Abnormal striatal and thalamic dopamine neurotransmission

Abstract

Objective:

To determine whether changes in D₂ receptor availability are present in carriers of genetic mutations for primary dystonia.

Methods:

Manifesting and nonmanifesting carriers of the DYT1 and DYT6 dystonia mutations were scanned with [¹¹C] raclopride (RAC) and PET. Measures of D₂ receptor availability in the caudate nucleus and putamen were determined using an automated region-of-interest approach. Values from mutation carriers and healthy controls were compared using analysis of variance to assess the effects of genotype and phenotype. Additionally, voxel-based whole brain searches were conducted to detect group differences in extrastriatal regions.

Results:

Significant reductions in caudate and putamen D₂ receptor availability were evident in both groups of mutation carriers relative to healthy controls (p < 0.001). The changes were greater in DYT6 relative to DYT1 carriers (−38.0 ± 3.0% vs −15.0 ± 3.0%, p < 0.001). By contrast, there was no significant difference between manifesting and nonmanifesting carriers of either genotype. Voxel-based analysis confirmed these findings and additionally revealed reduced RAC binding in the ventrolateral thalamus of both groups of mutation carriers. As in the striatum, the thalamic binding reductions were more pronounced in DYT6 carriers and were not influenced by the presence of clinical manifestations.

Conclusions:

Reduced D₂ receptor availability in carriers of dystonia genes is compatible with dysfunction or loss of D₂-bearing neurons, increased synaptic dopamine levels, or both. These changes, which may be present to different degrees in the DYT1 and DYT6 genotypes, are likely to represent susceptibility factors for the development of clinical manifestations in mutation carriers.

GLOSSARY

ANOVA

= analysis of variance;

BFM

= Burke-Fahn-Marsden;

CN

= caudate nucleus;

FWE

= family-wise error rate;

GPe

= external pallidum;

GPi

= interal pallidum;

RAC

= raclopride;

ROI

= region of interest;

SOR

= striato-occipital ratio;

THX

= trihexyphenidyl;

VL

= ventrolateral tier nuclei.

Multiple lines of evidence support the role of altered striatal DA neurotransmission in primary dystonia.^1,2 Nonetheless, imaging studies have revealed only minimal reductions in striatal D₂ receptor binding in patients with idiopathic dystonia^3,4 and in nonmanifesting carriers of the DYT1 dystonia mutation.⁵ Indeed, similar reductions have been described in DYT1 dystonia patients at postmortem.¹ It is not known whether similar abnormalities are also present in primary dystonia associated with other genetic mutations.

Experimental models of dystonia have yielded varied results with regard to the striatal DA dynamics. Increased striatal DA turnover was found in a transgenic murine DYT1 model regardless of behavioral phenotype.^6,7 However, in this model, striatal DA levels were reduced in affected but not unaffected animals.⁶ Abnormalities in DA turnover have also been described in mutant DYT1 knock-in and TOR1A knock-down mice.^8,9 By contrast, in a different murine mutant DYT1 model, basal striatal DA levels, as well as presynaptic and postsynaptic measures of DA signaling, were unchanged.^10,11 A conditional knockout mouse with selective torsinA deficits in the cerebral cortex had motor abnormalities reminiscent of dystonia, but no changes in striatal DA metabolism.¹² Similar inconsistencies have also been found in non-DYT1 experimental models of dystonia.^13–16

In the current study, we sought to identify changes in D₂ receptor availability linked to the presence of clinical manifestations in a cohort of DYT1 and DYT6 gene carriers. We expanded upon our earlier [¹¹C] raclopride (RAC) PET investigation of nonmanifesting DYT1 carriers⁵ to determine whether greater reductions in striatal D₂ receptor binding are present in manifesting DYT1 carriers, and whether comparable changes occur in manifesting and nonmanifesting carriers of the DYT6 haplotype.

METHODS

Subjects.

We studied the following groups of subjects:

1. DYT1: 12 nonmanifesting (NM DYT1, age 55.5 ± 14.8 years [mean ± SD]) and 9 clinically manifesting (MAN DYT1, 46.1 ± 15.8 years) mutation carriers.

2. DYT6: 4 nonmanifesting (NM DYT6, 38.4 ± 22.5 years) and 8 clinically manifesting (MAN DYT6, 31.7 ± 13.7 years) linkage carriers.

