity profile of SSOs.

[0057] The term “splicing” refers to an RNA processing mechanism in which a pre-m RNA is made into a mature mRNA. During splicing, introns are removed and exons are connected. Splicing is catalysed by the spliceosome complex. As used herein, the term “alternative splicing” is meant to include a process by which a gene can encode for multiple mRNA and protein products by differentially selecting which exons are to be included in a mature mRNA transcript. For example, alternative splicing can take the form of one or more skipped exons, variable position of intron splicing, or intron retention.

[0058] As used herein, the term “intron” refers to a segment of non-coding nucleic acid sequence that is transcribed and is present in the pre-mRNA but is excised by the splicing machinery and therefore not present in the mature mRNA transcript.

[0059] As used herein, the term “exon” refers to a segment of a nucleic acid sequence that is transcribed into mRNA and that is present in mature mRNA after splicing. The term “exon skipping” is meant to include the process by which an entire exon, or a portion thereof, is removed from a given pre-mRNA and is thereby excluded from being present in the mature mRNA. For example, the portion of the protein that is otherwise encoded by the skipped exon is not present in the expressed form of the protein.

[0060] As used herein, the term “splice site” is meant to include specific nucleic acid sequences that can be recognized by the splicing machinery as being suitable for excision and / or ligation with the corresponding splice site. The splice site defines the precise exonintron boundary that allows the excision of introns present in pre-mRNA transcripts. As used herein, the term “5’ splice site” (also known as donor splice site) refers to a nucleic acid sequence surrounding the exon-intron boundary at the 5’ end of an intron that marks the start of the intron and its boundary with the preceding exon sequence. The term “3’ splice site” (also known as acceptor splice site) as used herein refers to a nucleic acid sequence surrounding the intron-exon boundary at the 3’ end of an intron that marks the end of the intron and its boundary with the following exon sequence.

[0061] As used herein, the term “pre-mRNA” or “precursor mRNA” refers to a strand of messenger ribonucleic acid (mRNA), synthesized from a DNA template by transcription. Pre-mRNA is composed of exons, introns and untranslated sequences (before the first and after the last exons respectively). Generally, eukaryotic pre-mRNA exists only briefly before it is fully processed into mature mRNA.

[0062] The term “binding” as used in the context of an SSO is meant to include the hybridization of the SSO to a site within a target region on a pre-mRNA transcript. The term “hybridize” or “hybridization” may include the binding of a single-stranded nucleic acid or a locally single-stranded region of a double-stranded nucleic acid to another single-stranded nucleic acid or a locally single-stranded region of a double-stranded nucleic acid having a complementary sequence through the pairing of complementary nucleic acids. It is generally known to a person skilled in the art that binding or hybridization of one sequence to another does not require total complementarity of the sequences. For example, the sequence of the SSO may be completely complementary or partially complementary to the target region to which it binds.

[0063] Advantageously, the present disclosure provides various inventive aspects. In one embodiment, the present disclosure provides exemplary positional configurations (PCs) as shown in Figures 1 to 5.

[0064] In another embodiment, the present disclosure provides a regression model that correlates and predicts if a PC is “Good” or “Bad”. This is shown at “Linear regression model” under “Material and Methods”. In another embodiment, the present disclosure provides optimal LNA counts. The optimal LNA counts for a SSO sequence modified with OML can be estimated from its natural sequence’s melting temperature (Tm). Refer to Table 2. The optimal LNA counts for a SSO sequence modified with MOL is one less than OML. Refer to Table 3. In another embodiment, the present disclosure provides the Tm threshold, which is an upper Tm limit of a SSO’s natural sequence beyond which OML or MOL modification will not improve its potency. Refer to Table 2.

[0065] The invention illustratively described herein may suitably be practiced in the absence of any element or elements, limitation or limitations, not specifically disclosed herein. Thus, for example, the terms "comprising", "including", "containing", etc. shall be read expansively and without limitation. Additionally, the terms and expressions employed herein have been used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention claimed. Thus, it should be understood that although the present invention has been specifically disclosed by preferred embodiments and optional features, modification and variation of the inventions embodied therein herein disclosed may beresorted to by those skilled in the art, and that such modifications and variations are considered to be within the scope of this invention.

[0066] The invention has been described broadly and generically herein. Each of the narrower species and subgeneric groupings falling within the generic disclosure also form part of the invention. This includes the generic description of the invention with a proviso or negative limitation removing any subject matter from the genus, regardless of whether or not the excised material is specifically recited herein.

[0067] Other embodiments are within the following claims and non- limiting examples. In addition, where features or aspects of the invention are described in terms of Markush groups, those skilled in the art will recognize that the invention is also thereby described in terms of any individual member or subgroup of members of the Markush group.

[0068] MATERIAL AND METHODS

[0069] Cell culture and transfection

[0070] The human embryonic kidney 293T (HEK293T), 3T3 (mouse embryonic fibroblasts), B16F10 (murine melanoma), BJAB (Burkitt lymphoma B cell), HeLa (human cervical carcinoma), HepG2 (human hepatocellular carcinoma), and Huh7 (human hepatocellular carcinoma) cell lines were procured from the American Type Culture Collection (ATCC). The CT26 (mouse colon carcinoma) and patient derived iPSCs (SMN1 knock-out) cell lines were kindly provided by Dr. Manikandan Lakshmanan and Dr. Shi Yan NG respectively, of the Institute of Molecular and Cellular Biology in Singapore. These cell lines were propagated in accordance with ATCC guidelines.

[0071] To generate stable NATURA-expressing cell lines, 3 x 105cells were seeded in 6- well plates containing 1 mL of complete growth medium the day prior to transfection. Transfection was carried out using 4.5 pg of NATURA plasmid and 0.5 pg of PRP-mCherry- CAGhybase plasmid in conjunction with the Lipofectamine 3000 reagent after 16-20 hours of incubation. The following day, the medium was replaced with complete growth medium supplemented with Blasticidine antibiotic (1 :1000, Sigma-Aldrich, cat- 15205) and the cells were grown under selection for 14 days. Subsequently, the cells were sorted based on EGFP expression.

[0072] Reverse transfection in 96-well plate

[0073] 5 pL of ASO, at the required concentrations, was combined with 20 pL of Optimem(#31985062; Thermo Fisher Scientific) in a 96-well cell culture treated plate. For each well, a mixture of 25 pL of Optimem and 0.9 pL of Lipofectamine RNAiMAX transfection reagent (#13778030; Thermo Fisher Scientific) was aliquoted. The lipid-ASO complex was allowed toform by incubating the plate for 20 minutes at room temperature on a gentle shaker (300 rpm). Meanwhile, cells were trypsinized, washed in PBS, and resuspended in complete growth medium at the final concentration of 20,000 cells / 50 pL and 10,000 cells / 50 pL of media for 24 hours and 48 hours of incubation, respectively. 50 pL of the cell suspension was then added to the ASO-RNAiMAX 96-well plate, shaken gently for 10 seconds at 300 rpm, and then placed in a humidified incubator. Cells were collected and processed further after 24 hours, 48 hours, and 72 hours for SSO, GAPmers or siRNA, and CRISPR system analyses, respectively, unless stated otherwise.

[0074] High content imaging

[0075] AVA1 SSO candidates were screened by reverse transfecting 4 NATURA expressing cell lines (CT26, B16F10, 3T3 and Huh7) at 4 different concentrations- 6 nM, 15 nM, 37.5 nM and 93.75 nM. After 24hrs from transfection, cells were first fixed using 4% PFA and then stained with Hoechst 33342 stain (1 :3000) (#H3570; Invitrogen) for 15 mins at room temperature. Images were captured with Opera Phenix high content screening system and analyzed with Columbus software. The same pipeline was used for the analysis of all samples.

[0076] FACS analysis

[0077] Cells were trypsinized and washed in PBS. Collected cells were resuspended in ice-cold 5% FBS / PBS and analyzed within 1 hr using the Becton & Dickinson LSRII or Penteon-Novocyte Agilent L5 analyser. The HTS module was used for analysing 96-well plates.

