Antisense oligonucleotides for the treatment of fatty liver disease

Abstract

The present invention relates to antisense oligonucleotides (AONs) that can mediate RNA editing by binding to a target RNA nucleic acid molecule, preferably an RNA transcript molecule, in a cell and recruiting an endogenous deaminating enzyme in the cell to deaminate a target adenosine in the target RNA molecule to an inosine. The target RNA molecule is a transcript molecule form the human PNPLA3 gene that encodes the I148M variant of the Patatin-Like Phospholipase domain-containing protein 3 (PNPLA3), and the target adenosine is the adenosine in the AUG codon coding for methionine (M) at position 148 of the PNPLA3 I148M variant protein. The deamination of the adenosine changes the amino acid to a valine (V). The RNA editing of the adenosine will result in a reversion of the functionality of the PNPLA3 protein that is similar or equal to the functionality of the wildtype PNPLA3 protein, which will result in lowered fattening of the liver, thereby lowering the risk of suffering from disorders related to non-alcoholic fatty liver disease (NAFLD) caused by the I148M variant.

Description

[0001] ANTISENSE OLIGONUCLEOTIDES FOR THE TREATMENT OF FATTY LIVER DISEASE

[0002] TECHNICAL FIELD

[0003] This disclosure relates to the field of medicine, and in particular to the field of fatty liver disease, particularly Non-Alcoholic Fatty Liver Disease (NAFLD), for example steatosis, Non- Alcoholic Steatohepatitis (NASH), fibrosis, and cirrhosis. The disclosure describes antisense oligonucleotides that mediate nucleotide-specific RNA editing in the gene transcript encoding the Patatin-Like Phospholipase domain-containing protein 3 (PNPLA3) to bring about amino acid changes such that it influences its activity.

[0004] BACKGROUND

[0005] Excessive calory intake normally leads to fat accumulation in adipose tissue (in the form of triglycerides), but under circumstances that are not fully understood fat accumulation also occurs in other tissues, such as the liver. Excessive fat accumulation in the liver may give rise to so-called Fatty Liver Disease (FLD), which is divided in two types: non-alcoholic and alcoholic. Non-Alcoholic Fatty Liver Disease (NAFLD), exemplified by Non-Alcoholic Steatohepatitis (NASH), may lead to liver fibrosis, ultimately leading to liver failure and / or liver cancer. FLD and NASH and their more progressive stages have become very common in regions with a lifestyle characterized by excessive caloric intake, especially in the form of high fat diets and / or diets with a high glycaemic index, in combination with a sedentary lifestyle and a general lack of physical exercise. NAFLD affects approximately 25% of the population worldwide and is most prevalent in the Middle East and South America with the lowest incidence in Africa.

[0006] NAFLD is typically associated with obesity and insulin resistance. Its progression is characterized by four stages: steatosis (first) leading to NASH, a condition characterized by inflammation and ballooning (second). This condition may develop into organ impairment or cirrhosis (third) leading to the end stage of hepatocellular carcinoma (fourth) necessitating liver transplantation. NAFLD has multifactorial pathogenesis involving a close interplay of environmental factors and genetic determinants. Numerous studies suggest ethnic difference as the major cause of hepatic fat.

[0007] Although a lot remains to be elucidated concerning the etiology and the progression to subsequent stages, a strong correlation is found between the development and progression of FLD-related disorders in humans and variants in the gene coding for the Patatin-like Phospholipase domain-containing 3 protein (PNPLA3, also known as adiponutrin (ADPN) or as calcium-independent phospholipase A2-epsilon (IPLA2-epsilon)). As the name PNPLA3 suggests, this protein is a member of a family of proteins, the progenitor of which, patatin, has been identified in potato tubers, where it cleaves fatty acids from membrane lipids by its lipase activity.

[0008] Discovery of a common genetic variant in PNPLA3, generally referred to as I148M, located at rs738409 C>G, with a global prevalence of 30%-50%, was a ground-breaking finding in explaining individual genetic susceptibility to all stages of NAFLD (Romeo S et al. Nat Genet. 2008, 40:1461-1465). This variant alone explains 11% of the population attributable fraction of liver cirrhosis. Unlike NAFLD associated with the metabolic syndrome, the I148M variant predisposes to NAFLD independently of substrate excess and obesity and without altering insulin sensitivity, which implies that the mechanisms and thereby potential treatment of NAFLD attributed to insulin resistance and the PNPLA3 I148M variant are distinct. The variant relates to a substitution of methionine (M) for isoleucine (I) in amino acid position 148 and is therefore generally referred to herein and elsewhere as the I148M variant. The variant appears to be dominant over the wildtype version as carriers comprising a single allele of the variant sequence display the fibrotic phenotype.

[0009] PNPLA3 is a lipid droplet protein that is mainly expressed in the liver and adipose tissue and is markedly upregulated under postprandial conditions and downregulated during fasting. Hence, nutritional regulation of PNPLA3 is robust in humans. Low-calorie diet reduces PNPLA3 expression in the adipose tissue and gets upregulated on refeeding by both insulin and glucose. PNPLA3 is one among the nine members of patatin-like phospholipase domain containing proteins (PNPLA1-9). The sequence of a human PNPLA3 mRNA transcript can be found at National Center for Biotechnology Information (NCBI) RefSeq accession number NM_025225.3. PNPLA3 transfers unsaturated fatty acids from triglycerides to phospholipids. The phenotype of the I148M variant is best understood by assuming it is a gain-of-function variant, increasing its coenzyme A-dependent acylation of 1-acyl-sn-glycerol 3-phosphate to generate phosphatidic acid, which is a precursor for triglycerides and glycerophospholipids. Others have suggested that PNPLA3 has glycerolipid hydrolase activity that may be impaired by the methionine substitution, as the longer side chain of methionine in comparison to isoleucine would hamper substrate access. However, whether PNPLA3 has glycerolipid hydrolase activity in vivo remains controversial to this day. What is accepted is that PNPLA3 has different enzymatic activities. When expressed in Sf9 insect cells it demonstrates triglyceride lipase and acylglycerol transacylase activities. It was reported that the I148M variant showed an impaired triglyceride lipase activity suggesting a loss-of-function in the development of steatosis (Huang Y et al. J Biol Chem. 2011 , 286(43):37085-37093). In the same report, acyltransferase activity of the PNPLA3 wildtype or I148M variant was not observed. In another study, trigger fused soluble PNPLA3 was shown to have a lysophosphatidic acid acyl transferase activity which increased when the I148M variant was overexpressed, suggestive of a gain-of-function of the variant (Kumari M et al. Cell Metab. 2012, 15(5):691-702). It is generally accepted that the variant does not affect the localization of the protein. It is partitioned between the membrane and intracellular lipid droplets. Be that as it may, the PNPLA3 I148M variant strongly increases hepatic triglyceride accumulation in individuals carrying this mutation.

[0010] The present invention aims to provide alternative and / or improved compounds or compositions for use in the treatment of liver disease, such as NAFLD, NASH, fibrosis, and cirrhosis, caused by the PNPLA3 I148M variant, namely by editing the adenosine in the codon in the transcript encoding methionine at position 148 of the PNPLA3 variant protein, yielding a change to valine in the amino acid sequence of the protein.

[0011] SUMMARY OF THE INVENTION

[0012] Disclosed herein is an antisense oligonucleotide (AON) that is capable of recruiting an endogenous ADAR enzyme in a human cell after the AON has formed a double-stranded complex with a region of a target RNA nucleic acid molecule in the cell, wherein the cell is a liver cell, preferably a hepatocyte, wherein the transcript molecule is a pre-mRNA or an mRNA molecule, wherein the region comprises a target adenosine, wherein the nucleotide in the AON that is opposite the target adenosine is the orphan nucleotide, wherein the nucleotide numbering is such that the orphan nucleotide is number 0 and nucleotides are further positively (+) incremented towards the 5’ end and negatively (-) incremented towards the 3’ end, wherein the ADAR enzyme can deaminate the target adenosine into an inosine after binding to the double-stranded complex, wherein the target RNA nucleic acid molecule is a transcript molecule of the human PNPLA3 gene comprising the rs738409 single nucleotide polymorphism (SNP), and wherein the target adenosine is in the AUG codon coding for methionine (M) at position 148 of the encoded PNPLA3 protein.

[0013] In a preferred aspect, the orphan nucleotide is a deoxynucleotide comprising a cytosine or a cytosine analog. In a preferred aspect, the cytosine analog comprises a 6-amino- 5-nitro-3-yl-2(1 H)-pyridone nucleobase (also referred to herein and elsewhere as a “Benner’s base”), or a 5-aza-5,6-dihydro cytosine nucleobase (also herein and elsewhere referred to as an “E base”).

[0014] In a preferred aspect, the nucleotide at the -1 position is a deoxynucleotide comprising a hypoxanthine nucleobase (deoxyinosine; Id).

[0015] In a preferred aspect, the nucleotide at the +2 position comprises a cytosine nucleobase, and preferably further comprises a 2’-F substitution at the ribose sugar moiety.

[0016] In a preferred aspect, the nucleotide at the +3 position generates an A:C mismatch with the nucleotide in the PNPLA3 transcript molecule that is directly opposite the nucleotide at the +3 position. In a preferred aspect, the nucleotide at the +4 position generates a Watson-Crick G:C match with the nucleotide in the PNPLA3 transcript molecule that is directly opposite the nucleotide at the +4 position.

[0017] In a preferred aspect, the human PNPLA3 gene comprises the rs738408 SNP, and the nucleotide at the +5 position generates a G:ll wobble base pair with the nucleotide in the PNPLA3 transcript molecule that is directly opposite the nucleotide at the +5 position.

[0018] In one embodiment, the linkage numbering of the AON as disclosed herein is such that the linkage 5’ of the orphan nucleotide is linkage number 0 and linkage positions are further positively (+) incremented towards the 5’ end and negatively (-) incremented towards the 3’ end.

[0019] In a preferred aspect, the AON has the following structure:

[0020] 5’-N23N22N2l N20Nl9Nl8Nl7Nl6Nl5Nl4Nl3Nl2Nl l Nl0N9N8N7N6N5N4N3N2NlOMl M2M3M4M5M6-3’ wherein:

[0021] O is the orphan nucleotide at nucleotide position 0, which is: i) a deoxynucleotide comprising a 6-amino-5-nitro-3-yl-2(1 H)-pyridone nucleobase; or ii) a deoxynucleotide comprising a 5-aza-5,6-dihydro cytosine nucleobase;

[0022] Mi at position -1 is a nucleotide comprising a hypoxanthine nucleobase, preferably wherein the nucleotide is a deoxyinosine (Id);

[0023] M2 at position -2 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is a 2’-OMe modified adenosine (Am);

[0024] N1 at position +1 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is a deoxyadenosine (Ad) or a 2’-MOE modified adenosine (Ae);

[0025] N2 at position +2 is a nucleotide comprising a cytosine nucleobase, preferably wherein the nucleotide is a 2’-F modified cytidine (Of);

[0026] N3 at position +3 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is Ae;

[0027] N4 at position +4 is a nucleotide comprising a guanine nucleobase, preferably wherein the nucleotide is a 2’-MOE modified guanosine (Ge);

[0028] Ns at position +5 is a nucleotide comprising an adenine nucleobase or a guanine nucleobase, preferably wherein the nucleotide is 2’-F modified, more preferably wherein the nucleotide is a 2’-F modified guanosine (Gf); and wherein the linkage between M2 and M3 (linkage position -3) is a phosphorothioate (PS) linkage or a phosphodiester (PO) linkage, preferably a PO linkage.

[0029] In a preferred aspect, the AON has the structure as outline above, wherein the linkage between:

[0030] O and Mi (linkage position -1) is a PS or a mesyl phosphoramidate (PNms) linkage; Mi and M2 (linkage position -2) is a PS, a PNms, a (1 ,3-dimethylimidazolidin-2-ylidene) phosphoramidate (PNdmi), or a methylphosphonate (MP) linkage;

[0031] N1 and O (linkage position 0) is a PO, a PS, or a PNms linkage;

[0032] N2 and N1 (linkage position +1) is a PS linkage;

[0033] N3 and N2 (linkage position +2) is a PO or a PS linkage;

[0034] N4 and N3 (linkage position +3) is a PO or a PS linkage; and / or

[0035] Ns and N4 (linkage position +4) is a PNdmi or a PNms linkage.

[0036] In a preferred aspect, the linkage between the most terminal two nucleotides on the 5’ and / or 3’ terminus of the AON as disclosed herein is a PNdmi linkage or a PNms linkage.

[0037] In a preferred aspect, the nucleotide in the AON at position +19, +15, +13, and / or -6 generates a G:ll wobble base pair with the nucleotide in the human PNPLA3 target transcript molecule and that is directly opposite the nucleotide in the AON.

[0038] The present disclosure relates to an AON as disclosed herein, wherein the AON comprises or consists of an AON according to SEQ ID NO:402, 397, 416, 149, 150, 151 , 152,

[0039] 153, 154, 155, 156, 157, 158, 191 , 192, 196, 197, 237, 238, 239, 240, 241 , 242, 243, 244,

[0040] 245, 246, 247, 302, 304, 306, 313, 333, 340, 341 , 360, 398, 399, 403, 404, 412, 413, 414,

[0041] 415, 417, 418, 419, 420, 423, 424, 425, 432, 434, 435, 436, 437, 438, 439, 440, 441 , 442,

[0042] 443, 444, 446, 489, 508, 509, 510, or 511.

[0043] In a preferred aspect, the AON as disclosed herein is covalently or non-covalently, directly or through a linker, bound to a GalNAc moiety, preferably a tri-antennary GalNAc moiety, as disclosed herein.

[0044] The present disclosure also relates to a pharmaceutical composition comprising an AON as disclosed herein, and a pharmaceutically acceptable solvent, excipient, or carrier.

[0045] The present disclosure also relates to an AON as disclosed herein, for use in the treatment of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis.

[0046] The present disclosure also relates to a use of an AON as disclosed herein, in the manufacture of a medicament for the treatment of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis.

[0047] The present disclosure also relates to a method of editing a human PNPLA3 variant pre-mRNA or mRNA molecule in a liver cell, preferably a hepatocyte, wherein the PNPLA3 variant pre-mRNA or mRNA is encoded by a PNPLA3 gene comprising the rs738409 SNP, the method comprising contacting the PNPLA3 pre-mRNA or mRNA molecule with an AON capable of triggering an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a functionality that is equal or similar to the wildtype functionality, and wherein the AON is as disclosed herein.

[0048] The present disclosure also relates to a method of treating, ameliorating, or slowing down the progression of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis, in a human subject in need thereof, the method comprising administering to said subject an AON as disclosed herein, thereby contacting a PNPLA3 pre-mRNA or mRNA molecule encoded by the PNPLA3 gene comprising the rs738409 SNP in a cell of the subject, thereby effecting an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a functionality that is equal or similar to the wildtype functionality, thereby treating the subject.

[0049] The present disclosure also relates to an in vitro, ex vivo, or in vivo method for the deamination of a target adenosine in a PNPLA3 pre-mRNA or mRNA molecule encoded by a human PNPLA3 gene comprising the rs738409 SNP, in a liver cell, preferably a hepatocyte, the method comprising the steps of: i) providing the cell with an AON as disclosed herein; ii) allowing uptake by the cell of the AON; iii) allowing annealing of the AON to the PNPLA3 pre- mRNA or mRNA molecule; iv) allowing an endogenous ADAR enzyme to deaminate the target adenosine in the PNPLA3 pre-mRNA or mRNA molecule to an inosine; and optionally v) using a functional read-out to identify the presence of the inosine in the PNPLA3 pre-mRNA or mRNA molecule.

[0050] In all aspects of the present disclosure, the deamination of the adenosine in the AUG codon of the I148M mutant transcript version, changes the amino acid to a valine (V). Hence, the wildtype-to-variant situation is reflected by the I148M variant, and the AON as disclosed herein allows for the occurrence of an M148V variant, which should have PNPLA3 wild-type functionality (= equal to or similar functionality in comparison to the wildtype protein) even though isoleucine is not generated at the 148 position.

[0051] The present disclosure also relates to a vector, preferably a viral vector, more preferably an adeno-associated virus (AAV) vector, comprising a nucleic acid molecule encoding an AON as disclosed herein.

[0052] The present disclosure also relates to a nanoparticle delivery vehicle formulation comprising an AON as disclosed herein. Preferably, the nanoparticle delivery vehicle is a lipid nanoparticle (LNP).

[0053] The present disclosure also relates to a pharmaceutical composition comprising an AON, a vector, or a nanoparticle delivery vehicle formulation as disclosed herein, and a pharmaceutically acceptable carrier.

[0054] The present disclosure also relates to a method of editing a human PNPLA3 I148M variant pre-mRNA or mRNA molecule in a liver cell, preferably a hepatocyte, the method comprising contacting the PNPLA3 pre-mRNA or mRNA molecule with an AON capable of triggering an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a functionality that is equal or similar to the wildtype functionality, and wherein the AON is as disclosed herein.

[0055] The present disclosure also relates to a method of treating, ameliorating, or slowing down the progression of a disease caused by the I148M PNPLA3 variant in the liver, such as NAFLD, NASH, fibrosis, and cirrhosis, in a human subject in need thereof, the method comprising administering to said subject an AON, a vector, or a nanoparticle delivery vehicle formulation as disclosed herein, thereby contacting a PNPLA3 I148M variant pre-mRNA or mRNA molecule in a cell of the subject, thereby effecting an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a functionality that is equal or similar to the wildtype functionality, thereby treating the subject.

[0056] BRIEF DESCRIPTION OF THE DRAWINGS

[0057] One or more embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

[0058] Fig. 1 shows part of the 5’ to 3’ sequence of the human wildtype PNPLA3 mRNA transcript in which the AUG codon coding for isoleucine (I) at position 148 in the PNPLA3 protein is underlined (SEQ ID NO:1). In the I148M variant situation, at this position, there is an AUG codon (due to the C>G rs738409 SNP), wherein it then encodes methionine (M). The adenosine in capital A is the target for RNA editing as disclosed herein, which then results in a codon for a valine (V) residue (IUC / GUC) at this position after editing. The target sequence for the I148M variant is provided as SEQ ID NO:56. Notably, it has been shown that a further SNP (rs738408) is in tight disequilibrium with rs738409 (Najafi M. et al. Gene Reports 2022, 26:101472), this additional change in comparison to the sequence of SEQ ID NO:1 is shown as SEQ ID NO:179 and comprises rs738408 as a capital U, five nucleotides 3’ from the target A. Below the respective target sequences, the 5’ to 3’ sequences are provided of the 50 initial AONs (RM 107539 to RM 107588, respectively) that were designed to target the target adenosine in SEQ ID NO:1 , with their respective RM names and SEQ ID NO’s as indicated. RM 107578 to RM 107588 comprise a mismatch (underlined) with the target sequence of SEQ ID NO:1 at positions +3 and +4 in the AON. The AONs represented by RM 107578b to RM 107587b are identical to RM 107578 to RM 107587, respectively except for the +2 position, wherein RM 107578b to RM 107587b comprise a cytidine (Cf; bold and underlined) instead of a guanosine, which makes them fully complementary to the I148M variant target sequence of SEQ ID NO:56 at that position. The chemical modifications are as follows: “Ae” and “Ge” are 2’-MOE modified adenosine and guanosine, respectively; “m5Ce” is 2’-MOE modified 5- methyl-cytidine; “Cm”, “Am”, “Um”, and “Gm” are 2’-OMe modified cytidine, adenosine, uridine, and guanosine, respectively; “Gf”, “Cf”, “Af”, and “Uf” are 2’-F modified guanosine, cytidine, adenosine, and uridine, respectively; “Gl” and “Al” are locked nucleic acids with a guanine and adenine nucleobase, respectively; “Zd” (orphan nucleotide at position 0 in the AON) is a deoxynucleotide (deoxycytidine analog) carrying a Benner’s base; “Id” is deoxyinosine; “I” refers to a PNdmi linkage; “A” refers to a MP linkage; “*” refers to a PS linkage; “e” refers to a phosphodiester (PO) linkage.

[0059] Fig. 2 shows the editing percentage in Primary Human Hepatocytes (PHHs) in a ddPCR experiment, after transfection of 100nM of the indicated AONs (RM numbers provided).

[0060] Fig. 3 shows the editing percentage in PHHs in a ddPCR experiment, after transfection of 200nM of the indicated AONs (RM numbers provided).

[0061] Fig. 4 shows an additional set of 72 AONs that were subsequently designed based on the results with RM107578 (SEQ ID NO:41) shown in Fig. 2 and Fig. 3. The 72 AONs are also directed against the wildtype PNPLA3 target sequence depicted in SEQ ID NO:1 , with the AUG target codon. Of these 72 AONs, 36 AONs (SEQ ID NO:67 to 102) comprise a mismatch with that target sequence at the +3 and +4 position in the AON, whereas the further 36 AONs (SEQ ID NO: 103 to 138) are complementary with the target sequence at these two positions. The chemical modifications are as provided in Fig. 1 , wherein “m5Ue” is 2’-MOE modified 5- methyl-uridine; “Ad” is a deoxyadenosine; and “#” refers to a PNms linkage.

[0062] Fig. 5A and Fig. 5B show the editing percentage in PHHs in a ddPCR experiment, after transfection of 100nM of the indicated AONs as shown in Fig. 4 (RM numbers provided). RM 120392 to RM 120427 are equivalent to RM 120428 to RM 120463, respectively wherein each pair only differs in the presence (MM; black bars) or absence (Non-MM, grey bars) of the two mismatching nucleotides at positions +3 and +4 in the AON (here given as ‘EON’).

[0063] Fig. 6 shows a set of 40 AONs (SEQ ID NO’s and RM numbers provided). The four sets are: i) RM121965 to RM121974, ii) RM121975 to RM121984, iii) RM121985 to RM121994, and iv) RM121995 to RM122004. In this order, the ten AONs in each set are based on RM 107578 to RM 107587, respectively. All AONs are directed to the I148M variant encoding PNPLA3 target sequence depicted in SEQ ID NO:56 with the AUG target codon and therefore comprise a cytidine nucleotide at position +2 (underlined and bold). RM 121965 to RM121974 (set i)) comprise a mismatching nucleotide at position +3 and +4 (underlined). RM 121975 to RM 121984 (set ii)) comprise an A:C mismatching nucleotide at position +3 (underlined), but no mismatch at position +4. RM121985 to RM121994 (set iii)) comprise a C:C mismatching nucleotide at position +4 (underlined), but no mismatch at position +3. RM 121995 to RM 122004 are complementary with the target sequence at positions +3 and +4, and do not mismatch there. The chemical modifications are as shown in Fig. 1 and Fig. 4.

