Abstract
Objective: To study the clinical significance of metabolic alterations as measured in vivo with proton MRS in the striatum of patients with Huntington’s disease (HD).
Methods: Localized, single-voxel MRS was performed on the basal ganglia of 10 HD patients (4 presymptomatic gene carriers and 6 akinetic patients) and 5 age-matched healthy individuals. Metabolite quantification was performed by referring the areas of the respective spectral peaks to that of water in the analyzed voxel. The spectroscopic findings were correlated with motor and cognitive performance in several specific tests and with the length of the CAG repeat expansion normalized for age. Results: N-acetylaspartate (NAA) and creatine were reduced markedly in both groups of patients, particularly in the advanced group (approximately 60%), but the decrease was also significant in presymptomatic patients (approximately 30%) whose motor and cognitive performances were within the normal range. Both metabolites correlated highly with the motor score of the Unified Huntington’s Disease Rating Scale and with computed measurements of saccadic and tapping speed. Creatine reduction was also well correlated with performance in cognitive timed tasks and with the length of CAG expansion (r = −0.81).
Conclusion: The creatine signal appears to be an interesting marker for progression in HD and could be useful in assessing therapeutic outcome, particularly during the initial stages when most clinical indices are still within the normal range.
Huntington’s disease (HD) is a neurodegenerative disorder caused by an unstable CAG trinucleotide expansion in the Huntington gene (IT15).1 Striatal projection neurons and some cortical neurons are damaged selectively, despite wide expression of the abnormal protein huntingtin.2 To account for this selective vulnerability, energy metabolism failure and excitotoxicity have been considered as likely pathogenic mechanisms.3 Support for this theory comes from experimental models that reproduce either the clinical picture or the pathologic features of HD in animals,4,5 and from reports of mitochondrial defects in the caudate nucleus of HD patients.6,7 In vivo proton MRS provides a noninvasive means of assessing metabolites in brain. Jenkins et al.,8 using this technique, described an elevation of lactate—an indicator of energy metabolism impairment—in the occipital cortex of HD patients. The same group9 reported a favorable effect of coenzyme Q on cortical lactate levels, and more recently they reported10 a positive correlation in symptomatic patients between CAG length and several spectroscopic ratios. Taylor–Robinson et al.11 communicated significant alterations of the proton spectra in HD, including an increase in the glutamine-to-creatine ratio but no lactate increase. The increase of glutamine was interpreted as an indication of an excitotoxic mechanism. Apart from gaining insight into HD pathogenesis, the promise of proton spectroscopy would be to detect and to quantify metabolic dysfunction in relation to neurologic condition. With such a view, we investigated the correlation between the spectroscopic alterations of the basal ganglia of 10 HD patients and their performance on several motor and cognitive tasks.
Methods.
Patients.
Ten HD patients regularly attending the Movement Disorders Unit of Fundación Jiménez Díaz (University Autónoma, Madrid) and five age-matched control subjects (table 1) agreed to participate in this study. The protocol was approved by the ethics committee of the Fundación Jiménez Díaz, Madrid, Spain. Choreic patients were excluded deliberately; thus we had either presymptomatic gene carriers or advanced akinetic patients. Accordingly, two groups were defined: initial HD1 and advanced HD2. The HD1 group comprised four individuals considered asymptomatic (although in two of them soft motor signs were already apparent) who had undergone genetic predictive testing according to recommended guidelines of the World Federation of Neurology 1 to 3 years before entering this study. The HD2 group included six akinetic, young patients. As shown in table 1, HD2 patients had longer CAG expansions (>50) and earlier onset of the disease, and thus mean age was similar in both groups. None of the patients had ever received neuroleptic treatment.
Sample characteristics
Molecular analysis of the CAG expansion was performed in all patients as described previously.13
Neurologic assessment.
Motor impairment was rated according to the Unified Huntington’s Disease Rating Scale motor score.14 In addition, horizontal saccadic speed (with fixed predictable stimulus) measured with video nystagmo-oculography,15 and tapping speed on a computer keyboard (calculated for each hand as the time needed to make 40 pulsations) were computed.
Four timed tests sensitive to prefrontal dysfunction were selected from the full battery proposed in the Core Assessment Program for Intracerebral Transplantation in Huntington’s Disease protocol16:
1. Verbal fluency17: to measure the speed and ease of word generation for semantic (animals) and phonologic (p) categories
2. Trail-making tests A and B18: to measure the speed and flexibility of sequential shifting
3. Stroop test19: to rate the ability to inhibit automatic response and the capacity of shifting performance to changing demands
4. Symbol Digit Modalities Test20: to rate attention and visuomotor coordination with time limitation (90 seconds)
Folstein’s Mini-Mental State Examination21 was also included as a global indicator of intellectual function.
