Pedigree of the Texas family studied. Squares indicate male members; circles, female members; solid filled shapes, individuals with gene mutation; dotted shapes, individuals undergoing testing who had negative findings for mutation; and slashed shapes, deceased family members. Arrow indicates proband.
Analysis of the mutation of the guanosine triphosphate–cyclohydrolase (GCH1) gene. A, Sequence profile of the wild-type (wt) allele and that of the mutated (del) allele of the proband. The deleted 37 base pairs (bp) in exon 2 of the GCH1 gene are boxed in the wt sequence and marked with an arrow in the del allele. B, Polymerase chain reaction analysis of genomic DNA samples from individuals of the family (shown are III:3, IV:5, and IV:6 of the family pedigree). The primer set for exon 2 was used. The wt gene and the del gene migrate as a 315- and 278-bp band, respectively. Pedigree symbols are explained in the legand to Figure 1. C, Predicted translational reading frames of wt and del alleles of affected individuals. The 37-bp deletion in exon 2 leads to a shift of the reading frame, which results in a premature termination codon after amino acid 159. In boldface type are the deleted DNA sequence (marked in the wt allele) and the predicted amino acid sequence resulting from the deletion.
Hahn H, Trant MR, Brownstein MJ, Harper RA, Milstien S, Butler IJ. Neurologic and Psychiatric Manifestations in a Family With a Mutation in Exon 2 of the Guanosine Triphosphate–Cyclohydrolase Gene. Arch Neurol. 2001;58(5):749-755. doi:10.1001/archneur.58.5.749
Copyright 2001 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2001
To investigate the range of clinical features to correlate genotypic and phenotypic manifestations in hereditary progressive and/or levodopa-responsive dystonia due to a defect in the guanosine triphosphate–cyclohydrolase (GCH1) gene.
Design and Setting
A large family from Texas was studied in an ambulatory setting by clinicians in genetics, neurology, and psychiatry using structured interviews and examinations.
The family was selected after neurometabolic investigations of a young boy (proband) with foot dystonia and fatigue and his father, who had a long history of anxiety and depression. Results of metabolic studies showed decreased levels of metabolites of biopterin and biogenic amines in cerebrospinal fluid. Subsequently, a novel mutation (37–base pair deletion) in exon 2 of the GCH1 gene was demonstrated in 11 family members. There was no observed female sex bias, but there was a wide variability of motor dysfunctions in family members. Approximately 50% had clinical deafness and a similar number had significant psychiatric dysfunction, including depression and anxiety.
Study of additional families with hereditary progressive and/or levodopa-responsive dystonia using modern molecular methods will be necessary to confirm the neuropsychiatric spectrum of this disorder, in which important clinical features may be unrecognized and thus inappropriately managed.
FOLLOWING recognition that some patients with a clinically variable dystonic movement disorder (autosomal dominant hereditary progressive dystonia1) responsive to levodopa (levodopa-responsive dystonia2) had a single gene disorder with mutational defects in the guanosine triphosphate (GTP)–cyclohydrolase I (GCH1) gene,3 there has been considerable interest in understanding genotypic-phenotypic correlations using molecular methods.4,5
Guanosine triphosphate–cyclohydrolase I (GCH1) is the rate-limiting enzyme that catalyzes the first step in the biosynthesis of tetrahydrobiopterin (BH4), the natural cofactor for 3 aromatic amino acid monooxygenases. These are tyrosine hydroxylase and tryptophan hydroxylase, the rate-limiting enzymes in dopamine (DA) and serotonin (5-HT) biosynthesis, respectively, and phenylalanine hydroxylase, which is involved in phenylalanine metabolism. Mutations in the GCH1 gene have been identified in the following 3 clinically different neurometabolic disorders: (1) autosomal dominant hereditary progressive and/or levodopa-responsive dystonia2 that is characterized by childhood-onset dystonia with sustained clinical responsiveness to low doses of levodopa; (2) autosomal recessive GCH1-deficient hyperphenylalaninemia6 presenting in the first 6 months of life with a severe neurologic disorder (psychomotor retardation, convulsions, truncal hypotonia, and limb hypertonia); and more recently, (3) compound heterozygote mutations of the GCH1 gene with a neurologic disorder intermediate