Antisense Therapies in Neurological Diseases

Jean-Baptiste Brunet de Courssou; Alexandra Durr; David Adams; Jean-Christophe Corvol; Louise-Laure Mariani


Brain. 2022;145(3):816-831. 

In This Article

Clinical Applications of Antisense Therapies

Antisense therapies at the clinical stage of development are described in Table 2, Figure 3 and Supplementary Table 1.

Figure 3.

Timeline of the most advanced antisense therapies under development in neurology. Trials are showed beginning at the 'Actual Study Start Date' and ending at the Actual or 'Estimated Study Completion Date' as indicated in the database. First date of acceptance by either US Food and Drugs Administration (FDA) or European medicines Agency (EMA) was used when both agencies approved the drug. CMV = cytomegalovirus.

Diseases With Completed or Ongoing Phase 3 Trials and Marketing Authorization

Spinal Amyotrophy. Pathophysiology: Spinal muscular atrophy is a group of autosomal recessive diseases characterized by a loss of the lower motor neuron, located in the anterior horn of the spinal cord, leading to a progressive decrease in motor function. Its prevalence is around 1 for 6000–55 000 live births.[36] Spinal muscular atrophy is divided into four types according to disease severity. The most severe type 1 spinal muscular atrophy includes 60% of the patients. Children with type 1 spinal muscular atrophy never have the ability to sit unaided and 90% die before 20 months.[37]

The genetic cause is a homozygous deletion of exon 7 in SMN1 gene with a loss of function of the SMN1 protein. SMN2, a homologue gene of SMN1, is present in varying copy numbers. SMN2 and SMN1 differ by only five nucleotides. One of these differences is c.840C > T, in exon 7, that leads to alternate splicing of SMN2: one without exon 7 (85%) and others with exon 7 (15%), leading to a functional protein that can help mitigate the lack of SMN1 in patients.[38]

Owing to the known pathophysiology, two distinct approaches have been developed: gene therapies to express a new functional SMN1 gene, and splicing modification of SMN2 pre-mRNA with ASO to favour exon 7 expression.

Gene therapy is beyond the scope of this review per se, but this emblematic first successful strategy deserves to be mentioned, using an AAV9 viral vector to deliver a functional copy of the SMN1 gene in patients without genomic modification.[20]

SMN2 Splicing Modification: Nusinersen is a 2'-MOE and PS 18-mer ASO delivered intrathecally, pairing with a specific region (ISS-N1, the intronic splicing silencer N1 region) of intron 7 in SMN2 pre-mRNA. Pairing favours exon 7 conservation during splicing.

Two phase 3 trials have been completed. The ENDEAR (NCT02193074) trial included 122 children below 7 months of age with type 1 spinal muscular atrophy, treated by either intrathecal nusinersen or placebo with a 2:1 ratio. Intrathecal injection was performed at Days 1, 15, 29 and 64 followed by a maintenance dose at Days 183 and 302. Primary end points were a motor-milestone response using the Hammersmith Infant Neurological Examination (HINE) score section 2 and the event-free survival, corresponding to time to death or use of permanent assisted ventilation. The HINE score section 2 ranges from 0 to 26 with a higher score indicating better motor function. Motor response in this study was defined according to seven of eight categories composing the HINE score (voluntary grasp excluded), with an improvement in at least one category (increase ≥1 point in the head control, rolling, sitting, crawling, standing, or walking score or increase ≥2 points in the score for kicking) and more categories with improvement than with worsening. Interim analysis showed evident benefit: 41% of treatment group had a motor-milestone response compared to 0% in the placebo group (P < 0.001). Final analysis showed an improved survival in the treatment group. There was no serious adverse event with nusinersen, but an observed increase in constipation and various infections (respiratory and urinary tract).[39] An extension study is in progress (SHINE study, NCT02594124).

