Molecular genetics and treatment of narcolepsy

Abstract

Narcolepsy is a neurological disorder characterized by excessive daytime sleepiness and cataplexy. The hypocretin/orexin deficiency is likely to be the key to its pathophysiology in most of cases although the cause of human narcolepsy remains elusive. Acting on a specific genetic background, an autoimmune process targeting hypocretin neurons in response to yet unknown environmental factors is the most probable hypothesis in most cases of human narcolepsy with cataplexy. Although narcolepsy presents one of the tightest associations with a specific human leukocyte antigen (HLA) (DQB1*0602), there is strong evidence that non‐HLA genes also confer susceptibility. In addition to a point mutation in the prepro‐hypocretin gene discovered in an atypical case, a few polymorphisms in monoaminergic and immune‐related genes have been reported associated with narcolepsy. The treatment of narcolepsy has evolved significantly over the last few years. Available treatments include stimulants for hypersomnia with the quite recent widespread use of modafinil, antidepressants for cataplexy, and gamma‐hydroxybutyrate for both symptoms. Recent pilot open trials with intravenous immunoglobulins appear an effective treatment of cataplexy if applied at early stages of narcolepsy. Finally, the discovery of hypocretin deficiency might open up new treatment perspectives.

• Antidepressant
• autoimmune
• COMT
• environmental
• GHB
• genes
• IVIg
• modafinil
• narcolepsy
• stimulants

Narcolepsy is a disabling disorder characterized by excessive daytime sleepiness (EDS) and abnormal rapid‐eye‐movement (REM) sleep manifestations including cataplexy (sudden loss of muscle tone triggered by strong emotions), sleep paralysis, hypnagogic hallucinations, and sleep onset REM periods 1. Nocturnal sleep is usually disturbed with frequent parasomnias, obstructive sleep apneas, and periodic leg movements 2. Familial cases are rare (less than 10%) and contain only occasionally more than two affected individuals 3. However, the risk of narcolepsy in a first‐degree relative of a proband is 2%, 10–40 times higher than the prevalence observed in the general population (0.025%) 3. Twin studies indicate that only 6 monozygotic pairs, among the 17 described, are concordant for narcolepsy‐cataplexy 3–5. Since most cases are sporadic and monozygotic twins are usually discordant for narcolepsy, the development of narcolepsy should involve environmental factors acting on a specific genetic predisposition.

Narcolepsy is one of the most studied sleep disorders at the molecular level. Recent findings in animal models revealed that the hypocretin system plays a major role in the etiology of narcolepsy 6,7. A marked decrease in hypocretin‐1 levels in the cerebrospinal fluid (CSF) of narcolepsy patients 8,9 and the number of hypocretin neurons in post‐mortem brain tissues of narcolepsy subjects 10,11 were reported. The treatment of narcolepsy has evolved significantly over the last few years with the quite recent widespread use of modafinil for hypersomnia, antidepressants for cataplexy, and gamma‐hydroxybutyrate for both symptoms. In addition, recent pilot open trials with intravenous immunoglobulins appear also an effective treatment of cataplexy if applied at early stages of the disease 12,13. Although there is no cure for narcolepsy, new therapies based on pathophysiological mechanisms are currently being developed.

Key messages

• We still need to identify the steps of the pathophysiological process leading to hypocretin deficiency. We also need to identify new susceptibility genes and environmental factors that may cause the hypocretin/orexin deficiency in narcolepsy.

• Current treatment strategies are only symptomatic and new therapies based on underlying pathophysiological mechanisms (immune‐based therapies) are awaited in narcolepsy.

