Genetics of Restless Legs SyndromeJuliane Winkelmann, MD

Corresponding authorJuliane Winkelmann, MDHelmholz Center Munich, National Research Center for Environment and Health, Munich Institute of HumanGenetics, Ingolstaedter Landstrasse 1,D-85764 Munich-Neuherberg, Germany.E-mail: julianewinkelmann@t-online.de

Current Neurology and Neuroscience Reports 2008, 8:211–216Current Medicine Group LLC ISSN 1528-4042Copyright © 2008 by Current Medicine Group LLC

Restless legs syndrome (RLS) is a highly familial trait withheritability estimates of about 50%. It is a polygeneticdisorder in which a number of variants contribute to thephenotype. Linkage studies in families with RLS revealedseveral loci but have not yet led to the identificationof disease-causing sequence variants. Phenocopies,nonpenetrance, and possible intrafamilial heterogeneitymake it difficult to define the exact candidate region.Genome-wide association studies identified variantswithin intronic or intergenic regions of MEIS1, BTBD9,and MAP2K5/55 LBOXCOR1// . Carriers of one risk allelehad a 50% increased risk of developing RLS. MEIS1and LBXCOR1 are developmental factors and raise newpathophysiologic questions for RLS. These variants haveweak and moderate effects and increase the risk of devel-oping RLS. It is still possible that strong effects explainthe occurrence of RLS in families. Therefore, linkage andassociation studies should be used congruently to dissectthe complete genetic architecture of RLS.

IntroductionRestless legs syndrome (RLS) is a common neurologic dis-order with an age-dependent prevalence of 5% to 11% in the elderly population of Europe [1] and North America [2].RLS is a behavioral phenotype, and the diagnosis is basedon the report of symptoms by the patient. Diagnostic criteriawere defined by the International RLS Study Group [3] in 1995 and modified in 2003 [4]. There are four obligatorycriteria: 1) an urge to move the legs, usually accompanied or caused by uncomfortable sensations, 2) beginning or wors-ening during periods of rest or inactivity, 3) partial or totalrelief by movement, and 4) onset or worsening in the evening or at night. The diagnosis is further assisted by supportivecriteria that include a positive response to dopaminergics,the occurrence of periodic leg (limb) movements (PLM) dur-

ing wakefulness (PLMW) and sleep (PLMS), and a positive family history. Approximately 80% of RLS patients displayPLMs that are not disease-specific, vary night to night, and are age-dependent [5]. There is a broad phenotypic vari-ability of RLS and about 20% of the affected individualssuffer from severe RLS that requires treatment [6]. L-dopaand dopamine agonists are approved for RLS and provideprompt relief of sensory and motor symptoms [4]. Long-term treatment is limited due to adverse effects such as augmenta-tion [7••]. This complication occurs in RLS patients treated with dopaminergic agents and leads to an overall increase in symptom intensity, a spread of symptoms to other bodyparts, and loss of diurnal rhythmicity [7••].

Phenotype and Genetic ContributionRLS is a highly familial phenotype with heritabilityestimates of about 50%. Depending on the method andpopulation studied, 40% to 60% of RLS patients reporta positive family history [3,5,8,9]. RLS can be second-ary to uremia, iron deficiency, or pregnancy and is also associated with a variety of medical conditions such asmetabolic and hormonal alterations or neuropathy [10].Patients with primary RLS consistently show a higher familial component than patients with sporadic or second-ary RLS, which is often caused by uremia (12% familial) [9] or neuropathy (13% familial) [8].

Further evidence for a genetic contribution to the phe-notype comes from twin studies. Comparisons betweenmonozygotic and dizygotic twins can be used to evalu-ate the genetic and environmental proportion to a trait.A genetic influence is indicated when monozygotic twinshave a higher concordance rate than dizygotic twins. Astudy of 12 monozygotic twins revealed 10 pairs to beconcordant for RLS and pointed to the importance of genetic factors in RLS [11]. Based on a self-report surveyin a cohort of 4503 individual monozygotic and dizygotictwins using two diagnostic questions, the heritability wasestimated to be 0.6 [12]. A more accurate assumptioncomes from a French-Canadian study of 272 monozy-gotic and dizygotic twins who were investigated with avalidated diagnostic interview. There was a higher con-cordance rate in monozygotic than dizygotic twins, thusconfirming a significant genetic component in RLS [13].

