Abstract
Dysfunction of neostriatal medium spiny neurons (MSNs) is hypothesized to underlie late-stage motor complications of Parkinson disease (PD). The authors demonstrate shortened dendrite length of MSNs that was similar in four regions of neostriatum in late-stage PD. In contrast, MSN dendrite spine degeneration was unevenly distributed with the greatest loss in caudal putamen. The authors propose that these structural changes in MSN may contribute to late-stage motor complications of PD.
Nigrostriatal dopaminergic projections synapse largely on the shaft of dendritic spines of neostriatal medium spiny neurons (MSNs), whereas excitatory projections from cerebral cortex and thalamus synapse mostly on the heads of these same spines.1 The pivotal initiating event in Parkinson disease (PD) is degeneration of the nigrostriatal dopaminergic projections to MSNs, leaving unopposed excitatory input that may lead to secondary, late-stage degeneration of MSNs. One hypothesis is that MSNs may die or lose their axonal projections in late-stage PD; however, others have shown this does not occur to any significant extent.2 Alternatively, MSN degeneration in late-stage PD could result from loss of spines, removing the pharmacologic target for dopamine-replacement therapies, without overt MSN death. Only a single study has demonstrated MSN dendrite atrophy in advanced PD.3 We are unaware of any verification of this seminal finding or investigation of the regional distribution of MSN dendrite atrophy in PD. Changes in MSN spine density also have not been investigated, leaving open the question of whether MSN spine density also decreases in late-stage PD or whether there are compensatory increases along with dendrite atrophy. Here we address these issues.
Methods.
Patients and controls.
The University of Washington Human Subjects Committee approved all experiments. Inclusion criteria were patients diagnosed with PD and subsequently followed by a neurologist, treatment with l-dopa, no diagnosis of dementia, and neuropathologic changes of PD but that did not meet consensus criteria for dementia with Lewy bodies or Alzheimer disease (AD).4 Nine patients (three women and six men; 13 ± 4 years disease duration; 69 ± 10 years of age at death; 1,185 ± 127 g brain weight) with autopsies at the University of Washington between 1986 and 2003 met these criteria; all were included in this study. Inclusion criteria for controls were no diagnosis of neurologic or psychiatric illness and only age-related changes on post mortem examination of brain without brainstem or neocortical Lewy bodies or changes of AD.4 Seven controls (three women and four men; 75 ± 11 years of age at death; 1,198 ± 106 g brain weight) with autopsies at the University of Washington between 1987 and 2003 were matched for age and time in formalin but otherwise were randomly selected. The time in fixative for PD cases averaged 9.7 years (median, 12.5 years) and for controls averaged 9.6 years (median, 12.0 years; Mann–Whitney U test, p > 0.05). Post mortem intervals for PD cases and controls were 6.4 ± 3.1 hours and 8.6 ± 3.9 hours (Mann–Whitney U test, p > 0.05).
Animals.
The University of Washington IACUC approved all procedures. Approximately 8-week-old male C57Bl/6 mice were euthanized, and their brains quickly removed, fixed in formalin for 5 days, and then processed for Golgi–Cox staining exactly as human tissue described next.
Tissue staining and morphometry.
Portions of fixed putamen (~0.3 cm3) were taken from three consecutive tissue slabs: immediately caudal to the anterior commissure (precommissural), at the anterior commissure, and immediately rostral to the anterior commissure (postcommissural). Similarly sized tissue was dissected from the caudate nucleus at the level of the anterior commissure. Tissue was sectioned by vibratome in the coronal plane at 80 μm unless otherwise indicated. Golgi–Cox staining was performed exactly according to the method of Glaser and Van der Loos.5 Six to eight well-impregnated MSNs were randomly selected for morphometric analysis from each slide by an observer blinded to diagnosis, according to methods described previously.6 Morphometric measurements from single sections were made using Neurolucida (MicroBrightField, Williston, VT), according to the methods described by others.6
Results.
We observed interpretable and reproducible staining of human neostriatal MSNs in archived specimens with tissue sections that were up to 80 μm but not thicker. Using similar techniques, others have shown that quantification of dendritic arbor for human prefrontal cortex pyramidal neurons is independent of section thickness for the range of 101 to 200 μm.7 We pursued a similar analysis with our protocol in mice. Formalin-fixed cerebrums from four mice were sectioned at 80, 150, or 200 μm, stained, and analyzed exactly as human sections. Dendrite length and spine density (mean ± SEM) were 1,149 ± 35 μm and 17.4 ± 0.4 spines per 10 μm for 80-μm thickness, 1,102 ± 46 μm and 20.5 ± 1.8 spines per 10 μm for 150-μm thickness, and 1,103 ± 82 μm and 18.0 ± 2.0 spines per 10 μm for 200-μm thickness. One-way analysis of variance for dendrite length or spine density for these three thicknesses was not significant. All subsequent data are derived from analysis of 80-μm sections.
