Antisense oligonucleotides: absorption, distribution, metabolism, and excretion

ABSTRACT

Introduction

Antisense oligonucleotides (ASOs) have emerged as a promising novel drug modality that aims to address unmet medical needs. A record of six ASO drugs have been approved since 2016, and more candidates are in clinical development. ASOs are the most advanced class within the RNA-based therapeutics field.

Areas covered

This review highlights the two major backbones that are currently used to build the most advanced ASO platforms – the phosphorodiamidate morpholino oligomers (PMOs) and the phosphorothioates (PSs). The absorption, distribution, metabolism, and excretion (ADME) properties of the PMO and PS platforms are discussed in detail.

Expert opinion

Understanding the ADME properties of existing ASOs can foster further improvement of this cutting-edge therapy, thereby enabling researchers to safely develop ASO drugs and enhancing their ability to innovate.

Abbreviations

2′-MOE, 2′-O-methoxyethyl; 2′PS, 2 modified PS; ADME, absorption, distribution, metabolism, and excretion; ASO, antisense oligonucleotide; AUC, area under the curve; BNA, bridged nucleic acid; CPP, cell-penetrating peptide; CMV, cytomegalovirus; CNS, central nervous system; CYP, cytochrome P; DDI, drug–drug interaction; DMD, Duchenne muscular dystrophy; FDA, Food and Drug Administration; GalNAc3, triantennary N-acetyl galactosamine; IT, intrathecal; IV, intravenous; LNA, locked nucleic acid; mRNA, messenger RNA; NA, not applicable; PBPK, physiologically based pharmacokinetics; PD, pharmacodynamic; PK, pharmacokinetic; PMO, phosphorodiamidate morpholino oligomer; PMOplus, PMOs with positionally specific positive molecular charges; PPMO, peptide-conjugated PMO; PS, phosphorothioate; SC, subcutaneous; siRNA, small-interfering RNA; SMA, spinal muscular atrophy.

KEYWORDS:

• Absorption
• distribution
• metabolism
• excretion
• antisense oligonucleotides
• phosphorothioate
• phosphorodiamidate morpholino oligomer
• peptide-conjugated phosphorodiamidate morpholino oligomers

1. Introduction

The field of oligonucleotide therapeutics includes double-stranded small-interfering RNA (siRNA; ~14 kDa), CRISPR-Cas9 single-guide RNAs (~200 kDa), self-replicating messenger RNA (mRNA) (700–7000 kDa), and single-stranded antisense oligonucleotides (ASOs) (3–15 kDa). Of these, ASOs are considered one of the most advanced RNA-based therapeutic modalities, as evidenced by them having the highest number of New Drug Application approvals [1].

ASOs are complementary nucleic acid fragments with a range length of 12–25 nucleotides designed to specifically hybridize to a complementary endogenous pre-mRNA or mRNA to modulate a biological function; they do this either by degrading the mRNA through enzymatic cleavage, altering pre-mRNA splicing to include or exclude specific introns/exons, or changing the function of regulatory RNA [2]. Other potential ASO mechanisms of action include inducing translational arrest by steric hindrance of ribosomal activity, interfering with pre-mRNA maturation by inhibiting splicing, destabilizing pre-mRNA in the nucleus, or correcting cryptic splice sites [3–5]. The sequence of the ASO specifically targets the genetic abnormalities underlying disease pathology.

To date, 10 ASO drugs have been approved by the US Food and Drug Administration (FDA) for the treatment of genetic diseases, such as Duchenne muscular dystrophy (DMD), spinal muscular atrophy (SMA), and familial amyloid polyneuropathy (Table 1) [5–8].

Table 1. List of FDA-approved ASO drugs

Drug name	Brand name	Developer	FDA approval year	Therapeutic indication	Route of administration	Length (nucleotide)
Fomivirsen	Vitravene®	ISIS/Ionis Pharmaceuticals	1998	CMV retinitis in AIDS patients	Intravitreal injection	21 [9]
Pegaptanib	Macugen®	OSI Pharmaceuticals	2004	Treatment of neovascular age-related macular degeneration	Intravitreal injection	28 [10]
Mipomersen	Kynamro®	Ionis Pharmaceuticals	2013	Familial hypercholesterolemia	SC injection	20 [11]
Eteplirsen	Exondys 51™	Sarepta Therapeutics	2016	Treatment of Duchenne muscular dystrophy	IV infusion	30 [22]
Nusinersen	Spinraza®	Ionis Pharmaceuticals/Biogen	2016	Management of spinal muscular atrophy	IT injection	18 [77]
Defibrotide	Defitelio®	Jazz Pharmaceuticals	2016	Treatment of veno-occlusive disease in the liver	IV infusion	NA [12]
Inotersen	Tegsedi®	Akcea Therapeutics and Ionis Pharmaceuticals	2018	Treatment of nerve damage in adults with hereditary transthyretin-mediated amyloidosis	SC injection	20 [13]
Golodirsen	Vyondys 53™	Sarepta Therapeutics	2019	Treatment of Duchenne muscular dystrophy	IV infusion	25 [23]
Viltolarsen	Viltepso®	NS Pharma	2020	Treatment of Duchenne muscular dystrophy	IV infusion	21 [25]
Casimersen	Amondys 45™	Sarepta Therapeutics	2021	Treatment of Duchenne muscular dystrophy	IV infusion	22 [24]

