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/9780197549469.003.0005
Approximately 1% of 25,000 genes and 25 million single-nucleotide polymorphisms in the genome of modern humans are epilepsy variants. Epilepsy variants are the molecular basis of fatal progressive myoclonus epilepsies, epileptic encephalopathies, and other genetic epilepsies. This chapter annotates the early success and key turning points in the quest to cure fatal epilepsies. The first key turning point occurred in 2018 when “every-2-weeks” intraventricular infusion of cerliponase alfa, enzyme replacement therapy, halted progression of an early childhood progressive myoclonus epilepsy (CLN2 Batten disease), as reported by Schulz et al. In 2021, Schaeffer et al. reported 2 years of intraventricular cerliponase alfa delayed or prevented onset of symptoms in presymptomatic CLN2 infants. Earlier, in 2017, the Food and Drug Administration (FDA) and European Medicines Agency approved Nusinersen, an antisense oligonucleotide, and AVXS-101 (Zolgensma), a viral-mediated gene replacement therapy, as disease halting and potential cures for Werdnig Hoffmann type I spinal muscular atrophy of infancy (SMN1). Even now, breakthrough advances by others include (1) ex vivo HSPC transduction gene therapy in metachromatic keukodystrophy, and (2) genome editing with zinc finger nuclease opening DNA strands and inserting a correct alpha-L-iduronidase into the genome of Hurler mucopolysaccharidoses patients. Recently, the FDA proclaimed that they expect to approve 10–20 cell and gene therapies a year from 2025 onward. As clinical trials raise hopes for cures, vigilance for adverse effects and ethical concerns must not waiver.
The epilepsy genome constitutes approximately 1% of 25,000 genes contained in the genome of the modern human (Homo sapiens sapiens). Of 25 million common and rare single-nucleotide polymorphisms or variants (SNPs), about 1% are epilepsy variants. Epilepsy variants include its nucleotide sequences, coding and noncoding DNA (enhancers, promoters and terminators, transposable elements, and repetitive DNA), and mitochondrial DNA. Epilepsy genomics, the study of the epilepsy variants’ function, structure, analysis, regulation, editing, and evolution (Cristescu, 2019), together with the natural history of epilepsy syndromes, laid the platform for the five genetic and clinical ingredients necessary for a quest to cure epilepsies. The “concept of genome” came from Hans Winkler (1920).
This fifth edition of Jasper’s Basic Mechanisms of the Epilepsies will be remembered as the edition when the tools for the medical interpretation of the human epilepsy genome and discovery of epilepsy variants were in high gear, when epilepsy genomics laid the foundation for precision medicine in epilepsy, and when genomics-based therapeutics in epilepsy moved from symptomatic alleviation of disease to modification and halting disease progression and to the courageous quest for cures. One by one, genomics-based functional studies are defining disease mechanisms of fatal and nonfatal, nonprogressive genetic epilepsies of infancy, childhood, adolescence, and adulthood. The “single, one-and-done” gene replacement therapy (GRT, or introducing a wild-type sequence to replace the defective sequence) and disease mechanisms, targeted, down-regulated, or up-regulated by antisense oligonucleotides (ASOs), and disease mechanisms modified or edited by the genius of our researchers in basic mechanisms of epilepsies and the pharmaceutical industry provide the paths toward (a) the future curing of nonfatal common genetic epilepsies and (b) the beginning of the end of some fatal genetic epilepsies.
This chapter describes the progress in epilepsy genomics that transformed the quest to cure epilepsies by establishing five key genetic and clinical ingredients necessary for this quest. The chapter also identifies the key turning points when proof of principles were obtained in epilepsy clinical trials that a fatal epilepsy can have its disease progression halted, when symptoms can be reversed, and when a possible cure can be proven in presymptomatic individuals. The “every-3-months” intrathecal ASO injections are contrasted with the “single, one-in-a-lifetime” intrathecal injection of enduring viral-vector based GRT. At the start, the chapter provides a brief summary of the history of epilepsy variants and epileptogenesis concepts generated by disease mechanisms of epilepsy variants. In closing, the chapter emphasizes the need for continuing vigilance for unexpected adverse effects of novel therapeutics and highlights key social and ethical concerns in the quest to cure the epilepsies.
Epilepsy variants are the molecular basis of specific genetic epilepsy syndromes, causing their phenotypes and traits. Historically, the complexity and variability of phenotypes of age-dependent genetic epileptic syndromes and their overlapping clinical features limited the resolution of phenotype-based epilepsy classification and confounded epilepsy nosology.
Then came the first pioneers who discovered epilepsy genetic variants in the decade from 1995 to 2005 (see Table 5–1 for list of epilepsy genetic variants). In 1995, Steinlein and her team were among the first pioneers to show germline epilepsy causing variants/mutations in the alpha4 subunit of nACH receptor (CHRNA4) (Steinlein et al., 1995, 1997). But the story of curating the epilepsy genome started earlier than 1995 with Mark Leppert and Elving Anderson. In 1989, Leppert and his team at the University of Utah, in collaboration with Anderson’s group at the University of Minnesota, established genetic linkage of the long arm of chromosome 20 to benign familial neonatal seizures (BFNS). This was the first report of a genetic linkage of a chromosome site to an epilepsy syndrome (Leppert et al., 1989). Nine years later, in 1998, the same collaborating teams reported KCNQ2 as the potassium channel mutated in an inherited epilepsy of newborn. At the time it was considered a novel potassium channel gene. Now we know that KCNQ2 variants (more than 80 variants described) are the most common cause of BFNS (Singh et al., 1998; Charlier et al., 1998; Biervert & Steinlein, 1999; Biervert et al., 1998).
Some Epilepsy Genes and Their Phenotypes.
At about the same time, while the first epilepsy variants were being discovered in Germany and Utah, Scheffer and Berkovic, in 1997, taught us about a then unrecognized syndrome of generalized epilepsy with febrile seizures plus (GEFS+). Over the following 2–3 years, this Australian team showed that a variant in SCN1B could cause GEFS+, and that variants in SCN1A are the most common genetic causes of GEFS+ and febrile seizures (Scheffer and Berkovic, 1997; Wallace et al., 1998; Escayg et al., 2000, 2001).
As the twenty-first century turned, in 2002, Cossette with Rouleau and colleagues, in Montreal, identified, for the first time, that a variant in an inhibitory neurotransmitter receptor, namely GABA-A receptor (an Ala322Asp variant in GABRA1), could cause epilepsy in clinically affected members of a large multiplex French-Canadian family with juvenile myoclonic epilepsy (JME) (Cossette et al., 2002). Since then, variants in GABA-A gamma 3 subunit in chromosome 15q11 and variants in GABA-A receptor gamma 2 subunit in chromosome 5 and variants in GABA-A receptor beta 3 subunit have been reported in absence epilepsy syndromes. It is now accepted that Absence epilepsies, as part of the GEFS+ syndrome or as a separate and exclusive absence syndrome, are caused by GABR variants (Baulac et al., 2001; Wallace et al., 2001; Tanaka et al., 2008).
In 2003, one year after the first discoveries of variants in GABA-A receptors causing absence epilepsies and JME, Chen et al. (2003) reported association between variants in a calcium channel, CACN1A and childhood absence. Variants in CACN1A and CACNB4 have now also been observed in childhood absence with ataxia (Imbrici et al., 2004). Thus, in the first decade of epilepsy variants, “proof of principle” evidence was documented mainly for ion channelopathies and transmitter-receptoropathies.
In 2004, the first report of variants in a non-ion channel and non-transmitter receptor gene was made by Suzuki et al. (2004) in families with JME. The variants were in EFHC1, a novel gene, which encodes a microtubule-associated protein with an EF-hand motif whose variants disturb R-type VDCC and TRPM2-calcium currents. Among 17 cohorts with 54 EFHC1 variants, reported in literature by 2018, 23 were pathogenic or likely pathogenic. According to the National Human Research Institute and American College of Medical Genetics and Genomics, gene-level evidence, and variant-level evidence, EFHC1 is the first non-ion channel associated with JME. Since then, variants in other non-ion channel genes, such as intestinal cell kinase (ICK) and carboxypeptidase A6 (CPA6), have also been observed in JME (Bailey et al., 2017, 2018; Sapio et al., 2015). Ottman et al. (2010) were the first to report variants in a non-ion channel gene, leucine-rich glioma-inactivated 1 (LGI1), in a focal epilepsy known as autosomal dominant lateral temporal lobe epilepsy (ADLTLE with auditory features). About half of ADLTLE families have variants in LGI1 or epitempin (Kalachikov et al., 2002; Micheluchi et al., 2003).
