Craniocerebral Magnetic Resonance Imaging Measurement and Findings in Lesch-Nyhan Syndrome

Objective  To provide the first comprehensive magnetic resonance imaging (MRI) assessment of brain in a series of patients with Lesch-Nyhan syndrome (LNS), with emphasis on basal ganglia measurements.

Design  Routine readings of MRI studies, repeated reading in random order blinded to subject diagnosis, and 3-dimensional volumetric measures of basal ganglia regions.

Setting  The Johns Hopkins Hospital, Baltimore, Md.

Patients  Seven patients with LNS who have hypoxanthine guanine phosphoribosyltransferase levels less than 1.6% and characteristic clinical features of the disorder, which include hyperuricemia, cognitive impairment, and dystonic movement disorder, were compared with 7 age-matched control subjects. Five of the 7 patients demonstrated self-injurious behavior. MRI studies were performed using general anesthesia because of the severity of the movement disorder.

Main Outcome Measures  Measurement of brain regions from MRI-obtained images.

Results  Routine readings described mild cerebral atrophy in 2 of 7 patients, but no caudate or putamen abnormalities were reported. However, on the directed blinded rereading, small caudates were suspected in 5 of 7 cases, and abnormalities in cerebral size and cranium were identified. Volumetric studies of the patients with LNS confirmed a 34% decrease in caudate volume (P<.001), a 17% decrease in total cerebral volume (P<.03), and a 12% decrease in putamen volume (P=.19).

Conclusions  To our knowledge, this is the first demonstration of consistent neuroanatomic abnormalities in LNS. The findings of reduced basal ganglia volume are consistent with the dystonic movement disorder.

LESCH-NYHAN syndrome (LNS) is caused by the near absence of the purine salvage enzyme, hypoxanthine guanine phosphoribosyltransferase (HPRT). It is a rare sex-linked metabolic disorder with an estimated prevalence of 1:380000¹ with an unusual neurologic and behavioral phenotype. It was first described in 1964 in 2 brothers who presented with hyperuricemia, mental retardation, choreoathetosis, and self-destructive biting.² Affected individuals may appear to develop normally during the first months of life, although hypotonia may be present. Between 8 and 12 months of age, prominent motor signs suggestive of dysfunction of the basal ganglia become evident.³ Subsequently, behavioral features including aggression, self-injurious behaviors, and cognitive dysfunction are recognized.⁴ Most patients have choreoathetosis or dystonia, and ballismus has been reported. Other patients also display spasticity with hyperreflexia and abnormal extensor plantar reflexes.

Despite the severe motor disturbance, neuropathologic studies have failed to reveal any consistent neuroanatomic abnormalities in the cortex or basal ganglia, even at the electron microscopic level. There are few reported neuroimaging studies, in part due to the rarity of the disease but also due to technical difficulty in obtaining adequate images without general anesthesia. In descriptions of patients that have included neuroimaging studies, no consistent neuroanatomic abnormalities have been identified. In a thorough longitudinal clinical review of 8 patients with LNS, Watts et al³ described 3 who had computed tomographic (CT) scans. The images were considered within normal limits, including the cortex and basal ganglia. In another study, CT scans were reviewed from 12 patients with LNS.⁵ Mild to moderate diffuse cerebral atrophy was apparent in some, but not all, of these cases. The first study to use standard magnetic resonance imaging (MRI) in LNS reported on 4 patients.⁶ Two had normal studies, one had moderate cerebral atrophy, and one had decreased T₂ signal in the basal ganglia. In a subsequent study, routine MRI results were reported to be normal in 12 patients who participated in a positron emission tomography procedure.⁷

In the present study, we sought to identify subtle morphologic abnormalities in LNS. We evaluated standard brain MRI scans that were read routinely and, subsequently, by a neuroradiologist blinded to subject diagnosis in 7 classic LNS cases. We assessed changes in brain volumes using volumetric MRI in each of these patients. To our knowledge, there have been no published series that specifically focus on neuroradiologic findings. This is the first series of patients with LNS that provides a systematic description of cranial MRI measurements and that uses volumetric MRI procedures to measure basal ganglia volumes.

