Nucleic acid therapeutics in neurodegenerative disorders: emerging technologies and clinical landscape

Abstract

Neurodegenerative disorders have diverse etiologies from monogenic mechanisms such as toxic gain-of-function mutation in subsets of amyotrophic lateral sclerosis (ALS) to multifactorial cascades exemplified by Alzheimer’s disease (AD), which feature proteostasis failure, neuroinflammation, mitochondrial dysfunction, or age-associated cellular stress. While small molecules can modulate downstream pathways, nucleic acid therapeutics (NATs) target defined RNA sequences to directly reduce pathogenic gene expression, correct or shift isoform usage via splice modulation, and tune disease-relevant networks, positioning them as promising therapeutic modalities for neurodegenerative disorders. Recent clinical applications of NATs in the central nervous system (CNS) have made significant progress by optimized chemistries, improved tolerability, and established intrathecal administration routes, supporting sustained target engagement in relevant neural and glial compartments. This review summarizes three major NAT platforms—antisense oligonucleotides, small interfering RNAs, and aptamers—by highlighting their mechanisms of action, design principles to enhance their biodistribution, durability, and safety, and emerging novel delivery strategies for CNS targeting. We further discuss emerging translational opportunities across ALS, AD, polyglutamine disorders, prion diseases, and synucleinopathies by integrating disease biology with therapeutic rationale, and by summarizing regulatory approvals, ongoing clinical trials, and next generation NATs in preclinical development.

Introduction

The first antisense oligonucleotide (ASO) drug, fomivirsen (Vitravene), a 21 mer phosphorothioate-modified oligodeoxynucleotide, was approved by the United States Food and Drug Administration (FDA) in 1998 for the local treatment of cytomegalovirus retinitis to help AIDS patients [1]. This approval laid the groundwork for the field of nucleic acid therapeutics (NATs), which has advanced rapidly: as of July 2026, the FDA had approved 24 oligonucleotide therapeutics, and the broader nucleic acid therapeutic landscape now includes more than 29 approved products overall, with hundreds of ongoing clinical trials worldwide. Meanwhile, the first central nervous system (CNS) directed NAT, nusinersen (Spinraza) for spinal muscular atrophy (SMA), was approved in 2016, a milestone that remains relatively recent in the field. Applications of NATs in CNS disorders have expanded since then from rare diseases to more common conditions, including Alzheimer’s disease (AD), the most prevalent form of dementia.

We are cautiously optimistic about a new era of RNA therapeutics for CNS disorders that, while most of them are currently not curative, may slow disease progression.

In this review, we summarize ongoing efforts to develop nucleic acid therapies for CNS disorders, focusing on (i) mechanisms of action; (ii) recent technical advances, with particular emphasis on drug durability, tissue distribution, and CNS penetration; and (iii) clinically and preclinically validated therapies. Within the broader NAT field, this review focuses specifically on ASOs, small interfering RNAs (siRNAs), and aptamers. Because of the unique challenges of CNS therapeutics, these points may be critical determinants of whether NATs for CNS disorders can become increasingly successful as effective therapies in the future.

Programmable nucleic acid therapeutics and the role of chemical modifications

Nucleic acid therapeutics represent a rapidly expanding class of medicines that leverage programmable nucleic acid sequences or structures to target disease-causing genes, RNA transcripts, and associated molecular targets. In contrast to traditional small-molecule drugs, which act via fixed chemical structures to modulate protein activity, RNA-based therapeutics can be reprogrammable on demand, allowing precise, sequence-defined targeting of virtually any gene [2–4]. By altering nucleotide sequence rather than chemical scaffold, NATs can be readily redesigned to address disease-causing transcripts, including those driving neurodegenerative disorders characterized by aberrant gene expression, toxic protein accumulation, and limited druggability by conventional pharmacological approaches. Beyond target selection, the in vivo pharmacokinetics (PK) and pharmacodynamics (PD) of NATs are largely dictated by their chemical composition, which governs potency, delivery efficiency, tissue distribution, and durability within target cells. Unmodified nucleic acid components, such as synthetic ribonucleic acid (RNA) oligonucleotides, are intrinsically unstable in vivo and undergo rapid enzymatic degradation, often exceeding 50% within minutes, thereby severely limiting their therapeutic utility. Moreover, nucleic acids including phosphorothioate (PS)-modified DNA and double-stranded RNA, can activate innate immune pathways, resulting in unintended immunostimulatory and off-target effects [5–8].

These challenges have been substantially mitigated through the development of chemical modifications to ribose sugars, phosphate backbones, and nucleobases, which enhance metabolic stability, reduce immunogenicity, and improve overall PK/PD properties. Because several different oligonucleotide platforms engage distinct endogenous gene-regulatory protein machineries, stabilizing chemistries must be carefully selected and optimized to maintain compatibility with their respective mechanisms of action (Fig. 1). In addition, the large size and polyanionic nature of oligonucleotides impose intrinsic barriers to cellular uptake, tissue penetration, and distribution, particularly within the CNS [9–11]. Consequently, the developments of chemical strategies to enable efficient CNS delivery has also been an area of intense research activity. In the following sections, we provide a comparative overview of the structural features, mechanisms of action, and commonly used as well as emerging chemical modifications for ASOs, siRNAs, and aptamers, with particular emphasis on their clinical progress and remaining unmet needs in CNS disorders. Subsequent sections will focus on the therapeutic applications and clinical landscape of nucleic acid–based interventions across specific neurodegenerative diseases. Approaches relying on lipid nanoparticle–based formulation and similar class strategies are beyond the scope of this review and are discussed elsewhere [12–14].

Fig. 1.

Commonly used and emerging chemical modifications in ASO, siRNA, and aptamer-based therapeutics. (A) Sugar modifications, (B) Phosphate backbone modifications, (C) Base modifications, and (D) Morpholino oligonucleotides. Sugar, phosphate backbone, and nuclease modifications are frequently combined to tune biophysical and pharmacological properties of oligonucleotides. *Stereogenic centers at carbon or phosphorus atoms. R: 2′ substituents. R’: various functional groups

Antisense oligonucleotides

Mechanism

ASOs are short, single-stranded synthetic nucleic acids that bind to a sequence-complementary target RNA and modulate its function [2, 15]. Depending on chemistry and design, ASOs can be broadly classified into two major functional categories: (i) gapmer ASOs that recruit endogenous endonuclease, RNase H1, upon binding to the target selectively cleaving target transcripts (Fig. 2A); and (ii) steric-blocking oligonucleotides that modulate pre-mRNA splicing or other RNA processing events without inducing RNA cleavage (Fig. 2B).

Fig. 2.

Mechanisms and representative chemistries of CNS-targeted ASOs. (A) Gapmer ASOs hybridize to target RNA and recruit endogenous RNase H1, resulting in cleavage of the RNA strand and transcript degradation. Schematic examples of a 2′-MOE gapmer and a PS/PO mixed-backbone design are shown. (B) Schematic of steric-blocking ASOs bind to a splicing enhancer or silencer element, the ASOs sterically interfere with spliceosome and/or RNA-binding protein assembly, thereby promoting exon inclusion or exon skipping and resulting in splice modification. Representative 2′-MOE-modified and PMO-based steric-blocking designs are shown

Structure and chemistry for gapmer ASOs

Gapmer ASOs are designed to reduce target RNA levels through RNase H1-mediated cleavage (Fig. 2A) [2, 15]. A typical gapmer comprises chemically modified 5′ and 3′ “wings” flanking a central DNA gap; upon hybridization to the target RNA via Watson–Crick base pairing, the DNA region in the gapmer forms DNA/RNA heteroduplex that recruits RNase H1, which triggers the cleavage of the RNA strand and thereby suppresses gene expression [16]. To enhance the stability of gapmer oligonucleotides, a range of chemical modifications has been employed, including 2′-sugar modification such as 2′-O-methyl (2′-OMe), 2′-fluoro (2′-F), and 2′-O-methoxyethyl (2′-MOE), as well as phosphorothioate (PS) backbone substitution. These modifications are widely incorporated into FDA-approved ASOs and dozens of late clinical trials.

Structure and chemistry for steric-blocking ASOs

In contrast to gapmer-type ASOs, steric-blocking ASOs modulate RNA function without inducing RNA cleavage (Fig. 2B) [17]. In CNS therapeutics, steric-blocking ASOs are most commonly deployed as splice-switching oligonucleotides (SSOs), which bind to specific pre-mRNA sites to modulate spliceosome assembly, thereby promoting exon inclusion or exon skipping [17]. This steric blocking approach can also be used to alter non-coding RNA, such as microRNA (miRNA), to directly inhibit the function of disease-causing miRNA expression. Because this approach does not require interactions with endogenous protein machineries to be functional, a broader range of chemical modifications can be tolerated. Therefore, in addition to the chemistry used for gapmer ASOs, other chemical modifications such as phosphorodiamidate morpholino oligonucleotides (PMOs), which consist of a six-membered morpholine moiety and non-charged phosphorodiamidate backbone, are also applied for clinical development. While PMOs have significantly high metabolic stability, a major drawback is their low bioavailability.

FDA approvals of CNS-targeted ASO drugs

Both gapmer and steric blocker ASOs have received FDA approval for clinical use. The first (and only) approved gapmer ASO in the CNS is tofersen (Table 1), a 20-mer with partial PS backbone and 2′-MOE modified wings, used to treat amyotrophic lateral sclerosis (ALS) driven by mutations in superoxide dismutase 1 (SOD1). It is intrathecally administered and targets mutated SOD1 mRNA, inhibiting production of toxic SOD1-protein (see details in SOD1-ALS section) (Table 1) [18–20]. The first and only approved splice modulation ASO is nusinersen, an 18-mer fully PS- and 2′-MOE–modified SSO for spinal muscular atrophy (SMA), an autosomal recessive neuromuscular disorder characterized by degeneration of lower motor neurons in the spinal cord and brainstem [21] (Table 1). SMA is caused by homozygous deletions or loss of function mutations in Survival of Motor Neuron 1 (SMN1), leading to insufficient expression of SMN protein [21]. Nusinersen increases functional SMN protein by promoting inclusion of SMN2 exon 7 via blockade of an intronic splicing silencer [22, 23]. In 2016, nusinersen was approved by the US FDA to treat SMA [24]. Building on this mechanism, salanersen (BIIB115/ION306), a next-generation SSO incorporating 2′-O-[2-(methylamino)−2-oxoethyl] (NMA) modification [25], is being evaluated for once-yearly intrathecal administration, with preliminary phase 1 findings showing sustained improvements in biomarkers and motor function in patients previously treated with gene therapy [26].

Table 1.