The second phase 3 trial, the CHERISH trial (NCT02292537), included 126 children aged 2–9 years with type 2 spinal muscular atrophy and treated on days 1, 29, 85 and 274 with 12 mg intrathecal nusinersen or placebo with a 2:1 ratio. The primary end point was the mean change from baseline to 15 months in the Hammersmith Functional Motor Scale–Expanded, a 33-item measure of motor function ranging from 0 to 66 points, with higher scores indicating better motor function and a change of at least 3 points considered clinically relevant. The interim analysis showed improvement in the treated group (+4.0 points versus −1.9 points). The final analysis confirmed stabilization in 57% of the treated children compared to 26% in the untreated group, less drastic than for spinal muscular atrophy type 1 children.[40]

To date, data in adults with milder forms (types 3 and 4) of spinal muscular atrophy are scarce. Treatment with nusinersen during 14 months in 16 adults with types 3 and 4 spinal muscular atrophy prospectively enrolled suggest an improvement in hand grip strength, hand motor function and Medical Research Council (MRC) sum scores.[41] Favourable results with good safety profile was also reported in 54 patients with types 1–3 spinal muscular atrophy.[42] Similarly, in prospective cohorts, with seven patients with types 2 or 3 spinal muscular atrophy treated for at least 10 months, and with six adults with type 3 spinal muscular atrophy treated for 12 months, results suggested symptom improvement.[43,44]

Long-term efficacy and side effects of this treatment are unknown. Some cases of non-communicating hydrocephalus have been reported[45,46] with unknown mechanism, as described with other ASOs in Huntington's disease.[47]

Duchenne Muscular Dystrophy. Pathophysiology: Duchenne muscular dystrophy is the most frequent X-linked muscular dystrophy with an incidence of 1 per 5000 males. Symptom onset occurs during childhood in most cases with delayed acquisition of motor developmental milestones such as delayed walking and an impossibility to run, followed by a slow motor decline starting at a mean age of 7, leading to premature death by respiratory or cardiac complications. A milder form exists, Becker's muscular dystrophy, where disease starts later in life. Mutations in the DMD gene lead to loss of dystrophin, a protein anchoring myocytes' cytoskeleton to the extracellular matrix.

Gene Therapy: The size of the DMD gene is a major hurdle, because it is 2.2 million base pairs long and its mRNA is 14 kb, when viral vectors used for gene therapy have a load capacity of roughly 4.5 kb. A proposed solution is to provide a shortened dystrophin protein or 'microdystrophin', to mimic Becker's milder form. Microdystrophin is currently being tested with an AAV9 vector with PF-06939926 therapy (ongoing phase 1 study NCT03362502, then planned phase 3 study NCT04281485) and in the IGNITE DMD phase 1/2 study (NCT03368742). Microdystrophins are partially functioning proteins and these approaches will be insufficient to completely halt clinical evolution.

Another solution is to rescue DMD deficiency by overexpressing other genes. GALGT2 gene overexpression using an AAVrh vector is in progress (NCT03333590).[48]

Exon Skipping: Most mutations in Duchenne muscular dystrophy patients lead to a reading frame shift by either base deletion or insertion, or larger gene rearrangements like duplication or deletion, occurring preferentially in some hotspots along the gene. Excluding the exon that harbours the mutation prevents this shift and leads to a shortened dystrophin protein. ASOs are being developed for exon 51 skipping, for 10–15% of Duchenne muscular dystrophy patients.[49–51]