ASO, antisense oligonucleotide; CMV, cytomegalovirus; FDA, US Food and Drug Administration; IT, intrathecal; IV, intravenous;

NA, not applicable; SC, subcutaneous.

2. Engineering chemically modified ASOs for therapeutic applications

Stephenson and Zamecnik were the first to use ASOs for therapeutic purposes and they demonstrated that DNA-based ASOs can inhibit viral replication in vitro [14]. These findings were not sustained in vivo due to degradation of the phosphodiester backbone by nucleases and poor protein binding that prohibited efficient tissue distribution. Consequently, extensive research efforts were directed toward making chemical modifications to the ASO that would improve stability, tissue targeting and uptake, pharmacokinetics, and pharmacodynamics.

Modification of the phosphodiester backbone led to new chemistries, such as phosphoramidate, phosphorothioate, boranophosphate, methylphosphonate, and phosphonate analogs (Figure 1a), whereas the ribose unit was modified with 2′-O-methyl, 2′-fluoro, or 4′-thioribosyl, or substituted with a morpholino ring (Figure 1b). Substitution of DNA/RNA nucleobases with unnatural or modified nucleobases, such as hypoxanthine, 7-deaza-G, 2,6-diamino-purine, 4-thiouracil, and difluorotoluene, was also evaluated (Figure 1c) for improved metabolic stability and enhanced hybridization affinity [3]. The two major ASO platforms with clinical utility are the phosphorodiamidate morpholino oligomers (PMOs) and the phosphorothioates (PSs) (Table 2).

Figure 1. Examples of chemical modifications on different parts of the antisense oligonucleotide (ASO): (a) the inter-nucleotide linkages; (b) the ring that connects the ASO backbone to the nucleobase; and (c) the nucleobases that form the complementary Watson–Crick DNA base pairing with target RNA

Table 2. Summary of different chemical modifications of ASO therapeutics

ASO	Modification(s)	Charge	Metabolizing enzymes	Mode of action	Example
Phosphodiester	-	Negative	Nucleases phosphatase	Steric hindrance RNase H cleavage	Defibrotide
PMO	Replace backbone oxygen with dimethyl amine Replace ribose with morpholino	Uncharged	-	Steric hindrance	Eteplirsen
PPMO	Peptide-conjugated PMO	Positive	Proteases	Steric hindrance	SRP-5051
PMOplus	Add positionally specific positive molecular charges	Positive	-	Steric hindrance	AVI-7288
PS	Replace backbone oxygen with sulfur	Negative	Nucleases	Steric hindrance RNase H cleavage	Fomivirsen
2′ modified PS	Add substituents to the 2′ carbon	Negative	Nucleases	Steric hindrance RNase H cleavage	Nusinersen
BNA/LNA	Add bridging linker to between the 2′C and 4′C	Negative	-	Steric hindrance	AZD9150

ASO, antisense oligonucleotide; BNA, bridged nucleic acid; LNA, locked nucleic acid; PMO, phosphorodiamidate morpholino oligomer; PMOplus, PMOs with positionally specific positive molecular charges; PPMO, peptide-conjugated PMO; PS, phosphorothioate.

2.1. Phosphorodiamidate morpholino oligomers

Substituting the phosphodiester backbone with a carbamate or phosphorodiamidate backbone and replacing the ribose sugar with a morpholino subunit resulted in a new class of ASOs, known as PMOs (Figure 2a) [15–19]. PMOs are uncharged, biologically active antisense molecules that vary in length and have superior metabolic stability, sequence specificity, and absence of off-target effects [17,20,21].