FDA approved ASO-based therapeutics on neurodegenerative disorders

Drug name	Commecial name	FDA approved date	Disease application	Gene target	Modality/Chemcial modification	Dose and delivery system	Company	Relevent reference
Nusinersen	Spiniraza	May-17	Spinal muscular atrophy	SMN2	18-mer SSO, PS backbone, 2'-MOE	12.5 mg, once every 4 months, IT	Ionis	[24]
Tofersen	Qalsody	Apr-23	Amyotrophic lateral sclerosis-SOD1	SOD1	20-mer, gapmer ASO, PS/PO backbone, 2'-MOE modified wings	100 mg, every 14 days for 3 times, then every 28 days, IT	Ionis, Biogen	[18, 19]

Chemical determinants of CNS distribution and tolerability of ASOs

Single-stranded PS-modified ASOs are conformationally flexible and bind extensively to proteins in biofluids and tissues, which can slow extracellular clearance, promote tissue retention, and support cellular uptake in vivo even without formulations or uptake ligands [27–29]. For the CNS indications, where delivery is constrained by the blood–brain barrier (BBB), most programs rely on local intrathecal administration, and approved CNS ASOs have so far been delivered as unformulated (“naked”) oligonucleotides. This underscores that PS-ASOs can achieve sufficient exposure and broad pharmacologic activity. Importantly, recent work confirms that CNS distribution of ASOs is generally broad and robust across neuronal and glial cell types, largely independent of ribose sugar modifications [30]. However, chemical features, particularly 2′-sugar chemistries, strongly influence the magnitude and durability of pharmacologic activity, with higher target RNA affinity chemistries (e.g. cEt, LNA) generally producing more durable target knockdown than 2′-MOE ASOs, as reflected by CSF biomarkers of neuronal engagement [30]. More broadly, because ASO uptake is dependent on protein binding, and different ASO sequences adopt different conformations and thus protein binding properties, each sequence needs to be checked to ensure appropriate pharmacokinetics before proceeding to clinical development.

The high protein-binding affinity of PS-ASOs is a double-edged sword: while it supports tissue retention and uptake, it may also increase non-specific protein interactions in the CNS, potentially contributing to toxicities [27, 31–33]. Importantly, many of CNS toxicities are attributed to the PS backbone itself, as conversion to a phosphodiester (PO) backbone has been shown to markedly reduce toxicity [31–34], albeit often at the expense of uptake [34]. To preserve PS-mediated uptake and distribution while mitigating PS-driven adverse interactions, CNS-targeting ASO designs increasingly employ mixed-backbone patterns (Fig. 2B), in which PS linkages are strategically retained where needed, while other positions are converted to PO linkages. In addition to PS modifications, sequence features also influence tolerability: in particular, G-rich motifs have been associated with indirect effects on gene expression [35] and CpG-containing motifs with immunostimulatory effects [36], whereas G residues at the 3′ end are considered undesirable from a safety perspective [37], and are avoided when possible.

Beyond backbone composition and sequence, a broad range of 2′ -sugar chemistries has been explored for CNS-targeted ASOs. In many CNS programs, 2′-MOE-based designs have served as a gold standard for both gapmer and steric-blocking modalities, and approved CNS ASOs to date are 2′-MOE-based. In parallel, higher-affinity wing chemistries have been developed for gapmer ASOs, including constrained ethyl (cEt) [38, 39] as well as 2′,4′-bridged nucleic acids such as locked nucleic acid (LNA) [40, 41] and ethylene-bridged nucleic acid (ENA, more detail in AD section) [42], in which the 2′-O and 4′-C positions are covalently linked to form a bicyclic sugar that markedly increases RNA-binding affinity and nuclease resistance. These modifications can improve potency relative to 2′-MOE gapmers in some settings, and indeed, an LNA-containing ASO (rugonersen) has advanced from a phase 1 trial to phase 3 development for Angelman Syndrome [43], but high-affinity bicyclic chemistries seem to have a higher propensity to cause toxicities [44–46]. This underlines the need for careful sequence selection and validation in the CNS. In contrast, PMOs provide an alternative steric-blocking scaffold (Fig. 2B) [47]; however, their charge-neutral backbone and low protein binding can limit productive cellular uptake in the CNS without additional delivery assistance. Accordingly, these non-MOE chemistries require careful evaluation in CNS applications, particularly with respect to tolerability (e.g., cEt/LNA/ENA) and productive cellular uptake (PMOs).

Small interfering RNA

Mechanism

siRNAs are negatively charged, double-stranded RNA molecules of approximately 21–23 nucleotides that silence gene expression through the endogenous RNA interference (RNAi) pathway (Fig. 3A) [48]. Following cellular uptake, the siRNA duplex is loaded directly onto AGO2, forming the core of the RNA-induced silencing complex (RISC), where the sense strand is discarded and the antisense strand directs sequence-specific binding, enabling AGO2-mediated degradation of target RNAs (Fig. 3A) [49].

Fig. 3.

siRNA architectures and RNAi mechanism. (A) RNA interference (RNAi) mechanism: The siRNA duplex is loaded into the RNA-induced silencing complex (RISC), the sense strand is discarded, and the antisense strand directs AGO2-mediated cleavage of complementary target RNA, leading to RNA degradation. (B) Approach 1: Incorporating PS-modified single-stranded regions and Approach 2: Increasing the size of molecules. Schematic comparison of a conventional siRNA with architectures that introduce PS-modified tails to enhance pharmacokinetics and pharmacodynamics. Examples include an asymmetric siRNA with a shortened sense strand, a divalent siRNA format in which two siRNA duplexes are linked, and siRNA-ACO in which a PS-modified single-stranded oligonucleotide is appended via a linker. (C) Approach 3: Conjugating hydrophobic or amphiphilic ligands. Representative ligand-conjugated siRNA designs for CNS delivery are shown. C16-siRNA incorporates an internal C16 hydrophobic moiety on the sense strand, whereas dendrimer-siRNA is covalently linked to a dendrimeric, amphiphilic scaffold

Structure and chemistry for siRNAs

Clinically, all approved siRNA therapeutics incorporate AGO2-compatible 2′-OMe and 2′-F ribose modifications, and with terminal PS modifications to enhance stability. To date, these drugs target the liver, predominantly exploiting efficient and selective hepatocyte uptake via N-acetylgalactosamine (GalNAc) conjugation to siRNAs to promote asialoglycoprotein receptor (ASGPR) binding highly expressed in hepatocyte, in some cases resulting in more than 6-months of duration effect by single administration in humans [50].

Double-stranded nucleic acids are structurally rigid and hydrophilic due to their highly charged, poly-anionic nature. While even unconjugated PS-modified ASOs have a relatively flexible nature and high binding affinity for a variety of proteins, as described above, siRNAs are dependent on formulation or conjugation approaches for cellular uptake in vivo. In the CNS, similarly to ASOs, siRNA delivery is constrained by the BBB, and most programs therefore rely on local administration, such as intrathecal dosing. However, even with local delivery, conventional siRNAs are rapidly cleared from the cerebrospinal fluid (CSF) due to their low protein binding and physicochemical properties, limiting cellular uptake and penetration into the brain parenchyma. In addition, the CNS lacks an established, high-capacity uptake receptor comparable to ASGPR in the liver. Consequently, substantial effort has focused on engineering siRNAs to improve tissue retention, cellular uptake, and productive cytosolic delivery following local administration.

Chemical determinants of CNS distribution and tolerability of siRNAs

These strategies can be broadly categorized into the following approaches: (i) incorporating PS-modified single-stranded regions, (ii) increasing the size of the molecule, and (iii) conjugating hydrophobic or amphiphilic ligands.

In the first approach, the siRNAs are designed with flexible, single-stranded PS tail regions, effectively imparting PS-ASO-like properties to exploit the inherently higher bioavailability of ASOs. This approach was motivated by observations that conventional siRNA duplexes with 2-nucleotide 3′ overhangs exhibit suboptimal in vivo pharmacodynamics, particularly outside the liver [51]. Further optimization involves designing siRNA in a more asymmetric configuration, featuring a short sense strand and thus an extended PS-modified tail on the antisense strand to retain stability and cellular uptake (Fig. 3B) [51, 52]. The second strategy is to make these asymmetric duplexes in dimeric configuration (Fig. 3B) where two siRNA molecules are covalently connected via sense strands, which boosts broad CNS distribution, providing 6-month of duration of effect by single intracerebroventricular (ICV) administration in rodents and non-human primates (NHP). The increased molecular weight of these divalent constructs can also slow CSF clearance, thereby prolonging CNS exposure [53] (Fig. 3B). Another example that engages a similar principle is siRNA-ACOs, which consist of a conventional siRNA duplex appended with an “accessory oligonucleotide “ that potentially enhances cellular uptake while further prolonging CNS exposure [54–56] (Fig. 3B).

The third strategy is to attach lipophilic or amphiphilic ligands to siRNAs to improve their CSF retention and cellular uptake within the CNS (Fig. 3C). These strategies often leverage lipid ligands [57] and albumin-binding ligands [58, 59]. For example, C16 siRNAs incorporate a linear 16-carbon alkyl chain (2′-O-hexadecyl, “C16”) appended to a nucleotide within the duplex, which increases lipophilicity and supports activity and distribution across CNS regions and cell types with sustained RNAi activity [57]. As another example, albumin-binding dendrimer ligands employ a branched, amphiphilic dendritic scaffold that binds albumin with high affinity [60], thereby prolonging CSF residence and promoting widespread CNS delivery and target gene suppression following local administration [59]. In general, highly hydrophobic lipids such as cholesterol or docosanoic acid (DCA) with saturated C22 alkyl chain are associated with significant toxicity and limited CNS distribution near the injection site [61].

FDA-approvals and clinical trials of siRNA in CNS

As of early 2026, eight FDA-approved siRNA therapeutics form a coherent class of “information drugs” that selectively silence disease-causing genes, primarily in the liver. Patisiran (Onpattro) targets transthyretin in hepatocytes to treat hATTR amyloidosis using a lipid nanoparticle (LNP)-based delivery system, while all subsequent approvals, such as Givosiran (acute hepatic porphyria), Lumasiran and Nedosiran (primary hyperoxaluria), Inclisiran (hypercholesterolemia), Vutrisiran (hATTR amyloidosis), Fitusiran (hemophilia A/B), Plozasiran (familial chylomicronemia syndrome) share a common GalNAc-siRNA chemistry platform [2, 62–64]. These drugs employ the same core chemistry (2′-OMe and 2′-F sugar modifications with terminal PS linkages) and exploit GalNAc conjugation for targeted uptake by hepatocytes via the ASGPR receptor. These GalNAc-conjugated siRNAs share a similar structure and use the same mechanism of uptake, and so their ability to target such a wide variety of diseases simply by changing base sequence highlights a core advantage of oligonucleotide-based drugs: the steady stream of approvals illustrates how a single, validated chemistry platform can be efficiently adapted to treat numerous genetically defined disorders. Clinical development of siRNA therapeutics for CNS disorders is now emerging, progressing from preclinical proof-of-concept studies to first-in-human trials (see disease-focused sections). While siRNA drugs have not yet been approved for CNS indications, multiple programs using advanced CNS-targeted siRNA chemistry, such as Alnylam’s C16-conjugated siRNAs and Atalanta Therapeutics’ divalent-siRNA platforms, are actively advancing toward clinical proof-of-concept. These efforts provide the first clinical evidence that sequence-based RNAi can be delivered safely and effectively to the brain and spinal cord, overcoming the historical challenges of CNS delivery.