Molecular genetics

Sporadic cases

Human leukocyte antigen (HLA) genes

Numerous studies have demonstrated that narcolepsy is associated with HLA 14–19. In 1983, a first study reported 100% association between narcolepsy and the HLA‐DR2 haplotype in Japanese patients 15, a finding that was confirmed immediately in Caucasians 16. Four alleles corresponding to DRB1*1501, DRB5*0101, DQA1*0102, and DQB1*0602 constitute the susceptibility haplotype associated with human narcolepsy in Caucasian populations. However, shortly after this extraordinary finding, it was demonstrated that the DR2 association is dependent on the ethnic origin, with African‐American narcoleptics presenting a weaker (60%–65%) association 17. In this ethnic group the strongest association is found with DQA1*0102, DQB1*0602 haplotype typically in linkage disequilibrium with DRB*1503, suggesting that the susceptibility gene should be closer to the DQ loci 14,18. The principal predisposing allele is actually DQB1*0602, an allele found in 85%–95% of narcolepsy patients with cataplexy. Furthermore, HLA‐DQB1*0602 homozygosity doubles or quadruples the risk for narcolepsy 19 and the relative risk for narcolepsy varies in heterozygote subjects according to the allele associated with DQB1*0602 14. As in other HLA‐associated disorders, the associations are complex, and other DR and DQ alleles have a protective or predisposing effects. For example, DQB1*0301 increases susceptibility, whereas alleles such as DQB1*0601 and DQB1*0501 are protective 14. A recent finding is the possibility of additional effects of DRB1*0407. These effects are nevertheless far weaker than those of DQB1*0602, and the global contribution of the class II HLA system to the total genetic risk of narcolepsy is only partial 14.

Twelve to 38 percent of the general population caries the HLA DQB1*0602 while only a small fraction has narcolepsy, which may indicate that this allele is neither necessary nor sufficient to trigger narcolepsy 14,18, especially in narcolepsy without cataplexy and in familial cases of narcolepsy. Accordingly, other non‐HLA genes could also confer susceptibility to narcolepsy.

Hypocretin genes

It has been shown that in Doberman pinschers and Labrador retrievers, narcolepsy is transmitted as a recessive autosomal trait with complete penetrance. After intensive work over the past 15 years on the genetics of canine narcolepsy at Stanford University, Mignot's group identified, through linkage analysis and positional cloning, mutations in the hypocretin‐2 receptor as the cause of canine narcolepsy 6. Hypocretin‐1 and ‐2 are hypothalamic neuropeptides acting on two receptor subtypes and first found to be involved in feeding behavior 20. Simultaneously, Yanagisawa's group discovered in the mouse a phenotype similar to canine and human narcolepsies after a targeted deletion of the prepro‐hypocretin gene 7. Moreover, hypocretin/ataxin 3 transgenic mice (with apoptosis of hypocretin neurons after birth) present a narcolepsy phenotype as well 21.

In humans, a mutation in the prepro‐hypocretin gene was identified in an atypical case of narcolepsy (HLA DQB1*0602 negative, very young age at onset and severe phenotype), with a G to T substitution resulting in a Leu to Arg substitution in the peptide signal 10. No pathogenic mutations have been observed in the hypocretin receptor 1 and 2 genes. In addition, several recent studies demonstrated that the loci for prepro‐hypocretin, hypocretin receptor 1 and 2 do not contribute to susceptibility to narcolepsy in any significant way 10,22,23. The absence or strong reduction in the number of hypocretin neurons have, however, been demonstrated in human narcolepsy with a selective destruction of these neurons being the most probable etiology 10,11. Therefore, the causal implication of the hypocretin system in narcolepsy is now established. Most of sporadic narcoleptic patients with clear‐cut cataplexy and with positive HLA DQB1*0602 have undetectable CSF hypocretin‐1 levels. However even in this group, rare patients may also have normal or intermediate levels 8,9,24. In addition, in patients affected with narcolepsy without cataplexy or with positive familial history of narcolepsy, the association with low hypocretin is not as clear‐cut 9,24. Together with the tight association with the HLA antigens and the young bimodal distribution of age at onset 25, the most likely cause of hypocretin deficiency in narcolepsy might be an autoimmune process resulting in degeneration of hypocretin‐containing neurons in the hypothalamus. Because over 90% of narcolepsy subjects have no family history of narcolepsy and monozygotic twins are mostly discordant, environmental factors might also play an important role 4. The environmental factor(s) might trigger narcolepsy by inducing an autoimmune reaction that targets hypocretin neurons in genetically susceptible individuals.