[0078] The laser settings forthe EGFP and tRFP channels were kept consistent throughout the experiments and across different cell lines. For experiments reguiring absolute count, events / sec were recorded while keeping the volume of sample constant.

[0079] Luciferase assay

[0080] Live cells were stained by adding 0.4% trypan blue (1 :1 , #15250061 ; Gibco) and counted using Countess II FL automated cell counter (Thermo Fisher Scientific). Luciferase assay (#E1960; Promega) was performed by lysing 1 x 105cells in a flat white- 96 well plate with 22 pl of 1x passive lysis buffer at room temperature for 15 mins. Subsequently, the sample was incubated with 100 pl of luciferase assay substrate for 5 mins in humidified incubator. Readings were taken with Tecan multimode spark 10M plate reader.

[0081] MTS assay

[0082] Cells were trypsinized and resuspended in 80ul of complete growth medium with 16 pl of CellTiter 96® AQueous One Solution Reagent (#G3580; Promega). Absorbance was measured at 490nm after 1 hr of incubation in a humidified incubator at 37°C using Tecan multimode spark 10M plate reader.

[0083] Genescan

[0084] gDNA was extracted from cells with Qiagen DNeasy kit (#69506; Qiagen) as per the manufacturer’s instructions. Fragment length analysis of the PCR was performed by running the amplicons through QSEP 100 capillary electrophoresis apparatus.

[0085] SSO toxicity assay in BJAB cells

[0086] In a 96-well plate, 0.5 x 106BJAB cells were incubated with 2 pM of SSO in 100 pl of cell culture medium for 24 hours. The next day, RNA was extracted following the Trizol protocol (as per manufacturer instructions), and 500 ng of RNA were reverse-transcribed into cDNA using the Maxima kit (#K1642; Thermo Scientific). The cDNA obtained was then analyzed using a SYBR green qPCR Assay (#A25742; Applied biosciences) with primers and probes specified in Table 6. Each reaction mix consisted of 10 pl of 2x SYBR green master mix, 2 pl of primer mix (10 pM forward and reverse primers), and 8 pl of cDNA (3.33 ng / pl). The qPCR assays were performed on a Biorad CFX96 Real-time system using the following cycling conditions: 10 minutes at 95°C, followed by 40 cycles of 15 seconds at 95°C and 1 minute at 55°C. The data were analyzed using GraphPad Prism 9.

[0087] Differentiation of motor neurons

[0088] On Day 0 of the differentiation process, 1 x 106pluripotent IPSCs were placed in a standard 10 cm dish pre-coated with Matrigel, using iPS brew (#130-107-086; Miltenyi Biotec) with 5 pM ROCK inhibitor Y-27632. The next day, the media was switched to Neural Induction Media (NIM), composed of a basal mixture (50% DMEM / F12 (#11320-033; Gibco), 50% Neurobasal media (#130-093-570; Miltenyi Biotec), 1x GlutaMAX (#35050-061 ; Gibco), 1x non-essential amino acids (#11140-050; Gibco), 1x N2 (#17502048; Gibco), 1x Neurobrew (#130-093-566, Miltenyi Biotec)), and supplemented with 0.5 pM LDN193189 (#72147; Stem Cell Tech) and 4.25 uM CHIR99021 (#130-104-172; Miltenyi Biotec). Starting from Day 3, NIM was further enriched with 1 pM retinoic acid (RA; R2625; Sigma). Motor neuron progenitor cells were expanded from Day 10 onward in motor neuron progenitor expansion medium (MNPE), which included basal medium supplemented with 1 pM RA and 1 pM purmorphamine (#130-104-465; Miltenyi Biotec). Beginning on Day 17, the cells were cultured in motor neuron maturation medium, which contained basal medium supplemented with 10 ng / ml GDNF (#130-129-546; Miltenyi Biotec), 10 ng / ml BDNF (#130-103-435; Miltenyi Biotec), and 200 pM ascorbic acid (#A4544; Sigma). The motor neurons became ready for use in experiments starting from Day 25.

[0089] Taqman assay

[0090] The cDNA obtained was analysed by Taqman Assay (#4304437; Thermo Fisher Scientific) using primers and probes specified in Table 6.

[0091] Each reaction mix contained 10 pl of 2x Taqman master mix, 1 pl of NATURA primer-probe mix ( 2.5 pM probes and 10 pM primers), 1 pl of Cyclophilin A primer-probe mix ( 2.5 pM probes and 10 pM primers) and 8 pl of cDNA (3.33 ng / pl). The assays were run on Biorad CFX96 Real time system using cycling conditions- 2 mins at 95°C followed by 40 cycles of 5 secs at 95°C and 40 secs at 60°C with no dissociation curve. The data was analysed using Graphpad prism 9.

[0092] tPSI Calculation

[0093] tPSI (percentage spliced-in) measures the fraction of FS-containing NATURA mature transcripts (Figure 10, Figure 12), which is derived from Taqman qPCR by calculating the abundance of the qPCR probe spanning FS to Exon 3A (PR2-3A) over the sum of all the probes (PR2-3A + PR2-3B + PR1-3A). Analogously, pPSI (EGFP) measures the fraction of EGFP protein (translated from FS-containing NATURA transcripts) that is obtained from the mean fluorescence intensities (MFIs) of EGFP and tRFP proteins (translated from NATURA transcripts with skipped FS exon), i.e. pPSI (EGFP) = MFI(EGFP) / [MFI(EGFP) + MFI(tRFP)]

[0094] Linear regression model

[0095] A numerical score was assigned to a MX based on the presence of LNA substituted at the first, second, and third 5’ positions (pF1 , pF2, and pF3 respectively), and at the last, second last, and third last 3’ positions (pL1 , pL2, and pL3 respectively). pMid denotes the number of LNA substitutions in between the first three and last three positions. The PC score below correlated with the MX efficiency (N02_model_fitting.model_visualization). By assigning the “lateral pattern” to oligos with the positional score higher than 5, the present inventors observed a significant clustering of the best performing oligos in this category (Figure 2A, FigurePC sco+ (1.4602 * pL3) 1- (-0.4381 * pL2) + (2.4611 * plA) + (pFl * pF2 * —7.1709)

[0096] Statistical Analysis

[0097] The statistical analysis for each figure is listed in the figure legend. All the statistical analyses were performed with GraphPad Prism Software (Version 10.2).

[0098] EXAMPLES

[0099] Non-limiting examples of the invention and comparative examples will be further described in greater detail by reference to specific Examples, which should not be construed as in any way limiting the scope of the invention.

[0100] Example 1 : Enhanced SSO potencies from specific positional configurations (PCs) of ribose chemistry

[0101] The NATURA system and the SSO tool library (AVA1s and AVA2s, Table 1) were used to investigate effects of PCs on a SSO’s potency and its cellular toxicity. Every ribose sugar in AVA1 , labelled hereafter as AVA1 A1 , was applied either with 2’-OMe or LNA moieties (OML MiXmers, or OML MX for short), or with 2’-MOE or LNA moieties (MOL MiXmers, or MOL MX for short). To avoid confounding effects from considerable variations in melting temperature (Tm), every mixmer contains three to four LNAs (15% to 20% of the total length). 44 mixmers, each representing a PC permutation were considered in the initial study, with 22 in OML and 22 in MOL chemistry. For every mixmer and its mono-chemistry (MC) counterpart (OMe MC or MOE MC), the dose response of FS exon skipping efficiency, as measured by pPSi, and the cellular toxicity, as measured by ceil count, were determined in biological duplicates using a high content imaging confocal microscope at 24h, over four concentrations of SSO transfected on NATURA expressing cell line. To quantitate the effects of a PC on potency and cellular toxicity over the entire range of SSO concentrations, AAUC, defined as the cumulative difference of the area under the curve (AUC) in the dose responses of pPSI or cell viability between a mixmer and untreated (UTR) cells (see Material and Methods, Figure 11), was determined.

