Debate: Does genetic information in humans help us treat patients?
PRO—genetic information in humans helps us treat patients
CON—genetic information does not help at all
Address correspondence to Antonio V. Delgado-Escueta, M.D., David Geffen School of Medicine at UCLA, Epilepsy Genetics/Genomics Laboratories, VA Greater Los Angeles Healthcare System, West Los Angeles, CA 90073, U.S.A. E-mail: email@example.com
In the past decade, genotyping has started to help the neurologic practitioner treat patients with three types of epilepsy causing mutations, namely (1) SCN1A, a sodium channel gene mutated in Dravet’s sporadic severe myoclonic epilepsy of infancy (SMEI and SMEB); (2) laforin (dual specificity protein phosphatase) and malin (ubiquitin E3 ligase) in Lafora progressive myoclonic epilepsy (PME); and (3) cystatin B in Unverricht-Lundborg type of PME. Laforin, malin, and cystatin B are non–ion channel gene mutations that cause PME. Genotyping ensures accurate diagnosis, helps treatment and genetic counseling, psychological and social help for patients and families, and directs families to organizations devoted to finding cures for specific epilepsy diseases. In SCN1A and cystatin B mutations, treatment with sodium channel blockers (phenytoin, carbamazepine, oxcarbazepine, lamotrigine) should be avoided. Because of early and correct diagnosis by genotyping of SCN1A mutations, the avoidance of sodium channel blockers, and aggressive treatment of prolonged convulsive status, there is hope that Dravet’s syndrome may not be as severe as observed in all past reports. Genotyping also identifies nonsense mutations in Lafora PME. Nonsense mutations can be corrected by premature stop codon readthrough drugs such as gentamicin. The community practitioner together with epilepsy specialists in PME can work together and acquire gentamicin (Barton-Davis et al., 1999) for “compassionate use” in Lafora PME, a generalized lysosome multiorgan storage disorder that is invariably fatal. In Unverricht-Lundborg PME, new cohorts with genotyped cystatin B mutations have led to the chronic use of antioxidant N-acetylcysteine and combination valproate clobazam or clonazepam plus antimyoclonic drugs topiramate, zonisamide, piracetam, levetiracetam, or brivaracetam. These cohorts have minimal ataxia and no dementia, questioning whether the syndrome is truly progressive. In conclusion, not only is genotyping a prerequisite in the diagnosis of Dravet’s syndrome and the progressive myoclonus epilepsies, but it also helps us choose the correct antiepileptic drugs to treat seizures in Dravet’s syndrome and Unverricht-Lundborg PME. Genotyping also portends a brighter future, helping us to reassess the true course, severity, and progressive nature of Dravet’s syndrome and Unverricht-Lundborg PME and helping us craft a future curative treatment for Dravet’s syndrome and Lafora disease. Without the genotyping diagnosis of epilepsy causing mutations we are stuck with imprecise diagnosis and symptomatic treatment of seizures.
Genotyping of epilepsy may help to better understand the genetics of epilepsy, to establish an etiology in a patient with epilepsy, to provide genetic counseling, and to confirm a clinical diagnosis. However, critical analysis reveals that genotyping does not contribute to an improved treatment for the patients. In order to improve treatment, genotyping would have to (1) improve our ability to select the drug of choice for a given epilepsy or epileptic syndrome; (2) improve our ability to predict the individual risk of adverse reactions to certain drugs; (3) improve our ability to avoid unnecessary treatments or treatments that could aggravate seizures.
Many example illustrate the lack of impact of genetic information on the treatment outcome: we do not treat Dravet syndrome more successfully since SCN1A testing became available; we do not treat Lafora disease more successfully since testing for laforin and malin became available; we do not need to know the genetic nature of Unverricht-Lundborg disease or test for the cystatin B mutation in order to select or avoid certain drugs; we do not treat Rett syndrome more successfully since MECP2 testing became available; we do not treat JME more successfully since we know its genetic origin; we do not treat autosomal dominant nocturnal frontal lobe epilepsy more successfully since we know its genetic origin and can test for its mutation.
The clinical characteristics as well as the response to treatment of these epilepsy syndromes have been well established before genotyping became available. It can not be argued that genotyping is necessary for establishing a diagnosis or ensure accurate diagnosis. Since not all individuals with given syndromes have been shown to have the corresponding mutation, the clinical diagnosis must have been based on well-established clinical criteria. In addition, the presence or absence of the mutation in a given patient has never been shown to specifically predict the response to any form of treatment, positive or negative. Finally, the appropriate psychological and social help in a given patient will not depend on the identification of a mutation. This does not leave any role for genotyping in epilepsy for the sole reason of improving treatment of the patient. Claiming that the result of genotyping predicts optimal treatment in certain epilepsies is equivalent to stating that genotyping for diabetes has become available and that, based on this breakthrough, insulin can now be selected as the treatment of choice in those who test positive.
Pro: Genetic Information in Humans Helps Us Treat Patients
Antonio V. Delgado-Escueta
SCN1A and the Community Neurologic Practitioner
Why genotype SCN1A?