Curating the epilepsy genome sharpened the diagnostic view of the medical community to genetic testing of epilepsy syndromes. But it was the 2001 discovery of somatic recurrent de novo mutations in a sporadic epilepsy syndrome first described by Dravet (Claes et al., 2001) that sealed the deal for genetic testing—for example, in the alpha1 subunit of the sodium channel (SCN1A) arising mainly from paternally derived chromosomes, the beta1 subunit of sodium channel (SCN1B) (Wallace et al., 1998; Escayg et al., 2000, 2001), and in protocadherin 19 (PCDH19) in X-linked females with Dravet syndrome (Depienne et al., 2009; Dibbens et al., 2008). This recognition of de novo mutations in sporadic epilepsies led to whole exome sequencing (WES) of developmental disabilities mixed with infantile epileptic encephalopathies, such as infantile spasms in West syndrome and Dravet syndrome. In 2013, WES revealed de novo variants in GABRb3, CACNA1A, CHD2, FLNA, GABRa1, GRIN1, GRIN2B, HNRNPU, IQSEC2, MTOR, and NEDD4L (Allen et al., 2013) and, in 2017, de novo variants in NTRK2, GABRB2, CLTC, DHDDS, NUS1, RAB11A, GABBR2, and SNAP25 (Hamdan et al., 2017).
With association of disease-causing variants with specific epileptic syndromes also came realization of specific antiseizure drugs (ASDs) as “keys that fit the lock” of disease mechanisms of epileptic syndromes. The concept of “antiseizure drugs that act as sodium channel blockers worsen genetic epilepsies” had to be modified. De novo SCN1A variants Dravet syndrome or epileptic encephalopathy is worse with ASD sodium channel blockers (Dravet et al., 1992, 2005), but de novo SCN8A or SCN2A variants Dravet syndrome can respond favorably with ASD sodium channel blockers. Some KCNT1-deficient malignant migrating partial epilepsies in infancy respond to quinidine (Abdelnour et al., 2018) and some to carbamazepine. Sodium channel blockers can also be effective in variant KCNQ2, while ketogenic diet has diverse efficacy in variant SCN1A, CDKL5, KCNQ2, STXBP1, and SCN2A (Ko et al., 2018).
There are now 90 to 172 disease-causing variants primarily expressed as epilepsies or epileptic encephalopathies targeted in gene-panel sequencing platforms by commercial genome centers (Table 5–1) and about 40 genetic diseases whose phenotype significantly includes epilepsy (Table 5–1). There is an ever-growing list of epilepsy genes, and a complete list here and now would not be possible.
The concepts of epilepsy syndromes as epilepsy diseases (Roger et al., 1985), their appearance as developmental epileptic encephalopathies, and their association with autism spectrum disorder (Specchio et al, 2022) were bolstered when the epilepsy causing pathogenic variants/mutations and disease mechanisms were established.
These, in turn, led to new concepts of epileptogenesis:
Some Genetic Diseases Whose Phenotype Includes Epilepsy.
Epilepsy genomics in studying structure and function of epilepsy variants, their epilepsy-causing mechanisms in knockout and knock-in mouse models of epilepsy phenotypes, and the rescue of epilepsy and disease symptomatology of epilepsy syndromes in conditional knock-in mouse models helped identify the key and critical genetic ingredients.
Epileptologists who specialize in rare epilepsies by studying the natural history of these rare epilepsy syndromes can identify the critical window or time when treatment should start in their quest to cure the disease. Additional elements in the road to the quest are the earliest clinical, electroencephalogram (EEG), and biological biomarkers that can be used to gauge onset and progression of disease.
These key elements are as follows:
The monogenic fully penetrant epilepsy-causing variant must be defined and its structure and disease mechanism(s) offered as a therapeutic target.
Transgenic mouse models for the clinical epilepsy must be available for the preclinical trials of potentially curative therapeutic agents.
Potentially curative therapeutic agents must cross the blood–brain barrier (BBB); otherwise intrathecal or intracerebral-intraventricular route will be route of administration.
The critical window to start treatment in clinical trials for proof of a cure. Natural history studies of the clinical epilepsy syndrome must define the following:
The onset and trajectories of disease progression (but not only this)
The EEG spike wave indices, sleep architecture, and cognitive conditions of the presymptomatic states
When clinical disease starts versus when pathology starts and the optimal time to start treatment to prove a cure
A population of genotyped probands and affected family members must be available for the actual clinical curative trial (see Table 5–3).
Key Turning Points in the Quest for Cures.
There are several key turning points in epilepsy genomics, specifically in human clinical trials and in knockout or knock-in mouse models that provided proof of principle that a progressive epilepsy disease can be halted, symptoms reversed, and a possible cure obtained.
All the five above genetic and clinical ingredients were present in the quest to cure neuronal ceroid lipofuscinoses (NCL), a previously incurable epilepsy/neurodegenerative disease of childhood. There are 13 CLN pathogenic variants in genes that process lysosomal enzymes and receptors and that are transmitted as an autosomal recessive trait. There is one pathogenic CLN variant that is inherited as an autosomal dominant trait (CLN4). The CLN4 pathogenic variant is in DNAJC5 gene, which encodes the presynaptic co-chaperone cysteine string protein. Disruption of lysosome functions lead to the buildup of cellular aggregates composed by lipofuscin, fine yellow and brown pigments with autofluorescent granules—hence the name, lipofuscinoses. Cell aggregates of lipofuscin accumulate in neurons of the central nervous system (CNS), retinal and ganglion cells of the eye, liver, heart, spleen, adrenals, and kidney. Clinically, NCL patients share common signs of epileptic seizures, visual and cognitive decline, and regression of motor skills and balance. Age at disease onset varies according to the pathogenic variant.
A major problem of these NCL diseases is that the disease mechanisms of the 14 pathogenic variants known to cause their clinical phenotypes are not completely understood and remain unknown in some. Because of this, most research efforts in NCL diseases in the last decade were devoted to developing human gene therapy. Gene therapy as defined by the Food and Drug Administration (FDA) (Cellular & Gene Therapy Guidances, July 20, 2018) and the European Union Commission (Directive 2001/83/EC, Part IV of Annex I) is a “biological medicinal product containing recombinant nucleic acid used in or administered to a human to regulate, repair, replace, add or delete a genetic sequence with the aim to cure or treat diseases” (Jensen et al., 2021).
Gene therapy includes (a) “in vivo” vector-mediated gene replacement, (b) “ex vivo” cell transduction gene therapy, and (c) genome editing. Most research programs in NCL chose an in vivo viral-vector-mediated GRT program that targets loss-of-function mutation diseases. This GRT approach reached high hopes for a cure when the FDA and European Medicines Agency (EMA) both approved Zolgensma (onasemnogene abeparvovec) for in vivo adeno-associated virus mediated GRT for spinal muscular atrophy. More recently, EMA approved an ex vivo gene therapy with lentivirus vector transduced autologous CD-34- positive stem cells for therapy of metachromatic leucodystrophy (Libmeldy). FDA approval of cerlifonase alfa (Brineura) for CLN2 disease has given more hope for families and parents who have children with NCL.
In 2018, Schulz et al. were the first to show a halt to disease progression for a neurodegenerative epilepsy syndrome, neuronal ceroid lipofuscinosis type 2 (CLN2) (Schulz et al. 2018). CLN2 disease has a rapidly progressive and invariably fatal outcome, which was modified by intraventricular enzyme replacement therapy. CLN2 disease is caused by pathogenic variants in the gene encoding lysosomal enzyme tripeptidyl peptidase 1 or TPP1. CLN2-deficient children are normal until 2 to 4 years; then the disease starts as an epilepsy (tonic-clonic or partial seizures in 70%) and delayed language development (57%). Slowed psychomotor development may be followed by a clinical standstill, after which delayed language development is recognized. Cognitive, motor, and visual functions regress, leading to death by age 12 years due to accumulation of lysosomal storage material in the CNS and retina.
Diagnosis is confirmed by homozygote or compound heterozygote mutations in the TPP1 gene and deficiency or absence of TPP1 enzyme activity. EEG shows slow and attenuated background activities, epileptiform activities in occipital regions, and photoparoxysmal responses at low frequencies (1 to 2 Hz) of photic stimulation. Brain MRI shows early evidence of cerebral and cerebellar atrophy. Optic coherence tomography may show retinal degeneration and accumulation of hyperreflective material, abnormal visual evoked potentials, and diminished electroretinographic amplitudes.