[0064] Fig. 7A and Fig. 7B show the editing percentages in HepG2 cells in a ddPCR experiment after gymnotic uptake of 5pM AON using a pulse treatment with saponin. The results are arranged according to the presence or absence of the mismatching nucleotides at positions +3 and +4 in the AON in comparison to the sequence of SEQ ID NO:56. The black bars represent the results with AONs that only comprise the +3 mismatch (MM @3). The other bars represent the results AONs that only comprise the +4 mismatch (MM @4), both mismatching nucleotides (MM @3 and @4) and no mismatches (no MM). It should be noted that all these AONs (depicted in Fig. 6) have an additional mismatching nucleotide at position +5 (in all cases Gf) with the associated rs738408 polymorphism present in proximity with the rs738409 polymorphism (five nucleotides 3’ from the target A) that causes the occurrence of the I148M variant. HepG2 cells also carry the rs738408 polymorphism. Fig. 7A shows the results of the first experiment and Fig. 7B shows the results of the second experiment using a different set of AONs, as shown. The negative controls RM4266 (SEQ ID NO:298), Mock treatment and no-treatment (NT) did not show any editing. Also, the controls RM 107579 and RM 107578 showed low editing percentages, which was not unexpected because these are directed at editing the target A in a ’wildtype’ environment as shown in SEQ ID NO:1 , rather than on a target sequence comprising the rs738409 alteration, causing I148M (see SEQ ID NO:56) and the associated rs738408 SNP, as shown in SEQ ID NO:179.

[0065] Fig. 8 shows lipid droplet sizes (mean area of Bodipy staining per positively transfected cell) in pm2after transfection with: i) a pCDNA3.1 expression plasmid expressing human PNPLA3-wildtype (PNPLA3-WT) with a FLAG tag at the C-terminus (C-1481) or at the N- terminus (N-1481), represented by the left bar in each panel; ii) a pCNDA3.1 expression plasmid expressing the human I148M PNPLA3 variant (PNPLA3-148M) with a FLAG tag at the C-terminus (C-148M) or at the N-terminus (N-148M), represented by the middle bar in each panel; and iii) a pCDNA3.1 expression plasmid expressing the human M148V PNPLA3 variant (PNPLA3-148V) with a FLAG tag at the C-terminus (C-148V) or at the N-terminus (N- 148V), represented by the right bar in each panel. The left panels in the upper and lower row represent the lipid droplet sizes in positively transfected cells after treatment with Oleic Acid (OA). The middle panels in the upper and lower row represent the lipid droplet sizes in positively transfected cells after treatment with Linoleic Acid (LA). The right panels in the upper and lower row represent the lipid droplet sizes in positively transfected cells after treatment with alpha-Linolenic Acid (aLA). The difference indicated by ‘ns’ is non-significant. One-way ANOVA statistical analysis was performed and showed a significance (****) with p<0.0001 , indicative of the strong effect observed with the 1481 and 148V variants in comparison to the 148M variant.

[0066] Fig. 9 shows the exact same 40 AONs as shown in Fig. 6, except that the Gf at position +5 in all AONs is replaced by an Af. As outlined herein, this calculation + (plus) towards the 5’ terminus, and - (minus) towards the 3’ terminus) is based on the orphan nucleotide (Zd) being position 0. RM numbers and SEQ ID NO’s are provided. Chemical modifications are as provided in Fig. 1. Mismatching nucleotides at positions +3 and +4 are underlined. The nucleotide at +2 that is opposite the guanosine in the AUG codon (in the I148M variant) is provided in bold and is underlined. Fig. 10 shows the editing percentages in HepG2 cells in a dPCR experiment after gymnotic uptake of 5 pM AON using a pulse treatment with saponin. The results are arranged according to the presence or absence of the mismatching nucleotides at positions +3 and +4 in the AON like what is shown in Fig. 7A and Fig. 7B. The negative controls RM4266, Mock treatment and no-treatment (NT) did not show any editing. Also, the controls RM 107579 and RM 107578 showed low editing percentages, which was not unexpected because these are directed at editing the target A in a ’wildtype’ environment as shown in SEQ ID NO:1 , rather than on a target sequence comprising the rs738409 alteration, causing I148M (see SEQ ID NO:56) and the associated rs738408 SNP, as shown in SEQ ID NO:179.

[0067] Fig. 11 shows six sets of eleven AONs each (SEQ ID NO’s and RM numbers provided): i) RM122801 to RM122811 that comprise an A:C mismatch at nucleotide position +3, a C:C mismatch at nucleotide position +4, and a G:ll wobble at nucleotide position +5; ii) RM122812 to RM 122822 that comprise an A:C mismatch at nucleotide position +3 and a G:ll wobble at nucleotide position +5; iii) RM122823 to RM122833 comprise a C:C mismatch at nucleotide position +4, and a G:ll wobble at nucleotide position +5; iv) RM 122834 to RM 122844 comprise an A:C mismatch at nucleotide position +3 and a C:C mismatch at nucleotide position +4; v) RM122845 to RM122855 comprise an A:C mismatch at nucleotide position +3; and vi) RM122856 to RM122866 comprise a C:C mismatch at nucleotide position +4. Mismatching and wobble positions are underlined. The chemical modifications are as shown in Fig. 1 , Fig. 4, and Fig. 6. “ L005” is a tri-antennary GalNAc moiety conjugated to the 3’ terminus of the oligonucleotide, and that is represented by Formula IX(b) herein. Nucleotides provided in bold represent a wobble walk in which nucleotides were incorporated that form a wobble base pair with the nucleotide opposite this nucleotide in the human PNPLA3 target transcript molecule, as shown in SEQ ID NO:179.

[0068] Fig. 12 shows the editing percentages in HepG2 cells in a dPCR experiment after gymnotic uptake of 5 pM AON using a pulse treatment with saponin. The results are arranged according to the presence or absence of the mismatches / wobbles at positions +3, +4, and +5 in the AON in line with the six groups of AONs depicted in Fig. 11. The negative controls RM4266, Mock treatment and no-treatment (NT) did not show any editing.

[0069] Fig. 13A shows the editing percentages in primary human hepatocytes that were engrafted and isolated from a mouse (acquired from Yecuris) using a set of six AONs as indicated and in comparison, to these six AONs, respectively, when conjugated to a tri- antennary GalNAc moiety at the 3’ terminus. Fig. 13B shows the editing percentages in primary human hepatocytes that were from a male liver donor (acquired from BiolVT) again using the set of six AONs as indicated and in comparison, to these six AONs, respectively, when conjugated to a tri-antennary GalNAc moiety at the 3’ terminus. RM4266, RM4777, a non-treated sample (NT), and an AG1856 alone sample served as negative controls. S&T = seed and treat. S&T AG1856 pulse is the same seed and treat procedure with a pulse treatment with AG1856. AG1856 pulse is a procedure without seeding in the presence of AON (which was administered after seeding).

[0070] Fig. 14 shows the editing percentage in human HepG2 cells using the AONs as indicated in a Seed & Treat treatment with and without pulse treatment with the AG1856 saponin. RM4266, RM4777, a non-treated (NT) sample, and a sample that was only treated with AG 1856 served as negative controls.

[0071] Fig. 15 shows the editing percentage in human HepG2 cells using the AONs as indicated that comprise a variety of G:ll wobble base pair generating nucleotides at a variety of positions (wobble walk). A mock treated sample and a non-treated (NT) sample served as negative controls.

[0072] Fig. 16 shows a set of AONs (SEQ ID NO’s and RM numbers provided) that was designed based on the best performing AONs from previous experiments, to test for further chemical modifications and the introduction of additional and / or different mismatching nucleotides and wobble base pair generating nucleotides. The chemical modifications are as shown in Fig. 1 , Fig. 4, Fig. 6, and Fig. 11. “Ed” represents a deoxynucleotide comprising a 5-aza-5,6-di hydro cytosine nucleobase (also known as E base, E nucleobase, or simply as ‘E’; see below). “Xd” represents an a-basic deoxynucleotide. “Im” is a 2’-0Me modified nucleotide comprising a hypoxanthine nucleobase. “If” is a 2’-F modified nucleotide comprising a hypoxanthine nucleobase. “le” is a 2’-MOE modified nucleotide comprising a hypoxanthine nucleobase.

[0073] Fig. 17A shows the editing percentages in primary human hepatocytes (PHH) that were from a male liver donor (acquired from BiolVT), using a set of AONs (EONs) that comprise a 5-aza-5,6-dihydro cytosine nucleobase (E base) at the orphan position. RM123994 and RM 123995 (with a Benner’s base at the orphan position) served as positive controls. An unrelated AON (RM4899), a mock treatment and a non-treated (NT) sample served as negative controls. Two procedures were used: a Seed & Treat procedure (S&T) and a Seed & Treat procedure followed by a AG1856 pulse treatment (S&T AG1856 pulse). Fig. 17B shows the editing percentage in the same type of cells using a set of AONs (EONs) that comprise a deoxyguanosine (Gd) at nucleotide position -1 instead of a deoxyinosine (Id). RM123994 and RM123995 (with Id at -1) served as positive controls. RM4266, a mock treatment and a non-treated (NT) sample served as negative controls.

[0074] Fig. 18 shows the editing percentages in PHH that were from a male liver donor (acquired from BiolVT), using a set of AONs (EONs) that comprise a variety of chemical modifications. RM 123994 and RM 123995 served as positive controls. An unrelated AON (RM4899), a mock treatment and a non-treated (NT) sample served as negative controls. Two procedures were used: a Seed & Treat procedure (S&T) and a Seed & Treat procedure followed by a AG1856 pulse treatment (S&T AG1856 pulse).

[0075] Fig. 19 shows the editing percentages in PHH that were from a male liver donor (acquired from BiolVT), using a set of AONs (EONs) that comprise a variety of linkage modifications in the Editing Enhancing Region (EER), which is the region surrounding the orphan nucleotide. RM 123994 and RM 123995 served as positive controls. An unrelated AON (RM4899), a mock treatment and a non-treated (NT) sample served as negative controls. Two procedures were used: a Seed & Treat procedure (S&T) and a Seed & Treat procedure followed by a AG1856 pulse treatment (S&T AG1856 pulse).

[0076] Fig. 20 shows the editing percentages in PHH that were from a male liver donor (acquired from BiolVT), using a set of AONs (EONs) that varied in length and symmetry (calculated from the orphan nucleotide position). RM 123994 and RM 123995 served as positive controls. An unrelated AON (RM4899), a mock treatment and a non-treated (NT) sample served as negative controls. Two procedures were used: a Seed & Treat procedure (S&T) and a Seed & Treat procedure followed by a AG1856 pulse treatment (S&T AG1856 pulse).

[0077] Fig. 21 shows the editing percentages in PHH that were from a male liver donor (acquired from BiolVT), using a set of AONs (EONs) that comprise a further variety of chemical modifications. RM 123994 and RM 123995 served as positive controls. An unrelated AON (RM4899), a saponin-only treatment and a non-treated (NT) sample served as negative controls. Two procedures were used: a Seed & Treat procedure (S&T) and a Seed & Treat procedure followed by a AG1856 pulse treatment (S&T AG1856 pulse).

[0078] DETAILED DESCRIPTION

[0079] The present disclosure relates to using antisense oligonucleotide (AONs) and the cell’s own nucleic acid editing machinery in order to specifically target the adenosine in the AUG codon encoding the methionine in the I148M variant within the human PNPLA3 transcript, thereby providing a PNPLA3 protein with normalized function, namely wherein the AUG codon is changed to IUG, which is read as GUG by the translation machinery (see below) and thereby encodes a valine residue (V). Hence, the RNA editing changes the methionine in the I148M variant to a valine, and thus the editing ensures the occurrence of a M148V change in human subjects affected with an I148M variant. The technology that the present disclosure relates to is generally referred to as ‘RNA editing’.

[0080] RNA editing is a natural process through which eukaryotic cells alter the sequence of their RNA molecules, often in a site-specific and precise way, thereby increasing the repertoire of genome encoded RNAs by several orders of magnitude. RNA editing enzymes have been described for eukaryotic species throughout the animal and plant kingdoms, and these processes play an important role in managing cellular homeostasis in metazoans from the simplest life forms (such as Caenorhabditis elegans) to humans. Examples of RNA editing are adenosine (A) to inosine (I) conversions and cytidine (C) to uridine (U) conversions, which occur through enzymes called Adenosine Deaminases acting on RNA (ADAR) and APOBEC / AID (cytidine deaminases that act on RNA), respectively. ADAR is a multi-domain protein, comprising of a catalytic domain and two to three double-stranded RNA recognition domains, depending on the enzyme in question. Each recognition domain recognizes a specific double-stranded RNA (dsRNA) sequence and / or conformation. The catalytic domain does also play a role in recognizing and binding a part of the dsRNA helix, although the key function of the catalytic domain is to convert an A into an I in a nearby, predefined, position in the target RNA, by deamination of the nucleobase. Inosine is read as guanosine by the translational machinery of the cell, meaning that, if an edited adenosine is in a coding region of an mRNA or pre-mRNA, it can recode the protein sequence. A-to-l conversions may also occur in 5’ non-coding sequences of a target mRNA, creating new translational start sites upstream of the original start site, which gives rise to N-terminally extended proteins, or in the 3’ untranslated region (UTR) or other non-coding parts of the transcript, which may affect the processing and / or stability of the RNA. In addition, A-to-l conversions may take place in splice elements in introns or exons in pre-mRNAs, thereby altering the pattern of splicing. As a result, exons may be included or skipped. The enzymes catalysing adenosine deamination are within an enzyme family of ADARs, which include human deaminases hADARI and hADAR2, as well as hADAR3. However, for hADAR3 no deaminase activity has been demonstrated.

[0081] The use of oligonucleotides to edit a target RNA, applying adenosine deaminase, has been described (e.g., Woolf et al. Proc Natl Acad Sci USA 1995, 92:8298-8302; Montiel- Gonzalez et al. Proc Natl Acad Sci USA 2013, 110(45): 18285-18290; Vogel et al. Angewandte Chemie Int 2014, Ed 53:267-271). A disadvantage of the method described by Montiel- Gonzalez et al. (2013) is the need for a fusion protein consisting of the boxB recognition domain of bacteriophage Lambda N-protein, fused to the adenosine deaminase domain of a truncated natural ADAR protein. It requires target cells to be either transduced with the fusion protein, which is a major hurdle, or that target cells are transfected with a nucleic acid construct encoding the engineered adenosine deaminase fusion protein for expression. The system described by Vogel et al. (2014) suffers from similar drawbacks, in that it is not clear how to apply the system without having to genetically modify the ADAR first and subsequently transfect or transform the cells harboring the target RNA, to provide the cells with this genetically engineered protein. US 9,650,627 describes a similar system. The oligonucleotides of Woolf et al. (1995) that were 100% complementary to the target RNA sequences, suffered from severe lack of specificity: nearly all adenosines in the target RNA strand that were complementary to the AON were edited.

[0082] It is known that ADAR may act on any dsRNA. Through a process sometimes referred to as ‘promiscuous editing’, the enzyme will edit multiple A’s in the dsRNA. Hence, there was a need for methods and means that circumvent such promiscuous editing and only target specific adenosines in a target RNA molecule to become therapeutic applicable. Vogel et al. (2014) showed that such off-target editing can be suppressed by using 2’-O-methyl (2’-OMe) modified nucleosides in the oligonucleotide at positions opposite to adenosines that should not be edited and used a non-modified nucleoside directly opposite to the specifically targeted adenosine on the target RNA. However, the specific editing effect at the target nucleotide has not been shown to take place without the use of recombinant ADAR enzymes having covalent bonds with the AON. Several publications have now shown that the recruitment of endogenous ADAR (hence without the need for an exogenous and / or recombinant source) is feasible while maintaining a specificity in which a single adenosine within a target RNA molecule can be targeted and deaminated to an inosine. WO2016 / 097212 discloses AONs for the targeted editing of RNA, wherein the AONs are characterized by a sequence that is complementary to a target RNA sequence (therein referred to as the ‘targeting portion’) and by the presence of a stem-loop I hairpin structure (therein referred to as the ‘recruitment portion’), which is preferably non-complementary to the target RNA. Such oligonucleotides are referred to as ‘self-looping AONs’. The recruitment portion acts in recruiting a natural ADAR enzyme present in the cell (endogenously present) to the dsRNA formed by hybridization of the target sequence with the targeting portion. Due to the recruitment portion, there is no need for conjugated entities or presence of modified recombinant ADAR enzymes. WO2016 / 097212 describes the recruitment portion as being a stem-loop structure mimicking either a natural substrate (e.g., the GluB receptor) ora Z-DNA structure known to be recognized by the dsRNA binding domains, or Z-DNA binding domains, of ADAR enzymes. A stem-loop structure can be an intermolecular stem-loop structure, formed by two separate nucleic acid strands, or an intramolecular stem loop structure, formed within a single nucleic acid strand. The stem-loop structure of the recruitment portion as described is an intramolecular stem-loop structure, formed within the AON itself, and are thought to attract (endogenous) ADAR. Similar stemloop structure-comprising systems for RNA editing have since then been described in WO2017 / 050306, W02020 / 001793, WO2017 / 010556, US11 ,390,865, W02020 / 246560, and WO2022 / 078995.

[0083] WO2017 / 220751 and WO2018 / 041973 describe a next generation type of AONs that do not comprise such a stem-loop structure but that are (almost fully) complementary to the targeted area, and that appeared still capable of attracting endogenous ADAR enzymes. In one embodiment, one or more mismatching nucleotides, wobbles, or bulges exist between the oligonucleotide and the target sequence. A sole mismatch may be at the site of the nucleoside opposite the target adenosine, but in other embodiments AONs (or “RNA editing oligonucleotides” - even though the deamination reaction is carried out by the ADAR enzyme - and often abbreviated to ‘EONs’) were described with multiple bulges and / or wobbles when attached to the target sequence area. It appeared possible to achieve in vitro, ex vivo and in vivo RNA editing with AONs lacking a stem-loop structure and with endogenous ADAR enzymes when the sequence of the AON was carefully selected such that it could attract / recruit ADAR. The ‘orphan nucleoside’, which is defined as the nucleoside in the AON that is positioned directly opposite the target adenosine in the target RNA molecule, was a nucleotide with an unmodified cytosine nucleobase and that did not carry a 2’-0Me modification. The orphan nucleoside can be a deoxyribonucleoside (DNA), wherein the remainder of the AON could still carry 2’-0-alkyl modifications at the sugar entity (such as 2’- OMe), or the nucleotides directly surrounding the orphan nucleoside contained chemical modifications (such as DNA in comparison to RNA) that further improved the RNA editing efficiency and / or increased the resistance against nucleases. Such effects could even be further improved by using sense oligonucleotides (SONs) that ‘protected’ the AONs against breakdown upon delivery to the cells (described in WO2018 / 134301 and US11 ,274, 300).

[0084] The use of chemical modifications and particular structures in oligonucleotides that could be used in ADAR-mediated editing of specific adenosines in a target RNA have been the subject of numerous disclosures in the field, such as WO2019 / 111957, WO2019 / 158475, W02020 / 165077, W02020 / 201406, W02020 / 211780, WO2021 / 008447, WO2021 / 020550,

[0085] WO2021 / 060527, WO2021 / 117729, WO2021 / 136408, WO2021 / 182474, WO2021 / 216853,

[0086] WO2021 / 242778, WO2021 / 242870, WO2021 / 242889, W02022 / 007803, W02022 / 018207,

[0087] WO2022 / 026928, and WO2022 / 124345. The use of specific sugar moieties has been disclosed in for instance W02020 / 154342, W02020 / 154343, W02020 / 154344,

[0088] WO2022 / 103839, and WO2022 / 103852, whereas the use of stereo-defined linker moieties (in general for oligonucleotides that for instance can be used for exon skipping, in gapmers, in siRNA, or specifically for RNA-editing oligonucleotides, related to a wide variety of target sequences) has been described in WO2011 / 005761 , W02014 / 010250, WO2014 / 012081 ,

[0089] WO2015 / 107425, WO2017 / 015575 (HTT), WO2017 / 062862, W02017 / 160741 ,

[0090] WO2017 / 192664, WO2017 / 192679 (DMD), WO2017 / 198775, WO2017 / 210647,

[0091] WO2018 / 067973, WO2018 / 098264, WO2018 / 223073 (APOC3), WO2018 / 237194,

[0092] WO2019 / 032607 (C9orf72), WO2019 / 055951 , WO2019 / 075357 (SMA / ALS),

[0093] W02019 / 200185 (DM1), WO2019 / 217784 (DM1), WO2019 / 219581 , W02020 / 118246 (DM1), W02020 / 160336 (HTT), WO2020 / 191252, W02020 / 196662, WO2020 / 219981 (USH2A), WO2020 / 219983 (RHO), WO2020 / 227691 (C9orf72), WO2021 / 071788 (C9orf72), WO2021 / 071858, WO2021 / 178237 (MAPT), WO2021 / 234459, WO2021 / 237223, WO2022 / 099159, WO2021 / 030778, WO2022 / 174053, and WO2023 / 278589. Next to these disclosures, an extensive number of publications relate to the targeting of specific RNA target molecules, or specific adenosines within such RNA target molecules, be it to repair a mutation that resulted in a premature stop codon, or other mutation causing disease. Examples of such disclosures in which adenosines are targeted within specified target RNA molecules are W02020 / 157008 and WO2021 / 136404 (USH2A); WO2021 / 113270 (APP); WO2021 / 113390 (CMT1A); W02021 / 209010 (IDUA, for Hurler syndrome); WO2021 / 231673 and WO2021 / 242903 (LRRK2); WO2021 / 231675 (ASS1); WO2021 / 231679 (GJB2);

[0094] WO2019 / 071274 and WO2021 / 231680 (MECP2, for RETT syndrome); WO2021 / 231685 and WO2021 / 231692 (OTOF, for autosomal recessive non-syndromic hearing loss); WO2021 / 231691 (XLRS); WO2021 / 231698 (argininosuccinate lyase deficiency); W02021 / 130313 and WO2021 / 231830 (ABCA4, for Stargardt disease); WO2023 / 152371 (PCSK9); and WO2021 / 243023 (SERPINA1 ; for Alpha-1 -Antitrypsin deficiency; see also WO2016 / 097212, WO2017 / 220751 , and WO2018 / 041973). It is noted that WO2018 / 223056 and WO2018 / 223081 disclose the use of oligonucleotides, such as ssRNAi to lower the expression of PNPLA3. WO2022 / 256283 discloses RNA editing to target transcripts of pathogenic proteins, including ATP7B, ABCB11 , ACADM, ABCC6, MECP2 and the I148M variant of PNPLA3 (to target the single nucleotide polymorphism (SNP) that relates to the change of methionine at position 148 to valine). The present disclosure provides an alternative approach with alternative improved compounds to provide the M148V change as tool for the treatment of NAFLD, liver failure, liver cirrhosis, NASH, and other liver disorders caused by the PNPLA3 I148M variant.

[0095] The AONs as disclosed herein can recruit deaminating enzymes, such as ADAR1 and / or ADAR2 that are endogenously present (= naturally present) in a cell. An AON as disclosed herein can mediate RNA editing of a target adenosine present in a target RNA molecule after it is bound to the target RNA molecule, since the deaminating enzymes are recruited to the double-stranded AON / target RNA molecule complex and subsequently deaminate the target adenosine into an inosine. The AONs of the present disclosure are preferably conjugated to a saponin and to a GalNAc moiety either directly or indirectly via linkers. These conjugations provide an improved delivery and endosomal release after the AON has entered the target cell. In another preferred aspect the present disclosure provides AONs that are conjugated to a saponin and subsequently ‘packaged’ in lipid nanoparticles (LNPs) through which the AON can be delivered very efficiently to the target cells and subsequently released from the endosome as soon as the AON has entered the cell. These combinations, together with improved combinations of chemical modifications within the AON itself, and the choice of length, symmetry, and presence / absence of mismatching nucleotides ensure an improvement over the current available medicaments to treat FLD caused by the I148M variant of PNPLA3.