Figure 2. (a) Examples of backbone with carbamate, carbamate morpholino, and phosphorodiamidate morpholino backbones. (b) The structure of the most advanced phosphorodiamidate morpholino oligomer (PMO) prototypes; PMOs, peptide-conjugated PMOs (PPMOs), and PMOs having positionally specific positive molecular charges (PMOplus)

Eteplirsen, indicated for treatment of DMD, is the first example of an FDA-approved PMO for the treatment of DMD. Golodirsen, viltolarsen, and casimersen have followed, targeting different mutations of the DMD gene [22–25]. Eteplirsen, casimersen, vitolarsen, and golodirsen have 30, 22, 21, and 25 morpholino subunits, respectively, which are connected together by dimethylamino phosphorodiamidate inter-subunit linkages.

2.2. Peptide-conjugated phosphorodiamidate morpholino oligomers

Peptide-conjugated phosphorodiamidate morpholino oligomers (PPMOs) are a next-generation PMO-based platform consisting of a cell-penetrating peptide (CPP) that is covalently linked to the PMO sequence, thus triggering enhanced tissue uptake mediated by electrostatic interactions between cationic charges of the CPP and anionic charges of the cell-surface proteoglycans during endocytosis (Figure 2b) [26–29]. SRP-5051 is an investigational PPMO that is currently in phase II clinical trial for DMD (ClinicalTrials.gov Identifier: NCT03675126).

2.3. PMOs with positionally specific positive molecular charges

PMOs with positionally specific positive molecular charges (PMOplus) is another PMO chemistry that contains positively charged piperazinyl groups at selected PMO inter-nucleoside linkages (Figure 2b). AVI-7288 is an example of a PMOplus oligomer that was investigated for the treatment of Marburg virus, and which acts to interfere with viral replication [30,31].

2.4. First- and second-generation phosphorothioates

The first-generation PS-based ASO that demonstrated improved metabolic stability and relatively high protein binding was built by replacing an oxygen of a phosphodiester bond with a sulfur atom (Figure 3) [32,33]. Further, substituting two non-bridging phosphate oxygen atoms by two sulfur atoms led to the development of a second-generation PS-based ASO with a 1000-fold higher affinity toward target RNA with no reported off-target effects [34]. Stability of PS can also depend on the stereochemistry. It has been reported that the Sp stereoisomer of PS is more stable than the Rp stereoisomer. This difference in stability has been attributed to the orientation of the sulfur group in the Sp stereoisomer that results in displacement of the metal ion in the nuclease active site [35,36]. Fomivirsen, which consists of 21 nucleotides, is one of the first examples of an FDA-approved ASO in the PS platform for the treatment of cytomegalovirus retinitis in immunocompromised patients [37,38].

Figure 3. Examples of the most advanced chemistry backbones in the antisense oligonucleotide field. 2′-MOE, 2′-O-methoxyethyl; BNA, bridged nucleic acid; LNA, locked nucleic acid; PMO, phosphorodiamidate morpholino oligomer; PS, phosphorothioate

2.5. 2′-Modified phosphorothioate

Modifications of the 2′ position on the ribose sugar of PS led to 2′-O-methyl and the 2′-O-methoxyethyl (2′-MOE)–containing ASOs that further improved their plasma pharmacokinetic (PK) and safety profile (Figure 3) [39–42]. However, in order to increase their ability to recruit RNase H to degrade their target, a chimeric ASO was developed that consisted of adding gapmer 2′ nucleotide modifications on the terminal end of the 2ʹ-OM or 2ʹ-MOE ASO [43–45].

Nusinersen and mipomersen are 2′ modified PS-based ASOs approved for the treatment of SMA and homozygous familial hypercholesterolemia, respectively.

2.6. Bridged nucleic acid

The bridged nucleic acid (BNA) platform, also known as locked nucleic acid, improves the PS-based ASO properties through ‘locking’ the conformation of the ribose by coupling the 2′ substituent to the 4′ carbon on the ribose (Figure 3) [46]. BNAs can be added to the two ends of a specific gapmer, thus facilitating the degradation of the target RNA while protecting the ASO from nuclease degradation [45]. AZD9150 is a clinical candidate that utilizes BNA chemistry [47].

3. ADME characteristics of ASO

Understanding ADME properties allows the generation of preclinical PK/pharmacodynamic (PD) predictions to support the determination of efficacious dose levels and dosing frequency in clinical studies.

3.1. Absorption and distribution

Absorption of ASO-based therapeutics depends on their different physiochemical properties, including ionization, acid dissociation constant (pKa), hydrophobicity, and the route of delivery. Drug distribution depends on the route of delivery, free-drug concentration, tissue blood perfusion, tissue binding (e.g. because of lipid content), regional pH, and permeability of cell membranes [48].