SCN1A is currently the most clinically relevant epilepsy gene for the community neurologic practitioner because a mutation underpins its clinical diagnosis. It is perhaps the most important advance in ion channelopathies because the mutation is de novo, and the epilepsy is sporadic SMEI and SMEB and infrequently familial. More than 200 novel mutations spread throughout the gene; 95% are de novo (Mulley et al., 2005; Ferraro et al., 2006). The de novo mutations in SCN1A include both missense and truncation mutations and account for 70–80% of Dravet’s SMEI and SMEB (Claes et al., 2001; Oguni et al., 2001; Ohmori et al., 2002; Fujiwara et al., 2003;Nabbout et al., 2003; Wallace et al., 2003; Oguni et al., 2005; Harkin et al., 2007; Marini et al., 2007).
In practice, we now confirm the clinical diagnosis of Dravet’s SMEI and SMEB syndromes by the presence of SCN1A mutations. Community neurologic practitioners can genotype SCN1A through commercial genotyping laboratories. Sequencing of SCN1A is the primary test as it detects mutations in 70–80% of Dravet’s syndrome. Small chromosomal rearrangements involving SCN1A and contiguous genes are also on record to be associated with Dravet’s syndrome. The alterations range from large deletions including SCN1A and several contiguous genes to single SCN1A exon deletions. SCN1A cryptic genomic deletions are rare and range in size from 607–4.7 Mb. Multiplex ligation-dependent probe amplification (MLPA) is a rapid, highly sensitive and relatively economical diagnostic tool that detects all copy number variations of SCN1A from the megabase range to one exon. Sequencing fails to detect an SCN1A mutation in about 10%. In the latter clinical situation, MLPA can be positive and clinch the diagnosis.
SCN1A has 26 exons in 100-kb genomic DNA. Fifty-two percent to 70% of mutations are nonsense mutations and frame shifts; 27% are missense mutations. The nonsense and frame-shift mutations cluster in the C terminus and loop between segments five and six of the first three domains of SCN1A. Another 12% are missense mutations that occupy the voltage sensor of SCN1A. Ceulemans et al. (2004) and Kanai et al. (2004) observed that missense mutations around the pore-forming region and around the voltage sensor region were more likely to produce the most severe phenotype. Five percent of SCN1A mutations are also found in other family members. Such family members more commonly have the GEFS+ phenotypes; less frequently do they also have SMEI. Mosaic mutations of SCN1A have been rarely observed in several patients with SMEI who manifest a milder form of the SMEI phenotype.
In what clinical settings does the community neurologic practitioner genotype SCN1A?
Dravet’s SMEI and SMEB
Originally considered rare, SMEI was originally reported to occur in 1 in 20,000 to 1 in 40,000 of the population (Yakoub et al., 1992). By 1999, SMEI was found in 8% of seizures during the first year of life (Dravet et al., 2005, 2006). SCN1A should be genotyped by the community neurologic practitioner in four clinical settings of Dravet’s SMEI when: (1) febrile convulsions start between 6 and 12 months of age are repeated or prolonged, or evolve to convulsive status epilepticus (SMEI); (2) intractable grand mal tonic–clonic seizures start during infancy and into the second year of life (SMEB); (3) an epileptic encephalopathy is blamed on vaccine encephalopathy as in post pertussis vaccine encephalopathy; and (4) the start and origins of refractory epilepsy and mild mental retardation in an adult can be traced to infancy, and the early story during infancy is consistent with SMEI, that is, febrile and afebrile prolonged seizures, hemiclonic seizures, myoclonias absences, and focal seizures (adults with SMEI or SMEB).
Febrile seizures and SMEI
Aggressive treatment of prolonged and repeated febrile seizures in an otherwise normal infant between 4 and 24 months may be the key to preventing cognitive decline in Dravet’s syndrome. Seizures in Dravet’s syndrome begin before one year of age in all cases. Febrile seizures are often prolonged, generalized, unilateral motor (clonic or tonic) and less commonly tonic–clonic. Nonconvulsive status epilepticus is frequent. Only later do clinicians suspect Dravet’s syndrome when afebrile myoclonic seizures appear together with absences, atypical absences, and complex partial seizures. By then the child is in the second and third years and cognitive decline and developmental delay become evident. Severe-to-moderate mental retardation follows with attention deficit and hyperactivity. Delayed speech development and ataxia become prominent as do erratic myoclonus and long tract signs (Dravet et al., 2005). The mortality rate is high. In the 1992 series of Dravet, with patients followed from 3 to 27 years, 15.9% died. In a series from Tokyo, almost 6% (5.9%) will have passed away by 11 years of age (Ohki et al., 1997). The electroencephalography (EEG) findings may be normal at the beginning, but eventually fast polyspike wave complexes, multifocal spikes, and photoconvulsive responses in 50% of patients with slow background rhythms suggest the diagnosis of SMEI in infants with myoclonias. At this stage, SMEI is very resistant to all forms of antiepileptic drug treatment (Dravet et al., 1992).