MRI and MRS.
Acquisition.
MRI and MRS were performed on a 1.5-T Signa Advantage unit (General Electric Medical Systems, Milwaukee, WI) using the standard quadrature head coil as both the transmitter and the receiver. MR studies and clinical evaluation were performed within the same month.
T2-weighted axial images of the basal nuclei were obtained with a fast spin echo sequence with a repetition time (TR) of 4,400 msec, an echo time (TE) of 119 msec, and a field of view of 22 cm. The slice orientation was parallel to the orbitomeatal line. The axial plane that cut through the mid-axial section of the caudate nucleus was chosen to measure the bicaudate ratio (BCR) as described.22,23
Axial images were used to prescribe regions of interest (ROIs) on the left striatum from which proton spectra were subsequently acquired. Care was taken to include in the ROI as much caudate nucleus as compatible with the absence of contamination from the adjacent ventricular horn (figure 1). Mean volume of the ROI was 4.54 cm3. Atrophy of the caudate nucleus forced us to define slightly smaller volumes in advanced patients (mean volume, 3.27 cm3) to minimize contribution from other structures.
Figure 1. MR images depict the selected voxel for spectroscopic acquisition (above) and the resulting proton spectra in a representative case of each group (below). (A) Control subject. (B) Asymptomatic gene carrier (patient with initial Huntington’s disease (HD1). (C) Akinetic patient (HD2). Marked atrophy of the striatum and reduction of metabolite signals can be observed in HD2 (C) and milder changes but already notable in HD1 (B). The thickness of the spectroscopic voxel was slightly larger than that of the depicted MRI slices, extending into the adjacent slice above and below. These two slices were examined to ensure the absence of CSF contamination in the volume of interest. Cho = choline-containing compounds; Cr = creatine; NAA = N-acetylaspartate.
Proton spectra were acquired with the PROBE procedure of the General Electric spectroscopic package (release 5.6, Milwaukee, WI), working in a double-echo spin mode (probe p) and using a TR of 1,500 msec and a TE of 288 msec.24 In all subjects, prescan automatic optimization resulted in both 6 Hz or less of full width at half height of the unsuppressed water peak and percentages of water suppression higher than 98%. A total of 256 or 512 acquisitions were accumulated, depending on ROI dimensions. In addition, the probe procedure accumulated automatically 16 nonwater-suppressed signals that were of use as an internal reference (discussed later).
Data processing.
Acquired signals consisting of two frames of raw data (nonwater suppressed and water suppressed) were transferred to a Sun workstation (SPARC station 2, SunMicrosystems, Mountain View, CA) and processed with the spectral analysis software from General Electric (SA/GE version 3.0, Milwaukee, WI). Processing of water-suppressed data consisted of eddy current correction25 (using the nonwater-suppressed peak as a reference), apodization (exponential multiplication, line broadening, −3.5 Hz; followed by hemigaussian multiplication line broadening, 10 Hz), zero filling from 2 to 8 K data, fast Fourier transformation, zero-order phase correction, and frequency assignment of 304 Hz (4.7 ppm) to the residual water peak. No baseline correction was required except for one HD1 patient in whom a linear tilt was performed. Processing of nonwater-suppressed data was performed in the same way, excluding the eddy current correction step.
The resulting spectra were analyzed by the fitting routine included in the SA/GE package (Marquardt– Levenberg algorithm), obtaining for each patient the series of gaussian curves that best fit either the metabolic peaks at 2.0 ppm (N-acetylaspartate), 3.0 ppm (creatine/phosphocreatine), and 3.2 ppm (choline-containing compounds) in the water-suppressed spectrum, or the peak from water in the corresponding nonwater-suppressed spectrum. For each ROI, the areas of the curves fitted to the metabolite peaks were referred to that of the respective water peak, which was used as an internal reference.26
Statistics.
Statistical analyses were performed with the aid of the Statistical Package for Social Sciences (SPSS; release 7.5 for Windows, SPSS Inc., Chicago, IL). Comparisons among groups (control subjects, HD1, and HD2) were performed using analysis of variance followed by Bonferroni’s multiple comparison post hoc test. Excluding control subjects, bivariate correlations were calculated between spectroscopic data, BCR, and clinical scores. The product of triplets in excess (CAGn − 35.5) × years of age was used27 to correlate neurologic and spectroscopic data with the severity of the genetic defect.
