Corticospinal Physiology in Patients With Prader-Willi Syndrome: A Transcranial Magnetic Stimulation Study

Background  Prader-Willi syndrome (PWS) is a genetic developmental disorder, mostly caused by a deletion on the paternal chromosome 15 or by a maternal uniparental disomy 15. Some PWS clinical and neurochemical features suggest an involvement of the corticospinal motor structures.

Objective  To explore the corticospinal physiology of PWS by transcranial magnetic stimulation.

Setting  A community-based hospital.

Methods  We studied motor evoked potentials in the first dorsal interosseous muscle of 21 young-adult patients with PWS. Thirteen patients had a deletion at chromosome 15; 8 had a uniparental disomy. We measured the following variables: relaxed motor threshold, central motor conduction time, duration of the central silent period, and short-interval intracortical inhibition and facilitation. We also recorded F waves in the first dorsal interosseous muscle. We had 11 normal controls.

Results  In the whole PWS group, motor threshold was higher as compared with controls (P<.05). The central motor conduction time, central silent period, and F waves were normal. Intracortical facilitation was reduced significantly (P<.001). Patients with PWS and a deletion had a weaker intracortical inhibition as compared with patients with PWS and a uniparental disomy (P<.05).

Conclusions  Transcranial magnetic stimulation changes in patients with PWS suggested a hypo-excitability of the motor cortical areas. Defective neurogenesis of the cortical tissue and multiple transmitter alterations are the putative causes. Impaired intracortical inhibition might represent an electrical marker for a deletion defect.

Prader-Willi syndrome (PWS) affects both sexes, all races, and 1 in every 10 000 to 15 000 live births.¹ Its neurophysiology has been poorly investigated,^2,3 although the central nervous system is one of the main targets of the underlying genetic defect.⁴ In fact, apart from the characteristic facial aspect, the main clinical features of PWS are hyperphagia, obesity, short stature, and hypogonadism. A hypothalamic involvement has been proposed to explain these features. Other features suggest an involvement of the central nervous system motor structures: central hypotonia in infancy, delayed psychomotor development, difficulties in motor aspects of language, and obsessive-compulsive behavior.¹ The original PWS genetic defect leads to an altered coding for proteins that control the neuronal growth/differentiation in widespread brain regions, such as necdin.⁵ Cortical dysgenesis with atrophic aspects was also reported.⁶ Moreover, coding of some subunits (α-5, β-3, and γ-3) of the γ-aminobutyric acid type-A (GABA_A) receptors is defective.⁷ These receptors become hyposensitive in many cortical and subcortical areas. In response, circulating GABA levels increase abnormally.⁸ Thus, a complex derangement of the neurotransmitter balance arises,^9,10 which may well involve the motor cortical regions and their projections.

For these reasons, we explored the corticospinal system in a group of patients with PWS by means of transcranial magnetic stimulation (TMS). This is a noninvasive, safe, and painless probe of the corticospinal excitability and conductivity. Transcranial magnetic stimulation variables are sensitive to the transmitter balance in the cortex.^11,12

Prader-Willi syndrome develops from a failure to express paternally derived genes in the q11-q13 region of chromosome 15.¹³ Deletion in the paternally derived chromosome 15 occurs in 60% to 70% of patients with PWS.¹⁴ Almost 20% to 30% of patients inherit 2 maternally derived (ie, inactive) chromosomes 15; this is termed uniparental disomy(UPD).¹³ Other rare genetic defects can occur.¹⁴ With the hypothesis that a different genotype might produce different changes in the corticospinal physiology, we analyzed the TMS findings as a function of this variable.

We examined 21 right-handed patients with PWS (9 men; mean ± SD age, 24.6 ± 6.2 years; range, 15-39 years). All met the clinical and genetic criteria for a PWS diagnosis (average Holmscore ¹⁵ ± SD, 10.4 ± 1; range, 9-12.5). Each case was initially studied with high-resolution chromosome analysis. The presence of a deletion was confirmed by fluorescent in situ hybridization. If fluorescent in situ hybridization studies did not show a deletion, UPD and methylation were investigated.¹⁶ Thirteen patients had a deletion (6 men); 8 had a UPD (3 men). None had other unusual molecular findings. Informed consent was obtained from the patients or from their parents or legal guardians if the patients were underaged. In no case was the patient’s IQ below 69. The experimental procedure, which had the approval of the local ethics committee, was explained in detail to the patient. No patient was taking neuroactive drugs. We had a normal control group of 11 subjects, matched for age and sex.

