CLINICAL REVIEW

Genetics of restless legs syndrome

Juliane Winkelmanna,b,*, Luigi Ferini-Strambic

aInstitute of Human Genetics, GSF-National Research Center for Environment and Health,Ingolstadter Landstrasse 1, D-85764 Munich-Neuherberg, GermanybMax Planck Institute of Psychiatry, Munich, GermanycSleep Disorder Center, Vita-Salute San Raffaele University, Milano, Italy

10do

PsEnstTe

(J

KEYWORDSGenetics;Phenotype;Linkage;RLS

87-0792/$ - see front matter Q 2006i:10.1016/j.smrv.2006.01.003

* Corresponding author. Address:ychiatry, Munich and GSF-Nationvironment and Health, Institute ofadter Landstrasse 1, D-85764 Municl.: C49 89 3187 3525; fax: C49 89E-mail address: julianew

. Winkelmann).

Summary Several studies on restless legs syndrome (RLS) have suggested asubstantial genetic contribution in the etiology of this sleep disorder. Clinical surveysof idiopathic RLS patients have shown that 40–90% report a positive family history.The clinical features have been compared between familial and sporadic cases andthe only difference found was a younger age-at-onset in familial RLS.

Despite several reports suggesting a genetic contribution to the etiology ofidiopathic RLS, few molecular genetic studies have been carried out attempting toidentify genes that can predispose to this disorder. In particular, genes encoding forthe GABA A receptor subunits, the gene for the alpha1 subunit of the glycinereceptor, and genes involved in dopaminergic transmission and metabolism havebeen analyzed, however no significant findings have been reported.

Genomewide linkage analysis studies using microsatellite markers have identifiedthree loci for RLS: on chromosome 12q, on chromosome 14q and on chromosome 9p.It is important to investigate whether further RLS families show linkage to one ofthese loci to discuss the contribution of these loci and to provide a prerequisite of amutational screening and identification of the RLS genes.Q 2006 Published by Elsevier Ltd.

Clinical description and familialaggregation

Restless legs syndrome (RLS) is one of the mostcommon neurological disorders with an age depen-dent prevalence ranging from 9 to 14.2% in femalesand 5.4–9.4% in males. A female predominance has

Published by Elsevier Ltd.

Max Planck Institute ofal Research Center forHuman Genetics, Ingol-

h-Neuherberg, Germany.3187 3474.

inkelmann@t-online.de

been demonstrated.1–4 The diagnosis is based on thepresence of four clinical ‘essential criteria’ whichhave been set up in 1995 by the International RLSStudy Group5 and have later been modified.6 Four‘essential criteria’ are obligatory to diagnose thedisease and comprise: (1) an urge to move the legs,usually accompanied or caused by uncomfortableand unpleasant sensation, (2) the symptoms beginor worsen during rest or inactivity, (3) thesymptoms are partially or totally relieved bymovement, such as walking or stretching, at leastas long as the activity continues, and (4) thesymptoms are worse in the evening or at night. Inaddition, the diagnosis is supported by: (1) apositive response to dopaminergic medication, (2)the occurrence of periodic leg movements in sleep(PLMS) and, last but not least, (3) a positive family

Sleep Medicine Reviews (2006) 10, 179–183

www.elsevier.com/locate/smrv

J. Winkelmann, L. Ferini-Strambi180

history. Associated features are the following: (1)the fluctuating progressive clinical course with longperiods of remisson at the beginning, (2) sleepdisturbances due to the leg discomfort and the needto move, and (3) the neurological examination inidiopathic and familial forms of the syndrome isusually normal.

The first description of RLS as a ‘hereditarydisorder’ dates back to Oppenheim in his Textbookof Nervous Diseases published in 1923.7 A strongfamilial aggregation has also been reported sinceEkbom’s original clinical description. Ekbom esti-mated the frequency of hereditary RLS compared tosporadic cases as ‘one-third’ and described familieswith apparent autosomal-dominant inheritance aswell as concordant monozygotic twins.8 Severalstudies investigated the frequency of familial RLSand confirmed the early observations demonstrat-ing that 40–90% of the patients report a positivefamily history.9–12 This large variance is probablyrelated to the different methods used in thesestudies. Not all of them investigated the occurrenceof RLS by personal interviews of relatives of theindex patient in order to verify the diagnosis andthe definite occurrence of familial cases.