First- and second-generation PS oligomers, which have negatively charged backbones, can exhibit a high binding capacity (>85%) to plasma proteins, specifically albumin [49]. Mipomersen (a PS-based ASO) shows high protein binding in humans (95%) and mice (85%). PMO-based ASOs, on the other hand, are uncharged and have low protein binding in humans (6–17%), thus limiting their tissue absorption [50]. Eteplirsen, a PMO, has human plasma protein binding of between 6.1% and 16.5% [22].

Most ASOs are administered intravenously (IV), to maximize bioavailability; this leads to rapid distribution to highly vascularized organs (liver, kidney, and spleen), but slower distribution to tissues such as the heart, muscle, and lungs, thereby necessitating more frequent dosing. In clinical trials, weekly IV infusions of eteplirsen exhibited an average volume of distribution of 601 ml/kg at doses of 30 mg/kg, suggesting higher distribution in peripheral and muscle tissues [22]. Peak plasma concentrations occurred near the end of single or multiple IV infusions and dose-proportionality and linearity in PK properties were observed following multiple doses in the ranges of 0.5–20 mg/kg/week and 30–50 mg/kg/week in efficacy trials [51]. In order to improve organ-specific distribution, coupling ASOs with fatty acids or peptides is currently under way [27,52]. A clinically tested conjugate is the triantennary N-acetyl galactosamine (GalNAc3) conjugates of ASO [53]. Such conjugates have greatly enhanced targeted delivery to hepatocytes, which has resulted in significant improvements in potency [54]. GalNAc3-ASOs undergo active uptake by hepatocytes through the asialoglycoprotein receptor [55]. Another successful method for enhancing penetration of ASOs has been through conjugation with CPPs. The positive charge of the peptide facilitates coupling to the uncharged PMO ASO. It has been suggested that CPPs enhance PPMO entry to the cell through an endocytosis mechanism that involves interactions between cationic charges of the CPP and the anionic charges of the proteoglycans on the cell surface [56]. CPPs have been used to improve the cellular uptake of PMOs into muscle tissues [57].

Subcutaneous (SC) injections have been shown to be a feasible alternative to IV infusions in both PMO-based and PS-based ASOs. ASO distribution half-lives are usually longer after SC administration compared with IV injection, possibly due to gradual absorption from the site of injection. A study investigating plasma exposure in non-human primates dosed with eteplirsen 320 mg/kg via IV bolus or SC administration showed that the plasma PK with each route of administration was comparable, with 100% bioavailability after the SC administration and peak concentrations at approximately 7 h [58]. Mipomersen has also been shown to be fully absorbed from the SC injection site in monkeys, with bioavailability of 100% and a peak plasma concentration within 3–4 h [49].

To target sites not accessible by IV or SC administration, direct ASO delivery has been shown to be successful. In the case of fomivirsen, local delivery into the eye via intravitreal injection requires a very small amount of drug and allows direct distribution to the retina [59,60].

Intrathecal administration has also been researched to circumvent the physiological blood–brain barrier and allow ASO delivery to the central nervous system (CNS). Despite its invasive nature, intrathecal administration has been shown to be effective in increasing ASO bioavailability in the brain and spinal cord with reduced systemic exposure when compared to the IV and SC routes of administration. Intrathecal dosing of nusinersen has demonstrated the potential use of ASOs in the CNS field and opened the door for targeting new indications, such as Huntington’s disease and amyotrophic lateral sclerosis [2,61]. A single intrathecal bolus injection in adult cynomolgus monkeys resulted in widespread distribution throughout the spinal cord and accumulation to levels predicted to be pharmacologically active. Moreover, tissue analysis 7 days post dose showed that the accumulation of nusinersen in CNS tissues was dose-dependent, with greater accumulation in the spinal cord and cortex. It was shown that even at the lowest dose (1 mg/kg), the drug levels at the target tissue were three- to eight-fold higher than the concentration that gives half-maximal response (~1 µg/g) [62]. The localized intrathecal bolus dose resulted in better drug distribution throughout the CNS compared with IV administration.

Intranasal delivery has also been explored as a mode of ASO delivery for CNS indications. It has been suggested that after intranasal administration, ASO molecules can be transported along the olfactory and trigeminal nerve pathway and the rostral migratory stream [63,64]. CNS delivery may be further enhanced using a CPP. This approach has been successful in siRNAs where a CPP conjugated to a glycol–polycaprolactone copolymer was able to deliver siRNAs to the brain via the intranasal route [65].