[0102] in NATURA-expressing CT26 cells, both OML and MOL MXs had lower cellular toxicities than their respective OMe and MOE MCs (Fig.lA horizontal axis). The effect on potency was mixed (Fig.lA vertical axis) - whereas all MOL MXs were less efficient (i.e. lower AAUCs) than MOE MC, the relative efficiency of OML MXs were either augmented or diminished compared to OMe MC. Amongst all features analysed (% of GC containing LNA, LNA nucleotide compositions and combinations; Fig. 7A, 7B, 7C, 7D), only the relative LNA- substituted ribose positions within the MXs could correlate with the efficiencies for OML and MOL MXs, independently. MXs modified with lateral PCs, which specify LNA placements at the 5’ first and / or second ribose (Fig.lB), have higher skipping efficiency (higher AAUCs) than MXs modified with Central PCs whose LNA are placed elsewhere (Fig.lC, Fig. 7E). This and the following observations hold true for both OML and MOL MXs (Fig.4A). While Central PC MXs tend to have higher melting temperatures (Tm), no correlation with efficiency (both in pPSi and AAUC) was observed (Fig. 4B, 4C, 4D). Although MXs harbouring either PC patterns have similar cellular toxicity profiles, they are less toxic than the MC counterparts. As anillustrative example, the pPSI dose response curves of OML AVA1A1.12 (highest AAUC, Lateral PC), AVA1A1.6 OML (lowest / AAUC, Central PC) and AVA1A1 (OMe MC) demonstrates the superior efficiency of AVA1A1.12 across the SSO concentrations consistently (Fig.lD), which was further confirmed from repeated transfection and mRNA quantification by Taqman qPCR (Fig.4E). Cell viability from AVA1A1.12 and AVA1A1.6 transfection was also improved over the MC counterpart (Fig.4F).

[0103] The clustering feature of Lateral and Central PCs was found to be conserved as the MXs were screened on five more NATURA-expressing cell lines (B16F10, 3T3, HepG2, Huh7 and HeLa) (Fig.4G). To validate the on-target activity of MX and also to characterize the sensitivity of NATURA system in vivo, saline suspended AVA1A1.12 (OML MX) and AVA1A1 (OMe MC) were subcutaneously administered only singly on NATURA mice whose livers dissected after 48 hours were dissociated and stained for ASGR1 hepatocytes marker. Consistent with cell culture observations, AVA1A1.12 switched EGFP to trip in hepatocytes at 4.7 and 1.9folds higher than its sonochemistry counterpart at low (35mg / kg) and high dose respectively (100 mg / kg, Fig.l E, Fig.4H, 41, Fig.1 F). In the latter dose, RNA was extracted from a portion of whole liver, and an average of 4.5foids higher switch (tPSI) for the MX than the MC was detected by TaqMan PCR. The higher switch measured from topsy in liver compared to pPSI in hepatocytes (Fig.4J) could be due to uptake from other cell types, longer half-life of EGFP protein compared to mRNA, or potential underestimation of the switching efficacy caused by autofluorescence in the GFP channel. Nevertheless, a correlation between pPSI in hepatocytes and tPSI for the whole liver was discerned (Fig.4K). Lastly, no acute hepatotoxicity was observed at both doses (Fig.4L).

[0104] To expand the PCs' adaptability on more sequences and base compositions, and to investigate PCs over a range of substituted LNAs, three additional AVA1 (AVA1 B1 , AVA1 B2 and AVA1C2) were rationally designed. Both AVA1B1 and AVA1 B2 bind to FS exon 3’SS (acceptor) while AVA1C2 binds to the corresponding 5’SS (donor). Together with AVA1A1 (which bind to an ESE), the four SSOs were systematically modified with either Lateral or Central PCs, with one to eight LNA substitutions as OML MXs (Fig.5A). Their efficiencies and cellular toxicities were determined and analysed as before, with the fluorescence and cell count measured by flow cytometry for increased throughput and sensitivity. With linear regression methods applied on the enlarged dataset (Fig.5B, Material and Methods), the MX efficiencies are again differentiated only by LNA positions, while other parameters like SSO length, Tm, GC%, number of LNA residues or LNA base combinations are not associated with potency (Fig. 7F, 7G, 7H, 7I, 7 J, 7K, 7L, 7M). In detail, while LNA substitutions at the first three 5’ ribose remain the biggest contributor, substitutions at the last two or last 3’ ribose aresynergistic (Fig.5B). Thus, the clustering feature of Lateral (refined) and Central PCs was corroborated in the secondary screen (Fig.2A, Fig.5C. Fig.12), with the most efficient SSO for each AVA1 sequence modified with a Lateral PC. The only exception (out of 24 experimental conditions) is AVA1A1 in B16F10 cells in which Lateral PC AVA1A1.10 is the third most efficient. However, the most efficient MXs (AAUCs pPSI = 60.6 and 59.9), both with Central PC, are only marginally more efficient than AVA1A1.10 (AAUC pPSI = 56.6) but are considerably more toxic (AAUCs cell viability = -38.5 and -25.9 versus -12.2) (Fig. 5C, first panel). Notably, Lateral PC MXs have superior efficiency than their MC counterparts with up to 500folds (Fig.5E, AVA1 B2 in B16F10 cells). The results were further confirmed as the screen was repeated under free uptake in calcium chloride-enriched media (CEM, Fig.2B. Cell viability not shown as no significant toxicity was observed for all the SSOs). Fig.2C depicts the in-depth dose response of representative MXs modified with Lateral PCs (AVA1A1.12, AVA1 B2.2 and AVA1B2.3) and their respective MCs, performed to validate the CEM screening. Notably, among MXs modified with Lateral PCs, their relative order sorted by efficiencies was generally consistent between CEM and transfection conditions. On the other hand, no association between cell viability and number of substituted LNAs was discerned (Fig.5D)

[0105] The relative potency of a MX was derived from the ratio of its AAUC(pPSI) to the MC counterpart. The median potencies of OML MXs modified with Lateral PCs were always higher than MXs modified with Central PCs (except for AVA1 B2 in CT26 cells) and 2’-OMe MCs (Fig.5E). Their magnitudes however depend on the cell types and the SSO sequences - for instance, AVA1 B2 MXs recorded the highest relative potencies and the widest magnitude range (>18folds) among cell lines. For each SSO, the optimal number of LNAs was inferred, in all cell lines tested, from MXs with relative potencies within 10% from the most potent MX. The respective optimal residues for AVA1A1 , AVA1 B1 , AVA1 B2 and AVA1C2 are 3, 5, 5 and 4 LNAs (Table 2). This suggests the optimal LNA residues and the relative MX potency are both inversely correlated with Tmof the natural SSO sequences - 66, 57.84, 50.62 and 67.34°C respectively. To test the hypothesis, AVA1 A2, which overlaps AVA1 A1 sequence but has the highest predicted Tmat 69.53°C, was designed and validated as before. None of the OML MXs were more potent than the MC counterpart except in Huh7 cells for which a paltry median relative potency of 1.1 was achieved with two substituted LNAs (Fig.5F and Table 2). This hints to a Tmthreshold beyond which LNA substitution is futile. This could explain why MOL MXs of AVA1A1 were less efficient than its 2’-MOE MC counterpart in the initial screen (Fig.lA, Fig.4A), given that Tmis raised by 0.9 to 1.7 degrees Celsius with each 2’-MOE modification, which is higher than a 2’-OMe nucleoside. Hence, a corollary is that for a givensequence, the optimal LNA residues for MOL MXs will not be greater than in the corresponding OML MXs. The secondary screen was subsequently repeated with MOL MXs under CEM condition in Huh7 cells, and further validated the superior efficiency of MOL MXs modified with Lateral over Central PCs (Fig.5G). The optimal LNA residues inferred from AVA1B1 and AVA1 B2 MOL MXs are 4 and 3 respectively (Table 3), which are expectedly less than the corresponding OML MXs. As also anticipated from the higher Tmof 2’-MOE nucleosides, MOL MXs of AVA1A1 were not more potent than its 2!-MOE MC (median relative potency of MXs with Lateral PC is 1.04, Fig.5H), while MXs of AVA1C2 managed to improve its ineffective MC counterpart (Fig.5G and Table 3).