Eteplirsen (also known as 'AVI-4658' or 'EXONDYS 51') is a 30-nucleotide PMO delivered intravenously that has received the US Food and Drugs Administration (FDA) approval but has not yet received European Medicine Agency (EMA) clearance. The open-label phase 2 study in 13 Duchenne muscular dystrophy patients ranging from 6 to 13 years old did not report any serious adverse event and in seven patients showed a doubling of dystrophin-related fluorescence on muscle biopsy, raising from 8.9% at baseline to 16.4% of normal levels after treatment.[52] A double-blind randomized controlled study was conducted in 12 Duchenne muscular dystrophy patients ranging from 7 to 13 years of age and confirmed a significant increase of 22.9% in dystrophin levels compared to baseline when treated with 30 mg/kg for 24 weeks. The extension study for a total cumulated treatment time of 3 years confirmed treatment safety, with occasional transient proteinuria. Regarding the secondary end points, the 6-minutes walking test seems to worsen less when treated: there was a decrease of 68.4 m (±37.6) at Week 48 compared to baseline for the placebo cohort and an increase of 21 m (±38.2) in the 50 mg/kg eteplirsen cohort.[53] This possible effect is also seen when comparing the 12 patients to historical cohorts.[54] A phase 3 trial (MIS51ON trial, NCT03992430) is actively recruiting to further prove clinical efficacy.

Similarly, studies are in progress to assess SRP-5051, a peptide-conjugated ASO designed to improve muscle delivery of the ASO, also aiming at exon 51 skipping. The phase 1 Single Ascending Dose trial (NCT03375255) in 15 patients was completed in August 2019 with pending results and the phase 2 Multiple Ascending Dose trial (NCT04004065) in 70 patients is ongoing.

Weekly intravenous injections of drisapersen—a 2'-MOE and PS 20-mer ASO—were administered in a phase 2 double-blind study on 51 patients. Safety data were unremarkable and some clinical secondary criteria such as the 6-minutes walking test were encouraging.[55] However, a 48-weeks phase 3 controlled randomized double-blind study of drisapersen 6 mg/kg weekly in 186 Duchenne muscular dystrophy males did not confirm efficacy on walking improvement and led to drug development discontinuation.[56] Of note, 64% of patients had renal adverse events, mainly proteinuria as seen with eteplirsen. This proteinuria seems to be a frequent adverse effect of these ASO. In rhesus and cynomolgus monkeys, minimal to moderate tubular degeneration were present only at the higher doses of PS-ASO, although renal accumulation of ASO was seen at all doses.[57,58] A more recent study in cynomolgus monkeys and cell culture using drisapersen led to the hypothesis that ASO-induced proteinuria is a transient functional change due to a decrease in receptor-mediated endocytosis in proximal tubules and is not indicative of tubular damage.[59]

Other ASO therapies relying on exon skipping are under development in Duchenne muscular dystrophy:

  1. exon 53 exclusion with golodirsen (also known as 'SRP-4053', phase 3 ESSENCE study—NCT02500381) or viltolarsen (also 'NS-065' or 'NCNP-01', phase 3 RACER53 study—NCT04060199 following positive phase 2 study[60]). Of note, viltolarsen already received its first approval in Japan since March 2020.

  2. exon 45 exclusion with casimersen (also known as 'SRP-4045', phase 3 ESSENCE study—NCT02500381) and DS-5141b (phase 1/2, NCT02667483);

  3. exon 51 exclusion with suvodirsen (NCT03907072, phase 2/3 DYSTANCE 51 study). Results are pending, but the trial is registered as terminated due to lack of efficacy.

  4. in Duchenne muscular dystrophy with exon 2 duplication, the phase 1/2 study NCT04240314 used an AAV9 vector to deliver a small nuclear RNA (snRNA) construct to skip exon 2 during splicing.