[0096] The oligonucleotides are herein abbreviated to “AONs”, but sometimes also referred to as ‘editing oligonucleotides’, or ‘EONs’, even though the RNA editing event is performed by the deamination enzyme and the action of the oligonucleotide only triggers the RNA editing to take place. There is a constant need for improving the pharmacokinetic properties of the AONs without negatively affecting the efficiency in which the target adenosine is edited in the target RNA, and / or without negatively affecting the stability of the AON itself, which is constantly prone to breakdown because of nucleases present in a natural cell. Many chemical modifications are available for the generation of AONs (and many have been applied in the art). However, many of these properties are not always compatible with the desire of achieving efficient RNA editing. In the search for better pharmacokinetic properties, it was found earlier that a 2’-O-methoxyethyl (or 2’-methoxyethoxy, or 2’-M0E) modification of the ribose of some, but not all, nucleotides surprisingly appeared compatible with efficient ADAR engagement and editing (WO2019 / 158475). In a similar fashion, it was found earlier that a PS linkage at some, but not all, internucleoside linkages surprisingly appeared compatible with efficient ADAR engagement and editing (WO2019 / 219581). Also, it was found earlier that phosphonoacetate linkage modifications and / or unlocked nucleic acid (UNA) ribose modifications of some, but not all, positions in the AON appeared compatible with efficient engagement of an enzyme with nucleotide deamination activity and with subsequent deamination (W02020 / 165077). Whereas the properties of phosphonoacetate and UNA modifications were known as such, the compatibility thereof with engagement of enzymes with nucleotide deamination activity and with the deamination reaction was not known.

[0097] The present disclosure relates to AONs that mediate RNA editing, using endogenous (naturally present) ADAR enzymes in the host cell, preferably hepatocytes, of an adenosine present in the transcript of the I148M variant of the PNPLA3 gene.

[0098] It was realized that RNA editing could not revert the I148M variant back to wild type, because human ADAR1 and ADAR2 cannot edit a guanosine to a cytidine. However, based on the structural model described of PNPLA3 (He S et al. J Biol Chem. 2010, 285(9):6706- 6715), a valine substitution in position 148 would likely not interfere with substrate access to the catalytic centre of PNPLA3, formed by the Asp166-Ser47 dyad, as the valine side chain is much shorter than the methionine side chain and only differing from the isoleucine by one methyl group. As described by He et al. (2010), the I148M substitution does not affect the distance between the two residues - Asp166 and Ser47 - forming the catalytic dyad, which strongly suggests that a valine substitution would not affect the catalytic centre either, while it would restore substrate access. Hence, it is envisioned that changing the methionine at position 148 to a valine would mimic the wild-type situation in which an isoleucine is present in a sufficient manner to take away, or at least sufficiently lower the detrimental and pathogenic activity exerted by the I148M variant.

[0099] Definitions

[0100] Whenever reference is made to an oligonucleotide, oligo, ON, ASO, oligonucleotide composition, antisense oligonucleotide, AON, (RNA) editing oligonucleotide, EON, and RNA (antisense) oligonucleotide, both oligoribonucleotides and deoxyoligoribonucleotides are meant unless the context dictates otherwise. Potentially the oligonucleotide may completely lack RNA and DNA nucleotides (as they appear in nature) and may consist completely of modified nucleotides. Whenever reference is made to an ‘oligoribonucleotide’ it may comprise the bases A, G, C, II, or I. Whenever reference is made to a ‘deoxyoligoribonucleotide’ it may comprise the bases A, G, C, T, or I. However, an AON as disclosed herein may comprise a mix of ribonucleotides and deoxyribonucleotides. When a deoxyribonucleotide is used, hence without a modification at the 2’ position of the sugar, the nucleotide is often abbreviated to dA (or Ad), dC (or Cd), dG (or Gd), dl (or Id), or T in which the ‘d’ represents the deoxy nature of the nucleoside, while a ribonucleoside that is either normal RNA or modified at the 2’ position is often abbreviated without the ‘d’, and often abbreviated with their respective modifications and as explained herein.

[0101] The term ‘nucleoside’ refers to the nucleobase linked to the (deoxy)ribosyl sugar, without phosphate groups. A ‘nucleotide’ is composed of a nucleoside and one or more phosphate groups. The term ‘nucleotide’ thus refers to the respective nucleobase- (deoxy)ribosyl-phospholinker, as well as any chemical modifications of the ribose moiety or the phospho group. Thus, the term would include a nucleotide including a locked ribosyl moiety (comprising a 2’-4’ bridge, comprising a methylene group or any other group), an unlocked nucleic acid (UNA), a threose nucleic acid (TNA), a nucleotide including a linker comprising a phosphodiester (PO), phosphonoacetate, phosphotriester, PS, phosphoro(di)thioate, MP (or MeP), methyl thiophosphonate, phosphoramidate linkages, PNdmi, and a linkage according to the structure of formula (I) as described herein. Sometimes the terms nucleobase, nucleoside and nucleotide are used interchangeably, unless the context clearly requires differently, for instance when a nucleoside is linked to a neighbouring nucleoside and the linkage between these nucleosides is modified. As stated herein, a nucleotide is a nucleoside plus one or more phosphate groups. The terms ‘ribonucleoside’ and ‘deoxyribonucleoside’, or ‘ribose’ and ‘deoxyribose’ are as used in the art.

[0102] Sometimes the terms adenosine and adenine, guanosine and guanine, cytidine and cytosine, uracil and uridine, thymine and thymidine / uridine, inosine, and hypoxanthine, are used interchangeably to refer to the corresponding nucleobase on the one hand, and the nucleoside or nucleotide on the other. The nucleobase thymine (T) is also known as 5- methyluracil (m5U) and is an uracil (U) derivative; thymine and 5-methyluracil can be interchanged throughout the document text. Likewise, the nucleotide thymidine is also known as 5-methyluridine and is a uridine derivative; thymidine and 5-methyluridine can be interchanged throughout the document text.

[0103] Whenever reference is made to nucleotides in the oligonucleotide, such as cytosine, 5-methylcytosine, 5-hydroxymethylcytosine, 5-formylcytosine, 5-acetylcytosine, 5- hydroxycytosine, and p-D-glucosyl-5-hydroxymethylcytosine are included. Whenever reference is made to adenine, N6-methyladenine, 8-oxo-adenine, 2,6-diaminopurine and 7-methyladenine are included. Whenever reference is made to uracil, di hydrouracil, isouracil, N3-glycosylated uracil, pseudouracil, 5-methyluracil, N1-methylpseudouracil, 4-thiouracil and 5-hydroxymethyluracil are included. Whenever reference is made to guanine, 1-methylguanine, 7-methylguanosine, N2,N2-dimethylguanosine, N2,N2,7- trimethylguanosine and N2,7-dimethylguanosine are included. Whenever reference is made to nucleosides or nucleotides, ribofuranose derivatives, such as 2’-deoxy, 2’-hydroxy, and 2’- O-substituted variants, such as 2’-O-methyl (2’-0Me), are included, as well as other modifications, including 2’-4’ bridged variants. Whenever reference is made to oligonucleotides, one or more linkages may be a naturally occurring PO linkage, whereas the remaining linkages between two mononucleotides may be a modified linkage. Examples of such modified linkages are phosphonoacetate, phosphotriester, PS, phosphoro(di)thioate, MP, phosphoramidate linkages, phosphoryl guanidine, thiophosphoryl guanidine, sulfono phosphoramidate, PNdmi and the linkage structure according to formula (I), further outlined in detail below.

[0104] The term ‘comprising’ encompasses ‘including’ as well as ‘consisting of’, e.g., a composition ‘comprising X’ may consist exclusively of X or may include something additional, e.g., X + Y. The term ‘about’ in relation to a numerical value x is optional and means, e.g., x+10%.

[0105] The word ‘substantially’ does not exclude ‘completely’, e.g., a composition which is ‘substantially free from Y’ may be completely free from Y. Where relevant, the word ‘substantially’ may be omitted from the definition of the invention.

[0106] The term ‘conducive to’ or ‘mediate’ can be used interchangeably with ‘capable of facilitating’. When used in the context of an AON that is conducive to ADAR editing (or can mediate ADAR editing), this means that the AON, after entry into the cell, interacts with the target RNA sequence, thereby forming a double stranded structure which is recognized by the ADAR enzyme, which can then deaminate the target adenosine into an inosine. Hence, the AON itself does not have the enzymatic function (the ADAR enzyme has), but it can trigger, induce, cause, organize, mediate, provide, give, produce, facilitate, result in RNA editing after binding to the target RNA molecule.

[0107] The term ‘mismatch’ is used herein to refer to opposing nucleotides in a double stranded RNA complex which do not form perfect base pairs according to the Watson-Crick base pairing rules. In this historical sense, mismatches are G:A, C:A, U:C, A:A, G:G, 0:0, and U: U. In some embodiments AONs as disclosed herein comprise fewer than four mismatches with the target sequence, for example 0, 1 or 2 mismatches. ‘Wobble’ base pairs are G: U, I : U, l:A, and l:C base pairs. When a II is placed opposite the target A, there is no mismatch, and the AON may be 100% complementary. When a C is placed opposite the target A, there is at least 1 mismatch between the AON and the target sequence. A-basic nucleotides (nucleotides lacking a nucleobase) within the AON mismatch with their opposite nucleotide in the target sequence (because there can be no base ‘pair’). Although a G:G pairing would be considered a mismatch, that does not necessarily mean that the interaction is unstable, which means that the term ‘mismatch’ may be somewhat outdated based on the current disclosure where a Hoogsteen base-pairing may be seen as a mismatch based on the origin of the nucleotide but still be relatively stable. An isolated G:G pairing in duplex RNA can for instance be quite stable but still be defined as a mismatch. Analysis of natural targets of ADAR enzymes has indicated that these generally include mismatches between the two strands that form the RNA helix edited by ADAR1 or 2. It has been suggested that these mismatches enhance the specificity of the editing reaction (Stefl et al. 2006. Structure 14(2): 345-355; Tian et al. 2011. Nucleic Acids Res 39(13):5669-5681). Characterization of optimal patterns of paired / mismatched nucleotides between the AONs and the target RNA also appears important to the development of efficient ADAR-based AON therapy.

[0108] The term ‘complementary’ as used herein refers to the fact that the AON hybridizes under physiological conditions to a second nucleic acid strand. Examples are (i) when the AON as a first nucleic acid strand (= guide oligonucleotide) forms a heteroduplex RNA editing oligonucleotide complex with second complementary nucleic acid strand in vitro), or (ii) when it forms a double stranded complex with the target RNA molecule. The term does not necessarily mean that each nucleotide in a nucleic acid strand has a perfect pairing with its opposite nucleotide in the opposite sequence. In other words, while an AON may be complementary to a target sequence, there may be mismatches, wobbles and / or bulges between the AON and the target sequence, while under physiological conditions that AON still hybridizes to the target sequence such that the cellular RNA editing enzymes can deaminate the target adenosine to an inosine. The term ‘substantially complementary’ therefore also means that despite the presence of the mismatches, wobbles, and / or bulges, the AON has enough matching nucleotides with the target sequence that under physiological conditions the AON hybridizes to the target RNA molecule. As shown herein, an AON may be complementary, but may also comprise one or more mismatches, wobbles and / or bulges with the target sequence, if under physiological conditions the AON is able to hybridize to its target.

[0109] The term ‘orphan nucleotide’ relates to the nucleotide in the AON that is directly opposite the target adenosine, which is the adenosine that is deaminated by the deaminating enzyme. The orphan nucleotide may be a natural cytidine or deoxycytidine, or a uridine or deoxyuridine. It may also be a chemically modified nucleotide, as further described in detail below, or a known or chemically modified analog of a natural (deoxy)cytidine, such as a nucleotide carrying a Benner’s base, or a known or chemically modified analog of a natural (deoxy)uridine, such as iso-uridine, as further outlined in detail below. A ‘nucleotide analog’ refers to an analog of a nucleic acid nucleotide. The nucleotide analog is an analog of adenosine, guanosine, cytidine, thymidine, uridine, deoxyadenosine, deoxyguanosine, deoxycytidine, deoxythymidine or deoxyuridine.

[0110] The term ‘downstream’ in relation to a nucleic acid sequence means further along the sequence in the 3' direction; the term ‘upstream’ means the converse. Thus, in any sequence encoding a polypeptide, the start codon is upstream of the stop codon in the sense strand but is downstream of the stop codon in the antisense strand. The same holds true for the AONs as disclosed herein. Nucleotides that are upstream of the orphan nucleotide in the antisense oligonucleotide are located towards the 5’ terminus, and nucleotides that are downstream of the orphan nucleotide are located towards the 3’ terminus.

[0111] The nucleotide ‘numbering’ in an AON as disclosed herein is such that the orphan nucleotide is number 0 and the nucleotide 5’ from the orphan nucleotide is number +1. Counting is further positively (+) incremented towards the 5’ end and negatively (-) incremented towards the 3’ end, wherein the first nucleotide 3’ from the orphan nucleotide is number -1 . The internucleoside linkage numbering in the AON is such that linkage number 0 is the linkage 5’ from the orphan nucleotide, and the linkage positions in the oligonucleotide are positively (+) incremented towards the 5’ end and negatively (-) incremented towards the 3’ end.

[0112] References to ‘hybridisation’ typically refer to specific hybridisation and exclude non-specific hybridisation. Specific hybridisation can occur under experimental conditions chosen, using techniques well known in the art, to ensure that most stable interactions between probe and target are where the probe and target have at least 70%, preferably at least 80%, more preferably at least 90% sequence identity.

[0113] The term ‘splice mutation’ relates to a mutation in a gene that encodes fora pre-mRNA, wherein the splicing machinery is dysfunctional in the sense that splicing of introns from exons is disturbed and due to the aberrant splicing, the subsequent translation is out of frame resulting in premature termination of the encoded protein. Often such shortened proteins degrade rapidly and do not have any functional activity.

[0114] Whenever a ‘naked’ form in relation to the AON as disclosed herein is referred to, it means that the AON is manufactured in a laboratory or manufacturing facility, through which it is generally chemically modified to prevent it from rapid degradation after it enters the mammalian body or a tissue, or cell, upon administration. A naked form of an AON is therefore different from a form in which the AON is encoded (and delivered) by a viral genome or within a plasmid vector. When such viral vectors or plasmid vectors are administered, the encoded AON is expressed from the viral vector genome or from the plasmid in the cell to which the viral vector or plasmid vector is delivered. Consequently, the AON is then not chemically modified and comprises solely naturally occurring RNA nucleotides. The length of the AON as disclosed herein, and when delivered in a naked form is preferably 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 , 32, 33, 34, 35, 36, 37, 38, 39, 40, 41 , 42, 43, 44, 45, 46, 47, 48, 49, 50, 51 , 52, 53, 54, 55, 56, 57, 58, 59, or 60 nucleotides in length. However, when the AON as disclosed herein is to be delivered through the expression of a viral vector, then the AON may be longer, such as 70, 80, 90, 100, 150, or 200 or more nucleotides in length.

[0115] The term ‘HEON’ refers to a heteroduplex double-stranded complex molecule wherein an AON as disclosed herein is hybridized to a partially or fully complementary, partially of fully overlapping sense oligonucleotide. Because the AON as disclosed herein often has specified chemical modifications that are different from the chemical modifications in the sense strand, the two strands form such a heteroduplex RNA editing oligonucleotide complex. The sense strand may be chemically modified almost in its entirety, similar or different to what is performed in the AON as disclosed herein, for example by providing nucleotides with a ribose sugar moiety carrying a 2’-OMe substitution, a 2’-F substitution, or a 2’-MOE substitution. It is to be understood that the sense strand present in the HEON is a different entity in comparison to the target RNA molecule in the cell. The sense strand in an HEON is preferably 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 , 32, 33, 34, 35, 36, 37, 38, 39, 40, 41 , 42, 43, 44, 45, 46, 47, 48, 49, 50, 51 , 52, 53, 54, 55, 56, 57, 58, 59, or 60 nucleotides in length. The HEON is often generated in vitro and used as a delivery tool to protect the AON from degradation when administered to the cell. In other words, the HEON is preferably formed before the AON is administered to the cell.

[0116] Embodiments

[0117] The present disclosure relates to an AON that is capable of recruiting an endogenous ADAR enzyme in a human cell after the AON has formed a double-stranded complex with a region of a target RNA nucleic acid molecule in the cell, wherein the cell is a liver cell, preferably a hepatocyte, wherein the transcript molecule is a pre-mRNA or an mRNA molecule, wherein the region comprises a target adenosine, wherein the nucleotide in the AON that is opposite the target adenosine is the orphan nucleotide, wherein the nucleotide numbering is such that the orphan nucleotide is number 0 and nucleotides are further positively (+) incremented towards the 5’ end and negatively (-) incremented towards the 3’ end, wherein the ADAR enzyme can deaminate the target adenosine into an inosine after binding to the double-stranded complex, wherein the target RNA nucleic acid molecule is a transcript molecule of the human PNPLA3 gene comprising the rs738409 single nucleotide polymorphism (SNP), and wherein the target adenosine is in the AUG codon coding for methionine (M) at position 148 of the encoded PNPLA3 protein. In a preferred aspect, the orphan nucleotide is a deoxynucleotide comprising a cytosine or a cytosine analog, wherein the cytosine analog preferably comprises a 6-amino- 5-nitro-3-yl-2(1 H)-pyridone nucleobase (Benner’s base) or a 5-aza-5,6-di hydro cytosine nucleobase (E base).

[0118] In a preferred aspect, the nucleotide at the -1 position is a deoxyinosine (Id).

[0119] In a preferred aspect, the nucleotide at the +2 position comprises a cytosine nucleobase, and preferably further comprises a 2’-F substitution at the ribose sugar moiety.

[0120] In a preferred aspect, the nucleotide at the +3 position generates an A:C mismatch with the nucleotide in the human PNPLA3 transcript molecule that is directly opposite the nucleotide at the +3 position.

[0121] In a preferred aspect, the nucleotide at the +4 position generates a Watson-Crick G:C match (base pairs) with the nucleotide in the human PNPLA3 transcript molecule that is directly opposite the nucleotide at the +4 position.

[0122] In all aspects of the present disclosure, it is preferred that the human PNPLA3 gene comprises the rs738408 SNP (together with the rs738409 SNP), and wherein the nucleotide at the +5 position generates a G:ll wobble base pair with the nucleotide in the human PNPLA3 transcript molecule that is directly opposite the nucleotide at the +5 position.

[0123] In one aspect, the linkage numbering in the AON as disclosed herein is such that the linkage that is directly 5’ of the orphan position (nucleoside) is linkage position 0, and the linkages are positively (+) incremented towards the 5’ terminus and negatively (-) invremented towards the 3’ terminus of the AON.

[0124] In one particularly preferred aspect of the present disclosure, the AON is 30 nucleotides in length and has the following structure, wherein the ‘N’ nucleotides are the positive (+) positions from the orphan nucleotide and the ‘M’ nucleotides are the negative (-) positions from the orphan nucleotide: 5’-N23N22N2l N20Nl9Nl8Nl7Nl6Nl5Nl4Nl3Nl2Nl l Nl0N9N8N7N6N5N4N3N2NlOMl M2M3M4M5M6-3’ wherein:

[0125] O is the orphan nucleotide at nucleotide position 0, which is: i) a deoxynucleotide comprising a 6-amino-5-nitro-3-yl-2(1 H)-pyridone nucleobase; or ii) a deoxynucleotide comprising a 5-aza-5,6-dihydro cytosine nucleobase;

[0126] Mi at position -1 is a nucleotide comprising a hypoxanthine nucleobase, preferably wherein the nucleotide is a deoxyinosine (Id);

[0127] M2 at position -2 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is a 2’-OMe modified adenosine (Am);

[0128] N1 at position +1 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is a deoxyadenosine (Ad) or a 2’-MOE modified adenosine (Ae); N2 at position +2 is a nucleotide comprising a cytosine nucleobase, preferably wherein the nucleotide is a 2’-F modified cytidine (Cf);

[0129] N3 at position +3 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is Ae;

[0130] N4 at position +4 is a nucleotide comprising a guanine nucleobase, preferably wherein the nucleotide is a 2’-MOE modified guanosine (Ge);

[0131] Ns at position +5 is a nucleotide comprising an adenine nucleobase or a guanine nucleobase, preferably wherein the nucleotide is 2’-F modified, more preferably wherein the nucleotide is a 2’-F modified guanosine (Gf); and wherein the linkage between M2 and M3 (linkage position -3) is a phosphorothioate (PS) linkage or a phosphodiester (PO) linkage, preferably a PO linkage.

[0132] In a preferred aspect, the AON of the present disclosure has the structure as provided above, and the linkage between:

[0133] O and Mi (linkage position -1) is a PS or a mesyl phosphoramidate (PNms) linkage;

[0134] Mi and M2 (linkage position -2) is a PS, a PNms, a (1 ,3-dimethylimidazolidin-2-ylidene) phosphoramidate (PNdmi), or a methylphosphonate (MP) linkage;

[0135] N1 and O (linkage position 0) is a PO, a PS, or a PNms linkage;

[0136] N2 and N1 (linkage position +1) is a PS linkage;

[0137] N3 and N2 (linkage position +2) is a PO or a PS linkage;

[0138] N4 and N3 (linkage position +3) is a PO or a PS linkage; and / or

[0139] Ns and N4 (linkage position +4) is a PNdmi or a PNms linkage.

[0140] In a preferred aspect, the linkage between the most terminal two nucleotides on the 5’ and / or 3’ terminus of the AON is a PNdmi linkage or a PNms linkage.

[0141] In a preferred aspect, the nucleotide in the AON at position +19, +15, +13, and / or -6 generates a G:ll wobble base pair with the nucleotide in the human PNPLA3 target transcript molecule and that is directly opposite the nucleotide in the AON.

[0142] In a particularly preferred embodiment, the AON as disclosed herein comprises or consists of an AON (with its sequence and chemical modifications) as provided by SEQ ID NO:402, 397, 416, 149, 150, 151 , 152, 153, 154, 155, 156, 157, 158, 191 , 192, 196, 197, 237, 238, 239, 240, 241 , 242, 243, 244, 245, 246, 247, 302, 304, 306, 313, 333, 340, 341 , 360, 398, 399, 403, 404, 412, 413, 414, 415, 417, 418, 419, 420, 423, 424, 425, 432, 434, 435, 436, 437, 438, 439, 440, 441 , 442, 443, 444, 446, 489, 508, 509, 510, or 511.

[0143] Since the AON as disclosed herein preferably is used for the targeting of liver cells to treat liver a liver disease caused by a human PNPLA3 gene encoding the I148M mutant form, the AON as disclosed herein is preferably covalently or non-covalently, directly or through a linker, bound to a GalNAc moiety, preferably a tri-antennary GalNAc moiety. Preferred GalNAc moieties are outlined in detail herein.

[0144] The present disclosure also relates to a pharmaceutical composition comprising an AON as disclosed herein, and a pharmaceutically acceptable solvent, excipient, or carrier. Such pharmaceutically acceptable solvents, excipients, and carriers are known to the person skilled in the art.

[0145] The present disclosure also relates to an AON as disclosed herein, for use in the treatment of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis.

[0146] The present disclosure also relates to a use of an AON as disclosed herein, in the manufacture of a medicament for the treatment of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis.