Comparing the clinical characteristics of familial andnonfamilial patients, a younger age at onset was a consis-tent finding in familial cases [5,8,9,14]. The course is more

212 Genetics

slowly progressive for early-onset rather than late-onsetRLS [14]. No difference could be identified with respectto other disease characteristics, such as distribution andcharacter of symptoms, diurnal and seasonal fluctuations,aggravating/alleviating factors, presence of growing pains or attention-deficit/hyperactivity disorder during child-hood, different medical conditions, drug treatment, orinsomnia. Also, polysomnographic parameters did notreveal any differences between familial and nonfamilialpatients [5]. Female familial patients more frequently reported a worsening during pregnancy [9].

Families with RLS and Genetic ModelsSince the early 1980s, single families with RLS have been investigated in more detail. A broad intrafamilial pheno-typic variability has been a consistent finding across allfamilies [15]. Symptoms within a single family can be veryvariable, with individuals suffering from RLS only few days of their life up to severely affected persons with pro-found restlessness and insomnia. Age at onset is usuallyin early adulthood, and the course of the disease is either fluctuating or stationary with slow progression. Based onobservations of how RLS segregates in single families, itwas initially suggested that the disease is a monogenic dis-order and follows a mendelian autosomal dominant modeof inheritance with high penetrance [15–17]. One strikingcharacteristic in a subset of RLS families is the numberof affected persons within a single generation, which isfar above the expected 50% assumed for the offspring ina dominant disease. Although an affected status of up to 100% per generation does not necessarily argue againstan occurrence of a dominant mode of inheritance, it ispossible that additional factors contribute to the accumu-lation of affected siblings in the offspring generation.

A more accurate assumption about the underlying genetic model comes from segregation analysis in large populations of RLS families. Investigating the likely modeof inheritance and underlying genetic parameters usingstatistical methods in RLS families with a total of 900members suggested a bimodal distributed mode of inheri-tance depending on the age at onset [18]. Families with an earlier age at onset showed evidence for the involvement of a major gene segregating in a dominant fashion, in additionto a multifactorial component, which adds further geneticvariance to the model. A familial aggregation analysis tried to quantify the total genetic contribution in 15 large NorthAmerican RLS families with a total of 453 members. It esti-mated a heritability of 0.6 (which indicates again that RLS is a highly heritable trait) [19]. A single-locus, mendelian-dominant model with a complete penetrance had the bestfit in an analysis of American families, with two underlying distributions of age of onset and a possible dichotomy at26.3 years. Lack of evidence for a major gene controlling age of onset indicates that nongenetic causes of RLS mayexist and that RLS is a complex disorder [20].

Linkage StudiesLinkage studies are family based and analyze a cosegregationwith a genetic marker and a specific phenotype. Usually, this method is used to investigate monogenic mendelian traits.

Based on the assumption that the disease follows arecessive mode of inheritance, a French-Canadian family showed linkage to a region on chromosome 12q (RLS1)[21]. The genetic model used included a high frequency of disease alleles. As a consequence, the possibility of mat-ings among homozygous and heterozygous individualscould lead to the statistical occurrence of 50% RLS occur-rence in the offspring. This is normally the likelihood in dominant models and, therefore, is called a pseudodomi-nant pattern of inheritance. Five additional families of French-Canadian origin also showed linkage to this locusand further narrowed down the linkage interval [22]. The existence of RLS1 is further supported by a linkage signalin the Icelandic population [23] and in a subset of Ger-man families using the transmission disequilibrium test [24]. The confirmation within additional Canadian fami-lies did not clearly overlap with the original linkage peak,and analysis of the families did not reveal a consistenthaplotype segregating in all families [22]. This illustrates the problems of RLS linkage studies. It is likely that phe-nocopies and nonpenetrants can make it difficult to detect a common segregating haplotype shared between families[22], and defining the exact candidate region is difficultdue to intrafamilial, allelic, and nonallelic heterogeneity.

