Therapeutic Oligonucleotides: An Outlook on Chemical Strategies to Improve Endosomal Trafficking

The potential of oligonucleotide therapeutics is undeniable as more than 15 drugs have been approved to treat various diseases in the liver, central nervous system (CNS), and muscles. However, achieving effective delivery of oligonucleotide therapeutics to specific tissues still remains a major challenge, limiting their widespread use. Chemical modifications play a crucial role to overcome biological barriers to enable efficient oligonucleotide delivery to the tissues/cells of interest. They provide oligonucleotide metabolic stability and confer favourable pharmacokinetic/pharmacodynamic properties.

Chemical modifications of oligonucleotides have proven to be crucial to improve oligonucleotide drug properties. The diligent efforts on oligonucleotide chemistry for more than half a century have resulted in increasing oligonucleotide on-target activity and metabolic stability while decreasing off-target effects and immunogenicity. As of today, there are more than 15 FDA-approved oligonucleotide-based drugs mainly of small interfering RNAs (siRNAs) and antisense oligonucleotide (ASO) classes; latter including, in particular, splice switching oligonucleotides (SSOs), and encompassing either phosphorothioate (PS) oligonucleotides or phosphorodiamidate morpholino oligomers (PMOs), to target various disease-causing genes in the liver, CNS, and muscles.

Biological barriers preventing therapeutic activity of oligonucleotides.

Local delivery of oligonucleotides to the eye, spinal cord, and brain has been successfully achieved via intravitreal (IVT) and intrathecal (IT) administration, respectively. However, systemic administration of oligonucleotides has been less successful due to poor tissue distribution and uptake. While both lipid nanoparticles and N-acetylgalactosamine (GalNAc) conjugates are clinically validated and approved delivery strategies for liver targets, efficient delivery of oligonucleotides beyond the liver is a fundamental obstacle preventing their clinical utility. Alternative approaches such as lipid conjugation, exosome loading, enveloped virus, spherical nucleic acids (SNAs), DNA cages, and smart materials show great promise to improve oligonucleotide delivery into specific tissues.

Recently, it has been reported using quantitative NanoSIMS microscopy that only 1–2% of N-acetylgalactosamine (GalNAc)-conjugated ASOs escape endosomes to engage with the target in vivo. As demonstrated from clinical successes of oligonucleotide therapeutics, <1% of endosomal escape is sufficient to achieve robust and long-lasting activity in the liver when using GalNAc conjugation. A single subcutaneous administration of GalNAc-conjugated siRNAs can induce a clinical benefit that lasts up to 6–12 months in humans. However, the mechanistic understanding of how the oligonucleotides escape from endosomes is still not understood. Various hypotheses have been proposed in the literature: 1. The repetitive, spontaneous, and short-lived, small breaches that occur in the endosomal lipid bilayer can allow the escape of oligonucleotides into cytosol. 2. Oligonucleotides can escape through a temporary breach generated during the fusion events between endosomes, multivesicular bodies (MVBs), and lysosomes. 3. Oligonucleotides may escape via retro-transport from the Golgi.

Different chemical modifications used in oligonucleotide therapeutics

* May also include modifications other than PN backbone modification.

Chemical Modifications in ASOs

ASOs are short (about 20 mer long) and single-stranded synthetic nucleic acids. They have been successfully used for modulating the expression of transcripts, either by direct cleavage of the target RNA through the RNase H1 endonuclease, or by acting as steric blockers by interfering with the translation machinery of the transcript.

A range of modifications have been introduced to enhance ASO stability, efficacy, and delivery. The most common modifications consist of (i) substituting the phosphodiester (PO) backbone with a phosphorothioate (PS); (ii) modifying the 2′ position of the ribose; (iii) incorporating 5-methyl group on pyrimidine nucleobases; or (iv) introducing an artificial biopolymer scaffold, such as a PMO or peptide-nucleic acid (PNA) .

PS-ASOs have a longer circulation time in vivo that correlates with an enhanced binding to plasma proteins. This “simple” modification leads to a productive cellular uptake that is not observed with the canonical PO and thus ignited the clinical potential of ASOs. The PS backbone also generates a broader interactome between the ASO and membrane-bound proteins and, depending of the protein-ASO pair, initiates different endocytotic pathways. The latest work suggests a dynamic covalent exchange between the PS and the thiols/disulfides from proteins, resulting in a “thiol-mediated uptake”. While the PS modification improves ASO stability and bioavailability, it also reduces the affinity for the target RNA. Therefore, incorporation of additional modifications is likely required to enhance ASO activity. 2′-modifications (e.g., 2′-O-methyl (2′-OMe), 2′-O-methoxy-ethyl (2′-MOE), and 2′-fluoro (2′-F) modifications) or constrained ribose and bridged nucleic acid modifications (e.g., locked nucleic acid (LNA) and tricyclo-DNA (tcDNA) modifications) enable superior binding affinity to RNA, ASO nuclease resistance, and can have a significant impact on ASO distribution. Furthermore, introduction of the 5-methyl group on cytosine reduces the immunostimulatory profile of certain DNA oligonucleotides and enhances nuclease stability.