Sequence design and selection of ASO and siRNA

A recent preprint introduced “ASO Atlas”, a large patent-derived dataset of 188,521 RNase H-mediated ASOs spanning 334 genes, and used it to train OligoAI, a machine learning (ML) model that jointly encodes target RNA context, oligo sequence, sugar/backbone chemistry, and dose to predict in vitro knockdown efficacy [65]. Both ASO Atlas and OligoAI are made freely accessible through an online web tool, with the aim of enabling rapid, model-guided prioritization of candidate ASOs for experimental testing and accelerating optimization of ASO design (https://sitlabs.org/oligoai). In contrast, for SSOs, candidate binding sites are often inherently constrained to the area including and immediately flanking the exon of interest; therefore, a small screen may suffice to find active compounds, and AI-based design and prioritization tools may not always be required. Nonetheless, computational tools such as eSkip-Finder (https://eskip-finder.org/cgi-bin/input.cgi) [66] are used in practice to support SSO design. Platforms for artificial intelligence (AI)/ML-assisted siRNA sequence design have also been developed. For example, the Khvorova laboratory has released a web-based siRNA sequence search/prioritization interface (https://sirna.ai/). These approaches integrate large-scale sequence datasets with experimental screening data from fully chemically modified siRNA (2′-OMe/F with terminal PS modifications) to prioritize siRNA sequences with strong silencing activity and favorable specificity. Importantly, since the optimal sequences can differ between native RNA and modified RNA, the dataset used as the basis for this server made use of fully chemically modified siRNA (2′-OMe/F with terminal PS modifications). ML models typically incorporate features such as sequence composition, thermodynamic properties, and target-site accessibility; however, predictive performance can be limited by biological complexity, including the strong influence of protein-target interactions in vivo, modification chemistry, delivery context, and cell- and tissue-dependent differences in RNAi activity and RISC function [3]. Accordingly, AI-driven approaches currently serve as candidate-prioritization tools that complement empirical screening using clinically relevant siRNA scaffolds and delivery systems.

New chemical approaches to further enhance PK/PD and reduce toxicity of ASO, SSO, and siRNA

Despite multiple FDA approvals and a growing pipeline of nucleic acid therapeutics, there remains considerable interest in chemical modifications to improve durability, target selectivity, and tissue distribution. While FDA-approved oligonucleotide drugs have created a perception of “maturity” in ASO and siRNA chemistry, this success largely reflects efficacy in a single tissue context, primarily liver. Extrahepatic tissues, including the CNS, still present significant challenges for improvement in achieving effective delivery and therapeutic activity, highlighting the need for new chemical innovation (Fig. 1). The months-long activity of FDA-approved oligonucleotide drugs is likely due, in part, to endo/lysosomal entrapment, where “deposited” oligonucleotide slowly leaks from intracellular compartments to reach their targets, enabling prolonged activity from a single administration. After circulation and accumulation in tissues, these drugs are gradually metabolized over months, ultimately becoming inactive. Further stabilization of oligonucleotides could therefore extend their durability and therapeutic effect. To achieve this, a variety of ribose sugar modifications have been actively investigated for therapeutic utility; a new class of bridged (or locked) nucleic acids such as S-constrained ethyl (S-cEt) [67], spirocyclopropylene bridged nucleic acid (scpBNA) [68], guanidine-bridged nucleic acid (GuNA) [69, 70], 5′-cyclopropyl (5′-CP) [71], extended nucleic acid (exNA) [72], and NMA [25]. NMA is incorporated into the clinical-stage SMN2-targeting SSO salanersen and has also enhanced the potency of an SCN1a-targeting SSO for Dravet syndrome, supporting its broader applicability to SSOs [25]. Acyclic-type modifications such as (3′ → 2′)-L-α-threose nucleic acid (TNA) [73, 74], serinol nucleic acid (SNA) [75, 76], and glycol nucleic acid (GNA) [77, 78] have also been actively explored for various therapeutic RNA modalities and purposes. Phosphate backbone modifications have also been explored, including two phosphoramidate analogues [mesyl-phosphoramidate (MsPA) [79], and a cyclic phosphoryl guanidine (PN) [80, 81]], phosphorodithioate (PS2) [82], methoxypropylphosphonate (MOP) [83], amide (AM) [84, 85], and a variety of phosphotriester modifications [86, 87]. These chemical innovations have demonstrated enhanced stability, potency, and prolonged activity for ASOs and/or siRNAs, providing a foundation for the next generation of nucleic acid therapeutics. In the SSO field, other than PN backbone exploration, thiophosphoramidate morpholino oligonucleotide (TMO) [88, 89] has recently been developed to improve the poor in vivo PK of PMO, which arises from their uncharged backbone and poses challenges in chemistry, manufacturing, and controls (CMC). TMOs are morpholino oligonucleotides with PS linkages, combining the PMO backbone with PS chemistry to improve tissue distribution, plasma stability, and nuclear accumulation. Minimizing neurotoxicity is a critical aspect of therapeutic development, and chemistry-driven approaches have been particularly active in this area. Remarkably, very local and minimal structural alterations can drastically modulate the toxicity profile of oligonucleotides. For example, a single 2′-OMe-modification at gap position 2 of gapmer ASOs significantly reduces overall protein binding and mitigates toxicity associated with paraspeckle protein delocalization, with minimal impact on target knockdown [90]. A single 5′-cyclopropylene (5′-CP) modified nucleotide at the wing/gap junction of ASO-gapmer also markedly reduces late-onset neurotoxicity without compromising on-target potency [34]. The toehold approach uses a transiently bound peptide nucleic acid (PNA) oligonucleotide to mask not fully but partially the ASO strand, forming ASO/PNA double helix, which prevents non-specific toxic protein interactions to the single-stranded PS-ASO architecture. The masking PNA is displaced upon ASO’s binding to the fully complementary RNA target, allowing therapeutic activity while minimizing off-target protein interactions [91].

Chemical innovation has also advanced for the mitigation of off-target effects due to unintended transcript binding, particularly in siRNAs. For instance, incorporation of GNA in the siRNA seed region substantially reduces miRNA-like off-target activity while preserving on-target engagement in vivo [92, 93]. Another strategy involves introducing an intentional mismatch base outside the seed region of siRNA guide strand, which enhances single-mismatch discrimination, and improves selectivity for single-nucleotide polymorphism (SNP) discrimination in vivo [94], which discrimination properties are further improved with structurally constrained internucleotide E-VP incorporation in the siRNA guide strand [95].

Aptamers

Aptamers, structured single-stranded short DNA or RNA selected for high-affinity molecular recognition of target molecules, have emerged as powerful tools for the development of biomarkers, tissue/cell targeting, agonist/antagonist, and enzymatic regulators (Fig. 4) [96]. Aptamers recognize molecular targets through tertiary folding of single-stranded nucleic acids, enabling affinities comparable to monoclonal antibodies but with substantially smaller molecular footprints. Their small size, chemical versatility, low immunogenicity, and compatibility with rational engineering strategies have positioned them as synthetic analogs to antibodies with potential advantages for CNS applications. However, in contrast to clinically validated oligonucleotide therapeutic platforms such as siRNA and ASO-based drugs, CNS-targeted aptamer therapeutics remain an emerging and largely experimental modality. Therefore, their current role should be interpreted primarily as a promising discovery, targeting, diagnostic, and preclinical delivery platform rather than as a clinically proven CNS therapeutic class. Particularly notable are their adaptability to diverse diagnostic formats, their capacity to distinguish pathogenic conformers of misfolded proteins, and their potential to cross or exploit the BBB for therapeutic delivery. Since their inception alongside the SELEX (Systematic Evolution of Ligands by Exponential Enrichment) method in 1990 [97, 98], aptamers have expanded into a sophisticated platform capable of targeting proteins, peptides, ions, viruses, cells, and complex pathological structures [99]. Their in vitro evolution via SELEX has undergone substantial methodological refinement, incorporating microfluidics, cell-based SELEX, in vivo SELEX, and machine-learning–guided optimization [100–102]. SELEX using engineered polymerases and xeno nucleic acids (XNAs) has emerged as an exciting recent advance [103]. Aptamers built from 2′-F-arabinonucleic acid (FANA), (3′ → 2′)-L-α-threofuranosyl nucleic acid (TNA), and other XNA chemistries are particularly attractive because they are highly resistant to degradation and may enable access to expanded folding and functional “sequence space” [104–106]. Chemical synthesis ensures batch reproducibility and provides opportunities for structural reinforcement (2′-OMe, 2′-F, PS linkages), PEGylation, fluorescent tagging, and attachment to nanoparticles or drug cargos. Collectively, these features underpin the modularity and tunability central to aptamer function. At the same time, the translational performance of aptamers in the CNS depends on overcoming nuclease degradation, rapid renal clearance, limited tissue exposure, and the need to maintain target binding after chemical modification or cargo conjugation [107]. Advances in backbone engineering, including L-nucleic acid aptamers (Spiegelmers) [108] as well as nanoparticle encapsulation and ligand-mediated BBB transport strategies have substantially improved pharmacokinetic in experimental systems. Nevertheless, major challenges remain for chronic dosing paradigms in neurodegenerative disease, and the technological foundation continues to advance toward clinical validation.

Fig. 4.

Aptamer folding and target recognition. A single-stranded nucleic acid aptamer (left) adopts a defined secondary/tertiary structure (middle) that creates a specific binding interface. The folded aptamer then engages its molecular target (right; shown in green), forming a high-affinity, shape- and chemistry-complementary complex. This technology is adapted for detection of biomarkers, cell/tissue targeting, agonist/antagonist development, and enzymatic regulations

CNS receptors, channels, and neurotransmission modulating aptamers

Aptamers can also act as functional modulators of neuronal receptors. For example, NMDA receptor–targeting RNA aptamers can attenuate excitotoxic calcium influx [109]. Additional aptamers recognizing neurotransmitters like dopamine and serotonin support real-time neurochemical sensing [110]. In spinal cord injury models, aptamers against myelin-associated glycoprotein (MAG) promote neurite outgrowth by neutralizing MAG-mediated inhibition [111]. These applications underscore aptamers as both analytical and functional tools for probing synaptic physiology and neuronal repair.

Systemic strategies for CNS targeting of NATs

BBB penetrant ASOs/siRNAs

Most ASOs and siRNAs for CNS indications are currently delivered by intrathecal bolus or infusion, which achieves broad distribution in the neuraxis but entails invasive procedures and specialized infrastructure [112]. For chronic diseases such as AD, PSP, or ALS, systemic administration with efficient BBB penetration would be more scalable and acceptable to patients. Importantly, this section distinguishes clinically validated oligonucleotide therapeutic modalities from BBB-penetrant delivery technologies that remain under active development. Several ASO and siRNA drugs have established clinical precedent, but systemic BBB-crossing strategies for CNS delivery still require further human validation.

Receptor-mediated transcytosis across the BBB has emerged as a leading strategy. Bispecific or engineered antibodies that bind both a BBB receptor and a therapeutic target can shuttle cargos such as enzymes or antibodies into the brain. Transferrin receptor 1 (TfR1) and CD98 heavy chain (CD98hc) are among the most advanced targets, with distinct kinetics and distribution profiles [113, 114]. Recently, a dual transport vehicle (dual-TV) platform that simultaneously engages TfR and CD98hc was shown to achieve higher and more sustained brain exposure of large molecules than either receptor alone [115]. Although this dual-TV has so far focused on proteins, the same principles could be applied to ASO or siRNA conjugates via Fc fusion or other scaffolds. Among BBB shuttle approaches for oligonucleotides, TfR1-based delivery has been most developed in the context of ASOs. An engineered TfR1-binding oligonucleotide transport vehicle (OTV) conjugated to ASOs enabled broad CNS biodistribution and robust target RNA knockdown following intravenous administration in rodents and nonhuman primates [116]. This study establishes TfR1 shuttling as a powerful platform for systemic CNS delivery of ASOs, although clinical validation in humans has not yet been reported. Subsequent refinements, including strategies to reduce non-specific protein interactions of PS-modified ASOs, further improved the specificity and efficiency of TfR-mediated transport [117], highlighting key design principles for next-generation OTV systems. Similarly, conjugation of TfR-targeting antibodies [118] or peptides [119] has been shown to facilitate brain delivery of siRNAs and achieve gene knockdown in rodent models. However, these approaches remain at an early, proof-of-concept stage and have not yet demonstrated durable CNS target engagement beyond rodents. Collectively, integrating BBB transport strategies with ASO/siRNAs payloads could enable systemic administration of precision oligonucleotide therapeutics for neurodegenerative diseases. As an alternative approach, cholesterol-conjugated heteroduplex oligonucleotides (cholesterol-HDOs) have been explored for systemic CNS delivery. HDOs are double-stranded ASO constructs in which a DNA antisense strand is hybridized to a complementary RNA strand [120]. This design maintains the RNase H-dependent activity of the DNA antisense strand, while the complementary RNA strand can be functionalized with diverse ligands, enabling improved in vivo potency compared with conventional single-stranded ASOs or ligand-conjugated single-stranded ASOs [120]. In rodents, cholesterol-conjugated HDOs enabled BBB crossing after systemic dosing and led to broad distribution to the brain with substantial knockdown of multiple CNS targets, whereas cholesterol-conjugated single-stranded ASOs showed limited CNS activity [121]. However, key limitations include sequence dependence of BBB penetration as well as the relatively high systemic doses to achieve CNS exposure, which can increase peripheral exposure and may be associated with toxicity [121, 122], thereby potentially constraining the therapeutic window. Similarly, lipid conjugation has also been reported for siRNA-based CNS delivery. In one study, a phosphocholine-docosahexaenoic acid (PC-DHA)-conjugated, fully modified siRNA achieved brain target knockdown after systemic administration only when combined with transient osmotic BBB disruption [123], indicating the difficulty of crossing an intact BBB with lipid conjugation alone for siRNA modalities.