Correlation of time in fixative with dendrite length or spine density in all three regions of putamen in controls yielded correlation coefficients of 0.24 (p > 0.05) or 0.05 (p > 0.05). Similarly, no significant correlation was found between time in fixative and dendrite length or spine density in each region of putamen in patients with PD (correlation coefficients ranged from 0.01 to 0.30; p > 0.05 for all correlations) or in caudate (not shown).
Dendrite length was reduced in all regions in patients with PD compared with controls (p < 0.001; figure 1). Spine density in these same MSNs remained essentially constant across regions of putamen in controls but progressively decreased across precommissural to postcommissural regions in patients with PD. Bonferroni-corrected post-tests showed that spine density was less in patients with PD than in controls for commissural and postcommissural regions (p < 0.001) but not precommissural region (p > 0.05). Thus, MSNs in putamen of patients with PD show relatively uniform dendrite shortening but reduction in spine density only in caudal regions.
Figure 1. Morphometric values from three different regions of putamen from patients with Parkinson disease (PD) or controls. Six to eight neurons were evaluated in each section, and the average value was obtained for each region in each individual. These individual averages were then combined to yield a grand average (±SEM) for each region in patients with PD and controls. (A) Dendrite length: two-way analysis of variance (ANOVA) for dendrite length (F[42,2,1]) had p < 0.001 for “Control vs PD” but p > 0.05 for “region”; interaction between these two terms had p < 0.005. Bonferroni-corrected post-tests showed that dendritic length was less in patients with PD than in controls for all regions (p < 0.001). (B) Spine density: two-way ANOVA for spine density (F[42,2,1]) had p < 0.001 for “Control vs PD,” p < 0.01 for “region,” and p < 0.05 for interaction between these two terms. Bonferroni-corrected post-tests showed that spine density was less in patients with PD than in controls for commissural and postcommissural regions (p < 0.001) but not precommissural region (p > 0.05).
Dendrite length for controls and patients with PD in caudate was similar to the corresponding region of putamen (p > 0.05; figure 2). Importantly, unlike putamen, caudate MSN spine density in patients with PD was not different from controls (p > 0.05). Thus, caudate MSNs at the level of the anterior commissure in patients with PD show dendrite atrophy, like their counterparts in putamen, but no decrease in spine density.
Figure 2. Morphometric values from the caudate at the level of the anterior commissure from patients with Parkinson disease (PD) or controls. Six to eight neurons were evaluated in each section, and the average value was obtained for each individual. These individual averages were then combined to yield a grand average (±SEM) for patients with PD and controls. (A) Dendrite length: two-way analysis of variance (ANOVA) for dendrite length (F[25,1,1]) had p < 0.0001 for “Control vs PD” and p > 0.05 for “caudate vs putamen”; interaction between these two terms had p > 0.05. Bonferroni-corrected post-tests also showed that dendrite length was no different in caudate vs putamen for controls (p > 0.05) and patients with PD (p > 0.05). (B) Spine density: two-way ANOVA for spine density (F[25,1,1]) had p < 0.001 for “Control vs PD,” p < 0.001 for “caudate vs putamen,” and p > 0.05 for interaction between these two terms. Bonferroni-corrected post-tests of spine density in controls vs patients with PD had p > 0.05 for caudate and p < 0.01 for putamen.
Discussion.
The mechanisms that underlie late-stage motor complications of PD are not fully clarified; however, there is wide agreement that persistent alterations in striatal output by MSNs are central to their development.1 Here we have demonstrated structural changes in MSN dendrites in late-stage PD that include uniform dendrite atrophy and regional loss of dendritic spines, providing data that may explain in part the decreased efficacy of dopamine-replacement therapies in advancing PD.
In 1988, McNeill et al.3 presented the first and only demonstration of MSN dendritic atrophy in PD. Animal models of PD have also reported that striatal dopamine denervation results in decreased MSN dendrite length and spine density.8 Despite using different tissue thicknesses and image analysis methods than the present study, our work confirms and significantly extends their observations by including spine density and dendritic length and by determining changes in dendrites across different neostriatal regions. We found a similar degree of dendritic atrophy across all examined neostriatal areas in patients with late-stage PD. In contrast, spine density showed a topographic gradient of degeneration with no significant change in caudate or rostral putamen to severe spine loss in the caudal putamen.
We have hypothesized that after loss of mesostriatal dopaminergic input, MSN spines experience unopposed glutamatergic input that renders them susceptible to subsequent excitotoxic damage. From this perspective, it is noteworthy that the gradient we observed for severity of MSN dendritic spine degeneration roughly matches that for dopaminergic denervation in PD.9 Dysregulation of basal ganglia signaling, perhaps related to the structural changes described here, is amenable to manipulation by gene transfer and currently is undergoing clinical evaluation.10
Footnotes
-
Supported by the Nancy and Ellsworth Alvord Endowment, the Cheng-Mei Shaw Endowment, and P01 NS044282.
Received July 28, 2004. Accepted in final form September 30, 2004.
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
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.