Oral administration of ASOs has been challenging due to limited gastrointestinal absorption. This can vary depending on the chemistry, charge, and size of the ASO. The bioavailability of PS-based ASO ISIS 14725 was approximately 5.5% in rats after administration of 30 mg/kg intraduodenally [66], whereas oral bioavailability of PMO-based AVI-4472 was reported to be between 41% and 79% in rats [67]. However, there have been no follow-up reports to assess the oral delivery of PMOs in humans. Defibrotide, a mixture of ASOs with the natural phosphodiester backbone, was reported to show bioavailability in rats in the range of 58–71% [68]. Despite the reported high bioavailability of defibrotide, it is clinically administered as a 2-h IV infusion for the treatment of adult and pediatric patients with hepatic veno-occlusive disease [69].

Once absorbed, ASOs enter the systemic circulation and bind reversibly to proteins and lipids in plasma and tissues. They next enter the cytoplasm via diffusion [70] or endocytosis (Figure 4) [71], escape endocytic vesicles, and, depending on whether their target is in the cytoplasm or nucleus, may need to further translocate into the nucleus at sufficient concentrations to produce the desired biological response. Once inside the cells, ASOs can exhibit long half-lives (1–4 weeks) and prolonged biological activity [54,72].

Figure 4. The intracellular distribution of antisense oligonucleotides (ASOs) and the subsequent target engagement for mipomersen, eteplirsen, and nusinersen. Note that the three drugs have different target cells and tissues (mipomersen-all tissues aside from brain, eteplirsen-muscle, nusinersen-brain. (a) ASOs with different chemistries interact with the cell membrane and enter the cells through the endocytosis process. (b) ASOs can escape endocytosis and the phosphorothioate backbone may bind to target messenger RNA (mRNA) and facilitate downregulation through RNase H recruitment (c), or the phosphorothioate may be metabolized by nucleases (d). (e) Stable ASOs can distribute to the nucleus to bind to the pre-mRNA and correct transcription through (f) exon exclusion with eteplirsen or (h) intron inclusion with nusinersen; then, (g) and (i) are post-transcriptional modifications of spliceosomal RNAs

3.2. Metabolism

ASOs are metabolized by nucleases ubiquitously expressed by cells in most tissues, but their metabolic stability varies among the different platforms and species, and primarily depends on the chemistry backbone (Table 2). At therapeutic doses, modified nucleotides liberated from ASO metabolism have not been shown to pose a risk of genotoxicity [73,74]. Whereas exonuclease cuts the terminal sites of the ASO (recognizes 3′-hydroxy group for efficient cleavage), endonuclease cuts the internal part of the ASO, as it recognizes the 2′-hydroxy group on the ribose [75]. ASOs lacking ribose modifications at their 3′ and 5′ ends are primarily degraded by exonucleases generating 3′ or 5′ shortened fragments, which may still possess antisense activity, but also by endonucleases in tissues. Defibrotide is rapidly metabolized by plasma exonucleases and multiple DNA degradation enzymes to release the free 2′-deoxyribose sugar, purine, and pyrimidine bases. As a result, defibrotide exhibits a half-life of 14 min after a 2-h IV infusion [76]. Other PS-based ASOs, such as the 2′ modified PS and gapmers, can also be metabolized by endo- and exonucleases, but stability is improved compared with the first-generation PS-based ASO, with a prolonged terminal half‐life ranging from several days to weeks. Mipomersen and nusinersen are second-generation PS-based ASOs and their metabolism occurs at the phosphorothioate backbone via nucleases [77]. Mipomersen metabolism is mediated by endo- and exonucleases to form shortened sequence metabolites that no longer retain pharmacological activity [49]. Metabolism of nusinersen delivered intrathecally to the brain is mediated by exonuclease 3′- and 5′-mediated hydrolysis that cleaves the terminal nucleotides, resulting in the production of shortened ASO metabolites. Because of the lower metabolic activity of the brain, nusinersen was detected in the brain 1 year after administration, reflecting its long half-life in the cerebrospinal fluid (135–177 days). Following intravitreal administration, the PS-based ASO fomivirsen is slowly cleared with a half-life of approximately 55 h in humans. Preclinical studies demonstrated that fomivirsen distributed to the retina was slowly metabolized by exonuclease digestion [60]. In general, 2′ modified PS-based ASOs protected at the 3′ and 5′ ends from exonuclease degradation by chemical modifications are initially metabolized by endonucleases in tissues, leading to short fragments, which may be further degraded by exonucleases. Metabolites of these ASOs are too short to possess antisense activity [78].