Transthyretin-related Hereditary Amyloidosis. Pathophysiology: Transthyretin-related hereditary amyloidosis[61] is a rare lethal disease due to misfolded transthyretin (TTR, or prealbumin) aggregation. The aggregating TTR protein is mainly made of mutant amyloid TTR (TTRv) and at lower degree of wild-type TTR (TTRwt). In transthyretin-related hereditary amyloidosis (ATTRv), the aggregating TTR protein has an abnormal sequence. It is an autosomal dominant genetic disease with over 130 causing mutations identified in the TTR gene,[62] favouring TTR aggregation and partly explaining phenotypic variability,[63,64] mainly neurological (ATTRv-PN) or sometimes cardiological (ATTRv-CM). Common symptoms in ATTRv-PN are length-dependent sensory-motor neuropathy and autonomous neuropathy, non-length-dependent polyneuropathy mimicking chronic inflammatory demyelinating polyneuropathy, and troncular neuropathy—namely bilateral carpal tunnel.[63] Extra-neurological symptoms are mainly cardiomyopathy, nephropathy, ocular and involuntary weight loss.[65] ATTRv-PN is the most severe and disabling hereditary peripheral neuropathy of adult onset, inducing degeneration of sensory, motor and autonomic nerves. TTR is synthetized in the liver and transports thyroxin and retinol in the serum. Hepatic transplantation is an efficient treatment in ATTRv-PN suppressing the main source of mutant TTR, halting progression at early stages and increasing the survival.[66–69] However, neuropathy response to transplantation is inconsistent, with possible deterioration due to accumulation of TTRwt in peripheral nerve and heart in late-onset patients after liver transplantation.[70–72] Less-invasive treatments have been developed, such as diflunisal or tafamidis, both small molecules stabilizing TTR in a non-aggregation prone conformation. The effect on neuropathy is again inconsistent, likely positive in early neuropathy stages but rather uncertain in later stages.[65,72,73] Because of this unmet medical need for more consistent effects in this monogenic autosomal dominant hereditary disorder, and in order to knockdown both mutant and wild-type TTR, ASOs and iRNAs were developed.

Inotersen: Inotersen (previously IONIS-TTRRx or ISIS 420915) is a 2'-MOE and PS ASO with a 5–10–5 gapmer design delivered by weekly subcutaneous injections, triggering RNAse H-mediated mRNA degradation. The phase 3 placebo-controlled trial NEURO-TTR included 173 patients with polyneuropathy and TTR-related hereditary amyloidosis treated for 64 weeks. Randomization was stratified on disease stage, Val30Met mutation status and previous treatment by tafamidis or diflunisal. The primary end point was a difference >2 points in the modified Neuropathy Impairment Score + 7 (mNIS + 7, range 0 to 304, higher score indicating a more severe impairment). Two points difference considered as clinically significant for the original NIS score.[74] Inotersen 300 mg given subcutaneously three times on alternate days in the first week and then once-weekly for 64 weeks performed significantly better than placebo, with a lower increase of 5.8 points (95% CI: 1.6 to 10.0) from baseline in treated patients versus 25.5 points (95% CI: 20.2 to 30.8) under placebo. This clinical improvement mirrored the drop observed in TTR serum levels. Adverse events were glomerulonephritis in three patients (3%) receiving inotersen and progressive thrombopaenia in 60 patients (54%) receiving inotersen with a nadir at 3–6 months. This thrombopenia reached levels below 25 × 109/l in three patients (3%) with normalization in two of them after stopping inotersen treatment and with steroid therapy. An intracerebral haemorrhage leading to death occurred in the third patient. Weekly platelet count was implemented afterwards, leading to dose decrease (150 mg weekly) in two patients without further thrombopaenia below 50 × 109/l.[75] The interim analysis of the open-labelled extension study at 2 years with patients treated by inotersen after initially receiving placebo was in favour of a clinical improvement without reaching the magnitude observed in the initial treatment group, stressing the importance of an early treatment and a putative neuroprotective effect.[76] The exact mechanism of inotersen-induced thrombopaenia is unknown. Antiplatelet immunoglobulin G antibodies were detected at baseline in 5 of 31 inotersen-treated NEURO-TTR study subjects, four of whom eventually developed grade 1 or 2 thrombocytopaenia, and also with anti-GPIIbIIIa in two of them while on the drug, suggesting a predisposing immunological state[77] or possibly due to the PS-ASO chemical modifications modifying platelet activation.[78,79]