[0106] It is worth noting that all the top MX performers in Lateral PC did not elicit a CCL22 expression in the BJAB assay (Fig. 5I).

[0107] Lastly, AVA2 MXs modified with Lateral PCs confirmed to induce higher EGFP-to- luciferase luminescence than their MC counterpart (Fig.2D, Fig.12), hinting that the Lateral PC could be applied to SSOs targeting other exons other than NATURA’s FS.

[0108] Lateral PCs were then applied to three SSO therapeutics targeting genes involved in human diseases. The first example shows how mixmers can simultaneously boost potency and reduces the molecular weight of a SSO that downregulates GLDC gene expression for anticancer therapy. The reported candidate is a 27-mer with full 2’-OMePS modification that induces exon 7 exclusion, resulting in frameshifted GLDC transcripts that are degraded by nonsense-mediated decay (NMD). The 5’ was trimmed to a 17-mer, upon which Lateral and Central PCs were applied and validated in A549 and Huh7 cells. Consistently, OML MXs modified with Lateral PCs were more efficient than MXs modified with Central PCs, and with the most potent MX notably more efficient than the published candidate (Fig.3A, Fig.3B, Fig.6A, Fig. 19A). In MOL composition, MXs modified with Lateral PCs were again more efficient than MXs modified with Central PCs and 2'-MOE MC (Fig.6B, Fig.19B). To explore the adaptability of S-cET sugar moiety to the PCs, OML and MOL MXs were converted to OMc and MOc MXs respectively by substituting the LNAs within. In both OMc and MOc compositions, MXs modified with Lateral PCs were also observed to be more efficient than MXs modified with Central PCs (Fig.6C, Fig. 19C). Except for transfected OMc MXs however, none of the MXs was more efficient than their respective 2’-OMe or 2’-MOE MCs.

[0109] The second example is Nusinersen, a 2’-MOE-PS SSO that promotes SMN2 exon 7 inclusion as an approved therapy for spinal muscular atrophy. Given the high baseline potency of Nusinersen, which induces exon 7 inclusion in almost 90% of transcripts even at low dose, the most potent MOL MX, with four substituted LNAs, is up to 6.9% more efficient (Fig.3C, Table 4). When Nusinersen sequence was converted to OML MXs, the most potentMX, which contains six substituted LNAs, is up to 35.3% more efficient than 2’-0Me MC, aibeit is iess efficient than the 2’-M0E counterpart (Fig.3C). Marginal improvement in cell survival from MOL or OML MXs were noted (Fig.SD).

[0110] The third example demonstrated that a substantial gain in efficiency is achievable on a clinical candidate, Drisapersen, a 2’-OMePS SSO that induces dystrophin exon 51 exclusion as therapy for Duchenne muscular dystrophy. The most potent OML MX, which contains three substituted LNAs, is up to 12.8folds more efficient than Drisapersen (Fig.3D, Table 5), with no effect on cell viability (Fig.6E). The results were replicated by transfection in a dose-response assay (Fig.6F). Interestingly, this gain has been observed in vitro and in vivo by another group, where the highest exon 51 skipping in mouse heart attained by Lateral OML MXs, especially Drisa.202 (5folds increase over MC, named AON-B8). Similarly, as the Drisapersen sequence was converted to MOL MXs, a modest but consistent higher efficiency of Lateral PC MXs compared to 2'MOE MC was observed (Fig.6G).

[0111] In conclusion, the PCs elucidated in the study were successfully adapted on and were agnostic on nine SSO sequences, with no bias on specific base compositions. The nucleobase composition is evenly distributed (45% G, 49% C, 51% A and 59% U).

[0112] Example 2: Superior potency and improved cellular toxicity of steric blocker antisense oligonucleotides substituted with two classes of chemically-modified ribose sugars

[0113] The present disclosure provides a set of design rules that specifies the relative positions of ribose sugar, in a single-stranded ribonucleic acid molecule, to be substituted by either one of the two classes of modified sugar moieties namely 2’-C modified and constrained sugar ribose. This results in superior on-target potency and improved cellular toxicity of steric blocker antisense oligonucleotides, while keeping the sequence length of the drug compound at between 15 to 22-mers inclusive, which are critical parameters that confer a wide therapeutic index.

[0114] Natural nucleic acids do not possess attributes that are essential for therapeutic applications. The recurring issue inflicting every nucleic acid therapeutic modality is vulnerability to degradation by ubiquitous endonucleases and exonucleases, resulting in short plasma half-lives. Moreover, for modalities whose mechanism of action (MoA) is through Watson-Crick binding to complementary DNA or RNA target sequence, the slow binding and unbinding thermodynamics of natural nucleic acids results in low on-target potency and poor selectivity respectively. A high dose is required hence to demonstrate therapeutic effect that is frequently associated with toxicity effects, which narrows or eliminates the therapeutic index.

[0115] To overcome the aforementioned issues, the biopharma industry and scientific community have over the last 30 years invented chemistry moieties to substitute variousstructures of nucleic acids. Briefly, three components of a nucleotide are amenable to modifications namely deoxyribose and ribose sugars, backbone linkages, and nucleotide bases (Figure 8). Modified backbone linkages conferred nucleic acids resistance to endonucleases, resulting in significantly longer terminal plasma half-lives than their natural counterparts. On the other hand, nucleic acids whose sugars were partially or entirely substituted with specific classes of sugar chemistry have demonstrated superior levels of on- target potency and / or selectivity relevant for therapeutic application. These pharmacological improvements are validated by the subsequent FDA approval of 5 siRNAs, 5 steric blockers (or splice-switching oligonucleotides or SSOs) and 3 GAPmers. The aim and benefits of chemical modifications are nevertheless specific between therapeutic modalities due to differences in their MoAs, as explained below.

[0116] The use of modified sugar moieties have been most advanced for siRNA therapeutics wherein a mix of chemistry are applied at different ribose sugars on both the sense and antisense strands of the double-stranded molecule. Given the processing and subsequent sequestration of an siRNA by the RISC complex to its target is crucial for its MoA, chemical modifications mainly serve to confer resistance to nucleases and / or to improve selectivity, while striking the balance not to significantly disrupt RISC’s processing of the modified siRNA. By contrast, approved steric blocker and GAPmer therapeutics are composed of sugars modified with one type of chemical moiety. As they both engage their target without sequestration aid from the cellular machinery, the binding and unbinding thermodynamics crucially affect their on-target potency, which can be modulated with chemical modifications. Therefore, besides nuclease resistance and selectivity, chemical modifications can result in superior potency for both steric blockers and GAPmers. However, a very limited number of sugar therein a GAPmer can be chemically modified without losing its MoA to induce RNaseH activity.

[0117] Steric blockers, on the other hand, are most versatile in both extent and types of chemical modification since its MoA is simply steric blockage against RNA-binding proteins from binding to their cognate sequences. A steric blocker constructed from a natural ribonucleic acid can never possess favourable thermodynamics, and even as every sugar is modified, its thermodynamics while superior to the natural form remains sub-optimal. Moreover, substitution of the natural phosphodiester (PO) bond with the phosphorothioate (PS) bond in the backbone linkages, which confers the desired resistance to endonucleases, reduces the drug-target melting temperature, and which negatively impact the thermodynamics. The following approaches can further optimize the thermodynamics properties, albeit with undesirable effects that often offsets the benefits:• Longer sequence length, anecdotally 25 to 30-mers, compensates for the reduced melting temperature from PS bonds. But it is associated with sequence-independent toxicity and immunogenicity due to elevated unspecific binding of the PS bonds to circulating proteins. Furthermore, a larger molecular weight potentially impedes the rate of uptake by cells. Majority of FDA approved siRNA, GAPmer and steric blocker drugs are 20-mers.• Higher GC content, besides compensating for the reduced melting temperature from PS bonds, improves thermodynamics. However, it has been broadly agreed in the field that GCs promotes immune response activation through toll-like receptors, and should be avoided.• Less PS bonds to negate the drop in melting temperate. Besides no improvement in thermodynamics, doing so inadvertently reverts a steric blocker’s vulnerability to endonuclease.