Monoaminergic genes

Consistent with the imbalance between monoamines and acetylcholine in narcolepsy, two studies have looked for an association between polymorphisms in monoaminergic candidate genes and sporadic narcolepsy. An early study reported an association between a variable number of tandem repeats of monoamine oxydase A (MAO A) and narcolepsy 26, although this was not confirmed by us 27. A sexual dimorphism in the functional polymorphism of the gene encoding the catechol‐O‐methyltransferase (COMT) as well as an effect on the severity of daytime sleepiness 27 and the response to stimulant treatment (modafinil) 28 were detected in narcolepsy, suggesting a more critical alteration in the dopaminergic/noradrenergic than in the serotoninergic neurotransmission. In addition to a potential association between narcolepsy and the functional COMT polymorphism, a recent study revealed positive associations with genomic regions related to neurotransmission pathways such as the dopamine receptor D2 (DRD2), γ‐aminobutyric acid receptor β‐1 (GABABR1), and the serotonin receptor 2A (5HTR2A) 29. The role of the serotoninergic gene system remains controversial since another study reported the absence of association between 5HTR2A or tryptophan hydroxylase (TPH) genes and narcolepsy 30.

Immune‐related genes

Several studies sought for an association between narcolepsy and immune‐related genes with potential pathophysiological interests. A few association studies looked for the presence of single nucleotide polymorphism in the tumor necrosis factor alpha (TNF) gene promoter, with conflicting results, although TNF may be a good candidate gene, particularly in association with HLA‐DRB1*1501 31,32. An association between the tumor necrosis factor receptor 2 (TNFR2) gene and human narcolepsy has also been reported 33, indicating the possibility of an additive effect (relationship between TNFR2 and TNF) on the susceptibility to narcolepsy. Finally, in relation to apoptosis/immunity hypothesis targeting hypocretin neurons in narcolepsy, the possibility of an association between apolipoprotein E4 and narcolepsy has been tested but without any positive result 34.

In summary, the above association studies in sporadic narcolepsy, although preliminary and needing confirmation and replication in independent samples, support a complex pathogenetic model including disturbances in both the neurotransmission and the immune/autoimmune pathways (Table I).

Table I. Gene targets in human narcolepsy.

Genes	Mutation	Polymorphisms
Sporadic cases:
Hypocretin	One mutation (prepro‐hypocretin gene)	No pathogenic polymorphism
HLA DQB1*0602	No mutation	Pathogenic polymorphism
MAO‐A	No mutation	No pathogenic polymorphism
COMT	No mutation	Sexual dimorphism and correlation with severity of the phenotype
5HT2A and TPH	No mutation	No pathogenic polymorphism
TNF promoter and receptor 2	No mutation	Pathogenic polymorphism?
Apolipoprotein E4	No mutation	No pathogenic polymorphism
Familial cases:
HLA DQB1*0602	No mutation	Pathogenic polymorphism
Genome‐wide linkage (Japanese population)		Linkage to 4p13‐q21
Genome‐wide linkage (French population)		Linkage to 21q

Familial cases

Multiplex families with full‐blown narcolepsy‐cataplexy are rare while families with members affected either with narcolepsy‐cataplexy or with EDS without cataplexy are more common. In 10%–30% of narcolepsy probands, first or second degree relatives are affected with attenuated phenotypes mainly with isolated recurrent daytime naps and/or lapses into sleep 3,35.

A first genome‐wide mapping study reported a suggestive linkage (LOD score of 3.09) to chromosome 4p13‐q21 in eight small multiplex Japanese families affected with narcolepsy‐cataplexy 36. The identified region contains the candidate genes circadian locomotor output cycles kaput (CLOCK) and GABRB1, but no mutation search has been conducted (Table I). We recently reported a genome‐wide linkage analysis in a large French family with four members affected with narcolepsy‐cataplexy and ten others with isolated recurrent naps or lapses into sleep 37. Our linkage and direct sequencing results excluded the involvement of the hypocretin/orexin genes, the candidate regions containing major monoaminergic genes, and the 4p13‐q21 region. Only three regions showed LOD scores >1 in two‐point linkage analysis (D6S1960, D11S2359, and D21S228). Genotyping additional markers provided support for linkage to nine markers on chromosome 21 (maximum two‐point LOD score = 3.36 at D21S1245). The multipoint linkage analysis provided further evidence for linkage to the same region (maximum parametric LOD score = 4.00 at 21GT26K). A single haplotype was shared by all affected individuals and informative crossovers indicated that the gene that confers susceptibility to narcolepsy in this family is likely to be located between markers D21S267 and ABCG1, in a 5.15 Mb region of 21q. A systematic search of the human genome draft sequence has identified 23 known and 11 putative genes in this region with several potential candidate genes expressed and/or having a clear role in the brain 35. A systematic search for mutations or polymorphisms in these genes or changes in their expression is under investigation. A recent report on an atypical HLA DQB1*0602 narcolepsy‐cataplexy patient associated with Down's syndrome 38 further strengthens our finding that there is a narcolepsy susceptibility gene on 21q.