Patisiran: Patisiran (ALN-TTR02) is the first commercialized RNAi treatment. It is a 2'-OMe 21-nucleotide siRNA complementary to a sequence in the 3' UTR region of both wild-type and pathological TTR in transthyretin-related hereditary amyloidosis. The siRNA is incorporated in lipid nanoparticles, which protects RNA from degradation in the bloodstream and favours hepatic uptake.[80] The phase 1 trial tested siRNA in two distinct first- and second-generation formulations of lipid nanoparticles. The second-generation ALN-TTR02 suppressed more efficiently the production of both mutant and wild-type forms of transthyretin and established the clinical proof of concept for RNAi therapy targeting mRNA transcribed from a disease-causing gene.[81] These promising results were confirmed in a phase 2 trial for patisiran, which was generally well tolerated and resulted in significant dose-dependent knockdown of transthyretin protein in patients with ATTRv-PN with infusions every 3 weeks.[81,82] A phase 3 placebo-controlled trial included 225 patients, 148 of whom received intravenous patisiran 0.3 mg/kg every 3 weeks for 18 months. There was a significant difference between both groups on the primary end point: at 18 months, mNIS + 7 score had decreased by 6.0 ± 1.7 points in the patisiran group, compared to an increase of 28.0 ± 2.6 points in the control group. Moreover, 56% of patients receiving patisiran improved their mNIS + 7 score compared to 4% under placebo. Patisiran was efficient whatever the stage of the disease and the variant of TTR. Interestingly, clinically pertinent secondary end point on the locomotion (10 Metre Walk Test) and autonomic dysfunction (Composite Autonomic Symptom Score 31) improved compared to baseline, suggesting some reversal of the disease. The median decrease in serum TTR level was 81% at 18 months with patisiran. Adverse events, more common with patisiran than placebo, were peripheral oedema and infusion-related reactions, corresponding to back pain, flush, nauseas, abdominal pain, arthralgias and dyspnoea. Seven patients (5%) had to stop treatment owing to adverse events in the treatment group, and nine (14%) in the placebo group. Contrary to inotersen, no thrombopaenia was reported[24] and there was no drug related mortality. There was no impact on liver function test, renal function or haematology. The 1-year APOLLO open-label extension allowed confirmation of improvement of mNIS + 7 and quality-of-life scores in the patisiran group, and halted progression in placebo group after switching to patisiran.[83]

Amyotrophic Lateral Sclerosis. Pathophysiology: Amyotrophic lateral sclerosis is an upper and lower motor neuron degeneration in adults, with an incidence of 2/100 000/year leading to fast-progressive weakness and death in 2 to 4 years. The most frequently identified causal mutations are located in the SOD1, FUS, TARDBP and C9orf72 genes,[84] the latter carrying a GGGGCC intronic pathological expansion.[85,86] This pathological expansion accounts for <10% of apparently sporadic amyotrophic lateral sclerosis and 40% of familial forms.[87]

ASOs targeting SOD1 and C9orf72 transcripts are being tested to date. SOD1 encodes for the superoxyde dismutase 1 enzyme, a metalloprotease protecting against oxidative stress. Pathogenic variants in SOD1 are found in 10% of dominantly inherited forms and lead to dysregulated enzyme activity.

C9ORF72 seems to be implicated in autophagy and endosomal trafficking. The expansion appears to lead to a toxic gain of function of the protein or RNA itself.[84,88]

Antisense Oligonucleotides: Tofersen (BIIB067) is a 2'-MOE and PS ASO with a 5–11–4 gapmer design developed for SOD1-related amyotrophic lateral sclerosis, triggering both mutated and wild-type SOD1 mRNA RNAse H-mediated degradation, thus decreasing protein production. A phase 1 study on 21 patients found no serious adverse event with tofersen, but neither did it show any difference in CSF SOD1 protein levels. Such a decrease was, however, found in animal models, along with an increased survival.[89] The phase 1/2 blinded and placebo-controlled ascending-dose trial in adults (VALOR study, NCT02623699) assessed the safety of intrathecal injection of tofersen 20, 40, 60 or 100 mg in five injections over 12 weeks. At the highest dose, the change in SOD1 protein concentration in CSF at Day 85 compared to baseline was −33% (95% CI: −47 to −16). The most common adverse events were related to lumbar puncture. Exploratory outcomes point towards a slowing of clinical deterioration in the 100 mg group compared to placebo, but this clinical efficacy is further explored in the ongoing part C (phase 3 study) of the VALOR study.[90]

