An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Noebels JL, Avoli M, Rogawski MA, et al., editors. Jasper's Basic Mechanisms of the Epilepsies. 5th edition. New York: Oxford University Press; 2024. doi: 10.1093/med/9780197549469.003.0075
Most development and epileptic encephalopathies (DEEs) are genetic in origin. Drug-resistant epilepsy is often the most notable feature of the DEEs, although numerous other symptoms are present that have significant impact on patients’ quality of life. Despite novel, third-generation antiseizure drug treatment options becoming available over the last several years, seizure freedom is often not attained and nonseizure symptoms remain. Therapeutic approaches such as antisense oligonucleotides and adeno-associated virus delivered gene modulation offers realistic hope for seizure freedom in some DEEs with an underlying genetic etiology, with several approaches demonstrating preclinical success and now transitioning to clinical trials. Several of these therapeutic strategies risk the exacerbation of gain-of-function variants and may not be reversible, thus emphasizing the need for functional testing of variants and more reliable prediction tools for pathogenicity. Dravet syndrome, Rett syndrome, and Angelman syndrome are presented here as examples of how these techniques are being applied and the nuances and challenges of each unique genetic disorder. Additionally, with so many gene regulatory therapeutic options on the horizon, there will be a need to understand how to select appropriate patients for each treatment, whether treatments are complementary or adverse to each other, and long-term risks of the treatment. Nevertheless, precision therapeutics hold tremendous potential to provide long-lasting seizure freedom and even complete cures for this devastating disease.
Dravet syndrome (DS), a severe developmental and epileptic encephalopathy (DEE), is estimated to occur in 1/15,700 live births (Wu et al., 2015). Clinically, DS is a drug-resistant epilepsy presenting in the first year of life with prolonged seizures in the setting of fever or temperature changes, often hemi-clonic in nature, followed by unprovoked seizures of varying etiologies. Genetic testing has revealed pathogenic variants in the gene, SCN1A, in 70%–80% of children suspected to have DS (Wu et al., 2015; Depienne et al., 2009; Zuberi et al., 2011; also see Chapter 44, this volume). While DS is largely a monogenic disease, variants in other genes have been reported. Over the past two decades, DS has had the benefit of intense research, resulting in novel therapeutic approaches that target seizure control through genetic modulation, which we will discuss here. In this chapter we highlight an antisense oligonucleotide (ASO) gene modulatory approach to DS as an example of how advances in technology are being applied to the genetic epilepsies.
SCN1A, encoding the voltage-gated sodium channel Nav1.1 α subunit, was first reported to be associated with DS in 2001 (Claes et al., 2001), but it is important to note that not all children with SCN1A variants develop DS and not all children with a clinical diagnosis consistent with DS have an identified variant in SCN1A. Additionally, correlation of pathogenicity with variant type is not as reliably predictive of disease as would be preferred.
Over 3,000 monoallelic pathogenic DS variants in SCN1A have been reported, including missense and truncation mutations, microdeletions, and gene duplications(2021a). Most pathogenic variants occur de novo, with less than 10% inherited (Scheffer and Nabbout, 2019). In most cases, haploinsufficiency of Nav1.1 appears to be sufficient to produce clinical disease. Nav1.1 is expressed most prominently in inhibitory neurons in the brain; therefore, channel dysfunction is postulated to lead to disinhibition, a subsequent increase in network hyperexcitability, and epileptogenesis (Catterall, 2018). Most pathogenic variants in SCN1A result in loss-of-function (LOF) and nonsense-mediated decay (NMD) of the mutant allele (Catterall, 2018), although some pathogenic variants have been reported to cause gain of function (GOF) and may be associated with a more severe and earlier onset phenotype (Berecki et al., 2019).
As gene modulatory therapies are developed, the need for clarification of variants with functional testing will increase to determine GOF versus LOF as well as to clarify which variants are pathogenic. Indeed, to reduce the level of intellectual impairment associated with DS, early treatment will be required to lead to improved outcome. Thus, early-life genetic screening followed by functional testing of individual patient variants to predict pathogenicity will be essential to initiating effective treatment.
DS demonstrates clinical evolution over the life span, starting with prolonged seizures in the first year of life that are often hemi-clonic and associated with temperature changes such as fever or ambient temperature (e.g., temperature changes associated with bathing). Multiple types of spontaneous seizures follow shortly after and can include myoclonic seizures, though myoclonic seizures are not present in all children despite the early naming of this syndrome as severe myoclonic epilepsy of infancy. In the toddler years, seizures are often prolonged episodes of status epilepticus leading to high use of rescue medications and frequent admissions to the intensive care unit. Seizures then begin to evolve into clusters, often during sleep or on the borders of sleep, as children enter the school-age years. There has been a recent increase in the knowledge about adults living with DS, with seizures described as predominantly generalized tonic-clonic, occurring in clusters but with reduced frequency, and individual seizures that are short in nature (Akiyama et al., 2010; Genton et al., 2011; Takayama et al., 2014). Periods of nonconvulsive status are often reported in adolescents and young adults. Treatment goals are generally focused on seizure reduction with balanced side effects, as seizure freedom with antiseizure drugs (ASDs) is often futile. Sodium channel blockers such as lamotrigine and carbamazepine are reported to exacerbate seizures and are thus contraindicated, further limiting ASD choices in this patient population (Wirrell et al., 2017; Wirrell and Nabbout, 2019; Cross et al., 2019).