RLS2 on chromosome 14q13-21 was mapped in afamily of northern Italian origin and confirmed in a fam-ily of Canadian origin [25,26]. RLS3 on chromosome9p24-22 was identified by the investigation of 15 Americanfamilies that included 134 people with RLS [19]. Under the assumption of intrafamilial heterogeneity and stratifica-tion according to a phenotype with an early age at onset, this region was narrowed by investigating a German fam-ily [27]. The robustness of RLS3 is further supported by a recent finding in a second family originating from Germany that is linked to a region overlapping the centromeric part of RLS3 [28,29]. RLS4 was identified on chromosome2q33 in three South Tyrolean families descending from a common founder using dominant parametric and nonpara-metric linkage analysis [30]. RLS5 on chromosome 20p13was detected again in a French-Canadian pedigree using anautosomal dominant model [31]. Suggestive evidence for linkage comes from another family of Italian origin map-ping to 19q [32]. Finally, two further loci were identifiedon chromosome 4q and chromosome 17p in a single RLS family originating from Bavaria [33], under the assump-tion of an autosomal dominant model. The latter findingwas replicated in an independent, family-based linkage andassociation study in the EU-RLS-GENE trios originatingfrom eight European countries [33].

RLS2 and RLS3 were investigated by performing afamily-based association study of European RLS trios.Transmission disequilibrium tests were employed to

Genetics of Restless Legs Syndrome Winkelmann 213

analyze the data and showed a significant association to RLS2. In a subgroup analysis including only southern and central European trios, marginally significant asso-ciations were found to chromosome 9p (P = 0.0086 and P= 0.0077, respectively). These variable results observed infamilies of different origin further corroborate the geneticcomplexity of RLS [34], and the large number of loci iden-tified strongly point to genetic heterogeneity. So far, all of these genome-wide linkage analyses of RLS families havemet with limited success and have not led to the identi-fication of a disease-causing gene. Sequencing of genes annotated in the region did not lead to the identification of a disease-causing mutation. Further families have beeninvestigated, but analyses have either failed or resulted inonly borderline-significant logarithm of odds scores. Thefinding that the highest logarithm of odds scores achievedwere below the simulated scores in the respective familiesprovided indirect evidence for the complexity of RLS.

Association StudiesAssociation studies analyze the frequencies of alleles atthe site of interest and compare them in a case and controlsample (although family-based designs can also be used).A higher frequency of the allele tested in cases is taken asevidence that the allele or genotype is associated with an increased risk for the disease [35].

As a first approach, several candidate genes havebeen investigated. A candidate gene is selected either bya location in a linkage region or because the product hasbiochemical or other reasonable function in connectionto the assumed pathophysiology of the disorder. Basedon the improvement of RLS symptoms by dopaminergicagents, association studies investigated putative functionalsingle nucleotide polymorphisms (SNPs) within eightgenes coding for receptors and enzymes related to thedopaminergic transmission (dopamine receptors D1–D5,dopamine transporter, tyrosine hydroxylase, and dopa-mine -hydroxylase) [36]. No association was found whencomparing 92 French-Canadian RLS patients with 182 controls of the same origin. Within the same population,polymorphisms of genes encoding monoamine oxidase Aand B were genotyped [37]. These enzymes are involved in the dopamine catabolism through oxidative deamina-tion. Certain polymorphisms of these genes are correlatedwith different enzyme activities [37]. An association to the high-activity allele of the monoamine oxidase A in women could contribute to the susceptibility of RLS. Incontrast, no association was found with the high-activityor low-activity allele in men. The number of individuals investigated for such a heterogeneous disease was small,and replication in multiple independent samples has notbeen performed so far.