Subjects included 7 patients with LNS. The diagnosis was based on the demonstration of HPRT enzyme activity of less than 1.6% in erythrocyte lysates and intact fibroblast cultures performed by Theodore Page, PhD, and William Nyhan, MD, PhD, La Jolla, Calif.⁸ All patients displayed a characteristic phenotype that included delayed motor development with hypotonia, early-onset choreoathetosis or dystonia, dysarthria, cognitive impairment, aggressive behavior, and hyperuricemia. Five of the patients also demonstrated typical self-injurious behaviors. The other 2 (HPRT levels of 0.9% and 1.4%) were atypical in that self-injury was not observed. All subjects were male; 6 were white and 1 was African American. They ranged in age from 22 to 35 years (mean age, 25.8 years), with a mental age of 7 years or greater. Consent for the studies was obtained from parents or guardians, and assent was obtained from each of the patients in accordance with a Johns Hopkins institutional review board–approved protocol. The control population consisted of 7 healthy individuals matched for sex and age.

Standard MRI studies on each of the 7 patients with LNS and 7 age-matched controls were completed on a 1.5-T MRI scanner (GE Signa, Milwaukee, Wis) using the following protocol: 5-mm sagittal and 4-mm axial T₁-weighted images; 5-mm axial proton-density and conventional (not fast spin-echo) T₂-weighted images. The axial images were obtained parallel to the anterior commissure/posterior commissure (AC-PC) line. Axial contiguous 1.5-mm SPoiled GRASS images (described below) were then obtained.

Analysis of the MRI scans consisted of 3 phases. In phase 1, all patients had standard cranial MRI, and interpretations were supplied by an experienced neuroradiologist. In phase 2, scans from patients and controls were blindly reviewed by a neuroradiologist (R.R.L.), who was asked specifically to comment on the size and configuration of the basal ganglia and cortex. In phase 3, volumetric studies were completed.

Studies were read routinely as part of the normal clinical workflow. At a later time, the 7 patients with LNS and 7 age-matched controls (a total of 14 studies) were presented in random order and blindly reviewed by a single neuroradiologist (R.R.L.), without knowledge of how many of the 14 cases were affected with LNS and how many were healthy control subjects. The blind review of both LNS and healthy control MRI scans focused on volumes of the caudate, putamen, and associated structures, and evaluated the presence of cerebral atrophy. The cranium was also evaluated.

The 7 patients with LNS and 7 control subjects were then studied with volumetric MRI. The volumes of the basal ganglia were measured from T₁-weighted 3-dimensional SPGR (SPoiled GRASS-Gradient Refocused Acquisition in the Steady-State) MRI scans (repetition time, 35 milliseconds; echo time, 5 milliseconds; 1 excitation; 45° flip angle) that consisted of 128 contiguous axial 1.5-mm slices through the entire brain. The lateral ventricles (excluding the temporal horns), the head and body of the caudate nucleus, the putamen, and the globus pallidi were manually outlined by a trained operator. Three-dimensional volumetric images were reconstructed with standard clinical software (ISG, Mississauga, Ontario). Blinded repeated measurements of these structures demonstrated variability of less than 10%.

Because of the severity of the movement disorder, an anesthesia protocol was established to maintain immobility during the MRI procedure. Patients were anesthesized with either intravenous propofol (150-250 g/kg per minute) or rectal methohexital (30 mg/kg), supplemented by 70% nitrous oxide in oxygen. Pancuronium bromide (0.1 mg/kg) was used to facilitate endotracheal intubation and to keep the patients immobile during the procedure. Patients were mechanically ventilated (end-tidal CO₂, 40±5 mm Hg) by the anesthesiologist throughout the procedure through an endotracheal tube. Heart rate, blood pressure, and oxygen saturation did not vary by more than 10% from the baseline.

Statistical analysis of the volumes of the basal ganglia and brain in patients with LNS and control subjects were analyzed using SPSS for Microsoft Windows (SPSS, Chicago, Ill), Release 6.1. Significance testing was performed using 2-tailed Student t tests.

Routine reading of the 7 LNS cases as part of the normal clinical workflow identified few abnormalities. Only 2 patients were thought to have cortical atrophy, and there were specific comments about the basal ganglia in only a single patient.