[0118] Through exploiting steric blockers’ versatility for both extent and types of chemical modification, the approach taken in the present study relies on the application of two simultaneous classes of sugar chemistry that not only boasts the on-target potency significantly but also reduce cellular toxicity, and without any of the associated undesirable effects aforementioned, all of which are critical parameters for a wide therapeutic index of this modality.

[0119] For a single-stranded ribonucleic acid molecule at between 15 to 22-mers (inclusive) linked chemically by phosphorothioate (PS) backbones, every ribose sugar was subsequently substituted by either type of moieties selected each from the two distinct classes of modified sugars (Figure 9). The modifications create an steric blocker antisense oligonucleotide, since it is not capable of activating the RNA / pathway and RNaseH enzymatic activity.

[0120] The CALIMERO platform (also known as NATURA) was used for high-throughput screening of different mixmer versions of AVA1-SSO, SSO#2045; AVA1-SSO is a 20-mer splice-switching steric blocker antisense oligonucleotide validated to switch the expression of the CALIMERO gene from the EGFP to tRFP isoforms (Figure 10).

[0121] Each SSO#2045 mixmer is an unique positional configuration of sugars substituted by either 2’-OMe or LNA, with no sequence differences. From a version that is composed entirely of 2’-OMe modified sugars (labelled as 2045_2OM), different mixmers were constructed by using one to eight LNAs to replace the 2’-OMe sugars at different positions (labelled as 2045.xxx_OML where xxx is a running number and OML is a short-name for 2’- OMe + LNA). 61 mixmers were considered and their potencies and cell viability effects werecompared against the singly-modified version, 2045_2OM, at 4 increasing concentrations (6nM, 15nM, 37.5nM and 93.75nM) on 4 cell lines with biological duplicates. Dose response curve for every SSO version was constructed from a total of 1 ,984 data points obtained (Figure 11). To compare the potencies between an OML and its 2OM counterpart over the full range of concentration from OnM to 93.75nM, the area between their respective dose response curves, ABC (Area Between Curves), were computed, as explained in Figure 11.

[0122] Analyses of the ABC data uncover the strong association between potency and LNA substitution or the lack of at the first three sugars at 5’ (+1 , +2 and +3 positions) and / or at the first two sugars at 3’ (-1 and -2 positions). Permutations of positional configuration of LNA-substituted sugars defined as “good” are found on mixmers that were observed to be most potent, whereas mixmers with “bad” positional configuration of LNA-substituted sugars are less potent than its singly-modified version. On the other hand, configurations labelled as “intermediate” contains some features of both “good” and “bad” configurations, and the effect on a mixmer is not as predictable - while its potency was never the most potent, it falls between mixmers carrying the “good” or “bad” configurations. Hence, for therapeutic applications, one should preferably focus on the “good” configurations.

[0123] The efficacies of every mixmer collected at 6nM, 15nM, 37.5nM and 93.75nM were sorted in ascending order (decreasing pPSI) and plotted for each of the cell lines tested (CT26, B16F10, HuH7 and HepG2), with the top 10 most potent mixmers and 2045_2OM tabulated as an inset. The most potent mixmers contains 2 to 5 LNAs, which suggest an optimal range of LNAs for SSO#2045.

[0124] To investigate the association between number of LNAs on mixmers’ potency and toxicity, ABC versus cell viability plot was each generated for mixmers having 1 to 8 LNAs. When the number of LNAs is less than optimal, LNA substitution on any ribose improved potency (i.e. ABC is negative for any mixmer). Within the optimal range of LNA substitutions however, positional configuration of LNA-substituted sugars in a mixmer determines the superiority (ABC < 0) or inferiority (ABC > 0) of its potency, as compared to the singly-modified version. On the other hand, ABC of mixmers with more LNAs than the optimal range either trends to zero (marginal improvement) or is positive (inferior potency).

[0125] For quantitative comparison of potency between a mixmer and its singly-modified version, the pPSI fold change was computed at each SSO concentration. As expected, the fold change magnitudes are cell-type specific. Interestingly, other than for HepG2 cells, the maximum fold change observed did not occur at the highest SSO concentration, which indicated that the peak potency superiority of mixmers modified with “good” positional configurations began at low SSO concentrations.

[0126] In summary, the results of the present study showed that the specific positional configuration of sugars resulting in the highest potencies are corroborated in other cell lines and at different mixmer concentrations. To validate that these positional configurations are independent of SSO sequences and target genes, these positional configurations were applied on other AVA1 SSOs and on four therapeutic SSO leads.

[0127] The positional sugar configuration were applied on the following 4 SSOs each targeting the FS exon of the CALIMERO gene at different sites - both AVA1 b1 (SSO#2100) and AVA1b2 (SSO#2101) target the acceptor splice site whereas both AVA1c1 (SSO#2102) and AVA1c2 (SSO#2103) target the donor splice site (Figure 10). Together with AVA1 (SSO#2045), the 5 SSOs provide variation in the nucleobase sequences. The potencies of mixmers derived from each SSO were obtained similarly as above. Our results showed that the specific positional configuration of sugars resulting in the highest potencies are corroborated in different SSO sequences across cell lines and concentrations.

[0128] It was observed that the optimal number of LNAs for each SSO are respectively #2100 (2 to 8), #2101 (4 to 8), #2102 (2 to 3), and #2103 (3 to 7). It was also observed that poor potency of both #2101_2OM and #2102_2OM versions can be restored by mixmers carrying the positional configurations. In fact, substituting any LNA improves the efficiency of #2101_2OM.

[0129] AVA2-SSOs are the last SSOs that validated on the CALIMERO platform. Both SSO#2087 (33-mer) and SSO#2167 (21-mer) were each designed to switch the expression of the CALIMERO gene from the EGFP to luciferase isoforms (Figure 12). SSO#2167 is a shortened version of SSO#2087 (more on the shortening of SSO sequences in the next section). The OML mixmer 2167.2 induced higher level of luciferase than SSO#2087 and SSO#2167, both of which were modified with 2OMe sugars (Figure 12).

[0130] Example 3: M-Score

[0131] Given the consistency of the preceding positional configurations in predicting the relative potency of mixmers that were modified with “good”, “intermediate” or “bad” rules, the present inventors then sought to represent them with a quantitative score which would enable statistical analyses to be performed. The present inventors defined the M-Score as the average minima sugar positions in a mixmer that are substituted with the second ribose sugar moiety, where the “second” moiety refers to the ribose chemistry with the lower incidence in a mixmer. The M-Score is calculated by dividing the sum of minima LNA positions by the number of LNAs. For example, for a 20-mer OML mixmer composed of 6 LNAs and 14 2’OMe ribose, the M-Score = 20 / 6 = 3.333.

[0132] The M-Score was calculated on every mixmer of AVA1 SSOs (SSO#2045, SSO#2100, SSO#2101, SSO#2102, SSO#2103, and SSO#2099) modified with “good”, “intermediate” or “bad” rules. There is indeed statistical significance in the M-Score differences between each group whether by non-parametric Wilcoxon Rank Sum test or by parametric t- test.