In a mouse model of SOD1 amyotrophic lateral sclerosis, intracerebroventricular injection of a 25-mer PMO ASO targeting an overexpressed microRNA, miR-129-5p, resulted in a significant increase in survival and improved the neuromuscular phenotype. This could constitute an additional option to target SOD1-related amyotrophic lateral sclerosis.[91]

Regarding C9orf72-related amyotrophic lateral sclerosis, a phase 1 trial (NCT03626012) is ongoing to assess the safety of intrathecal BIIB078 (or IONIS C9Rx), an ASO triggering RNAse H-mediated degradation of transcripts containing the hexanucleotide expansion.

Diseases With Ongoing Phase 1 or 2 Clinical Trials

Huntington's Disease. Pathophysiology: Huntington's disease is an autosomal dominant inherited neurological disorder characterized by behavioural changes, cognitive decline and movement disorders (most often chorea, dystonia and bradykinesia). The disease is caused by a pathological CAG repeat expansion in exon 1 of the huntingtin (HTT) gene exceeding 35 units.[92] The longer the repeat, the earlier the disease onset of symptoms.[93] Thus, modest expansions of 40–42 repeats in HTT are associated with the appearance of psychiatric, cognitive and motor disturbances in late adulthood, whereas large expansions of over 80 repeats cause childhood onset. Other genes have been described to cause Huntington's disease phenocopies, with patients sharing the same or a close phenotype to Huntington's disease but without the HTT pathological expansion.[94]

The HTT gene contains 67 exons and has two mRNA transcripts, with alternative splicing occurring in rare circumstances. Two antisense approaches are currently being developed: global downregulation of HTT and an allele-specific approach that aims to target only mutant HTT.

Antisense Oligonucleotides Targeting Both Normal and Pathological HTT. Tominersen is a 2'-MOE and PS ASO with a 5–10–5 gapmer design, triggering RNAse H-mediated degradation of HTT mRNA. ASO with similar structure showed to be effective in three distinct Huntington's disease rodent models, with up to an 80% drop in HTT mRNA levels and up to a 75% decrease in HTT protein levels.[19] In wild-type rhesus monkeys, intrathecal ASO injection led to a decrease in HTT mRNA production of 47–63% in cortical areas, compared to about 20% in deep brain nuclei.[19]

A phase 1–2a randomized double-blind placebo-controlled trial included 46 patients.[95] Tominersen was administered intrathecally every 4 weeks at doses of 10 to 120 mg depending on the randomization group. There was no increase in adverse events compared to placebo and no severe adverse events in this study with a total of four administrations per patient. Treatment led to a dose-dependent decrease of mHTT level in the CSF compared to baseline, ranging from 20% at 10 mg/injection to 42% at 60 mg/injection, inferring target engagement.[95] Interestingly, a dose-dependent ventricular enlargement in the treated group and neurofilament light chain transitory increase has been noted.[95,96]

A phase 3 trial including >800 patients with Huntington's disease was recently halted for dosing based on independent data monitoring committee recommendations (GENERATION HD1 study, NCT03761849). The preliminary analysis showed an unfavourable efficacy profile, as those who received an intrathecal injection of tominersen every 8 weeks clinically worsened compared to those who received placebo injections.

Antisense Oligonucleotides Targeting SNPs Associated With mHTT: To only decrease the mHTT allele expression, current clinical trials apply an indirect approach targeting single nucleotide polymorphisms (SNPs) associated with the mHTT gene (Figure 4).

Figure 4.