In the study by Schulz et al. (2018), 24 children had their CLN2 disease documented by TPP1 enzyme activity and CLN2 genotype. Median age at enrollment was 5 years, range 3.3 years to 9 years, with subjects classified as having mild to moderate disease progression as measured by a combined score of 3–6 on the motor and language domains of the CLN2 clinical rating scale.
Subjects received during the stable dose phase, 300 mg of intraventricular cerliponase alfa (a recombinant proenzyme form of human TPP1, also called zymogen, developed by BioMarin pharmaceuticals) every 2 weeks for at least 96 weeks, after initial doses of 30 mg, 100 mg, or 300 mg. Intraventricular infusion rate at 2.5 ml per hour for 4 hours through a surgically implanted Ommaya or Rickham ventricular reservoir is meant to replace the deficient enzyme in affected brain tissues. Cerliponase alfa is activated upon uptake into the acidic pH environment of lysosomes in vivo and becomes the mature proteolytic form of rhTPP1 that catabolyzes lysosomal aggregated storage material. Concentrations peak in CSF at the end of ~4-hour infusion and 8 hours thereafter in plasma.
The primary efficacy outcome was the time until an “unreversed” first two-point decline in the motor and language domains of the CLN2 clinical rating scale, occurred, compared to 42 historical controls. The treated children had slower decline in motor and language function than that in CLN2 historical control patients. Adverse effects consisted of convulsions, pyrexia, vomiting, hypersensitivity reactions, and device-related complications, including grade-3 infection, leakage, and an increase in white cell count in cerebrospinal fluid in half the patients. There was no apparent correlation between cerebrospinal fluid and plasma peak concentrations and incidence of adverse effects. There was no relationship between magnitude of csfexposure and changes in CLN2 score, indicating maximum benefit across the range of exposure with 300 mg every 2 weeks.
This clinical trial that started as the first in human phase I/II study and one subsequent phase I/II extension study still in progress were made possible by a safety and efficacy study and pharmacology/pharmacokinetics study in young dogs with spontaneous homozygous TPP1 deficiency and safety studies in monkeys. Cerliponase alfa clears lysosomal storage material and preserves neuron morphology and reduces brain inflammation in CLN2-deficient dogs. Treated dogs had delayed onset and slower progression of neurologic signs, brain atrophy, preserved cognitive functions, and an extended lifespan (Schulz et al., 2018).
Nusinersen provided the first proof of principle that disease progression can be halted, symptoms decreased, and risk for death reduced in a non-epilepsy disorder (spinal muscular atrophy [SMA] rarely manifests with myoclonus epilepsy) (Chen, 2020; Finkel et al., 2016). SMA is caused by a loss-of-function variant in the “survival motor neuron1 or SMN1 gene.” SMN1 gene encodes a protein essential for survival of alpha motor neurons in the spinal cord. Most commonly, this is a homozygous deletion of exon 7 of the survival motor neuron 1 gene resulting in SMN protein deficiency. The human genome has a second and similar gene, SMN2, which is different from SMN1 only by a few nucleotides, notably, a nucleotide variant 7. This variant in SMN2 leads to exclusion of exon 7, resulting in an unstable truncated protein without exon 7. In SMA, a pathogenic variant in SMN1 is present and a few copies of SMN2, which is not enough to transcribe/translate into functional SMN proteins.
Clinically, four separate syndromes of SMA are present. Type Zero or Type 0 with one SMN2 copy starts in utero. Infants are dependent on mechanical ventilation at birth, and death usually follows before 6 months of age. Type 1 SMA or Werdnig Hoffmann disease with one to three SMN2 copies is the most common clinical syndrome, accounting for approximately 60% of all SMA cases. Type 1 SMA starts before 6 months of age, as hypotonia and muscle weakness require posture, respiratory, and feeding support. Median age to ventilation is at 10.5 months and to exitus is at 13.5 months. Type 2 SMA or Dubowith disease with one to two SMN2 copies is often included in the clinical trials because it also starts early, after 6 months of age, usually 7 to 18 months, with infants able to sit up but unable to walk independently, also eventually needing respiratory and feeding support. Life expectancy may reach adulthood. Type 3 Kugelberg-Welander disease, with three to four SMN2 copies, starts at 18 months with inability to stand and walk. Eventually gaining strength, patients have a normal life expectancy. Life expectancy is also normal in Type 4 with four SMN2 copies with adult-onset mild muscle weakness.
In 2006, an intron-splicing silencer N1 sequence in intron 7 of the SMN2 gene, which enhances skipping of exon 7 and produces full and functional SMN2 protein, was identified. ASOs were then designed that bind to the intron-splicing silencer N1 region of the SMN2 pre-messenger promoter exon 7, upregulating transcription of full-length SMN2. This led to the original open-label phase 1/2 clinical trial in 2016 for “the every-3-to-4-months” intrathecal nusinersen (brand name Spinraza) that was followed by an open-label extension study (Chen, 2020). Then, in 2017, Finkel et al. (2017) led a double-blind, placebo-controlled trial with type 1 and type 2 SMA patients.
All trials showed that multiple intrathecal nusinersen reduced symptoms, halted disease progression, and reduced the risk of death or permanent ventilation by 47%. All three clinical trials also showed a favorable safety profile. Among the three clinical trials, a total of 194 type 1 SMA or Werdnig-Hoffman type SMA were treated with nusinersen with similar positive results of improvement (Finkel et al., 2016, 2017; Aragon-Gawinska et al., 2018; Pechmann et al., 2018).
Since its first reports, in 2018, cerliponase alfa (Brineura) had its first global approval by the FDA in the United States for CLN2 patients that are over 3 years of age and in the EMA for CLN2 patients of all ages. Real-world experience of effectiveness of cerliponase alfa in CLN2 patients has been reported in Colombia (eight patients) and in France (seven patients) with similar success in halting disease progression and even improvement of clinical signs when patients with a CLN2 score greater than 3 were enrolled for intraventricular enzyme GRT.
Remarkably, Schaefers et al. (2021) reported that a 23-month-old presymptomatic CLN2 individual was treated with intraventricular cerliponase alfa for 2 years and demonstrated a delay in onset of symptoms. The individual had walked at 14 months, running and climbing stairs with fine motor skills. He spoke two-word sentences, had normal vocabulary for age, and had good nonverbal communication skills. He had no symptoms of CLN2 disease at the start of treatment. This 23-month-old subject had maximum scores (total score of 12) on all subdomains of the CLN2 rating scale, which remained the same over the 2-year course of treatment. An increase in cognitive development from 21 to 39 months was observed, and cerebral and cerebellar atrophy was not observed on MRI after 2 years of treatment when the subject reached 4 years of age. Remaining asymptomatic at 4 years of age, the subject did have a prolonged P100 latency time on visual evoked potential studies. Results, nonetheless, suggest that disease onset has been delayed and a cure could be possible with continued treatment. Studies with a larger population of presymptomatic individuals with pathogenic variants in CLN2 are needed to truly establish a cure, but this single case report is most encouraging.
Pilot programs testing if presymptomatic SMN1 newborns can be cured by intrathecal nusinersen support the studies on presymptomatic CLN2 children and FDA and EMA approval of AVXS-101 or Zolgensma for in vivo adeno-associated viral-mediated GRT for SMN1 patients represent groundbreaking milestones in the quest to cure epilepsy and neurodegenerative diseases (Jensen et al., 2021; Aragon-Gawinska et al., 2018; Bertini etal., 2017; Mendell et al., 2021; Chiriboga et al., 2016; Mercuri et al., 2018; Day et al., 2021).
About a year after Nusinersen was shown to be effective in type 1 SMA, AveXis pharma (recently acquired by Novartis Pharmaceuticals) with Mendell et al. (2017) introduced AVXS-101 or onasemnogene abeparvovec (brand name Zolgensma), a “one-and-done, single intravenous injection” of GRT (Jensen et al., 2021). AVXS-101 is contained in a self-complementary AAV9 that crosses the BBB when administered before 2 years of age. AVXS 101 also has a constitutively active promoter that provides persistent expression of SMN1 protein. Three open-label, single-arm, single-dose phase 3 trials were conducted in the United States, Europe, and Asia after a one-time intravenous infusion of AVXS-101 in 22 type 1 SMA infants younger than 6 months. All three clinical trials established safety and efficacy (Mendell et al., 2017, 2021; Al-Zaidy al., 2019a,b; Lowes et al., 2019; Finkel et al., 2019, 2020). End-of-study analysis showed AVXS-101 improved permanent ventilation-free survival and improved motor function and milestones in infants with SMA type 1. At 14 months, 91% survived free from permanent ventilation support. At 18 months, 59% were able to sit independently.
