Progressive supranuclear palsy (PSP) is a neurodegenerative disease with motor, cognitive, and behavioral symptomatology. Cholinergic dysfunction is thought to underpin several key symptoms. There is known pathologic involvement of the corticobasal ganglia-thalamocortical loops in PSP, but little attention has been focused on potential thalamic dysfunction. Using autoradiography, we measured muscarinic M2 and M4 receptors in specific thalamic nuclei involved in the limbic and motor loops in patients with PSP (n = 11) and compared results from brain tissue of subjects with Lewy body dementias (including dementia with Lewy bodies and Parkinson disease with dementia, n = 31), Alzheimer disease (n = 22) and normal elderly control subjects (n = 27). In the thalamus M2 receptors were more abundant than M4 receptors and were most densely concentrated in the anteroprincipal (AP) and mediodorsal (MD) nuclei, which connect to limbic cortices. M2 receptor binding was reduced in the AP nucleus in PSP compared with control subjects and those with Lewy body dementias. M4 receptors were markedly reduced in the MD nucleus in those with PSP compared with control subjects. M4 receptors were also reduced in the subthalamic nucleus in patients with PSP. M4 receptor binding was reduced in the MD nucleus in the Lewy body dementia and Alzheimer disease groups compared with control subjects. There were no significant changes in the ventrolateral nucleus (motor). Cholinergic dysfunction within the AP and MD nuclei of the thalamus may contribute to behavioral and cognitive disturbances associated with PSP.
Progressive supranuclear palsy (PSP) is a neurodegenerative disease with motor and cognitive features for which there is currently no available treatment. Pathologically abnormal tau deposition is seen within the basal ganglia and related structures, including the thalamus, subthalamic nucleus, and brainstem, notably the pedunculopontine nucleus (PPN) (1-3). Although cholinergic deficits are believed to be responsible for some of the cognitive symptomatology in PSP, little is known about the cholinergic receptor status in a number of structures in the corticobasal ganglia-thalamic loop or their clinical significance. Five functionally distinct loops are involved in the clinical expression of PSP (4). Each loop remains segregated, and anatomically distinct thalamic nuclei mediate specific clinical features. The ventrolateral (VL) nucleus is a motor relay nucleus, whereas the anteroprincipal (AP) and mediodorsal (MD) nuclei are involved in limbic functions. There is significant atrophy and neuronal loss within the motor thalamus in PSP (2, 5), and functional imaging shows reduced thalamic glucose metabolism (6). The PPN and laterodorsal tegmental nucleus (LTN) send cholinergic projections to the thalamus and show neuronal loss in PSP. We are not aware of any neurochemical studies that have examined cholinergic receptor function in the PSP thalamus.
Muscarinic M1 receptors are found predominantly in the cortex and are in low density in the thalamus, whereas M3 and M5 receptors have a very restricted brain distribution, being mainly localized in the peripheral nervous system. We therefore measured muscarinic M2 and M4 receptors in the thalamus (AP, MD, and VL nuclei) and subthalamic nucleus (STN) in postmortem PSP brains and compared the findings with cases of Lewy body dementia (LBD) (a composite group comprising dementia with Lewy bodies [DLB] and Parkinson disease with dementia) and Alzheimer disease (AD) and controls. LBD and AD were chosen for comparison as they also have cholinergic deficits underpinning their key clinical features.
Materials and Methods
PSP cases (n = 11) were selected from the Newcastle Brain Tissue Resource (Newcastle University, UK) (n = 5) and Sara Koe PSP Research Centre (London, UK) (n = 6). All cases were of classic "Richardson's” phenotype antemortem (7). Control (n = 27), LBD (n = 31), DLB (n = 15), Parkinson disease with dementia (n = 16), and AD (n = 22) were selected from the Newcastle Brain Tissue Resource. Control cases had no record of neurologic or psychiatric disease and no significant pathologic changes. Demographic details are shown in Table 1. Clinical data were obtained either by prospective collection, for LBD and AD, or retrospective case note review for PSP cases. Consent for brain donation was obtained in accordance with the Local Research Ethics Committee (Newcastle and North Tyneside) and the London Multi-Centre Research Ethics Committee.