[0133] Example 4: Cellular toxicity of mixmers

[0134] The cellular toxicity were inferred from the normalized cell viability, which was obtained by counting the cells upon SSO treatment and normalized it with the cell counts from untreated samples for each SSO concentration and cell line tested. Boxplots of normalized cell viability of OML mixmers were depicted for all AVA1 SSOs (SSO#2045, SSO#2100, SSO#2101, SSO#2102, SSO#2103, and SSO#2099) grouped by “good”, “intermediate” and ’’bad” positional configurations, and GEN2 (the singly 2’OMe modified version). Only boxplots in which the difference in normalized cell viability between any two groups is statistically significant were depicted.

[0135] With the exception of HuH7 cells, there is at least one statistically significant difference in the normalized cell viability between groups in the other three cell lines tested, which suggests that the effect and extent on cell viability is cell-type dependent. Notably, the median normalized cell viability in the “good” mixmer group was never lower than both the “bad” mixmer and GEN2 groups. In fact, the normalized cell viability in the “good” group is statistically higher than the “bad” mixmer group under many cases.

[0136] Example 5: Validation of positional configuration of sugars on therapeutic SSO leads

[0137] The positional sugar configurations were applied to four validated therapeutic SSO leads with the following objectives.• To shorten these leads, which are composed of 27 to 34 bases, to within 15- to 22- mer without losing their on-target potencies; for each SSO lead, its shortened version was obtained by trimming at either the 5’, 3’ or both ends.• To investigate whether mixmers of the shortened versions can attain superior potency over the current respective leads.• To investigate if the positional sugar configurations apply to other dual combinations of modified sugar moieties

[0138] 1. GLDC SSO

[0139] The SSO lead, SSO#181 (2OM), which is 27-mer whose every sugar is substituted with 2’-OMe ribose, was shortened by 37% to 17-mer, SSO#2013. As expected, whenmodified entirely with 2’-0Me ribose, the on-target efficiency of SSO#2013(2OM) was almost obliterated (Figure 13A).

[0140] OML mixmers

[0141] Ribose of SSO#2013 was substituted with either 2’-OMe or LNA with the preceding “good”, “bad” and “intermediate” positional sugar configurations. As depicted in Figure 13A, the most potent OML mixmers were modified with “good” rules tested in all cell lines, concentrations and under transfection, calcium-enriched medium (“OEM”, which stimulates in vivo free uptake by cells) and in the absence of any delivery reagent (“Free uptake”). They were all more potent than the SSO#2013(2OM) monomer (denoted with light grey circles) and its parent SSO#181(2OM) lead (denoted with grey circles), except on A549 cells via free uptake.

[0142] By transfection, the most efficient OML mixmer from each respective condition is between 2.5X to 5X more potent than the monomer counterpart, SSO#2013(2OM). Under OEM, the most efficient OML mixmer is at least 3X more potent than SSO#2013(2OM). On the other hand, in the absence of a delivery agent, the most efficient OML mixmer is at least 1.4X more potent than SSO#2013(2OM) on TS32 tumor sphere cells.

[0143] MOL mixmers

[0144] Ribose of SSO#2013 was substituted with either 2’-O-MOE (Figure 9) or LNA with the preceding “good”, “bad” and “intermediate” positional sugar configurations. As depicted in Figure 13B, the most potent MOL mixmers were modified with “good” rules tested in all cell lines, concentrations and under transfection, OEM and in the absence of any delivery reagent, except on A549 cells via free uptake. They were all more potent than the SSO#2013(2MOE) monomer (denoted with light grey circles).

[0145] By transfection, the most efficient MOL mixmer from each respective condition is between 3X to 8X more potent than the monomer counterpart, SSO#2013(2MOE). Under OEM or in the absence of a delivery agent, the most efficient MOL mixmer from each respective condition is at least 1.5X and up to 3.6X more potent than SSO#2013(2MOE).

[0146] OMc mixmers

[0147] Ribose of SSO#2013 was substituted with either 2’-OMe or (S)-cET (Figure 9) with the preceding “good”, “bad” and “intermediate” positional sugar configurations. Among the mixmers, the most potent OMc mixmers were modified with “good” rules tested in every condition tested (Figure 13C), which again validates the positional sugar configurations. When compared to SSO#2013(2OM) monomer (denoted with light grey circles), the most potent mixmers were more potent by transfection but not by OEM or in the absence of delivery agents. When compared to the parent SSO#181(2OM) lead (denoted with gray circles), the mixmerswere generally less potent. This could be attributed to the higher melting temperature of OMc mixmers than both OML and MOL mixmers, which requires further shortening of the SSO sequence.

[0148] MOc mixmers

[0149] Ribose of SSO#2013 was substituted with either 2’-O-MOE or (SJ-cET with the preceding “good”, “bad” and “intermediate” positional sugar configurations. Among the mixmers, the most potent MOc mixmers were modified with “good” rules tested in every condition tested (Figure 13D), which again validates the positional sugar configurations. However, none of the mixmers is more potent than the SSO#2013(2MOE) monomer (denoted with light grey circles), which suggests that MOc mixmers have even higher melting temperatures than the OMc counterparts.

[0150] 2. SLC25A13 SSO

[0151] The SSO lead, SSO#2008(2OM), which is 33-mer whose every sugar is substituted with 2’-OMe ribose, was shortened by 42% to 19-mer, SSO#2034. When synthesized as entirely 2’-OMe or 2’-O-MOE ribose, the on-target efficiency of both SSO#2034(2OM) and SSO#2034(2MOE) monomers were almost abrogated respectively (Figure 14).

[0152] OML mixmers

[0153] Ribose of SSO#2034 was substituted with either 2’-OMe or LNA with the preceding “good”, “bad” and “intermediate” positional sugar configurations. As depicted in Figure 14A, the most potent OML mixmers were modified with “good” rules tested in all cell lines, concentrations and by transfection and under CEM condition. They were all more potent than the SSO#2034(2OM) monomer (denoted with light grey circles) and its parent SSO#2008(2OM) lead (denoted with grey circles).

[0154] By transfection, the most efficient OML mixmer from each respective condition is between 8X to 30X more potent than the monomer counterpart, SSO#2034(2OM). Under CEM, the most efficient OML mixmer from each respective condition is between 7X to 76X more potent than SSO#2034(2OM).

[0155] MOL mixmers

[0156] Ribose of SSO#2034 were substituted with either 2’-O-MOE or LNA with the preceding “good”, “bad” and “intermediate” positional sugar configurations. As depicted in Figure 14B, the most potent MOL mixmers were modified with “good” rules, and were all more potent than the SSO#2034(2MOE) monomer (denoted with light grey circles).

[0157]

[0158] By transfection, the most efficient MOL mixmer is at least 3X more potent than the monomer counterpart, SSO#2034(2MOE). Under CEM, the most efficient MOL mixmer from each respective condition is 1.9X to 3X more potent than SSO#2034(2MOE).

[0159] 3. PTC SSO

[0160] The SSO lead, SSO#2021(2OM), which is 34-mer whose every sugar is substituted with 2’-OMe ribose, was shortened by 38% to 21-mer, SSO#2044. When synthesized as entirely 2’-OMe ribose, the on-target efficiency of SSO#2044(2OM) was significantly reduced as compared to the parent SSO#2021 (2OM) lead (Figure 15).

[0161] OML mixmers

[0162] Ribose of SSO#2044 was substituted with either 2’-OMe or LNA with the preceding “good”, “bad” and “intermediate” positional sugar configurations. As depicted in Figure 15A, the most potent OML mixmers were modified with “good” rules tested in all cell lines, concentrations and by transfection and under CEM condition. They were all more potent than the SSO#2044(2OM) monomer (denoted with light grey circles).

[0163] By transfection, the most efficient OML mixmer from each respective condition is between 1.6X to 4.6X more potent than the monomer counterpart, SSO#2044(2OM). Under CEM, the most efficient OML mixmer from each respective condition is at least 4.9X more potent than SSO#2044(2OM).