Several associated DS symptoms can impact quality of life as much as, if not more than, seizures (Fig. 75–1). These symptoms affect several organ systems and include constipation, feeding issues, sleep disruption, lack of perception of pain, behavioral issues, intellectual impairment, short stature, progressive gait abnormalities, balance issues, parkinsonism, and peripheral neuropathy (de Lange et al., 2019; Licheni et al., 2018; Lagae et al., 2018; Villas et al., 2017; Olivieri et al., 2016; Connolly, 2016; Knupp et al., 2017; Eschbach et al., 2017; Berg et al., 2020; Nabbout et al., 2019) and have been reported by parents and caregivers to be as concerning to families as seizures (Villas et al., 2017; Knupp et al., 2017). The severity of many of these symptoms appears to be independent of seizure burden and likely due to the underlying pathophysiology created by the disruption of SCN1A expression. Cognition continues to plateau regardless of the level of seizure control (Jansson et al., 2020; Brown et al., 2020; Verheyen et al., 2020; Ceulemans, 2011) and may be predictable by variant type (Ishii et al., 2017; Riva et al., 2009). Behavioral issues are major concerns for families and appear to be independent of seizure control (Berg et al., 2020, Villas et al., 2017). Gait abnormalities emerge during early adolescence at a time period when seizures are slowing (Rilstone et al., 2012; Rodda et al., 2012; Black and Gaebler-Spira, 2016). Additionally, DS patients have a significant risk for sudden unexpected death in epilepsy (SUDEP), up to 20%, which is higher than reported in the general epilepsy population (Cooper et al., 2016; Shmuely et al., 2016). All of these associated symptoms of DS impact quality of life but are neither well studied nor used as outcome measures to assess treatment response. There are new small-molecule ASDs that are leading to improved seizure control in this population, but until people living with DS are seizure free and have resolution of associated symptoms, the disease is not “cured.” Resolution of seizures, while important, should not be the only measure of “intractable Dravet syndrome,” this definition will need to be expanded to consider all associated symptoms of DS that impact quality of life.
DS presents with a complex array of symptoms in addition to seizures. Current medications target seizures, but not other presentations.
Despite the development of novel ASDs (Loscher et al., 2020), most DS patients have drug-resistant epilepsy and associated symptoms that impact quality of life, underscoring the unmet need for precision approaches that directly target the underlying genetic cause, SCN1A haploinsufficiency in the central nervous system (CNS). SCN1A is one of a group of genes expressed in the brain that have naturally occurring, alternative exons containing a premature termination codon (PTC) (Carvill and Mefford, 2020; Lim et al., 2020; Fig. 75–2).
Identification of protein-coding genes with NMD-inducing events: 7,757 unique genes containing at least one nonproductive alternative splicing events, of which 1,246 unique genes are disease-associated. The left panel depicts the types of alternative (more...)
These exons, called NMD, or poison, exons are differentially expressed during brain development in a cell-type-specific fashion to control the levels of critical proteins involved in neuronal excitability, like ion channels. For example, early studies in 1997 by the Meisler group showed that the sodium channel gene, SCN8A, contains developmental and cell-type-regulated alternative exons (Plummer et al., 1997). Fetal brain and adult non-neuronal brain tissues, including glia, express the SCN8A exon 18N, containing a PTC. As the brain develops, neurons express the alternative exon, 18A, resulting in full-length SCN8A transcript. In 1998, Oh and Waxman identified an alternative, PTC-containing Scn1a NMD exon that is expressed in rat spinal cord astrocytes (Oh and Waxman, 1998). They predicted that astrocytes may use this alternative splicing strategy to control sodium channel gene expression. Five pathogenic DS patient genetic variants were later identified in the corresponding human SCN1A NMD exon, which was designated 20N (Carvill et al., 2018). The authors of this study proposed that disease variants within 20N may lead to its increased inclusion, resulting in increased levels of NMD, reduced levels of full-length SCN1A transcript, and Nav1.1 haploinsufficiency. In 2019, Steward et al. used long-read human brain-derived RNA-sequencing techniques to identify novel exons in human disease-associated genes and to reclassify these exons as coding or putative NMD. Their work resulted in the identification of two additional NMD exons in SCN1A (Steward et al., 2019).
Antisense oligonucleotides (ASOs) were introduced by Dr. Adrian Kraner and colleagues as an innovative strategy to inhibit the expression of NMD exons (reviewed in Nomakuchi et al., 2016). Recent innovations in ASO chemistry have introduced modifications that increase nuclease resistance, reduce immune response, and increase binding affinity of ASOs, including a uniform phosphorothioate backbone and methoxyethyl at the 2′ ribose position (2′MOE-PS), effectively paving the way for their use in clinical practice (Scharner and Aznarez, 2021; Lim et al., 2020). The first precision therapy for SCN1A-linked DS to reach FDA-approved human clinical trials was STK-001, an ASO introduced by Stoke Therapeutics (Han et al., 2020). STK-001 was developed using targeted augmentation of nuclear gene output (TANGO) technology (Lim et al., 2020; Scharner and Aznarez, 2021), targeting the SCN1A NMD exon 20N (Carvill et al., 2018) to specifically reduce levels of nonproductive mRNA and increase levels of productive mRNA and Nav1.1 sodium channel protein (Fig. 75–3). This strategy, which prevents exon 20N expression, ultimately upregulates expression of the wild-type allele, but not the NMD allele to compensate autosomal dominant SCN1A haploinsufficiency.
TANGO ASOs prevent NMD to increase productive mRNA and protein. A. Example of a cell expressing a pre-mRNA that is alternatively spliced to generate a productive mRNA and a nonproductive mRNA. While the nonproductive mRNA is degraded by NMD, the productive (more...)