[0147] The present disclosure also relates to a method of editing a human PNPLA3 variant pre-mRNA or mRNA molecule in a liver cell, preferably a hepatocyte, wherein the PNPLA3 variant pre-mRNA or mRNA is encoded by a PNPLA3 gene comprising the rs738409 SNP, the method comprising contacting the PNPLA3 pre-mRNA or mRNA molecule with an AON capable of triggering an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a functionality that is equal or similar to the wildtype functionality, and wherein the AON is as disclosed herein.

[0148] The present disclosure also relates to a method of treating, ameliorating, or slowing down the progression of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis, in a human subject in need thereof, the method comprising administering to said subject an AON as disclosed herein, thereby contacting a PNPLA3 pre-mRNA or mRNA molecule encoded by the PNPLA3 gene comprising the rs738409 SNP in a cell of the subject, thereby effecting an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a functionality that is equal or similar to the wildtype functionality, thereby treating the subject.

[0149] The present disclosure also relates to an in vitro, ex vivo, or in vivo method for the deamination of a target adenosine in a PNPLA3 pre-mRNA or mRNA molecule encoded by a human PNPLA3 gene comprising the rs738409 SNP, in a liver cell, preferably a hepatocyte, the method comprising the steps of: i) providing the cell with an AON as disclosed herein; ii) allowing uptake by the cell of the AON; iii) allowing annealing of the AON to the PNPLA3 pre- mRNA or mRNA molecule; iv) allowing an endogenous ADAR enzyme to deaminate the target adenosine in the PNPLA3 pre-mRNA or mRNA molecule to an inosine; and optionally v) using a functional read-out to identify the presence of the inosine in the PNPLA3 pre-mRNA or mRNA molecule.

[0150] In all aspect of the present disclosure, the deamination of the adenosine changes the amino acid from a methionine (M) to a valine (V) at position 148 in the human PNPLA3 protein. Deamination by ADAR can take place in the nucleus as well as in the cytoplasm. Preferably, the orphan nucleotide is a deoxynucleotide comprising a cytosine, a cytosine analog, an uracil, or an iso-uracil. In one aspect, the AON comprises one or more modifications in the linkage moiety, which is each independently selected from the group consisting of: PS, phosphonoacetate, phosphorodithioate, MP, sulfonylphosphoramidate, PNdmi, and a linkage according to formula (I) below, more preferably PNms. In one aspect, the AON comprises one or more nucleotides comprising a mono- or di-substitution at the 2', 3' and / or 5' position of the ribose, each independently selected from the group consisting of: -OH; -F; substituted or unsubstituted, linear or branched lower (C1-C10) alkyl, alkenyl, alkynyl, alkaryl, allyl, or aralkyl, that may be interrupted by one or more heteroatoms; -O-, S-, or N-alkyl; -O-, S-, or N-alkenyl; -O-, S-, or N-alkynyl; -O-, S-, or N-allyl; -O-alkyl-O-alkyl; -methoxy; -aminopropoxy; - methoxyethoxy; -dimethylamino oxyethoxy; and -dimethylaminoethoxyethoxy.

[0151] In one aspect, the AON is covalently or non-covalently, directly or through a linker, bound to a triterpene glycoside, such as AG1856. As disclosed in Inti. Patent Application Publication No. WO2024 / 153801 , it is very efficient to increase RNA editing, when an AON is connected 1 :1 with a saponin, for instance when the saponin is AG1856. Hence, to increase the endosomal release (intracellularly) of the AON and make it available for RNA target hybridization, the AON can be attached / conjugated (non-covalently, but preferably covalently) to a saponin before administration to the cell or the subject to be treated. In another aspect, the AON is bound to a triterpene glycoside, such as AG1856, as well as to a GalNAc moiety, either directly or indirectly, via linkers.

[0152] The present disclosure also relates to a vector, preferably a viral vector, more preferably an adeno-associated virus (AAV) vector, comprising a nucleic acid molecule encoding an AON as disclosed herein. When an encoding vector is applied, and the AON is encoded by the viral vector (genome), the orphan nucleotide is a cytidine. Furthermore, since the AON is transcribed from the viral vector (genome), no chemical modifications are present in the backbone, linkages, and sugar ribose.

[0153] The present disclosure also relates to a nanoparticle delivery vehicle formulation comprising an AON as disclosed herein. In a preferred aspect, the nanoparticle delivery vehicle is a lipid nanoparticle (LNP). LN P’s that can be used in the context of the AON of the present disclosure are those that have been used in the art for the delivery of small and large RNA molecules, such as those applied in the delivery of mRNA-based vaccines like those against Covid-19 coronaviruses. LNP’s that have been applied for the delivery of other types of RNA, such as siRNA can also be applied for the delivery of the AON as disclosed herein. If an LNP is applied or any other similar type of carrier, the AON is still considered naked because it is not transcribed from an encoding polynucleotide (such as in the case of a plasmid or a vector, in which the AON is not regarded as ‘naked’). So, even though a chemically modified AON is encapsulated by a carrier, preferably an LNP, it is still seen as naked, as it has been manufactured as such in a laboratory setting and encapsulated thereafter in the carrier using methods known to the person skilled in the art.

[0154] The present disclosure also relates to a pharmaceutical composition comprising an AON, a vector, or a nanoparticle delivery vehicle formulation as disclosed herein, and a pharmaceutically acceptable carrier. In one aspect, an AON as disclosed herein is in a naked form. In one aspect, an AON as disclosed herein is in a circular format. In one aspect, an AON as disclosed herein is not in a naked form but is expressed from the genome of a viral vector. In one aspect, an AON as disclosed herein is not in a naked form but is expressed from an expression vector such as a plasmid. In one aspect, when the AON as disclosed herein is not in a naked form, the AON is 15 to 60 nucleotides in length as indicated above, or in another embodiment, from 61 to 300 nucleotides in length. Although it is preferred to use ‘naked’ AONs that have chemical modifications as outlined herein, AONs that are delivered through other means, for instance through AAV vector expression, or editing molecules that are circular, or have hairpin structures (recruiting portions, e.g., as disclosed in WO2016 / 097212, WO2017 / 050306, W02020 / 001793, WO2017 / 010556, W02020 / 246560, and WO2022 / 078995) are also encompassed by the present disclosure because these can also be applied to edit adenosines in the target PNPLA3 I148M variant RNA nucleic acid molecule to generate a PNPLA3 protein with a reverted wild-type functionality. The person skilled in the art understands that when a delivery moiety, or attachment to the AON is used (such a GalNAc moiety to target hepatocytes in the liver) that the AON is still seen as naked as well, also when a GalNAc-AON is encapsulated in a delivery vehicle such as an LNP. The present disclosure also relates to an AON as disclosed herein, for use in the treatment of a disease caused by the I148M PNPLA3 variant in the liver, such as NAFLD, NASH, fibrosis, and cirrhosis. The present disclosure also relates to a use of an AON, as disclosed herein, in the manufacture of a medicament for the treatment of a disease caused by the I148M PNPLA3 variant in the liver, such as NAFLD, NASH, fibrosis, and cirrhosis. The present disclosure also relates to a method of editing a human PNPLA3 I148M variant pre-mRNA or mRNA molecule in a liver cell, preferably a hepatocyte, the method comprising contacting the PNPLA3 pre-mRNA or mRNA molecule with an AON capable of triggering an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 I148M variant pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a reverted wild-type functionality in which the methionine at position 148 is changed to a valine residue, and wherein the AON is as disclosed herein. The present disclosure also relates to a method of treating, ameliorating, or slowing down the progression of a disease caused by the I148M PNPLA3 variant in the liver, such as NAFLD, NASH, fibrosis, and cirrhosis, in a human subject in need thereof, the method comprising administering to said subject an AON, a vector, or a nanoparticle delivery vehicle formulation as disclosed herein, thereby allowing the AON to hybridize to a complementary part of a region comprising a target adenosine in a PNPLA3 1148M variant pre-mRNA or mRNA molecule in a cell of the subject, thereby effecting an ADAR-mediated adenosine to inosine deamination of the target adenosine, thereby changing the PNPLA3 pre-mRNA or mRNA molecule such that it encodes a PNPLA3 protein with a reverted wild-type functionality, thereby treating the subject.

[0155] In one aspect, a method as disclosed herein comprises the step of administering a triterpene glycoside before, after or simultaneously with administering the AON, wherein in a preferred aspect, the triterpene glycoside is AG1856. However, in a preferred aspect the triterpene glycoside (or ‘saponin’ as it is often referred to) is physically bound, or conjugated, to the AON. In yet another embodiment, the AON is bound to a GalNAc moiety and is administered separately from the saponin moiety, which may also be conjugated to a GalNAc moiety for liver cell targeting purposes. When two such formulations: (i) AON-GalNAc and (ii) saponin-GalNAc, are used, they may be encompassed within a kit-of-parts. The formulations in such kit-of-parts (that may further comprise a guide on how to use the formulations) may be combined to provide a single formulation for a single administration or may be administered separately, consecutively or separate within a certain time span, which time span may be determined typically during clinal trials.

[0156] In one aspect, an AON as disclosed herein comprises a linkage moiety with the structure according to formula (I) wherein: X = O or S; and R = an aryl, a substituted aryl, a heterocycle, a substituted heterocycle, an aromatic heterocycle, a substituted aromatic heterocycle, a Ci-Ce alkoxy, a substituted Ci-Ce alkoxy, a C1-C20 alkyl, a substituted C1-C20 alkyl, a Ci-Ce alkenyl, a Ci-Ce substituted alkenyl, a Ci-Ce alkynyl, a substituted Ci-Ce alkynyl, or a conjugate group. In a preferred embodiment, X = O and R = methyl and the linkage modification is referred to as mesyl phosphoramidate, MsPA or “PNms”. In one embodiment, a PNms linkage is used instead of the MP and / or PNdmi linkages.

[0157] Chemical modifications

[0158] Various chemistries and modifications are known in the field of oligonucleotides that can be readily used in accordance with the invention. All chemical modifications listed herein that may be used in the AON as disclosed herein may also be used for a sense strand that is complementary to the AON, when the AON and the complementary strand form a HEON complex, such as described in Inti. Patent Application Publication No. W02024 / 084048 and as disclosed above, except that the opposite sense strand does not have an orphan nucleotide. Hence, the modification related to the orphan nucleotide relate only to the AON as disclosed herein, but all other modifications relate to the AON as disclosed herein and any (protecting) sense oligonucleotide that may be used together with the AON in a pharmaceutical product. This includes the use of hydrophobic moieties (such as tocopherol and cholesterol) and cell-specific ligands (such as GalNAc moieties), that have also been described herein, and in detail in Inti. Patent Application Publication No. W02024 / 084048, which may either be bound to the AON or its opposite strand, or both. Preferred GalNAc moieties that can be used in the context of the AONs as disclosed herein are disclosed in WO2022 / 271806.

[0159] The skilled person knows that an oligonucleotide, such as an AON as outlined herein, generally consists of repeating monomers. Such a monomer is most often a nucleotide or a chemically modified nucleotide. The most common naturally occurring nucleotides in RNA are adenosine monophosphate (A), cytidine monophosphate (C), guanosine monophosphate (G), and uridine monophosphate (II). These consist of a pentose sugar, a ribose, a 5’-linked phosphate group which is linked via a phosphate ester, and a T-linked base. The sugar connects the base and the phosphate and is therefore often referred to as the “scaffold” of the nucleotide. A modification in the pentose sugar is therefore often referred to as a ‘scaffold modification’. The original pentose sugar may be replaced in its entirety by another moiety that similarly connects the base and the phosphate. It is therefore understood that while a pentose sugar is often a scaffold, a scaffold is not necessarily a pentose sugar. Examples of scaffold modifications that may be applied in the monomers of the AON as disclosed herein are disclosed in W02020 / 154342, W02020 / 154343, and W02020 / 154344.

[0160] A nucleoside in the AON as disclosed herein may be a natural nucleoside (deoxyribonucleoside or ribonucleoside) or a non-natural nucleoside. It is noted that for RNA editing, in which double-stranded RNA is generally the substrate for enzymes with deamination activity (such as ADARs), ribonucleosides are considered ‘natural’, while deoxyribonucleosides may then be, for the sake of argument, considered as non-natural, or modified, simply because DNA is not present in the RNA-RNA double stranded (natural) substrate configurations. The skilled person appreciates that when the nucleotide has a natural ribose moiety, it may still be non-naturally modified in the base and / or the linkage.

[0161] It is recognized in the art that common limiting factors in oligonucleotide-based therapies are the oligonucleotide’s ability to be taken up by the cell (when delivered per se, or ‘naked’ without applying a delivery vehicle such as a viral vector or plasmid), the biodistribution and the resistance to nuclease-mediated breakdown. The skilled person is aware, and it has been described in detail in the art, that a variety of chemical modifications can assist in overcoming such limitations. Examples of such now commonly used chemical modifications are the 2’-0Me, 2’-F and 2’-MOE modifications of the sugar and the use of PS linkages between nucleosides, as described herein.

[0162] Scaffold modifications (ribose)

[0163] The ribose 2’ groups in all nucleotides of the AON as disclosed herein, except for the ribose sugar moiety of the orphan nucleotide that has certain limitations in respect of compatibility with RNA editing, can be independently selected from 2’-H (i.e., DNA), 2’-OH (i.e., RNA), 2’-0Me, 2’-MOE, 2’-F, or2’-4’-linked (for instance a locked nucleic acid (LNA), see for instance RM 107588 in Fig. 1), or other ribosyl T-substitutions, 2’ substitutions, 3’ substitutions, 4’ substitutions or 5’ substitutions. The orphan nucleotide in the AON that comprises no other chemical modifications to the ribose sugar, the base, or the linkage preferably does not carry a 2’-OMe or 2’-MOE substitution when the nucleobase is a naturally occurring cytosine, but may carry a 2’-F, a 2’,2’-difluoro (diF), or 2’-ara-F (FANA) substitution or may be DNA. W02024 / 013360 discloses the modification of the 2’ position of the ribose sugar moiety of the orphan nucleotide by a 2’,2’-disubstituted substitution such as diF, which is also applicable to what is disclosed here. The 2’-4’ linkage can be selected from many linkers known in the art, such as a methylene linker, amide linker, or constrained ethyl linker (cEt).

[0164] An AON as disclosed herein may comprise one or more nucleotides carrying a 2’-MOE ribose modification. Also, an AON as disclosed herein may comprise one or more nucleotides not carrying a 2’-MOE ribose modification, or wherein the 2’-MOE ribose modifications are at positions that do not prevent the enzyme with adenosine deaminase activity from deaminating the target adenosine. An AON as disclosed herein may comprise a 2’-OMe ribose modification at a position that does not comprise a 2’-MOE ribose modification. An AON as disclosed herein may comprise deoxynucleotides at positions that do not comprise a 2’-MOE or a 2’-OMe ribose modification, or other 2’ ribose substitution. An AON as disclosed herein may comprise one or more nucleotides comprising a 2’ substitution comprising a 2’-MOE, 2’-OMe, 2’-OH, 2’-deoxy, TNA, 2’-fluoro (2’-F), 2’,2’-difluoro (diF) modification, 2’-fluoro-2’-C-methyl modification, or a 2’-4’-linkage (i.e., a bridged nucleic acid such as a locked nucleic acid (LNA or examples mentioned in e.g. WO2018 / 007475)). Other nucleic acid monomers that may be used in an AON as disclosed herein are arabinonucleic acids and 2’-deoxy-2’-fluoroarabinonucleic acid (FANA), for instance for improved affinity purposes. The 2’-4’ linkage can be selected from linkers known in the art, such as a methylene linker or constrained ethyl linker. A wide variety of 2’ modifications that may present in an AON as disclosed herein are known in the art, including but not limited to the modifications outlined in detail in WO2016 / 097212, WO2017 / 220751 , WO2018 / 041973, WO2018 / 134301 , WO2019 / 219581 , WO2019 / 158475, and WO2022 / 099159. In all cases, the modifications should be compatible with RNA editing such that the AON fulfils its role as an oligonucleotide that can form a double stranded complex with the target RNA and by generating this double-stranded nucleic acid complex, recruit a deaminating enzyme, which can subsequently deaminate the target adenosine. Where a monomer in an AON as disclosed herein comprises an unlocked nucleic acid (UNA) ribose modification, that monomer can have a 2’ position comprising the same modifications discussed above, such as a 2’-MOE, a 2’-OMe, a 2’-OH, a 2’-deoxy, a 2’-F, a 2’,2’-diF, a 2’- fluoro-2’-C-methyl, an arabinonucleic acid, a FANA, or a 2’-4’-linkage (i.e., a bridged nucleic acids such as an LNA). In an aspect, the AON as disclosed herein comprises at least one nucleotide comprising a threose nucleic acid (TNA) ribose modification. In one aspect, the AON as disclosed herein comprises at least one nucleotide with a sugar moiety that comprises a 2’-fluoro (2’-F) modification.

[0165] Base modifications

[0166] A base, sometimes called a nucleobase, is generally adenine, cytosine, guanine, thymine or uracil, or a derivative thereof. A nucleobase is defined as a moiety that can bond to another nucleobase through H-bonds, polarized bonds (such as through OF moieties) or aromatic electronic interactions. Cytosine, thymine, and uracil are pyrimidine bases, and are generally linked to the scaffold through their 1 -nitrogen. Adenine and guanine are purine bases and are generally linked to the scaffold through their 9-nitrogen. The terms ‘adenine’, ‘guanine’, ‘cytosine’, ‘thymine’, ‘uracil’ and ‘hypoxanthine’ as used herein refer to the nucleobases as such. The terms ‘adenosine’, ‘guanosine’, ‘cytidine’, ‘thymidine’, ‘uridine’ and ‘inosine’ refer to the nucleobases linked to the (deoxy)ribosyl sugar. The nucleobases in an AON as disclosed herein can be adenine, cytosine, guanine, thymine, or uracil or any other moiety able to interact with another nucleobase through H-bonds, polarized bonds (such as CF) or aromatic electronic interactions. The nucleobases at any position in the AON as disclosed herein can be a modified form of adenine, cytosine, guanine, or uracil, such as hypoxanthine (the nucleobase in inosine), pseudouracil, pseudocytosine, isouracil, N3- glycosylated uracil, 1 -methylpseudouracil, orotic acid, agmatidine, lysidine, 2-thiouracil, 2- thiothymine, 5-substituted pyrimidine (e.g., 5-halouracil, 5-halomethyluracil, 5- trifluoromethyluracil, 5-propynyluracil, 5-propynylcytosine, 5-aminomethyluracil, 5- hydroxymethyluracil, 5-formyluracil, 5-aminomethylcytosine, 5-formylcytosine), 5- hydroxymethylcytosine, 7-deazaguanine, 7-deazaadenine, 7-deaza-2,6-diaminopurine, 8- aza-7-deazaguanine, 8-aza-7-deazaadenine, 8-aza-7-deaza-2,6-diaminopurine, 8-oxo- adenine, 3-deazapurine (such as a 3-deaza-adenosine), pseudoisocytosine, N4- ethylcytosine, N2-cyclopentylguanine, N2-cyclopentyl-2-aminopurine, N2-propyl-2- aminopurine, 2,6-diaminopurine, 2-aminopurine, G-clamp and its derivatives, Super A, Super T, Super G, amino-modified nucleobases or derivatives thereof; and degenerate or universal bases, like 2,6-difluorotoluene, or absent like abasic sites e.g. 1 -deoxyribose, 1 ,2- dideoxyribose, 1-deoxy-2-O-methylribose, azaribose). Modified bases comprise synthetic and natural bases such as inosine, xanthine, hypoxanthine and other -aza, deaza, -hydroxy, -halo, -thio, thiol, -alkyl, -alkenyl, -alkynyl, thioalkyl derivatives of pyrimidine and purine bases that are or will be known in the art. Purine nucleobases and / or pyrimidine nucleobases may be modified to alter their properties, for example by amination or deamination of the heterocyclic rings. The exact chemistry and formats may vary from oligonucleotide construct to oligonucleotide construct and from application to application, and may be worked out in accordance with the wishes and preferences of those of skill in the art.

[0167] A scaffold modification indicates the presence of a modified version of the ribosyl moiety as naturally occurring in RNA {i.e., the pentose moiety), such as bicyclic sugars, tetrahydropyrans, hexoses, morpholinos, 2’-modified sugars, 4’-modified sugar, 5’-modified sugars and 4’-substituted sugars. Examples of suitable modifications include, but are not limited to 2’-O-modified RNA monomers, such as 2’-O-alkyl or 2’-O-(substituted)alkyl such as 2’-0Me, 2’-O-(2-cyanoethyl), 2’-M0E, 2’-O-(2-thiomethyl)ethyl, 2’-O-butyryl, 2’-O-propargyl, 2’-O-allyl, 2’-O-(2-aminopropyl), 2’-O-(2-(dimethylamino)propyl), 2’-O-(2-amino)ethyl, 2’-O-(2- (dimethylamino)ethyl); 2’-deoxy (DNA); 2’-O-(haloalkyl)methyl such as 2’-O-(2- chloroethoxy)methyl (MCEM), 2’-O-(2,2-dichloroethoxy)methyl (DCEM); 2’-O-alkoxycarbonyl such as 2’-O-[2-(methoxycarbonyl)ethyl] (MOCE), 2’-O-[2- / V-methylcarbamoyl)ethyl] (MCE), 2’-O-[2-( / V, / V-dimethylcarbamoyl)ethyl] (DCME); 2’-halo e.g. 2’-F, FANA; 2'-O-[2- (methylamino)-2-oxoethyl] (NMA); a bicyclic or bridged nucleic acid (BNA) scaffold modification such as a conformationally restricted nucleotide (CRN) monomer, a locked nucleic acid (LNA) monomer, a xy / o-LNA monomer, an a-LNA monomer, an a-l-LNA monomer, a p-d-LNA monomer, a 2’-amino-LNA monomer, a 2’-(alkylamino)-LNA monomer, a 2’-(acylamino)-LNA monomer, a 2’- / V-substituted 2’-amino-LNA monomer, a 2’-thio-LNA monomer, a (2’-O,4’-C) constrained ethyl (cEt) BNA monomer, a (2’-O,4’-C) constrained methoxyethyl (cMOE) BNA monomer, a 2’,4’-BNANC(NH) monomer, a 2’,4’-BNANC(NMe) monomer, a 2’,4’-BNANC(NBn) monomer, an ethylene-bridged nucleic acid (ENA) monomer, a carba-LNA (cLNA) monomer, a 3,4-dihydro-2 / 7-pyran nucleic acid (DpNA) monomer, a 2’- C-bridged bicyclic nucleotide (CBBN) monomer, an oxo-CBBN monomer, a heterocyclic- bridged BNA monomer (such as triazolyl or tetrazolyl-linked), an amido-bridged BNA monomer (such as AmNA), an urea-bridged BNA monomer, a sulfonamide-bridged BNA monomer, a bicyclic carbocyclic nucleotide monomer, a TriNA monomer, an a-l-TriNA monomer, a bicyclo DNA (bcDNA) monomer, an F-bcDNA monomer, a tricyclo DNA (tcDNA) monomer, an F-tcDNA monomer, an alpha anomeric bicyclo DNA (abcDNA) monomer, an oxetane nucleotide monomer, a locked PMO monomer derived from 2’-amino LNA, a guanidine-bridged nucleic acid (GuNA) monomer, a spirocyclopropylene-bridged nucleic acid (scpBNA) monomer, and derivatives thereof; cyclohexenyl nucleic acid (CeNA) monomer, altriol nucleic acid (ANA) monomer, hexitol nucleic acid (HNA) monomer, fluorinated HNA (F- HNA) monomer, pyranosyl-RNA (p-RNA) monomer, 3’-deoxypyranosyl DNA (p-DNA), unlocked nucleic acid UNA); an inverted version of any of the monomers above. All these modifications are known to the person skilled in the art.