BBB penetrant aptamers

Crossing the BBB remains one of the most persistent hurdles in CNS therapeutics. Aptamers engineered to bind endogenous transport receptors, particularly the TfR, have demonstrated promising BBB-shuttling potential. TfR-binding aptamers such as FB4 have been used for targeted delivery of imaging agents and therapeutics to glioma tissue [124]. Aptamer-mediated transport of siRNAs and antisense oligonucleotides across the BBB has also been reported, with measurable gene knockdown in brain tissue following systemic administration [125]. In vivo SELEX approaches can identify aptamers that home to the brain without pre-defined target molecules [126], opening opportunities to discover new BBB transport pathways. However, these studies should be viewed as preclinical proof-of-concept rather than clinical validation.

Alzheimer’s disease (AD)

AD and primary tauopathies remain areas of major unmet medical need. While amyloid-beta (Aβ)-directed monoclonal antibodies have demonstrated statistically significant slowing of clinical decline in early symptomatic AD, the magnitude of benefit is modest, and safety monitoring, particularly for amyloid related imaging abnormalities (ARIA), adds complexity. This reinforces the need for complementary strategies that more directly target downstream neurodegeneration. In this context, RNA therapeutics are attractive since they can modulate disease pathways at the level of gene expression and RNA processing, enabling for precise control of proteins that are difficult to target with small molecules or antibodies. RNA-based modalities, particularly ASOs and siRNAs, provide programmable, sequence-specific intervention points that can be tailored for complete knockdown, isoform correction, or pathway rebalancing. Aptamers, often described as “chemical antibodies,” add a third modality with distinct strengths in high-affinity binding to conformational epitopes and potential use in both therapeutics and diagnostics.

ASO-based therapeutics for AD

Among RNA therapeutics, ASOs are the most clinically advanced for tau-related indications (Table 2). A central rationale is that tau pathology is closely linked to neuronal dysfunction and clinical progression, and that both quantitative (tau abundance) and qualitative (isoform balance, phosphorylation state, and aggregation propensity) properties of tau may be therapeutically actionable. Intrathecally administered tau-lowering ASOs have shown robust target engagement, with reductions in cerebrospinal fluid (CSF) total tau and phosphorylated tau species reported in early clinical studies in mild AD [127]. BIIB080/IONIS-MAPTRx, now known as diranersen, has been evaluated in a Phase 1b trial in mild AD (NCT03186989) [128] and, more recently, in the Phase 2 CELIA study in early AD (NCT05399888) [129]. Although CELIA did not meet its primary endpoint of demonstrating a dose–response relationship for change from baseline in CDR-SB at Week 76, Biogen and Ionis reported pre-specified analyses showing slowing of clinical decline across all tested doses, with the most pronounced effect observed at the lowest dose regimen, 60 mg every 24 weeks. Robust reductions in CSF tau biomarkers and tau PET signal were also reported across dose groups, supporting sustained target engagement and downstream biomarker effects. Based on these topline data, Biogen announced plans to advance diranersen into registrational development. In addition, Novartis’ MAPT-targeting gapmer ASO NIO752 [130] is in Phase 1 for progressive supranuclear palsy (PSP) (NCT04539041) and has initiated early AD studies (NCT05469360; NCT06372821). These results support the feasibility of sustained pharmacodynamic effects in the CNS, and ongoing studies aim to connect biomarker changes to clinically meaningful outcomes. In parallel, clinical programs are extending beyond AD to primary 4R tauopathies such as progressive supranuclear palsy (PSP) and related disorders, where MAPT biology provides a strong mechanistic rationale [131]. From a development perspective, the near-term challenge is not only demonstrating target engagement but also establishing that the degree and duration of tau lowering are sufficient, and appropriately localized, to modify disease trajectories. For tauopathies such as PSP, where deep brain structures (for example, basal ganglia and brainstem nuclei) are central to symptom generation, the ability to achieve adequate exposure and pharmacology in these regions becomes particularly important.

Table 2.

Cliniclal evaluation of ASO-based therapeutics on neurodegenerative disorders

Drug name	Disease application	Gene target	Modality/Chemcial modification	Company	CT identifier	Phase	Stage	Key clinicla findings	Relevent reference
Diranersen (BIIB080; IONIS-MAPTRx)	Alzheimer's disease	MAPT	18-mer gapmer ASO, PS backbone, 2'-MOE modified wings	IONIS/Biogen	NCT03186989	Phase Ib	Completed	Dose- and time-dependent reductions in CSF total tau and p-tau181, and reduced tau PET signal	[128 ]
Diranersen (BIIB080; IONIS-MAPTRx)	Alzheimer's disease	MAPT		IONIS/Biogen	NCT05399888	Phase II (CELIA)	Active, not recruiting (placebo-controlled Phase 2 completed; Long-term extension ongoing)	Primary dose–response endpoint not met; robust CSF tau (50–65%) and tau-PET reductions observed across doses. At 60 mg, clinical decline was slowed by 26% (CDR-SB), 42% (ADAS-Cog13), and 50% (MMSE) at 18 months. Phase 3 development planned.	[129]
NIO752	Progressive supranuclear palsy	MAPT	20-mer gapmer ASO, PS backbone, 2'-MOE modified wings	Novartis	NCT04539041	Phase I	Completed	Safety concern with certain doses. Higher levels of certain CSF inflammatory indicators (e.g., white blood cell counts)	[130]
NIO752	Alzheimer's disease, for safety PK/PD tiral	MAPT		Novartis	NCT05469360	Phase I	Active, recruiting	N/A	[130]
NIO752	Alzheimer's disease, to determine clinical benefit using stable isotope labeling kinetics (SILK)	MAPT		Novartis	NCT06372821	Phase I	Active, recruiting	N/A	[130]
Ulefnersen (ION363)	FUS-ALS	FUS	20-mer gapmer ASO, PS backbone, 2'-MOE modified wings	IONIS	NCT04768972	Phase III	Active, not recruiting	Early open-label studies have demonstrated reductions in CSF FUS and suggest possible slowing of functional decline in some patients	[168, 169]
BIIB078	C9ORF72-ALS	C9ORF72	18-mer gapmer ASO, PS/PO backbone, 2'-MOE modified wings	Biogen	NCT04288856	Phase I	Terminated	Persistent repeat RNA and dipeptide repeat inclusions in the brain and spinal cord despoite reduced CSF poly(GP)	[171]
WVE-004	C9ORF72-ALS		20-mer gapmer ASO, PS/PN backbone, 2'-MOE/OMe modified wings	Wave	NCT04931862	Phase I/II	Terminated	No clinicla benfit was found
QRL-201	Sporadic ALS and C9ORF72-ALS	STMN2	SSO	QurAlis	NCT05633459	Phase I	Active, not recruiting	Based on the PK data analysis of the dose escalation phase, the CSF exposure level of QRL-201 met or exceeded the target range	[178]
LY4256984	Sporadic ALS	UNC13A	SSO	QurAlis/Eli Lilly	NCT07100119	Phase I	Active, recruiting	N/A	[181]
BIIB105	Sporadic ALS	ATXN2	20-mer gapmer ASO, PS backbone, 2'-MOE modified wings	Biogen	NCT04494256	Phase I/II	Terminated	No clinical benfit was found	[183]
ION717	Prion disease	PRNP	ASO	IONIS	NCT06153966	Phase I/II	Active, not recruiting	N/A	[186, 187]
RG6042(Tominersen)	Huntington disease	HTT	20-mer gapmer ASO, PS/PO backbone, 2'-MOE modified wings	IONIS/Roche	NCT02519036	Phase I/Iia	Completed	Dose-dependent lowering of CSF mutant huntingtin protein	[191]
RG6042(Tominersen)	Huntington disease	HTT		Roche	NCT03761849	Phase III	Terminated	No clinical benfit was found and there was safety concern	[193]
RG6042(Tominersen)	Early Huntington disease	HTT		Roche	NCT05686551	Phase II	Active, not recruiting	N/A
WVE-003	Hungtinton disease	exapnded CAG repeat, mutant HTT(SNP rs362273)	20-mer gapmer ASO, PS/PN backbone, 2'-MOE/OMe modified wings	Wave Life Science/Takeda	NCT05032196	Phase Ib/IIa	Completed	Reduced CSF mutant HTT by a mean of 46%	[200]
VO659	Spinocerebellar ataxia types 1 and 3, Hungtinton disease	expanded CAG repeat	SSO	Vico Therapeutics B. V	NCT05822908	Phase I/II	Active, recruiting	~ 28% mean reduction in CSF mutant huntingtin at day 85 with good tolerance	[213]
BIIB094(ION859)	Parkinson disease	LRRK2	gapmer ASO	IONIS	NCT03976349 (REASON)	Phase I	Completed	Generally safe and well tolerated	[230]
BIIB101 (ION464)	Multiple system atrophy	SNCA	gapmer ASO	IONIS/Biogen	NCT04165486(HORIZON)	Phase I	Active, not recruiting	Generally safe and well tolerated	[237]

In parallel with these clinical-stage programs, preclinical ASO strategies are broadening in two notable directions: (i) splice‑switching to reprogram RNA processing and protein output, and (ii) pathway‑oriented knockdown to modulate disease‑relevant nodes (Table 3).

Table 3.