Perspectives in the genetic field

Even if narcolepsy seems a complex, polygenic, and environmentally modulated disorder, the hypocretin/orexin deficiency is likely to be the key to its pathophysiology. The identification of new susceptibility genes, of environmental triggering factors, and the study of gene‐environment interactions will constitute the major next steps. In sporadic cases, most of susceptibility genes are anticipated to be involved in an autoimmune process targeting hypothalamic hypocretin/orexin neurons in conjunction with or independent of HLA. Finally, new genome‐wide linkage analyses are needed in narcoleptic families to confirm the chromosome 21 localization, and candidate gene analyses within the identified region need also to be performed.

Treatment

There have been several advances in the treatment of either EDS or cataplexy in the last few years. Because no cure is available for narcolepsy, its management relies on several classes of drugs, namely stimulants for EDS and irresistible episodes of sleep, antidepressants for cataplexy and other REM‐associated symptoms, and hypnotics for poor nocturnal sleep. Recently, a single drug named gamma‐hydroxybutyrate (GHB) seems to be effective in narcolepsy for controlling EDS, cataplexy, and the disturbed nocturnal sleep. An effort in standardizing the treatment of narcolepsy was initiated in the US four years ago 39.

Treatment of sleepiness and irresistible sleep episodes

Modafinil, methylphenidate, amphetamine and amphetamine‐like drugs are common stimulants for the treatment of EDS in narcolepsy. However, the availability of modafinil for almost 20 years under controlled distribution in France and since 1998 in the US has dramatically changed the prescription of stimulants in narcolepsy. Recently the reemergence of GHB as a major treatment for narcolepsy may also change future prescription practices.

Modafinil

Modafinil (2‐diphenyl‐methyl‐sulfinyl‐acetamide) is a wakefulness‐promoting drug whose mechanism of action is unclear but clearly differs from amphetamines. Several studies have outlined its possible action on dopaminergic 27,40–42 adrenergic and noradrenergic 43,44 and serotonergic/GABAergic systems 42. Elimination half‐life is 15 hours, and maximal concentration is achieved in 2–4 hours. Most of randomized studies 45–48 using 200–400 mg per day reported a significant reduction of irresistible episodes of sleep and of severe somnolence as measured by the Epworth Sleepiness Scale (ESS), a significant improvement in maintaining wakefulness measured by the Maintenance Wakefulness Test (MWT) and in decreasing sleepiness judged by the Multiple Sleep Latency Test (MSLT). We have recently analyzed the response to stimulant treatment in our narcoleptic sample, and found a strong effect of COMT genotype on response to modafinil 28. This first pharmacogenetic study in narcolepsy suggests that female narcoleptics are more frequently of low COMT enzyme activity genotype, less severely affected, and respond better to modafinil treatment 28. In addition, the optimal daily dose of modafinil was significantly lower in all narcoleptics with the low COMT enzyme activity.

Undesirable side effects are relatively rare and include, mainly at the onset of treatment, headache, irritability, and insomnia. There is no clear evidence of tolerance or abuse potential in contrast to amphetamine. The possibility of induction of human hepatic cytochrome P450 enzymes is well known. Therefore, modafinil may increase the metabolism of oral contraceptives and a product containing at least 50 µg of ethinylestradiol is highly recommended.