3. Controls: 13 healthy volunteers (C, 47.3 ± 17.1 years).

DYT1 and DYT6 mutation carriers were recruited and genetically tested through the Mirken Department of Neurology at Beth Israel Medical Center in New York. The controls were recruited by advertisement among North Shore University Hospital. This control group was demographically representative of these communities (including 25% nonwhite participation), and the inadvertent inclusion of DYT1 or DYT6 mutation carriers was unlikely by chance. We defined nonmanifesting mutation carriers as gene-positive individuals who did not display signs or symptoms of dystonia. In the DYT1 genotype, all subjects also exceeded the age of maximal likelihood for clinical onset. Written informed consent was obtained from all participants under protocols approved by the Institutional Review Boards of the participating institutions.

The clinical characteristics of the affected subjects are detailed in table e-1 on the Neurology® Web site at www.neurology.org. In brief, five out of the nine MAN DYT1 subjects had generalized dystonia, two had a multifocal phenotype, and the remaining two had segmental and focal dystonia, respectively. In MAN DYT6, three subjects had generalized dystonia, three subjects had focal dystonia, and two subjects had segmental dystonia. Burke-Fahn- Marsden movement scale scores were relatively higher in DYT1, with a median of 21 (range 1–53) compared to a median of 6 (range 2–51) in DYT6. However, this difference did not reach significance (p = 0.17, Kruskal-Wallis). Five of the MAN DYT1 and four of the MAN DYT6 subjects were on chronic trihexyphenidyl, which was withheld 12 hours prior to imaging.

Exclusion criteria for all subjects were as follows: 1) past history of additional neurologic illness; 2) prior or current exposure to neuroleptic agents or drug use; 3) past medical history of hypertension, cardiovascular disease, or diabetes mellitus; 4) abnormal MRI; and for controls and nonmanifesting subjects only: 5) current use of psychotropic medication; 6) abnormal neurologic examination; 7) past history of dystonic symptoms. Data from nine of the NM DYT1 carriers have been reported previously.⁵

There was a difference in age across groups (p = 0.04, analysis of variance [ANOVA]) with NM DYT1 younger than MAN DYT6 (p < 0.05, Tukey Kramer HSD); differences in age between the other groups were not significant. Because older age in nonmanifesting gene carriers is an inherent feature of the study design,¹⁷ an age-correction was incorporated for all group comparisons.

Positron emission tomography.

All subjects fasted overnight before PET imaging. Medications were withheld in the affected subjects (table e-1) for at least 12 hours prior to scanning. Subjects received 15 mCi of RAC by IV injection while positioned in the gantry. Three-dimensional dynamic RAC scans were acquired over 70 minutes (7 × 10 minutes postinjection) on a GE Advance PET camera (General Electric, Milwaukee, WI) as described previously.⁵ Data processing was performed using statistical parametric mapping (SPM99; Wellcome Department of Cognitive Neurology, London, UK) implemented in Matlab (Mathworks, Sherborn, MA). Dynamic scans from each subject were realigned with each other using a least squares approach and a six parameter rigid-body transformation. Each of the realigned dynamic images was then integrated into single slices covering the entire striatum. Regional binding measures were obtained from the single-slice image of the 10-minute frame between 60 and 70 minutes postinjection.

Regions of interest (ROIs) were placed bilaterally on the caudate nucleus, putamen, and the occipital cortex using an automated algorithm, blind to subject identity and genotype. Specific RAC binding in the caudate and putamen was estimated by the striato-occipital ratio (SOR) ([ROI activity − occipital activity]/occipital activity) with an explicit correction for age.⁵

Striatal SOR values for each subject were divided by the age-predicted normal mean such that a value of 1 equals 100% age-appropriate normal binding. Analogously, a value of 0.75 reflected a 25% reduction from the age-appropriate normal binding. Group comparisons were additionally performed using an entirely data-driven voxel-based approach in SPM99. SOR images were produced on a voxel basis over the same time interval as the ROI analysis and anatomically transformed into Talairach space based on the mean image from the early frames. Normalized SOR images of all groups were then compared voxel-wise with age as a covariate of no interest. Group differences were considered significant at p < 0.05, incorporating the family-wise error rate (FWE) correction at peak voxel as a means of controlling for type I error (see http://www.sph.umich.edu/∼nichols/Docs/FWEfNI.pdf). In addition to highly significant localized changes, more widespread differences in regional RAC binding were examined at a lower threshold (p < 0.01, uncorrected), with a cluster cutoff of 100 contiguous voxels. The results of these analyses were considered descriptive and were not treated as statistically significant findings. Coordinates were reported in the standard anatomic space developed at the Montreal Neurological Institute.