A systematic, hypothesis-free approach was performedin a case-control study in 918 independent cases and con-trols of European ancestry. A total of 1536 SNPs in 366

genes encompassing a 21-megabase region of the RLS1 locus were analyzed and significant association with theneuronal nitric oxide synthase (NOS1) was identified. Different allele frequencies with opposite directions werefound after analyzing single-associated SNPs within theNOS1 gene [38]. This implies that the same allele is a risk allele in one but a protective allele in the other sample. Therefore, further studies in independent populationsare needed to replicate and confirm this finding [39].The association of variants in the NOS1 gene and RLS, however, suggests involvement of the nitric oxide/argininepathway in the pathogenesis of RLS.

Current technology has made large-scale, high-density, genome-wide association studies a reality. These studiescombine the power of association studies with the system-atic nature of a genome-wide search. Two studies have taken advantage of this new methodology. In a study conductedin 1600 RLS patients and 2600 controls in the German and Canadian populations, three genomic regions havebeen identified encoding the genes MEIS1 and BTBT, andTTanother region has been identified that encodes the genesMAP2K5 and LBXCOR1 [40••]. Association was identi-fied with intronic variants, which suggests a functional rolein the expression or alternative splicing of the gene. Carri-ers of one risk allele had a 50% increased risk of developing RLS. Theoretically, carriers of all risk alleles have a 20-foldincreased risk of developing RLS. One haplotype was evenmore strongly associated with RLS (odds ratio > 2). In asimilar study conducted in the Icelandic and US populations,association was found with the identical variant in BDBD9[41••]. Both studies used different assessments of the phe-notype. In the German/Canadian study, all individualswere investigated by a face-to-face interview conducted bya neurologist. To minimize phenotypic heterogeneity, onlypatients with a positive family history (defined as having one first-degree family member with RLS) were included.The RLS patients were individuals who came to specializedRLS clinics and asked for treatment of their symptoms. The Icelandic study took advantage of a standardized diagnosticinterview conducted by a study nurse. In a preceding vali-dation study of the interview, a false-positive rate of about20% was elucidated. A combination of parameters was usedto define the phenotype more accurately. The presence of PLMs was used to support the diagnosis of RLS. The asso-ciation of BTBD9 was based on PLM-positive individuals. This still does not answer the question of whether PLM (aswell as PLMS and PLMW) should be included as a diagnos-tic criterion for RLS in genetic studies. A PLMS index greaterthan 5 is commonly considered pathologic and used as anobjective parameter to confirm the diagnosis. PLMs can alsooccur in other disorders, such as in isolated PLM disorder,or even in individuals without any sleep complaints. PLMs have also been documented in association with narcolepsy, Parkinson’s disease, or rapid eye movement–sleep behaviordisorder. It remains to be investigated whether this BTBD9 gene variant is also associated with individuals with these

214 Genetics

disorders and displaying PLMs. This could finally help to differentiate whether BTBD9 is more associated with PLMrather than RLS. Furthermore, an analysis of parametersinvolved in iron metabolism revealed that the risk allele wasassociated with a 13% decrease of the serum ferritin level. Whether or not this variant is also associated with serum ferritin levels in the German-Canadian study is not known.

A closer inspection of the known function of thegenes identified is surprising because some of them aredevelopmental factors and challenge our previous patho-physiologic concept of RLS. MEIS1 is a member of afamily of highly conserved TALE homeobox transcriptionfactors. Heterodimers of MEIS1 with PBX and HOX pro-teins augment the affinity and specificity of DNA bindingby HOX proteins [40••]. HOX genes are organized inclusters and are expressed along the body axis in a mannercorresponding to their position along the chromosome.Mutation in HOX genes results in morphologic trans-formation of the segmental structures in which a specificgene is normally expressed [42]. MEIS1 is part of a tran-scriptional regulatory network that specifies spinal motorneuron pool identity and connectivity and therefore may have a function in the motor part of RLS or PLM [43].A study in Xenopus showed that MEIS1 is also involved in neural crest development [44]. Pathologically, MEIS1is overexpressed in acute myeloid leukemia. Togetherwith HOXA9, it is a strong oncogenic collaborator in the induction and maintenance of acute myeloid leukemia.