in severity between the above disorders.7 Biochemical studies of patients and families with recessive mutations of the GCH1 gene (homozygous or compound heterozygous) have demonstrated severe or moderately severe defects in BH4metabolism that correlate with the severity of neurologic symptoms, low biogenic amine metabolite levels in cerebrospinal fluid (CSF), and partial responsiveness to neurotransmitter precursors (levodopa and 5-hydroxytryptophan) and cofactor administration. Patients with autosomal dominant levodopa-responsive dystonia have a dystonic movement disorder without mental retardation or convulsions. Biochemically, they have a milder defect in biogenic amine and BH4metabolism (based on results of CSF studies) than patients with the recessive form of the disease. With levodopa administration, the responsiveness of the motor dysfunctions in these patients is dramatically improved. A recent report documented that autosomal dominant levodopa-responsive dystonia is a disorder with a high degree of penetrance. The expressivity of this disorder, however, shows marked interfamilial and intrafamilial variability.4
In this study, we investigated the neurologic and biochemical as well as molecular variables in members of a large Texas family with an unusual frequency of neurologic and psychiatric symptoms. The starting point of the investigation was a young boy (the proband) with variable foot dystonia and fatigue, and his father, who had a long history of anxiety and depression. Low CSF levels of metabolites of DA (homovanillic acid [HVA]), 5-HT (5-hydroxyindoleacetic acid [5-HIAA]), norepinephrine (NE) (3-methoxy-4-hydroxy phenylethylene glycol [MHPG]), and tetrahydrobiopterin (neopterin and biopterin) were measured in both subjects. Molecular analysis demonstrated a heterozygous mutation in the GCH1 gene in both subjects. Other members of their large Texas family were invited to participate in these studies. A total of 11 members (8 male and 3 female) were identified as having a mutation in exon 2 of the GCH1 gene. All members with this mutation underwent a detailed neurologic and psychiatric history and examination to delineate the spectrum of neurologic and psychiatric manifestations in this single family with this unique mutation in the GCH1 gene (Table 1).
Once the proband and his father were identified, this large Texas family of more than 70 members was recruited after formal consent was obtained and after review of the clinical and molecular protocols by an institutional review board of the University of Texas–Houston Medical School. Paternal and maternal family members were of English Quaker origins from Pennsylvania. The family recognized that Parkinson disease was common in older members and that psychiatric manifestations, including anxiety, depression, obsessive-compulsive traits, and eating disorders, were present in family members. Once a mutational deletion (37 base pairs [bp]) was demonstrated in exon 2 of the GCH1 gene in several family members, the family underwent systematic evaluation, and the 11 genetically affected members were enrolled in a detailed study of personal, medical, social, and family history by a genetic counselor (M.R.T.), with a standardized neurologic and psychiatric assessment by a board-certified neurologist (I.J.B.) and psychiatrist (R.A.H.), respectively. A complete pedigree was obtained (Figure 1), and none of the family members carrying the gene mutation was born of a consanguineous union. Informed consent was obtained and most patients were examined (and videotaped) at home. All subjects underwent a clinical psychiatric interview. A psychiatrist board certified in general as well as child and adolescent psychiatry conducted this interview (R.A.H.). The interview screened for major psychiatric disorders and was adapted from the questions on the Schedule for Affective Disorders and Schizophrenia8 and the Schedule for Affective Disorders and Schizophrenia for School-age Children.9 Further symptoms were obtained from subjects who responded positively to the screening questions to determine if they met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria for a particular psychiatric disorder.10 Each subject also received the Mini-Mental State Examination11 to screen for cognitive difficulties. Coded samples of venous blood were obtained for genomic DNA preparation from peripheral leukocytes using a standard method (Puregene kits; Gentra Systems, Minneapolis, Minn). Member III:1 did not wish to be studied.