Pre-clinical evaluation of ASO-based therapeutics on neurodegenerative disorders

Target	CNS disease context & model	Modality & delivery	Mechanism/therapeutic strategy	Key outcomes in CNS model	Relevent reference
App	Alzheimer's disease, human cells and mouse model	SSO, in vitro (human), ICV (mice)	APP exon 17 skipping	Reduced Abeta 42 produdction in Down syndrome patients' fibroblast cells and wild type mouse brains	[132]
Mapt	4 repeat (4R)tauoptahy, FUS-silenced human tau knock-in mice	ENA modificed SSO, ICV	MAPT exon 10 skipping	Corrected 4R/3R ratio in FUS silenced humanized tau mice and human iPSC derived neurons	[133]
Ttbk1	Alzheimer's disease, PS19 tau mice	Gapmer ASO, ICV	RNase H1 mediated TTBK1 pre-mRNA degradation	Reduced phosphorylated tau in the mouse hippocampal tissue deteremined by ELISA and Immunofluroesence	[135]

First, splice‑switching ASOs can reshape pathogenic protein composition. For example, APP splice‑switching ASOs have been explored to reduce amyloid‑β–generating proteolytic processing by modifying APP splicing [132]. Isoform-correcting ASOs also aim to address tauopathy biology more mechanistically by adjusting MAPT splicing, particularly exon 10 inclusion, to normalize the 3R/4R tau ratio. This is especially relevant for 4R tauopathies, where 4R-dominant pathology is characteristic [133, 134]. Long-acting splice-modulating ASOs have shown preclinical efficacy in models of 4R tauopathy phenotypes, supporting isoform correction as a viable disease-modifying concept [133]. Notably, an ENA-modified ASO targeting MAPT exon 10 exhibited an approximately twofold longer brain half-life after a single bolus ICV injection than a sequence-matched MOE-modified ASO (~6 vs. ~ 3 months). Consistently, the ENA-modified ASO sustained splice modulation for up to ~100 weeks (~2 years), whereas the MOE-modified ASO showed attenuation by 24 weeks. Together, these data suggest that ENA chemistry may confer improved durability of splice-modulating activity in the CNS relative to MOE in this setting [133]. Second, ASOs are being developed against upstream modulators of tau pathology, including kinases that drive phosphorylation cascades. For example, antisense suppression of tau-tubulin kinase 1 (TTBK1) has been explored as a means to reduce pathological phospho-tau accumulation and related phenotypes in tauopathy models [135]. This approach is conceptually appealing because it may temper tau toxicity without requiring maximal reduction of tau itself, potentially preserving physiological tau functions while limiting pathological conversion.

Beyond tau-centric targets, preclinical ASO work also explores resilience and network-level modulation, including synaptic function, transcriptional regulation, and glial contributions. This reflects an emerging view that disease modification may ultimately require both suppression of toxic species and reinforcement of circuit stability [136, 137]. In PSP, integrative transcriptomic analyses have prioritized glial-enriched targets (e.g., DDR2, STOM, and KANK2), expanding the target landscape and providing a foundation for selecting PSP-oriented ASO candidates beyond MAPT [138].

siRNA/RNAi-based therapeutics for AD

siRNA therapeutics constitute a second major clinical approach. While siRNAs also aim to reduce pathogenic protein production, they act via RNA interference-mediated mRNA cleavage. In AD, clinical programs have focused on upstream drivers of amyloidogenesis, including strategies aimed at reducing APP expression and downstream Aβ production [139] (Table 4). Alnylam initiated a phase I study in 2022 to evaluate the safety and tolerability of ALN-APP, a C16-conjugated siRNA targeting APP in early onset AD patients with mild cognitive impairment or mild dementia. In 2025, Alnylam reported that a single injection at the highest dose (150 mg) showed sustained reduction of CSF sAPPβ (75%) after 6 months without any serious adverse events [140]. A phase 2 clinical trial of ALN-APP in cerebral amyloid angiopathy patients is already underway, reflecting accelerating clinical translation of CNS-targeted RNAi therapeutics in amyloid-spectrum disorders. Notably, tau-directed RNAi has advanced into clinical evaluation: Eli Lilly has initiated a first-in-human Phase 1 study of LY3954068, an intrathecally administered siRNA targeting MAPT (NCT06297590) [141]. Alongside these first clinical CNS RNAi programs, preclinical siRNA platforms are expanding into lipid metabolism and immune or glial pathways, including approaches aimed at genetic risk drivers such as APOE [142] (Table 5). Divalent siRNA designs and related architectures have enabled CNS gene silencing with encouraging biological effects in AD-relevant models, including changes in amyloid burden and immune response pathways [142]. These studies highlight siRNA’s potential to modulate disease susceptibility mechanisms, not only classical amyloid or tau nodes.

Table 4.

Clinical evaluation of siRNA-based therapeutics on neurodegenerative disorders

Drug name	Disease application	Gene target	Modality/Chemcial modification	Company	CT identifier	Phase	Stage	Key clinical findigns	Relevent reference
ALN-APP	Early onset Alzheimer's disease	APP	2′-O-hexadecyl(C16)- conjugated siRNA	Alnylam	NCT05231785	Phase I	Active, recruiting	N/A	[140]
LY3954068	Early symptomatic Alzheimer's disease	MAPT	siRNA	Eli Lilly	NCT06297590	Phase I	Active, recruiting	N/A	[141]
RAG-17	SOD1-ALS	SOD1	siRNA-accessory oligonucleotide	Beijing Tiantan Hospital	NCT05903690	Early Phase I	Completed	N/A	[55]
RAG-17	SOD1-ALS	SOD1	siRNA-accessory oligonucleotide	Ractigen Therapeutics	NCT06556394	Phase I	Active, recruiting	N/A	[164]
ALN-HTT02	Hungtinton disease	exon1-HTT	2′-O-hexadecyl(C16)- conjugated siRNA	Alnylam	NCT06585449	Phase1b	Active, recruiting	N/A	[53, 202]
LY3962681	Parkinson diseaes	SNCA	siRNA	Eli Lilly	NCT06565195	Phase I	Active, recruiting	N/A	[238]

Table 5.

Pre-clinical evaluation of siRNA-based therapeutics on neurodegenerative disorders

Target	CNS disease context & model	Modality & delivery	Mechanism/therapeutic strategy	Key outcomes in CNS model	Relevent reference
Apoe	Alzheimer's disease, 5XFAD mice	siRNA (di-siRNA or GalNAc conjugated), ICV	endogenous mouse Apoe knock down	Reduced Apoe expression and amyloid burden without affecting systemic cholesterol	[142]
Prnp	Prion disease, humanized prion over expressing mouse with inoculation of prion proteins. Non-human primates	siRNA (di-siRNA), ICV	Human PrP knock down	Profound suppression of PrP in humanized mouse models, significantly extends survival in prion-infected mice. Demonstrates favorable CNS exposure and safety in non-human primates	[188]

Aptamer-based therapeutics for AD

In contrast, aptamer-based therapeutics for AD and tauopathies are not yet prominent in late-stage clinical pipelines, even though aptamers have long-standing conceptual appeal and tested in stroke and brain tumor treatment (Table 6). However, aptamers have proven valuable in the areas of AD-related diagnostics and biosensing, leveraging the ability of aptamers to recognize specific assemblies (for example, oligomers or protofibrils) and conformers that can be challenging for conventional reagents [143–146]. Aptamer-based biosensors using electrochemical and Raman-based readouts have achieved high sensitivity for Aβ detection in biofluids [147]. Additionally, enzyme-targeting aptamers, such as those directed against BACE1, have shown reductions in amyloid-related readouts in cellular models [148].

Table 6.

Clinical evaluation of aptamer-based therapeutics in CNS disorders

Drug name	Molecular target	CNS indication/disease setting	Phase & design	CT identifier	Key clinical findings (CNS-relevant)	Relevent reference
Olaptesed pegol (NOX-A12)	CXCL12 chemokine (Spiegelmer L-RNA aptamer)	Newly diagnosed, MGMT-unmethylated glioblastoma with incomplete resection or biopsy only; first-line radiotherapy backbone (GLORIA trial)	Phase I/II, multicenter, dose-escalation single-arm study; NOX-A12 + radiotherapy with expansion cohorts adding bevacizumab or pembrolizumab	NCT04121455	NOX-A12 with radiotherapy was safe with no dose-limiting toxicities, and evidence of microenvironmental and vascular remodeling around the tumor	[259]
ARC1779	von Willebrand factor (VWF) A1-domain–binding RNA aptamer	Perioperative cerebral microembolism in patients undergoing carotid endarterectomy (CEA); stroke prevention surrogate	Randomized, double-blind, placebo-controlled trial in CEA patients evaluating perioperative IV ARC1779 vs placebo	NCT00742612	Significantly reduced cerebral microembolic signals, supporting the concept that VWF inhibition can reduce cerebral thromboembolism risk in high-grade carotid disease	[260]
BT200 (rondaptivon pegol) – first-in-human	VWF A1-domain–binding pegylated RNA aptamer	Stroke prevention–oriented antithrombotic mechanism, but tested initially in healthy volunteers (PK/PD and safety); mechanistically aimed at secondary stroke prevention by partial VWF inhibition	Phase I, randomized, double-blind, placebo-controlled SAD/MAD study in healthy volunteers	NCT04103034	Well tolerated and showed predictable, dose-dependent modulation of VWF/FVIII levels and platelet function	[261]
BT200 – stroke patient PD study	VWF A1-domain	Acute ischemic stroke (large-artery atherosclerosis); ex vivo/early clinical PD trial in stroke patients’ blood	Randomized, placebo-controlled pharmacodynamic study; BT200 added ex vivo or in a small controlled setting to examine VWF activity and platelet function in blood from acute stroke patients	Not clearly registered/no CT number reported in primary paper	Effectively inhibited VWF activity and VWF-dependent platelet function in blood from acute stroke patients	[262]

For the pre-clinical evaluation of aptamer-based therapeutics, some aptamers can bind oligomeric and hyperphosphorylated tau species with high specificity (Table 7). Early RNA aptamers demonstrated selective recognition of tau oligomers [149], while DNA aptamers targeting hyperphosphorylated tau or tau oligomers [150, 151] provided additional molecular tools to differentiate pathological from physiological tau. Such aptamers have been incorporated into highly sensitive electrochemical biosensors for tau detection in CSF and serum-like matrices [152]. These technologies support ongoing efforts to develop blood-based or minimally invasive diagnostic assays for AD, complementing established immunoassay and PET-based approaches. Therapeutic translation is active but still earlier than ASO and siRNA programs, with stability, brain exposure, and scalable manufacturing representing persistent hurdles. Accordingly, the therapeutic strategies highlighted below are largely preclinical. Within this preclinical space, aptamers are being explored for several molecular targets.

Table 7.

Pre-clinical evaluation of aptamer-based therapeutics on neurodegenerative disorders

Target/Aptamer	CNS disease context & model	Aptamer format & delivery	Mechanism/therapeutic strategy	Key outcomes in CNS model	Relevent reference
BACE1 – DNA aptamer A1	Alzheimer’s disease; Tg6799 (5xFAD-like) transgenic mice	DNA aptamer against BACE1; ICV injection	High-affinity BACE1 inhibitor; reduces amyloidogenic APP processing	↓ BACE1 and sAPPβ, ↓ Aβ₄₂ levels and plaque load; improved spatial learning and memory	[157]
AChE – DNA aptamer Ob2	Alzheimer’s disease; 5 × FAD transgenic mice	DNA aptamer to human brain AChE; repeated ICV infusion	Potent AChE inhibition (mouse and human); secondarily ↓ BACE1 and Aβ production	Restored cognitive performance, ↓ Aβ plaques, altered secretase profile and astrocyte activation	[158]
PrPc – peptide aptamer PA8	Alzheimer’s disease; 5XFAD transgenic mice	Peptide aptamer (PA8) binding cellular prion protein; chronic ICV infusion	Blocks Aβ oligomer–PrPc interaction and downstream Fyn activation; targets Aβo–PrP–Fyn neurotoxic axis	↓ Aβ oligomer binding to PrP, ↓ Aβ plaques and oligomer burden, reduced Fyn phosphorylation, gliosis and apoptotic neurodegeneration; improved learning and memory	[159]
α-Synuclein aggregates – DNA aptamers (incl. F5R2) in RVG-exosomes	Parkinson’s disease; α-syn preformed fibril (PFF)–induced mouse model	DNA aptamers recognizing α-syn PFFs packaged into neuron-targeting exosomes; ip administration	Directly bind α-syn aggregates; RVG-exosomes deliver aptamers across BBB into neurons to neutralize PFF-induced pathology	↓ pathological α-syn aggregates, rescue of synaptic protein loss and neuronal death; improved motor behavior; demonstrates exosome-aptamer platform	[243]
STAT3 – Gint4.T–STAT3 aptamer–siRNA chimera	Glioblastoma (CNS tumor); PDGFRβ⁺ GBM xenografts in mice	PDGFRβ-binding RNA aptamer Gint4.T covalently linked to STAT3 siRNA; systemic ip administration	Aptamer mediates selective delivery to PDGFRβ⁺ GBM cells; siRNA silences STAT3, a key oncogenic transcription factor in GBM	Efficient STAT3 knockdown in GBM cells; ↓ viability and migration, reduced tumor growth and angiogenesis in mouse GBM xenografts	[258]

Amyloid-β (Aβ) in AD

Aβ remains one of the most widely explored aptamer targets in neuroscience. Aptamers selected against monomeric, oligomeric, or fibrillar Aβ species show diverse binding profiles, and several have demonstrated capacity to inhibit fibrillization and reduce cytotoxicity [153, 154]. Aptamers that bind to toxic oligomeric assemblies can interfere with nucleation or neutralize toxicity in vitro, serving as a complementary approach to antibodies, especially for conformationally specific soluble species [155]. Separately, aggregation-inhibiting aptamers reduce Aβ fibrillization and cytotoxicity in vitro [156].