PMOs are generally very stable in the presence of nucleases. For example, eteplirsen did not form any metabolites when incubated with plasma and liver subcellular fractions [22]. For PPMOs, however, the CPP sequence can introduce a metabolic soft spot that is susceptible to degradation by ubiquitous proteolytic enzymes. When a PMO and corresponding PPMO were incubated separately with tissue lysates or plasma, the PMO was shown to be metabolically stable [26]. However, the CPP portion of the PPMO was degraded more rapidly in the plasma incubation than in the tissue lysate incubation. The CPP portion of the PPMO was also degraded in the presence of tissue lysates with longer incubation. Interestingly, this PPMO was more stable in the liver lysates than in the heart and kidney lysates, which may suggest higher protease activity in the mouse kidney and heart compared with the liver. Metabolism of the liver and kidney occurred between adjacent L-arginine and 6-aminohexanoic acid residues of the CPP with mass-to-charge ratio peaks corresponding with XB-(RXR) and XB(RXR)2 (Figure 5a) [26].

Figure 5. (a) Peptide-conjugated phosphorodiamidate morpholino oligomers (PPMOs) can be metabolized by proteolytic enzymes to form active metabolites that result from sequential metabolism of the peptide sequence. The PPMO portion is metabolically stable. (b) Example of different 3H-ION-681257 metabolites in rats that form the metabolism of the triantennary N-acetyl galactosamine conjugate

Metabolism of other conjugated ASOs has also been evaluated. In the case of 3H-ION-681257, a PS-based ASO conjugated to GalNAc3, extensive distribution to the liver was observed following administration to rats [79]. The identified major metabolic pathway was oxidation to form 14 linker-associated metabolites on GalNAc3 and free ASO. Biliary-excreted metabolites were the same as the renally excreted metabolites, except for the oxidized linear and cyclic linker metabolites. 3 H-ION-681257 was metabolized by N-acetyl-β-glucosaminidase, deoxyribonuclease II, alkaline phosphatase, and alcohol/aldehyde dehydrogenases. Unconjugated PS-based ASOs were slowly metabolized to chain-shortened oligonucleotide metabolites by nucleases and eliminated in urine. The three major metabolites that were detected in monkeys are shown in Figure 5b [79].

Consistent with their physicochemical properties such as their large molecular weight/size and hydrophilicity, ASOs are not CYP450 substrates; therefore, no CYP-mediated metabolites have been reported for any of the approved ASOs. Moreover, owing to their low affinity toward CYP enzymes, ASOs have not been shown to result in drug–drug interactions (DDIs) with other small molecules. This could be explained by the uptake mechanism of ASOs―as ASOs enter cells through endocytosis, there is limited free intracellular ASO interaction with cytosolic CYP enzymes, which are located within the endoplasmic reticulum [80].

Several clinical trials aimed to assess the potential DDIs of ASOs with warfarin (CYP2C9/3A4), simvastatin (CYP3A4), ezetimibe (multiple uridine diphosphate-glucuronosyltransferases), rosiglitazone (CYP2C8/2C9), glipizide (CYP2C8/2C9), metformin (renal), cisplatin (renal), and gemcitabine (nucleoside kinases), and none of these investigations suggested DDI risk [81–84].

A recent study looking at ISIS 304801, ISIS 420915, ISIS 681257, and ISIS 396442 demonstrated that none of these tested ASOs showed potential inhibition of CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, or CYP3A4 [85]. In addition, none of the tested compounds showed induction of CYP1A2, CYP2B6, or CYP3A4 at the enzyme activity level or mRNA level in cryopreserved primary human hepatocytes. These data are consistent with the previously reported lack of interaction between ASOs and CYP enzymes [85].

3.3. Excretion

Owing to their charge, polarity, and hydrophilicity, ASOs are primarily excreted in urine or feces as unchanged drugs, or are metabolized in the form of cleaved short fragments [86]. However, the extent of excretion in urine or feces depends on the ASO chemistry, route of administration, animal species, and conjugation. The extent of protein binding affects elimination. Binding to circulating proteins spares ASOs from glomerular filtration and slows renal clearance (the major route for clearance of ASOs). It has been suggested that the low protein binding of PMO contributes to its high renal clearance and, subsequently, short plasma half-life [26].

In healthy volunteers with normal renal function, approximately 10–14% of defibrotide (natural phosphodiester backbone) is excreted unchanged in urine, with 19% excreted in feces. However, in patients with impaired renal function, the area under the curve (AUC) increases by 50–60%. Kidney-impaired patients require special monitoring and dose adjustments [87].