Preclinical work showed that intracerebroventricular (ICV) administration of STK-001 to wild-type C57BL/6J mouse brain in vivo increased the expression of productive, full-length Scn1a mRNA and Nav1.1 protein (Han et al 2020). A single ICV dose of STK-001 at postnatal day 2 (P2) in the Scn1aTmKea (F1:129S-Scn1a+/– x C57BL/6J) heterozygous mouse model of DS (Mistry et al., 2014), in which exon 1 of Scn1a is deleted, increased productive Scn1a mRNA and Nav1.1 protein expression. Importantly, this single-dose treatment also prevented SUDEP in 97% of DS mice monitored to 90 days following the single injected dose (Fig. 75–4, I). Single-dose ASO treatment of DS mice at P14, closer to the time of seizure onset in this model, resulted in a less robust, but significant, effect on mouse survival. Infrared-video monitoring of 19 DS mice injected with STK-001 at P2 showed a single tonic-clonic seizure followed by SUDEP in only 1 animal, with no behavioral seizures detected in the other 18. Electroencephalogram (EEG) recording of DS mice injected with STK-001 at P2 showed a reduction in seizure frequency with a prolonged latency to first seizure (Fig. 75–4, II). Importantly, while STK-001 administration significantly reduced seizure frequency and latency in mice, the infrequent seizures that did occur were as electrographically and behaviorally severe as those observed in untreated animals.
I. A single ICV injection of 20 µg ASO-22 at P2 results in reduced SUDEP incidence and increased NaV1.1 protein expression in DS mice. A. Experimental design for target engagement, pharmacology, and efficacy study in DS and WT mice. (more...)
Intrathecal lumbar bolus administration of STK-001 was subsequently evaluated at two different dosages in good laboratory practice (GLP) studies of nonhuman primates (NHPs) for safety, brain biodistribution, target engagement, and pharmacodynamics (Liau et al., 2019). The drug was reported to be well tolerated at both dosages with no changes noted on physical and neurological examination, no changes in food intake, body weight, hepatic function, or platelet counts, and no abnormal histopathology in brain, spinal cord, liver, or kidney. In non-GLP studies, at doses well above the no observed adverse effect level (NOAEL) dose, acute, transient hind-limb paresis and acute convulsions were observed. Biodistribution of STK-001 was observed in all brain regions except the pons and thalamus, with highest levels found in cortex. Levels of brain Nav1.1 protein in treated wild-type animals were increased up to three-fold over controls, with highest levels observed in the motor cortex, occipital cortex, parietal cortex, and prefrontal cortex. Thus, this study showed that a single, intrathecal lumbar bolus injection of STK-001 in NHPs was safe and pharmacologically active at dosages below the NOAEL.
Preclinical work in mice and NHPs led to initiation of the Butterfly multicenter, longitudinal, prospective study of 2- to 18-year-old DS patients, which showed, in addition to seizures, substantially decreased neurocognitive abilities despite the use of multiple ASDs, widening from normal levels in overall intellectual development that increased with age, severe developmental delays, and gaps in adaptive functioning from age-equivalent peer groups (Sullivan, 2020). Together, this body of work led to FDA approval of two Phase 1/2a clinical trials for STK-001 in the United States as well as authorization by the UK Medicines and Healthcare products Regulatory Agency (MHRA) of a similar trial in the United Kingdom. All three trials were designed as open-label studies of children and adolescents ages 2 to 18 with established diagnosis of DS linked to a confirmed pathogenic variant in SCN1A (Laux et al., 2020) to assess the safety and tolerability of STK-001, as well as to characterize human pharmacokinetics. Secondary endpoints included changes in seizure frequency and quality of life measures. In the United States, both single and multiple ascending dose protocols (SAD and MAD, respectively) up to 30 mg per dose were approved. The MONARCH trial, testing both SAD and MAD protocols, launched in August 2020 testing SAD, and the first MAD patient was dosed in January 2021. The SWALLOWTAIL study was subsequently approved as an open-label extension of MONARCH, in which DS patients received multiple, but not ascending, doses of the last dose they received in MONARCH. The ADMIRAL Phase 1/2a open-label study of MAD up to 70 mg in DS patients from 2 to 18 years of age was authorized by the MHRA in the United Kingdom in 2021. Importantly, TANGO therapy is contraindicated for DS patients with missense SCN1A variants that result in the generation of Nav1.1 polypeptides, which may have maladaptive GOF or dominant-negative effects (Sadleir et al., 2017; Berecki et al., 2019), as TANGO-mediated increases in mRNA and protein expression would likely increase disease severity. Nevertheless, introduction of this ASO drug was a major advance in precision therapeutics for DS patients.
In addition to NMD exons, targeting SCN1A genomic regulatory elements may hold therapeutic potential for DS patients. SCN1A, like a large proportion of mammalian genes, has multiple, coactive promoters. In SCN1A, this structure results in a variable 5′ untranslated region containing one of two transcription start sites (TSSs), designated 1a and 1b (Haigh et al., 2021). Haigh and colleagues used a pool of single-guide RNAs (sgRNAs) targeted to the human 1b sequence, delivered with the inactivated Cas9-histone acetyltransferase fusion protein, dCas9-p300, to induce SCN1A expression 2.5-fold in HEK293 cells compared to nontransfected controls. This strategy highlights the importance of including noncoding sequences in genetic screening of DS patients and suggests that synthetic transcriptional activation may be a potential therapeutic strategy in DS.
In addition to STK-001, other precision therapeutic strategies for DS are being developed, including other ASOs, viral gene delivery, and virally delivered CRISPR-Cas9-based strategies. This body of work suggests that SCN1A regulatory sequences, as well as other genes, including those encoding other sodium channel α and β subunits, may be effectively targeted to provide benefit to DS patients.