Potential use of antisense therapies for allele specific and non-allele specific targeting of mRNA or pre-mRNA, taking the example of a CAG expansion disease like Huntington's disease. (A) Non-specific allele targeting: targeting a conserved sequence present in both mutant and wild-type mRNA or pre-mRNA leads to a similar decrease in both RNAs. For example, tominersen in Huntington's disease. (B) CAG repeat targeting: targeting the CAG repeats leads to preferential mHTT decrease, as more binding sites are available in the mRNA or pre-mRNA harbouring the pathological expansion. Examples are at preclinical stage only. (C) Allele-specific targeting: targeting a SNP associated with the pathological CAG expansion, leading to a specific mHTT decrease. Example: WVE-120101 and WVE-120102 in Huntington's disease.

Phase 1b/2a studies were ongoing for ASO targeting selectively mHTT production by targeting the rs362307 SNP for ASO WVE-120101 (PRECISION-HD1 study, NCT 03225833) or the rs362331 SNP for ASO WVE-120102 (PRECISION-HD2 study, NCT02225846). There was no significant decrease of the mHTT, which motivated discontinuation, with pending results. A third SNP will be targeted in a new study that has just opened to recruitment, planning to include 36 patients (NCT05032196). Perhaps the strategy aiming at maximizing the decrease of mHTT should be considered. Indeed, we could speculate that the dose of tominersen was too high and reached the cortex too intensely, causing a drop of Huntingtin protein of more than 80%.

Antisense Oligonucleotides Targeting CAG Expansion in mHTT. ASO targeting the CAG repeat have been developed. For instance, locked nucleic acid oligomers targeting CAG repeats were indeed able to preferentially inhibit mutant HTT synthesis in vitro (Figure 4) without affecting other tested genes that contain CAG repeats.[97,98] A 21-mer PS 2'-OMe ASO with a seven CUG triplet sequence, delivered during six weekly intracerebroventricular injections, showed an in vivo efficacy in two mouse models of Huntington's disease (R6/2 and Q175) with a decrease in mutant HTT and a phenotypic improvement. There was no significant decrease in the levels of proteins coded by genes with shorter non-expanded CAG repeats and mouse huntingtin.[99,100]

Interestingly, many other autosomal dominant neurological disorders are caused by the pathological expansion of unstable CAG repeats over a threshold, such as spinocerebellar ataxias types 1 (ATXN1), 2 (ATXN2), 3 (ATXN3), 6 (CACNA1A), 7 (ATXN7), 17 (TBP) and dentatorubral-pallidoluysian atrophy (ATN1). ASO targeting CAG repeats could be useful in other neurological diseases with unstable CAG repeats than Huntington's disease such as these seven forms of dominant ataxias.[101] This common feature could indeed lead to a common treatment, targeting the pathologically expanded but not the normal CAG expansion, polymorphic in size in the general population (for example varying between nine and 34 for HTT).[102] However, it should be kept in mind that the threshold for normal CAG repeats in one gene is pathological in another gene, i.e. CAG repeats around 21 in CACNA1A are pathological for spinocerebellar ataxia type 6, but in ATXN3 the number exceeds 53 to become pathological for spinocerebellar ataxia type 3. In addition, the human genome encodes about 200 mRNAs harbouring CAG repeats longer than six repeated units. This raises concerns about potential 'off-target' effects of ASO or iRNA targeting the CAG repeats as a general non-specific target.[103]

Interfering RNA Treatments. Treatments based on RNAi in Huntington's disease are currently at the preclinical development stage. Selective mHTT silencing has been confirmed in vitro with miRNA-like siRNA,[104–107] and this approach appears also effective in vitro in other CAG repeat diseases.[103,108–110]