The role of protein-binding in excretion is also evident with mipomersen, a second-generation PS-based ASO. The negatively charged 2′ modified PS is bound to albumin, which protects the drug from elimination via renal filtration. After a 25-mg/kg SC administration in mice, 23% of mipomersen was excreted in urine within 24 h compared with 1% after a 5-mg/kg SC administration. Mipomersen demonstrated higher in vivo clearance in mice (plasma clearance, 674 ml/h/kg; half-life, 0.33 h) than in humans (plasma clearance, 40.9 ml/h/kg; half-life, 1.26 h), which was attributed to the lower protein binding in mice [49]. In another study, 3 H-mipomersen (3 H-ISIS 301012) was administered in rats as a 5-mg/kg IV bolus. Twenty-six percent of total radioactivity was recovered in urine compared with 4.2% in feces within 2 weeks of administration [88].

3 H-ION-681257 is a second-generation PS ASO conjugated to a THA-C6-triantennary GalNAc3 for hepatocyte-specific delivery via the asialoglycoprotein receptor. Owing to the specific accumulation in the liver, 3 H-ION-681257 was rapidly metabolized and excreted with 26% and 71% of radioactivity recovered in urine and feces, respectively, within 48 h post dose. This could have been driven by the high metabolism of the GalNAc side chains compared to the PS ASO backbone.

Although renal clearance is a major route of elimination for most ASOs, nusinersen (a second-generation PS ASO) did not undergo significant renal elimination. Approximately 0.5% of unchanged drug was detected in urine on day 85 following the third dose; since its metabolites were not tracked; however, it cannot be excluded that renal elimination was the main route of excretion [77]. This low renal elimination is most likely owing to the prolonged tissue half-life (over 100 days after administration in the spinal cord and brain of monkeys) and slow elimination phase from central nervous system tissues to the systemic circulation [62].

Clinical trials have shown that eteplirsen PMO is excreted primarily unchanged in urine, with 69.4% of the dose excreted after a 50-mg/kg IV infusion. The total clearance of eteplirsen after a 50-mg/kg IV infusion was 5.3 ml/min/kg, which is considered low. The renal clearance of eteplirsen was 3.9 ml/min/kg at the same dose, accounting for two-thirds of the total systemic clearance [22].

AVI-6002 and AVI-6003 are combination therapies previously under evaluation for post-exposure prophylaxis of Ebola virus and Marburg virus, respectively. AVI-6002 is a combination consisting of AVI-7537 and AVI-7539, and AVI-6003 is a combination consisting of AVI-7287 and AVI-7288. The four components belong to the PMOplus platform that contains positionally specific positive molecular charges along the backbone. When the four components were administered to healthy volunteers during clinical trials, the PMOplus drugs were all excreted in urine within the first 24 h. Dose-dependent renal clearance was observed for the four drugs: 44% of AVI-7537, 31% of AVI-7539, 39% of AVI-7287, and 52% of AVI-7288 were excreted unchanged [89].

3.4. ASO pharmacokinetic translation

Plasma PK exposure for ASOs within the same platform is independent of sequence and, in many cases, similar across species. When three PMO clinical candidates were dosed in monkeys, they each exhibited similar plasma exposure [90]. PS oligomers have also been shown to exhibit very similar plasma profiles when dosed in different animals. However, there are only a few reports related to prediction of human doses. Mahmood [91] evaluated the predictive performance of interspecies scaling of oligonucleotides to predict clearance and volume of distribution at steady state in humans. Analysis of the literature on eight oligonucleotides from the phosphorothioate platform (PS and 2′ modified PS) concluded that it is possible to predict human PK parameters for these classes of oligonucleotides with reasonable accuracy using the principles of allometry and at least three animal species. In addition, Mahmood reported that single-species scaling failed in predicting the human PK, and this approach should therefore not be considered.

In contrast, Yu et al. reported that human PK parameters can be predicted based on a single-species analysis, such as the monkey or mouse [92]. They analyzed nine different 2′-MOE ASOs and concluded that monkeys can predict human plasma exposure based on body weight-adjusted dose ratios. For instance, mipomersen showed comparable dose-normalized AUCs of 10.3 and 13.3 for monkeys and humans, respectively. Interestingly, eight of the nine compounds exhibited similar normalized AUC values between humans and monkeys within the two-fold acceptance criteria. However, when the same comparison was made between mice and humans, the mouse doses needed to be multiplied by five to give a reasonable human exposure at steady state. There is no similar comparison in the literature for the PMO platform to determine whether the observed trend is specific to the 2′-MOE platform or whether it could be extended across other platforms. Much work is still needed before recommendations can be made regarding the optimal approach for human-dose predictions.