Sodium channel Nav1.6, encoded by SCN8A, is critical for the regulation of neuronal excitability in the CNS (also see Chapter 44, this volume). GOF variants in SCN8A are linked to DEE13, which has a similar disease presentation as DS (Meisler, 2019). In fact, work in mice showed that Scn8a is a genetic modifier of Scn1a-linked DS. Reduction of Scn8a expression by intercrossing Scn8a+/med-jo mice with Scn1a+/– DS mice to produce double heterozygous animals rescued premature lethality and extended life span compared to Scn1a+/– animals (Martin et al., 2007). Meisler and colleagues expanded on this work using ASO technology (Lenk et al., 2020). They developed a Scn8a-specific gapmer ASO (Scharner and Aznarez, 2021) that reduced Scn8a expression, delayed seizure onset, and increased survival in the Scn8aR1872W/+ GOF mouse model of DEE (Fig. 75–2). They then tested this Scn8a ASO in the Scn1aTmKea (F1:129S-Scn1a+/– x C57BL/6J) mouse model of DS (Mistry et al., 2014). DS mice treated with a single ICV dose of Scn8a ASO at P2 survived beyond 5 months of age without behavioral seizures or SUDEP. While, as expected, Scn8a gapmer ASO administration resulted in 50% reduction of Scn8a transcript in brain and spinal cord of DS mice, it had no effect on the level of Scn1a transcript despite its effects on seizures and survival in the Scn1a DS mouse model. Furthermore, while ASO treatment of Scn8aR1872W/+ mice required repeated Scn8a ASO administration to be effective long term, 100% of DS mice survived for 5 months following a single Scn8a ASO dose at P2. This result, combined with the observation that a single dose of STK-001 at P2 prevented SUDEP in 97% of DS mice during a 90-day observation period (Han et al., 2020), raised the possibility that single-dose ASO administration during the early critical period of postnatal brain development may provide long-term seizure control in DS patients. A critical future experiment will be to determine whether reduction of Scn8a expression via Scn8a gapmer ASO administration is effective in DS mouse models expressing missense Scn1a variants that result in the generation of Nav1.1 polypeptides. If so, then SCN8A ASO therapy may provide seizure relief for a wider range of DS patient variants than ASOs like STK-001, which are targeted to NMD exons in SCN1A.
Adeno-associated virus (AAV) vectors currently lead the therapeutic field for gene delivery to the brain (Zhu et al., 2021). However, because standard AAV vectors used in gene therapy cannot accommodate large payloads like sodium channel α subunit encoding genes (Wu et al., 2010), investigators in the DS field have developed novel strategies to modulate SCN1A gene expression through the modulation of SCN1A gene regulatory regions or activities of smaller gene products like transcription factors and sodium channel accessory subunits, as discussed below. While high-capacity adenoviral vectors, with expanded cloning capacities up to 37 kb, are in development (Ricobaraza et al., 2020), their utility for SCN1A expression remains under investigation.
A number of neuronal cell types are affected in DS (Yu et al., 2006; Mistry et al., 2014; Liu et al., 2013; Goff and Goldberg, 2019; Tran et al., 2020; Ogiwara et al., 2013); however, GABAergic neurons, especially parvalbumin (PV)-positive, fast-spiking interneurons, are particularly vulnerable, and thus disinhibition of PV neurons has been proposed to be a primary contributor to DS mechanisms (Catterall, 2018; also see Chapter 44, this volume). Encoded Therapeutics has taken a gene therapy approach in the development of ETX101, an adeno-associated virus serotype 9 (AAV9) vector-based, GABAergic neuron-selective, therapeutic agent expressing an engineered transcription factor that upregulates endogenous SCN1A gene expression. Unlike ASO therapies, which must be readministered to patients throughout their lifetime, AAV9-based approaches are designed for one-time administration to permanently alter gene expression in human tissues. Preclinical work published in abstract form showed that ETX101 administration to the brain, similar to STK-001, increased Nav1.1 protein expression, prolonged survival, and decreased the occurrence of spontaneous and hyperthermia-induced seizures in a DS mouse model (Belle et al., 2020). A single, unilateral ICV injection of ETX101 in NHPs resulted in transgene expression that was limited to the CNS and included the cerebral cortex and hippocampus. ETX101 was reported to be well tolerated in NHPs with no detectable changes in clinical data, including body weight and body temperature. No macroscopic or microscopic abnormalities were reported in NHP brain, liver, dorsal root ganglion, or spinal cord, and all animals survived until necropsy. Serum titers of AAV9 neutralizing antibodies, which are significant obstacles to AAV-based gene therapy, increased 28 days post ETX101 administration in NHPs, while cerebrospinal fluid neutralizing antibody titers were not different from pretreatment levels. Importantly, and similar to STK-001, the clinical utility of ETX101 is likely limited to DS patients with SCN1A variants that result in Nav1.1 haploinsufficiency through NMD, and contraindicated for DS patients with missense SCN1A variants that result in the generation of Nav1.1 polypeptides, which may have maladaptive GOF or dominant-negative effects (Sadleir et al., 2017; Berecki et al., 2019), as increases in mutant channel expression may increase disease severity. Encoded Therapeutics launched the ENVISION prospective natural history study, without gene therapy, designed to understand the seizure, neurodevelopmental, motor, and behavioral characteristics of SCN1A-linked DS in children aged 6 to 60 months. The first clinical trial of ETX101 gene regulation therapy is planned to initiate in 2022.
Sodium channel non-pore-forming β1 subunits (Navβ1), encoded by SCN1B, are multifunctional (O’Malley and Isom, 2015; Bouza et al., 2021). Hampson and colleagues hypothesized that viral-driven overexpression of Navβ1 subunits in CNS GABAergic neurons might enhance Nav1.1 plasma membrane expression and thus excitability, countering the effects of SCN1A haploinsufficiency (Niibori et al., 2020). They developed an AAV9-based vector driving Navβ1 cDNA expression via the Gad1 promoter (AAV-Navβ1) and injected this agent into the cerebral spinal fluid of Scn1aTmKea (F1:129S-Scn1a+/– x C57BL/6J) wild-type and DS mice (Mistry et al., 2014) at P2. Interestingly, in their study, they observed that untreated female DS mice showed a higher degree of SUDEP than males, raising a previously unknown sexual dimorphism in DS mice. AAV-Navβ1-treated DS mice displayed modest, but significant, increases in survival compared to untreated mice, and this effect was more pronounced in females over males. Male, but not female, DS mice treated with AAV-Navβ1 showed significantly reduced spontaneous seizures. However, AAV-Navβ1 treatment had no effect on febrile seizure susceptibility in either sex. Behavioral analyses showed that male, but not female, DS mice displayed motor hyperactivity and performed abnormally in tests of fear, anxiety, and learning and memory. AAV-Navβ1 treatment of male, but not female, DS mice normalized their motor activity and performance in a test for anxiety. While this work may suggest a new therapeutic strategy, the broad multifunctionality of Navβ1 and modest effects of AAV-Navβ1 in DS mice must be more carefully considered. In addition, their work uncovered previously unappreciated sexual dimorphisms in DS, at least in mice.