However, because of early and correct diagnosis by genotyping of SCN1A mutations, the avoidance of sodium channel blockers, and aggressive treatment of prolonged convulsive status, there is hope that Dravet’s syndrome may not be as severe as observed in all past reports. A new cohort of Dravet’s syndrome with proven SCN1A mutations and in whom sodium channel blockers are avoided should be started. Today, once the presence of SCN1A mutations is confirmed in the first year of life, sodium channel blockers, such as phenytoin, carbamazepine, oxcarbazepine, and lamotrigine, are avoided. These antiepileptic drugs are known to aggravate seizures and cause status epilepticus. Status epilepticus in turn is suspected to contribute to if not cause the severity of mental retardation and death in Dravet’s SMEI. Status epilepticus is listed as a cause of death in the series of Dravet et al. (1992) and in the series of Ohki et al. (1997). Aggressive and acute seizure treatment with intravenous benzodiazepines, rigorous treatment of fever, prevention of hyperthermia, and maintenance of chronic treatment using combination antiepilepsy drug treatment, such as valproate plus stiripentol or valproate plus levetiracetam or valproate plus topiramate or valproate plus ketogenic diet are recommended (Dravet et al., 2006; Guerrini et al., 1998a, 1998b; Striano et al., 2007; Ceulemans et al., 2004; Korff & Nordli, 2007; Striano et al., 2007). Because one-half to two-thirds of mutations are nonsense mutations, because of 6–15.9% mortality rate, and because all patients have some cognitive deficits, a study trial with a premature stop codon drug, such as gentamicin or a form of PTC 124 fused with human insulin receptor that crosses the blood–brain barrier, should be conducted in the future.
It is important to point out that SCN1A mutations are not found in index cases with simple febrile seizures or febrile seizures after 2 years of age, as when Marini et al. (2007) studied 132 patients with febrile seizures. Marini et al. (2007) found SCN1A mutations only in Dravet’s SMEI and SMEB and GEFS+ phenotypes. Seventy percent of SMEI have mutations in SCN1A. However, there are family members of pedigrees with GEFS+ who may have febrile seizures only and have the SCN1A mutation. Mantegazza et al. (2005) reported one family with only febrile seizures who had SCN1A mutations. Of some 200 mutations identified so far in SCN1A of SMEI, one-half to two-thirds are nonsense, splice sites or microdeletions truncating the channel protein and causing loss of function. This contrasts with missense mutations in SCN1A that are present in 11% of families with GEFS+ or generalized epilepsy with febrile seizures plus.
SCN1A and other epilepsy syndromes
Although single case reports or family reports exist where SCN1A mutations in specific syndromes have been found, there is not enough information that supports routine genotyping of SCN1A in simple febrile seizures and temporal lobe epilepsy, in Rasmussen’s encephalitis, and in atypical Panayiotopoulos syndrome (Colosimo et al., 2007; Grosso et al., 2007; Osaka et al., 2007). Genotyping SCN1A is not recommended in all febrile seizures.
The concept of borderline forms: SMEB and intractable grand mal seizures during infancy
Many infants and children who have the distinctive pattern of onset, varying seizure types, and clinical pictures and course similar to SMEI of Dravet, but are not afflicted with myoclonias, have been designated as peripheral or borderline SMEI or SMEB. According to Dravet, other features that separate SMEI from SMEB are “no atypical absences, few EEG epileptiform abnormalities, or only occasional multifocal or diffuse spike waves during the first stage of the disease and rare photosensivity.”Seino and Higashi (1978) and Wada et al. (1983) both described generalized and/or unilateral clonic and tonic–clonic seizures in normal infants with normal EEG findings in the first year of life. There are no other seizure types that appear in the course of illness. Later, EEG epileptiform discharges, generalized or multiple spike waves appear, and psychomotor development slows down. Therefore, the beginning of the illness is very similar to SMEI, and Wada et al. (1983) called this syndrome “high voltage slow wave–grand mal syndrome.” Later on, Kanazawa reported that SMEI, “high voltage slow wave grand mal and variant forms did not differ much clinically, all having a poor prognosis,” and concluded that all may be part of one syndrome and called it “infantile refractory grand mal syndrome.”Fujiwara et al. (2003) and Kanazawa (1992) also called this condition “intractable childhood epilepsies with frequent generalized tonic clonic seizures (grand mal) or refractory grand mal seizures with onset in infancy.”Ogino et al. (1989) proposed the name “polymorphous convulsive epilepsy beginning in infancy.”
More recently, Ohmori et al. (2003), Fujiwara et al. (2003), Fujiwara (2006), and Oguni et al. (2005) found SCN1A mutations in these patients. Fujiwara (2006) and Rhodes et al. (2005) claim that mutations in SCN1A are found in SMEB in similar rates and in similar locations to SMEI. Marini et al. (2007) found that patients with SMEB were more likely to have missense mutations (62.5%), whereas SMEI patients had truncating, splice site, and microdeletion mutations. Fujiwara et al. (2003) reported that 69% of these infants with intractable grand mal have SCN1A mutations.
SCN1A mutations in alleged vaccine encephalopathy
In all series reported so far, febrile or afebrile seizures start before one year of age in Dravet’s syndrome. In the original and updated cohort of Dravet, afebrile seizures usually occurred in the context of a vaccination or an infection, or after a bath. In 80%, febrile seizures then followed. Nieto-Barrera et al. (2000) emphasized the coincidence between the first seizure and DPT (diphtheria–pertussis–polio) vaccination. Berkovic et al. (2006) retrospectively studied 14 patients with alleged vaccine encephalopathy in whom the first seizures occurred within 72 h of vaccination. SCN1A mutations (five truncations and six missense) were identified in 11 of 14 patients (70%).