The right cerebral hemisphere was fixed and used for pathologic diagnosis, and the left cerebral hemisphere was sliced into 1-cm-thick coronal blocks, rapidly frozen in liquid Arcton cooled over liquid nitrogen, and stored at −70°C. Cases acquired in the last 3 years have been quickly frozen on copper plates maintained at −80°C. There is no difference in tissue quality between the 2 freezing methods. Sections were cut through the thalamus at coronal levels 16 to 18 to include the relevant nuclei (8). For autoradiography, 20-μm cryostat sections were cut and dried onto Vectabond (Vecta, Peterborough, UK)-coated slides before storage at −70°C. Before assay, slides were warmed to room temperature and air dried for 1 to 2 hours.
Autoradiographic Ligand-Binding Assays
Total and selective binding was determined in triplicate contiguous sections and nonspecific binding was established in one section by the addition of 2 μmol/L atropine (Sigma, Poole, UK). At room temperature sections were prewashed in a buffer (10 mmol/L KH2PO4 and 10 mmol/L Na2HPO4, pH 7.4) for 15 minutes to remove any endogenous ligand, for example, acetylcholine or residual potentially competing drugs. M2 and M4 combined receptor density was measured using [3H]AFDX 384 (PerkinElmer, Beaconsfield, UK), which labels both receptors (total binding), and in adjacent sections, selectively blocking M4 binding with dicyclomine (Sigma). After prewashing 4.8 nmol/L [3H]AFDX 384 either alone, with dicyclomine (10 nmol/L), or with atropine, was added and the solution was incubated for 1 hour at room temperature. Sections were washed in the same buffer 2 times for 2 minutes at room temperature before being dipped in ice-cold water, dried, and apposed to tritium-sensitive Hyperfilm (Amersham, GE Healthcare, Little Chalfont, UK) for 7 weeks.
Developing and Analysis
Films were developed after warming to room temperature (2 hours) before development using 500 mL of D19 for 5 minutes, stopped using 500 mL of 1% aqueous acetic acid for 1 minute, fixed using 500 mL of 25% Unifix for 6 minutes, and washed for 20 minutes in running water. Films were dried and binding was assessed by comparison to 3H autoradiographic microscale standards (Amersham) using MCID Elite image analysis (Imaging Research Inc, GE Healthcare, Interfocus, Linton, UK) to give binding in fmol/mg. The specific binding for combined M2 and M4 receptors was calculated by subtracting the nonspecific binding from the total. Similarly, M2 binding was determined by subtracting the nonspecific binding from the total plus dicyclomine. M4 was then found by taking the M2 binding value away from the total. Nonspecific binding was low at <5% of total binding.
Statistical analysis was performed using MINITAB (version 13). To ensure normality of data, results were logged before statistical analysis and the means were antilogged. Descriptive data are presented as geometric means and SDs. Statistical tests included one-way analysis of variance between disease groups, with post hoc analysis using Tukey's pairwise comparisons where indicated. Regression analysis was used to explore potential correlations between clinical variables, demographics, and neurochemical data.
In controls, M2 receptor binding was greatest in the AP and MD nuclei, with less avid binding in the VL nuclei and minimal binding in the STN (Table 2; Figs. 1, 2). There was reduced M2 binding in the AP nuclei in PSP compared with controls (−30%, p < 0.05) and LBD (−33%, p < 0.01). The other thalamic nuclei showed no difference in M2 binding among groups. M2 binding in the STN was lower in PSP than in controls and AD, but this difference was not statistically significant. The number of cases measured in different nuclei varied, as not all coronal levels were available in every case due to limitations in the supply of postmortem tissue.