Within 2 years of introduction, a one-time intravenous dosing of AVXS-101 in three presymptomatic infants (age up to 42 days) was started, realizing that a “one-and-done, single intravenous injection” of AVXS-101 will cross the BBB in infants. Results are not yet available. Newborn screening for SMN1 variants has also started in some countries in anticipation of intravenous dosing with AVXS-101 (Jensen et al., 2021).
Eventually, pilot programs for presymptomatic newborns with three copies of SMN2 began in Taiwan, New York, Belgium, and France. Twenty patients have been enrolled (Bertini et al., 2017). Twelve infants have reported typical normal development so far and have been presented in international meetings (Gidaro and Servais, 2019). Long-term efficacy and normality until adulthood remain to be established.
Contrasting “one single lifetime” intrathecal injection with enduring results versus every 2 weeks intraventricular GRT led to ongoing clinical trials testing “one single lifetime” intrathecal injections of CLN2, CLN3, CLN6, and CLN7 ceroid lipofucsinoses.
With “every-2-weeks” intraventricular cerliponase alfa GRT considered a promising treatment to cure CLN2 disease, a “single-injection treatment strategy of GRT” was developed for CLN2 and other genotypes of CLN with the ultimate goal of achieving a long-term, enduring, if not lifelong cure. The single-treatment strategy of GRT has been made possible by the newer generation of AAV gene therapy vector, self-complementary AAV (scAAV). scAAV vectors deliver long-term expression of the packaged gene which has been demonstrated to be present for as long as 10 years in humans and 15 years in nonhuman primates (Sehara et al., 2017; Chu et al., 2020; Jiang et al., 2006). AAV vectors induce expression in both dividing and nondividing cells and remain within the cell nucleus in episomal form (Salganik et al., 2015). scAAV has a desirable safety profile with less vector-associated toxicity, lack of pathogenicity, and the least immunogenic in humans. scAAV has strong neuronal tropism, is effective in transduction, and is unable to self-replicate. Self-complementary AAV carries a genome that resembles a double-stranded DNA template ready for transcription upon infection. Improved transgene expression cassette designs, and the use of split-vector systems have overcome its small genome size (4.7 kb) packaging.
After two phase1/2 trials using AAVrh10-CNL2 were completed in a mice model and safety profiles in rats and nonhuman primates deemed acceptable, clinical trials using “single injection treatment strategy of intrathecal GRT with the ultimate goal of achieving a long-term, if not life-long cure” soon followed (Sondhi et al., 2007).
Clinical trials.gov now lists such trials in CLN2, CLN3, CLN6, and CLN7 disease testing single injection of intrathecal GRT. Clinical trials.gov lists “Gene Therapy for Children with CLN3 Batten Disease” sponsored by Amicus Therapeutics (NCTO3770572). The human trial is described as an open-label, single-dose escalation clinical trial to evaluate safety and efficacy of AT-GTX-502 (AAV9 .P546.-CLN3) delivered intrathecally in the lumbar spinal cord region of subjects with CLN3 Batten disease. The “one-time intrathecal injection” has seven participants and completion date is estimated as September 2023.
CLN3 in chromosome 16p12.1 encodes a transmembrane 438-amino acid hyrophobic protein (Battenin), which is trafficked to and expressed on lysosomal membranes. Pathogenic variants in CLN3 lead to accumulation and buildup of autofluorescent lipofuscin-like-deposits with lysosomes of affected neurons. CLN3 juvenile CLN is the most prevalent of CLN and presents with rapidly progressive retinal degeneration between 4 and 7 years. In the ensuing decade, seizures, language problems, and ataxia lead ultimately to death during the second or third decades of life.
Results of the open-label, single intrathecal injection of low-dose (6 ×1013) vector genomes and high-dose (1.2 × 1014) vector genomes of AT-GTX-502 were presented by Amicus Therapeutics as a poster in the 17th Annual World Symposium, held online February 8–12, 2021. Early safety and efficacy results included four children; three received low dose and one at high dose. Analyses indicated AT-GTX-502 was safe and well tolerated. Four reported side effects consisted of elevated levels of liver enzymes that resolved.
Intraocular gene therapy with AAV-meditated expression of CLN3 in a mice model of CLN3 eye disease prevented decline in inner retinal function resulting from death of bipolar disease (Klein Holthaus et al., 2020).
Clinical trials.gov now also lists as ongoing “Gene Therapy for Children with CLN6 Batten Disease” sponsored by Amicus Therapeutics (NCTO2725580) and a 15-year follow-up study to evaluate long-term safety and efficacy (NCT04273243). The human trials are described as an open-label, single-dose escalation clinical trial to evaluate safety and efficacy of AT-GTX-501 (AAV9-CNL6) delivered intrathecally in the lumbar spinal cord region of subjects with CLN6 Batten disease. The “one-time intrathecal injection” has 13 human participants with variant late infantile/early juvenile NCL. Preliminary results were presented at the joint meetings of the 16th Child Neurology Congress and 49th Annual Child Neurology Society. Interim analyses for efficacy showed stabilization of motor and language function in treated individuals versus individuals in the natural history cohort (see (see Section 7, Chapter by Kayani et al. and Chapter by Sarah Mole, 2022).).
CLN6 disease is caused by one of the 70 well-characterized pathogenic variants in a 311-amino acid protein with seven predicted transmembrane domains, predominantly localized in the endoplasmic reticulum. Most pathogenic variants lead to a complete loss of CNL6 protein or a truncated nonfunctional CNL6 protein. Two clinical phenotypes result: (1) variant late infantile/early juvenile NCL, which is the relatively more common syndrome, and (2) adult type A Kufs disease. In the variant late infantile syndrome, the disease usually starts at 16 months to 6 years with impaired language and delayed motor and cognitive development in early childhood. Visual, motor, and mental deterioration lead to death at 12 to 15 years of age.
Cain et al. (2019) dosed three 4-year-old cynomolgus macaques with intrathecal sc AAV serotype 9.chicken B-actin.CLN6 and followed these macaques for 6 months. High levels of transgene expression were found in brain and spinal cord of all animals. No adverse effects or pathology were observed. Cain et al. (2019) also delivered scAAV9.CB.CLN6 into cerebrospinal fluid by a single postnatal intracerebroventricular injection into the Cln6nclf mouse model. The single injection prevented accumulation of autofluorescent storage material, including ATP synthase subunit C, prevented reactive gliosis, and loss of dendritic spines. It also prevented motor, memory, and learning deficits and improved survival of the Cln6nclf mouse. These results strongly supported the initiation of the clinical trial (NCTO2725580) for the single-dose AT-GTX-501 (AAV9-CNL6) for the variant late infantile/early juvenile NCL syndrome.
Clinical trials.gov now also lists NCTO4737460 as enrolling and ongoing “Gene Therapy for Children with CLN7 Batten Disease.” The human trial is described as an open-label, single-dose clinical trial to evaluate safety and efficacy of intrathecal AAV9-CNL7. The “one-time intrathecal injection” is in the enrollment stage.
CLN7 disease is caused by pathogenic variants in the major facilitator superfamily of transporter genes with domain containing 8 (MSFD8), which encodes a lysosomal membrane protein of unknown function. Two common mutations are known that causes disease at ages 2 to 7 years of age. Epilepsy, developmental delay, and regression with myoclonus, ataxia, loss of vision, and speech impairment worsen until death at 11.5 years of age.
CLN7 disease recently received attention and merited publication in the New England Journal of Medicine as an example of true personalized medicine, a novel concept known as “N of 1 treatment” where a specific ASO was developed to correct mis-splicing of the MSFD8 gene. The patient was compound heterozygous for an insertion of an SVA retrotransposon and c.1102G>C.RNA sequencing and reverse-transcriptase-PCR revealed missplicing of exon 6 into a cryptic splice-acceptor site in MSFD8 intron 6, about 119 bp upstream from the SVA insertion site. This missplicing was predicted to lead to premature translation termination. A 22-nucleotide antisense oligonucleotide was designed to target the intron 6 SA cryptic splice-acceptor site and nearby splicing enhancers (dubbed as Milasen). RNA sequence from the patient’s fibroblasts showed that Milasen more than tripled the amount of normal (exon6-exon7) splicing. Clinically, after intrathecal dose escalation followed by maintenance doses of 42 mg every 3 months for 1 year, seizures were substantially decreased and several neurologic and neuropsychological subscores stabilized after 7 months of treatment (Kim et al., 2019).