Samples of CSF obtained at lumbar puncture were collected on ice and stored at −80°C until analyzed. Samples were collected in the morning in glass tubes (2 mg of ascorbic acid per milliliter of CSF), and concentrations of neurotransmitter metabolites HVA, 5-HIAA, and MHPG were determined by means of a modification of a gas chromatography–mass spectroscopy procedure.12 Total CSF neopterin and biopterin levels (oxidized plus reduced forms) were determined by means of high-performance liquid chromatography with fluorescence detection after oxidation as previously described.13
For mutational analysis, exons 1 through 6, including splice junctions, were amplified from genomic DNA using polymerase chain reaction (PCR) analysis. Primer sequences used for amplification of exons were the same as those reported previously.14 At first, a combined single-stranded conformational polymorphism (SSCP) and heteroduplex analysis approach was used under optimized conditions for a subset of DNA samples of family members. The DNA samples were amplified in PCR buffer containing 2.2-mmol/L magnesium chloride and α-phosphate 32–labeled deoxycytidine-5′-triphosphate for 35 cycles at 95°C for 45 seconds, 58°C for 1 minute 30 seconds, and 72°C for 45 seconds. Products were diluted 1:3 in a stop solution and denatured at 95°C for 10 minutes, and 2.5 µL was loaded on the gel. The gel formulation was as follows: 6% acrylamide:Bis(2.6% cross-linking) 10% glycerol at room temperature, 45W. Heteroduplexes were identified at the bottom of the gels, and SSCPs were identified from the single-stranded region. Next, PCR products with SSCP and heteroduplex variants were subcloned in the PCR 2.1 vector (Invitrogene, Carlsbad, Calif) and sequenced using an automatic DNA sequencer (Applied Biosystems 370A, Foster City, Calif) and a commercially available kit (Taq Dye Deoxy Terminator Cycle Sequencing kit; Perkin Elmer, Norwalk, Conn). A 37-bp deletion was identified by sequencing of the coding and complementary strands. Subsequently, all genomic DNA samples from individuals of the family were amplified with the primer set for exon 2 (5′-GTA ACG CTC GCT TAT GTT GAC TGT C-3′ and 5′-ACC TGA GAT ATC AGC AAT TGG CAG C-3′), and the PCR products underwent electrophoresis on a 3% agarose gel.
Eleven subjects were determined to have the mutation and agreed to take part in the clinical family study (Figure 1). Neurologic, psychiatric, and other medical manifestations of these family members are summarized inTable 1. Ages at evaluation ranged from 10 to 73 years. Six subjects had a motor disorder, and 6 had psychiatric manifestations, with 3 of these having a combination of motor and psychiatric manifestations. Only 2 subjects were considered clinically unaffected. Deafness was clinically apparent in 6 subjects but was not evaluated by audiometry. Results of DNA mutational analysis were conveyed to those subjects who requested this on their consent form. Genetic and medical counseling was given. In several patients, levodopa treatment was recommended and initiated after appropriate consultation with local physicians. Levodopa in combined with carbidopa (Sinemet tablets) was administered to 6 subjects with subjective (eg, fatigue) and objective symptoms (eg, dystonia and parkinsonism). Improvement was observed in all 6, with the usual dose being 2 to 3 tablets (25 mg carbidopa–100 mg levodopa tablet) daily. Antidepressants (usually fluoxetine hydrochloride) were administered to 4 subjects.
The proband (IV:5) and his father (III:3) had low CSF levels of HVA, 5-HIAA, and MHPG metabolites, compared with age-matched control subjects (Table 2). Levels of neopterin and biopterin in CSF were also markedly decreased in the proband and his father, whereas all metabolite levels were normal in the unaffected sister (IV:6) of the proband (Table 2).
The proband and his father were heterozygous for an SSCP variant in exon 2 that was not present in related family members (data not shown). Sequencing analysis of exon 2 of the proband demonstrated a 37-bp deletion mutation in 1 allele of the GCH1 gene (Figure 2A). With the use of PCR-based screening with the primer set for exon 2 (Figure 2B), the deletion was detected in 11 individuals in the family. The deletion shifts the translational reading frame of the GCH1 gene at amino acid 138 and predicts a premature stop codon at position 160 (Figure 2C).
Torsion dystonia is clinically characterized by early, exercise-induced or spontaneous intense and sustained muscle contractions of trunk and extremity muscles with torsional components. Age of onset is in the first 2 decades of life. Clinicians in the 1950s observed dramatic clinical improvement in some patients with torsion dystonia when they were treated with modest doses of an anticholinergic medication, trihexyphenidyl hydrochloride.15 With the advent of levodopa administration for Parkinson disease, clinicians observed a similar dramatic improvement in approximately 5% to 10% of children with torsion dystonia such that a trial of levodopa is now empirically recommended in all children with torsion dystonia.2 Further clinical features observed in such patients include ataxia, choreoathetoid and spastic-type cerebral palsy states, and parkinsonism, often with worsening of motor symptoms toward the end of the day.1 Earlier studies emphasized the prominent diurnal fluctuations of the motor disorder, the continuing therapeutic response to levodopa, and clustering in families, although a pattern of recessive or dominant inheritance with variable penetrance was clinically difficult to determine.