[0168] The orphan nucleotide

[0169] Mutagenesis studies of human ADAR2 revealed that a single mutation at residue 488 from glutamate to glutamine (E488Q), gave an increase in the rate constant of deamination by 60-fold when compared to the wild-type enzyme (Kuttan and Bass. Proc Natl Acad Sci USA 2012, 109(48): 3295-3304). During the deamination reaction, ADAR flips the edited base out of its RNA duplex, and into the enzyme active site (Matthews et al. Nat Struct Mol Biol. 2016. 23(5):426-433). When ADAR2 edits adenosines in the preferred context (an A:C mismatch) the nucleotide opposite the target adenosine is often referred to as the ‘orphan nucleotide’ (or ‘orphan cytidine’ as the case may be), as indicated above. The crystal structure of ADAR2 E488Q bound to double stranded RNA (dsRNA) revealed that the glutamine (Gin; Q) side chain at position 488 can donate an H-bond to the N3 position of the orphan cytidine, which leads to the increased catalytic rate of ADAR2 E488Q. In the wild-type enzyme, wherein a glutamate (or glutamic acid; Glu; E) is present at position 488 instead of a glutamine (Gin) the amide group of the glutamine is absent and is instead a carboxylic acid. To obtain the same contact of the orphan cytidine with the E488Q mutant would then, for the wild-type situation, require protonation for this contact to occur. To make use of endogenously expressed ADAR2 to correct disease relevant mutations, it is essential to maximize the editing efficiency of the wild type ADAR2 enzyme present in the cell. WO2020 / 252376 discloses the use of AONs with modified RNA bases, especially at the position of the orphan cytidine to mimic the hydrogenbonding pattern observed by the E488Q ADAR2 mutant. By replacing the nucleotide opposite the target adenosine in the AON with cytidine analogs that serve as H-bond donors at N3, it was envisioned that it would be possible to stabilize the same contact that is believed to provide the increase in catalytic rate for the mutant enzyme. Two cytidine analogs were of particular interest: pseudoisocytidine (also referred to as ‘piC’; Lu et al. J Org Chem 2009. 74(21):8021-8030; Burchenal et al. (1976) Cancer Res 36:1520-1523) and Benner’s base Z (also referred to as ‘Zd’; Yang et al. Nucleic Acid Res 2006. 34(21):6095-6101) that were initially selected because they offer hydrogen-bond donation at N3 with minimal perturbation to the shape of the nucleobase. Benner’s base is also referred to as a 6-amino-5-nitro-3-yl- 2(1 H)-pyridone nucleobase. The presence of the cytidine analog in the AON may exist in addition to modifications to the ribose 2’ group. The ribose 2’ groups in the orphan nucleotide can be independently selected from 2’-H (i.e., DNA), 2’-OH (i.e. , RNA), 2’-OMe, 2’-MOE, 2’-F, or 2’-4’-linked (i.e., a bridged nucleic acid such as an LNA), or other 2’ substitutions. The 2’-4’ linkage can be selected from linkers known in the art, such as a methylene linker or constrained ethyl linker.

[0170] The orphan nucleotide in the AON as disclosed herein is preferably a cytidine or analog thereof (such as a nucleotide carrying a Benner’s base) or a uridine or analog thereof (such as iso-uridine). The orphan nucleotide, whether it is a cytidine or analog thereof, or a uridine or analog thereof, preferably comprises a deoxyribose (2’-H; = DNA) but may also comprise a di F modification at the 2’ position of the sugar. In an embodiment at least one and in another embodiment both the neighbouring (directly adjacent) nucleotides flanking the orphan nucleotide do not comprise a 2’-OMe modification. Complete modification wherein all nucleotides of the oligonucleotide hold a 2’-OMe modification (including the orphan nucleotide), with natural bases, results in a non-functional oligonucleotide as far as RNA editing goes (known in the art), presumably because it hinders the ADAR activity at the targeted position. In general, an adenosine in a target RNA can be protected from editing by providing an opposing nucleotide with a 2'-OMe group (at least when there are no other chemical substitutions or modifications within the nucleotide), or by providing a guanine or adenine as opposing base, as these two nucleobases are also able to reduce editing of the opposing adenosine.

[0171] Another preferred cytidine analog is a deoxyribonucleoside comprising a cytosine analog nucleobase according to the structure or any of its tautomeric forms, wherein: R1, R2, R4and / or R5is H; OH; SH; =0; NH2; a halogen; a linear or branched lower (C1-C10) alkyl; or a Ci-Ce cycloalkyl, wherein the alkyl and / or cycloalkyl is optionally interrupted by one or more heteroatoms; and R3is H; OH; SH; =0; NH2; or a halogen. If R1, R2, R3, R4and R5are H, which is a preferred embodiment, the cytosine analog is referred to as a 5-aza- 5,6-dihydro cytosine nucleobase (also known as E base, E nucleobase, or simply as ‘E’), or any of its enol or keto tautomeric forms thereof. Examples of enol tautomers of the E base are enol-amino and enol-imino tautomers. Examples of keto tautomers of the E base are ketoamino and keto-imino tautomers.

[0172] Linkage modifications

[0173] A nucleoside is generally connected to neighboring nucleosides through condensation of its 5’-phosphate moiety to the 3’-hydroxyl moiety of the neighboring nucleotide monomer. Similarly, its 3’-hydroxyl moiety is generally connected to the 5’-phosphate of a neighboring nucleotide monomer. This forms PO bonds. The PO and the scaffold form an alternating copolymer. The bases are grafted on this copolymer, namely to the scaffold moieties. Because of this characteristic, the alternating copolymer formed by linked scaffolds of an oligonucleotide is often called the ‘backbone’ of the oligonucleotide. Because PO bonds connect neighboring monomers together, they are often referred to as ‘backbone linkages’. It is understood that when a phosphate group is modified so that it is instead an analogous moiety such as a PS, such a moiety is still referred to as the backbone linkage of the monomer. This is referred to as a ‘backbone linkage modification’. In general terms, the backbone of an oligonucleotide comprises alternating scaffolds and backbone linkages.

[0174] As outlined in detail herein, naked AONs as disclosed herein comprise at least one, preferably multiple linkage modifications. It is generally preferred that the AON as disclosed herein comprises linkage modifications at most, and potentially all positions if the AON is capable of mediating RNA editing through the deamination enzyme when the AON is bound to the target RNA nucleic acid molecule. A linkage modification can be, but is not limited to, a modified version of the PO present in RNA, such as PS, chirally pure PS, ( ?)-PS, (S)-PS, MP (or MeP), chirally pure MP, (R)-MP, (S)-MP, phosphoryl guanidine (such as PNdmi), chirally pure phosphoryl guanidine, (R)-phosphoryl guanidine, (S)-phosphoryl guanidine, phosphorodithioate (PS2), phosphonacetate (PACE), phosphonoacetamide (PACA), thiophosphonoacetate, thiophosphonoacetamide, methyl phosphorohioate, methyl thiophosphonate, PS prodrug, alkylated PS, H-phosphonate, ethyl phosphate, ethyl PS, boranophosphate, borano PS, metyl boranophosphate, methyl borano PS, methyl boranophosphonate, methyl boranophosphothioate, phosphate, phosphotriester, aminoalkylphosphotriester, and their derivatives. Another modification includes phosphoramidite, phosphoramidate, N3’->P5’ phosphoramidate, phosphorodiamidate, phosphorothiodiamidate, sulfamate, diethylenesulfoxide, amide, sulfonate, siloxane, sulfide, sulfone, formacetyl, alkenyl, methylenehydrazino, sulfonamide, triazole, oxalyl, carbamate, methyleneimino (MM I), and thioacetamide nucleic acid (TANA); and their derivatives. Various salts, mixed salts, deprotonated, protonated, tautomeric, and free acid forms are also included, as well as 3’->3’ and 2’->5’ linkages. An AON as disclosed herein may also comprise one or more linkage modifications according to the structure of formulas (I), (II), (III), IV), or (V).

[0175] In a preferred aspect, the AON as disclosed herein comprises an internucleoside linkage of the structure of formula (I), wherein X = O and R = CH3, which linkage is generally referred to herein as a PNms linkage. In other preferred aspects, R equals one of the following structures (a), (b), (c), (d), (e), (f), (g), (h), or (i):

[0176] The one or more PN linkages as depicted in formula (I), present in an AON as disclosed herein, can be independently of each other of R or SP chirality, or stereo random.

[0177] The one or more PN linkages as depicted in formula (I), in an AON as disclosed herein, can be in tautomeric and / or pH-dependent (de)protonated form, including but not limited to the structures (A), (B), (C), (D), and (E): wherein X and R are as indicated above for formula (I).

[0178] Disclosed herein is also an AON that is able to mediate adenosine deamination by recruitment of a deaminating enzyme in a cell after the AON has formed a double-stranded complex with a region of a target RNA nucleic acid molecule in a cell, wherein the region comprises a target adenosine, wherein the deaminating enzyme can deaminate the target adenosine into an inosine, and wherein the AON comprises a moiety at one and / or both termini with a structure according to formula (II): wherein: X = O or S;

[0179] Y = O' or S' ; and

[0180] R = an aryl, a substituted aryl, a heterocycle, a substituted heterocycle, an aromatic heterocycle, a substituted aromatic heterocycle, a Ci-Ce alkoxy, a substituted Ci-Ce alkoxy, a C1-C20 alkyl, a substituted C1-C20 alkyl, a Ci-Ce alkenyl, a Ci-Ce substituted alkenyl, a Ci-Ce alkynyl, a substituted Ci-Ce alkynyl, or a conjugate group. In a preferred embodiment, X = O and R = methyl. A preferred internucleoside linkage modification that is used in the AON as disclosed herein has the structure of formula (III): which is also referred to as a PNms linkage. PNms linkages and their application in oligonucleotides and as replacement of PS linkages has been described (Chelobanov BP et al. Russ J Bioorganic Chem. 2017. 43(6): 664-668; DOI: 10.1134 / S1068162017060024; Klabenkova K et al. Molecules. 2021. 26(17):5420; Miroshnichenko SK et al. Proc Natl Acad Sci USA. 2019. 116(4):1229-1234), and for instance in oligonucleotides that may provide splice switching (Hammond SM et al. Nucleic Acid Ther. 31 (3): 190-200).

[0181] An AON as disclosed herein may comprise a substitution of one of the non-bridging oxygens in the PO linkage. This modification slightly destabilizes base pairing but adds significant resistance to nuclease degradation. A preferred nucleotide analogue or equivalent comprises PS, phosphonoacetate, phosphorodithioate, phosphotriester, aminoalkylphosphotriester, H-phosphonate, methyl and other alkyl phosphonate including 3'- alkylene phosphonate, 5'-alkylene phosphonate and chiral phosphonate, phosphinate, phosphoramidate including 3'-amino phosphoramidate and aminoalkylphosphoramidate, thionophosphoramidate, thionoalkylphosphonate, thionoalkylphosphotriester, selenophosphate or boranophosphate. Particularly preferred are internucleoside linkages that are modified to contain a PS. Particularly preferred are internucleoside linkages that are modified to contain a PNms. Particularly preferred are internucleoside linkages that are modified to contain a PNdmi. The regular internucleoside linkages between the nucleotides may be altered by mono- or di-thioation of the PO bonds to yield PS esters or phosphorodithioate esters, respectively. Other modifications of the internucleoside linkages are possible, including amidation and peptide linkers. The skilled person can determine for what target RNA nucleic acid molecule the AON comprises a certain linkage modification at each linkage position of the AON as disclosed herein to generate the most effective and stable oligonucleotide compound.

[0182] Many of the non-naturally occurring modifications of the linkage, such as PS, are chiral. This means that there are Rp and Sp configurations, known to the person skilled in the art. In one embodiment, the chirality of the PS linkages is controlled, which means that each of the linkages is either in the Rp or in the Sp configuration, whichever is preferred. The choice of an Rp or Sp configuration at a specified linkage position may depend on the target sequence and the efficiency of binding and induction of causing RNA editing of the target adenosine. However, if such is not specifically desired, a composition may comprise AONs as active compounds with both Rp and Sp configurations ata certain specified linkage position. Mixtures of such AONs are also feasible, wherein certain positions preferably have either one of the configurations, while for other positions such does not matter. In one aspect, the AON as disclosed herein comprises one or more (chirally pure or chirally mixed) PS linkages. In one aspect, the AON as disclosed herein comprises one of more (chirally pure or chirally mixed) phosphoramidate (PN) linkages. In one aspect, the AON as disclosed herein comprises one or more (chirally pure or chirally mixed) PNms linkages. In one aspect, a PNdmi or PNms linkage connects the terminal two nucleotides on either end of the AON. AONs as disclosed herein may also comprise linkage modifications at all positions that are not chirally controlled. The AON as disclosed herein may also comprise one or more naturally occurring internucleoside linkages. The choice and number of modified linkages may depend on the specific target, the sequence, the length, and the stability of the AON observed in a particular cell type of interest, which can be assessed by methods known to the person skilled in the art. In one aspect, at least one, at least two, at least three, or at least four internucleoside linkages between the 5’ and / or the 3’ terminal two, three, four, or five nucleosides respectively of the AON as disclosed herein are modified internucleoside linkages. In one aspect, the AON as disclosed herein comprises at least one MP internucleoside linkage according to the structure of formula (IV):

[0183] As was noted in the art, a preferred position for an MP linkage in an AON is linkage position -2, thereby connecting the nucleoside at position -1 with the nucleoside at position - 2. Inti. Patent Application Publication No. W02024 / 200278 discloses that PNms at this position renders the AON more stable than when a MP linkage is present. Hence, in a preferred embodiment, this -2 position, in an AON as disclosed herein, comprises a linkage modification according to the structure of formula (I), more preferably a linkage modification according to the structure of formula (III), instead of an MP linkage. W02020 / 201406 discloses the use of MP linkage modifications at certain positions surrounding the orphan nucleotide in the first nucleic acid strand. Although the presence of MP linkages is compatible with RNA editing by human ADAR enzymes, introducing MP linkages during the manufacturing of oligonucleotides is also challenging in view of additional manufacturing (purification) steps in the coupling and decoupling process. In one aspect, the AON does not comprise an MP linkage, and the linkage position at -2 is a PNms linkage.

[0184] In one aspect, the AON as disclosed herein comprises at least one PNdmi linkage, preferably linking the most terminal two nucleosides at the 5’ and / or 3’ end of the AON. A PNdmi linkage that can be used in an AON as disclosed herein has the structure of formula (V):

[0185] In one aspect, at either end or both termini of an AON as disclosed herein, inverted deoxyT or dideoxyT nucleotides are incorporated. Other internucleoside linkages that may be used in the AONs as disclosed herein are those that are disclosed in WO2023 / 278589.

[0186] In one aspect, the AON as disclosed herein comprises at least one phosphonoacetate and / or at least one phosphonoacetamide internucleoside linkage.

[0187] Conjugate chemistries

[0188] In one aspect, the AON as disclosed herein, or the sense strand to which it may be annealed before entering a target cell (in an HEON as disclosed herein), is bound to a hydrophobic moiety, such as palmityl or an analog thereof, cholesterol or analog thereof, or tocopherol or analog thereof. It is preferably bound to the 5’ terminus. In case a hydrophobic moiety is bound to the 5’ terminus as well as to the 3’ terminus, such hydrophobic moieties may the same or different. The hydrophobic moiety bound to the oligonucleotide may be bound directly or indirectly mediated by another substance. When the hydrophobic moiety is bound directly, it is sufficient if the moiety is bound via a covalent bond, an ionic bond, a hydrogen bond, or the like. When the hydrophobic moiety is bound indirectly, it may be bound via a linking group (a linker). The linker may be a cleavable or an uncleavable linker. A cleavable linker refers to a linker that can be cleaved under physiological conditions, for example, in a cell or an animal body (e.g., a human body). A cleavable linker is selectively cleaved by an endogenous enzyme such as a nuclease, or by physiological circumstances specific to parts of the body or cell, such as pH or reducing environment (such as glutathione concentrations). Examples of a cleavable linker comprise, but is not limited to, an amide, an ester, one or both esters of a PO, a phosphoester, a carbamate, and a disulfide bond, as well as a natural DNA linker. Cleavable linkers also include self-immolative linkers. An uncleavable linker refers to a linker that is not cleaved under physiological conditions, or very slowly compared to a cleavable linker, for example, in a PS linkage, modified or unmodified deoxyribonucleosides linked by a PS linkage, a spacer connected through a PS bond and a linker consisting of modified or unmodified ribonucleosides. There is no restriction on the chain length, when a linker is a nucleic acid such as DNA, or an oligonucleotide. However, it may be usually from 2 to 20 bases in length, from 3 to 10 bases in length, or from 4 to 6 bases in length. There is no restriction on the length or composition of a spacer that is connects the ligand and the oligonucleotide, and may include for example ethylene glycol, triethylene glycol (TEG), HEG, alkyl chains, propyl, 6-aminohexyl, or dodecyl. In one embodiment, a GalNAc moiety is bound to the AON as disclosed herein via a TEG linker. One or more other types of molecules may be bound to the AON through one or more linkers, including peptides, sugars, vitamins, polymers, aptamers, (fragments of) antibodies, small molecules, and the like.

[0189] General

[0190] In addition to the specific preferred chemical modifications at certain positions in compounds as disclosed herein, AONs as disclosed herein may comprise one or more (additional) modifications to the nucleobase, scaffold and / or backbone linkage, which may or may not be present in the same monomer, for instance at the 3’ and / or 5’ position. In one aspect, the AON as disclosed herein comprises at least one internucleoside linkage according to the structure of formula (I), and / or the AON further comprises at least one nucleotide with a sugar moiety that comprises a 2’-OMe modification, and / or the AON comprises at least one nucleotide with a sugar moiety that comprises a 2’-MOE modification, and / or the AON comprises at least one nucleotide with a sugar moiety that comprises a 2’-F modification, and / or the AON comprises an orphan nucleotide that carries a 2’-H in the sugar moiety and is therefore referred to as a DNA nucleotide, even though additional modifications may exist in its base and / or linkage to its neighbouring nucleosides. In one aspect, the orphan nucleotide carries a 2’-F in the sugar moiety. In one aspect, the orphan nucleotide carries a diF substitution in the sugar moiety. In one aspect, the orphan nucleotide carries a 2’-F and a 2’- C-methyl in the sugar moiety. In one aspect, the orphan nucleotide comprises a 2’-F in the arabinose configuration (FANA) in the sugar moiety.

[0191] In one aspect, the AON is an antisense oligonucleotide that can form a double stranded nucleic acid complex with a target RNA molecule, wherein the double stranded nucleic acid complex can recruit an adenosine deaminating enzyme for deamination of a target adenosine in the target RNA molecule, wherein the nucleotide in the AON that is opposite the target adenosine is the orphan nucleotide, and wherein the orphan nucleotide has the structure of formula (VI): wherein: X is O, NH, OCH2, CH2, Se, or S; B is a nitrogenous base selected from the group consisting of: cytosine, uracil, isouracil, N3-glycosylated uracil, pseudoisocytosine, 8-oxo- adenine, and 6-amino-5-nitro-3-yl-2(1 H)-pyridone; R1 and R2 are both selected, independently, from H, OH, F or CH3; R3 is the part of the AON that is 5’ of the orphan nucleotide, consisting of 7 to 30 nucleotides; and R4 is the part of the AON that is 3’ of the orphan nucleotide, consisting of 4 to 25 nucleotides. The nucleotide 3’ and / or 5’ from the orphan nucleotide may be DNA, more preferably at least the nucleotide at the 3’ location (position -1).

[0192] Other chemical modifications of the AON as disclosed herein include the substitution of one or more than one of any of the hydrogen atoms with deuterium or tritium, examples of which can be found in e.g., WO2014 / 022566 or WO2015 / 011694. Again, in all cases, the modifications should be compatible with editing such that the AON fulfils its role as an oligonucleotide that can, after binding to its target sequence, recruit an adenosine deaminase enzyme because of the double-stranded nucleic acid entity that arises. In all aspects of the disclosure, the enzyme with adenosine deaminase activity is preferably ADAR1 , ADAR2, or ADAT.

[0193] In one embodiment, the AON as disclosed herein comprises a delivery moiety. Preferably, for delivery to liver cells, preferably hepatocytes, the delivery moiety is an / V- Acetylgalactosamine (GalNAc) moiety. Preferred GalNAc moieties are tri-antennary GalNAc moieties as shown in Formula VII, VIII, IX(a) and (b), and XII (below). Preferred mono- antennary GalNAc moieties are as shown in Formula X and XI (below).

[0194] Formula VII, includes connection point E (see WO2022 / 271806):

[0195]

[0196] Formula VIII, includes a linker and a connection point F which conjugates to the AON (see WO2014 / 179620):

[0197] Formula IX(a) and IX(b) include a linker and a connection point F which conjugates to the AON (see W02009 / 073809):

[0198] Formula IX(a);

[0199] Formula IX(b)

[0200] Formula X:

[0201] Formula XII (see WO2011 / 104169), wherein the squiggly line indicates a connection point, optionally via linker and / or a spacer, to the AON:

[0202]

[0203] When the GalNAc moiety of Formula VII is applied, the GalNAc moiety is preferably conjugated to the AON at its 3’ terminus, via connection point E of Formula VII, optionally via a linker and / or a spacer, for example as depicted in Formula XIII (below).

[0204] Formula XIII:

[0205] When the GalNAc moiety of Formula VIII is applied, the GalNAc moiety is preferably conjugated to the AON at its 5’ terminus, via connection point F of Formula VIII, optionally via a linker and / or a spacer.

[0206] AONs as disclosed herein preferably do not include a 5’-terminal O6-benzylguanosine or a 5’-terminal amino modification and preferably are not covalently linked to a SNAP-tag domain (an engineered O6-alkylguanosine-DNA-alkyl transferase). An AON as disclosed herein preferably does not comprise a boxB RNA hairpin sequence. In one aspect, an AON as disclosed herein comprises 0, 1 , 2 or 3 wobble base pairs with the target sequence, and / or 0, 1 , 2, 3, 4, 5, 6, 7, or 8 mismatching base pairs with the target RNA sequence. No mismatch exists when the orphan nucleotide is uridine, which may be defined differently when the orphan nucleotide is a uridine analog or derivative. One alternative for uridine is positioning an isouridine opposite the target adenosine, which likely does not pair like G pairs with II. Preferably, the target adenosine in the target sequence forms a mismatch base pair with the nucleoside in the AON that is directly opposite the target adenosine.