What about other epilepsies? Is a cure coming for them also in the next 10 years? Recently, the FDA proclaimed that they expect to be approving 10–20 cell and gene therapies a year from 2025 on (Jensen et al., 2021).
Ex vivo hematopoietic stem and progenitor cell (HSPC) transduction gene therapy has been central in the treatment of lysosomal storage disorders that include, aside from CNS symptomatology, widespread systemic and major organ signs of disease such as metachromatic leukodystrophy (MLD) and mucopolysaccharidoses (MPS).
Epilepsy specialists encounter MLD as a rare cause of seizures in late infancy, early childhood, and even more rarely as seizures in adults. With cerebral demyelination and neuronal death in central and peripheral nervous systems caused by a loss-of-function mutation in arylsulfatase A (ARSA), muscle weakness and wasting plus dementia and seizures characterize MLD. To correct the loss of function of ARSA, three presymptomatic children had their stem cells isolated and transduced ex vivo with a lentivirus-vector carrying the ARSA gene. The autologous CD34-positive HSPCs (OTL-200 from Orel Therapeutics and San Raffaele-Telethon Institute for Gene Therapy) were then reinfused and transplanted into the bone marrow of three children (Biffi et al., 2013).
Progression of MLD was arrested in the three children with stable engraftment of transduced HSPCs; ARSA activity was reconstituted in CSF (Sessa et al., 2016).
As a result of these encouraging results in treatment during the presymptomatic state, 33 other children symptomatic with early-onset MLD were treated with OTL-200. With up to 7.5 years of follow-up, long-term stabilization of motor functions was reported during the 46th Annual Meeting of the European Society for Blood and Marrow Transplantation. These preliminary results were encouraging, together with stable engraftment of OTL-200 in bone marrow and restoration of ARSA activity in cerebrospinal fluid (Calbi et al., 2020; Fumagalli et al., 2022). In a short time, namely, the same year of 2020, the EMA approved OTL-200 (trade name Libmeldy) for treatment of MLD.
Ex vivo transduced CD34-positive HSPCs with a Lenti virus containing N-sulfoglucosamine sulfohydrolase (SGSH, also known as heparan-sulfatase) has been administered to patients with MPS III or Sanfilippo syndrome A (NCT04201405). Sanfilippo syndrome A is the most common and the most severe of MPS with the lowest survival rate. With the study still ongoing, no results are available.
Mucopolysaccharidosis I (MPS I), an autosomal recessive lysosomal storage disorder, also called Hurler MPS I, is one of the severe forms of MPS I. It is caused by alpha-L-iduronidase enzyme deficiency (IDUA); IDUA is located in chromosome 4 and catalyzes degradation of glycosaminoglycans (GAGs). The enzyme deficiency results in long chains of sugar carbohydrates (GAGs heparan and dermatan sulfates) accumulating in cartilage, tendons, cornea, skin, and connective tissues. Infants develop bone deformities months after birth. Children remain behind in development, with many bodily system disorders such as mental retardation, speech impediments, short stature, stiff joints, heart, and respiratory systems affected. Patients usually succumb to heart failure or pneumonia.
Genome editing using zinc finger nucleases (ZFN) platform studies where AAV8 virus vectors were injected intravenously to genome edit living cells described encouraging findings in rodents (Ou et al., 2020). ZFN binds to triplet DNA sequences and opens up DNA strands with its intrinsic nuclease activity. At the present time, ZFN genome editing uses an intravenous-injected AAV6 vector that inserts a corrective copy of the IDUA (alpha-L- iduronidase) into the patient’s hepatocytes is in clinical trials (NCT02702115). Permanent liver-specific expression of IDUA is expected, and it is hoped that high blood levels of IDUA will cause passage through the BBB into brain. A 10-year follow-up for safety studies is also ongoing (NCTO4628871). Human treatment induced immunogenicity and “on-target” and “off-target” complications, including oncogenesis, is of particular importance in recent attempts at genome editing (Shim et al., 2017; Mills et al., 2006). Ex vivo and in vivo genome editing in MPS I and MPS II reached clinical trials testing, and concerns arose about immunogenic response to the bacterial protein in the Cas9 based on in vivo genome editing (Wang et al., 2015, 2020; Merkel et al., 2015; Poletto et al., 2020) and off-target effects with CRISPR/Cas9 in vivo editing (Lubroth et al., 2021).
The critical genetic and clinical ingredients necessary for a quest to cure the ultra-rare Lafora disease (LD) or Lafora progressive myoclonus epilepsy (Lafora PME) ( de Haan et al., 2004 ) have been available. Two studies of the clinical natural history: (a) an international/global study of a 45-patient cohort led by UCLA and (b) a 32-patient cohort studied in Madrid, Bologna, Dallas, and Los Angeles. Genotyped LD patient cohorts that are essential to the quest for a cure in LD are now waiting for the clinical trials to start. These advances have been made possible by the Lafora Epilepsy Cure Initiative (LECI), a National Institutes of Health (NIH)-funded program project.
Initially described, in 1911, by Gonzalo Lafora (Lafora & Glueck, 1911), scientific breakthroughs were made in 1998–2003, with the identification of EPM2A and EPM2b as the genetic basis for LD (Minassian et al., 1998; Serratosa et al., 1999; Chan et al., 2003). PTG depletion or downregulation of GS1 removes Lafora bodies and rescues the fatal epilepsy of Lafora disease (Turnbull et al., 2011; Tagliabracci et al., 2008). Using epm2a and epm2b KO models:
Gys1-ASO is an antisense oligonucleotide that targets the mRNA of the brain-expressed glycogen synthase 1 as synthesized by Ionis pharma (Ahonen et al., 2021; see Chapter 56, this volume). The preclinical trial, testing efficacy, and adverse effects of “every-3-months” intrathecal Gys1-ASO has been completed in epm2a and epm2b knockout mouse models. Gys1-ASO prevented Lafora body formation in young mice that had not yet formed them and inhibited further LB accumulation in older mice. Astrogliosis was prevented, and markers of neuroinflammation were reduced. Testing for safety and adverse effects has also been completed in nonhuman simians. The study design for a human clinical trial testing “every-3-months” intrathecal Gys1-ASO in presymptomatic individuals and LD patients with “seizures only” is projected for late 2022 to early 2023.
Canine safety testing pancreatic alpha-amylase bound to a cell penetrating monoclonal antibody (VAL-0417), an antibody-enzyme fusion delivered by intracerebral/intraventricular injection, has completed preclinical tests in epm2a KO models and progressed to canine safety testing (see Chapter, this volume; Austin et al., 2019, Brewer et al., 2019, Sun et al., 2021). The pharma company that has developed this therapeutic agent has obtained an IND number from the FDA and plans to proceed with the novel concept known as “N of 1 treatment.” The pancreatic alpha-amylase bound to a cell penetrating monoclonal antibody (VAL-0417) was developed to digest and degrade the Lafora inclusion bodies.
A single, once-in-a-lifetime intracerebral/intraventricular dose, a GYS1-targeting/ downregulation by microRNA packaged in an “scAAV9” vehicle, has also completed preclinical trials in epm2a and epm2b KO models. Safety concerns, adverse effects, and the effects of varying doses will be tested in a nonhuman simian prior to a clinical trial in humans (see Section VII by Kayani et al., 2022 Chapter 52, this volume).
A single, once-in-a-lifetime dose by intrathecal injection of a wild-type gene to replace a mutated gene or GRT is also being developed for LD. Two scAAV9 constructs, one containing human EPM2A and the other human EPM2B gene, are undergoing preclinical trials in epm2a and epm2b knockout models as a “single, one-time intrathecal injection” in Dallas, Madrid, and Lexington (Kentucky).
Oral small molecules or preclinical small molecules downregulating GYS1 conceived as a potential oral drug against LD were developed by Drs. Peter Roach and Anna DePauli-Roach, Tom Hurley, and David Watt (Tang et al., 2020). This preclinical success was built upon by MAZE pharma, which developed a small molecule that is administered as an oral pill. This small-molecule downregulating GYS1 is being tested now in a clinical trial in Pompe disease where clinical experience will be invaluable for LD clinical trials.