Results.
Neurologic data.
The results of the motor evaluation and cognitive performance are summarized in table 2. The scores of the HD1 group were not significantly different from control subjects. On the other hand, the HD2 patients had a poor performance on every test. Considering all HD patients, most clinical scores showed a significant decline correlated with the length of the expansion corrected for age (CAGn − 35.5) × years of age. Significant correlation coefficients between neurologic scores and CAG are presented in table 2.
Neurologic scores and their correlation with CAG, creatine, and NAA in Huntington’s disease (HD)
Spectroscopic data.
Representative images and spectra for each group are shown in figure 1, illustrating the atrophy of the striatum and the reduction of metabolite signals in the HD2 group (see figure 1C), and similar although milder changes in the HD1 group (see figure 1B). The results of spectroscopic metabolite quantification and BCR measurements are represented in figure 2, and the statistics are summarized in table 3.
Figure 2. Individual values of metabolites and bicaudate ratio (BCR) measurements. Metabolite values are ratios to water signal (×103). NAA = N-acetylaspartate; HD1 = patient with initial Huntington’s disease (HD); HD2 = patient with akinetic HD.
Spectroscopic and bicaudate ratio (BCR) statistics
NAA was decreased significantly in all HD patients (see figure 2). Mean reduction was 36% for the HD1 group and 66% for the HD2 group. NAA levels correlated with caudate atrophy as measured by BCR (r = −0.73, p = 0.01). NAA decrease was well correlated with motor scores but not with cognitive scores (see table 2) or CAG length (r = −0.45, p = 0.06).
Creatine levels were reduced significantly in HD patients, although individual values showed some overlap (see figure 2). The HD1 group showed a mean decrease of 30% and the HD2 group exhibited a mean decrease of approximately 60% of control levels. The decrease of creatine showed significant associations with all clinical scores excluding the mean computed tapping speed, the Trail Making Test, and mini mental status. Significant correlations are shown in table 2. Creatine correlated strongly with the length of the CAG expansion normalized for age (r = −0.81, p < 0.005).
With regard to the choline signal, only the HD2 group showed a significant decrease (50%) because the scatter of the data was considerable (see figure 2). No clear association with clinical or genetic indices was found for this metabolite.
BCR was grossly enlarged in the HD2 group (+138% of control subjects). The increase in the initial group (HD1) was also evident (+64% of control subjects), even if it did not reach statistical significance (mean difference, 0.042 ± 0.02; p = 0.1). As mentioned earlier, BCR correlated inversely with NAA but not with creatine (r = −0.5) or choline. The relationship with clinical scores was not significant.
Discussion.
We observed a remarkable reduction of both NAA and creatine in the basal ganglia of HD patients (see figure 2). Interestingly, metabolite changes were already significant in the group of patients without relevant neurologic impairment (HD1 patients; see tables 2 and 3⇑). Moreover, the changes in creatine levels correlated significantly with several motor and cognitive neurologic scores (see table 2) and with the length of CAG expansion.
Although increased CSF content in the localized volumes cannot be ruled out completely, it does not seem to represent a major contribution to the observed reductions in spectral peak intensities because relevant CSF contamination would affect all metabolites in parallel, which was not the case (see figure 2). In addition, the absence of obvious partial volume effects from ventricles was made certain in every study by visual inspection of all MRI slices covered by the spectroscopic voxel (see figure 1). Lastly, the values reported in table 3 for the different metabolites correspond with the ratios between their respective signal intensities and that of the water from the same voxel (see the discussion of data processing in Methods). In this way, the calculated metabolite-to-water ratios can be compared among the different subjects by taking the water signal as a valid internal reference. To assess metabolite changes it is mandatory to perform absolute quantification because using peak ratios can yield ambiguous results. For such a purpose the use of the water signal as an internal standard has been proposed,25 but calculating absolute metabolite concentrations in this way could result in some error because the concentration of water standard is not known precisely.28 In addition, because it is impractical to collect fully relaxed spectra, correction for T1 and T2 relaxation times are necessary. To our knowledge no striatal metabolite relaxation times have been reported. This led us to refer the intensity of metabolite signals to that of water, obtaining ratios that have an absolute value in themselves (although they still require the above-mentioned correction for striatal T1 and T2 to translate them into concentrations).