Group comparisons of the ROI values were performed using ANOVA, followed by post hoc tests incorporating the Tukey-Kramer HSD adjustment for multiple comparisons. These tests, with age as a covariate, were also used for post hoc analysis of the significant clusters extracted from voxel-based analyses. Patient characteristics for the two genotypes were compared with Fisher exact test. Symptom severity was assessed according to the Burke-Fahn-Marsden (BFM) Dystonia Rating Scale. These ratings and the duration of symptoms were correlated with RAC binding values by computing Pearson product moment correlation coefficients. In each analysis, results were considered significant for p < 0.05. All statistical analyses were conducted on PCs running JMP software (SAS Institute Inc, Cary, NC) for PC and SPSS (SPSS, Inc., Chicago, IL) software.

RESULTS

ROI analysis.

Age-corrected RAC binding values for left and right caudate and putamen (ROIs) in the MAN and NM DYT1 and DYT6 groups as well as the control group are presented in table 1. No significant differences were present between left-sided and right-sided measurements of homologous striatal subregions in all of the subgroups (p > 0.5 for all paired t tests within and across subgroups for caudate and putamen, respectively). Left and right caudate/putamen values were highly intercorrelated in both controls and mutation carriers (controls: caudate: R = 0.88, p < 0.0001; putamen: R = 0.94, p < 0.0001; DYT1/DYT6 mutation carriers: caudate: R = 0.75, p < 0.0001; putamen: R = 0.82, p < 0.0001). Thus, the left-right means of age-corrected RAC binding values were used for subsequent group comparison.

Table 1 [¹¹C]-raclopride binding in dystonia mutation carriers and controls

Two-way ANOVA (figure 1) revealed a main effect of genotype (DYT1, DYT6, C) for both regions (caudate: F [2,41] = 18.9, p < 0.001; putamen: F [2,41] = 19.3, p < 0.001). By contrast, the effect of phenotype (MAN, NM, C) across groups did not reach significance (caudate: F [2,41] = 3.56, p = 0.07; putamen: F [2,41] = 2.86, p = 0.10) and there was no phenotype × genotype interaction (p > 0.9). Post hoc analysis (figure 1) showed that the RAC binding reductions were most pronounced in DYT6 carriers relative to controls (p < 0.001 for both caudate and putamen; Tukey-Kramer HSD), with DYT1 carriers taking an intermediate position (DYT1 vs C: p < 0.01 for both regions; DYT1 vs DYT6: p < 0.001 for both regions).

Figure 1 Age-corrected measures of D₂ receptor availability (mean ± SE) in caudate and putamen regions of interest (see Methods and table 1)

Significant reductions in D₂ receptor availability were evident in manifesting (MAN) and nonmanifesting (NM) carriers of the DYT1 and DYT6 dystonia mutations.

In manifesting mutation carriers, age-corrected caudate/putamen RAC binding measures did not correlate with symptom duration or with BFM dystonia severity ratings (R² < 0.25, p > 0.2 for all correlations). In these patients, we also assessed the effect of chronic anticholinergic medication on the RAC binding measures. Patients on chronic trihexyphenidyl (THX) treatment tended to have lower caudate binding (p = 0.13) than their untreated counterparts. This difference was not evident for putamen binding (p > 0.3). The frequency of treatment with THX was similar for patients of either genotype (5/9 in DYT1 and 4/8 in DYT6; p > 0.5). This indicates that the effects of genotype on striatal RAC binding were not driven by medication. Likewise, BFM motor sum scores did not differ across genotypes (p > 0.2).

Voxel-based analysis.

Voxel-based searches of the RAC PET scans over the whole brain confirmed and expanded upon the results of the ROI analysis. Reductions in D₂ receptor availability (p < 0.05, FWE-corrected) were present in the combined group of DYT1 and DYT6 carriers relative to controls (table 2A). At this stringent peak threshold, the striatal changes (p < 0.01, FWE-corrected) were localized to the lateral putamen bilaterally (figure 2, top). There was also evidence of reductions in RAC binding (p = 0.002, FWE-corrected) in the right ventrolateral thalamus (figure 2, bottom). Post hoc analysis of binding values for this region revealed an effect of genotype (F [2,39] = 10.31, p = 0.003) but not phenotype (F [2,39] = 0.09, p > 0.7). Thalamic RAC binding was reduced (p < 0.05, Tukey Kramer HSD) in both MAN and NM carriers of each of the two mutations. Moreover, the thalamic RAC binding abnormalities were relatively greater in DYT6 as compared with DYT1 carriers (p < 0.05, Tukey Kramer HSD).