A novel study design that maximizes likelihood of demonstrating proof for halting progression, reversal of symptoms, and a cure will consider the five clinical and EEG stages of LD. During stage zero when the individual is still clinically asymptomatic, 3.5 Hz-multispike-slow waves appear in fragments and sleep architecture is normal. During Stage I “Visual Seizures Only” (6–14 years) and Stage II “Beginning Mild Cognitive Decline” (12–14 years), overnight sleep video-EEGs record loss of REMs. The 3.5 Hz-multispike-slow wave index is 0.33% to 4% in awake, and 2% to 9% in N1 and N2 sleep stages, but is paradoxically suppressed in stage N3 delta sleep. During Stage III “Established Dementia and Status Epilepticus” (15 years), overnight sleep video-EEG show not only loss of REMs, but loss of EEG desynchrony and decrease to loss of N3 delta sleep stage. The 3.5 Hz-multispike-slow wave index rises to 3% to 30% to 80% as absence status and obtundation status dominate clinically. During Stage IV “Myoclonic Encephalopathy” (16–20 years) with no sleep spindles and vertex waves, EEG cannot differentiate wakefulness from sleep. High-voltage 16 Hz polyspikes index dominate with indices rising from 2.6% to 80% as myoclonic status and myoclonic-tonic-clonic status recur. During Stage V, progressive neurologic deterioration (>21 years) leads to exitus.
An open-label study design should treat the EPM2A and EPM2B clinically presymptomatic person and the LD patient when seizures are the only or main phenotype and prevent increases in clinical seizures, EEG 3.5 Hz-multispike-slow wave indices, and EEG high-voltage 16 Hz polyspikes indices and prevent cognitive decline, loss of REMs, and suppression of polyspike-slow waves during N3 delta sleep.
With more clinical trials in NCL disease assessing efficacy of the “single, one-time, intrathecal injection GRT,” research units addressing Lafora disease also tilted in favor of GRT, more recently. However, this was not always the case at first, when the excitement involving nusinersen (Spinraza) for spinal muscular atrophy (SMA) encouraged a similar program for ASOs in Lafora PME. Hence, ASOs were the first therapeutic agents tested in epm2a–/– and epm2b–/– mice.
In Lafora disease, pathogenic variants in EPM2A (glycogen dual specificity phosphatse) and EPM2B (E3 ubiquitin ligase) form longer-than-normal glucan chains and less branch points rendering the glycogen molecule insoluble and forming what has been defined as PAS+ Lafora inclusion bodies. ASO downregulation of GYS1 transcripts halts the formation of these PAS+ inclusion bodies, neuroinflammation, and progression of disease in epm2a–/– and epm2b–/– mice. An ASO will likely be the first therapeutic agent tested in a clinical trial for Lafora disease, projected for 2022/2023.
Because intrathecal ASOs have to be delivered every 3 to 4 months and probably require lifelong treatment, a “one-time” intrathecal injection of a microRNA which targets and downregulates GYS1 and packaged in an “AAV9 vehicle” was developed and is ongoing as a preclinical trial in epm2a–/– and epm2b–/– mice (Gumusgoz et al., 2022).
“Which one should my child take?” The success of AVXS-101 or onasemnogene abeparvovec (brand name Zolgensma), a “one-and-done, single intravenous injection” of GRT in type 1 SMA, in three open-label, single-arm, single-dose phase 3 trials in 22 type 1 SMA infants younger than 6 months (Mendell et al., 2017, 2021; Al-Zaidy al., 2019a,b; Lowes et al., 2019) prompted the next study in presymptomatic individuals. Within 2 years of introduction, a one-time intravenous dosing of AVXS-101 in three presymptomatic infants (age up to 42 days) was started, realizing that a “one-and-done, single intravenous injection” of AVXS-101 will cross the BBB in infants. Results are not yet available. Newborn screening for SMN1 variants has also started in some countries in anticipation of intravenous dosing with AVXS-101.
Now, parents of patients with Lafora type progressive myoclonic epilepsies and parents of infants with Wednig Hoffman type 1 SMA ask, “Which treatment should we choose—ASO or GRT? Can I start with ASO and then also get GRT for my child? Or vice-versa?”
In 2019, Dabbous et al. (2019) compared results of AVXS-101 one-time intravenous infusion treatment (clinical trial NCT02122952) with “every-3-to-4-months” intrathecal injection of nusinersen (clinical trial NCT02193074) using “frequentist and Bayesian” approaches. Results suggest that AVXS-101 has an efficacy advantage over nusinersen concerning overall survival, independence from permanent assisted ventilation, motor functions, and milestones.
To evaluate the cost-benefit of AVXS-101 versus nusinersen, results from the clinical trials were also used to create a model that assessed probabilities of patient dying and quality of life. Probabilities showed that 50% of SMA infants treated with AVXS-101 will survive until the age of 3 years. SMA infants treated with nusinersen, only 50% survival to the age of 3 years. Quality-of-life scores also favored AVXS-101 at 15.65 versus 5.29 for nusinersen.
Perhaps these results, comparing these nusinersen versus AVXS-101 clinical trials, convinced the research teams on Lafora disease in Madrid and Lexington (Kentucky) to conduct their preclinical trials testing of intrathecal GRT consisting of EPM2A or EPM2B contained in scAAV9 in epm2a–/– and epm2b–/– mice.
The same five key genetic and clinical ingredients would need to be present for a quest to cure de novo loss-of-function, X-linked, and gain-of-function variants in the classic DEEs of infantile spasms, Lennox-Gastaut syndrome, and Dravet syndrome. Epilepsy variants have been defined in some DEEs, but the molecular mechanisms of many of these epilepsy variants still need to be fully defined. In 2013, Epi4K and EPGP investigators searched for de novo mutations in two “classic” forms of EE: infantile spasms (IS) or West syndrome and Lennox-Gastaut syndrome (LGS). Exomes of 264 trios were sequenced and 439 putative de novo mutations were identified. Sanger sequencing confirmed 329 de novo mutations. In these 264 trios, nine genes with de novo SNV mutations were present in two or more probands (SCN1A, n = 7; STXBP1, n = 5; GABRB3, n = 4; CDKL5, n = 3; SCN8A, n = 2; SCN2A; n = 2; ALG13, n = 2; DNM1, n = 2; and HDAC4, n = 2). Among these nine genes, SCN1A, STXBP1, SCN8A, SCN2A, and CDKL5 have previously been associated with EE (Allen et al., 2013).
Investigators claim that they had implicated, for the first time, “GABRB as a single gene cause of EE and provide the strongest evidence yet available for any epilepsy” based on (1) four unique de novo mutations in GABRB3, (2) the absence of similar de novo mutations in 610 control exomes, and (3) a computational modeling of these genes associated with EE having low tolerance for protein-disrupting mutations and producing adverse phenotypic consequences. They used the same reasoning to implicate ALG13, an X-linked gene encoding a subunit of the uridine diphosphate-N-acetylglucosamine transferase, which had been reported in two earlier studies of severe intellectual disability and seizures (Allen et al., 2013).
The next year in 2014, the Epi4K consortium combined their 264 trios with 356 trios of the EuroEPINOMICS-RES consorium and analyzed their exome sequencing results. All trios had a diagnosis of epileptic spasms and Lennox Gastaut syndrome. With the expanded cohorts, they found 429 de novo mutations, including de novo variants in DNM1 in five persons and de novo variants in GABBR2, FASN, and RYR3 in two persons. The combined cohort provided a sigificant likelihood analysis (p = 8.2 × 104) (EuroEPINOMICS-RES, Epilepsy Phenome/Genome Project and Epi4K Consortium, 2014).
Hamdan et al. (2017) combined three strategies during a search for de novo mutations in persons with DEE: (1) they identified candidate variants with de novo mutations during whole genome sequencing of 197 individuals with both epilepsy and intellectual disability, (2) they identified additional individuals with the same candidate variants in another series, and (3) they mined several datasets of developmental and epileptic encephalopathy and intellectual disability. By combining these strategies, these authors suggested a causal link between DEE and NTRK2, GABRB2, CLTC, DHDDS, NUS1, RAB11A, GABBR2, and SNAP25.
Epileptic seizure types, a varying degree of intellectual disability and developmental delay, sometimes with dysmorphic features or nystagmus, have also been associated with variants in GABRA3 encoding the α3-subunit of the GABAA receptor on chromosome Xq28 (Niturad et al., 2017).