Transcranial magnetic stimulation was performed through a round coil (outer diameter, 14 cm) centered at the vertex. We used an anticlockwise current direction to stimulate the left hemisphere preferentially. Motor evoked potentials (MEPs) were recorded from the right first dorsal interosseous muscle by silver-silver chloride surface electrodes. Data were collected, amplified, stored, and analyzed by a Neuroscan machine (Neuroscan Laboratory, Sterling, Va). The signal was amplified with a bandpass of 3 Hz to 3 kHz, a sweep duration of 10 to 50 ms/division, and a gain of 0.1 to 1 mV/division. Patients sat comfortably in a chair. Muscle relaxation was essential for most recordings, but patients were kept awake. Motor evoked potential variables were assessed following the guidelines of the International Federation of Clinical Neurophysiology.¹⁷ Using a single stimulator (Magstim 200; Magstim Co, Whitland, Wales), we determined the relaxed motor threshold (rT) and the central motor conduction time. To estimate the peripheral conduction time, we used the F wave latency. The central silent period was the period of electromyographic silence produced by a TMS pulse set as 1.5 × rT, and the first dorsal interosseous muscle was activated at the maximum force level. The length of the central silent period extended from the stimulus artifact to the consistent reappearing of electromyographic activity. With the paired-pulse technique,¹⁸ we finally studied intracortical inhibition (ICI) and intracortical facilitation (ICF) during complete muscle relaxation. Two Magstim 200 stimulators were coupled with a Bistim device (Magstim Co). The intensity of the conditioning stimulus was 0.8 × rT. The test stimulus was 1.2 × rT, and it was slightly adjusted to evoke MEPs sized 0.5 to 1 mV. We studied 2 inhibitory (2 and 3 ms) and 2 facilitatory (14 and 16 ms) interstimulus intervals. For each interstimulus interval, we recorded at least 12 control and 12 conditioned MEPs at random. The effect of conditioning was the ratio of the average conditioned MEP to the average control MEP.

We evoked 16 F waves from the first dorsal interosseous muscle by supramaximum electrical stimulation of the ulnar nerve at the wrist (1Hz, 0.1 ms). We measured their minimum latency and average peak-to-peak size.

We used the software package SPSS 6.0 for Windows (SPSS Inc, Chicago, Ill) to analyze the TMS variables. First, we compared the whole PWS group with controls. Intracortical inhibition was the average inhibition (percentage of control MEP) at interstimulus intervals of 2 and 3 ms. Intracortical facilitation was the average facilitation at interstimulus intervals of 14 and 16 ms. We used unpaired t tests with a Bonferroni correction of P values. We then split patients into 2 subgroups (deletion and UPD) according to their genotype. These were compared with controls by a 1-way analysis of variance model. Post hoc Bonferroni tests were performed to determine significant (P < .05) differences.

The Table shows numerical data. The average rT value was higher in the PWS group than in the control group (P < .05) (Figure 1A). The central motor conduction time (corrected by height) was normal in the PWS group. The average central silent period duration and ICI values did not differ between patients and controls as well. Intracortical facilitation showed an average reduction in the PWS group (P<.001) (Figure 1B). The F wave size of patients with PWS was normal.

The threshold was enhanced, and ICF was decreased to a similar extent in both PWS subgroups in comparison with controls (P<.05). Intracortical inhibition was weaker (ie, its nominal value was higher) in the deletion subgroup (24.1% ± 11.9%) than in the UPD subgroup (11.4% ± 5.6%) (P<.05) (Figure 1C and Figure 2).