CRISPR-Cas9 (clustered, regularly interspaced, short palindromic repeats and CRISPR-associated protein) genome editing technology, originally discovered in prokaryotes, has transformed all of biology. Dr. Emmanuelle Charpentier and Dr. Jennifer Doudna were awarded the 2020 Nobel Prize in Chemistry for this remarkable work. CRISPR-Cas9 offers powerful new precision gene therapy approaches for the treatment of genetic diseases in humans, including neurodevelopmental disorders (Doudna and Charpentier, 2014; Jinek et al., 2012; Gasiunas et al., 2012). Here, we discuss reports from two groups, each using a modified version of the CRISPR-Cas9 system, CRISPRa, which features nuclease-deficient Cas9 (dCas9) and guide RNAs (gRNAs) complementary to promoter regions of a target gene to activate its transcription, have suggested that CRISPR-Cas9 technology may be useful in developing gene therapies to treat DS (Yamagata et al., 2020; Colasante et al., 2020).
In the first example, Broccoli and coworkers screened sgRNAs for their ability to stimulate Scn1a transcription in association with the dCas9 activation system, a CRISPRa strategy. Similar to other work in HEK cells (Haigh et al., 2021), this group identified a specific sgRNA, sg1P, targeting a sequence near the Scn1a proximal promoter, that selectively increased Scn1a mRNA and Nav1.1 protein expression in cell lines and primary mouse cortical interneurons, including neurons isolated from Scn1a+/R1407X DS pups that are haploinsufficient for Nav1.1 (Colasante et al., 2020; Ogiwara et al., 2007). Importantly, no changes were observed in the expression levels of other sodium channel gene transcripts in response to this treatment. In contrast to the STK-001 ASO work, in which increased Nav1.1 protein levels were observed in treated wild-type mouse brain (Han et al., 2020), no changes were observed in the levels of Nav1.1 protein in treated Scn1a+/+ neurons using this strategy. The Scn1a-dCas9 activation system was delivered ICV to P0 Scn1a+/R1407X DS and Scn1a+/+ pups using a dual AAV-9-based system that included the mDlx5/6 promoter to drive expression in forebrain GABAergic neurons. In response to treatment, DS mouse interneuron firing rates, measured electrophysiologically in brain slices, were increased. In addition, hyperthermia-induced seizures, recorded from 1-month-old pups using EEG, were attenuated, with increased threshold for seizure induction. Spontaneous seizures and SUDEP rates were not reported in the study. Limitations to the utility of this strategy included the relatively low coinfection efficiency (~20%) of the two separate AAVs in the interneuron population along with the size of the SpCas9, which required the use of two independent AAVs. In the future, the use of smaller Cas9 orthologs may improve delivery and efficacy. Notably, this transcriptional activation strategy does not distinguish between wild-type and mutant Scn1a alleles. Despite the PTC variant, transcript expression of the R1407X truncated allele was stimulated along with the wild-type allele, although truncated Nav1.1 protein was not detected. Thus, like STK-001, this approach may be contraindicated for DS patients with GOF or dominant-negative SCN1A variants.
In a second example, Yamagata and colleagues used a CRISPRa strategy to achieve inhibitory, Vgat-positive, neuron-selective Scn1a gene activation (Yamagata et al., 2020). A complex experimental system was developed using triple mutant mice (floxed-dCas9-VPRVPR/+/Vgat-CreCre/+/Scn1aR1407X/+) injected with AAV harboring four, synergistically effective mouse gRNAs. A high level of toxicity leading to lethality in the triple mutant mouse model confounded the experiments. Nevertheless, IV delivery of AAV particles containing the gRNAs into surviving pups at 4 weeks of age, following seizure onset at ~P18, resulted in an increased temperature threshold for hyperthermia-induced seizures measured at 12 weeks of age. In addition, latency to clonic seizures, wild-jumps, and generalized tonic-clonic seizures was prolonged when the body temperature of the animals was held at 43˚C. Electrocorticography recorded in freely moving animals showed a decrease in the frequency of abnormal spike discharges in CRISPRa-treated mice. Behavioral analyses, including hyperactivity, thigmotaxis, and anxiety, showed partial phenotypic rescue. Interestingly, while Scn1a mRNA and Nav1.1 protein levels were increased in total brain lysates prepared from CRISPRa-treated mice, the amount of plasma membrane-associated Nav1.1 in CRISPRa-treated Scn1aR1407X/+ brains was not different from that measured in untreated brains, suggesting that, even though overall levels of Nav1.1 protein were increased, the excess channels may not have been capable of functioning in ion conduction. Consistent with this result, immunofluorescence imaging of brain slices with anti-Nav1.1 antibodies revealed excess cytoplasmic staining in the CRISPRa brains rather than in the axon initial segment, as normally measured in wild-type mice. Taken together, while the significance of the CRISPRa work for Scn1a was limited, and more efficient delivery systems must be developed, the results suggest that gene therapy treatment to increase SCN1A expression in DS patients after seizure onset may effectively treat some aspects of their disease.