Adults with Dravet’s syndrome
Jansen et al. (2006) studied 14 adults with refractory epilepsy and intellectual disability whose disorder could be traced to infancy and whose clinical course suggested Dravet’s syndrome. Seizure types were heterogenous with nocturnal tonic–clonic seizures, the most common seizure type. Ten patients had SCN1A mutations; one patient had a mutation in GABRG2. Ebach et al. (2005) studied cryptogenic generalized or focal epilepsy that were traced to infancy. Twenty percent had mutations in SCN1A. Harkin et al. (2007) also found SCN1A mutations in 24% of cryptogenic generalized epilepsy and 22% of cryptogenic focal epilepsy.
GEFS+ and SCN1A
Hidden among the common forms of febrile convulsions is the syndrome of familial generalized epilepsy with febrile seizures plus (GEFS+) with SCN1A mutations in chromosome 2q23-31, SCN1B in chromosome 19q13, and GABRgamma2 in chromosome 5q. GEFS+ is a familial syndrome of febrile convulsions and febrile seizure plus in which the affected members have various epilepsy phenotypes (Scheffer & Berkovic, 1997). Among the seizure phenotypes of affected family members are myoclonic seizures, myoclonic astatic, absences, focal seizures (predominantly temporal lobe epilepsy), atonic and hemiclonic seizures, and SMEI. GEFS+ was first mapped by Wallace et al. (1998) to chromosome 19q, where a missense mutation in the beta1 subunit gene (SCN1B) of the neuronal sodium channel was found. Three reports of GEFS+ kindreds identified a second locus, namely, chromosome 2q23–q31, where mutations in the alpha 1 subunit (SCN1A) resided in evolutionarily conserved residues, changing amino acids within or close to transmembrane segments of the protein (Wallace et al., 1998; Moulard et al., 1999; Peiffer et al., 1999; Escayg et al., 2000, 2001; Baulac et al., 2001; Wallace et al., 2001a, 2001b). The phenotypes in two families reported by Baulac et al. (1999) and by Moulard et al. (1999) included myoclonic, hemiclonic, hemicorporeal, and atonic seizures in addition to absences and tonic–clonic. These phenotypes were similar to those of the Australian kindred reported by Scheffer & Berkovic, 1997. The phenotype in the family reported by Peiffer et al. (1999) was reported as “febrile seizures” (FS), but one member had generalized atonic seizures and the phenotype was typical of GEFS+, according to Lopes-Cendes et al. (2000). Since these studies more than 13 additional missense mutations of SCN1A have been reported accounting for approximately 10% of GEFS families tested.
Genotyping GABR mutations and epilepsy syndromes with absence seizures
Mutations in GABR, for example, GABRG2, GABRA1, GABRD, and GABRb3 subunits of GABAA receptor, have the commonality of causing epilepsy syndromes with absence seizures except that GABRG2 mutations have also been reported in SMEI. These include febrile seizures plus absence in chromosome 5q34, autosomal dominant JME with absence in 5q34, childhood absence in 5q34, and remitting childhood absence in 15q11-12 (Wallace et al., 2001a,b; Baulac et al., 2001; Cossette et al., 2002; Kananura et al., 2002; Dibbens et al., 2004; Tanaka et al., 2008). There are but few patients so far reported with GABR mutations, and any recommendations to genotype these syndromes routinely would need further experience with larger cohorts. Nonetheless, clinical experience in these few patients to date suggests that clonazepam and clobazam are useful in stopping absence seizures.
Why Genotype Adolescent Progressive Myoclonic Epilepsy
Genotyping adolescent progressive myoclonic epilepsy (PME) will separate Lafora disease from Univerricht-Lundborg PME and from JME. In their early stages, Lafora disease, Univerricht-Lundborg PME, and JME are clinically similar (see Table 1). Genotyping patients with PME will identify mutations in one of three known genes, namely, EMP2A/laforin or EMP2B/malin, which cause the majority of Lafora progressive myoclonus epilepsy or cystatin B, which cause almost all of Unverricht-Lundborg PME (Delgado-Escueta, 2007a, 2007b). A second reason for genotyping adolescent PME is to identify nonsense mutations in Lafora disease that can be treated with a premature stop codon readthrough drug, such as intravenous gentamicin used for “compassionate scientific reasons” in an invariably fatal disease. (Barton-Davis et al., 1999; Clancy et al., 2001; Wagner et al., 2001; Wilschanski et al., 2003; Brooks et al., 2006; Sermet-Gaudelus et al., 2007; Welch et al., 2007).