[0207] As outlined above, an AON as disclosed herein makes use of specific nucleotide modifications at predefined spots to ensure stability as well as proper ADAR binding and activity. These changes may vary and may include modifications in the backbone of the AON, in the sugar moiety of the nucleotides as well as in the nucleobases or the phosphodiester linkages, as outlined in detail herein. They may also be variably distributed throughout the sequence of the AON. Specific modifications may be needed to support interactions of different amino acid residues within the RNA-binding domains of ADAR enzymes, as well as those in the deaminase domain. For example, PS linkages between nucleotides or 2’-0Me or 2’-M0E modifications may be tolerated in some parts of the AON, while in other parts they should be avoided so as not to disrupt crucial interactions of the enzyme with the phosphate and 2’-OH groups. Specific nucleotide modifications may also be necessary to enhance the editing activity on substrate RNAs where the target sequence is not optimal for ADAR editing. Previous work has established that certain sequence contexts are more amenable to editing. For example, a target sequence 5’-UAG-3’ (with the target A in the middle) contains the preferred nearest-neighbor nucleotides for ADAR2, whereas a 5’-CAA-3’ target sequence is disfavored (Schneider et al. Nucleic Acids Res 2014, 42(10):e87). The structural analysis of ADAR2 deaminase domain hints at the possibility of enhancing editing by careful selection of the nucleotides that are opposite to the target trinucleotide. For example, the 5’-CAA-3’ target sequence, paired to a 3’-GCU-5’ sequence on the opposing strand (with the A-C mismatch formed in the middle), is disfavored because the guanosine base sterically clashes with an amino acid side chain of ADAR2. The guanosine opposite the C in such circumstances (and as outlined herein) is preferably replaced by an inosine (hence, at the -1 position within the AON), preferably a deoxyinosine (Id, see Figs. 1 and 4).

[0208] The AON as disclosed herein, in contrast to what has been described for siRNA, or gapmers and their relation towards RNase breakdown and the use of such gapmers in doublestranded complexes (see for instance EP 3954395 A1), does not comprise a stretch of DNA nucleotides which would make a target sequence (or a sense nucleic acid strand) a target for RNase-mediated breakdown. It is not desired that the target transcript molecule is degraded through the binding of the AON to the transcript molecule. In one embodiment, the AON does not comprise four or more consecutive DNA nucleotides anywhere within its sequence. In an embodiment, the AON is composed of as much (chemically) modified nucleotides as possible to enhance the resistance towards RNase-mediated breakdown, while at the same time being as efficient as possible in producing an RNA editing effect. This means that the orphan nucleotide and several other nucleotides within the AON may be DNA, but also that there is no stretch of four or more consecutive DNA nucleotides within the AON. Hence, the AON as disclosed herein is not a gapmer. A gapmer reduces the expression of a target transcript but does not produce RNA editing of a specified adenosine within the target transcript. A gapmer is in principle a single-stranded nucleic acid consisting of a central region (DNA gap region with at least four consecutive deoxyribonucleotides) and wing regions positioned directly at the 5’ end (5’ wing region) and the 3’ end (3’ wing region) thereof. In contrast, the AON as disclosed herein may be any oligonucleotide that produces an RNA editing effect in which a target adenosine in a target RNA molecule is deaminated to an inosine and accordingly is resistant to RNase-mediated breakdown as much as possible to yield this effect and to allow the mRNA transcript being translated into a protein.

[0209] The AONs as disclosed herein may also be administered in the context of aids that will increase the entry of the AON into the target cell and / or its endosomal escape as soon as it is in the cell. Moieties that can be applied for such applications are for example a set of chemical compounds (generally purified from nature) referred to as “saponins” or “triterpene glycosides”, as outlined infra. A saponin that can be used in the methods as disclosed herein is AG1856, disclosed in Inti. Patent Application Publication No. WO2021 / 122998 and further described for use with RNA editing producing oligonucleotides in Inti. Patent Application Publication No. WO2024 / 153801.

[0210] Disclosed herein is also a pharmaceutical composition comprising the AON as disclosed herein, and further comprising a pharmaceutically acceptable carrier, solvent, diluent, and / or other additive (such as a saponin or triterpene glycoside like AG1856 (as discussed above), which in fact may be conjugated to the AON, and may also be administered separately from the AON) and may be dissolved in a pharmaceutically acceptable organic solvent, or the like. Dosage forms in which the AON or the pharmaceutical composition are administered may depend on the disorder to be treated and the tissue that needs to be targeted and can be selected according to common procedures in the art. The pharmaceutical compositions may be administered by a single-dose administration or by multiple dose administration. It may be administered daily or at appropriate time intervals, which may be determined using common general knowledge in the field and may be adjusted based on the disorder and the efficacy of the active ingredient.

[0211] Although in a preferred embodiment, the AON as disclosed herein is a single-stranded oligonucleotide comprising an orphan nucleotide opposite the target adenosine, wherein the orphan nucleotide is chemically modified as disclosed herein, and wherein the remainder of the oligonucleotide is chemically modified to prevent it from nuclease breakdown also as disclosed herein, in another embodiment, disclosed is any kind of oligonucleotide or heteroduplex oligonucleotide complex, that may or may not be bound to hairpin structures (internally or at the terminal end(s)), that may be bound to ADAR or catalytic domains thereof, or wherein the oligonucleotide is in a circular format. In a preferred aspect, the AON as disclosed herein is a ‘naked’ oligonucleotide, comprising a variety of chemical modifications in the ribose sugar and / or the base of one or more of the nucleotides within the sequence, that preferably comprises at least one linkage according to the structure of formula (I) as disclosed herein, that can hybridize to the target transcript or a part thereof that includes the target adenosine, and can recruit endogenous (naturally present) ADAR in the target cell for the deamination of the target adenosine. In another aspect, the AON as disclosed herein, that is delivered in a ‘naked’ form, does not comprise a stem-loop structure for recruitment of the deaminating enzyme, which allows for a shorter AON and improved cellular delivery and trafficking.

[0212] It is known in the art that RNA editing entities (such as human ADAR enzymes) edit dsRNA structures with varying specificity, depending on several factors. One important factor is the degree of complementarity of the two strands making up the dsRNA sequence. Perfect complementarity of the two strands usually causes the catalytic domain of human ADAR to deaminate adenosines in a non-discriminative manner, reacting with any adenosine it encounters. The specificity of hADARI and 2 can be increased by introducing chemical modifications and / or ensuring several mismatches in the dsRNA, which presumably helps to position the dsRNA binding domains in a way that has not been clearly defined yet. Additionally, the deamination reaction itself can be enhanced by providing an oligonucleotide that comprises a mismatch opposite the adenosine to be edited. Following the instructions in the present application, those of skill in the art will be capable of designing the complementary portion of the oligonucleotide according to their needs.

[0213] It will be understood by a person having ordinary skill in the art that the extent to which the editing enzymes inside the cell are redirected to other target sites may be regulated by varying the affinity of the first nucleic acid strand for the recognition domain of the editing enzyme. The exact modification may be determined through some trial and error and / or through computational methods based on structural interactions between the AON and the recognition domain of the editing enzyme. In addition, or alternatively, the degree of recruiting and redirecting the editing enzyme resident in the cell may be regulated by the dosing and the dosing regimen of the AON. This is something to be determined by the experimenter in vitro) or the clinician, usually in phase I and / or II clinical trials.

[0214] Disclosed herein is the site-specific editing of target adenosines in RNA sequences in eukaryotic, preferably metazoan, more preferably mammalian, more preferably human cells, more preferably human liver cells, and most preferably human hepatocytes. The target cell can be located in vitro, ex vivo or in vivo. One advantage of the AON as disclosed herein is that it can be used with cells in situ in a living organism, but it can also be used with cells in culture. In some embodiments cells are treated ex vivo and are then introduced into a living organism (e.g., re-introduced into an organism from whom they were originally derived). The AON as disclosed herein can also be used to edit target RNA sequences in cells from a transplant or within a so-called organoid, e.g., a liver tissue organoid. Organoids can be thought of as three-dimensional in v / fro-derived tissues but are driven using specific conditions to generate individual, isolated tissues. In a therapeutic setting they are useful because they can be derived in vitro from a patient’s cells, and the organoids can then be re-introduced to the patient as autologous material which is less likely to be rejected than a normal transplant.

[0215] Without wishing to be bound by theory, the RNA editing through human ADAR2 for example is thought to take place on primary transcripts in the nucleus, during transcription or splicing, or in the cytoplasm, where e.g., mature mRNA, miRNA or ncRNA can be edited. Generally spoken, RNA editing may be used to create RNA sequences with different properties. Such properties may be coding properties (creating proteins with different sequences or length, leading to altered protein properties or functions), or binding properties (causing inhibition or over-expression of the RNA itself or a target or binding partner; entire expression pathways may be altered by recoding miRNAs or their cognate sequences on target RNAs). Protein function or localization may be changed at will, by functional domains or recognition motifs, including but not limited to signal sequences, targeting or localization signals, recognition sites for proteolytic cleavage or co- or post-translational modification, catalytic sites of enzymes, binding sites for binding partners, signals for degradation or activation and so on. These and other forms of RNA and protein “engineering”, whether to prevent, delay or treat disease or for any other purpose, in medicine or biotechnology, as diagnostic, prophylactic, therapeutic, research tool or otherwise, are encompassed by the present disclosure.

[0216] The amount of AON to be administered, the dosage and the dosing regimen can vary from cell type to cell type, the disease to be treated, the target population, the mode of administration e.g., systemic versus local), the severity of disease and the acceptable level of side activity, but these can and should be assessed by trial and error during in vitro research, in pre-clinical and clinical trials. The trials are particularly straightforward when the modified sequence leads to an easily detected phenotypic change, or a change in (the level of, or activity of) a specified biomarker (such as plasma levels of bile acids for example). It is possible that higher doses of AONs could compete for binding to an ADAR enzyme within a cell, thereby depleting the amount of the enzyme, which is free to take part in RNA editing, but routine dosing trials will reveal any such effects for a given AON and a given target. One suitable trial technique involves delivering the AON to cell lines, or a test organism and then taking biopsy samples at various time points thereafter. The sequence of the target RNA can be assessed in the biopsy sample and the proportion of cells having the modification can easily be followed. As mentioned above, plasma level concentrations of bile acids in a sample from a treated subject is a proper biomarker for assessing the function of certain proteins in the subject, before and after treatment, or with or without treating the subject with an AON as disclosed herein. After this trial has been performed once then the knowledge can be retained, and future delivery can be performed without needing to take biopsy samples. A method as disclosed herein can thus include a step of identifying the presence of the desired change in the cell’s target RNA sequence, thereby verifying that the target RNA sequence has been modified. This step will typically involve sequencing of the relevant part of the target RNA, or a cDNA copy thereof (or a cDNA copy of a splicing product thereof, in case the target RNA is a pre-mRNA), as discussed above, and the sequence change can thus be easily verified. Alternatively, as indicated above, the change may be assessed on the function of the protein before, during, and / or after treatment or assessing any other potential marker, which measurements are preferably performed in vitro on samples obtained from the treated subject.

[0217] After RNA editing has occurred in a cell, the modified RNA can become diluted over time, for example due to cell division, limited half-life of the edited RNAs, etc. Thus, in practical therapeutic terms a method as disclosed herein may involve repeated delivery of an AON until enough target RNAs have been modified to provide a tangible benefit to the patient and / or to maintain the benefits over time.

[0218] AONs as disclosed herein are particularly suitable for therapeutic use, and so disclosed is also a pharmaceutical composition comprising an AON as disclosed herein and a pharmaceutically acceptable carrier, solvent, or diluent. In some embodiments the pharmaceutically acceptable carrier can simply be a saline solution. This can usefully be isotonic or hypotonic, particularly for pulmonary delivery. The AON as disclosed herein is suitably administrated in aqueous solution, e.g. saline, or in suspension, optionally comprising additives, excipients and other ingredients, compatible with pharmaceutical use, at concentrations ranging from 1 ng / ml to 1 g / ml, preferably from 10 ng / ml to 500 mg / ml, more preferably from 100 ng / ml to 100 mg / ml. Dosage may suitably range from between about 1 pg / kg to about 100 mg / kg, preferably from about 10 pg / kg to about 10 mg / kg, more preferably from about 100 pg / kg to about 1 mg / kg. Administration may be by inhalation (e.g., through nebulization), intranasally, orally, by injection or infusion, intravenously, subcutaneously, intradermally, intramuscularly, intra-tracheally, intra-peritoneally, intrarectally, intrathecally, intra-cisterna magna, parenterally, and the like. Administration may be in solid form, in the form of a powder, a pill, a gel, a solution, a slow-release formulation, or in any other form compatible with pharmaceutical use in humans. In one embodiment, depending on the ultimate deamination effect of A to I conversion, the identification step of whether the editing has taken place, comprises the following steps: sequencing the target RNA; assessing the presence or absence of a functional protein (or at least the presence of a PNPLA3 protein that has a functionality that is close or identical to the wild-type protein); assessing whether splicing of the pre-mRNA was altered by the deamination; or using a functional read-out, because the target RNA after the deamination should encode a protein with a reverted, or regained, functionality that is equal or similar to the wild-type functionality. The identification of the deamination into inosine may be a functional read-out using a suitable biomarker. A functional assessment will generally be according to methods known to the skilled person. A suitable manner to identify the presence of an inosine after deamination of the target adenosine is of course dPCR or even sequencing, using methods that are well-known to the person skilled in the art. However, the person skilled in the art of liver disease will preferably apply tests to monitor certain biomarkers related to liver function(s) and in the present case the level of liver fattening.

[0219] In one embodiment, a method as disclosed herein comprises the steps of administering to the subject an AON or pharmaceutical composition as disclosed herein, allowing the formation of a double stranded nucleic acid complex of the AON with its specific complementary target nucleic acid molecule in a cell in the subject; allowing the engagement of an endogenous present adenosine deaminating enzyme, such as ADAR 1 or ADAR2; and allowing the enzyme to deaminate the target adenosine in the target nucleic target molecule to an inosine, thereby alleviating, treating, ameliorating, or slowing down progression of the liver disease, preferably NAFLD caused by the PNPLA3 I148M variant.

[0220] RNA editing molecules present in the cell will usually be proteinaceous in nature, such as the ADAR enzymes found in metazoans, including mammals. The ones of most interest are the human ADARs, hADARI and hADAR2, including any isoforms thereof. RNA editing enzymes known in the art, for which oligonucleotide constructs as disclosed herein may conveniently be designed, include the adenosine deaminases acting on RNA (ADARs), such as hADARI and hADAR2 in humans or human cells and cytidine deaminases. It is known that hADARI exists in two isoforms; a long 150 kDa interferon inducible version and a shorter, 110 kDa version, that is produced through alternative splicing from a common pre-mRNA. Consequently, the level of the 150 kDa isoform available in the cell may be influenced by interferon, particularly interferon-gamma (IFN-y). hADARI is also inducible by TNF-a. This provides an opportunity to develop combination therapy, whereby IFN-y or TNF-a and AONs as disclosed herein are administered to a patient either as a combination product, or as separate products, either simultaneously or subsequently, in any order. Certain disease conditions may already coincide with increased IFN-y or TNF-a levels in certain tissues of a patient, creating further opportunities to make editing more specific for diseased tissues. It will be understood by a person having ordinary skill in the art that the extent to which the editing entities inside the cell are redirected to other target sites may be regulated by varying the affinity of the first nucleic acid strand for the recognition domain of the editing molecule.

[0221] An AON as disclosed herein can utilise endogenous cellular pathways and naturally available ADAR enzymes to specifically edit a target adenosine in the target RNA sequence. An AON as disclosed herein is capable of recruiting ADAR and complex with it and then facilitates the deamination of a (single) specific target adenosine nucleotide in a target RNA sequence to which it is bound. Ideally, only one adenosine is deaminated. An AON as disclosed herein, when complexed to ADAR, preferably brings about the deamination of a single target adenosine.

[0222] An AON as disclosed herein, especially when it is in a naked form, is normally longer than 10 nucleotides, preferably more than 11 , 12, 13, 14, 15, 16, still more preferably more than 17 nucleotides. In one aspect the AON as disclosed herein is longer than 20 nucleotides. The AON as disclosed herein is preferably shorter than 100 nucleotides, still more preferably shorter than 60 nucleotides, still more preferably shorter than 50 nucleotides. In a preferred aspect, the AON as disclosed herein comprises 18 to 70 nucleotides, more preferably comprises 18 to 60 nucleotides, and even more preferably comprises 18 to 50 nucleotides. Hence, in a particularly preferred aspect, the AON as disclosed herein comprises 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 , 32, 33, 34, 35, 36, 37, 38, 39, 40, 41 , 42, 43, 44, 45, 46, 47, 48, 49, 50, 51 , 52, 53, 54, 55, 56, 57, 58, 59, or 60 nucleotides. In one embodiment, the AON is 27, 28, 29, or 30 nucleotides in length.

[0223] EXAMPLES

[0224] Example 1. RNA editing of the PNPLA3 target transcript using a variety of AONs.

[0225] A set of 50 AONs was designed to target the adenosine in the AUG codon coding for isoleucine at position 148 in the human wildtype PNPLA3 protein, which serves as a model for targeting the adenosine in the AUG codon coding for methionine in the PNPLA3 I148M variant. The design and chemical modifications of these AONs (SEQ ID NO:2 to 51), together with the target sequence in the human wildtype PNPLA3 transcript are provided in Fig. 1.

[0226] These AONs were tested as follows: On day 0, Primary Human Hepatocytes (PHHs; 5.0x104cells / well) were transfected with either 100nM or 200nM AONs, each concentration in triplicates, using Lipofectamine® RNAiMAX Reagent at the same time of seeding, following the protocol of the manufacturer. The plates containing cells, medium and AON were held at 37°C, 5% CO2 for 72 hrs, during which the medium was refreshed 24 hrs after transfection / plating. On day 3 (72 hrs post transfection / plating) the supernatants were discarded, and subsequent analysis was performed as follows. Cells were collected and used for RNA isolation using a RNeasy 96 Kit (Qiagen-74182) according to the manufacturer’s instructions. Extracted RNA was treated with DNasel (ThermoFisher-EN0521) according to manufacturer’s protocol. Samples were incubated at 37°C for 30 min and then 1 pL 50 mM EDTA was added and further incubated at 60°C for 2 min. The total RNAs were then reverse-transcribed using the Maxima Reverse Transcriptase (Thermo-EP0742) kit with oligo-dT primer, random Hexamer Primer, and dNTP Mix (10 mM each). A quantitative PCR was then performed with the Digital PCR System (Bio-Rad, QX200) in 22 pl aliquots of reaction mixtures containing cDNA, appropriate pairs of primers and ddPCR Supermix for Probes (no dUTP) (Bio-Rad-1863024). The primers for this target site in the PNPLA3 (pre-) mRNA (Table 1) were used with a PCR program that was as follows: 10 min at 95°C; 40 x 30 sec at 94°C and 60 sec at 60°C, 10 min at 98°C and a hold step at 4°C.

[0227] Table 1: Primers and probes with their respective sequences for quantitative PCR assays. The SEQ ID NO of each primer or probe is given between brackets following the respective sequence in the middle column. The + indicates a Locked Nucleic Acid (LNA) on the 3' side of the symbol. Then the plate was placed into the droplet QX200 reader to measure the number of positive droplets. The editing percentage was calculated by pooling the three replicates for each transfection for all A and G counts and then scored as follows: score = SUM(G) / (SUM(A+G) * 100

[0228] The editing percentage results with the indicated AONs using the 100nM transfections are shown in Fig. 2. Not all samples provided valid results (invalid data with RM107563 to RM 107566 not shown). However, editing up to almost 40% was obtained, in which RM 107578, RM 107579, RM 107580, RM 107581 , RM 107582, RM 107583, RM 107584, RM 107585, RM 107586, and RM 107587 (SEQ ID NO:41 to 50, respectively) performed best. RM 107588, which contains LNA modifications at certain positions did not cause detectable editing. The editing percentage results with the indicated AONs using the 200nM transfections are shown in Fig. 3. All samples provided valid results. Editing above 45% was obtained, in which again RM 107578, RM 107579, RM 107580, RM 107581 , RM 107582, RM 107583, RM 107584, RM 107585, RM 107586, and RM 107587 (SEQ ID NO:41 to 50, respectively) performed best, and in which RM 107578 (SEQ ID NO:41) outperformed the others, albeit not always significantly better in comparison to the other nine good performers.

[0229] It is noted that the +2 position of the AONs of SEQ ID NO:2 to 25 all contain a 2’-OMe modified guanosine (Gm) at position +2, whereas the AONs of SEQ ID NO:26 to 50 all contain a 2’-F modified guanosine (Gf) at position +2, and it is noted that RM 107563 to RM 107587 also have an ‘asymmetrical’ design, in which the orphan nucleotide is located more towards the 3’ terminus of the AON than towards the 5’ terminus. As can be seen in Fig. 2 and Fig. 3, the AONs of SEQ ID NO:26 to 50, and especially those of SEQ ID NO:41 to 50, outperform the AONs of SEQ I D NO:2 to 25, which is an indication that the nucleotide at position +2 and / or its 2’ sugar ribose modification and / or the symmetry of the AON in respect of the location of the orphan nucleotide plays an important role in the ADAR-mediated deamination of the target adenosine in the target PNPLA3 transcript nucleic acid molecule.

[0230] Strikingly, it should be noted that RM 107578 to RM 107587 all comprise a A:C mismatch at the +3 position and a C:C mismatch at the +4 position with the wildtype target sequence of SEQ ID NO:1 and with the I148M variant target sequence of SEQ ID NO:56 (see underlined nucleotides in Fig. 1), which unexpectedly adds significantly to the editing efficiency brought about by these AONs.

[0231] Example 2. RNA editing of the PNPLA3 target transcript using a variety of AONs with and without mismatches at +3 and +4. Based on the unexpected results with the mismatches on nucleotide positions +3 and +4 in the AONs, see Example 1 , a further set of 36 AONs was designed based on the sequence of RM 107578 that also contain the mismatches on positions +3 and +4 (see SEQ ID NO:67 to 102 in Fig. 4; mismatches underlined), together with 36 AONs as a further set that have the exact same chemistries, respectively, but do not comprise these two mismatches (SEQ ID NO:103 to 138 in Fig. 4). In other words, these 36 + 36 AONs form 36 pairs that can be compared one-to-one in respect of the mismatches at positions +3 and +4.

[0232] In a next experiment, these pairs were compared with each other using the same transfection RNA isolation and ddPCR as discussed in Example 1. The editing percentage results are provided in Fig. 5A and Fig. 5B, wherein for instance RM 120392 (with the two mismatches) is compared to RM120428 (without the two mismatches) and so further. In almost each comparison the AON with the two mismatches outperforms the AON that is fully complementary at +3 and +4. Especially RM 120392, RM 120393, RM 120398, RM 120399, RM120400, RM120404, RM120409, and RM120426 outperformed their comparison partner. This suggests strongly that it is beneficial to use an AON for editing of the PNPLA3 target site as outlined herein, wherein the AON comprises a mismatch at nucleotide position +3 and / or +4 with their respective nucleotide located opposite in the target sequence. It is noted that the AONs of Fig. 4 are complementary at the +2 position with the opposite cytidine in the target sequence when the human PNPLA3 target sequence is wildtype (AUG, encoding isoleucine), but it is to be understood that the AONs of Fig. 4 can also be used for the targeting of the I148M variant (pre-) mRNA in which the codon for methionine at position 148 is AUG. If such is the case (and it is in the diseased variant), the guanosine that is in SEQ ID NO:67 to 138 on position +2 (opposite the C in the AUG codon for isoleucine in the wildtype situation) is preferably replaced by a cytidine, preferably a m5Ce or a Of nucleotide to become complementary at that position.

[0233] Example 3. RNA editing of the PNPLA3 I148M variant target transcript using a variety of AONs in HepG2 cells.