Table 2 Brain regions with significant group differences in [¹¹C]-raclopride binding

Figure 2 Voxel-based comparison of D₂ receptor availability of the combined group of DYT1 and DYT6 mutation carriers and control subjects

Significant regional differences were identified in the putamen and thalamus (left). Bar diagrams (right) illustrate binding values from each significant cluster (arrows) in nonmanifesting (NM) and manifesting (MAN) DYT1 and DYT6 carriers and healthy controls (C). A statistical parametric map (SPM) of significant binding differences between gene carriers and controls (DYT1 + DYT6 > C) was superimposed on a single-subject T1-weighted MRI template. The color scale represents a T threshold of 4.2, p < 0.05 (FWE-corrected). Age was used as a nuisance variable in the SPM analysis so an explicit age correction was not employed in the post hoc comparisons.

At a less stringent peak threshold, but using a conservative spatial threshold (p < 0.01, uncorrected, with a cluster cutoff of 100 contiguous voxels), there was evidence of more widespread reductions in D₂ binding. In the combined group of mutation carriers, these changes extended from the putamen to include the caudate nucleus as well as the globus pallidus and ventral thalamus (table 2B).

Separate comparisons of each of the four subgroups (MAN and NM DYT1; MAN and NM DYT6) with controls using the stringent peak threshold revealed reductions in putamen and thalamus D₂ receptor binding in both MAN and NM DYT6 genotype carriers (table 2A). In the DYT1 genotype, these reductions were only present at the more lenient peak threshold (table 2B). Similarly, direct comparison of the DYT1 to the DYT6 genotype showed more pronounced RAC binding deficits in DYT6 in the posterior putamen (table 2A, figure 3). From a spatial standpoint, the reductions were also more widespread in DYT6 carriers, with a total of 10,775 abnormal voxels (table 2B). By contrast, in DYT1 mutation carriers, the volume of reduced binding encompassed only 477 voxels. As in the ROI-based analyses, there was no effect of phenotype either within or across genotypes. There were also no regions with increased D₂ receptor binding in DYT1 and DYT6 mutation carriers relative to controls.

Figure 3 Voxel-based comparison of D₂ receptor availability in DYT1 relative to DYT6 dystonia mutation carriers

Significantly greater reductions in putamen binding were present in DYT6 relative to DYT1 carriers. A statistical parametric map of significant regional differences between the two groups of mutation carriers was superimposed on a single-subject T1-weighted MRI template. The color scale represents a T threshold of 4.2, p < 0.05 (family-wise error rate–corrected).

DISCUSSION

In this study we show that reductions in striatal and thalamic D₂ receptor binding reflect primarily trait features of primary dystonia. In an earlier report we described reductions in striatal D₂ receptor availability in NM DYT1 carriers.⁵ The current study revealed that these reductions were not influenced by the presence or absence of clinical manifestations or by their severity. Notably, significant reductions in striatal and thalamic binding were also present in DYT6 haplotype carriers. These changes were relatively greater than those observed in DYT1 carriers. Likewise, in DYT6, these reductions were not influenced by clinical penetrance. In aggregate, the findings support the notion that alterations in DA neurotransmission in dystonia gene carriers are a susceptibility factor for the development of disease manifestations.

Our findings of abnormal reductions in striatal D₂ receptor binding in primary dystonia patients are consistent with prior PET and SPECT studies.^3,4 These observations are also compatible with postmortem analysis of brain tissue from a DYT1 patient,¹ as well as with findings from several experimental models.^6,7 Although significant, striatal D₂ receptor binding reductions are not profound. It has therefore been suggested that dysfunction of D₂-bearing striatal projection neurons to the external globus pallidus (i.e., the classic indirect pathway) represents “a permissive factor” for the development of primary dystonia.² In other words, an additive or potentiating effect of several subthreshold lesions (i.e., “a dual hit”) leads to the development of symptoms.¹⁸ This notion has recently received further support from experimental studies in genetic and pharmacologic animal models, suggesting that penetrance is mediated by an abnormal interaction of striatal and cerebellar functional activity.^13,19 For DYT1, the combined effects of mutant torsinA on neuronal maturation²⁰ and on the resistance of DA neurons to stress²¹ may be needed for symptoms to occur. Indeed, we have shown abnormalities linked to both processes in human carriers of dystonia mutations. Microstructural alterations in cerebellar outflow pathways, presumably developmental in origin, have been found to differentiate penetrant and nonpenetrant mutation carriers.¹⁹ That said, abnormalities in DA neurotransmission in mutation carriers, as demonstrated in the current study, may constitute a susceptibility factor for the development of clinical manifestations in these subjects.