Table 1. Lafora and Unverricht-Lundborg PMEs and JME are similar at onset of illness
|Laforin-deficient||6–19 (80% at 10–12 years)||Yes (in 1/3)||10–14||10–14||17–20||16–20||18–20||20–25||Death in 2–10 years|
|Malin-deficient||7–15||Yes (in 1/3)||16–27||16–27||26–32||26–32||26–37||26–37||Death in 10–30 years|
|Unverricht-Lundborg PME (cystatin B deficient)a||7–16 (86% at 9–13 years)||None||None||Yes||None||None||None||None||Usually mild/chronic; few are severe|
|Juvenile myoclonic epilepsy (JME)a (Herpin-Rabot-Janz type)||8–26 (75% at 12–18 years); (mean age onset 14 yearsb)||None||None||None||None||None||None||None||Good|
Mutations in Lafora disease genes (laforin and malin) and in Univerricht-Lundborg PME genes (cystatin B) are inherited in an autosomal recessive fashion. The gene EPM2A makes an enzyme called dual specificity protein phosphatase or laforin, and the gene EPM2B makes another enzyme named ubiquitin E3 ligase or malin. Therefore, most cases of Lafora PMEs are enzymopathies, are classified as lysosomal diseases, and produce a system-wide glycogen storage disorder. A minority of cases of Lafora disease are caused by as yet unidentified genes. Cystatin B is a cysteine protease inhibitor and its deficiency causes abnormal activation of cathepsin S. G1qB-chain of complement, beta2-microglubulin, glial fibrillary acidic protein, apolipoprotein D, fibronectin 1, and metallothionein II, which are factors involved in proteolysis, apoptosis, and glial activation (Leiuallen et al., 2001). The lysosome-associated functions of cystatin B are also associated with the pathogenic mutations, suggesting that Unverricht-Lundborg PME may also be a lysosomal disease (Alakurtti et al., 2005).
Most commonly, Lafora PME, Unverricht-Lundborg PME, and JME, start as myoclonic, tonic–clonic grand mal and absence seizures at late childhood or adolescence. Rarely, Lafora PME begins in 5–6-year-old children as a learning disorder (Ganesh et al., 2002).
Can early Lafora PME and Unverricht-Lundborg PME be clinically distinguished from JME? Yes. A clinician experienced in PME can look for subtle differences between these epilepsies, but clinicians can still mistake one condition for the other, early in their courses, especially when seizures are not stimuli sensitive, action myoclonus is absent, and there are no occipital seizures. Genotyping will solidify and clarify diagnosis. Moreover, genotyping is crucial for early diagnosis in Lafora PME so that treatment of nonsense mutations can begin.
Unlike JME and Univerricht-Lundborg PME, Lafora PME can have visual occipital seizures in addition to myoclonias, grand mal, and absences at the start of illness. Myoclonias can be present in facial muscles in both Unverricht-Lundborg PME and in Lafora PME, and are never present in facial muscles in JME. Myoclonias are stimuli sensitive in both Lafora and Unverricht-Lundborg PME, and merely tapping the knee jerk or shaking hands can be enough to induce a large myoclonia. The symptom that makes the difference in PME is action myoclonus, which may not become evident, even many years after grand mal and myoclonic seizure onset. Family studies in Lafora PME reveal the appearance of EEG irregular polyspike waves before patients become symptomatic (Van Heycop Ten Ham, 1974). At onset, fragments of incomplete isolated spike waves and diffuse irregular 3–6 Hz polyspike wave complexes appear in both JME and PME against a normal EEG background activity. In the short period of weeks to months, the EEG background activities usually slow down in PME, and focal occipital irregular sharp waves and spikes may appear in Lafora disease. The normal stages of sleep and physiologic sleep patterns disappear in PME. Somatosensory evoked potentials are increased to giant sizes in PME, even in the early stages, and brainstem-evoked potentials show prolonged latencies (Genton et al., 2005). Of course, when a sibling or first cousin has passed away from Lafora disease that has been proven by skin biopsy and/or autopsy, the diagnosis is made easy for the community practitioner. Taking a good family history is, therefore, crucial. There is a higher incidence of Lafora disease in children and adolescents with ancestral origins from the Middle East, Mediterranean or Southern Europe (Spain, France, and Italy), South Asia (India and Pakistan), and North Africa. In these countries consanguineous marriages are not uncommon. Unverricht-Lundborg PME is most common in patients of Northern European origin, for example, the Baltic region and Scandinavia, and is found worldwide because of human migration. Although Lafora PME and Unverricht-Lundborg PME are rare diseases, JME is a common epilepsy that is present in all continents and racial–ethnic groups.
Laforin-deficient PME is more rapidly fatal than malin-deficient PME. In laforin-deficient PME, cognitive decline and poor school performance usually starts with seizures or follows seizures in 1–2 years. Ataxia and spasticity is present by 14–16 years, and dementia and mutism by 17–20 years. Anorexia and swallowing difficulties compel gastrostomy by 18–22 years. Respiratory assistance is necessary usually by 20–25 years. In malin-deficient PME, mild cognitive decline does not start until 16–27 years, ataxia and spasticity at 16–27 years, dementia and mutism at 26–32 years, and gastrostomy and respiratory assist between 26 and 37 years (see Table 1).
Because of its progressive and invariably fatal nature, Lafora PMA should be genotyped by the community neurologic practitioner in its early stages, when it is indistinguishable from JME and Unverricht-Lundborg PME, before ataxia and dementia set in, before independent daily living activities are lost, and certainly before the need for gastrostomy is realized. If genotyping does not show mutations in laforin or malin, a skin biopsy should be performed in all cases of adolescent PME. Biopsy of sweat glands in the axilla should show the disease causing inclusion bodies that stain with periodic acid–Schiff (PAS) inside eccrine sweat duct cells or apocrine myoepithelial cells. These inclusion bodies are made of abnormally branched glycogen called polyglucosan that are present in excessive amounts in various major organs including brain. If PAS+ inclusion bodies are present in the skin biopsy but mutations are absent in laforin or malin, a rarer form of Lafora PME is present caused by the as-yet unidentified third or fourth gene for Lafora PME.