Twin studies are a common method to furtherestimate the hereditary component of a disease andto study the significance of genotype and environ-ment interactions. In a study of 12 monozygotictwin pairs, Ondo et al.13 found that 10 of the 12pairs were concordant for RLS symptoms. Despitethe high concordance rate (83%), the RLS severityand age at onset often varied between twins. Inspite of the small sample size and the exclusiveinvestigation of monozygotic twins, this studyfurther underlines the significant genetic contri-bution in RLS. Besides the highly familial componentno diseases causing genes have been identified sofar. On the other hand, the high familiarity makes itlikely that RLS is a heterogeneous phenotype andgenetic heterogeneity complicates the process ofthe identification of disease causing genes.

Familial phenotype

The clinical features, such as age-at-onset, varia-bility of signs and symptoms and associatedfeatures, such as sleep and polysomnographicvariables, have been compared between familialand sporadic cases. Consistently, there was nodifference in the clinical features, except for ayounger age-at-onset in familial RLS: 25.9 vs. 29.2years in the study by Montplaisir et al.,10 and 35.4 vs.47.2 years, respectively, in the study by Winkelmann

et al.11 In the so-called ‘Night-Walker Survey’,Walters et al. interviewed by telephone 105 patientsand subdivided them into two groups: patients withan age-at-onset of !20 years and patients with anage-at-onset O20 years. In the younger group, 81%reported a positive family history, in contrast to theolder group, in which only 58% were aware of otheraffected family members.9 Investigating the age-atonset of the disease, Allen et al.14,15 showed abimodal age-at-onset distribution with a breakbetween 45 years of age. The same authorsalso demonstrated a correlation of an earlier age-at-onset and ‘familial’ RLS.14 Moreover, Allen andEarley compared the effects of current ageand serum ferritin on RLS severity for early- andlate-onset RLS15 by classifying the subjects into twogroups: patients with an age-at-onset!45 years andpatients with an age-at-onsetO45 years. Theauthors demonstrated an age effect in the early-onset form of RLS that slowly progresses with ageand has a limited relation to serum iron. On the otherhand, late-onset RLS appeared to occur lesscommonly in families, progressed rapidly with age,and had a stronger relation to iron status. Takentogether, the results of these studies suggest a newconcept for RLS: the age of symptom onset can beused to better define the RLS phenotype.

Mode of inheritance, and possibilityof anticipation?

Investigations of single families with RLS havesuggested an autosomal dominant mode of inheri-tance with variable expressivity.16–18 Performing aone-factor analysis of variance of ages of onsetbetween generations and segregation ratios calcu-lated for each generation showed an autosomal-dominant mode of inheritance and a male:femaleratio of 1:1.4.19 A formal complex segregationanalysis in a German population of RLS familiesshowed evidence for a single major gene acting in adominant mode of inheritance in early age at onsetfamilies.20 The authors investigated 238 RLSpatients and as many of their first-degree relativesas possible. Assessments were based on direct,personal, standardized diagnostic interviews.Respondents were classified as RLS ‘affected’ or‘non-affected’. Complex segregation analysis wasperformed with the families stratified into twogroups according to the mean age-at-onset of thedisease within the families: age-at-onset upto 30years of age and older than 30 years. In the youngergroup, segregation analysis strongly suggested amajor gene acting as autosomal-dominant with a

Genetics of restless legs syndrome 181

multifactorial component. In contrast, no evidencefor a major gene could be proven in the older group,in which the most likely mode of inheritance was afree transmission probability.20 The segregationpattern found in these families argues for a majorgene autosomal allele acting dominantly in RLSfamilies with an early age-at-onset of symptomsand altogether suggests that RLS is an etiologicallyheterogeneous disease. Similar patterns of inheri-tance have been documented in other diseases withcomplex inheritance, such as Parkinson’s diseaseand Alzheimer’s disease.

The possibility of anticipation has been discussedin four RLS families with a decreasing age at onset inthe following generations.19,21 On the molecularlevel anticipation is caused by an unstable expan-sion of trinucleotides. However, whether thisphenomenon plays a role in RLS is not known. Onehas to consider that the symptoms of the diseaseare better known in the following generations andthe observation of an earlier age of onset could alsobe an artefact.

Interestingly, in this context the co-occurrenceof RLS with diseases known to be caused by unstabletrinucleotide repeat expansions has been observed:RLS is significantly more frequent in patients withSCA322 (Spinocerebellar ataxia) and SCA1 andSCA223 than in the general population. A recentstudy demonstrated that RLS patients in compari-son to controls show an expansion of the trinucleo-tide of SCA3 within the normal range and this makesa causal relationship unlikely for RLS as a whole.24 Itcould be possible, however, that the rare SCA allelecontributes to the RLS phenotype in very rare cases.