[0164] MOL mixmers

[0165] Ribose of SSO#2044 were substituted with either 2’-O-MOE or LNA with the preceding “good” and “bad” positional sugar configurations. As depicted in Figure 15B, the most potent MOL mixmers were modified with “good” rules, and were all more potent than the SSO#2044(2MOE) monomer (denoted with light grey circles). By CEM, the most efficient MOL mixmer is 1.25X more potent than the monomer counterpart, SSO#2044(2MOE).

[0166] Example 6: Mixmers with combinations of other modified riboses

[0167] The aforementioned positional sugar configurations, which were validated to boast the on-target potency of AVA1A1 mixmers, were further tested on 7 more types of ribose moieties namely ENA, 2’-Fluoro, 2’-FANA, 2’-O-NMA, TNA, HNA and MNA (Figure 17A). As shown in Figure 17B, in each combination of ribose moiety with 2’-OMe, 8 permutations of positional configurations (PCs) were considered. They include both lateral (101 , 301 , 401 , 402, 501 and 601) and central (333 and 444) PCs.

[0168] The on-target potency of each AVA1A1 mixmer was evaluated, as described above, for concentrations 10nM, 50nM and 250nM at timepoints 24 h, 48 h and 96 h post transfection. The splice-switching efficiency at each condition was normalized for #2045.101 (OML), the best-performing AVA1 A1 modified with 2’-OMe and LNA at a lateral PC (top left box in Figure18); higher numbers indicate higher exon-skipping efficiencies and a number larger than unity indicates higher efficiency than #2045.101 (OML). Two notable observations are evident. First, mixmers modified with the combination of 2’-OMe with either ENA or HNA were more efficient than #2045.101 (OML) at most of the lateral PCs tested (Figure 18), under the conditions of 10 nM and 50 nM concentrations at 48 h and 96 h timepoints. Second, when comparing among mixmers modified with the same moiety combination (same moiety grouped in the same rows in Figure 18), lateral PCs resulted in higher efficiency than central PCs (the latter are highlighted with grey boxes, oligos .333 and .444). This showed that the PC rules are applicable to these 7 types of ribose moieties.

[0169] To summarize, the positional sugar configurations hold for the 11 SSOs applied thus far:• CALIMERO-AVA: SSO#2045, SSO#2100, SSO#2101 , SSO#2102, SSO#2103, SSO#2167, and SSO#2099 (data not shown).• GLDC: SSO#2013• SLC25A13: SSO#2034• OTC: SSO#2044• BRD4: SSO#2067(data not shown)

[0170] In various embodiments, the sequence length of the SSOs ranges from 15 to 22 bases.

[0171]

[0172] On mixmers with the following combination of ribose moieties:• OML: 2’-OMe and LNA• MOL: 2’-O-MOE and LNA• OMc: 2’-OMe and (S)-cET• MOc: 2’-O-MOE and (S)-cET• 2’-OMe and ENA (Figure 17A)• 2’-OMe and 2’-Fluoro (Figure 17A)• 2’-OMe and 2’-FANA (Figure 17A)• 2’-OMe and 2’-O-NMA (Figure 17A)• 2’-OMe and TN A (Figure 17A)• 2’-OMe and HNA (Figure 17A)• 2’-OMe and MNA (Figure 17A)

[0173] In various embodiments, the results were validated on the following cell lines: CT26, B16F10, A549, TS32, HepG2, HuH7, Bel7402, via transfection and free uptake by cells (CEM and non-CEM).

[0174] In various embodiments, key technical features of the present disclosure lie in a set of positional sugar configurations that specifies the permutations of sugar positions at which they can be substituted with one of the following combinations of ribose moiety to boost the on-target potency of steric blocking antisense oligonucleotides:• OML: 2’-OMe and LNA• MOL: 2’-O-MOE and LNA• OMc: 2’-OMe and (S)-cET• MOc: 2’-O-MOE and (S)-cET• 2’-OMe and ENA (Figure 7A)• 2’-OMe and 2’-Fluoro (Figure 17A)• 2’-OMe and 2’-FANA (Figure 17A)• 2’-OMe and 2’-O-NMA (Figure 17A)• 2’-OMe and TN A (Figure 17A)• 2’-OMe and HNA (Figure 7A)• 2’-OMe and MNA (Figure 17A)

[0175] Example 6: Embodiments of rules of positional configurations

[0176] The simultaneous use of two classes of modified ribose at specific sugars in a single-stranded splice-switching oligonucleotide improves its on-target potency, efficiency and selectivity. The first 2’-modified sugar ribose includes 2’-OMe and / or 2’MOE. These must be used in combination with the second class of chemical modifications that may be labelled as constrained nucleotides (CN), which include LNA, cET, ENA, 2’-O-NMA, TNA, HNA, and MNA. 2’-Fluoro, and 2’-FANA, which are also 2’ ribose modifications, can also be considered as belonging to the CN class.

[0177] Good Rules:• A CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +1 or +2 or +3.• First CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +1 or +2 AND a second CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, ending at either -2 or -1 .• One CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +1 or +2 AND one CN nonadjacent to the CN cluster and NOT at -2 or - 1.• First CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +1 or +2 AND a second CN cluster, which consists of one or moreconsecutive CNs for up to 8 CNs, ending at either -2 or -1 AND one CN nonadjacent to both of the CN ciusters,• First CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +1 AND a second CN ciuster, which consists of one or more consecutive CNs for up to 8 CNs, ending at -1 AND two adjacent or nonadjacent CNs NOT adjacent to both of the CN ciusters.• CN at any of the foilowing sets of position - {+3, -2} and {+3, -1}.

[0178] intermediate Rules:• First CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +3 AND a second CN cluster, which consists of two or more consecutive CNs and up to 8 CNs, ending at either -2 or -1 .• One CN ciuster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +3 AND one CN nonadjacent to the CN cluster and NOT at -2 or -1.• First CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +3 AND a second CN ciuster, which consists of one or more consecutive CNs for up to 8 CNs, ending at either -2 or -1 AND one CN nonadjacent to both of the CN clusters.• One CN ciuster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +1 or +2 AND two adjacent or nonadjacent CNs NOT adjacent to the CN cluster.• First CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +1 AND a second CN ciuster, which consists of one or more consecutive CNs for up to 8 CNs, ending at -2 AND two adjacent or nonadjacent CNs NOT adjacent to both of the CN clusters.• First CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, beginning at +2 AND a second CN cluster, which consists of one or more consecutive CNs for up to 8 CNs, ending at -2 or -1 AND two adjacent or nonadjacent CNs NOT adjacent to both of the CN clusters.

[0179] As used herein, the term “cluster” can refer to a modification on one nucleotide as well as modifications on two or more consecutive nucleotides.

[0180] Example 7: Exemplary sequences of the present disclosure

[0181] Nucleotides with a 2’-O-Methyl RNA (2’OMe) are indicated with “m”. Nucleotides with a 2:'O-methoxyethyl (2:MOE) RNA are indicated with " / MOEr / ”. 2:MOE’modified thymidine is used in place of 2’MOE-modified uridine. Nucleotides with a locked nucleic acid(LNA) are indicated wildLNA-modified thymidine is used in place of i..NA-modified uridine Nucleotides joined by a phosphorothioate (PS) bond to the foliowing nucleotide are indicated with It would be generally known to the person skilled in the art that the number of phosphorothioate bonds is one less than the number of bases. “2OM’:indicates that the SSO is modified with 2’-O-methyl RNA. “2MOE” indicates that the SSO is modified with 2’-O- methoxyethyl RNA. “2OML” indicates that the SSO includes 2'-O-methyl RNA and locked nucleic acid modifications. “2MOL” indicates that the SSO includes 2’-O-methoxyethyl RNA and locked nucleic acid modifications.

[0182] Table 1 : NATURA SSO sequencesWAiereAll oligos are 5' -> 3'

[0183] Table 2: Tabulation of relative potency of OML MXs.