Boosting protein synthesis from <1% to as little as 5% of normal levels reduces severity of disease phenotype. (Ramalho et al., 2002; Kerem, 2004)
Genotyping in Lafora PME will show if deletions, frame shifts, missense mutations, or nonsense mutations are present in laforin or malin. If nonsense mutations are present, the practitioner should work with a specialist in PME and consider using intravenous gentamicin, a premature stop codon readthrough drug that is clinically justified for “compassionate use” in a fatal disorder (Barton-Davis et al., 1999; Clancy et al., 2001;Wagner et al., 2001; Politano et al., 2003;Wilschanski et al., 2003; Brooks et al., 2006; Welch et al., 2007). PTC124, a new orally bioavailable nontoxic, small molecule that selectively induces ribosomal readthrough of premature but not terminal codons and is superior in potency to gentamicin is not yet commercially available. PTC124 does not penetrate the blood brain barrier.
Because Lafora PME is a lysosomal disease and presently has no therapeutic options, high concentrations of intravenous gentamicin can be justified for compassionate use, especially before the needs for gastrostomy and respiratory assist ensue. Lafora PME accumulates polyglucosan bodies in liver, retina, cardiac muscles, esophageal skeletal muscles, and diaphragm muscles producing anorexia, malaise, visual problems, cardiac arrhythmias, and swallowing and breathing difficulties. Boosting protein synthesis of laforin and malin which function to purge polyglucosan inclusion bodies from cells should slow the disease and alleviate the peripheral pathologies responsible for anorexia, malaise, and swallowing and breathing difficulties. High concentrations of intravenous gentamicin can also be expected to cross the blood–brain barrier because of continuing and frequent seizures and because endothelium and neuronal cell death break open the blood–brain barrier. Selective and specific readthrough of disease causing premature termination codons in endothelium and neurons should boost protein synthesis of laforin and malin, which purge the neurons of Lafora inclusion bodies and alleviate the central pathology of Lafora PME. Potential renal and otic toxicities produced by high concentrations of intravenous gentamicin limit their clinical use to invariably fatal diseases like Lafora PME.
Gentamicin, a drug approved by the U.S. Food and Drug Administration (FDA), is an aminoglycoside, and consists of three amino sugars joined in glycosidic linkage to a hexose nucleus, which is in a central position. Rapidly bactericidal, its primary intracellular action is on the 30S ribosomal subunit, which consists of 21 proteins and a single 16S molecule of RNA. In nonsense mutations produced by human genetic diseases, gentamicin also works on translation, binding to the 30S ribosomal subunit. Because it reads through the premature stop codon induced by nonsense mutations, gentamicin only recognizes the true stop codon and allows 30S to 50S complexes to complete translation of mRNA and synthesis of normal protein, for example, laforin. The proof of principle for use of gentamicin as a stop codon readthrough drug was established in Duchenne dystrophy, cystic fibrosis, ataxia telangiectasia, and mucopolysaccharidoses (Barton-Davis et al., 1999; Clancy et al., 2001; Wagner et al., 2001; Wilschanski et al., 2003; Brooks et al., 2006; Sermet-Gaudelus et al., 2007;Welch et al., 2007).
Genotyping Cystatin B in Unverricht-Lundborg PME
Unverricht-Lundborg PME is the most common single cause of PME. Found more in Finland and western Mediterranean countries, it is underdiagnosed in many countries including the United States (Joensuu et al., 2008). Clinicians have known for the last 25 years, at least since the report of Eldridge et al. (1983), that phenytoin aggravates ataxia and dementia of Unverricht-Lundborg PME. More recent experience has shown that sodium channel blockers (carbamazepine, oxcarbazepine, and lamotrigine) and GABAergic drugs (gabapentin, tiagabine, and vigabatrin) also aggravate myoclonus, ataxia, and dementia of Unverricht-Lundborg PME (Genton et al., 2005; Medina et al., 2005). In March 2007, during an International Workshop on Progressive Myoclonus Epilepsies in Sarlat, France, Genton et al. (2006) validated the experience of clinicians that myoclonus, dementia, and ataxia in Unverricht-Lundborg PME can be iatrogenic and caused by specific antiepileptic drugs. A cohort of Unverricht-Lundborg PME who had expansions of CCC-CGC-CCC-GCG-3′ dodecamer in the 5′ untranslated region of cystatin B did not receive phenytoin or carbamazepine or oxcarbazepine or lamotrigine. Treatment relied on valproate plus a benzodiazepine such as clonazepam or clobazam or plus topiramate or zonisamide. Patients also received high doses of piracetam or levetiracetam or combined piracetam and levetiracetam and, most recently, brivaracetam (Koskiniemi et al., 1998; Genton et al., 1999; Genton & Gelisse, 2000;Genton & van Vleyman, 2000; Magaudda et al., 2004; Genton et al., 2006). None of the patients had dementia, and ataxia was at most minimal. Genton (2007) claims that these patients were clinically indistinguishable from those with JME because they had no dementia and ataxia was minimal. For this clinical reason, Genton (2007) suggests that patients with adolescent onset myoclonias and grand mal seizures should be genotyped for cystatin B mutations to separate Unverricht-Lundborg type PME from JME (see Table 1).