The decrease in NAA—the neuronal marker in the proton spectra—correlated inversely to the BCR, a good index of caudate atrophy in HD patients.22 Atrophy of the caudate nucleus as measured by MRI volumetry has been demonstrated to occur before clinical onset.29-31 In this study we found a significant reduction of NAA in the initial group (see table 3), but this metabolite showed no significant relationship to either the repeat length by age or by cognitive performance. A similar lack of association between caudate/basal ganglia atrophy and several neuropsychological scores has been reported.32,33
The available information from recent clinical, pathologic, and experimental studies supports a major role of the CAG expansion in the progression of the disease.27,34-37 Accordingly, most clinical scores measured in this study appeared associated significantly with the length of the expansion by age. These tests measure processing speed, which is considered a reliable indicator of disease severity and progression.
The creatine peak proceeds from the combined contribution of creatine plus phosphocreatine. Switching between both forms—from dephosphorylated to phosphorylated creatine and vice versa—constitutes an efficient way of reserving and obtaining energy as required. Several lines of evidence point to energy impairment as a pathogenic mechanism of neuronal death in HD.3 The early and progressive reduction of creatine/phosphocreatine levels associated with clinical and molecular evolutive indices (see figure 2 and table 2) supports the theory of energy impairment underlying striatal dysfunction. This has also been argued as an explanation for the increased levels of lactate found in proton spectra of HD patients.8-10 However, we did not find increased lactate contribution in any spectrum acquired from either the basal ganglia (this study) or the occipital lobe (unpublished data) in HD patients. Jenkins et al.8-10 has repeatedly reported increased lactate levels in HD patients, but in their recent study10 they acknowledge that “lactate levels are found to fall within the normal range in some patients with clinical HD” (p 1359).10 They report normal levels in all occipital and in five of eight striatal spectra obtained in presymptomatic gene carriers (none of them on psychoactive drugs). On the other hand, of 31 patients they include data from 8 patients on neuroleptics and 7 patients on antidepressants. Thus one factor that might be relevant is that our patients had never received neuroleptics and were not taking any psychoactive drug at the time of the study. In another study Taylor–Robinson et al.11 found no abnormalities in lactate in either the striatum, thalamus, or the temporal or occipital cortex. Their patients were also drug free for more than 3 months at the time of the study.
Another interesting result from the study by Jenkins et al.10 is a significant correlation between the length of the CAG expansion and NAA and lactate levels measured as ratios to creatine in symptomatic HD patients. There is a major methodological difference—preventing further comparisons—between both studies, which is the fact that they use creatine as a reference when, according to our results, creatine is affected profoundly in HD and the changes in this metabolite might account for significant variations in the ratios. Nevertheless, despite apparent discrepancies, both studies share a common conclusion: MRS in HD can detect metabolic abnormalities that correlate with the severity of the molecular CAG defect and that point to an alteration of energy metabolism. Very recently a detailed biochemical study in HD brains38 has confirmed a selective deficit of aconitase activity (a sensitive marker of excitotoxic cell damage) in the striatum (8 to 27% of normal) and cortex (52%), further supporting the role of energy failure and excitotoxicity in the pathogenesis of this disorder.
Considering the major spectroscopic alterations reported herein, the decrease of neuronal density (NAA), and the impairment of energy metabolism (creatine), proton spectroscopy appears to be a valuable tool for obtaining both structural and functional information from the basal ganglia in HD patients. As in other studies combining functional and volumetric measurements,39,40 atrophy and metabolic dysfunction seem closely related but might progress independently because metabolic markers appear better correlated with neurologic condition. We propose that the creatine signal might become a useful outcome measure, particularly in early HD stages when other indices are normal or only slightly impaired, making the assessment of progression or recovery very difficult.
Acknowledgments
Supported in part by a fellowship of the Conchita Rábago Foundation (R.S.-P.).
Acknowledgment
The authors thank the patients for their participation in this study and the technical staff of the MR unit for their assistance. We also thank Vicenta Sánchez Bernardos, who did the tapping registration; Francisco Esperón, who performed the oculographic measurements; and Pedro García Ruíz, for sharing these data with us. R.S.P. gratefully acknowledges a fellowship of the Conchita Rábago Foundation.
- Received October 6, 1998.
- Accepted in final form April 10, 1999.
References
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Hemiplegic Migraine Associated With PRRT2 Variations A Clinical and Genetic Study
Dr. Robert Shapiro and Dr. Amynah Pradhan