Beta-site APP cleaving enzyme 1 (BACE1)

Preclinical studies increasingly demonstrate that aptamers can modulate key pathogenic pathways in CNS disorders with high molecular precision. In AD, BACE1-targeting DNA aptamer A1 delivered intracerebroventricularly in Tg6799 mice reduced BACE1 activity, sAPPβ production, and Aβ42 burden, leading to attenuated plaque load and gliosis [157] (Table 7). Complementing this strategy, the AChE-targeting DNA aptamer Ob2 improved cholinergic signaling in 5 × FAD mice, restoring cognitive performance and reducing oxidative and apoptotic markers [158] (Table 7). A peptide aptamer (PA8) directed against cellular prion protein blocked Aβ oligomer binding, thereby reducing synaptic toxicity and enhancing spine density in 5XFAD mice [159] (Table 7), highlighting aptamer potential for intercepting receptor-mediated Aβ neurotoxicity.

Microtubule-associated protein tau (Tau) in AD

Aptamers selected for functional properties have been reported to inhibit tau oligomerization or aggregation and modulate pathological processes in experimental systems [151]. Early work also identified inhibitory RNA aptamers that delay tau oligomerization and confer neuroprotective effects under proteotoxic stress [149]. Notably, a circular DNA aptamer targeting both tau and transferrin receptor (TfR) penetrates the BBB and suppresses tau accumulation in a TBI-related tauopathy mouse model [160]. Though no aptamer-based AD therapeutic has yet advanced to clinical trials, advancing chemical modifications and delivery platforms strengthen the translational trajectory.

Amyotrophic lateral sclerosis (ALS)

ALS is a progressive neurodegenerative disease characterized by the selective loss of upper and lower motor neurons [161]. Clinically, ALS presents with progressive muscle weakness, atrophy, and paralysis, typically leading to respiratory failure within a few years of onset [161]. The disease is genetically and biologically heterogeneous: approximately 5–10% of cases are familial, caused by mutations in genes such as SOD1, Fused in sarcoma (FUS), and Chromosome 9 open reading frame 72 (C9ORF72), whereas the majority of cases are sporadic [161]. Until recently, treatment options were limited to minimally effective drugs; however, the emergence of gene-targeted therapies has begun to alter the therapeutic landscape of ALS.

ASO-based therapeutics for SOD1-ALS

SOD1 was the first gene identified to cause familial ALS and encodes a ubiquitously expressed antioxidant enzyme, with pathogenic mutations leading to toxic gain-of-function [162]. SOD1-associated ALS provides the strongest clinical evidence that targeted molecular therapies can modify ALS disease biology. Tofersen is an intrathecally delivered ASO designed to lower SOD1 mRNA and protein [18–20] (Table 1). In the phase 3 trial, tofersen produced robust, dose-dependent reductions in CSF neurofilament light chain (Nfl), a biomarker linked to neuroaxonal injury, though it did not meet its primary endpoint of ALSFRS-R change at 28 weeks. Importantly, longitudinal analyses incorporating the open-label extension showed that patients who initiated tofersen earlier experienced slower functional decline and better preservation of respiratory function compared with those who began treatment later [163]. Based on these biomarker data, with supportive longer-term functional findings, tofersen received accelerated approval for SOD1-ALS in several regions, establishing proof-of-principle that targeted suppression of a disease-causing gene in the CNS can alter ALS disease biology.

siRNA-based therapeutics for SOD1-ALS

In parallel with ASO-based approaches, RNAi strategies targeting SOD1 have advanced toward the clinic. One of the most mature programs, RAG-17, is an intrathecally administered siRNA-ACO (Fig. 3B and Table 4) [55]. In a small investigator-initiated, open-label study (NCT05903690), intrathecal RAG-17 was well tolerated and showed target engagement, including substantial reductions in CSF SOD1 and Nfl [55]. A randomized, double-blind, placebo-controlled phase 1 trial is now ongoing to further evaluate safety, tolerability, and pharmacokinetics/pharmacodynamics (NCT06556394) [164]. Together, tofersen and emerging siRNA approaches position SOD1-ALS as the most advanced subtype for RNA-targeted therapeutic intervention.

ASO-based therapeutics for FUS-ALS

Fused in sarcoma (FUS) is an RNA-binding protein involved in multiple aspects of RNA metabolism, including transcription, splicing, and RNA transport, and ALS-associated mutations are involved in its nuclear depletion and cytoplasmic aggregation [165, 166]. Mutations in FUS cause a rare but particularly aggressive, often early-onset form of familial ALS driven by toxic gain-of-function from mislocalized FUS protein [167]. In preclinical mouse models, non-allele-specific reduction of Fus using ASOs prolongs survival, preserves motor neurons, and reduces insoluble protein accumulation, indicating that partial lowering of FUS is tolerated and therapeutically beneficial [168]. Consistent with these findings, early open-label studies have demonstrated reductions in CSF FUS and suggest possible slowing of functional decline in some patients [169]. Together, these preclinical and early clinical observations enabled the clinical development of the FUS-lowering ASO ION363 (ulefnersen/jacifusen) [168] (Table 2), which is currently being evaluated in a randomized phase 3 trial (NCT04768972).

ASO-based therapeutics of C9ORF72-ALS/FTD

Hexanucleotide (G₄C₂) repeat expansions in C9ORF72 are the most common genetic cause of ALS and ALS–frontotemporal dementia (FTD). Because C9ORF72 haploinsufficiency and loss-of-function mechanisms may also contribute to disease, antisense strategies have largely focused on selectively suppressing repeat-containing transcripts while sparing the normal C9ORF72 transcript. This selective approach aims to reduce RNA foci and dipeptide repeat protein (DPR) production without exacerbating potential loss-of-function effects. In patient-derived neurons and C9-BAC transgenic mouse models, these approaches robustly suppress toxic RNA species and rescue cellular and behavioral phenotypes [45, 170], and notably, one study further reported that the ASO treatment reduced CSF poly(GP) levels in one ALS patient [45]. However, two intrathecal ASOs (BIIB078 and WVE-004) failed to demonstrate clinical benefit in symptomatic patients despite reducing CSF poly (GP) [171](Table 2). The reasons for these negative outcomes remain uncertain and are likely multifactorial. Potential explanations include: (i) insufficient modulation of disease-relevant pathology in vulnerable CNS tissues, as postmortem analyses of BIIB078-treated participants showed persistent DPR and phosphorylated TDP-43 pathology [172]; (ii) the inability of sense-targeting approaches to address antisense repeat RNA and antisense-derived DPRs, which may exert substantial pathogenic effects [173] [174]; (iii) initiation of treatment after downstream pathological cascades had already become established [173]; and (iv) contributions from additional disease mechanisms not directly ameliorated by sense repeat RNA lowering, including C9orf72 loss of function [175].

ASO-based therapeutics for sporadic ALS

In contrast to genetically defined familial ALS, therapeutic strategies for sporadic ALS have focused on downstream consequences of TDP-43 proteinopathy. Although no single causative gene has been identified, loss of nuclear TDP-43 is seen in > 95% of sporadic ALS cases and is regarded as a hallmark pathology.

Stathmin-2 (STMN2)

Nuclear TDP-43 depletion induces widespread cryptic exon mis-splicing, and among the transcripts most affected by TDP-43 loss is Stathmin-2 (STMN2), a protein critical for axonal maintenance and regeneration. Loss of TDP-43 permits inclusion of a cryptic exon in STMN2 pre-mRNA, introducing a premature stop codon and leading to reduced full-length STMN2 protein together with production of a truncated STMN2 transcript [176, 177]. Accordingly, splice-modulating ASOs that mask the cryptic splice/polyadenylation site restore full-length STMN2 mRNA and protein in human motor neurons with TDP-43 dysfunction and rescue defects in axonal outgrowth and regeneration [178]. These preclinical data underpin the ongoing ANQUR trial of QRL-201, the first STMN2-targeting ASO in ALS (NCT05633459) (Table 2).

Unc-13 homolog A (UNC13A)

In parallel, UNC13A has emerged as another therapeutically actionable TDP-43 loss of function-linked target; ALS-associated UNC13A variants that promote inclusion of a TDP-43-dependent cryptic exon provide genetic support for the disease relevance of this missplicing pathway [179, 180]. Splice-modulating ASOs that target the UNC13A cryptic exon can restore UNC13A protein levels and rescue synaptic deficits in human iNeurons following TDP-43 knockdown [181]. Building on this rationale, QurAlis and Eli Lilly are conducting a first-in-human study of an intrathecally administered ASO targeting UNC13A (NCT07100119) (Table 2). Trace Neuroscience has also disclosed an UNC13A-restoration ASO program and has indicated plans for a Phase 1/2 clinical trial.

Ataxin 2 (ATXN2)

ATXN2 has also been investigated as a modifier of ALS risk and TDP-43 pathology [182]. Although preclinical studies showed that lowering Atxn2 improves survival and motor function in TDP-43 mouse models [183], the ATXN2-targeting ASO BIIB105 failed to demonstrate clinical or biomarker benefit in a phase 1/2 ALS trial and was discontinued (NCT04494256) (Table 2).

Prion diseases

ASO-based therapeutics for prion disease

Prion diseases are rapidly progressive, fatal neurodegenerative disorders in which propagation of misfolded prion protein (PrP) critically depends on the presence of endogenous cellular PrP [184, 185]. This unique biology makes endogenous PrP lowering an especially compelling therapeutic strategy, as even partial reduction of PrP is predicted to slow or halt disease progression. Initial proof-of-concept was established using Prnp-targeting gapmer ASOs, which produced dose-dependent suppression of brain PrP, marked delays in disease onset, and extensions of survival in multiple prion-infected mouse models [186, 187]. Building on these data, the first human PrP-lowering ASO, ION717, has advanced into a phase 1/2a clinical trial (NCT06153966) (Table 2) in symptomatic patients with prion disease. This study is designed to assess safety, pharmacokinetics, and cerebrospinal fluid PrP reduction as a pharmacodynamic marker.

siRNA-based therapeutics for prion disease

In parallel, RNAi-based approaches have rapidly advanced, with divalent siRNA emerging as the most advanced siRNA modality for prion disease. A PrP-lowering divalent siRNA series has been identified and optimized, culminating in candidate 2439-s4, which achieves deep and durable suppression of PrP across mouse and humanized mouse models, significantly extends survival in prion-infected mice, and demonstrates favorable CNS exposure and safety in non-human primates [188](Table 5). Based on these encouraging data, 2439-s4 has now been nominated as a clinical development candidate, with an FDA-cleared IND for a first-in-human trial in patients with prion disease.