The third region encoded the MAP2K5 gene, a mem-ber of the mitogen-activated protein kinase family, and the adjacent LBXCOR1 gene. LBXCOR1 is annotateddownstream of MAP2K5 and acts as a transcriptionalcorepressor of LBX1. This homeobox gene plays a critical role in the development of sensory pathways in the dorsalhorn of the spinal cord, which relays pain and touch [45]. Based on its function, it is more likely LBX1 is involved in RLS. Both genome-wide association studies showed asso-ciation to BTBD9. Little is known about BTBD9 otherthan that it belongs to the BTB(POZ) domain–containingproteins. In Drosophila, BTB (broad complex, tramtrack, and bric à brac) genes are required for embryonic develop-ment, cell fate determination in the eye, metamorphosis,and pattern formation in the limbs [46] Functions of BTB(POZ) proteins include transcription repression,cytoskeleton regulation, tetramerization and gating of ionchannels, and protein ubiquitination/degradation [46].The modular nature of this protein and the universaloccurrence of the particular domains of BTBD9 makeassignment of a specific function difficult at present.

One important factor for further genetic studies inRLS is the availability of large and thoroughly phenotyped patient samples. Analysis of endophenotypes can give abetter understanding about the genes involved in associa-tion with a specific symptom of RLS. It is very likely thatinvestigations in a larger sample size are going to identifyfurther variants associated with RLS. Analysis of pooled

data will provide the basis for a combined analysis of several thousand samples genome-wide. This will allow researchers to identify even smaller effects. The European RLS Study Group has initiated a project named EU-RLS-GENE, whose aim is to identify RLS susceptibility genesby investigating RLS families, sibling pairs, and trios.

Is RLS a Neurodegenerative Disorder?The phenomenon of anticipation refers to a younger age at onset of the disease and a more severe phenotype in sub-sequent generations. Anticipation can correlate with anexpansion of an unstable oligonucleotide repeat sequence(eg, CAG) in the coding or noncoding region of a gene.Based on the observation of a decreasing age at onset insubsequent generations, it was suggested that RLS maybe a repeat expansion disorder [47,48]. RLS also occurswith diseases known to be caused by trinucleotide repeatexpansions. RLS is significantly more frequent in patientswith spinocerebellar ataxia type 3 (SCA3) [49], SCA1, andSCA2 [50] than in the general population. The frequencyof neuropathy did not differ between SCA patients with or without RLS, suggesting that RLS is not simply a second-ary effect due to neuropathy [51]. In a single family with SCA3, an association was observed between RLS and anintermediate length of the CAG repeat (mean: 53.4 repeats; normal: 12–44 repeats; pathologic: 60–85 repeats) [52]. But a comparison of RLS patients and controls showed an expansion of the trinucleotide in the SCA3 gene within thenormal range, rendering a causal relationship unlikely [53].A comparison of expansions within the SCA1 or SCA2genes has never been investigated. It is still possible thatthe trinucleotide expansion displays a weak effect of a rareallele in very rare cases of RLS.

A parkin gene mutation was found in 10 of 20 patients in two possibly related families with idiopathic RLS butwas not considered causative. The clinical phenotype didnot differ between RLS patients with and without a par-kin mutation [54].

ConclusionsThe identification of genetic risk variants has been alarge step forward in unraveling the genetics of RLS. Aquestion to solve is whether RLS has components of adevelopmental disorder and whether the genes identifiedplay a role in early embryonic days or have a completely different function in the elderly and in association withRLS. It is now possible to investigate the relation of thedopaminergic system and to better understand the modeof action of dopaminergic drugs in RLS.

A possible concept for the disease is whether specificnongenetic triggering factors, such as environment orlifestyle, are required to develop RLS in an individualthat possesses a genetic susceptibility. This could applyto women who develop RLS only during their pregnancy.