4. Conclusions and future directions

ASO therapeutics have considerable potential to contribute significantly to the future of medicine, as evidenced by the clinical benefit observed in SMA, familial hypercholesterolemia, and DMD indications.

The clinical development of new ASOs for the treatment of a range of human diseases is progressing at a rapid pace. Knowledge of ASO ADME properties is essential for drug development and the safe use of medicines. Careful consideration and the optimization of factors such as ADME, safety, efficacy, target selection, and delivery technologies in clinical trial design are critical to the future of successful drug production.

5. Expert opinion

ASO therapeutics have demonstrated potential to treat many genetic diseases that were once considered untreatable or for which there were limited treatment options. The field has advanced over the past 3 decades and has seen the approval of six new products in the past 5 years. This became possible due to advancements in chemistry that have improved the metabolic stability of the ASO and allowed it to reach target sites and elicit therapeutic benefit. Challenges remain in ASO drug development, however, and these are centered around achieving delivery to specific tissues, circumventing endosomal trapping, and understanding preclinical PK/PD translation for human dose predictions.

Over the next few years, research will continue toward the development of new ASO conjugates that will facilitate the targeted delivery of ASOs to specific tissues beyond the liver. The chemistry of all ASO platforms allows the conjugation of ASOs to lipids, peptides, or antibodies, to enhance cellular uptake and promote cell-specific targeting. Furthermore, ASOs may utilize nanocarriers such as lipid and polymeric nanoparticles, spherical nucleic acids, liposomes, and exosomes to improve the target tissue-specific delivery and increase the therapeutic index. Development of conjugates that will facilitate ASO uptake across the blood–brain barrier could transform the treatment landscape for CNS indications.

Once delivered to the target tissue, ASOs need to escape the endosomes and lysosomes to engage with the intracellular target. The mechanism of endosomal/lysosomal escape is currently unknown and only a small percentage of ASOs escape the endosomes. Therefore, understanding the mechanism of endosomal/lysosomal escape and modifying ASO chemistry to increase endosomal/lysosomal release will potentially impact the ASO dose levels and the overall therapeutic index.

Finally, the scarcity of published reports on the preclinical translational aspect of ASOs hampers our ability to translate PK/PD behavior from preclinical models into humans. Part of the challenge is that plasma PK for ASOs does not always correlate with the PD due to the indirect PK/PD relationship. Physiologically based PK (PBPK)/PD modeling could be used as an alternative approach. PBPK modeling provides a mechanistic approach to study and predict the PK and PD of drugs at the target tissue based on physiologic and anatomic characteristics, such as biological processes, organ function, tissue volume, and blood flow. This modeling approach has been successfully applied to other modalities and it is expected to gain more traction within the ASO field.

Focusing on the aforementioned areas of research will help advance the ASO platform further. As the ASO platform continues to evolve, its application in more therapeutic areas will become important for expanding its utility. New technologies, such as artificial intelligence, advanced computer modeling, and machine learning, may be useful in enabling the identification of new genetic targets that could be amenable to ASO-targeting therapy. Such methodologies could also help in providing mechanistic insights on drug PK or in simulating head-to-head comparisons of ASOs to generate valuable information that would not otherwise be as easily obtained.

Article highlights

• The two major chemistry backbones that are widely used to build the most advanced ASO drugs are the phosphorodiamidate morpholino oligomers (PMOs) and the phosphorothioates (PSs).

• PMOs and PS oligomers demonstrate superior metabolic stability, sequence specificity, and absence of off-target effects that allow specific RNA targeting in an increasing number of clinical therapeutic areas.

• ASOs can be clinically administered via several routes of administration, such as intravenous, subcutaneous, intravitreal, and intrathecal injections, depending on the target tissues.

• Poor tissue uptake presents one of the biggest challenges in the ASO field; ASO delivery has become an active area of research, focusing on optimizing ASO delivery to target tissues and increasing the cellular uptake.

• ASOs are not CYP substrates and no drug–drug interactions have been reported for any clinical ASO drug.

• PK/PD translation of ASOs from preclinical species into humans is a critical area of research and more data are needed.

This box summarizes key points contained in the article.

Declaration of interest

M Shadid is an employee of Sarepta Therapeutics Inc. M Badawi is an employee of AbbVie. Editorial support, provided by Joseph DeSisto Alling of Eloquent Scientific Solutions, was utilized in the production of this manuscript and funded by Sarepta. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Acknowledgments

Django Andrews, Marc Evans, John Haberman, Aaron Novack, and Ihor Sehinovych of Sarepta Therapeutics, Inc. contributed their review, suggestions, and comments.

Additional information

Funding

This review was funded by Sarepta Therapeutics Inc.