Gamma‐hydroxybutyrate (GHB)

GHB, or sodium oxybate in its most recent designation, is proposed to be a metabolite of GABA that acts as a natural neurotransmitter through its own receptors (GHB receptors) and/or via the stimulation of GABAB receptors. GHB may also modulate the activity of dopamine neurons with an attenuation of dopamine neurotransmission 49. Previous studies reported its efficacy in the management of REM‐related symptoms including cataplexy, hypnagogic hallucinations, and sleep paralysis but also for EDS and irresistible episodes of sleep 50,51. GHB has recently reemerged as a major treatment of narcolepsy with a clear reduction of EDS and an increased level of alertness 52,53. Optimal dose is 6–9 g/night in two separate doses (at bedtime and 2.5–4 hours later). Elimination half‐life is 40 to 60 min, but its benefic effects persist much longer. Side effects are dizziness, headache, nausea, pain, depressive mood, enuresis, and sleepwalking. The major problem with GHB is its nonmedical use to elicit altered states of consciousness with important social and legal implications and there are many reports of high abuse potential with risks of seizures and coma with overdose in nonnarcolepsy subjects. Therefore, GHB is a very controlled substance and there are huge rigorous regulations for patients using this drug both in the US and in Europe.

Amphetamines and amphetamine‐like CNS stimulants

The first drug used in 1931 for the treatment of sleepiness associated with narcolepsy was ephedrine 54, followed by amphetamines benzedrine 55, d‐amphetamine 56, l‐amphetamine 57, and methamphetamine 58. These compounds promote monoamine (catecholamines but also serotonin) release, blocks monoamine reuptake, and inhibit slightly the monoamine oxydase. Amphetamines are very effective against sleepiness in narcolepsy. The d‐isomer is more specific to dopaminergic transmission and appears four times more potent than the l‐isomer. Methamphetamine is more lipophilic than d‐amphetamine and therefore has more affinity for the central nervous system and fewer peripheral side effects than d‐amphetamine. Elimination half‐life of amphetamines varies between 10 and 30 hours, depending on their metabolites. Several adverse effects include irritability, aggressiveness, insomnia, hypertension, and abnormal movements at doses from 15 to 60 mg/day. Serious toxic effects including anxiety, aggressiveness, and psychotic reactions may occur at dose > 60 mg/day 59,60. Abuse potential is theoretically important but has rarely been reported in narcolepsy patients. However, tolerance may occur in one‐third of patients 59,61.

Methylphenidate

Methylphenidate (methyl‐phenyl‐2‐piperidineacetate) was introduced in 1959 62. Methylphenidate blocks the reuptake of monoamines but has only a slight effect on monoamine release in contrast to amphetamine. Clinical experience clearly indicates an improvement in daytime sleepiness in narcolepsy patients with daily dose between 10 and 30 mg 63. Elimination half‐life is six hours. Side effects are similar to amphetamines but irritability, insomnia, and hypertension are less frequent with methylphenidate. Tolerance and abuse potential are rare. A long‐acting form is now available that may be of interest in some cases.

Mazindol

Mazindol is an imidazolidine derivative with dopamine and adrenaline reuptake blocker properties. Elimination half‐life is ten hours. There were a few reports on the use of mazindol in narcolepsy patients with significant improvement of sleepiness in up to 75% of patients 64. Clinical experience suggests a low starting dosage of 1 mg/day and optimal daily dose of 2–3 mg/day. Side effects include dry mouth, nervousness, constipation, and, less frequently, nausea, vomiting, headache, dizziness, and tachycardia. Mazindol has less potential for abuse and tolerance than amphetamines 65.

Others (rarely used)

Pemoline

Pemoline is an oxazolidine derivative that selectively blocks the dopamine reuptake. Elimination half‐life is 16 hours. There were a few reports on the use of pemoline in narcolepsy patients with moderate improvement in sleepiness in up to 65% of patients 66. However, due to reports of liver toxicity, the medication has been stopped in several countries, and in the others the use of pemoline in narcolepsy patients is rare 66.

Monoamine oxidase inhibitors (MAOIs)

MAOIs increase monoamine signaling by inhibiting the mitochondrial enzyme monoamine oxydase that metabolizes the three classes of amines: noradrenaline, dopamine, and serotonin. Selegiline is a methamphetamine derivative and a potent irreversible MAO‐B selective inhibitor, with several metabolites such as desmethyl selegiline, amphetamine, and methamphetamine. Elimination half‐life varies between 2 and 20 hours, depending on metabolites. There were a few reports on the use of selegiline in narcolepsy patients with a significant improvement of daytime sleepiness and a reduction of irresistible episodes of sleep at 10–40 mg/day 67,68. Phenelzine is an irreversible nonselective MAOI with positive effects on daytime sleepiness in severe and resistant cases to conventional therapy 69. Optimal daily doses are 60–90 mg. Side effects of MAOIs are rare including dry mouth, headache, insomnia and dizziness. The main limitation of MAOIs use is the risk of serotonin syndrome and hypertensive crisis, in cases of co‐ingestion of tyramine‐containing food or sympathomimetic agents. Therefore, treatment with MAOIs cannot be combined with tricyclic antidepressant, serotonin specific reuptake inhibitors (SSRIs) or sympathomimetic medications. Finally, MAOIs were rarely used in narcolepsy.