Other studies relied on non-allele-specific targeting of HTT by RNAi. In wild-type rhesus monkeys, miRNA or shRNA in an AAV2 viral vector injected directly in the putamen reduced HTT production by 45% at 6 months.[111,112] Similar results were obtained in mice, rats, minipigs and sheep using AAV viral vectors.[113–120] Similarly, striatal injection of cholesterol-conjugated siRNA or siRNA loaded in exosomes appeared effective in different mouse models.[121–123] More recently, divalent siRNA, composed of two fully chemically modified PS siRNAs connected by a linker, targeting HTT and delivered via intracerebroventricular injection, led to a robust decrease of HTT expression in both the BAC-HD-ΔN17 mouse model of Huntington's disease and normal cynomolgus macaque. The HTT decrease was still present throughout the brain at 1 month. There was no reported adverse event.[124] Non-invasive routes of administration are also being developed, for instance using the nose-to-brain route with siRNA encapsulated in nanoparticles containing chitosan in the YAC128 mouse model of Huntington's disease.[125]

In addition to this classical downregulation of the HTT gene approaches, other metabolic pathways can be targeted in the hope of increasing the metabolic rate in the brain upregulating CYP46A1 expression.[126]

Synucleinopathies. Synucleinopathies encompass neurodegenerative diseases that are characterized neuropathologically by the presence of proteinaceous inclusions consisting, among other proteins, of α-synuclein aggregates. Among these synucleinopathies, Parkinson's disease is the most common. Other less-common synucleinopathies include dementia with Lewy bodies and multiple system atrophy.[127–129]

Both data on cell-to-cell transmission of misfolded α-synuclein in a prion-like manner and identification of mutations or duplications/triplications of the alpha-synuclein gene (SNCA) suggest a toxic effect of the protein as causative for Parkinson's disease in dominant familial forms and led to the development of treatments targeting α-synuclein.[130–133]

An ASO targeting the SNCA pre-mRNA is being evaluated in a phase 1 trial (HORIZON study, NCT04165486) with intrathecal injections, but its recruitment is currently on hold. The same route of administration is being used in a phase 1 single and multiple ascending dose study (REASON study, NCT03976349) evaluating an ASO targeting LRRK2 in Parkinson's disease patients, which alleviated motor symptoms in a mouse model.[134]

Tauopathies. The Tau protein (or microtubule-associated protein tau, MAPT) has a role in cytoskeleton dynamics. Aggregation of misfolded tau protein is the neuropathological hallmark of a group of neurodegenerative disorders regrouped under the term 'tauopathies'.[135]

ASO targeting the tau mRNA efficiently decreased tau protein levels in the brain of cynomolgus monkeys after intrathecal injection.[136,137] Clinical studies are ongoing with this approach in Alzheimer's disease (phase 1–2, NCT03186989, active but not recruiting) and progressive supranuclear palsy (phase 1 multiple ascending doses, NCT04539041, actively recruiting).

Other Neurological Diseases With Ongoing Preclinical or Clinical Development. Several neurological diseases are at the preclinical and clinical stage, reflecting the dynamic sector that is ASO development. While a complete review of preclinical data is out of scope, we present some preclinical data in Supplementary Table 2, either relative to ongoing phase 1 trials (Table 2 and Supplementary Table 1) or illustrating the diversity of neurological diseases and of therapeutic approaches. The potential clinical applications of antisense therapies in neurological diseases are therefore plentiful and could even help patient-tailored approaches.[138]

Such a patient-tailored treatment has been successfully implemented in a short time. Milasen is a 22-mer PS and 2'-MOE ASO that was specifically designed and delivered in a 6-year-old female presenting with neuronal ceroid lipofuscinosis 7. This child was compound heterozygous in the MFSD8 (or CLN7) gene, with a known c.1102G->C pathologic mutation and a SINE-VNTR-Alus (SVA) retrotransposon insertion. Within a year, Kim and colleagues were able to demonstrate that this insertion leads to a mis-splicing of exon 6, designed candidate splicing-modifier ASOs, tested them in the child's fibroblast and assessed the most efficient ASO—milasen—safety in rats before proceeding with the intrathecal treatment of the patient. Milasen was well tolerated and decreased seizure duration and frequency, but did not prevent continued brain volume loss.[138]