Linkage studies

So far, three genomewide linkage analysis usingmicrosatellelite markers have been published andidentified three different loci for RLS. Based on theassumption that the disease follows a recessivemode of inheritance in a French-Canadian familylinkage to chromosome 12q was identified.25 Theauthors discussed the possibility of pseudodomi-nance. This means that due to a high allelefrequency and consequent homo- and heterozygotematings, the pattern of inheritance can appeardominant although the underlying model is reces-sive. It was assumed that 25% of the populationcarry the RLS allele. In such a population, a patientwould have a 1:4 chance of marrying a disease allelecarrier, and each of their children a 25% chance ofthe inheritance of two defective alleles. This is thesame probability like the inheritance in a dominant

trait. The linkage to chromosome 12q could beconfirmed in five further French-Canadianfamilies.26 It could be hypothesized that a foundereffect may play a role in these families.26 TheFrench-Canadian population descends from a smallnumber of founders from the 17th century and canbe regarded as an isolated, genetically homo-geneous population. However, possible phenoco-pies and non-penetrants made it difficult to detecta common segregating haplotype segregating in theFrench-Canadian families.26 The chromosome 12locus was excluded in a family from South Tyrol27

but could not be excluded in 12 RLS familiesoriginating from Germany. The authors suggestedthat it could be a modifier locus in these families.28

Furthermore, the chromosome 12 locus was repli-cated in the icelandic population.29 Altogether thereplication of this locus in independent familiespoints to the robustness of these results and makesit possible that the RLS-genes on chromosome 12contributes not only in the French-Canadian butalso in other populations to the RLS phenotype.

Based on a dominant mode of inheritance asecond locus was identified on chromosome 14q inan Italian RLS family.30 In this study, the PLMS indexwas used to better classify the phenotype of RLS.The authors supported the hypothesis that theoccurrence of PLMS could be a manifestation of amilder RLS phenotype. An independent confir-mation of this locus was identified in one family ofCanadian origin.31 Using a dominant mode ofinheritance, a third locus for RLS was identified intwo American RLS families on chromosome 9p.Possible candidates like the Multi-PDZ DomainProtein and the excitatory amino acid carrier 1(EAAC), a neuronal and epithelial glutamatetransporter were screened for RLS relatedmutations but no disease causing mutation wasfound.32

Investigating 12 RLS families on the known locishowed genetic heterogeneity.28 Altogether it isimportant to investigate whether further RLSfamilies show linkage to one of these loci to discussthe contribution of these loci and to provide aprerequisite of a mutational screening and pos-itional cloning of the RLS genes. Defining the exactcandidate region in all of these studies was difficultdue to possible intrafamilial heterogeneity.Suggested locus heterogenetity seems to play arole in RLS genetics and it appears possible thatseveral genes contain several disease associatedvariants contributing different effects on thephenotype. There is evidence that the genetics ofRLS is likely more a complex disorder witha contribution of multiple genes each addingdifferent effects to the phenotype.

*

*

Practice points

† RLS is a highly familial disorder† Even within a single RLS family the pheno-

type can be very variable† Familial cases have an earlier age-at-onset

comparing to sporadic cases† Three loci for RLS have been published

(RLS1, Chr.12; RLS2, Chr.14; RLS3, Chr.9)under the assumption of a recessive anddominant mode of inheritance

Research agendaIn order to have a prerequisite for identifyingRLS genes we need:

† A better understanding of the variablephenotype

† To investigate the phenotype to be able toclassify subgroups of possible endopheno-types

† Large patient populations and controlswhich are carefully phenotyped for RLS

† Further investigations on RLS families to seewhether they show linkage to one of theknown loci. The knowledge about thecontribution of these loci is a prerequisiteto replicate and confirm these

* The most important references are denoted by an asterisk.

J. Winkelmann, L. Ferini-Strambi182

Association studies

One association study investigated functionalsingle nucleotide polymorphisms (SNPs) withineight genes coding for receptors and enzymesrelated to the dopaminergic transmission (dopa-mine receptors D1–D5, dopamine transporter,tyrosine hydroxylase and dopamine b-hydroxyl-ase).33 An association in 92 French-Canadian RLSpatients and 182 controls was not found. Associ-ation was also negative after stratificationsaccording to age-at-onset or occurrence ofperiodic limb movements. Analysing a subgroupof 44 patients with a positive family historyshowed that those individuals carrying a serineinstead of a glycine allele in the D3 receptordisplay a higher PLM index but this did not reachstatistical significance after correction for mul-tiple testing. Nevertheless, this study does notexclude the causative or modifying role of thegenes investigated. In another study with the samepopulation polymorphisms of two mitochondrialgenes coding for the enzymes Monoaminooxidase Aand B (MAOA and MAOB) were genotyped.34 Theseenzymes are involved in the dopamine catabolismthrough oxidative deamination. A variation of thelength of these alleles is correlated with adifferent enzyme activity. The high activity alleleof the MAOA gene resulted in an elevated enzymeactivity and contributed to the susceptibility ofRLS in females. In contrast, no correlation wasfound with the low activity allele. In males noassociation was found neither with the high norwith the low activity allele. Finally, it has to bementioned that none of these association studieshave been replicated in independent populationsand the samples size were possibly not largeenough to detect subtle effects. Thus, the resultsof these studies should be interpreted withcaution.