In conclusion, we may propose that the first line treatment of daytime sleepiness for the management of narcolepsy is modafinil with an optimal daily dose of 100–400 mg (in two doses per day) with in a few cases the possibility to increase the dose up to 600 mg per day. An alternative treatment, GHB, based on randomized controlled trials, may become available in the near future (Table II). In circumstances of intolerance or insufficient efficacy, an alternative stimulant should be proposed: methylphenidate (10–30 mg), mazindol (2–4 mg), or selegiline (10–40 mg) (Table II). Rare cases where neither of these drugs is active can benefit from d‐amphetamine or methamphetamine, under close control. Although co‐prescription of stimulants is common in resistant narcolepsy patients, there is no clear evidence to recommend a particular combination. Finally, behavioral treatment (recommendation to take voluntary naps during the day) is often necessary but without clear guidelines to support the number, duration and timing of naps.

Table II. Pharmacological treatment for excessive daytime sleepiness and irresistible sleep episodes.

Stimulants:
Modafinil (100–400 mg/day)	First‐line treatment
Gamma‐hydroxybutyrate (6—9 g/night) or Methylphenidate (10–60 mg/day)	Second choice if persistent difficulties
Mazindol (1–4 mg/day) or Selegiline (20–40 mg/day)	Third choice
Dextroamphetamine (5–60 mg/day) or Methamphetamine (20–25 mg/day)	In rare cases of no response
Future therapies:
Histaminergic H3 antagonist/inverse agonists
Thyrotropin‐Releasing Hormone (TRH) analogues
GHB long release
Modafinil long release
Hypocretin agonists
Hypocretin cell transplantation

Treatment for cataplexy and other REM‐associated phenomena

GHB is the first and only medication specifically indicated for the treatment of cataplexy in both US and Europe. GHB has been considered as effective on the number of cataplectic attacks in previous studies 51,70. Recently, the US Xyrem Multicentric Study Group confirmed 52,53 the clear reduction in the number of cataplectic attacks at doses between 3 g and 9 g nightly in two administrations. The positive effect was significant at 4 weeks, maximal after 8 weeks with maintenance of efficacy in patients from 7 to 44 months 71. Patients showed no evidence of tolerance. Adverse effects are already mentioned (see Treatment of sleepiness section). Contrary to antidepressant‐based therapies for cataplexy, GHB interruption does not result in a rebound of cataplexy.

A variety of medications representing several major central nervous system drug classes (aminergic reuptake inhibitors) may also improve cataplexy but with only poor level of evidence for efficacy. Drugs with efficacy on cataplexy are also clearly associated with reduction in hypnagogic and hypnopompic hallucinations and sleep paralysis symptoms. The anticataplectic effects of these drugs differ from their antidepressant effects with a rapid response (less than a week) and variable dose to control cataplexy with low doses for tricyclic agents but high doses for serotonin specific reuptake inhibitors (SSRIs).

Tricyclic antidepressants are used since 1960 to treat cataplexy with complete abolition or decrease in severity and frequency of attacks 72–76. Clomipramine, the most potent anticataplectic drug, has serotonergic, noradrenergic, and dopaminergic reuptake inhibitory and anticholinergic properties. Daily doses have to be titrated according to their anticataplectic effects and potential side effects, and may vary between 10 and 75 mg, but low doses (10–20 mg) are often effective. Adverse effects are frequent, especially anticholinergic effects (dry mouth, urinary retention, nausea, constipation, blurred vision and tachycardia) but also orthostatic hypotension, anorexia, diarrhea, weight gain, tiredness, and decrease of libido. In addition, an important rebound in cataplexy or status cataplecticus may occur after rapid discontinuation of the drug 76. Several other tricyclic antidepressants such as imipramine and desipramine have also been tested to treat cataplexy with less efficacy and same side effects than clomipramine.