[0234] Then, to study the effect of the mismatching nucleotides at positions +3 and +4, a further four sets of AONs (see Fig. 6) was designed based on the ten AONs RM 107578 to RM 107587, but now carrying a cytidine at position +2 to be complementary to the guanosine in the transcript carrying the I148M variant (see SEQ ID NO:56), in which: i) there is a mismatch at position +3 and +4 (RM121965 to RM121974); ii) there is only a mismatch at position +3 (RM121975 to RM121984) and not at +4; iii) there is only a mismatch at position +4 (RM 121985 to RM 121994) and not at +3; and iv) there are no mismatches (RM 121995 to RM 122004) at +3 and +4 with the opposite target sequence.

[0235] To determine the effect of having a single mismatch at +3, a single mismatch at +4, both mismatches, and no mismatches, the AONs of Fig. 6 were tested for their editing efficiency on human PNPLA3 (pre-) mRNA carrying the AUG codon for methionine at position 148. To achieve this, human hepatocellular carcinoma cell line HepG2 were used, which are homozygous for the PNPLA3 I148M sequence variant (Pingitore P et al. Int J Mol Sci. 2019, 20(7): 1629). Similarly, liver organoids generated from HepG2 can be tested to determine the effect of the mismatch or the complementarity of the nucleotide at position +2 in the AON and the opposing nucleotide in the target sequence, and how it influences the efficiency of editing. During the experiment, the target sequence in HepG2 was sequenced and in line with earlier published data (Najafi et al. 2022), it turned out that HepG2, in a homozygous manner, comprises the single nucleotide polymorphism (SNP; identifier rs738408) located five nucleotides 3’ from the target adenosine (a T instead of a C in the PNPLA3 gene), see SEQ ID NO:179 in Fig. 1 , wherein the Aug codon for methionine is underlined (including the C>G mutation, also referenced as SNP rs738409), wherein the target adenosine is a capital A, and wherein the associated SNP rs738408 (five nucleotides 3’ from the target A) is in bold with a capital U. The change from COG to CCU in codon 149 is silent and does not change the amino acid residue at this position (proline) but is also associated with NAFLD like the rs738409 mutation due to its proximity. The presence of the homozygous rs738408 SNP in HepG2 means that the AONs shown in Fig. 6 - at least when used on HepG2 cells - all comprise a further G:U wobble base pair with the target sequence at nucleotide position +5 in the AON sequence. Further AONs were designed that comprise an adenosine at position +5 in the AONs, based on this knowledge, to circumvent the wobble. This further set (with the Watson- Crick matching adenosine at position +5) is shown in Fig. 9.

[0236] To investigate editing, two experiments with HepG2 were performed, with a first experiment using the following 7 sets of AONs (depicted in Fig. 6):

[0237] RM121965 - RM121975 - RM121985 - RM121995 RM121966 - RM121976 - RM121986 - RM121996 RM121967 - RM121977 - RM121987 - RM121997 RM121968 - RM121978 - RM121988 - RM121998 RM121969 - RM121979 - RM121989 - RM121999 RM121970 - RM121980 - RM121990 - RM122000 RM121971 - RM121981 - RM121991 - RM122001 RM 107579 was taken along as a control in the first experiment, together with an Actin B (ActB) specific AON (RM4266; SEQ ID NO:298), a mock procedure and a non-treated sample. In the second experiment, the following additional 3 sets of AONs were tested:

[0238] RM121972 - RM121982 - RM121992 - RM122002 RM121973 - RM121983 - RM121993 - RM122003 RM121974 - RM121984 - RM121994 - RM122004 Here, RM107578 was taken along as a control, together with the ActB AON RM4266, a mock procedure and a non-treated sample.

[0239] HepG2 cells were seeded in a 24-well plate with a density of -75,000 cells / well. Cells were cultured overnight in MEM advanced (Gibco), complemented with 10% FBS (Biowest) 1 % P / S (Gibco) and 2 pM L-glutamine (Gibco) at 37°C and 5% CO2. AONs were added in technical triplicates to the cells with a final concentration of 5pM to allow gymnotic uptake for 24h under the above-mentioned culture conditions. A validated ActB editing AON, as well as a non-treated (NT) control were taken along as negative controls. After 24 h half of the medium was removed and replaced with medium complemented with a purified AG 1856 triterpenoid saponin (see Inti. Patent Application Publication No. WO2024 / 153801), resulting in a final concentration of 0.5 pM AG1856. After 4 h of AG1856 treatment the complete media was refreshed (additionally also an AG1856 mock control was included). The cells were then cultured for an additional 72 h. Subsequently, the cells were washed once with 500 pl PBS (Gibco) prior to total RNA isolation (Maxwell, Promega, according to manufactures protocol). RNA quality and concentration was determined using Nanodrop (Thermo Scientific). 1 pg of total RNA was used for cDNA conversion using Maxima Reverse transcriptase (Thermo Scientific). A ddPCR (Qiagen) assay was performed using the set of specific primer / probes shown in Table 2 to identify and quantitatively determine A to I (G) conversion.

[0240] Table 2: Primers and probes with their respective sequences for quantitative PCR assays. The SEQ ID NO of each primer or probe is given between brackets following the respective sequence in the middle column. The + indicates a Locked Nucleic Acid (LNA) on the 3' side of the symbol.

[0241] Total expression of PNPLA3 and A to I (G) conversion of Actin B using an ActB specific AON (RM4266) was also determined (data not shown). PNPLA3 editing was determined by calculating the original (A) copies / ng RNA and edited (G) copies / ng RNA followed by calculating the edit percentage using the following formula (G copies / ng RNA) I (G copies / ng RNA + A copies / ng RNA) * 100%). Statistics were performed in Graphpad.

[0242] Fig. 7A shows the results of the first experiment with the 7 sets of AONs, arranged from best performer (RM121976) to worst performer (RM121971). Control AON RM107579 provided hardly any detectable editing, suggesting that when the AON comprises a guanosine at position +2, which is opposite the cytidine in the AUG codon in a wildtype situation, that such is not beneficial for editing the adenosine in an AUG codon (in the I148M variant situation), and that the +2 position is preferably a cytidine. No editing was observed with the negative control AON, the mock and NT controls. Strikingly, as can be observed with the shading of the bars, in which the black bars represent the AONs that only comprise a mismatching nucleotide at position +3 (MM @3; see RM121975 to RM121984 in Fig. 6), AONs RM121975 to RM121981 outperformed a) the AONs with two mismatches (MM @3 and @4), b) the AONs with only a mismatch at +4 (MM @4) and c) the AONs without mismatches (no MM), although, as discussed above, in all AONs, the nucleotide at position +5 forms a G:U wobble with the opposite nucleotide in the HepG2 PNPLA3 target sequence. This finding was further supported by the second experiment with the remaining three sets as shown in Fig. 7B. Again, the AONs with the mismatch at position +3 (MM @3; black bars) outperformed the other AONs, and again the control (here RM 107578) did not show significant editing levels. The best performing AONs (with >30% editing in these HepG2 cells) were RM121976 (SEQ ID NO:150), RM121977 (SEQ ID NO:151), and RM121982 (SEQ ID NO:156). From these experiments it becomes clear that it is preferred to have at least a complementary nucleotide at +2 in the AON, while it is preferred that the nucleotide at position +3 is not complementary and represents preferably an A:C mismatch. Even more preferably, when the position +3 is an A:C mismatch, the position at +4 is complementary to the target sequence.

[0243] Example 4. In-cell lipase activity by PNPLA3 (wt), PNPLA (148M), and PNPLA3 (148V) in HeLa cells.

[0244] As outlined in detail herein, it was envisioned that - whilst the PNPLA3 I148M variant strongly increases hepatic triglyceride accumulation - replacing the methionine at position 148 in the PNPLA3 protein with a valine residue (M148V) would re-instate the PNPLA3 wildtype activity such that its lipase activity is restored. To investigate this, 2x3 expression plasmids were generated for over-expression of PNPLA3-WT, PNPLA3-148M, and PNPLA3-148V protein, using the pCDNA3.1 expression plasmid backbone. FLAG tags for protein determination (immunofluorescence as well as western blot detection) were placed either at the N-terminus or on the C-terminus, thereby resulting in six different expression plasmids. Generation of the respective cDNAs and cloning of these plasmids were performed by methods known to the person skilled in the art. Expression levels were determined after transfection in 96-wells plates for 24-48 hr, using Lipofectamine 3000 in HeLa cells and in HEK-293T cells followed by fluorescence CQ1 confocal microscopy and western blotting. Expression levels and attachment of cells appeared better in HeLa cells than in Hek-293T cells (data not shown). Both N-terminal and C-terminal FLAG-containing proteins were expressed as observed by confocal microscopy (data not shown). Western blots showed a somewhat higher expression of the PNPLA3 protein containing the FLAG at the C-terminus than at the N-terminus for all three variants (data not shown).

[0245] After determination of the best conditions for transfecting the three protein varieties and the best expressing cell line (HeLa), transfected cells were used in an in-cell lipase activity in which, 8hr after start of the transfection, the cells were refreshed with fresh medium and incubated with three types of triglyceride: Oleic Acid (OA), which is not a natural substrate for PNPLA3, and Linoleic Acid (LA) and alpha-Linolenic Acid (a-LA) that are both natural substrates for PNPLA3, each in a concentration of 250 pM, overnight for 16 hrs. Cells were stained after 24 hr after start of the transfection because it was found that after 48 hr, FLAG tag signals had significantly reduced (data not shown). Hence, after 24 hr after the start of the transfection, intracellular lipid droplet sizes and intensity assessment using CQ1 confocal microscopy was performed on a cell-by-cell basis to investigate the effect of lipid droplet size after over-expression of the three PNPLA3 variants. Lipids were stained with Bodipy (Thermo Fisher) such that the size of the lipid bodies inside the cells could be determined. Clearly, not all cells were transfected and therefore, only FLAG-positive cells were analysed, ranging from several dozens to a few hundred cells per transfection.

[0246] Results are shown in Fig. 8. The upper three panels show the results with the FLAG tag attached to the C-terminus using PNPLA3-WT (C-1481), PNPLA3-148M (C-148M), and PNPLA3-148V (C-148V) in (from left to right) Oleic Acid (OA), Linoleic Acid (LA), and alpha- Linolenic Acid (a-LA) treated cells, displaying the mean area of Bodipy staining per positively transfected cell (pm2). The lower three panels display the same, but for the proteins in which the FLAG tag was positioned at the N-terminus. In each instance, the wildtype version of PNPLA3 (PNPLA3-WT; C-1481 or N-1481) reduced the size of the Oleic Acid (OA) droplets to approximately 10 pm2and completely reduced the size of the droplets after Linoleic Acid (LA) and alpha-Linolenic Acid (a-LA) treatment, which was as expected. The sizes of the droplets after transfection with PNPLA3-148M were significantly larger with ranges of approximately 35 pm2(OA), 43 pm2(LA), and 30 pm2(a-LA), see the upper three panels. As outlined herein, it was envisioned that replacing the methionine at position 148 in the I148M with a valine residue, thereby in fact placing a valine at position 148 instead of a methionine, due to RNA editing (M148V) of the PNPLA3 transcript molecules, the wildtype functionality of the PNPLA3 protein should be restored. As can be shown in all six panels in Fig. 8, this appeared to be a correct thought: in each case, be it with the FLAG tag at the C-terminus or at the N-terminus, the PNPLA3-148V (C-148V and N-148V) reduced the lipid droplet sizes in the transfected cells, especially in the case of LA and a-LA, to an extent that is comparable to what was found with the wildtype protein (right bars in each of the six panels, in comparison to the left bars in each of the six panels, respectively). This indicates that, even though the I148M variant cannot be reversed by ADAR-mediated RNA editing, it is appropriate to deaminate the first adenosine in the AUG codon for methionine to generate a codon for valine (IUG / GUG) at this position, because that generates a PNPLA3 variant that has a lipase activity that is comparable to the wildtype version.

[0247] Example 5. RNA editing of the PNPLA3 I148M variant target transcript using a further set of AONs in HepG2 cells.

[0248] Further investigations were performed to reveal whether it is beneficial to have complementarity at the +5 position in the AON, since all AONs of Fig. 6 generate a wobble base pair with the opposite nucleotide in the target sequence at that position, because of the additional rs738408 SNP present in the PNPLA3 gene in HepG2, as outlined above. The 40 AONs shown in Fig. 6 (that all comprise a Gf nucleotide at position +5) were then manufactured with an Af nucleotide at that position. This new set of 40 AONs is provided in Fig. 9. Using these AONs, the same experiment as discussed in Example 3 was performed, except for the probes in the quantitative dPCR, that were replaced by the probes provided in Table 3.

[0249] Table 3: Probes with their respective sequences for quantitative PCR assays. The SEQ. ID NO of each probe is given between brackets following the respective sequence in the middle column. The + indicates a Locked Nucleic Acid (LNA) on the 3' side of the symbol.

[0250] The results are shown in Fig. 10 and confirmed what was already observed in the results provided in Fig. 7A and Fig. 7B, wherein the shading of the bars show that the black bars represent the AONs that only comprise a mismatching nucleotide at position +3 (MM @3; see RM 122670, RM 122671 , RM 122675, RM 122676, RM 122669, RM 122673, and RM 122672 in Fig. 9) again outperformed: a) the AONs with two mismatches (MM @3 and @4); b) the AONs with only a mismatch at +4 (MM @4); and c) the AONs without mismatches (no MM), in which, as discussed in this Example, in all AONs the nucleotide at position +5 no longer mismatched with the opposite nucleotide in the HepG2 PNPLA3 target sequence. Strikingly, the best performing AONs in this second screen (with the Af at position +5) were the equivalents of the best performing AONs in the first screen. For example, RM 122670 (SEQ ID NO: 191) and its earlier equivalent AON RM 121976 (SEQ ID NO: 150) were both very effective in RNA editing. The same holds true for RM 122671 (SEQ ID NO: 192) and its earlier equivalent RM121977 (SEQ ID NO:151). The same holds true for RM122676 (SEQ ID NO:197) and its earlier equivalent RM121982 (SEQ ID NO:156). The same holds true for RM122675 (SEQ ID NO:196) and its earlier equivalent RM121981 (SEQ ID NO:155).

[0251] Example 6. RNA editing of the PNPLA3 I148M variant target transcript using a further set of AONs in HepG2 cells.

[0252] Based on the results obtained in Examples 3 and 5, the variety of mismatches and / or wobbles at positions +3, +4, and +5 were studied together in one experiment in HepG2 cells as described above, using a new set of AONs that was designed based on the results observed with RM121976 (SEQ ID NQ:150), RM121977 (SEQ ID NO:151), RM121981 (SEQ ID NO: 155), RM 121982 (SEQ ID NO: 156) and their equivalents studied in Example 5. These four AONs (that performed best) all generate an A:C mismatch at nucleotide position +3, a G:ll wobble base pair at position +5, and a Watson-Crick complementary G:C base pair at nucleotide position +4. The primers as provided in Table 2 and the probes as provided in Table 3 were again used for the dPCR. The newly designed AONs are provided in Fig. 11 .

[0253] RM122801 to RM122811 comprise an A:C mismatch at nucleotide position +3, a C:C mismatch at nucleotide position +4, and a G:ll wobble at nucleotide position +5.

[0254] RM 122812 to RM 122822 comprise an A:C mismatch at nucleotide position +3 and a G:ll wobble at nucleotide position +5 (like RM121976, RM121977, RM121981 , RM121982, RM122670, RM122671 , RM122675, and RM122676).

[0255] RM122823 to RM122833 comprise a C:C mismatch at nucleotide position +4, and a G:ll wobble at nucleotide position +5.

[0256] RM122834 to RM122844 comprise an A:C mismatch at nucleotide position +3 and a C:C mismatch at nucleotide position +4.

[0257] RM122845 to RM122855 comprise an A:C mismatch at nucleotide position +3.

[0258] RM 122856 to RM 122866 comprise a C:C mismatch at nucleotide position +4.

[0259] The editing of the adenosine in the AUG codon for methionine at position 148 in the human PNPLA3 protein using this new set of AONs in human HepG2 cells, in comparison to RM121976, RM121977, RM121981 , and RM121982 is shown in Fig. 12. It is immediately obvious that, whereas good editing efficiencies are observed with these four earlier lead AONs (speckled bars on the right), the set of AONs that also comprise a mismatching nucleotide at positions +3 and a wobble at position +5, but not at position +4 (RM122812 to RM122822), again outperform the AONs with other mismatching combinations, with RM122818 (SEQ ID NO:243), RM122813 (SEQ ID NO:238), RM122812 (SEQ ID NO:237), RM122820 (SEQ ID NO:245), RM122816 (SEQ ID NO:241), RM122821 (SEQ ID NO:246), RM122817 (SEQ ID NO:242), and RM122814 (SEQ ID NO:239) performing best. This shows that - when it comes to targeting and deaminating the adenosine in the AUG codon for position 148 in the human PNPLA3 protein - it is preferred that the AON does not mismatch at position +4, but in contrast preferably comprises mismatching / wobbling nucleotides at positions +3 and +5.

[0260] Example 7. RNA editing of the PNPLA31148M variant target transcript using human liver cell lines.

[0261] The experiments of Examples 3, 5, and 6 were performed in HepG2 cells. To investigate the editing efficiencies in cells that would represent ‘normal’ liver cells rather than a cancerous cell line that has been continuously cultured over decades, and by applying a GalNAc conjugated to the AON for targeting to the asialoglycoprotein receptor (ASGPR) found on liver cells and to investigate such increased targeting effect, the best performing AONs (see above) were manufactured with a tri-antennary GalNAc moiety attached to the 3’ terminus. The AONs RM121976 (SEQ ID NQ:150), RM121977 (SEQ ID NO:151), RM122816 (SEQ ID NO:241), RM122818 (SEQ ID NO:243), RM122820 (SEQ ID NO:245) and RM 122821 (SEQ ID NO:246) were selected and resulted respectively in RM 123992 (SEQ ID NO:292), RM123993 (SEQ ID NO:293), RM123994 (SEQ ID NO:294), RM123995 (SEQ ID NO:295), RM123996 (SEQ ID NO:296) and RM123997 (SEQ ID NO:297), shown in Fig. 11 , wherein ‘L005’ represents the tri-antennary GalNAc moiety provided by Formula IX(b) herein.

[0262] Two types of liver cells were used for this purpose, in which both are homozygous for the human PNPLA3 I148M mutation. The first are HepaCur human hepatocytes that were engrafted in and isolated from mice (Yecuris; cat# 20-0001 ; donor# HHF13022) and the second are male human cryoplateable hepatocytes (BiolVT; cat# M00995-P; donor# IXL). Both cell types carry the rs738408 and rs738409 SNPs (confirmed by sequencing; data not shown), like HepG2, and as expected because of their proximity.

[0263] Three different AON treatment procedures were used in this experiment:

[0264] A. Seed and T reat (S&T): the human hepatocytes were seeded in presence of 5 pM AON; after 24 h medium was refreshed; cells were harvested 72 h after start of the AON treatment.

[0265] B. Seed and Treat pulse (S&T + AG1856 pulse): the human hepatocytes were seeded in presence of 5 pM AON; 20 h after seeding 0.5 pM AG1856 was added; 4 h later cells were refreshed with new medium; cells were harvested 72 h after start of the AON treatment. C. AG1856 pulse: the human hepatocytes were seeded and left for 24 h to adhere; after 24 h cells were treated with 5 pM AON; 20 h later cells were treated with 0.5 pM AG1856; 4 h later cells were refreshed with new medium; cells were harvested 72 h after start of the AON treatment.

[0266] Negative controls were a non-treated sample (NT), a sample that only received AG1856 but no AON, and two samples with unrelated AONs (RM4266; SEQ ID NO:298; and RM4777; SEQ ID NO:299). Yecuris and BiolVT primary human hepatocytes were thawed according to the manufacturer’s protocol and seeded in 96-well collagen coated plates with a density of -50,000 cells / well for all the above-described methods, using INVITGRO CP media (Yecuris) and INVITGRO HI media (BiolVT). Each AON was tested in six consecutive wells. The cells were washed once with 200 pl PBS prior to total RNA isolation. Per well 50 pl TG / homogenization solution (Promega) was added per well. Per P / VPLA3-targeting AON two wells were pooled, ending up with three wells each containing 100 pl of cell lysate per EON treatment (technical triplicate). RNA isolation, cDNA production, and dPCR was carried out as described in Example 5, with the primers as provided in Table 2 and the probes as provided in Table 3.

[0267] The results with the primary human hepatocytes from Yecuris are shown in Fig. 13A, whereas the results with the primary human hepatocytes from BiolVT are shown in Fig. 13B. It is instantly clear that a gymnotic uptake of the AONs without the assistance of the endosomal escape enhancer (AG1856) results in relatively low editing percentages in the Yecuris cells (S&T dark grey bars; with RM123994 scoring the highest with -10% editing), although relatively higher levels were observed in the S&T samples in the BiolVT cells with the AONs comprising a GalNAc moiety (dark grey bars; again with RM123994 scoring best with -23% editing). However, a clear improvement in editing efficiency was observed in the S&T AG1856 pulse treated cells (light grey bars), reaching approximately 55-56% editing in the Yecuris hepatocytes using RM123997. In both cell types the use of the GalNAc moiety appeared to be beneficial, whereas the effect of the GalNAc conjugation was more pronounced in the BiolVT hepatocytes. These experiments show that the six selected and highly preferred AONs can induce significantly high levels of PNPLA3 editing, targeting the adenosine in the AUG codon of a PNPLA3 transcript carrying both rs738408 and rs738409 SNPs, wherein it is preferred that the AON comprises a GalNAc moiety, attached to the 3’ or 5’ terminus of the AON, that is preferably a tri-antennary GalNAc moiety, to increase the efficiency of targeting to the ASGPR protein on liver cells. Even though the experiments outlined here are in vitro experiments using cultured cells, the S&T AG1856 pulse set-up allows for a conclusion that the ASGPR protein is available on the seeded cells and increases the rate of cell entry of the AON when it is conjugated to a GalNAc moiety, which is significantly important for in vivo delivery in targeting the liver. In such an in vivo setup it is also preferred to conjugate the AG1856 saponin to a GalNAc moiety, to have that co-targeted to liver cells and assist the AON in endosomal escape and more efficient intracellular delivery and subsequent pre-mRNA and mRNA targeting, followed by the deamination of the target adenosine through the activity of an endogenous ADAR enzyme.

[0268] Example 8. RNA editing of the PNPLA3 I148M variant target transcript using a subset of AONs in HepG2 cells.

[0269] Based on the results shown above, the best performing AONs were also tested in a Seed & Treat + / - AG1856 pulse treatment as outlined in Example 7, using HepG2 cells. The following AONs were selected: RM121976 (SEQ ID NO:150), RM121977 (SEQ ID NO:151), RM122816 (SEQ ID NO:241), RM122818 (SEQ ID NO:243), RM122820 (SEQ ID NO:245), and RM122821 (SEQ ID NO:246). Negative controls were RM4266 (SEQ ID NO:298) and RM4777 (SEQ ID NO:299). HepG2 cells were seeded in a 96-well plate with a density of -12,500 cells / well. The same experimental procedure with AON treatment and saponin pulses as outlined in Example 7 was used. The same methods for RNA isolation, cDNA manufacturing and dPCR were as outlined in Example 6.