The main findings of the present study are an increased rT and a reduced ICF in patients with PWS. Physiologically, these findings correspond to an overall corticospinal hypo-excitability.¹¹

Many factors may affect the rT to transcranial magnetic stimulation.¹¹ To extract the cortical responsibility in a given threshold change, many authors evaluated the spinal excitability by means of F waves.¹⁹ This method, however, has some limitations.¹⁷ In our PWS group, F waves were of normal size. Thus, we suggest that hypo-excitability was mainly a motor cortical phenomenon. Changes in rT reflect changes in the voltage-gated ion channel function in the neuron membrane.¹² In PWS, some of the pivotal genetic defects interfere with neural differentiation in the cortex.⁶ This may well cause abnormalities in the membrane properties of the area 4 neurons targeted by TMS shocks.¹¹ Interestingly, threshold is also increased in the Angelman syndrome,²⁰ a disorder that results from a genetic defect in the maternally inherited chromosome 15.¹³

Paired-pulse TMS is now used to test neurotransmission within the motor cortex^11,12 in health and disease.^21,22 Reportedly, dopaminergic drugs increase ICI, while antidopaminergic, anticholinergic, and serotonergic drugs reduce it. Antidopaminergic and anticholinergic drugs increase ICF. Moreover, N-methyl D-aspartate receptor blockers can increase ICI and reduce ICF. In turn, enhancement of GABAergic transmission mainly decreases ICF and, to a lesser extent, strengthens ICI. However, not all GABAergic drugs have the same action on ICF. Vigabatrin and the benzodiazepines, the typical GABA_A receptor antagonists, reduce ICF when given acutely. Conversely, tiagabine leads to an ICF increase through presynaptic GABA_B receptors.¹²

Neurochemically, the PWS genetic defect would imply a malfunction of GABA_Aergic inhibition, due to hyposensitivity of GABA_A receptors. This, in turn, would release an excess serotonergic, dopaminergic, glutamatergic, and peptidergic transmission.^9,10 Moreover, the PWS genetic defect would imply, as a feedback response, enhanced GABA plasma levels.⁸ On this basis, we should have expected an excess ICF (and a weaker ICI) in our patients. The actual data went in the opposite direction, denying the validity of the model. In addition, the complex nature and distribution of the GABA_A receptors do not support simplistic explanations.²³ Prader-Willi syndrome affects 3 subunits (α-5, β-3, and γ-3) of the GABA_A receptor. They aggregate into isoforms, which represent no more than one third of the overall GABA_A receptors in the brain.²⁴ Thus, high levels of circulating GABA might even result in a paradoxical overstimulation of the surviving GABA_A (or GABA_B) receptors. Overstimulation of GABA_A would nicely explain our finding of a defective ICF.

The central silent period is an index of motor cortical inhibition, and GABA transmission might affect it.²⁵ However, its pharmacology turned out to be contradictory.¹² Perhaps this is why we found no significant central silent period changes in our patients, even if GABA changes were expected.

Incidentally, the TMS features of our patients with PWS were nearly opposite to those of patients with obsessive-compulsive disorder.²⁶ Yet, PWS reportedly implies a proneness to obsessive-compulsive behavior,¹ whose physiology would thus seem peculiar.

Finally, the amount of ICI was less in patients with PWS and a deletion than in patients with PWS and a UPD. Deletion, as compared with UPD, would thus imply a lower cortical GABAergic tone and possibly an enhanced serotonergic/glutamatergic tone.¹² This neurochemical change would be more akin to the proposed model of PWS.^9,10 A similar, selective involvement of ICI was found in patients with subtypes 2 and 3 of spinocerebellar ataxia. Selective involvement of ICI was seen as a typical genotype-related electrophysiological difference because of the expression of the mutant protein in the cortex.²⁷ We then hypothesize that impaired ICI is a candidate electrical marker for patients with PWS and a deletion as opposed to patients with PWS and a UPD. Reportedly, seizures are more common in patients with a deletion,^28,29 which would fit the lessened inhibition nicely.²² However, this is a preliminary observation, which awaits reproduction in larger patient cohorts.

Correspondence: Roberto Cantello, MD, PhD, Dipartimento di Scienze Mediche, Via Solaroli 17, 28100 Novara, Italy (cantello@med.unipmn.it).

Accepted for Publication: March 2, 2004.

Author Contributions:Study concept and design:Civardi, Grugni, and Cantello. Acquisition of data: Civardi andVicentini. Analysis and interpretation of data:Civardi and Cantello. Drafting of the manuscript: Civardi and Cantello. Critical revision of the manuscript for important intellectual content: Civardi, Vicentini, and Grugni. Statistical expertise: Civardi and Cantello. Administrative, technical, and material support: Cantello. Study supervision: Cantello.

Funding/Support: Drs Civardi and Cantello were supported by a research grant from the Istituto Auxologico Italiano, Milano, Italy.