Huntington’s disease (HD)

Huntington’s disease is an autosomal dominant neurodegenerative disorder characterized by progressive motor dysfunction, cognitive decline and psychiatric disturbance [189]. HD is caused by CAG repeat expansion in the huntingtin (HTT) gene, which encodes an expanded polyglutamine tract in the huntingtin protein. Therapeutically, lowering HTT expression has been pursued through both non-allele-selective and allele-selective strategies.

ASO-based therapeutics for HD

Non-allele-selective (total HTT lowering) ASOs: Non-allele-selective HTT lowering (total HTT lowering) has substantial preclinical evidence supporting its feasibility and potential benefit. Early work using RNA interference-based silencing demonstrated that partial suppression of both mutant and wild-type HTT in the adult rodent brain can be tolerated and confer benefit [190]. Importantly, consistent with this proof-of-concept, preclinical benefit has also been demonstrated with nucleic acid therapeutics, including HTT-targeting gapmer ASOs, which showed durable HTT lowering and phenotypic improvement in HD models [191].

Building on these preclinical data, the first major clinical ASO trial for HD evaluated tominersen (IONIS-Roche), a gapmer ASO that lowers HTT in a non-allele-selective manner [192](Table 2). In this phase 1/2a multiple-dose-ascending study, intrathecal administration across dose cohorts (up to 120 mg) produced dose-dependent reductions of mutant HTT in CSF in early-stage HD participants and was generally well tolerated [192]. In 2021, a subsequent Phase 3 tominersen trial was discontinued due to an unfavorable overall benefit-risk profile, with a lack of evident clinical benefit compared with placebo [193]. However, Roche initiated a new phase 2 study in prodromal or earlier stage participants in 2023, based on the analyses from the halted program, which suggested potential benefit in early-stage HD patients (GENERATION HD2, NCT05686551). This study is expected to be completed in April 2027. Collectively, these preclinical and early clinical observations have motivated continued clinical evaluation of non-allele-selective approaches. However, the overall clinical benefit–risk profile of non-allele-selective HTT lowering remains to be fully defined [194, 195]. A key consideration is that wild-type huntingtin is neuroprotective and essential for normal cellular functions, and broader suppression of total HTT has raised safety concerns in some contexts, highlighting the need to balance efficacy with preservation of wild-type HTT function.

Allele-specific ASOs: Accordingly, many HTT-lowering efforts have pursued mutant allele-selective approaches designed to spare wild-type HTT [196]. One approach to allele selective HTT suppression is to target polymorphisms linked to the expanded CAG allele. Southwell et al. reported that ~93% of identified HTT SNPs are intronic [197], suggesting that ASO targeting of nuclear HTT pre-mRNA, an approach not generally achievable with siRNA, may be a practical strategy for allele-selective silencing [196]. The same team developed ASO targeting HD-enriched SNPs to preferentially suppress mutant HTT [198], helping establish SNP-based allele selective targeting in HD. In parallel, Wave Life Sciences’ phase 1b/2a trials of allele-selective ASO targeting HD-associated SNPs (SNP1 rs362307 and SNP2 rs362331), variants carried by 65 ~ 70% of individuals with HD of European ancestry, were terminated in 2021 (Wave) due to the inconsistent outcomes across the participants. More recently, a pre-clinical study reported that PS- and PN-containing ASO targeting the HD-associated SNP rs362273 (SNP3) selectively lowered mutant HTT while sparing wild type HTT [199]. Consistent with this concept, Wave (co-sponsored by Takeda) initiated a phase1b/2a clinical trial in 2021, which was completed in 2024 (WVE-003, NCT05032196) (Table 2). The company reported in 2025 that intrathecal administration of an SNP3 targeting ASO (30 mg, every 8 weeks, 28 weeks total study period including12 weeks follow up) reduced the CSF mutant HTT by a mean of 46% versus placebo, while wild-type HTT concentrations remained comparable to placebo/control groups [200]. The company has indicated plans to initiate a phase 2/3 trial of WVE-003 in HD patients [201].

siRNA-based therapeutics for HD

Non-allele-selective HTT-lowering concepts are also being pursued using siRNA modalities. Using the di-siRNA platform, Alterman et al. reported that a single ICV dose achieved broad and durable suppression of HTT mRNA and protein in mouse and non-human primate brain, consistent with long-lasting CNS target engagement [53] (Table 4). In addition, Alnylam’s C16-conjugated siRNA (ALN-HTT02) has shown broad CNS distribution and HTT lowering in NHP [202], and a phase 1b clinical trial with intrathecal single-ascending-dose is ongoing (NCT06585449) (Table 4). Of note, ALN-HTT02 targets a conserved exon 1 sequence and is intended to lower total HTT, including HTT1a (described below). Beyond HTT lowering, several additional disease-modifying strategies are being explored at the preclinical and early clinical stages. One such approach aims to slow or prevent somatic CAG repeat expansion by targeting the DNA mismatch repair gene MSH3 [203, 204], which is increasingly recognized as a key driver of HD pathogenesis [205, 206]. This strategy remains predominantly preclinical and has not yet reached clinical testing. Second, growing evidence suggests that an aberrant exon-1 HTT RNA/protein (often referred to as HTT1a) generates highly toxic HTT products [207–209], motivating exon-1-focused interventions [210]. Third, repeat-centric approaches seek allele-preferential engagement by targeting expanded CAG repeats. These include steric-blocking ASOs that bind expanded CAG repeat RNA to suppress mutant polyQ protein production without inducing mRNA degradation [211, 212]. Notably, this strategy has advanced into early clinical testing with the CAG repeat–targeting steric-blocking ASO [213] (Vico Therapeutics, VO659, NCT05822908) (Table 2).

Parkinson’s disease

Parkinson disease (PD) is the second most common neurodegenerative disease following AD and affects approximately 500,000 individuals in the United States. Neuropathological hallmarks of PD include misfolded alpha (α)-synuclein aggregation in Lewy bodies(LBs) and Lewy neurites, along with progressive loss of dopaminergic neurons in the substantia nigra pars compacta [214–216]. A genome wide association study in PD patients and control suggest that variants in the SNCA gene, encoding α-synuclein, and the LRRK2 gene, encoding leucine-rich kinase 2, modulate the risk of PD [217, 218]. Despite major advances in symptomatic therapy, PD still lacks approved disease-modifying treatments. Current pre-clinical or clinical therapeutic strategies largely focus on reducing α-synuclein burden by preventing α-synuclein aggregation and/or enhancing its clearance or degradation. We will highlight two approaches, ASOs targeting LRRK2 and SNCA, which are both aimed to reduce a-synuclein production.

ASO-based targeting of LRRK2

LRRK2 mutation is the most prevalent familial PD gene mutations. Previous studies reported that G2019S LRRK2 autosomal dominant mutation were identified with 6.6% of familial [219] and 1.6% sporadic PD [220]. The LRRK2 gene is located on chromosome 12, encoding a large 286 kDa multidomain protein, which functions as a kinase and guanosine triphosphatase (GTPase) [221]. Familial linked mutations of LRRK2 gene increase kinase activity by inducing both autophosphorylation and phosphorylation substrates [222]. LRRK2 inhibition is widely considered among the most promising therapeutic approaches for PD. LRRK2 deletion [223] and kinase inhibition [224, 225] have been shown to mitigate α-synuclein pathology and related neurodegeneration in multiple preclinical models. However, systemic LRRK2 kinase inhibition can reduce LRRK2 activity in multiple tissues and has been associated with off-target or peripheral phenotypes, including changes in the kidney [226] and lung [227]. CNS restricted delivery of ASO targeting LRRK2 could avoid peripheral phenotypes seen with systemic LRRK2 modulation. A previous pre-clinical study showed that ICV injection of a Lrrk2-targeting ASO in mouse brains reduced the Lrrk2 protein level and diminished fibril induced α-synuclein inclusion in mice [228]. Following this work, preclinical efficacy and safety data of a LRRK2 targeting-ASO in NHPs were reported [229]. ICV injection of LRRK2 targeting ASO in cynomolgus macaques demonstrated that ASO selectively and potently reduced LRRK2 mRNA and its protein levels across the brain regions. The lead ASO was well tolerated and, notably, did not reduce LRRK2 mRNA or produce pathology in the lung, a target organ for LRRK2. BIIB094 (ION859) is the most clinically advanced LRRK2-targeting ASO for Parkinson’s disease, administered intrathecally and designed to reduce LRRK2 mRNA. This ASO has been evaluated in the phase 1 REASON study (NCT03976349) (Table 2), which started in 2019 in participants with PD with or without LRRK2 variants. Pharmacodynamic analyses from the REASON study describe dose-dependent decreases in CSF LRRK2 and CSF phosphorylated Rab10, a downstream substrate of LRRK2 kinase activity, irrespective of LRRK2 variant status, alongside shifts in endolysosomal/lysosomal protein markers measured by LC–MS multiplex assays, supporting central target engagement and downstream pathway modulation in treated participants [230].

While gapmer ASOs aim to lower transcripts to reduce total protein, SSOs can be designed to specifically attenuate pathogenic signaling while preserving expression of other protein regions. LRRK2 exon 41 encodes a region of the kinase domain that includes Gly²⁰¹⁹, the residue mutated to serine in the common pathogenic G2019S variant. In transgenic mice expressing human LRRK2 (G2019S) or LRRK2 (WT), a single ICV injection of an exon 41 SSO modestly induced exon skipping for 10–25%, generating the kinase-deficient LRRK2 isoform, and partially reduced phosphorylation of the LRRK2 kinase substrate Rab10, while also modulating autophagy related markers in human LRRK2 WT-expressing mice [231]. Exon 41 splice switching ASO also resulted in recovered mitophagy function in PD-patient derived fibroblast cells. Exon 41 splice switching ASO also normalized altered endoplasmic reticulum (ER) calcium levels in PD patient induced pluripotent stem cell (iPSC)-derived neurons [232], suggesting potential translatability.

ASO-based targeting of SNCA

A second major axis is SNCA targeting ASOs intended to reduce α-synuclein production, which helps preventing accumulation of mutant or misfolded forms of the protein. In pre-clinical models, ASO targeting mouse Snca reduced α-synuclein protein production in pre-formed fibrils treated primary cultured neurons and increased neuronal survival [233]. Following this study, it has been shown that an ASO targeting mouse Snca reduced α-synuclein production in rodent preformed fibril models, prevented or removed established pathology, and mitigated dopaminergic dysfunction [234]. The same study also reported an ASO targeting human SNCA with broad distribution in NHP and associated lowering of CSF α-synuclein, supporting CSF α-synuclein as a candidate pharmacodynamic biomarker for clinical translation [234]. While promising, a, b, g triple synuclein knockout mice previously exhibited reduction in excitatory synapse size by ~30% both in vivo and in vitro, and old-age knockout mice showed significantly reduced survival rate [235]. A recent study also suggests that α-synuclein is critically involved in neurotransmitter release by modulating the process of synaptic vesicle clustering and docking [236]. Thus, SNCA-targeting ASOs needs careful optimization of the dosing to avoid side effects from exaggerated pharmacology.