However, under the assumption that RLS is morea complex phenotype, association studies usinga case–control design will be of importance inthe future. A prerequisite for such studies is largepatient populations with up to 1000 cases. In thebest case all of these individuals should be ofthe same geoethnical origin and recruited using thesame methods. Furthermore, it is likely thatthe exclusion of RLS cases in any control populationis necessary to detect subtle effects contributing tothe phenotype.

The discovery of RLS genes will give us a betterunderstanding of the pathophysiology of RLS and anidea of underlying molecular mechanisms which arestill completely unknown. However, as the genetics

of RLS turned out to be more and more complicatedin the past, it is hardly predictable when the firstgenes will be identified.

References

1. Berger K, Luedemann J, Trenkwalder C, John U, Kessler C.Sex and the risk of restless legs syndrome in the generalpopulation. Arch Intern Med 2004;164:196–202.

2. Allen RP, Walters AS, Montplaisir J, Hening W, Myers A,Bell TJ, Ferini-Strambi L. Restless legs syndrome prevalenceand impact: REST general population study. Arch Intern Med2005;165(11):1286–92.

3. Hogl B, Kiechl S, Willeit J, Saletu M, Frauscher B, Seppi K,et al. Restless legs syndrome: a community-based study ofprevalence, severity, and risk factors. Neurology 2005;64(11):1920–4.

4. Bjorvatn B, Holsten F, Skeidsvoll H. Prevalence, severity andrisk factors of restless legs syndrome in the general adultpopulation in two Scandinavian countries. Sleep Med 2005;6(4):307–12.

*

*

*

*

*

*

*

*

Genetics of restless legs syndrome 183

5. Walters AS. Toward a better definition of the restless legssyndrome. The International Restless Legs Syndrome StudyGroup. Mov Disord 1995;10(5):634–42.

6. Allen RP, Picchietti D, Hening WA, Trenkwalder C,Walters AS, Montplaisir J. Restless Legs Syndrome Diagnosisand Epidemiology workshop at the National Institutes ofHealth; International Restless Legs Syndrome Study Group.Restless legs syndrome: diagnostic criteria, special consider-ations, and epidemiology. A report from the restless legssyndrome diagnosis and epidemiology workshop at theNational Institutes of Health. Sleep Med 2003;4(2):101–19.

7. Oppenheim H. Lehrbuch der nervenkrankheiten. 7th ed.Berlin: Karger; 1923.

8. Ekbom KA. Restless legs syndrome. Neurology 1960;10:68–873.

9. Walters AS, Hickey K, Maltzman J, Verrico T, Joseph D,Hening W, et al. A questionnaire study of 138 patients withrestless legs syndrome: the ‘Night-Walkers’ survey. Neurol-ogy 1996;46:92–5.

10. Montplaisir J, Boucher S, Poirier G, Lavigne G, Lapierre O,Lesperance P. Clinical, polysomnographic, and geneticcharacteristics of restless legs syndrome: a study of 133patients diagnosed with new standard criteria. Mov Disord1997;12:61–5.

11. Winkelmann J, Wetter TC, Collado-Seidel V, Gasser T,Dichgans M, Yassouridis A, Trenkwalder C. Clinical character-istics and frequency of the hereditary restless legs syndromein a population of 300 patients. Sleep 2000;23:597–602.

12. Bassetti CL, Mauerhofer D, Gugger M, Mathis J, Hess CW.Restless legs syndrome: a clinical study of 55 patients. EurNeurol 2001;45(2):67–74.

13. Ondo W, Jankovic J. Restless legs syndrome: clinicoetiologiccorrelates. Neurology 1996;47:1435–41.

14. Allen RP, LaBuda MC, Becker PM, Earley CJ. Family historystudy of RLS patients from two clinical populations. SleepRes 1997;26:309.

15. Allen RP, Earley CJ. Defining the phenotype of the restlesslegs syndrome (RLS) using age-of-symptom-onset. Sleep Med2000;1:11–19.