[0270] The results are shown in Fig. 14 and indicate that while a Seed & Treat procedure without the use of saponin (procedure A) editing efficiencies were rather low, the editing percentages reached around 30% in the B) S&T pulse treatment as well as in the C) AG1856 pulse treated cells, with all selected AONs that all provided significant high levels of RNA editing. RM 122816 (SEQ ID NO:241) performed best with editing levels above 30%, confirming the results achieved earlier in HepG2 cells as well as the results obtained in the liver cells obtained from Yecuris and BioIVT.

[0271] Example 9. RNA editing of the PNPLA3 I148M variant target transcript using AONs generating a variety of G:U wobble base pairs in HepG2 cells (wobble walk).

[0272] Because the results obtained in the previous examples indicated that the presence of (additional) mismatches and wobble base pairs between the PNPLA3 editing AONs and the human PNPLA3 target transcript molecules carrying the rs738408 and rs738409 SNPs, a further study was performed using an additional set of AONs comprising a variety of nucleotides that would generate G:ll wobble base pairs at different positions. The AONs comprising the mismatching adenosine at position +3 (underlined) and the wobble base pair generating guanosine at position +5 (underlined), with these additional wobble base pair generating nucleotides (in bold) are RM 123137 to RM 123150 in Fig. 11 , in which their respective SEQ ID NO’s are also provided. The same set of AONs were manufactured, in which the same additional wobble base pair generating nucleotides were introduced, but in the absence of the wobble base pair generating guanosine at position +5. These AONs are RM123151 to RM123164 in Fig. 11 , again in which their respective SEQ ID NO’s are also provided. RM121976, RM122816, RM122818, RM122820, and RM122821 were also taken along in this experiment for comparison.

[0273] In this experiment, again HepG2 cells were seeded in 96-well collagen coated plates with a density of -12,500 cells / well. Cells were cultured overnight at 37°C and 5% CO2. 24 h post seeding media was replaced with 100 pL complete medium containing 5 pM AON. Cells were incubated for 24 h under the above-mentioned culture conditions. After 24 h half of the medium was removed and replaced with medium complemented with a final concentration of 0.5 pM AG1856. After 4 h of AG1856 treatment the complete media was refreshed, and the cells were cultured for an additional 48 h. A mock treatment and a non-treated (NT) sample served as controls. RNA isolation, cDNA generation and dPCR procedures were as described in Example 6.

[0274] The results are shown in Fig. 15. As observed before, the AONs comprising a guanosine nucleotide at position +5 (RM123137 to RM123150, RM121976, RM122816, RM122818, RM122820, and RM122821), generating a wobble G:ll base pairwith the opposite nucleotide in the target sequence generally performed better than the AONs with an adenosine at position +5 (RM123151 to RM123164). No significant improvement was observed overwhat was seen before with the best performing AON in HepG2 cells, which is RM122816, although it may be concluded that introducing an additional wobble base pair at positions +19 (RM123139; SEQ ID NO:302), +15 (RM123141 ; SEQ ID NQ:304), +13 (RM123143; SEQ ID NQ:306), and -6 (RM123150; SEQ ID NO:313) does not compromise RNA editing efficiency. In this experiment, RM123141 (SEQ ID NQ:304) with the additional wobble base pair at position +15 slightly outperformed RM122816.

[0275] Example 10. RNA editing of the PNPLA31148M variant target transcript using a human liver cell line and a variety of AONs comprising chemical modifications.

[0276] As shown in the previous examples, RM122818 (SEQ ID NO:243) was among the best performing AONs to trigger the deamination of the target adenosine in the AUG codon giving rise to the I148M mutation in human PNPLA3. Based on the design of RM122818 a further set of AONs was generated and tested for RNA editing in the human liver cells (from a male donor) that were acquired from BiolVT (see Example 7). These AONs are provided in Fig. 16. RM123994 (SEQ ID NO:294) and RM123995 (SEQ ID NO:295), which are the 3’-GalNAc versions of RM122816 (SEQ ID NO:241) and RM122818, respectively, were taken along in these experiments. All tests were performed using the Seed & Treat (S&T) protocol A, and the S&T pulse protocol B as outlined in Example 7. Cell culture, conditions, RNA isolation, cDNA generation, and dPCR procedures were all as described above in Example 7. In a first experiment, six AONs were manufactured that were based on RM123992 to RM 123997 (SEQ ID NO:292 to 297), but instead of a Zd at the orphan position comprised a deoxynucleotide comprising a 5-aza-5,6-di hydro cytosine nucleobase (abbreviated to Ed in Fig. 16). This was to test whether this alternative cytosine analog could replace the Benner’s base. These six AONs are provided as RM125052 to RM125057 (SEQ ID NO:328 to 333). Furthermore, six AONs were manufactured that were also based on RM 123992 to RM 123997, but instead of an Id at nucleotide position -1 comprised a deoxyguanosine (Gd). These six AONs are provided as RM 125058 to RM 125063 (SEQ ID NO:334 to 339). The results with RM 125052 to RM 125057 are provided in Fig. 17A. This shows that - generally spoken - the E base could replace the Benner’s base at the orphan position and may serve as a proper alternative for this modification. In any case, underthese conditions, in a S&T pulse procedure, RM125057 (SEQ ID NO:333) outperformed RM123994 and RM123995. The results with RM 125058 to RM 125063 are provided in Fig. 17B. This shows that, even though proper RNA editing percentages were achieved with all six AONs, it seems that it is preferred that the -1 position is a deoxynucleotide comprising a hypoxanthine nucleobase (Id), rather than a deoxyguanosine (Gd).

[0277] In a next experiment, a variety of AONs were manufactured based on the finding that G:ll wobble base pairs at nucleotide positions +15 and / or +19 (see Example 9) could be introduced without negatively impacting the editing efficiency. This further set of AONs is represented by RM125066 to RM125122 (SEQ ID NQ:340 to 396) shown in Fig. 16. The editing results (percentages) in the primary human hepatocytes are shown in Fig. 18 and indicate that the positive controls RM123994 and RM123995 provide editing levels of -35%, but that at least two alternative AONs, represented by RM 125066 (SEQ ID NQ:340) and RM125067 (SEQ ID NO:341) perform in a similar efficient manner, also providing -35% editing. RM125067 comprises a 2’-OMe modified uridine (Um) at position +15, which is in line with the good performing RM123141 oligonucleotide (SEQ ID NQ:304), see Example 9. RM125066 comprises a 2’-F modified guanosine (Gf) at position +19, which is in line with the good performing RM123139 oligonucleotide (SEQ ID NQ:302), see Example 9. Interestingly, RM125068 that has G:U wobble base pairing nucleotides at both positions (+15 and +19) did not perform as good, indicating that it is preferred to have at least one of the +15 and +19 positions generating a G:U wobble base pair, but not both. Both RM125066 and RM125067 comprise PNms linkages at the positions of the PNdmi linkages present in RM 123139 and RM123141. Next to that, RM125066 and RM125067 comprise a 3’-conjugated GalNAc moiety. Furthermore, RM125066 and RM125067 comprise PNms linkages at linkage positions 0, -1 , and -2, instead of the PS, PS, and MeP linkages, respectively, that are present in RM123139 and RM123141. PS linkages were also introduced replacing PO linkages at a variety of linkage positions. From this same experiment it could also be concluded that replacing the 2’-M0E modified adenosine at position +3 and the 2’-F modified guanosine at position +5 by a variety of differently modified nucleotides did not increase the editing efficiency, indicating that these positions are preferably occupied by Ae (at +3) and Gf (at +5). The results obtained with RM 125086 (SEQ ID NO:360) show that the nucleotide position +19 positively influences RNA editing efficiency also when a U : U mismatch is generated albeit not as efficient as when a G:ll wobble is generated.

[0278] In a next experiment, a variety of AONs was designed and generated to investigate editing efficiency based on the presence of different PO, PS, and PNms linkages at different positions. This further set is represented by RM125123 to RM125165 (SEQ ID NO:397 to 439) shown in Fig. 16. The editing results (percentages) in the primary human hepatocytes are provided in Table 4 and indicate a variety of aspects.

[0279] Table 4: Editing percentages of AONs (RM numbers provided) in primary human hepatocytes after a Seed & Treat (S&T) procedure and after a Seed & Treat saponin pulse (S&T pulse) procedure. Editing percentages of each of the three replicates in each procedure is provided together with the average (Avg) of the three replicates. Average percentages obtained after the S&T pulse procedure, and which are comparable to the positive controls RM123994 and RM123995 are given in bold. Names of the AONs that score >40% are also given in bold.

[0280] The AONs comprising a PO linkage at linkage position -3, such as RM125123 (SEQ

[0281] ID NO:397), RM125124 (SEQ ID NO:398), RM125125 (SEQ ID NO:399), RM125128 (SEQ

[0282] ID NQ:402), RM125129 (SEQ ID NQ:403), RM125130 (SEQ ID NQ:404), RM125138 (SEQ

[0283] ID NO:412), RM125139 (SEQ ID NO:413), RM125140 (SEQ ID NO:414), RM125141 (SEQ

[0284] ID NO:415), RM125142 (SEQ ID NO:416), RM125143 (SEQ ID NO:417), RM125144 (SEQ

[0285] ID NO:418), RM125145 (SEQ ID NO:419), RM125146 (SEQ ID NQ:420), RM125149 (SEQ

[0286] ID NO:423), RM125150 (SEQ ID NO:424), RM125151 (SEQ ID NO:425), RM125158 (SEQ

[0287] ID NO:432), RM125160 (SEQ ID NO:434), RM125161 (SEQ ID NO:435), RM125162 (SEQ

[0288] ID NO:436), RM125163 (SEQ ID NO:437), RM125164 (SEQ ID NO:438), and RM125165

[0289] (SEQ ID NO:439), all outperform the AONs that do not have a PO linkage at this position. This clearly indicates the preference for a PO linkage at linkage position -3. Certain AONs with a

[0290] PS linkage at linkage positions +8, +11 , and / or +14 outperformed RM123994 and RM123995, if there was a PO linkage at position -3.

[0291] In a next experiment, a set of AONs was designed and manufactured that comprise one or more PNms linkages surrounding the orphan nucleotide. This further set is represented by RM125178 to RM125182, and RM125184 to RM125198 (SEQ ID NQ:440 to 460) shown in Fig. 16. RM125158 (SEQ ID NO:445) is a duplicate of SEQ ID NO:432. The editing results (percentages) in the primary human hepatocytes are shown in Fig. 19 with the positive controls RM123994 and RM123995 shown on the left. RM125178 (SEQ ID NO:440), RM125179 (SEQ ID NO:441), RM125180 (SEQ ID NO:442), RM125181 (SEQ ID NO:443), RM125182 (SEQ ID NO:444), RM125158 (SEQ ID NO:445 and 432), and RM125184 (SEQ ID NO:446) performed similar to the positive controls, and outperformed all AONs in this set that did not comprise a PO linkage at position -3. This confirms the results shown above that for this position it is highly preferred that it is a PO linkage. The results also show that each of the linkage positions 0, -1 , and -2 can either be a PNms linkage or a PS linkage and that the choice between these two does not significantly influence the editing efficiency. The linkage position -2 can, besides being a PS or a PNms linkage, also be a MP linkage as shown before.

[0292] In a next experiment it was investigated how length would influence editing efficiencies, taking the knowledge of the best performing AONs along. The set of these length variable AONs is represented by RM125215 to RM125235 (SEQ ID NO:461 to 481) in Fig. 16. The editing results (percentages) in the primary human hepatocytes are shown in Fig. 20 with the positive controls RM 123994 and RM 123995 shown on the left. Certain AONs that were shortened on the 5’ terminus (in comparison to the positive control RM123995), such as RM125216 (SEQ ID NO:462; 2 nucleotides shorter), RM125217 (SEQ ID NO:463; 3 nucleotides shorter), and RM125218 (SEQ ID NO:464; 4 nucleotides shorter), performed better in the S&T procedure (without using saponin). The S&T pulse samples for RM 125227 to RM125231 were lost due to experimental / technical issues. RM125232 (SEQ ID NO:478) and RM 125233 (SEQ ID NO:479) outperformed the positive controls. Both AONs are 30 nucleotides in length, like RM123994 and RM123995, but with a shorter 5’ wing (calculated from the orphan nucleotide), and a longer 3’ wing. RM 125232 is 30 nt with a 19 nt 5’ wing and a 10 nt 3’ wing, whereas RM125233 is 30 nt with an 18 nt 5’ wing and an 11 nt 3’ wing. RM 123994 and RM 123995 are also both 30 nt, but both with a 23 nt 5’ wing and a 6 nt 3’ wing. From this it is concluded that when the AON is 30 nt in length the number of nucleotides in the 5’ wing (calculated from the orphan nucleotide) is preferably 18, 19, 20, 21 , 22, or 23 nt in length, whereas the number of nucleotides in the 3’ wing (calculated from the orphan nucleotide) is preferably 6, 7, 8, 9, 10, or 11 nucleotides in length. Hence, the AON, when it is 30 nucleotides long, has preferably the following code in respect of (a-) symmetrical design: 5’-18nt-orphan-11nt-3’, 5’-19nt-orphan-1 Ont-3’, 5’-20nt-orphan-9nt-3’, 5’-21 nt-orphan-8nt-3’, 5’-22nt-orphan-7nt-3’, or 5’-23nt-orphan-6nt-3’.

[0293] In a next experiment a small screen was performed with a set of AONs in which a variety of positions was selected to include a nucleotide comprising a hypoxanthine nucleobase, which in each case replaces a nucleotide comprising a guanine nucleobase. This set is represented by RM 125236 to RM 125240 and RM 125243 (SEQ ID NO:482 to 486, and SEQ ID NO:487, respectively), provided in Fig. 16. In the editing experiment that was performed as all other experiments in this example, none of the AONs outperformed the positive controls RM 123994 and RM 123995 in either one of the experimental procedures (data not shown).

[0294] In a further experiment, a large set of AONs was designed, all based again on the RM122818 AON, but including a variety of chemical modifications throughout the AON sequences. This set of AONs is represented by RM 125546 to RM 125555, RM 125557 to RM 125564, and RM 125566 to RM 125605 (SEQ ID NO:488 to 545, respectively) shown in Fig. 16. The editing results (percentages) in the primary human hepatocytes are shown in Fig. 21 with the positive controls RM 123994 and RM 123995 shown on the left. These results indicate that some positions, such as nucleotide positions +17 and +18 can perform well when a 2’-MOE modification is present and when either one, or both positions are held by nucleotides comprising a 2’-F modification in the sugar moiety (see for instance RM125568 and RM125569; SEQ ID NQ:508 and 509, respectively). Other good performing AONs were RM125547 (SEQ ID NO:489), RM125570 (SEQ ID NQ:510), and RM125571 (SEQ ID NO:511). RM125581 to RM125584 underperformed, which may be because these four AONs all comprise a 2’-MOE modified nucleotide comprising an adenine nucleobase (Ae) at nucleotide position -2, instead of a 2’-OMe-modified nucleotide comprising an adenine nucleobase (Am), which therefore appears to be the preferred modification at this position.

[0295] Although likely to occur, based on literature (Najafi et al. 2022), it is also investigated whether the rs738408 C>T SNP is associated (co-inherited) with the rs738409 C>G SNP (I148M) variant in liver disease patients, such as patients suffering from NAFLD, that carry this variant (Romeo S et al. 2008).

Claims

CLAIMS1. An antisense oligonucleotide (AON) that is capable of recruiting an endogenous ADAR enzyme in a human cell after the AON has formed a double-stranded complex with a region of a target RNA nucleic acid molecule in the cell, wherein the cell is a liver cell, preferably a hepatocyte, wherein the transcript molecule is a pre-mRNA or an mRNA molecule, wherein the region comprises a target adenosine, wherein the nucleotide in the AON that is opposite the target adenosine is the orphan nucleotide, wherein the nucleotide numbering is such that the orphan nucleotide is number 0 and nucleotides are further positively (+) incremented towards the 5’ end and negatively (-) incremented towards the 3’ end, wherein the ADAR enzyme can deaminate the target adenosine into an inosine after binding to the doublestranded complex, wherein the target RNA nucleic acid molecule is a transcript molecule of the human PNPLA3 gene comprising the rs738409 single nucleotide polymorphism (SNP), and wherein the target adenosine is in the AUG codon coding for methionine (M) at position 148 of the encoded PNPLA3 protein.

2. An AON according to claim 1 , wherein the orphan nucleotide is a deoxynucleotide comprising a cytosine or a cytosine analog.

3. An AON according to claim 2, wherein the cytosine analog comprises a 6-amino-5-nitro- 3-yl-2(1 H)-pyridone nucleobase or a 5-aza-5,6-dihydro cytosine nucleobase.

4. An AON according to any one of claims 1 to 3, wherein the nucleotide at the -1 position is a deoxyinosine.

5. An AON according to any one of claims 1 to 4, wherein the nucleotide at the +2 position comprises a cytosine nucleobase, and preferably further comprises a 2’-F substitution at the ribose sugar moiety.

6. An AON according to any one of claims 1 to 5, wherein the nucleotide at the +3 position generates an A:C mismatch with the nucleotide in the PNPLA3 transcript molecule that is directly opposite the nucleotide at the +3 position.

7. An AON according to any one of claims 1 to 6, wherein the nucleotide at the +4 position generates a Watson-Crick G:C match with the nucleotide in the PNPLA3 transcript molecule that is directly opposite the nucleotide at the +4 position.

8. An AON according to any one of claims 1 to 7, wherein the human PNPLA3 gene comprises the rs738408 SNP, and wherein the nucleotide at the +5 position generates a G:ll wobble base pair with the nucleotide in the PNPLA3 transcript molecule that is directly opposite the nucleotide at the +5 position.

9. An AON according to any one of claims 1 to 8, wherein the AON has the following structure:5’-N23N22N2l N20Nl9Nl8Nl7Nl6Nl5Nl4Nl3Nl2Nl l Nl0N9N8N7N6N5N4N3N2NlOMl M2M3M4M5M6-3’ wherein:O is the orphan nucleotide at nucleotide position 0, which is: i) a deoxynucleotide comprising a 6-amino-5-nitro-3-yl-2(1 H)-pyridone nucleobase; or ii) a deoxynucleotide comprising a 5-aza-5,6-dihydro cytosine nucleobase;Mi at position -1 is a nucleotide comprising a hypoxanthine nucleobase, preferably wherein the nucleotide is a deoxyinosine (Id);M2 at position -2 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is a 2’-OMe modified adenosine (Am);N1 at position +1 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is a deoxyadenosine (Ad) or a 2’-MOE modified adenosine (Ae);N2 at position +2 is a nucleotide comprising a cytosine nucleobase, preferably wherein the nucleotide is a 2’-F modified cytidine (Of);N3 at position +3 is a nucleotide comprising an adenine nucleobase, preferably wherein the nucleotide is Ae;N4 at position +4 is a nucleotide comprising a guanine nucleobase, preferably wherein the nucleotide is a 2’-MOE modified guanosine (Ge);Ns at position +5 is a nucleotide comprising an adenine nucleobase or a guanine nucleobase, preferably wherein the nucleotide is 2’-F modified, more preferably wherein the nucleotide is a 2’-F modified guanosine (Gf); and wherein the linkage between M2 and M3 (linkage position -3) is a phosphorothioate (PS) linkage or a phosphodiester (PO) linkage, preferably a PO linkage.

10. An AON according to claim 9, wherein the linkage between:O and Mi (linkage position -1) is a PS or a mesyl phosphoramidate (PNms) linkage;Mi and M2 (linkage position -2) is a PS, a PNms, a (1 ,3-dimethylimidazolidin-2-ylidene) phosphoramidate (PNdmi), or a methylphosphonate (MP) linkage;N1 and O (linkage position 0) is a PO, a PS, or a PNms linkage;N2 and N1 (linkage position +1) is a PS linkage;N3 and N2 (linkage position +2) is a PO or a PS linkage;N4 and N3 (linkage position +3) is a PO or a PS linkage; and / orNs and N4 (linkage position +4) is a PNdmi or a PNms linkage.

11. An AON according to any one of claims 1 to 10, wherein the linkage between the most terminal two nucleotides on the 5’ and / or 3’ terminus of the AON is a PNdmi linkage or a PNms linkage.

12. An AON according to any one of claims 1 to 11 , wherein the nucleotide in the AON at position +19, +15, +13, and / or -6 generates a G:ll wobble base pair with the nucleotide in the human PNPLA3 target transcript molecule and that is directly opposite the nucleotide in the AON.

13. An AON according to any one of claims 1 to 12, wherein the AON comprises or consists of an AON according to SEQ ID NO:402, 397, 416, 149, 150, 151 , 152, 153, 154, 155, 156,157, 158, 191 , 192, 196, 197, 237, 238, 239, 240, 241 , 242, 243, 244, 245, 246, 247, 302,304, 306, 313, 333, 340, 341 , 360, 398, 399, 403, 404, 412, 413, 414, 415, 417, 418, 419,420, 423, 424, 425, 432, 434, 435, 436, 437, 438, 439, 440, 441 , 442, 443, 444, 446, 489,508, 509, 510, or 511.

14. An AON according to any one of claims 1 to 13, wherein the AON is covalently or non- covalently, directly or through a linker, bound to a GalNAc moiety, preferably a tri-antennary GalNAc moiety.

15. A pharmaceutical composition comprising an AON according to any one of claims 1 to 14, and a pharmaceutically acceptable solvent, excipient, or carrier.

16. An AON according to any one of claims 1 to 14, for use in the treatment of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis.

17. Use of an AON according to any one of claims 1 to 14, in the manufacture of a medicament for the treatment of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis.

18. A method of editing a human PNPLA3 variant pre-mRNA or mRNA molecule in a liver cell, preferably a hepatocyte, wherein the PNPLA3 variant pre-mRNA or mRNA is encoded by a PNPLA3 gene comprising the rs738409 SNP, the method comprising contacting the PNPLA3 pre-mRNA or mRNA molecule with an AON capable of triggering an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a functionality that is equal or similar to the wildtype functionality, and wherein the AON is according to any one of claims 1 to 14.

19. A method of treating, ameliorating, or slowing down the progression of a liver disease caused by a PNPLA3 gene comprising the rs738409 SNP, such as NAFLD, NASH, fibrosis, and cirrhosis, in a human subject in need thereof, the method comprising administering to said subject an AON according to any one of claims 1 to 14, thereby contacting a PNPLA3 pre- mRNA or mRNA molecule encoded by the PNPLA3 gene comprising the rs738409 SNP in a cell of the subject, thereby effecting an ADAR-mediated adenosine to inosine deamination, thereby editing the PNPLA3 pre-mRNA or mRNA molecule to encode a PNPLA3 protein with a functionality that is equal or similar to the wildtype functionality, thereby treating the subject.

20. An in vitro, ex vivo, or in vivo method for the deamination of a target adenosine in a PNPLA3 pre-mRNA or mRNA molecule encoded by a human PNPLA3 gene comprising the rs738409 SNP, in a liver cell, preferably a hepatocyte, the method comprising the steps of:(i) providing the cell with an AON according to any one of claims 1 to 14;(ii) allowing uptake by the cell of the AON;(iii) allowing annealing of the AON to the PNPLA3 pre-mRNA or mRNA molecule;(iv) allowing an endogenous ADAR enzyme to deaminate the target adenosine in the PNPLA3 pre-mRNA or mRNA molecule to an inosine; and optionally(v) using a functional read-out to identify the presence of the inosine in the PNPLA3 pre-mRNA or mRNA molecule.

Images