When the 14 cases (patients and controls) were retrospectively blindly reviewed in randomized order, with explicit attention to the basal ganglia, several additional features were identified in the patients with LNS only. The MRI scans were blindly read as normal for 6 of 7 control subjects; the scan of the seventh subject showed mild prominence of the sulci of the cerebellar vermis but was read as otherwise normal.

Blindly reviewed neuroradiologic findings on the 7 LNS cases are summarized in Table 1. Three of the 7 patients had small caudates and 2 others had caudates that were believed to be probably small. This assessment was based on the shape of the caudate nucleus and the shape and size of the adjacent frontal horns (Figure 1); small caudates create less of a bulge into the ventricles, resulting in more prominent frontal horns, out of proportion to the rest of the ventricular system. The putamina were believed to be possibly small in only 2 patients. The bicaudate index (bicaudate diameter divided by the outer tables of the skull at the level of measurement)⁹ was determined and is described in Table 2. A significant difference (t₁₃=2.28; P =.04) was documented between the LNS and control groups.

Parenchymal volume loss, manifested by prominent cerebral sulci, was assessed in 4 patients. In 1 patient, there was focal bifrontal atrophy (Figure 1, B). Two of 7 patients had slightly prominent ventricles. In 4 or 5 patients, there was brainstem atrophy, manifested by prominence of the prepontine cistern and other basilar cisterns (Figure 2). One patient had a very large cerebellomedullary cistern, and 2 others were slightly large (Figure 2, A).

No foci of abnormal T₂ hyperintensity were observed in the brains of any patients or controls. However, 2 patients with LNS had suggestion of slightly decreased T₂ signal within the globus pallidi and substantia nigra on the T₂-weighted sequences, whereas all the controls had normal signal.

Rather striking abnormalities of the skull and sinus pneumatization were observed in 5 of 7 patients with LNS, whereas all controls had normal findings. Four patients had prominent thickening of the skull marrow, and 5 had abnormal T₁ signal within the marrow that was isointense to brain, rather than exhibiting normal bright fatty signal (Figure 3). The marrow of the clivus and visualized cervical vertebrae also demonstrated this signal abnormality. On T₂-weighted images, the marrow signal was not appreciably different from normal fatty marrow. Five patients with LNS also exhibited prominent pneumatization of the paranasal sinuses, especially the frontal sinuses, and the mastoid air cells.

The MRI volumetric findings are shown in Table 3 and Table 4 and Figure 4. The blinded retrospective review, suggestive of low basal ganglia volume in the majority of cases, was confirmed by volumetric studies. As shown in Table 4, when compared with the 7 age-matched controls studied using the same MRI volumetric sequences, the 7 patients with LNS demonstrated a significant 34% decrease in volume in the caudate nucleus (P<.001) and a 12% reduction (P=.19) in the putamen that was not significantly different from the controls. The total caudate volume (mean±SD) was 7.9±0.04 cm³ for control subjects and 5.2±0.097 cm³ for patients with LNS (P<.001). The total brain volume for the control subjects was 1163.4±79.2 cm³ and for the patients with LNS 963.1±213.96 cm³, a 17% reduction in total brain volume (P<.01). Although there was little variability in cerebral size in the control subjects, considerable variability was found in the patients with LNS (range, 799.9-1409.3 cm³). The greatest cortical volume was that of a patient with LNS whose total brain volume was 1409.3 cm³, a value that is 237 cm³ greater than the average of the control group. This patient had classic LNS with an HPRT level of 0.9% and reduction in caudate volume. His cognitive function was in the low normal range.

The standard MRI studies were read as unremarkable or as showing mild cortical atrophy, and the basal ganglia sizes were not specifically commented on by the neuroradiologists in the original readings. These results are consistent with findings in various case reports^4-6 that imaging studies are unremarkable in LNS. Yet when these studies of 7 patients with LNS and 7 controls were read blindly by a neuroradiologist with specific attention given to the basal ganglia (without knowledge of how many of the 14 cases were affected with LNS, or which cases were affected), reduced brain volume, particularly in the caudate, was suggested, as well as increases in ventricular size and mild cortical atrophy and/or brainstem atrophy in several patients. (Scans for controls were all read as normal in the blinded readings, except for mild vermian atrophy in 1 control subject.) One case (patient 1) demonstrated specific frontal lobe atrophy, manifested by more prominent frontal sulci than in the rest of the brain, and in this patient, the caudate nuclei were the smallest. Another case (patient 4) showed the greatest brain volume but also demonstrated small caudate nuclei. The main brain findings from the blinded directed readings of patients with LNS were reduction in total cerebral volume and reduced volumes in the caudate and putamen in some cases, when compared with control subjects.