Chemical Modifications in siRNAs

siRNAs are short double-stranded RNAs (20 to 25-mer RNA duplexes) naturally present in eukaryotic cells. Their primary role is to regulate gene expression via the RNA interference (RNAi) machinery. Extensive studies on their mode of action enable siRNA design optimization to utilize them as powerful lab tools and drugs for degradation of specific disease-causing mRNA.

Bulky substituents on the 2′-position of the ribose are generally not well tolerated, and chemists combine smaller modifications like 2′-F and 2′-Ome, which do not prevent RNA-induced silencing complex (RISC) assembly and function. Both modifications are approved in clinic and have demonstrated efficiency to enhance siRNA stability, increase affinity to the targeted mRNA, induce the duration of the effect, and reduce immunogenicity.

In contrast to PS-ASOs, chemical modifications are not sufficient to significantly increase the cellular uptake of siRNAs. Therefore, formulation and conjugation strategies are required for an efficient delivery into tissues/cells of interest. As of today, N-acetylgalactosamine (GalNAc) conjugation is the clinically dominant approach for siRNA delivery to hepatocytes. Given the wide therapeutic index and excellent safety profile of these compounds, four chemically modified GalNAc-siRNAs have been FDA-approved (Givosiran, Lumasiran, Inclisiran, and Vutrisiran) to treat liver-associated disorders.

Delivery Platforms to Enhance Oligonucleotide Therapeutics Intracellular Uptake

While PS-ASOs can spontaneously and productively enter cells without the need for a delivery system, they distribute mainly to the liver and kidneys after systemic administration, limiting their use to treat widespread diseases. On the other hand, siRNAs are too large and charged to enter cells unassisted and require a delivery system to target any tissues after systemic injection. While both lipid nanoparticles and GalNAc conjugates are clinically validated and approved delivery strategies for liver targets, efficient delivery of oligonucleotides beyond the liver and kidneys is a fundamental obstacle preventing their clinical utility.

Conjugation of a Glucagon-Like Peptide 1 Receptor (GLP1R) agonist to ASO enables efficient delivery to pancreatic beta cells, resulting in significant silencing of islet amyloid polypeptide in mice. The GLP1R conjugate shows not only cell selectivity, but also improves ASO productive uptake to intracellular compartments.

Lipid conjugation has emerged as a delivery platform for oligonucleotides for systemic administration. It has been demonstrated that the chemical structure of the lipids significantly impacts oligonucleotide clearance, lipoprotein binding, tissue distribution, and efficacy. Highly lipophilic lipids such as cholesterol tend to bind to low-density lipoprotein (LDL), driving oligonucleotide distribution to the liver. On the other hand, less lipophilic lipids such as docosahexaenoic acid (DHA) tend to bind to high-density lipoprotein (HDL) and accumulate in the kidneys. Even though the delivery to extra-hepatic tissues remains challenging, fatty acid conjugates including palmitic acid and docosanoic acid enable functional oligonucleotide delivery to the lungs, CNS, eyes, and muscles.

Antibody (Ab) conjugation is also applied to deliver both ASOs and siRNAs to extra-hepatic tissues. Conjugating an anti-transferrin receptor Ab (anti-TfR Ab) to PMO enables an increase in SMN2 mRNA level in the CNS when injected systematically to a spinal muscular atrophy mouse model. These data demonstrate that the cargo can cross the blood–brain barrier to deliver the drug efficiently. In addition, multiple PMOs linked to an anti-TfR Ab induce an increase in dystrophin expression in skeletal muscle in mdx mice. For siRNA delivery, the first example of antibody-mediated delivery was published in 2005 by the Liberman group against the HIV-1 capsid gene gag. Since then, many leading pharmaceutical companies such as Avidity, Dyne, Tallac, Denali, and Genna Bio have been developing Ab-conjugated siRNAs for the treatments of various muscle disorders. Avidity’s AOC 1001, an anti-TfR conjugated siRNA, has entered clinical trials to treat DM1 disease, demonstrating the value of such an approach. This compound shows >80% of mRNA silencing in muscles after a single-dose administration in mice. Furthermore, siRNAs-Ab conjugates are being used in some instances as complexes with cationic peptidic sequences, conjugated to the Ab itself. An Ab targeting a prostate specific membrane antigen, a marker found in most prostate cancers, complexes siRNAs through a protamine sequence (a 30–65 amino acids long sequence, with >50% of arginine). This system shows a significative inhibition in the proliferation and the growth of cancerous cells in a castration-resistant prostate cancer mouse model.