Genetics of Restless Legs Syndrome Winkelmann 215

Secondary cases of RLS, such as uremic RLS patients, may present genetically susceptible individuals with aclearly defined provoking factor. The variants identified confer small to moderate relative risk. MEIS1 is more associated with familial cases, but the occurrence of RLS in families is not fully explained. The identification of a large monogenic effect responsible for RLS in families isanother challenge in the research of RLS. Therefore, link-age and association studies should be used congruently infuture approaches to further dissect the genetic architec-ture of RLS.

An extended knowledge of endophenotypes anddetailed phenotyping would give us the possibility of stratifying and relating the variants to distinct features of the phenotype. The discovery of RLS genes will notonly provide new insights into the pathophysiology of this disorder, but will presumably also increase our understanding of the sleep-wake rhythms as well as othermovement disorders and the interplay of both.

DisclosureNo potential conflict of interest relevant to this article was reported.

References and Recommended ReadingPapers of particular interest, published recently, have been highlighted as:• Of importance•• Of major importance

1. Berger K, Luedemann J, Trenkwalder C, et al.: Sex and the risk of restless legs syndrome in the general population. Arch Intern Med 2004, 164:196–202.

2. Phillips B, Young T, Finn L, et al.: Epidemiology of restless legs symptoms in adults. Arch Intern Med 2000, 160:2137–2141.

3. Walters AS, and the International Restless Legs SyndromeStudy Group: Toward a better definition of the restless legssyndrome. Mov Disord 1995, 10:634–642.

4. Allen RP, Picchietti D, Hening WA, et al.; Restless LegsSyndrome Diagnosis and Epidemiology Workshop at theNational Institutes of Health, International Restless Legs Syndrome Study Group: Restless legs syndrome: diagnosticcriteria, special considerations, and epidemiology. A reportfrom the Restless Legs Syndrome Diagnosis and Epidemiol-ogy Workshop at the National Institutes of Health. Sleep Med 2003, 4:101–119.

5. Montplaisir J, Boucher S, Poirier G, et al.: Clinical, polysomnographic, and genetic characteristics of restlesslegs syndrome: a study of 133 patients diagnosed with newstandard criteria. Mov Disord 1997, 12:61–65.

6. Allen RP, Walters AS, Montplaisir J, et al.: Restless legs syndrome prevalence and impact: REST general population study. Arch Intern Med 2005, 165:1286–1292.

7.•• García-Borreguero D, Allen RP, Kohnen R, et al.; Inter-national Restless Legs Syndrome Study Group: Diagnosticstandards for dopaminergic augmentation of restless legs syndrome: report from a World Association of SleepMedicine-International Restless Legs Syndrome StudyGroup consensus conference at the Max Planck Institute.Sleep Med 2007, 8:520–530.

This is an important article.

8. Ondo W, Jankovic J: Restless legs syndrome: clinicoetio-logic correlates. Neurology 1996, 47:1435–1441.

9. Winkelmann J, Wetter TC, Collado-Seidel V, et al.: Clinicalcharacteristics and frequency of the hereditary restless legssyndrome in a population of 300 patients. Sleep 2000,23:597–602.

10. Trenkwalder C, Paulus W, Walters AS: The restless legs syndrome. Lancet Neurol 2005, 4:465–475.

11. Ondo WG, Vuong KD, Wang Q: Restless legs syndrome inmonozygotic twins: clinical correlates. Neurology 2000, 55:1404–1406.

12. Desai AV, Cherkas LF, Spector TD, Williams AJ: Genetic influences in self-reported symptoms of obstructive sleepapnoea and restless legs: a twin study. Twin Res 2004, 7:589–595.

13. Xiong L, Jang K, Montplaisir J, et al.: Canadian restlesslegs syndrome twin study. Neurology 2007, 68:1631–1633.

14. Allen RP, Earley CJ: Defining the phenotype of the restlesslegs syndrome (RLS) using age-of-symptom-onset. SleepMed 2000, 1:11–19.