Previous investigators documented low levels of HVA and biopterin metabolites in CSF from patients with torsion dystonia responsive to levodopa.16 Given the familial nature of the disorder and advances in understanding the role of BH4 in aromatic amino acid hydroxylation and biogenic amine metabolism, molecular biologists allied with clinicians were subsequently able to demonstrate mutations underlying this disorder. Mutations were detected in the rate-limiting enzyme of BH4 biosynthesis, GCH1, causing a dominantly inherited disorder with variable penetrance.3 Compound heterozygote mutations of the GCH1 gene cause severe psychomotor delay and dystonia responsive to levodopa.7 Mutations have also been found in the gene for tyrosine hydroxylase, which is the rate-limiting enzyme in DA biosynthesis. These mutations cause a recessive disorder with decreased levels of DA and NE metabolites.17
By means of biochemical and molecular methods, we characterized 11 members of a large Texas family heterozygous for a deletion mutation in exon 2 of the GCH1 gene. We observed a marked intrafamilial phenotypic variability in GCH1 heterozygotes. Dystonia was present in 2 patients (IV:5 and IV:21). Obvious parkinsonism was observed only in 1 patient (II:3), whose father (I:1, the great-grandfather of the proband of the study) died of parkinsonian-related complications. Late-afternoon fatigue was observed in 3 patients (IV:5, II:1, and III:3), which improved in all 3 with administration of moderate doses of levodopa. Tremor was observed in 3 patients (II:1, II:3, and IV:21) and torticollis in 2 patients (father and son, III:12 and IV:21, respectively), and in 3 patients (II:3, III:12, and IV:21) there was a cerebral palsy–like state (flexion of hips and legs affecting gait and posture). Extensor plantar responses (possible striatal toe18) were present in 3 patients (II:3, III:12, and II:5), ataxia in 2 patients (II:3 and II:5), and definite brisk deep tendon reflexes in 1 patient (III:12). In the patient with obvious Parkinson disease (II:3), dyskinetic features improved after a reduction in his dose of levodopa. Although the various phenotypic motor features of levodopa-responsive dystonia with parkinsonism have been recognized for many years,19 only with the advent of molecular techniques has it been possible to define phenotypically the motor manifestations with certainty in GCH1-deficient patients and families.20 A recent study of 5 families with levodopa-responsive dystonia4 demonstrated marked variation in expressivity, even between affected members of the same kindred. In our study we have demonstrated that there is indeed a marked intrafamilial variability in motor manifestations in levodopa-responsive dystonia with parkinsonism due to a specific new mutation in the GCH1 gene.
In previous studies of levodopa-responsive dystonia, investigators have postulated that, based on clinical manifestations and rapid and persistent responses to medications (anticholinergic agents and levodopa replacement), the various motor manifestation of GCH1 deficiency are DA mediated.19,21 In addition, the patients often show decreased levels of 5-HT (5-HIAA)22 and NE (MHPG)23 metabolites. By means of CSF analysis in 2 heterozygous subjects of the family (the proband [IV:5] and his father [III:3]), we were able to show that the mutation in the present study produced a defect in cerebral DA as well as cerebral 5-HT and NE biosynthesis (Table 2). Psychiatrists have increasingly implicated the role of 5-HT and NE in depressive and anxiety states.24 Using a standardized psychiatric interview process, a board-certified psychiatrist (R.A.H.) determined that depressive manifestations (requiring previous and current pharmacotherapy) were prominent in 4 patients (IV:5, III:3, III:16, and III:17), anxiety was variably present in 6 patients (IV:5, II:1, III:3, III:12, III:16, and III:17), and obsessive-compulsive traits in 1 patient (III:12). In general, psychiatric manifestations have been only occasionally mentioned in patients with levodopa-responsive dystonia with parkinsonism.25 However, the psychiatric manifestations (eg, depression and anxiety) of adult-onset Parkinson disease have been increasingly recognized26,27 and may have a similar neurochemical basis, given the known pathologic and neurochemical defects in the locus ceruleus (NE) and raphe nuclei (5-HT) in the brain regions of patients with Parkinson disease.28
Manifestation of juvenile-onset parkinsonism may not always be present in the patient and may only become apparent in later decades. Therefore, the differential diagnosis between degenerative Parkinson disease and levodopa-responsive (genetic) dystonic parkinsonism without family history or a demonstrable genetic defect is often a hard task for a clinician. The possibility of both conditions being present in older individuals is certainly feasible and of prognostic significance. In such patients, the decrements in DA functions with aging, including DA transporter (presynaptic) and dopamine (D2) receptor (postsynaptic) defects, as recently quantitated by means of positron-emission tomography imaging,29 may be further exacerbated by modest defects in DA synthesis in GTP1-deficient heterozygous and manifesting carrier subjects. The recent demonstration of increased D2 receptor density in the basal ganglia using carbon 11–labeled raclopride positron-emission tomography imaging in asymptomatic and symptomatic patients with levodopa-responsive dystonia is of interest, and presumed up-regulation of these receptors is related to functional synaptic DA deficiency.16 This up-regulation in DA receptors may explain the initial therapeutic sensitivity of patients with levodopa-responsive dystonia treated with even small doses of levodopa.