Kalviainen et al. (2008) further suggest that when action myoclonus is absent or very mild, Unverricht-Lundborg PME can be easily misdiagnosed as JME. Alternatively, during the course of a drug-resistant JME, a diagnosis of Unverricht-Lundborg PME should be considered and genotyping for cystatin B mutations should be discussed with the patient.
Virtually all Unverricht-Lundborg PME–affected individuals have an unstable expansion of a 12-nucleotide (dodecamer) repeat 5′-CCC-CGC-CCC-GCG-3′ (Lalioti et al., 1997a, 1997b). The expanded dodecamer repeat mutation accounts for approximately 90% of Unverricht-Lundborg PME found throughout the world. So far, nonsense mutations have been rarely reported in exon 3 of cystatin B (Pennachio et al., 1996; Lafreniere et al., 1997; de Haan et al., 2004). Its rarity in Unverricht-Lundborg PME should be discussed with the patient before genotyping is done and treatment with premature termination codon readthrough drugs is considered. On rare occasions a patient may be referred with a clinical diagnosis of Unverricht-Lundborg PME and their family has already been reading about N-acetylcysteine on the Internet and has already talked to families who swore to N-acetylcysteine and its ameliorating effects as an antioxidant in Unverricht-Lundborg PME (Hurd et al., 1996; Edwards et al., 2002). In such a clinical setting, obtaining an accurate diagnosis of cystatin B mutations would be important before treatment with N-acetylcysteine.
Genotyping JME Genes
Of 13 chromosome loci genetically linked to JME, three mendelian genes [α1-subunit of the GABAA receptor (GABRA1), chloride channel 2 gene (CLCN2), and myoclonin1/EFHC1] and two SNP-susceptibility alleles of putative JME genes in epistases [bromodomain-containing protein 2 (BRD2) and connexin (Cx)-36] have been identified to date (Delgado-Escueta, 2007a, 2007b). When should the practitioner ask the JME researchers (genotyping in JME is not yet commercially available, 2008) to genotype for known JME-causing genes? In certain rare clinical settings (as occurred with the author) genotyping for JME mutations has been clinically useful. (1) During the course of drug-resistant JME, MRI shows an atrophic cerebellum (that had been caused possibly by phenytoin) and genotyping for cystatin B mutations was negative, the family asked for genotyping JME mutations, fearful of a degenerative disorder. (2) In the course of drug resistant adolescent myoclonic epilepsy in a Hispanic American, and anxious parents are referred already showing negative results in genotyping for Lafora and Unverricht-Lundborg PME, the presence of a myoclonin/EFHC1 mutation can reassure parents. (3) When phase one presurgical evaluation has been completed and intracranial recordings/stereo-EEG is planned in a drug-resistant myoclonic syndrome suspected to be drug-resistant frontal lobe epilepsy, genotyping may solidify the diagnosis of JME and negate the need for intracranial recordings.
Because none of the five putative JME genes, so far, mandates a choice in antiepileptic drugs for treatment, or a change in management (because there are at least eight more genes to be identified in eight known chromosome loci for JME), genotyping is not routinely used in JME. The use of stop codon readthrough drugs cannot be justified for “compassionate use” to correct nonsense mutations in JME because JME is not a debilitating or fatal syndrome. When we obtain stop codon readthrough drugs that cross the blood–brain barrier and whose adverse effects are equal to or less than antiepileptic drugs, then perhaps such drugs can be used to treat nonsense mutations in JME.
Genotyping Helps Counseling, Psychological, and Social Support
Genotyping not only ensures accurate diagnosis and helps treatment, but it can also lead to more accurate genetic counseling, psychological, and social support for patients and families. Genotyping will direct families to organizations that have support groups, and similar empathetic families who are educated about the PMEs or Dravet’s syndrome and who are devoted to finding cures for their specific epilepsy diseases. Social and psychological help is of critical importance in coping with depression and the debilitating neurologic deficits of PME or the cognitive problems of children in Dravet’s syndrome.
In conclusion, not only is genotyping a prerequisite in the diagnosis of Dravet’s syndrome and the progressive myoclonus epilepsies, but it also helps us choose the correct antiepileptic drugs to treat seizures in Dravet’s syndrome and Unverricht-Lundborg PME. Genotyping also portends a brighter future, helping us reassess the true course, severity, and progressive nature of Dravet’s syndrome and Unverricht-Lundborg PME and helping us craft a future curative treatment for Dravet’s syndrome and Lafora disease. Without the genotyping diagnosis of epilepsy causing mutations, we are stuck with imprecise diagnosis and symptomatic treatment of seizures.
CON: Genetic Information does not help at all Blaise F.D. Bourgeois
All the preceding information on genotyping of epilepsy is of great interest for our understanding of the genetics of epilepsy, to help establish an etiology in a patient with epilepsy, to provide genetic counseling, and to confirm a clinical diagnosis. However, a closer look at the data reveals that they do not contribute to an improved treatment for the patients. Let us take the case of a patient with a form of epilepsy for which genotyping is available. The patient has the choice between two neurologists. They are both equally competent and knowledgeable. The only difference is that one knows nothing about the genetic information discussed previously in this report or has no access to genotyping (Neurologist A) and the other has access to all this genetic information and to genotyping (Neurologist B). Is the patient really going to be treated more successfully by the one with genetic knowledge? If yes, how and why? In order to answer this question, we first need to define the meaning of “Help us to treat patients.” The definition has to include that the genetic information will help to achieve (1) better seizure control, and/or (2) seizure reduction with less adverse effects, and/or (3) seizure reduction within a shorter time.” If genetic information cannot contribute to achieve one of these three goals, with equal or better success towards the other two goals, then it does not help at all with treatment. In addition, the genetic information would have to yield relevant clues for treatment that could not be obtained by any other currently available clinical tool.