Reductions in striatal RAC binding, such as those observed in this study, are usually interpreted as reflecting dysfunction or loss of dopaminoceptive striatal projection neurons. Nonetheless, these changes are also compatible with increases in synaptic DA in gene-positive individuals.^22,23 These possibilities can be differentiated through DA release studies. Such investigations, however, may not be an option in dystonia mutation carriers because of the potential emergence/exacerbation of symptoms in these subjects.^24,25 Given that reductions in striatal D₂ binding in dystonia have also been demonstrated with less displaceable radioligands,^3,4 it is possible that DYT1 and DYT6 carriers harbor different alterations in DA neurotransmission. The highly localized, significant reductions in caudate and putamen RAC binding that we observed are consistent with focal loss of dopaminoceptive neurons in these regions. By contrast, the presence of more widespread binding changes raises the possibility of abnormal elevations in synaptic DA, perhaps as a consequence of increased corticostriatal excitability. Both mechanisms may apply in our cohort of mutation carriers. Indeed, the relative contributions of these factors may differ for the DYT1 and DYT6 genotypes. It is tempting to speculate that because the DYT6 genotype is associated with greater reductions in putamen D₂ receptor availability as well as glucose metabolism,¹⁷ the substantial RAC binding changes seen in these subjects are likely to reflect actual receptor loss in this region. By contrast, the relatively modest RAC binding changes seen in DYT1 carriers, with relative putaminal hypermetabolism, may be better explained by abnormal increases in synaptic DA levels.

The voxel-based analysis confirmed and extended the finding of reduced striatal D₂ receptor availability that was observed using predetermined ROIs. This completely data-driven approach also revealed binding reductions in the ventrolateral thalamus in dystonia mutation carriers, irrespective of clinical penetrance. The role of abnormal thalamic DA neurotransmission in dystonia has not been studied, despite the critical role of this structure in the pathophysiology of the disorder.^26–29 Abnormalities of thalamic signaling and metabolic activity have been interpreted as reflecting functional alterations in afferents to this structure from the globus pallidus and related structures.^28–32 Indeed, the ventral thalamus has been the target of surgical interventions for medically refractory dystonia.^33,34 That said, the role of the thalamus in dystonia is by no means clear in that lesions in this area can also produce dystonic manifestations in both humans and experimental animal models.³⁵

Substantial dopaminergic input to the human ventral thalamus has been recognized.^36–38 In primates, DAT-positive terminals are evident at thalamic interneurons,³⁸ which are mainly inhibitory.³⁹ It is conceivable that these DA projections normally serve to reduce noise during sensorimotor processing. In dystonia, alterations in DA neurotransmission at this level may interfere with this function and facilitate the occurrence of abnormal movements. Interestingly, the dopaminergic innervation of the thalamus is markedly greater in the primate than in the rodent.³⁸ Although the observed thalamic D₂ binding reductions need to be confirmed, it is tempting to relate some of the limitations of transgenic animal models in mimicking human dystonia to the absence of a thalamic dopaminergic system in rodents. Moreover, we note that similar deficits in thalamic D₂ receptor binding are present in schizophrenia, which like dystonia, is thought to have a developmental etiology. These abnormalities may reflect impaired dopaminergic modulation of thalamocortical processing in the two conditions.

AUTHOR CONTRIBUTIONS

M.C. conducted all statistical analyses.

ACKNOWLEDGMENT

The authors thank Claude Margouleff for technical support and Toni Flanagan for help with study coordination and copyediting.

Address correspondence and reprint requests to Dr. David Eidelberg, Center for Neurosciences, The Feinstein Institute for Medical Research, North Shore-Long Island Jewish Health System, 350 Community Drive, Manhasset, NY 11030 david1@nshs.edu.

Supplemental data at www.neurology.org.

Supported by the NIH (R01 NS 047668) and the General Clinical Research Center of The Feinstein Institute for Medical Research (M01 RR018535).

Disclosure: The authors report no disclosures.

Received November 21, 2008. Accepted in final form March 13, 2009.