15. Walters AS, Picchietti D, Hening W, Lazzarini A: Variableexpressivity in familial restless legs syndrome. Arch Neurol 1990, 47:1219–1220.

16. Trenkwalder C, Seidel VC, Gasser T, Oertel WH: Clinicalsymptoms and possible anticipation in a large kindred of familial restless legs syndrome. Mov Disord 1996, 11:389–394.

17. Montplaisir J, Godbout R, Boghen D, et al.: Familialrestless legs with periodic movements in sleep: electrophysi-ologic, biochemical, and pharmacologic study. Neurology 1985, 35:130–134.

18. Winkelmann J, Muller-Myhsok B, Wittchen HU, et al.:Complex segregation analysis of restless legs syndromeprovides evidence for an autosomal dominant mode of inheritance in early age at onset families. Ann Neurol 2002,52:297–302.

19. Chen S, Ondo WG, Rao S, et al.: Genomewide linkagescan identifies a novel susceptibility locus for restless legssyndrome on chromosome 9p. Am J Hum Genet 2004, 74:876–885.

20. Mathias RA, Hening W, Washburn M, et al.: Segregation analysis of restless legs syndrome: possible evidence for amajor gene in a family study using blinded diagnoses. Hum Hered 2006, 62:157–164.

21. Desautels A, Turecki G, Montplaisir J, et al.: Identification of a major susceptibility locus for restless legs syndrome onchromosome 12q. Am J Hum Genet 2001, 69:1266–1270.

22. Desautels A, Turecki G, Montplaisir J, et al.: Restless legssyndrome: confirmation of linkage to chromosome 12q, genetic heterogeneity, and evidence of complexity. Arch Neurol 2005, 62:591–596.

23. Hicks AA, Rye DB, Kristjansson K, et al.: Population/basedconfirmation of the 12q RLS locus in Iceland [abstract]. Mov Disord 2005, 20:P117.

24. Winkelmann J, Lichtner P, Putz B, et al.: Evidence for fur-ther genetic locus heterogeneity and confirmation of RLS-1 in restless legs syndrome. Mov Disord 2006, 21:28–33.

25. Bonati MT, Ferrini-Strambi L, Aridon P, et al.: Autosomal dominant restless legs syndrome maps on chromosome 14q.Brain 2003, 126:1485–1492.

26. Levchenko A, Montplaisir JY, Dube MP, et al.: The 14q restless legs syndrome locus in the French Canadian popula-tion. Ann Neurol 2004, 55:887–891.

27. Liebetanz KM, Winkelmann J, Trenkwalder C, et al.:RLS3: fine-mapping of an autosomal dominant locus in a family with intrafamilial heterogeneity. Neurology 2006,67:320–321.

28. Lohmann-Hedrich K, Neumann A, Kleensang A, et al.:Evidence for linkage of restless legs syndrome to chromo-some 9p: are there two distinct loci? Neurology 2008, 70:686–694.

216 Genetics

29. Winkelmann J, Muller-Myhsok B: Genetics of restless legssyndrome: a burning urge to move [editorial]. Neurology 2008, 70:664–665.

30. Pichler I, Marroni F, Volpato CB, et al.: Linkage analysisidentifies a novel locus for restless legs syndrome on chromosome 2q in a South Tyrolean population isolate. AmJ Hum Genet 2006, 79:716–723.

31. Levchenko A, Provost S, Montplaisir JY, et al.: A novelautosomal dominant restless legs syndrome locus maps tochromosome 20p13. Neurology 2006, 67:900–901.

32. Kemlink D, Plazzi G, Vetrugno R, et al.: Suggestive evi-dence for linkage for restless legs syndrome on chromosome 19p13. Neurogenetics 2008 Jan 10 (Epub ahead of print).

33. Winkelmann J, Lichtner P, Kemlink D, et al.: New loci forrestless legs syndrome map to chromosome 4q and 17p[abstract]. Mov Disord 2006, P304.

34. Kemlink D, Polo O, Montagna P, et al.: Family-based association study of the restless legs syndrome loci 2 and 3in a European population. Mov Disord 2007, 22:207–212.