Serotonin specific reuptake inhibitors (SSRIs) are much more selective towards the serotonin transporter than tricyclic drugs, but have also clear affinities for other monoamine transporters. Higher doses are often necessary to control cataplexy when compared to antidepressant doses. A few studies reported the efficacy of fluoxetine (20–60 mg/day) 77, fluvoxamine (25–200 mg/day) 78, and citalopram (20–40 mg/day) 79 on cataplexy, but these are all less potent than clomipramine. Side effects are rare but may include sexual dysfunction. A similar risk of rebound cataplexy after a rapid discontinuation of SSRIs can occur. Norepinephrine reuptake inhibitors are also active on cataplexy as reported by a few studies using viloxazine (100 mg/day), reboxetine (2–10 mg/day), or atomoxetine (40–100 mg/day) 80,81. These treatments are generally well tolerated, with only minor adverse effects such as headache, dry mouth, hyperhydrosis, constipation, nausea, and dizziness. Norepinephrine/serotonin reuptake inhibitors such as venlafaxine at 75–300 mg/day are also effective on cataplexy although no well designed study has been published 82. Venlafaxine is commonly a well tolerated drug except for the high risk of nausea.

Mazindol has both anticataplectic and stimulant properties. According to a few studies and to personal experience, mazindol at doses of 1–3 mg/day clearly decreases the number of cataplexy in up to 85% of patients 65,83,84. Potential adverse effects have been reviewed above. Monoamine oxidase inhibitors such as phenelzine and selegiline are also effective in the treatment of cataplexy, in addition to their stimulant effects, as reported in a few studies with patients resistant to conventional therapies 67–69.

Immune‐based therapies such as high‐dose intravenous immunoglobulins (IVIg) have been applied by us and seem to be effective if applied at early stages of narcolepsy‐cataplexy development and act specifically on the number and severity of cataplexy 12,13. These recent findings, obtained only in five patients in open trials, although needing replication in well designed trials, support the efficacy of immune‐modulating treatments on cataplexy, treatments that may also modify the course of the disease.

In conclusion, the first line treatment for cataplexy was previously a low dose of tricyclic antidepressants (especially clomipramine at 10–20 mg/day) or SSRIs, or norepinephrine/serotonin reuptake inhibitor venlafaxine (Table III). However, those medications even widely used had no legal indication on ‘cataplexy’ and had only poor level of evidence for efficacy. We may propose actually, as a first choice treatment for cataplexy, GHB because of its efficacy based on randomized controlled trials (Table III). GHB could become the first line treatment of cataplexy in the near future; however, the important rigorous regulations for this drug will clearly complicate a widely prescription. Second or third line treatment could be norepinephrine reuptake inhibitors reboxetine and atomotexine, mazindol or selegiline sharing the advantage of being both stimulant and anticataplectic drugs (Table III). Finally, there is no established behavioral treatment of cataplexy although patients can learn to avoid situations likely to trigger cataplexy attacks.

Table III. Pharmacological treatment for cataplexy.

Anticataplectics
First‐line treatments:
Gamma‐hydroxybutyrate (6—9 g/day)
Tricyclic antidepressants (at low doses)
Clomipramine (10–20 mg/day)
Protriptyline (5–10 mg/day)
Serotonin reuptake inhibitors
Fluoxetine (20–60 mg/day)
Fluvoxamine (25–200 mg/day)
Citalopram (20–40 mg/day)
Noradrenaline and serotonin reuptake inhibitor
Venlafaxine (75–200 mg/day)
If persistent difficulties:
Tricyclic antidepressants (medium to high doses)
Clomipramine (30–100 mg/day)
Protriptyline (20–80 mg/day)
Noradrenaline reuptake inhibitors
Viloxazine (100–300 mg/day)
Reboxetine (2–10 mg/day)
Atomoxetine (40–100 mg/day)
Mazindol (1–4 mg/day)
Or Selegiline (20–40 mg/day)
Future therapies
IVIg and/or plasmapheresis at disease onset
New monoaminergic reuptake inhibitors
Gamma‐hydroxybutyrate (GHB) long‐release
Hypocretin agonists
Hypocretin cell transplantation

Note: GHB is the first and only medication with indication on ‘cataplexy’. Its efficacy based on randomized controlled trials contrasts with all other medications with only poor level of evidence for efficacy on cataplexy.