Although the relative decrease in caudate volume and cerebral volume in LNS are similar, the reduction in caudate volume is not merely a reflection of reduction in overall cerebral volume. If the small caudate volume merely reflected a globally small brain, the respective proportions of all structures would be preserved; there would be no relative prominence of the frontal horns, and the bicaudate index would be expected to be normal, not large as documented in Table 2. The overall loss of brain volume is thus composed of proportionately smaller caudates (and perhaps their cortical projections) superimposed on a small brain.

The reduction in brain size was accompanied by thickened calvaria and increased sinus pneumatization or sinus size. The increased thickness of the skull and in sinus size may reflect a failure of continued brain growth, probably beginning in infancy. When brain growth does not continue at the normal rate throughout the developmental years, a compensatory thickening in the skull and sinuses may occur.¹⁰

The abnormal thickening and abnormal signal intensity of the skull marrow, with isointense rather than normal fatty hyperintense T₁ signal, suggest abnormal increase in marrow hematopoietic activity, consistent with known anemia in LNS.¹¹

No foci of T₂ hyperintensity were found in any of the 7 patients or 7 controls, consistent with previously reported MRI results.⁶ Only 2 of 7 patients had a suggestion of possible mild T₂-dark signal in the globus palladi and substantia nigra, which was reported in 1 of 16 MRI scans of patients with LNS in previous literature.⁶

These results indicate that standard MRI readings involving patients with suspected or known LNS may be interpreted as normal unless specific attention is directed to the basal ganglia and the skull. Based on our findings, MRI readings of patients with LNS should include a review of skull thickness, paranasal sinus size, caudate and putamen size, ventricular size, and presence of cortical and/or brainstem atrophy. The second directed reading suggested the need for volumetric studies to substantiate differences in the basal ganglia size of these patients when compared with control groups.

Our subsequent volumetric MRI findings of reduced caudate as well as total cerebrum size are consistent with the movement disorder and cognitive impairment found in LNS. Lehericy et al¹² reported that MRI scans of patients with dystonic spasms had focal lesions in the striatopallidal complex, involving the putamen posterior to the anterior commissure and extending into the dorsolateral part of the caudate nucleus, the posterior limb of the internal capsule, or the lateral segment of the globus pallidus. Thalamic lesions associated with dystonia were located in the ventral intermediate and ventral caudal nuclei. Other authors have reported similar findings.^13,14 Our findings of significantly reduced caudate volume and a trend toward reduced putamen volume in patients with LNS are consistent with the striatopallidal form of dystonia rather than a thalamic form.

These changes differ from those found in Huntington disease or in Parkinson disease. In Huntington disease, a neurodegenerative disorder, there is a progressive reduction in basal ganglia volumes with age, whereas in our LNS series there is no apparent decline when the younger and older patients are compared. In fact, one of the younger patients (age 22 years) has a caudate volume similar to that of the 35-year-old patient. In Huntington disease, caudate atrophy has been documented with CT (bicaudate ratio),¹⁵ and caudate, putamen, and globus pallidus volume reductions have been reported on MRI¹⁶ in 10 gene marker-positive asymptomatic persons. Atrophy of basal ganglia structures has been associated with disease duration, severity of dementia, severity of movement disorder, and functional capacity. In patients with Huntington disease who had mild-to-moderate motor signs, Harris et al¹⁷ found that putamen volume was reduced by 50.1% and caudate volume was reduced by 27.7%. Aylward et al¹⁶ suggest that caudate volume may show an initial decline and remain fairly stable in the early stages of Huntington disease, while putamen volume continues to decline.