Nanocarriers represent an important class of oligonucleotide delivery systems. These nanotechnologies can be of many forms, going from liposomes [86], peptides, dendrimers, exosomes, spherical nucleic acids (SNAs), or DNA nanostructures, to name a few. One advantage of this strategy is the possibility to control and fine-tune the biophysical properties of the carrier to enhance oligonucleotide cellular uptake, intracellular trafficking, and endosomal escape. The control of the particle size is particularly interesting when considering oligonucleotide pharmacokinetic properties and elimination through renal clearance. Exosomes are also interesting as vehicles for oligonucleotide delivery, they are found endogenously and shown to be more resistant and less toxic compared to LNPs. They were successfully engineered and used to deliver large materials like single-guide RNAs and even plasmids.

Overcoming Endosomal Barrier to Maximize Oligonucleotide Therapeutic Activity

Both receptor and non-receptor targeted oligonucleotides are initially internalized by various endocytic mechanisms, depending upon cells and brought into the early endosomes (EEs). Oligonucleotides—whether non-conjugated, ligand-conjugated, or associated with nanocarriers—are gradually trafficked from the EEs to late endosomes (LEs) to downstream multivesicular bodies (MVBs) and finally to lysosome (LYs) for degradation. In this event, oligonucleotide therapeutics experience an environmental change in pH (7.4–4.5) as they travel along the endocytic pathway. In addition, while oligonucleotides are stable in endosomal compartments, small amounts can be recycled back to the extracellular space through exocytosis. It has been reported that only 1–2% of GalNAc-PS ASOs and less than 0.3% of GalNAc-siRNA conjugates escape endosomes to engage with the target in vivo. Even though ~99% of oligonucleotide therapeutics remain entrapped in endosomes or lysosomes, recent evidence suggests that trapped oligonucleotides serve as a depot, enabling slow oligonucleotide leakage over the course of time and thus long duration of the effect.

Strategies to Enhance Endosomal Escape

Several hypotheses on how drugs escape endosomes have been reported and summarized in a review by D. Pei and M. Buyanova. Briefly, endosomal escape can be triggered by either one or multiple simultaneous mechanistic events such as membrane fusion, pore formation, proton sponge effect, membrane destabilization, or vesicle budding and collapse, depending upon the chemical properties of different endosomal escape domains.

The endosomal pH is associated with the stage of endosomal maturation. It decreases progressively from the cell surface (pH ~7.4) to pH 6.5 in early endosomes and then, pH 5.5 in late endosomes. This pH change in the endocytic pathway can facilitate the endosomal escape of oligonucleotides into cytosol by using pH-sensitive scaffolds, which can interact with the endosomal membrane and disrupt it at an acidic pH. Taking advantage of pH differences between the extracellular matrix and endosomal compartments, a variety of endosomolytic agents such as endosomal buffering polymers, endosomolytic peptides, small molecules, ionizable lipids, and cationic liposomes have been developed and used in gene delivery.

Cleavable and Non-Cleavable Linkers in Oligonucleotide Conjugates

While a variety of deliver systems has been explored, direct conjugation of oligonucleotides to ligands—such as lipids, peptides, aptamers, antibodies, and sugars (e.g., GalNAc)—is emerging as the clinically dominant delivery strategy. Therefore, combining chemical modifications and conjugation approaches gives opportunities to maximize oligonucleotide delivery to the tissues of interest. It has been preferred to link conjugates at the termini of oligonucleotides to minimize base-pairing interference. In siRNA, the passenger strand is generally conjugated to the ligand, because, although the duplex is loaded into RISC complex, only the guide strand is functional for mRNA hybridisation and degradation.

Cleavable linkers are broadly classified into enzymatic, physicochemical, or chemically labile linkers depending on the conditions used for their cleavage. Unlike cleavable linkers, non-cleavable linkers do not have a particular weak bond in their structure that could be cleaved by enzymes, photo-irradiation, or change in pH.  The most used non-cleavable (i.e., amide, triazole, and maleimide linkages) and chemically cleavable (i.e., disulfide and pH-sensitive linkages) linkers in oligonucleotide therapeutics are described in more details in this section.