We have presented DS as an example of a genetic DEE that may benefit from TANGO treatments. However, it is important to note that not all DEEs will benefit from this strategy. Because the TANGO approach relies on upregulation of a functional gene product, the disease must present with haploinsufficiency to ensure the presence of a functional wild-type allele to upregulate. Therefore, autosomal recessive disorders are not candidates to be treated with this process. In addition, because the level of upregulation of gene expression in human brain in response to TANGO is currently unpredictable, it must be known whether the disease can be worsened or exacerbated by overexpression. Disorders that have clinical presentation with too little or too much gene product, the so-called Goldilocks phenomena, such as MECP2 and UBE3A, have unique challenges that must be considered. MECP2 disorders can present as Rett syndrome (too little gene product) or MECP2 duplication syndrome (too much gene product). Yet there is evidence that use of gene modulatory approaches may be successful in these DEEs.
Rett syndrome (RS) is an X-linked, progressive DEE present largely in females, that is characterized by loss of purposeful hand function, loss of expressive language, acquired microcephaly, and loss of ambulation in addition to intellectual impairment, seizures, and autism. Pathogenic variants in the X-linked gene MECP2 account for >90% of people afflicted with RS (Neul et al., 2010; Sajan et al., 2017). Symptoms usually emerge between 1 and 3 years of age. Breathing, including breath holding, and autonomic abnormalities are common and can make clinical differentiation between these events and seizures challenging. Similar to DS, behavioral problems and parkinsonian features can be present in RS patients and are more evident at older ages. Current care includes medical management of symptoms such as seizures, behavioral management, nutrition to ensure adequate calorie and vitamin D intake, augmented communication strategies, and physical and occupational therapies.
MECP2 protein functions as a transcriptional repressor that impacts a number of genes and signaling pathways (Ip et al., 2018; Ebert et al., 2013). Pathogenic variants are often de novo point mutations but also can be truncation mutations. Females are mosaic for the mutation, generally with cells that express wild-type MECP2 and cells that express pathogenic MECP2. Males with pathogenic MECP2 do not have wild-type MECP2 and therefore have a more significant presentation in the neonatal period that is often lethal.
Related to RS is MECP2 duplication syndrome, a clinical syndrome with some overlap features though a distinct syndrome from RS characterized by seizures, hypotonia, dysmorphic features, and regression of speech and motor skills, but at a later age of onset than RS. Hearing loss is prominent, and there is an increased risk of infection leading to death in early adulthood (Lim et al., 2017; Miguet et al., 2018; Giudice-Nairn et al., 2019).
The genetics of RS present some unique challenges to gene modulatory treatment that differ from DS. MECP2 influences a host of cellular processes and functions that vary by cell type (Vashi and Justice, 2019). Additionally, as described above, too little gene product and too much gene product each have clinical consequences. Similar to DS, clinical outcome measures can be challenging due to the progressive nature of the disease, number of symptoms involved, and heterogeneity in the clinical presentation (Kaufmann et al., 2016; Katz et al., 2016).
The genetics of RS present an opportunity for TANGO therapy in that the wild-type MECP2 allele can be reactivated. Experiments in animal models of RS have been able to reverse the disease process with gene “reinstatement” or reactivation of the wild-type MECP2 on the inactivated X chromosome in female mice (Guy et al., 2007; Gadalla et al., 2017). Future translation of this approach to human trials must consider the impact of developmental changes throughout the life span such that treatment may not be as successful in older patients. Additionally, this therapeutic approach may theoretically stimulate brain growth, which would present a unique challenge if this occurred after cranial sutures have fused.
Other therapeutic approaches for RS under consideration include gene replacement therapy; however, this approach would require precise delivery to ensure that the proper amount of gene is delivered (Gadalla et al., 2013; Garg et al., 2013), as MECP2 overexpression has been shown to lead to seizures and hypoactivity in animal models (Zhang et al., 2008; Luikenhuis et al., 2004) and would be expected to have a similar clinical presentation as MECP2 duplication syndrome in humans. Tillotson et al. demonstrated that gene replacement with a truncated Mecp2 construct delivered ICV via an AAV vector led to improvement in Mecp2-null mice (Tillotson et al., 2017). Trials are currently in the planning stages for a single-dose AAV9 treatment that would deliver complementary DNA to human brain, though the published literature does not yet address how overexpression will be avoided (2021b). Additionally, murine models of MECP2 duplication syndrome have been used to demonstrate that use of a gapmer ASO to suppress MECP2 expression can lead to seizure cessation and recovery in adult mice (Sztainberg et al., 2015).
Angelman syndrome (AS) is another DEE with a unique genetic presentation. AS is associated with maternal disruption of the gene UBE3A, which can be caused by a variety of genetic perturbations, including pathogenic variants of the maternal allele, imprinting abnormalities, and deletion syndromes. Clinically, children with AS present with developmental delay around 6–12 months of age followed by a plateau of development around 24–30 months, with abnormalities of fine and gross motor function, movement disorders, speech and cognitive function disorders, seizures, and sleep disturbances (Williams et al., 2006; Bindels-de Heus et al., 2020; Pelc et al., 2008). UBE3A is located on chromosome 15q11.2 within 15q11-13, a region that has been recognized to have frequent deletions and duplications. When deletions of the maternal chromosome occur, the clinical presentation is AS (Matsuura et al., 1997; Sutcliffe et al., 1997). Similar to RS, there are clinical disorders of UBE3A other than AS, which are associated with gene overexpression versus underexpression. Duplication of UBE3A is characterized by autism and increased risk for schizophrenia (LaSalle et al., 2015; Noor et al., 2015). The absence of paternal UBE3A is associated with Prader-Willi syndrome, a clinical syndrome that is characterized by hypotonia, short stature, and later development of hyperphagia (Cheon, 2016). Further complicating the clinical picture of UBE3A disorders is that the paternal or maternal origin of the gene impacts its expression in neurons, where only maternally derived UBE3A is typically expressed (Jones et al., 2016; Runte et al., 2004; Vu and Hoffman, 1997). Thus, much like RS, the complicated genetics of UBE3A disorders must be considered when developing gene manipulation strategies, as both over- and underexpression can lead to clinical symptoms.