[0184] Table 2 above shows the numerical values of relative potency of OML MXs of AVA1A1 , AVA1A2, AVA1B1, AVA1 B2 and AVA1C2, breakdown by number of LNA substitutions, obtained from every cell line tested. For each SSO and in each cell line, the most potent MX and MXs with relative potencies within 10% are indicated with asterisks (last column). The optimal number of substituted LNAs for each SSO was determined from the median number of LNAs from MXs marked with asterisks,

[0185] Table 3: Tabulation of relative potency of MOL MXs.

[0186] Table 3 above shows the numerical values of relative potency of MOL MXs of AVA1A1 , AVA1 B1 and AVA1 B2, breakdown by number of LNA substitutions, obtained from Huh7 via GEM condition. For each SSO, the most potent MX and MXs with relative potencies within 10% are indicated with asterisks.

[0187] Table 4: Spinraza MXs.

[0188] Table 5: Drisapersen MXs.

[0189] Table 6: PCR primers and probesPCR primers and probes

[0190] The present inventors conducted a systematic study to discover universal modification patterns that improve SSO potency, and to investigate their sequence (in)dependency. Using NATURA and five sequences in its SSO tool library, general modification patterns, defined as Lateral PCs, were discovered for OML and MOL MXs for which the most efficient MX per SSO sequence was consistently found to be applied with Lateral PCs across all cell lines tested. The finding was further corroborated on three published SSO therapeutics and therefore, it is likely sequence agnostic as no specific base compositions was discerned from the eight SSOs. Moreover, the rule may be generalisable to mixmers utilizing any combination of 2’-O-modified and constrained ribose including cETs. However, since the combination of ribose chemistries and the specific Lateral PC that confer the highest SSO potency depend on the sequence and cell type, it is important in drug discovery to screen for all possible Lateral PCs for each ribose chemistry combination. With the method of the present disclosure, this becomes a practical approach given the drastically diminished mixmer permutations, which can be further reduced by guidance from the inverse correlation between the optimal LNA residues with Tmof the unmodified SSO sequence and the 2 -O-modification. In addition, depending on the cell types, Lateral PC MXs were always observed to have either improved or identical cell survival as compared to their mono- chemically modified counterparts.

[0191] As defined in Lateral PCs, the placement of LNA substitutions at the flanks of a SSO with 5’ end more potent than at 3’ is likely to have a mechanistic basis, given that they are also agnostic to sequence, 2’-O-modification and cell type.

[0192] While embodiments of the invention have been particularly shown and described with reference to specific embodiments, it should be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the invention as defined by the appended claims. The scope of the invention is thus indicated by the appended claims and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced.

Claims

CLAIMS1 . A method of identifying a chemical modification pattern which confers an increased splice-modulation efficiency to a single-stranded RNA oligonucleotide with a predetermined nucleobase sequence, as compared to a single-stranded RNA oligonucleotide with the same predetermined nucleobase sequence having every nucleotide chemically modified with a 2’ ribose modification, the method comprising: (a) applying a plurality of chemical modification patterns (Patterns A to G) to a plurality of single-stranded RNA oligonucleotides respectively to result in a plurality of chemically-modified single-stranded RNA oligonucleotides, wherein each of the plurality of single-stranded RNA oligonucleotides has one of Patterns A to G applied to it, and wherein the plurality of single-stranded RNA oligonucleotides have the same predetermined nucleobase sequence,(i) wherein each of the plurality of chemical modification patterns comprises:(A) at least one nucleotide chemically modified with the 2’ ribose modification; and(B) one to eight nucleotides chemically modified with a constrained nucleotide (CN); and(ii) wherein the plurality of chemical modification patterns comprises the following positional configurations of CN-modified nucleotides:(A) Pattern A comprising a CN cluster comprising one or more consecutive CNs, beginning at position +1 , +2 or +3 of the singlestranded RNA oligonucleotide;(B) Pattern B comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the singlestranded RNA oligonucleotide, and a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide;(C) Pattern C comprising a CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the singlestranded RNA oligonucleotide, and an additional CN that is not adjacent to the CN cluster and that is not at position -2 or -1 of the single-stranded RNA oligonucleotide;(D) Pattern D comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the singlestranded RNA oligonucleotide, and a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to both the first and second CN clusters;(E) Pattern E comprising a first CN cluster comprising one or more consecutive CNs, beginning at position +1 , and a second CN cluster comprising one or more consecutive CNs, ending at position -1 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters;(F) Pattern F comprising a first CN at position +3 and a second CN at position -2 of the single-stranded RNA oligonucleotide; and(G) Pattern G comprising a first CN at position +3 and a second CN at position -1 of the single-stranded RNA oligonucleotide; and(b) measuring the splice-modulation efficiencies of the chemically-modified singlestranded RNA oligonucleotides and comparing these with the splice-modulation efficiency of the single-stranded RNA oligonucleotide having every nucleotide chemically modified with the 2’ ribose modification to determine the chemical modification pattern which confers the increased splice-modulation efficiency.

2. The method of claim 1 , wherein step (a) further comprises applying a further chemical modification pattern (Pattern H) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern H comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide, and a second CN cluster comprising two or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide.

3. The method of claim 1 or 2, wherein step (a) further comprises applying a further chemical modification pattern (Pattern I) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern I comprising the following positional configuration of CN-modified nucleotides: a CN cluster comprising (i) one or more consecutive CNs, beginning at position +3 of the single-stranded RNAoligonucleotide; and (ii) an additional CN that is not adjacent to the CN cluster and that is not at position -2 or -1 of the single-stranded RNA oligonucleotide.

4. The method of any one of claims 1 to 3, wherein step (a) further comprises applying a further chemical modification pattern (Pattern J) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern J comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +3 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and an additional CN that is not adjacent to both the first and second CN clusters.

5. The method of any one of claims 1 to 4, wherein step (a) further comprises applying a further chemical modification pattern (Pattern K) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern K comprising the following positional configuration of CN-modified nucleotides: a CN cluster comprising one or more consecutive CNs, beginning at position +1 or +2 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to the CN cluster.

6. The method of any one of claims 1 to 5, wherein step (a) further comprises applying a further chemical modification pattern (Pattern L) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern L comprising the following positional configuration of CN-modified nucleotides: a first CN cluster comprising one or more consecutive CNs, beginning at position +1 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters.

7. The method of any one of claims 1 to 6, wherein step (a) further comprises applying a further chemical modification pattern (Pattern M) on a single-stranded RNA oligonucleotide having the same predetermined nucleobase sequence, Pattern M comprising the following positional configuration of CN-modified nucleotides: a first CNcluster comprising one or more consecutive CNs, beginning at position +2 of the single-stranded RNA oligonucleotide, a second CN cluster comprising one or more consecutive CNs, ending at position -2 or -1 of the single-stranded RNA oligonucleotide, and a first additional CN and a second additional CN that are not adjacent to both the first and second CN clusters.

8. The method of any one of claims 1 to 7, wherein the 2’ ribose modification is selected from the group consisting of 2’-O-methyl (2’-0Me) and 2’-O-methoxyethyl (2’-MOE).

9. The method of any one of claims 1 to 8, wherein the constrained nucleotide (CN) is selected from the group consisting of locked nucleic acid (LNA), S-constrained ethyl ((S)-cET), ethylene-bridged nucleic acid (ENA), 2’-O-N-methylacetamido (2’-0-NMA), threose nucleic acid (TNA), hexitol nucleic acid (HNA), methyleneoxy nucleic acid (MNA), 2’-Fluoro and 2’-Fluoroarabino nucleic acid (2’-FANA).

10. The method of any one of claims 1 to 9, wherein all nucleotides that are not CN- modified comprise the 2’ ribose modification.

11. The method of any one of claims 1 to 10, wherein the single-stranded RNA oligonucleotide is between 15 and 22 nucleotides in length.

12. The method of any one of claims 1 to 11 , wherein the nucleobase sequence of the single-stranded RNA oligonucleotide has a melting temperature of 70°C or below.

Images