In controls, M4 receptor binding was lower than M2 binding in all nuclei except the subthalamic nucleus (Table 3; Figs. 1, 2). In the PSP group, M4 receptor binding in the MD nucleus was markedly lower than that in controls (−65%, p < 0.01). Binding was also reduced in the AP nucleus in PSP, but this value was not statistically significant (−45%, p = 0.08) as the standard deviations overlapped. In the VL nucleus, the reduction in binding in all disease groups was not significant. Although M4 receptors were low in number in the STN, there was a marked reduction in binding in PSP cases compared with controls (−79%, p < 0.001). In the LBD and AD groups, M4 binding was reduced in all areas compared with controls, reaching statistical significance in the MD nucleus (LBD, −54%, p < 0.01; AD,−51%, p < 0.05).
Clinical and Demographic Correlations
There was no significant difference between age or postmortem (PM) delay among the disease groups and controls in the comparison of any of the nuclei measured. In the control group, M4 receptor binding in the STN was negatively correlated with PM delay (p = 0.001), but this was not apparent for any of the thalamic nuclei or for M2 binding in any area. STN M2 and M4 receptor binding did not correlate with PM delay in any disease groups. In PSP and LBD groups, disease duration had no effect on binding. In AD cases, M2 binding in the AP nucleus increased with disease duration (p = 0.038).
We found a marked reduction in cholinergic M2 receptor binding in the AP nucleus and in M4 receptor binding in the MD nucleus of the thalamus in PSP, with normal receptor binding in the VL nucleus. The thalamus receives 90% of its cholinergic input from the PPN and LTN (9). The MD nucleus also receives projections from basal forebrain cholinergic neurons (10). Thalamic afferents from the globus pallidus are under tonic GABAergic inhibitory control (and excitatory influence from the subthalamic nucleus), and there are also dopaminergic projections from the substantia nigra pars compacta. M2 receptors are the predominant muscarinic receptors in the thalamus and are probably distributed mainly as presynaptic autoreceptors where they inhibit neurotransmitter release (11, 12). Thalamic M4 receptors are likely to be mainly postsynaptic, as they are in the striatum (13, 14).
The reduction in M2 binding in PSP in the AP nucleus is therefore likely to represent loss of cholinergic afferents. It is interesting to speculate, however, why we did not detect a more global loss of M2 receptors in thalamic nuclei, as might have been predicted. There is evidence to suggest that anterior thalamic nuclei are innervated predominantly by the LTN (15) and the other nuclei by the PPN (16). Cholinergic loss is higher in the LTN than the PPN (17) in PSP, so this could account for a relatively restricted loss of M2 receptors to the AP nucleus. Inhibitory M4 receptors, although less abundant in the thalamus, showed more marked changes in PSP, especially in the MD nucleus. Despite moderate reductions in M4 receptors in the AP nucleus in PSP and similar reductions in AD and LBD, the differences were not statistically significant, probably because of large standard deviations, reflecting heterogeneity of disease in the cases studied. The standard deviations are larger in the M4 receptor densities than the M2, which may reflect more variable neuronal loss and atrophy in the thalamus than the marked and more consistent cell loss in the brainstem.
In the control thalamus, choline acetyltransferase (ChAT) and acetylcholinesterase staining (reflecting cholinergic neurons) is seen predominantly in the anterior and MD nuclei (10, 18). Within the MD nucleus, there are areas of densely staining patches surrounded by less densely stained matrix (10). The medial matrix projects to the orbitofrontal cortex and the patches to the dorsolateral prefrontal cortex (19). In a previous study, 5 patients with PSP showed reduced ChAT staining within the MD nucleus by 75% in the matrix and 60% in the patches (20). The MD and AP nuclei are involved predominantly in the "limbic” circuits. The AP receives afferents from the mammillothalamic tract and projects to the cingulate gyrus, whereas the MD receives inputs from the amygdala and hypothalamus and projects to the dorsolateral prefrontal and orbitofrontal cortex (4, 21). Cholinergic dysfunction in limbic projections could therefore contribute to behavioral, cognitive, and apathetic symptoms in PSP. The MD nucleus also has connections to the premotor cortex (10) and to the frontal and supplementary eye fields (22), suggesting that its dysfunction could have more widespread clinical implications in PSP.