Several AS treatment options have been attempted using animal models, including restoration of protein by activation of paternal UBE3A in the brain, a concept called “unsilencing.” The UBE3A-ATS gene product silences the paternal allele of UBE3A (Meng et al., 2012; Chamberlain and Brannan, 2001; Landers et al., 2004), and therefore inactivation of its expression in neurons may lead to restoration of UBE3A. ASOs that reactivate the paternal UBE3A allele in mice have been shown to be effective (Meng et al., 2015). Another class of medications, topoisomerase I inhibitors (Huang et al., 2011; Lee et al., 2018), have also demonstrated success in unsilencing the paternal UBE3A in animal models. These drugs suppress the UBE3A-ATS transcript, permitting expression of paternal UBE3A in mouse and human neurons (King et al., 2013). Unfortunately, this approach has clinical challenges, as there can be off-target, long-lasting side effects through the suppression of other transcripts in human cells. Targeting UBE3A-ATS with ASOs (Meng et al., 2015) has led to rescue of symptoms in an AS murine model but, like other ASO treatments, required repeated administration. Additionally, AAV therapy using Cas9 technology was used successfully in mouse models to unsilenced the paternal UBE3A (Wolter et al., 2020). ASO clinical trials were recently started, with reported improvement in symptoms, but have been on hold due to side effects associated with the treatment (2020).
Another approach that has been considered is the AAV administration of one of the isoforms of UBE3A, but this represents a similar challenge as RS, regarding achievement of the right amount of expression rather than creating overexpression of the gene and mimicking duplication syndromes that are associated with both genes. Also critical will be selection of the appropriate UBE3A isoform. The protein is expressed in both the cytoplasm and the nucleus, but AS appears to be associated with mutations that impact the nuclear isoform (Avagliano Trezza et al., 2019). Additionally, an optimal promoter must be chosen to lead to UBE3A protein expression at the correct levels across the life span, without overshooting normal expression levels (Zylka, 2020).
There are several aspects of gene regulation therapy that must be considered regardless of the mechanism: viral vector delivery or ASO. Timing of treatment for many DEEs seems particularly important and related to outcomes, as there may be aspects of the disease that are not reversible. Significant overexpression or having “too much of a good thing” must not result in adverse effects, as there currently is little ability to regulate the amount of response in a single patient. Long-term side effects and stability of response are not yet known and will not be able to be answered in short-term clinical trials. Finally, CRISPR approaches can result in significant off-target effects, and viral-administered treatments in general are not reversible due to the single-administration limitations. Theoretically, there is the potential for gene modification of germ cell lines, though none have yet been reported. Regarding DS, current ASDs target symptoms, predominantly seizures, though the syndrome impacts many other organ systems. The promise of advanced, gene regulatory therapies targeting SCN1A expression has the potential to improve all of these aspects of DS.
Studies in animal models suggest that early treatment during critical periods of brain development, prior to onset of clinical symptoms, may lead to long-term improvement (Han et al., 2020). While this is an exciting prospect, it is currently challenging to implement in people. Early-life treatment would require prenatal testing and, for some genetic epilepsies, may require treatment prior to birth to maximally impact brain development. Phenotype-genotype correlates are not well established and thus predicting which children will develop disease and thus benefit from early treatment is challenging. For DS, not every child identified with a known or presumed pathogenic variant will develop clinical DS. Additionally, with so many new variants being identified, there is increased need for rapid functional testing to determine who may have GOF mutations, as many of the proposed gene regulatory therapeutic strategies risk disease exacerbation rather than improvement. Initiating treatment at very early stages of disease or prior to emergence of disease also creates challenges for assessing outcomes. Standard outcome measures usually include comparison of post-treatment symptoms to the patient’s baseline or that of historical controls. Treatment prior to symptom onset will create the need to rethink therapeutic strategies and may require new outcomes measures such as time to first seizure or developmental quotient compared to well-established historic controls that represent the entire scope of the disease presentation.
There are several challenges associated with viral vector therapies, including gene size limitations and the impact of the immune response. The size of the gene itself may hinder treatment, and this appears to be the case for SCN1A. Gene replacement is not currently practical via viral vector in many cases due to a gene size that will simply not fit in the carrier. Additionally, AAV administration induces an immune response from the body leading to many negative issues that must be considered. The immune response can lead to destruction of the AAV vector prior to reaching its target, as well as a systemic inflammatory response. Repeated exposures are expected to demonstrate a greater response; therefore, the vector can, in practice, only be used once. If antibodies to the viral vector are already present, this may prohibit treatment. This single-treatment requirement means that there must be a clinically meaningful response with the first delivery. As therapeutic strategies are improved, the presence of an immune response to AAV subtypes may preclude patients who have received one treatment from receiving subsequent new treatments. When treating very young infants, the presence of maternal antibodies may lead to a delay in treatment. Dosing must also be considered, and since there is only one opportunity, the dose must be large enough to have an effect but cannot be so large that it induces a strong Inflammatory response (Wilson, 2009). Viral vectors, even when administered directly in the CNS, can have a broad distribution, leading to increased uptake in the liver (Grimm et al., 2006; Bevan et al., 2011).
There are several reports of adverse reactions to existing AAV vector treatments. Some best examples are from the recently approved onasemnogene for use in children with spinal muscular atrophy (SMA). These include elevated transaminases in 90 of 100 patients (Chand et al., 2021); liver-associated adverse events were reported for 34 of 100 (Chand et al., 2021), and transient liver failure has been reported in 2 patients (Feldman et al., 2020). Similar events were also reported in animal models (Hinderer et al., 2018). The medication has a black-box warning for acute hepatoxicity, although there is some evidence that steroids may help (Stevens et al., 2020). Decreased platelets have been reported, though this effect was transient and did not require intervention (Waldrop et al., 2020). Thrombotic microangiopathy (hemolytic anemia, thrombocytopenia, and acute kidney injury), a very rare entity, has been reported in 3 of 500 children treated with onasemnogene (Chand et al., 2021). Virus is known to be shed in the stool; therefore, caution will need to be taken by caretakers who may be immunocompromised. Families that have more than one child with DS will also need to be cautious about exposure to the rest of the family, so that antibodies to the vector do not develop in the other affected family member and preclude treatment.