Clinically detectable deafness of a mild to severe degree was apparent in 6 heterozygous patients; however, the significance of this observation is uncertain because deafness has not been a feature in previously described patients with levodopa-responsive dystonia with parkinsonism and is rarely associated with dystonia.30 Recently, an X-linked form of dystonia-deafness syndrome was studied by means of mutational analysis of a novel X-linked gene, deafness/dystonia peptide (DFN-1:DDP).31 The association of clinically apparent hearing loss in 6 members of this family with a mutation in the GCH1 gene may warrant further clinical and molecular studies.
Previous investigators have demonstrated a sex-related bias with a female-male ratio of 4.3 in patients with levodopa-responsive dystonia. Penetrance of GCH1 mutations was 2.3 times higher in women compared with men.5 In the family described herein, 8 male and 3 female subjects were affected, and there was no apparent segregation by sex of motor and psychiatric manifestations (Table 1). Although we cannot explain this difference in sex bias in the family, the specific mutation within a family may be an important determinant for sex-related penetrance.5,32
From a therapeutic viewpoint, management of the movement disorder(s) component of levodopa-responsive dystonia with levodopa and a peripheral decarboxylase inhibitor has to be one of the most satisfying in neurologic practice.33,34 Levodopa may actually have a neurotrophic and neuroprotective role.35- 37 Thus, after the molecular identification of genetically affected individuals who are asymmptomatic, administration of levodopa may also have a beneficial role in disease prevention, particularly in later-onset parkinsonism. Levodopa administered alone (as is currently advocated) may interfere with the uptake of other aromatic amino acids (including phenylalanine, tyrosine, tryptophan, and 5-hydroxytryptophan) into neurons and thus interfere with precursor availability in neuronal synthesis of 5-HT.38 The impact on impaired neuronal synthesis of 5-HT by levodopa administration may be further exacerbated by the vulnerability of 5-HT biosynthesis by a reduction in active cofactor levels in patients with the defective GCH1 gene.39 Potentially defective brain synthesis of 5-HT could cause or exacerbate psychiatric conditions such as depression or obsessive-compulsive traits.
The study of this large Texas family with a newly described and unique heterozygous mutation in the GCH1 gene and with a dominant pattern of inheritance and variable penetrance shows the wide intrafamilial variability in clinical expression (phenotype) of this disorder. Although the proband may have had a rather typical presentation with focal foot dystonia, his history of muscle fatigue was sufficiently prominent to delay diagnosis. The immediate family of the proband recognized psychiatric disorders in family members, and parkinsonism was determined to be relatively common in family members. Furthermore, diurnal variation in motor manifestations was unusual in the family, except for increased afternoon fatigue in several members and increased foot dystonia in the afternoon in the proband. Variability of clinical expression is not unusual in GCH1 deficiency. Furthermore, family members may undergo initial evaluation by a clinical psychologist or psychiatrist for fatigue states, depression, anxiety, and obsessive-compulsive manifestations before the familial nature of a movement disorder is recognized.
From our observations of the clinical manifestions of the syndrome in this family, the current terminology emphasizing dystonia, parkinsonism, and levodopa responsiveness may be too restrictive. Even the criterion of levodopa responsiveness may be overly restrictive, given the potential impact of a deficit of BH4 cofactor on other hydroxylation steps. The combination of neurologic and psychiatric consultation, biochemical (CSF) analysis, possibe phenylalanine loading studies,40 and molecular methods should enable a more precise diagnosis of these conditions and the application of appropriate therapies.
Accepted for publication August 28, 2000.
Supported in part by Clinical Research Center grant RR02588 at the University of Texas–Houston Medical School (Dr Butler), and by grant 15956 from Shriners Hospital for Children, Tampa, Fla.
Corresponding author and reprints: Ian J. Butler, MD, Department of Neurology, PO Box 20708, Houston, TX 77225-0708.