In Parkinson disease, CT and MRI scans are usually normal or show minor degrees of atrophy¹⁸ that are consistent with involutional changes in Parkinson disease.¹⁶ In Rett syndrome, another neurodevelopmental disorder, quantitative MRI scans show reduced cerebral volume, greater loss of gray matter in comparison with white matter, regional variation in cortical gray matter, with frontal regions showing the largest decrease, and reduced volume of the caudate nucleus and midbrain (apparently taking into account the general reduction in brain size).¹⁹ Quantitative MRI studies of children and adults with Tourette syndrome have shown left-sided reductions in the caudate, putamen, and pallidum.^20,21

There are few neuroanatomic studies in LNS. The brains of 11 patients have been autopsied,^3,22-28 with variable reports regarding small brain size. Focal softenings were noted in 3 reports,^22,23,25 areas of gliosis in 3 reports,^22,25,28 and loss of cerebellar cells in 3 reports.^22,25,28 The brain size was considered small in 4 of these cases.^22,24,26,27 Less commonly noted findings have included demyelinative lesions,²² cerebral edema,²⁶ marked hydrocephalus,²⁹ and periodic acid–Schiff–positive inclusions in the olivary nuclei.²⁸ These abnormalities may be related to postmortem changes secondary to chronic renal failure in untreated patients or ischemia in patients dying of pneumonia and aspiration. It is noteworthy that neuropathologic examination results were completely normal in 3 patients.^24,25,27 Although the small brains observed in several prior neuroimaging and neuropathologic studies have been interpreted as showing diffuse atrophy, the term atrophy may be misleading since it implies a pathophysiologic process that has not been substantiated. In particular, the small brain may reflect a developmental, rather than a degenerative, process. Our findings suggest that future neuropathologic studies should focus on the caudate, putamen, and brainstem.

In contrast to Huntington disease, LNS is considered to be a neurodevelopmental disorder. Rather than degenerative atrophy, it is proposed that dopaminergic innervation of the basal ganglia is curtailed.³⁰ Reductions in dopaminergic innervation of the striatum have received confirmation with positron emission tomographic scanning using a presynaptic dopamine marker (WIN 35,428)³¹ and fluorodopa F 18.⁷ The cause of this dopamine reduction is unknown. Wong et al³¹ applied partial-volume corrections for caudate and putamen in 6 of the 7 LNS cases reported herein to ascertain reductions in presynaptic dopamine density. Our findings suggest that anatomic abnormalities may extend beyond presynaptic loss of dopamine fibers to account for these changes in volume. The relationship of subcortical atrophy and cortical atrophy has also been considered in Huntington disease,¹⁵ and cortical changes are thought to be secondary to subcortical ones. This is also a consideration in LNS.

In summary, we have documented that there are reductions in caudate volumes in the brain associated with LNS, using volumetric MRI. These findings must be considered when positron emission tomographic scan studies are conducted in patients with LNS since they make partial-volume correction necessary for appropriate interpretation. Future analyses should focus on ascertaining the relationship of these neuroanatomic changes to enzyme levels and the genetic defect. Moreover, it is likely that there are regional cortical changes in the patients, some of which might be predicted by the cortical projection areas of the caudate; these projection areas might be disproportionally smaller than other cortical areas. In the future, cortical brain analysis techniques³² may be used to obtain these measurements.

Accepted for publication August 27, 1997.

This work was supported in part by grants (HD23042, HD24061, and HD33095) from the National Institutes of Health, Washington, DC, and by Lesch-Nyhan Syndrome Children's Research Foundation, Lake Forest, Ill.

The authors thank Lowell Anderson, PhD, and William Nyhan, MD, PhD, for patient referrals and Wayne Fisher, PhD, and his staff in Behavior Management at the Kennedy Krieger Institute for their help. Special thanks to Hugo Moser, MD, the Johns Hopkins/Kennedy Krieger Institute Mental Retardation Developmental Disability Research Center, and George Breese, PhD, for their technical support.

Corresponding author: James C. Harris, MD, Director of Developmental Neuropsychiatry, Professor of Psychiatry and Pediatrics, Johns Hopkins Hospital, 600 N Wolfe St, CMSC 341, Baltimore, MD 21287 (e-mail: jamesharris@erols.com).