ASO-based targeting of SNCA for MSA therapy

A Phase 1 clinical trial of BIIB101, also known as ION464 began in July 2022 (HORIZON, NCT04165486) (Table 2) to evaluate ION464, an SNCA targeting ASO developed by Ionis and co-sponsored by Biogen, in ~40 participants with multiple system atrophy (MSA). Eligible participants were required to have evidence of dopaminergic terminal loss on DaTscan without cognitive impairment. The primary objectives are to assess the safety and tolerability of multiple ascending intrathecal doses of ION464 and to evaluate long-term safety/tolerability over 72 weeks. The study is expected to complete in September 2027 across 10 sites in Austria, France, Germany, and the U.K. Interim results presented at the September 2024 International Congress of Parkinson’s Disease and Movement Disorders reported that ION464 was safe and well tolerated, with no drug-related serious adverse events or deaths. On Feb 12, 2025, Biogen announced it would discontinue development of ION464/BIIB101[237]. As of October 2025, Ionis still listed the program in its pipeline.

siRNA-based targeting of SNCA

Currently, LY3962681 is an investigational intrathecal siRNA therapy designed to reduce α-synuclein by targeting SNCA mRNA. Publicly disclosed preclinical summaries report dose-dependent and sustained lowering of SNCA mRNA and α-synuclein protein in brain regions including substantia nigra and dorsal striatum of NHP, with α-synuclein reductions in CSF, and a tolerable GLP toxicology profile [238]. Clinically, LY3962681 is being evaluated in a Phase 1 single ascending dose and multiple ascending doses study (NCT06565195) (Table 4) in healthy volunteers and patients with Parkinson’s disease, primarily assessing safety/tolerability and PK/PD including CSF α-synuclein pharmacodynamics following intrathecal administration.

Aptamer-based targeting of SNCA

In Parkinson’s disease (PD), aptamers have shown utility in both detection and inhibition of α-synuclein aggregation. Aptamers can differentiate among α-synuclein fibril polymorphs derived from distinct synucleinopathies [239], enabling nuanced molecular distinction. Functional studies demonstrate inhibition of fibrillization and reduction of intracellular α-synuclein burden: aptamers such as F5R2 block fibril uptake and downstream toxicity [240], while other DNA aptamers interfere with early nucleation steps [241]. Additional work highlights aptamer-mediated facilitation of lysosomal clearance of α-synuclein aggregates [242], supporting therapeutic interest. In PD models, aptamers have demonstrated potent anti-aggregation effects on α-synuclein. High-affinity DNA aptamers identified by Zheng et al. inhibit fibrillization, reduce intracellular α-synuclein accumulation, and mitigate mitochondrial dysfunction [240]. DNA aptamers that recognize aggregated α-synuclein [243] (Table 7) reduced pathological burden and preserved synaptic and motor function when delivered via RVG-exosome carriers in α-syn preformed fibril mouse models, demonstrating an effective CNS-targeted delivery paradigm.

Key challenges and future directions of RNA therapeutics

Several factors will likely determine whether RNA therapeutics can achieve durable clinical impact in neurodegenerative disorders. First, achieving adequate exposure in relevant CNS cell types and brain regions, while maintaining tolerability, remains essential, particularly as doses and dose intervals are optimized for chronic administration. In the near term, IT delivery will likely remain the dominant route for CNS-focused RNA therapeutics in CNS disorders. However, dual-TV platforms, optimized dosing regimens (e.g., extended intervals based on long ASO half-life in brain), and less invasive lumbar delivery techniques may improve tolerability and accessibility. Ultimately, integrating receptor-mediated BBB shuttles with RNA payloads could enable subcutaneous or intravenous administration of precision RNA drugs for neurodegenerative diseases. Second, increased durability of target suppression could enable longer dosing intervals and improve feasibility for patients and health systems. Pre-clinical studies have reported sustained pharmacodynamic effects with divalent-exNA [72] or C16 conjugate RNAi scaffolds [57], as well as ENA modified ASOs [133] in rodents or NHPs. Combination strategies may ultimately be required to achieve larger and more sustained effects, for example pairing Aβ-lowering antibodies with tau-directed ASOs or integrating APOE-targeted RNAi with tau isoform correction. The third challenge is inter-individual variability, which may be amplified in aged and diseased brains by differences in neuroanatomy, blood–brain barrier integrity, cellular uptake pathways, comorbidities, and baseline inflammatory state; these sources of variability must be characterized to improve dose selection and predictability of target engagement. Fourth, chronic CNS RNA modulation raises safety considerations, including off-target hybridization, innate immune activation, and unintended network effects arising from sustained perturbation of gene expression programs. In this context, isoform-correcting strategies may offer a more physiologically nuanced alternative to maximal knockdown but require careful validation across heterogeneous biology [133]. Finally, patient selection and timing are likely to be decisive for clinical success. Experience from prior programs (e.g., tominersen in HD disease) and from disease modifying efforts in AD (e.g., Aβ immunotherapy) suggests that earlier intervention may yield larger and more consistent benefit in appropriately selected patients. To enable this, blood-based biomarkers are expected to play expanding roles in screening, stratification, and longitudinal monitoring [139, 244].

Pathway to clinical implementation

As detailed above, two RNA therapeutics are now approved for the treatment of neurological diseases, with dozens more in clinical development. For many common diseases and commercial applications, the path to clinical implementation remains time-consuming and resource intensive: after extensive compound screening, regulators normally require animal safety testing in two species. A traditional drug development path requires extensive compound screening in multiple steps. It typically begins with a broad screen of dozens to thousands of oligonucleotides across the full target space (namely, the whole genomic locus (pre-mRNA) in the case of RNase H ASOs, the targeted exon and proximal regulatory regions for splice modulating ASOs, or the mature mRNA in the case of siRNAs). Initial sequence selection for the screens can be aided by online servers for siRNAs, gapmers and SSOs as described above. When hits are identified, additional screens are carried out such as a “micro walk” (moving the target sequence slightly upstream and downstream) and optimization of the precise pattern of chemical modifications. Because each sequence typically responds slightly differently to chemical modification, this sequence-specific approach to the final pattern of chemical modification is important to turn a good lead into a truly optimal drug. Later stages of screening should be carried out in vivo, when possible, as the most potent compounds in cells are not always the best in vivo. After small and large animal testing, companies apply for permission from regulators to initiate human trials, typically beginning with healthy volunteers. This overall process is demanding and typically takes at least three years even if everything goes smoothly [245].

The fact that oligonucleotide drugs of each class have some shared properties across different sequences has allowed the development of patient-customized ASOs to be developed much more rapidly. For example, after a deep intronic mutation in CLN7 was discovered as the cause of Batten’s disease in a single pediatric patient, investigators developed a patient-customized SSO using the same chemical modification pattern as nusinersen, showed that it restored CLN7 expression in patient-derived cells and was safe in rodents, and succeeded in treating the patient within 10 months of her genetic diagnosis. The treatment was well tolerated and resulted in an > 80% reduction in seizures [246]. This experience began to shift the paradigm of drug development for severe neurological diseases: It became possible to envisage drug development that works at a suitable pace to treat the patient in front of us (rather than only some future patient) even in the case of rapidly progressing and uniformly fatal diseases. As such, non-profit groups such as the N = 1 Collaborative [247], n-Lorem [248], 1 Mutation 1 Medicine (1M1M) [249] and the Dutch Center for RNA Therapeutics [250], along with companies such as EveryONE Medicines, began to study how to scale this paradigm. As momentum built in this field, the US FDA released draft guidance regarding the clinical development of individualized ASOs for severe neurological diseases [251].

The need remains vast – with an estimated 10,000 recognized monogenic rare diseases – and the development of these therapies remains time- and resource-intensive. But to date, more than 25 single-patient trials have been carried out using customized n-of-1 or n-of-few RNA therapies [245]. With rare diseases, the lines are not always perfectly sharp between personalized drugs and more traditional development pathways. For example, in one case, a drug that started its clinical life as a rapidly developed single-patient trial for a young patient with aggressive FUS-driven ALS [168] was later adapted to be used in about 10 patients on an expanded access basis [169] before being transitioned to a traditional Phase III trial [252]. While ASOs are the most mature RNA-based therapeutic technology, they may not ultimately be the easiest to scale and adapt to new patients. The tolerability of CNS-directed ASOs can be highly dependent on sequence, chemistry, dose, formulation, and preclinical species [31, 34, 37, 253, 254]. PS-modified ASOs, particularly gapmers, can cause acute or delayed neurotoxicity in preclinical models [31, 253, 254], while ventricular enlargement/hydrocephalus and inflammatory neurological complications, such as radiculitis, have emerged as clinical safety signals in some CNS ASO programs [193, 255, 256]. These liabilities may constrain the therapeutic window and complicate clinical development. siRNAs, by virtue of their predictable double-stranded structure with the bases stacked together on the interior of the duplex, tend to show more predictable pharmacokinetics across diverse sequences than ASOs do [257]– this provides a crucial advantage if we are to minimize slow and resource-intensive animal testing and focus our efforts on rapid patient-specific sequence discovery and clinical translation.

Conclusion

RNA therapeutics are rapidly transitioning from conceptual promise to clinical reality for neurodegenerative disorders. ASOs currently lead the clinical landscape, with an already FDA-approved drug for SOD1-associated ALS and tau-lowering programs demonstrating robust target engagement in early AD and expansion toward primary tauopathies. A growing number of clinical trials are also evaluating the drug safety and efficacy of ASOs in patients with Huntington disease, prion diseases, and synucleopathies, with early signals of target engagement and potential therapeutic benefit. siRNA programs are progressing toward upstream and genetic risk pathways, with improving delivery technologies enabling broader target space. Although no siRNA therapeutics have yet received FDA approval for neurodegenerative indications, their favorable stability and potential for prolonged target suppression through chemical modulation and conjugation make them an increasingly compelling modality for CNS translation. Finally, Aptamers, while earlier in clinical development, offer a versatile platform with particular promise for conformer-selective targeting and diagnostic strategies. Across modalities, the next phase of progress will depend on rigorous linkage of molecular effects to clinically meaningful outcomes, enabled by increasingly sensitive and scalable biomarker platforms.

Authors contributions

Manuscript writing and editing; M.O., S.I., J.K.W., K.Y., T.I., and Seiko.I. All authors have read and approved the final version of the manuscript.

Funding

This work was funded in part by NIH RF1 AG082704 (Seiko.I., T.I), R01 AG054199 (T.I.), R01 AG066429 (T.I.), R01 AG067763 (T.I.), R01 AG072719 (T.I.), R01 AG079859 (Seiko.I.), Alzheimer’s Association ZEN-26–1436830 (T.I.), JSPS KAKENHI: JP25K02555 (S.I) AMED: Grant Numbers JP26ym0126184 and JP26wm0625522 (S.I.), MJFF-022463, Michael J Fox Foundation for Parkinson's Research, Translational Pipeline Program — Pre-Clinical (K.Y.), R01 NS111990 (J.W.)

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interest

Scientific advisory board of AAVINUE and consults Takeda and Otsuka (T.I.). Scientific advisory board of EnPlusOne Biosciences, EveryONE Medicines, M2DS Therapeutics, Nucyrna Therapeutics, PepGen and Sixfold Bioscience (J.W.). Seiko Ikezu is an Editorial Board Member for Molecular Neurodegeneration Advances and was not involved in the editorial review or decisions regarding this manuscript.