ASO treatment has reported side effects, although milder than with AAV. In patients treated with nusinerson for SMA, the largest population treated with ASOs to date, almost all patients in the trials reported adverse events, although only one was thought to be treatment related, post-procedural nausea (Darras et al., 2019). Elevation of hepatic enzymes, proteinuria, and thrombocytopenia have been reported across several ASO treatment trials (Jason et al., 2004; Chan et al., 2006); however, these are often mild and transient in nature. Recent trials of GTX-102 for Angelman syndrome have been paused due to lower extremity weakness developing in all treated patients within 1–4 weeks of administration of study drug. Two of the children were unable or unwilling to walk but have since had resolution. Treatment with intravenous immunoglobin and steroids was given, and there was reported clinical improvement. Despite these concerns, there was reported improvement in overall disease symptoms, and there are plans to amend the protocol and resume the study (2020).
For many gene regulatory therapy options, and particularly for those proposed for DS, GOF variants must be identifiable prior to consideration of treatment, as an approach that impacts the gene product directly will lead to an increase in expression of the pathogenic allele and thus exacerbate the underlying disease. GOF variants can be identified via phenotype in some situations but ideally must be confirmed with functional testing. While at this time, obtaining functional testing in clinical practice can be very challenging, increased functional testing will be necessary to understand the mechanisms of specific patient variants to determine which individuals would best benefit from specific treatment strategies.
In selecting gene-modifying treatments, several critical factors should be considered. The duration of action of these treatments is not yet known, but as discussed above, treatments that are administered via AAV will only be able to be administered once, whereas treatment with ASO allows and may require multiple doses. ASO therapy requires administration within the CNS, as ASOs do not cross the blood–brain barrier. AAV therapy has been administered intravenously, but current animal models have used ICV administration. This may be required for AAV treatment in certain diseases, and appears to be the case in DS, to ensure that highest possible level of drug reaches the target cell population.
There are several key aspects of study design that must be considered in clinical trials. Timing of treatment is likely critical. As discussed above, earlier treatment leads to better outcomes in animal models. However, treatment prior to clinical presentation is not as easy to replicate in the clinical setting as it is in the laboratory environment. Genetic testing is currently not performed until after clinical presentation in infants, and pathogenic variants do not correlate as closely to disease to allow prediction of disease progression.
Primary outcome measures in clinical trials have only thus far included seizure occurrence. This is an obvious outcome that is measurable both accurately and in a short period of time; however, very young children may not yet have seizures frequent enough to be measurable in a time-limited trial and, with the recent addition of newer ASDs, a larger percentage of children have had a reduction of seizures such that they may not qualify for a trial. There are several other possible outcomes that are concerning to families but may be challenging to measure consistently and in short time period, such as behavior, cognition, attention, and gait, with changes noticed over months to years rather than days to weeks. Additionally, outcome measures will need to be considered for long-term evaluation and safety that are applicable over the life span. Further, there is lack of clarity as to which aspects of genetic DEEs are reversible with gene regulatory therapy and which are not. Adolescents and adults may benefit from treatment in ways that are not yet clearly measurable, such as improvement in behavior, reduction in rigidity, or improvement in mobility.
There is great excitement regarding the several developing options to treat DS that will require AAV administration, but due to the immune response, unless unique vectors are used, an individual will only be able to receive one of these therapies. Therapeutic development will require consideration of how to use unique vectors that will allow for several different approaches to be used in a single person and how to select patients who are most likely to benefit from a specific therapy.
The cost of current gene therapy products on the market is very high, approaching millions of dollars per patient, and while uncomfortable to consider, cannot be ignored. Clinically, the cost of treatment is considered in the treatment of SMA where ethical discussions are ongoing regarding whether these treatments are cost-effective (Connock et al., 2020) and, if so, who should carry the burden of the cost of care. Regardless of treatment choice, the cost of care for a patient with a lifelong disability like SMA (Dangouloff et al., 2021) or DS (Campbell et al., 2018; Jensen et al., 2017; Whittington et al., 2018) is very expensive, and these discussions will certainly continue. As more long-term data are obtained, we will gain a better understanding of the true reduction in cost of care created by precision therapeutics.
There are exciting opportunities on the horizon for precision gene regulatory therapeutics to treat many of the genetic DEEs; however, translating preclinical data will require significant scientific work for each individual etiology, as each gene brings unique challenges. Careful consideration will need to be given to genetic analysis and characterization of functional consequences of variants to allow for development of improved animal models that represent these variants and appropriate selection of study participants. Development of reliable and valid outcome measures beyond seizures will improve our understanding of treatment responses across the life span. As these therapies are studied in clinical trials, there will be a need to determine the best candidates for each treatment approach, as well as the risks and benefits of multiple treatments in a single patient. Finally, we cannot forget that there is a need for treatment options in patients with unique variants such as GOF.
A portion of Dr. Isom’s research was funded by a research grant to the University of Michigan from Stoke Therapeutics. Dr Knupp has received research funding from Stoke Therapeutics, Encoded Inc., Zogenix Inc., Eisai, and West Pharmaceuticals and consulting funding from Stoke Therapeutics, Encoded Inc., Zogenix Inc., Biocodex, GW Pharmaceuticals, and West Pharmaceuticals.
This is an open access publication, available online and distributed under the terms of a Creative Commons Attribution-Non Commercial-No Derivatives 4.0 International licence (CC BY-NC-ND 4.0), a copy of which is available at https://creativecommons.org/licenses/by-nc-nd/4.0/. Subject to this license, all rights are reserved.
Your browsing activity is empty.
Activity recording is turned off.
See more...