Aristaless-Related homeobox X-linked gene (ARX), which results in defective GABAergic interneuronal migration, has also been observed in infants with West syndrome epileptic encephalopathy (Katsarou et al., 2017).
More recently, an autosomal-recessive DEE was associated with biallelic variants in PTPN23, which encodes a tyrosine phosphatase (Sowada et al, 2017). Myers et al. (2017) identified six individuals with DEE harboring a de novo mutation in PPP3Ca, which encodes the alpha isoform of calcineurin (a calcium and calmodulin-dependent serine/threonine phosphatase).
A cohort of 23 patients with epileptic encephalopathy carrying loss- or gain-of-function KCNA2 mutations was reported by Masnada et al. (2017). The more severe epilepsy, with developmental problems and ataxia, and atrophy of the cerebellum or even the whole brain, was observed in about half of the patients with gain-of-function mutations.
Individuals with pathogenic DNM1 variants suffer from infantile spasms and Lennox-Gastaut syndrome (EuroEPINOMICS-RES Consortium, Epilepsy Phrnomr/Genome Project and Epi4K Consortium, 2014). At leasy 20 heterozygous de novo variants have been reported in persons with drug-resistant seizures in infancy, profound intellectual disability, hypotonia of muscles, lack of speech, dystonia, and spasticity (von Spiczak et al., 2017; Li et al., 2019). DNM1 encodes a multimeric brain-specific GTPase, dynamin-1, which is located in the presynapse where it mediated endocytosis (Dhindsa et al., 2015). Aimiuwu et al. (2020) tested a Dnm1-targeted therapeutic microRNA contained in a self-complementaryAAV vector. The thrapeutic package is injected once into the cerebral-ventricular system of both cerebral hemispheres of the Dnm1Ftfl/Ftfl homozygotes at day 1 or at birth. The one-time injection reduced seizure severity, extended lifespan, improved growth, and abolished various developmental impairments. miDnm1, a type of treated mice, had significant reduction of Dnm1a mRNA 2 weeks postdelivery and did not show fibrillary gliosis and cell degeneration in CA1 of hippocampus. This preclinical study provided proof of principle that postnatal gene silencing by a microRNA targetted to Dmn1a can curb the fundamental symptomatology of DNM1 epileptic encephalopathy (Aimiuwu et al., 2020).
Meng et al. (2015) targeted the UBE3A-ATS with ASO treatment which led to RNase–H-mediated cleavage of the AS-RAN heteroduplex, thereby unsilencing the paternal UBE3A gene. Silva-Santos et al. (2015) and Sonzogni et al. (2020) then showed that unsilencing the paternal UBE3A gene around birth works best at improved behavioral phenotypes of Angelman syndrome. Two clinical trials, one by Hoffman-La Roche (NCTO442881) and the second by GeneTxBiotherapeutics (NCTO4259281), are currently active using the principle of unsilencing the paternal UBE3A gene (Elgersma, 2021).
The working hypothesis now is that febrile, afebrile generalized, unilateral, clonic and tonic clonic seizures in the first year of life, developmental delay evident in the second year of life, progressing cognitive decline, later followed by repeated bouts of status epilepticus, behavioral disorder, ataxia, and motor impairment in Dravet syndrome (Dravet, 1965, 1978, 1992, 2005a, 2005b; Delogu et al., 2011) are caused by loss-of-function variants in one allele of the SCN1A gene. This produces nonfunctional NaV1.1 units located in GaBAergic neurons, leading to epileptic seizures, the symptomatology of an encephalopathy, and risk of sudden death. ASOs are being explored as treatment in Dravet syndrome (Hill and Meisler, 2021; Han et al., 2020). Dravet syndrome, X-linked and found only in females is due to variants in protocadherin (Depienne et al., 2009).
Valassina et al. (2022) generated an Scn1a conditional knock-in mouse model (Scn1astop/+), in which Scn1a can be reactivated on demand during the mouse lifetime. Here, Scn1a gene was disrupted by inserting a STOP cassette arrayed between loxP sites in the intron between exons 6 and 7. The STOP cassette was conditionally removed upon Cre recombinase expression to reconstitute the functional Scn1a allele. Valessina et al. (2022) show that Scn1a reactivation when symptoms were already manifest at P30 or at P90 led to complete rescue of both spontaneous and thermic induced seizures, marked amelioration of behavioraal abnormalities, and normalization of hippocampal fast-spiking interneuron firing.
Colasante et al. (2020) developed nuclease-dead Cas9 (dCas9) fused to effector domains such as the V16 transcriptor activator domians for transcription gene regulation. This platform manipulates transcription without cleaving the target. They then delivered the Scn1a-dCas9 activation system to Dravet syndrome pups using an adeno-associated virus vector. In the package they also included the Dlx5/6 gene activation system that exclusively tartgets forebrain interneurons. With this, Scn1a gene activation in cortical interneurons reduced and even totally suppressed temperature-induced epileptic seizures. A similar gene-editing approach is suggested for human Dravet syndrome.
An ASO (STK-001) was developed to specifically target the SCN1A NMD or “poison exons” 20N and reduces nonproductive mRNA while increasing levels of productive mRNA and Nav1.1 sodium channel protein. (Hill & Meisler, 2021)The upregulated expression of wild-type SCN1A should compensatesfor the haploinsufficiency of SCN1A in Dravet syndrome. Preclinical studies had shown that a single intracerebral-ventricular injection at day 2 in the Scn1aTmKea heterozygous mice increased Scn1a mRNA and Nav1.1 protein expression, prevented SUDEP in 97% of DS mice, and suppressed seizure frequency and latency. (See Chapter 75, this volume.)
Early reports from an interim analysis hint that the intrathecal STK-001 medication may reduce seizure frequency among trial participants. As of November 23, 2021, it has so far raised no safety concerns among children taking part in the ongoing phase ½ MONARCH clinical trial. There is also an open-label extension study (NCT04740476), which will enroll 69 participants with an estimated study completion date of March 3, 2027.
Among the 40 to 60 million persons worldwide with epilepsies, focal epilepsies account for the greatest proportion. Even with optimal treatment with antiseizure medications, approximately 30% remain uncontrolled. Consequently, new treatment paradigms are being developed.
Snowball et al. (2019) codon-optimized the voltage-gated potassium channel Kv1.1 for human expression and mutated to accelerate recovery of the channel from inactivation. The engineered potassium channel gene was packaged in a nonintegratng lentiviral vector under the control of a cell-specific CAMK2A promoter. The potassium channel-lentiviral vector robustly reduced seizure frequency in a rat model of focal neocortical epilepsy during a blinded, randomized, placebo-controlled preclinical trial.
When packaged into an adeno-associated viral vector (AAV2/9), the engineered potassium channel gene was also effective at suppressing seizures in a male rat model of temporal lobe epilepsy. The authors conclude that the engineered potassium channel gene therapy was ready for clinical translation (Snowball et al., 2019).
Szczygiel et al. (2020) used a vector that encodes the combination of neuropeptide Y and its antiepileptic receptor Y2 and administered the package unilaterally into the dorsal ventral hippocampus of adult male rats. Previous studies had shown that this package has antiepiletic effects in in vivo rodents. Combined overexpression of NPY and Y2 in the hippocampus had exerted a superior seizure-suppressant effect compared to single transgene expression (Cattaneo et al., 2020; Woldbye et al., 2005, 2010). In the present study, the expression of both transgenes was found to be elevated and remain without a decline at 6 months post-injection. The package mediated expression selectively in hippocampal neurons without expression in astrocytes and oligodendrocytes. No effects were observed on body weight, or on short-term or long-term memory as tested by the Y-maze SA test or the Morris water-maze test. Authors conclude that the combination of neuropeptide Y and Y2 receptor gene therapy is ready for future gene therapy in pharmaco-resistant temporal lobe epilepsy patients (Szczygiel et al., 2020). CombiGene in Lund, Sweden, reports their intention to start a clinical trial and treat focal epilepsies delivering NPY and Y2 receptors through an AAV approach (Binder et al., 2020).