How could genetic information help us to achieve either one of these three goals?
- 1Genotyping improves our ability to select the drug of choice for a given epilepsy or epileptic syndrome.
- 2Genotyping improves our ability to predict the individual risk of adverse reactions to certain drugs.
- 3Genotyping improves our ability to avoid unnecessary treatments or treatments that could aggravate seizures.
Having established how genotyping might improve treatment, the next step in our analysis is to address some of the following questions:
- 1Do we treat Dravet’s syndrome more successfully since SCN1A testing became available?
- 2Do we treat Lafora disease more successfully since testing for laforin and malin became available?
- 3Do you need to know the genetic nature of Unverricht-Lundborg disease or test for the cystatin B mutation in order to select or avoid certain drugs?
- 4Do we treat Rett syndrome more successfully since MECP2 testing became available?
- 5Do we treat JME more successfully since we know its genetic origin?
- 6Do we treat autosomal dominant nocturnal frontal lobe epilepsy more successfully since we know its genetic origin and can test for its mutation?
- 7Does the identification of a ring chromosome 20 lead to more successful therapy of the epilepsy?
Let us focus on the three syndromes discussed earlier in greater detail: (1) Dravet’s syndrome and SCN1A; (2) Lafora disease and testing for laforin and malin; and (3) Unverricht-Lundborg disease and testing for the cystatin B mutation.
Why the availability of genotyping did not improve treatment in these epilepsies
- 1These syndromes have been described, defined by clinical criteria, and accurately diagnosed long before genotyping became available. The information regarding drugs that can reduce seizure frequency and drugs that can worsen the course of the disease became available long before genotyping, such as the unfavorable response to lamotrigine in severe myoclonic epilepsy (Guerrini et al., 1998a, 1998b). Therefore, claiming that the result of genotyping determines optimal treatment in certain epilepsies is equivalent to stating that genotyping for diabetes has become available and that, based on this breakthrough, insulin can now be selected as the treatment of choice for those who test positive.
- 2It could be argued that genotyping is necessary for establishing a diagnosis or that it ensures accurate diagnosis. We were provided with the proof that this is not the case. Indeed, it has been reported that 70–80% of patients with SMEI or SMEB have the SCN1A de novo mutation. Accordingly, in the other 20–30% of the cases, the authors of these reports felt comfortable to establish the diagnosis without genotyping, based on well-established clinical criteria. If genotyping were indeed necessary to establish the diagnosis, the mutation would have to be present in 100% of the cases, by definition.
- 3It could be argued that the result of genotyping (or knowing the underlying channelopathy) could help us select the best antiepileptic drug based on its mechanism of action, or predict those that may aggravate seizures. If this were true, then those patients with SMEI or SMEB who test positive for the SCN1A mutation (sodium channelopathy) should have a response to medications that is different from those who test negative. In addition, in patients who test negative, the sodium channel blockers phenytoin, carbamazepine, oxcarbazepine, and lamotrigine would not need to be avoided. No such observation has been reported and, until it is, there is no evidence that the result of the genotyping in SMEI matters for treatment choice or for predicting treatment response. A case has been made for genotyping for cystatin B in all adolescent-onset myoclonic and generalized tonic-clonic seizures, in order to distinguish between JME and Unverricht-Lundborg disease, because phenytoin, carbamazepine, oxcarbazepine, and lamotrigine (Genton et al., 2006) aggravate Unverricht-Lundborg disease. However, these very same drugs have been shown to aggravate JME also (Biraben et al., 2000; Genton et al., 2000; Carrazana & Wheeler, 2001), and they should be avoided in these patients regardless. So far, successful treatment of patients with Lafora progressive myoclonic epilepsy with intravenous gentamicin has not been published, and the evidence for N-acetylcysteine treatment of Unverricht-Lundborg disease is anecdotal (Edwards et al., 2002).
Genotyping has not provided new insight into the clinical presentation or the response to treatment in the corresponding syndromes. Genotyping is not necessary to establish a diagnosis, especially given that not all individuals with the syndrome have been shown to have the mutation. Finally, the presence or absence of the mutation in a given patient has never been shown to specifically predict the response to treatment, positive or negative. These conclusions do not leave any role for genotyping for the sole reason of improving treatment of the corresponding syndrome. Finally, it is not the identification of a mutation itself that will play a pivotal role in the implementation of the appropriate “psychological and social help” in a given patient.
There is little doubt that, in the foreseeable future, genetic discoveries will improve treatment by meeting one of the three criteria that have been outlined: “Achieve (1) better seizure control, and/or (2) seizure reduction with less adverse effects, and/or (3) seizure reduction within a shorter time.” But, as of now, a patient with epilepsy would be treated as successfully by competent neurologist A (who does not have access to genotyping) as by competent neurologist B (who has access to genotyping).
We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
Disclosure: The authors declare no conflicts of interest.