16. Walters A, Hening W, Cote L, Fahn S. Dominantly inheritedrestless legs with myoclonus and periodic movements ofsleep: a syndrome related to the endogenous opiates? AdvNeurol 1986;43:309–19.

17. Walters AS, Picchietti D, Hening W, Lazzarini A. Variableexpressivity in familial restless legs syndrome. Arch Neurol1990;47(11):1219–20.

18. Montagna P, Coccagna G, Cirignotta F, Lugaresi E. Familialrestless legs-syndrome: long-term follow-up. In:Guilleminault C, Lugaresi E, editors. Sleep/wake disorders:natural history, epidemiology, and long-term evolution.New York: Raven Press; 1983. p. 231–5.

19. Lazzarini A, Walters AS, Hickey K, Coccagna G, Lugaresi E,Ehrenberg BL, et al. Studies of penetrance and anticipationin five autosomal-dominant restless legs syndrome pedi-grees. Mov Disord 1999;14:111–6.

20. Winkelmann J, Muller-Myhsok B, Wittchen HU, Hock B,Prager M, Pfister H, et al. Complex segregation analysis of

restless legs syndrome provides evidence for an autosomaldominant mode of inheritance in early age at onset families.Ann Neurol 2002;52:297–302.

21. Trenkwalder C, Collado-Seidel V, Gasser T, Oertel WH.Clinical symptoms and possible anticipation in a largekindred of familial restless legs syndrome. Mov Disord1997;11:389–94.

22. Schols L, Haan J, Riess O, Amoiridis G, Przuntek H. Sleepdisturbance in spinocerebellar ataxias—is the SCA3 mutationa cause of restless legs syndrome? Neurology 1998;51:1603–7.

23. Abele M, Burk K, Laccone F, Dichgans J, Klockgether Th.Restless legs syndrome in spinocerebellar ataxia types 1,2,and 3. J Neurol 2001;248:311–4.

24. Desautels A, Turecki G, Montplaisir J, Brisebois K,Desautels AK, Adam B, Rouleau GA. Analysis of CAG repeatexpansions in restless legs syndrome. Sleep 2003;26(8):1055–7.

25. Desautels A, Turecki G, Montplaisir J, Sequeira A, Verner A,Rouleau G. Identification of a major susceptibility locus forrestless legs syndrome on chromosome 12 q. Am J HumGenet 2001;69:1266–70.

26. Desautels A, Turecki G, Montplaisir J, Xiong L, Walters AS,Ehrenberg BL, et al. Restless legs syndrome: confirmation oflinkage to chromosome 12q, genetic heterogeneity, andevidence of complexity. Arch Neurol 2005;62(4):591–6.

27. Kock N, Culjkovic B, Maniak S, Schilling K, Muller B, Zuhlke C,et al. Mode of inheritance and susceptibility locus forrestless legs syndrome on chromosome 12q. Am J HumGenet 2002;71:205–8.

28. Winkelmann J, Lichtner P, Putz B, Trenkwalder C, Hauk S,Meitinger T. Evidence for further genetic locus heterogen-eity and confirmation of RLS-1 in Restless Legs Syndrome.Mov Disord 2006; 21: 28–33

29. Reye DB, Hicks AA, Kristjansson K, Sigmundsson T,Sigurdssson A, Hakonarson H, et al. Population-basedconfirmation of the 12q RLS locus in Iceland. Sleep 2005;28(Abstract supplement):0824.

30. Bonati MT, Ferini-Strambi L, Aridon P, Oldani A, Zucconi M,Casari G. Autosomal dominant restless legs syndrome mapson chromosome 14q. Brain 2003;126:1485–92.

31. Levchenko A, Montplaisir JY, Dube MP, Riviere JB, St-Onge J,Turecki G, et al. The 14q restless legs syndrome locus in theFrench Canadian population. Ann Neurol 2004;55:887–91.

32. Chen S, Ondo WG, Rao S, Li L, Chen Q, Wang Q. Genomewidelinkage scan identifies a novel susceptibility locus for RLS onchromosome 9p. Am J Hum Genet 2004;74:876–89.

33. Desautels A, Turecki G, Montplaisir J, Ftouhi-Paquin N,Michaud M, Chouinard VA, Rouleau GA. Dopaminergicneurotransmission and restless legs syndrome: a geneticassociation analysis. Neurology 2001;57:1304–6.

34. Desautels A, Turecki G, Montplaisir J, Brisebois K,Sequeira A, Adam B, Rouleau GA. Evidence for a geneticassociation between monoamine oxidase A and restless legssyndrome. Neurology 2002;59(2):215–9.