35. Hirschhorn JN, Lohmueller K, Byrne E, Hirschhorn K: Acomprehensive review of genetic association studies. Genet Med 2002, 4:45–61.

36. Desautels A, Turecki G, Montplaisir J, et al.: Dopaminergic neurotransmission and restless legs syndrome: a genetic association analysis. Neurology 2001, 57:1304–1306.

37. Desautels A, Turecki G, Montplaisir J, et al.: Evidence for a genetic association between monoamine oxidase A andrestless legs syndrome. Neurology 2002, 59:215–219.

38. Winkelmann J, Lichtner P, Schormair B, et al.: Variants in the neuronal nitric oxide synthase (nNOS, NOS1) gene areassociated with restless legs syndrome. Mov Disord 2008,23:350–358.

39. Lin PI, Vance JM, Pericak-Vance MA, Martin ER: No geneis an island: the flip-flop phenomenon. Am J Hum Genet 2007, 80:531–538.

40.•• Winkelmann J, Schormair B, Lichtner P, et al.: Genome-wide association study of restless legs syndrome identifiescommon variants in three genomic regions. Nat Genet 2007, 39:1000–1006.

This is an important article.41.•• Stefansson H, Rye DB, Hicks A, et al.: Genetic risk factor

for periodic limb movements in sleep. N Engl J Med 2007, 357:639–647.

This is an important article.

42. Capecchi MR: Hox genes and mammalian development. Cold Spring Harb Symp Quant Biol 1997, 62:273–281.

43. Dasen JS, Tice BC, Brenner-Morton S, Jessell TM: A Hox regulatory network establishes motor neuron pool identity and target-muscle connectivity. Cell 2005, 123:477–491.

44. Maeda R, Mood K, Jones TL, et al.: Xmeis1, a protoonco-gene involved in specifying neural crest cell fate in Xenopusembryos. Oncogene 2001, 20:1329–1342.

45. Gross MK, Dottori M, Goulding M: Lbx1 specifies somato-sensory association interneurons in the dorsal spinal cord.Neuron 2002, 34:535–549.

46. Stogios PJ, Downs GS, Jauhal JJ, et al.: Sequence andstructural analysis of BTB domain proteins. Genome Biol 2005, 6:R82.

47. Lazzarini A, Walters AS, Hickey K, et al.: Studies of penetrance and anticipation in five autosomal-dominant restless legs syndrome pedigrees. Mov Disord 1999,14:111–116.

48. Trenkwalder C, Collado-Seidel V, Gasser T, Oertel WH:Clinical symptoms and possible anticipation in a largekindred of familial restless legs syndrome. Mov Disord 1997, 11:389–394.

49. Schols L, Haan J, Riess O, et al.: Sleep disturbance in spinocerebellar ataxias—is the SCA3 mutation a cause of restless legs syndrome? Neurology 1998, 51:1603–1607.

50. Abele M, Bürk K, Laccone F, et al.: Restless legs syndromein spinocerebellar ataxia types 1, 2 and 3. J Neurol 2001, 248:311–314.

51. Reimold M, Globas C, Gleichmann M, et al.: Spinocerebel-lar ataxia type 1, 2, and 3 and restless legs syndrome: striatal dopamine D2 receptor status investigated by[11C]raclopride positron emission tomography. Mov Disord 2006, 21:1667–1673.

52. Van Aften J, Sinke RJ, Twats MJ, et al.: Intermediate CAG repeat lengths (53.54) for MJD/SCA3 are accociated withan abnormal phenotype. Ann Neurol 2001, 49:805–808.

53. Desautels A, Turecki G, Montplaisir J, et al.: Analysis of CAG repeat expansions in restless legs syndrome. Sleep2003, 26:1055–1057.

54. Adel S, Djarmati A, Kabakci K, et al.: Co-occurrenceof restless legs syndrome and Parkin mutations in twofamilies. Mov Disord 2006, 21:258–263.