Treatment for poor sleep

Several studies with GHB 50–53,70,87 revealed a reduction in the number of nighttime awakenings and in the nocturnal disrupted sleep through enhancement of slow wave sleep. Therefore GHB, at 3–9 g/day, might become the first line treatment for consolidating the disrupted sleep as well. However, clinical experience indicates that regular sleep‐wake schedule, benzodiazepines or nonbenzodiazepine (zolpidem, zopiclone and zaleplon) hypnotics may also be effective in improving fragmented sleep at night. Finally, several patients treated with modafinil alone report rapidly an improvement of their sleep quality at night.

Treatment for other symptoms

Narcolepsy patients often report vivid dreams and REM sleep behavior disorder (RBD). Based on available information it is difficult to provide guidelines, other than to recommend conventional medications for treating parasomnias associated with narcolepsy. The use of clonazepam was frequently reported as successful. Based on several publications, the prevalence of obstructive sleep apnea/hypopnea syndrome is larger in narcolepsy patients than in the general population. Obstructive sleep apnea/hypopnea syndrome should be treated in the same way as in the general population. Periodic limb movements in sleep (PLMS) are also more prevalent in narcolepsy than in the general population 85. However, no documented beneficial effects of treating PLMS on EDS are available in narcolepsy patients. In most cases, there is no special need to treat PLMS in the absence of associated restless legs syndrome. Finally, depression is also reported to be more frequent in narcolepsy patients than in the general population 86. Antidepressant drugs and psychotherapy are indicated as in the nonnarcoleptic depressed patients.

Perspective on therapeutics in narcolepsy

Despite a major advance in our understanding of the neurobiological basis of narcolepsy, its current therapy is only symptomatic. Novel and experimental treatments are needed, and several are under development.

Symptomatic therapies for EDS are currently being developed: 1) R‐isomer of modafinil with a longer duration of action; 2) Histamine H3 antagonist that enhances wakefulness in narcoleptic mice and could be of interest in several types of human hypersomnias; 3) a longer duration of action of GHB; and 4) growth hormone releasing hormone (GHRH) antagonists. In addition, new antidepressants such as duloxetine, reboxetine, and atomoxetine may be of interest for the treatment of cataplexy.

Hypocretin‐based therapies, such as direct use of hypocretin peptides, hypocretin agonists, or hypocretin neuron transplantation, are part of the research projects for the development of new treatments in narcolepsy (so far in animal models).

Based on the autoimmune hypothesis in narcolepsy, immune therapies including IVIg, corticosteroids, and plasmapheresis have been tested in few cases. These treatments have the advantage that if applied at disease onset may favorably and definitely modify the course of the disease. Further randomized controlled trials are under investigation to clarify the beneficial potential of IVIg in narcolepsy.

Conclusion

The contribution of genetic components to sleep disorders is increasingly recognized as important. Narcolepsy is known to occur with higher frequency in certain families and with a higher rate of concordance in monozygotic than in dizygotic twins, thus suggesting the presence of predisposing genetic factors. Although the HLA‐DQB1*0602 remains the unique established genetic risk factor, polymorphisms and/or point mutations in other non‐HLA genes might be linked to narcolepsy and need to be discovered. Future approaches may focus on identifying genetic susceptibility factors that lead to hypocretin neuronal degeneration. Other genes that may explain the response to available treatments need also to be discovered. The clear variation in drug responses between narcolepsy patients should motivate future pharmacogenetic studies.

To date, there has been no cure in narcolepsy but several treatments have been proposed in the last few years. Current management of narcolepsy with cataplexy is based on the available knowledge and clinical experience, but new therapies based on underlying pathophysiological mechanisms (hypocretin‐based and immune therapies) will become available in the near future. Finally, the treatment of narcolepsy in children is still an area in great need of guidelines because of the absence of any available legal drug treatment.

Acknowledgements

The work in YD's laboratory is supported by the Association pour l'Etude du sommeil (Montpellier, France); MT's laboratory is supported by the State of Vaud and the Swiss National Science Foundation.