We found normal M2 receptors in the VL nucleus of the thalamus, with a nonsignificant reduction in M4 receptors in all diseases, despite reported cell loss and atrophy in the VL nucleus of thalamus in PSP (5, 23). The VL, along with the ventroanterior nucleus, forms part of the motor circuit linking the basal ganglia with the motor cortex. It receives GABAergic afferents from the globus pallidus (24) and sends projections to the supplementary motor (25), the premotor, and the motor cortex (26). One possible explanation for the apparently unchanged M2 binding is that neuronal loss and atrophy in the VL nucleus may be associated with increased synaptic density.
M4 receptors were reduced in the MD nucleus in LBD and AD. ChAT was previously reported to be reduced in the MD nucleus in AD by 34% in the matrix and 36% in the patches (20). In contrast, normal ChAT and acetylcholinesterase levels were found in the MD and anterior thalamic nuclei in another study in AD (18). No reductions were noted in ChAT levels in the MD nucleus in DLB, although there was reduced enzyme activity in Parkinson disease with dementia (27). In DLB, binding to nicotinic α6 cholinergic receptors was reduced in the VL nucleus (28), but, to our knowledge, there have been no other studies looking specifically at cholinergic transmission in this nucleus. Overall, there are moderate reductions in the M4 receptors in the limbic related thalamic nuclei in LBD and AD. The normality of M2 receptors, especially in the AP nucleus where PSP shows reduced binding, could, in part, explain the higher frequency of apathy found in PSP compared with AD and LBD. Compensatory changes, such as axonal sprouting, may take place in AD and LBD, but not in PSP, as evidenced by increasing numbers of M2 receptors with disease duration in the AP nucleus in AD (29).
The STN is involved in the indirect pathway of the basal ganglia. It receives afferents from the globus pallidus externa and cortex (for a summary, see Reference 30) and projects mainly to the pallidum and subthalamic nucleus in rat (31). It is known to be overactive in Parkinson disease (32). Our study found a significant reduction in M4 receptors in the STN in PSP, probably reflecting neuronal loss, but the overall binding in this nucleus is very low, suggesting that the cholinergic influence on the STN is not mediated via M2 or M4 receptors.
M4 binding in the STN decreased with increasing PM delay in controls, but this was not replicated in the thalamic nuclei, and, as there was no difference in PM delay among groups, we believe that this delay did not affect the analysis. Furthermore, PM delay has previously been shown not to affect muscarinic binding (33, 34).
This article is the first to examine thalamic cholinergic receptors in PSP. Overall, there is marked cholinergic receptor loss in specific thalamic nuclei involved in the limbic circuits in PSP, which may be associated with cognitive, behavioral, and possibly broader symptomatology. Our findings provide further evidence of widespread but specific cholinergic deficits in PSP. Prospective clinicopathologic studies are needed to further explore correlations between thalamic cholinergic receptor changes and specific clinical features in PSP. The generally negative clinical trial results to date with both cholinesterase inhibitors and nonselective muscarinic agonists (35) suggest that a simple single neurotransmitter replacement approach is unlikely to be beneficial in PSP, unlike more encouraging outcomes in AD and DLB. Although further trials of more specific cholinergic agents may be warranted, with particular assessment of limbic symptoms such as behavioral, motivational, and cognitive functions, the widespread degenerative changes and neurotransmitter loss in PSP may require a shift in therapeutic strategy toward addressing the fundamental pathophysiologic process before a significant therapeutic impact may be made in this disease.
We thank Mary Johnson for her help with the tissue preparation at the Newcastle Brain Bank; and also all those at the Sara Koe Research Centre, London, in particular Linda Parsons and Susan Stoneham, for their help with the tissue samples and data collection. We are very grateful to the patients and control subjects who donated their brains for this research.
- alzheimer's disease
- lewy body disease
- cell nucleus
- cholinergic agents
- lewy bodies
- neurodegenerative disorders
- muscarinic acetylcholine receptor
- subthalamic nucleus
- progressive supranuclear palsy
- thalamic nuclei
- thalamic structure
- dementia due to parkinson's disease
- cognitive impairment
- brain tissue