Recently, the US FDA proclaimed that they expect to be approving 10 to 20 cell and gene therapies a year from 2025 onward (Jensen et al., 2021). It has been 32 years (September 14, 1990) since Anderson, Blaese, and Culver conducted the first clinical trial of human gene therapy at NIH on a 4-year-old girl with severe combined immune deficiency (SCID). The girl, reportedly, continues to do well, even after three decades. However, 9 years after that first gene therapy, the “Gelsinger case” occurred (Wilson, 2009). Jesse Gelsinger from Arizona suffered from ornithine transcarbamylase deficiency. He was treated with GRT that used a replication defective human adenovirus type 5 vector. The gene therapy package was injected into the Gelsinger patient’s hepatic artery, which then found its led to the liver. Four days later, Jesse Gelsinger died from multiple organ failures triggered by an immunogenic reaction from the adenovirus vector (Raper et al, 2003; Lehrman, 1999; Verma, 2000).
The research community has not forgotten this early decade (1990 to 2000) of the “stop and go” history of gene therapy, the stop periods causing careers of scientists to be derailed by NIH and FDA grant suspensions or stopped altogether (e.g., an institute director at Penn confined to animal research and no human studies) and even costing medical school deanships. Of course, this was complicated by the legal community’s exploitation of these unfortunate and unforeseen complications of gene therapy. In the research community, only the brave of heart have “soldiered on” in spite of the “harsh” lessons on adverse effects of gene therapy learned during its first decade, during the turn of the century (Thompson, 2000).
Aside from the notable unfortunate story of the 1999 Gelsinger case from Arizona, there was the 2002 LMO2 insertional mutagenesis causing “T-cell acute lymphoblastic leukemia case” from Paris. Eleven children in the Necker hospital in Paris were enrolled for GRT to treat X-linked severe combined immunodeficienct (X-SCID). A gamma-retroviral vector was used to package the wild-type gene. Among these 11 children, two of the youngest boys developed leukemia. The retroviral vector inserted the retrovirus genome close to the promoter of the LMO2 gene in the two boys’ genome. The promoter of LMO2 gene encodes a transcription factor, which caused the continous expression of LMO2 protein, which, in turn, led to the development of T-cell acute lymphoblastic leukemia (Hacein-Bey-Abina et al., 2003).
These GRT complications emphasized the short-term and long-term complications of GRT, genome integration issues, and concerns about long-term sustained safety. These adverse effects overshadowed the successful results of GRT in two patients with SCID in the United States (Thompson, 2000) and four patients in the United Kingdom (Gaspar et al, 2004), in patients with Hemophilia B (Factor IX deficiency) and in patients with Canavan disease (aspartoacylase deficiency) (Jiang et al., 2006; McPhee et al., 2006). Despite these setbacks, there has been continuing high hopes for GRT over the last decade because of safer AAV vectors, as discussed above, in NCL diseases. Vigilance remains for unexpected complications such as thrombotic microangiopathy with AVXS-101 and intrinsic uncertainties of GRT. Genome integration issues must be better understood. As an example, wild-type AAV integrates into a specific site in human chromosome 19 (Kotin et al., 1990, 1991). Unexpected sudden deaths from liver failure have been reported in clinical trials using systemically administered AAV vectors for X-linked myotubular myopathy where hepatic disease had not been previously recognized (Shieh et al., 2020), and hepatic dysfunctions have been observed even in intracerebral delivery of AAV (Merkel et al., 2015). The genome of AAV vectors when inserted into dog genes involved in cell growth may potentially lead to malignancies (Nguyen et al., 2021). These short-term and long-term complications of GRT, genome integration issues, the insertion, deletion, or replacement of nucleotides during genome editing, the immunogenicity of their genome editing tools, on- and off–target modifications, and concerns about long-term sustained safety should be part of the discussions during the consenting process of parents and affected persons (Gaspar et al., 2004; Jensen et al., 2021; Wang et al., 2007; Wang & Gao, 2014).
The ascendance of therapeutic agents that can halt the progression of disease, reverse symptoms, and cure what was previously considered incurable fatal epilepsies raise a variety of issues on equity, ethics, and the consenting process with regard to the treating physician, the pharmaceutical industry, and the patients and their families (Krauss et al., 2000; Barns et al., 2000; Temkin, 1971; Jacoby, 2002, 2004; Shostak and Ottman, 2006; Ngugi et al., 2010; West, 1979).
The expensive and novel therapeutic agents are a first concern (Gehlert et al., 2000). For example, will antisense oligonucleotides (e.g., intrathecal STK-001; IONIS intrathecal Gys1-ASO), GRT (e.g., intracerebral-ventricular cerliponase alfa [Brineura]; Amicus single lifetime intrathecal AT-GTX-502), and special molecules such as the pancreatic alpha-amylase bound to a cell-penetrating monoclonal antibody (VAL-0417) be available to patients with rare fatal epilepsies who are located around the world, including countries with low socioeconomic status? This is an important first concern.
To make sense of the benefits and risks of the therapy being offered, parents will often rely on the treating doctor for advice. When making the final decision, parents often ask, “On- and off-targets . . . what is that?” “Do you mean I have to wait ten years to see if my child develops leukemia?” “What would you do if it were your child?” A significant addition to the burden of being diagnosed with a fatal form of epilepsy is the complexity of information on the therapeutic agent being offered as treatment. This can amplify the stress parents are undergoing, even though they feel relief that finally a curative agent has become available. Emerging information on adverse effects of the treatment offered is anxiety building enough, and the honest admission by the treating physician that unknown, previously unrecognized complications may be lurking somewhere during the long process of monitoring adverse effects (e.g., 10 years follow-up after a single lifetime intrathecal injection of GRT) can be too much to bear.
The child may carry the epilepsy-causing variant, but the full array of the therapy’s complications is still unknown. Because there is a lack of social scientific literature, guidelines, or policies from government agencies such as the National Human Genome Research Institute (NHGRI), societies representing medical genetics professionals (e.g., American Society of Human Genetics/American College of Medical Genetics), or the Recombinant DNA Advisory Committee (a forum for genetic engineering debates), which addresses the particular social dimensions that confront families, parents, and their affected child, the treating physician often turns to workshops on the subject of ethics in gene therapy and the experience of human ethics committees and institutional review boards (IRBs) of universities for discussion and advise.
Importantly, it is not a perfect science. It is difficult to determine whether parents and affected persons have a clear and total understanding of the burden and benefits of treatment. Can they understand what a seemingly elevated ammonia level without other laboratory signs of liver dysfunction means, and what signals of adverse effects in preclinical trials of nonhuman simians should contraindicate the human trial?
In the often-quoted Gelsinger case (Thompson, 2000; Verma, 2000), the FDA, in its investigation, claims that one contribution to Gelsinger’s death was “incomplete disclosure of the potential risks of Jesse’s treatment to his family. . . .Gelsinger’s high ammonia levels at the time of treatment should have excluded him from the study. . . . Two patients who had experienced serious side effects from gene therapy were not immediately reported and the deaths of monkeys given the similar treatment were never included in the informed consent discussion.” In those early years of gene therapy, it is difficult to blame the investigators when the antenna for possible complications signaled by the sole sign of an elevated ammonia level and the “dots/signs” of possible complications appearing during the preclinical trials and the “dots/signs” of possible complications during the human trial were not being connected. These were all done, of course, “post hoc” after liver organ failure caused the death of Gelsinger.
It is also difficult to decide what they, as parents, would consider acceptable results of treatment in the context of the neurologically debilitating and invariably fatal nature of their epilepsy syndrome being treated.
To incentivize biopharmaceutical companies to develop treatments for previously incurable rare diseases and neurodegenerative diseases such as Alzheimer disease, Parkinson disease, and amyotrophic lateral sclerosis, the FDA and EMA offer the prospect of fast-track designations and longer market exclusivity with their patents. For rare diseases and their small market, biopharmaceutical companies are forced to price their novel therapeutics with very high prices, many-fold higher than that offered for treatment modalities for common diseases. This necessarily limits the number of patients who can receive these novel therapeutics. Which brings us to the first concern discussed earlier, namely, equity (Jensen et al., 2021).
The COVID-19 pandemic has brought a new problem to the public and exacerbated the issue of equity. Some “start-up” pharmaceutical companies have gone bankrupt because of the pandemic. Even well-established large pharmaceutical companies have had to roll back or actually stop funds during preclinical experiments and clinical trials. The public asks, “Where should the responsibility lie?” Should it lie with the government agencies, like the NIH, or with the patient advocacy organizations and the nonprofit lay associations?
It will be important for the drug developers, pharma industry, investors, public agencies, and medical insurance companies, the advocacy organizations, and the families of patients with rare epilepsy diseaess and now the more frequent developmental epileptic encephalopathies to work together to mold new models of pricing and reimbursements to ensure that these novel curative therapeutics reach the patients in need (Jensen et al., 2021).
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...
