|Home | About | Journals | Submit | Contact Us | Français|
Huntington disease is characterized clinically by chorea, motor impairment, psychiatric manifestations and dementia. Atrophy of the striatum is the neuropathological hallmark of Huntington disease and previous studies have suggested that striatal atrophy correlates more closely with motor impairment than with chorea. Motor impairment, as measured by motor impairment score, correlates with functional disability in Huntington patients, but chorea does not. In this study, we investigate the relation between neuronal loss and these motor features.
We conducted neuropathological and stereologic assessments of neurons in putamen and subthalamic nuclei in Huntington patients and age-matched controls. In putamen, we estimated the total number and volume of medium spiny neurons labeled with dopamine- and cAMP-regulated phosphoprotein 32 kDa (DARPP32). In subthalamic nuclei, we estimated the total number of neurons on Hematoxylin/Eosin -Luxol/Fast Blue stains.
In putamen of Huntington disease, immunohistochemistry showed DARPP 32 neuronal atrophy with extensive disruption of neurites and neuropil; stereologic studies found significant decreases in both the number and size of DARPP32 neurons; we also detected a significant reduction of overall putamen volume in Huntington patients compared with controls. In subthalamic nuclei, there was a mild but significant neuronal loss in Huntington group. The loss of neurons in putamen and subthalamic nuclei and the putaminal atrophy were significantly correlated with the severity of motor impairment, but not with chorea.
Our findings suggest that neuronal loss and atrophy in striatum and neuronal loss in subthalamic nuclei contribute specifically to the motor impairment of Huntington, but not to chorea.
Huntington disease (HD) is caused by an expanded CAG repeat in the huntingtin gene, coding for a long stretch of polyglutamines in the huntingtin (htt) protein1–5. Factor analysis of the two major motor scales used in HD (the quantitative neurological exam and the motor UHDRS) indicate that there are two major components of the movement disorder--chorea and “motor impairment” 6, 7. Previous studies suggest that chorea often predominates early in the disease; while motor impairment is more characteristic of advanced HD and is associated with greater pathological severity and length of CAG repeats 8, 9.
Neuroanatomical and biochemical studies indicate that motor abnormalities of HD are due to degenerative changes of the striatum, globus pallidus, subthalamic nucleus (STN), and substantia nigra pars reticulata 10–12. Previous studies have examined the loss of neurons in these anatomical regions, but few have examined in detail the correlations between the loss of neurons in any of these specific nuclei and motor abnormalities 13, 14.
Neostriatal atrophy is the morphologic hallmark of HD and the basis for grading its neuropathological severity at autopsy 12. Nevertheless, the pathology of HD extends beyond the neostriatum to extrastriatal regions 4, 15. Despite a wealth of information on the neuropathology of HD, the pathological substrate of either chorea or motor impairment in HD remains unclear. We previously showed that the motor impairment observed in individuals with HD strongly correlated with the severity of striatal atrophy at autopsy measured on the Vonsattel scale 12, whereas chorea showed no significant correlation 9. These results strongly suggested that motor impairment in HD may be a reflection of neostriatal neuronal degeneration.
To test this hypothesis, we used design-based stereology to estimate the number and size of neurons in the putamen in autopsy cases of HD and age-matched controls. Because it is difficult to distinguish atrophic putaminal neurons from reactive astrocytes histologically, we labeled medium spiny neurons (MSNs) with an antibody against dopamine- and cAMP-regulated phosphoprotein 32 kDa (DARPP32), a marker for MSNs. Animal studies have shown that about 95% of MSNs are DARPP32 positive in striatum 16. Stereological estimates of absolute neuronal numbers in a nucleus require to examine the entire structure. For this reason, we chose to examine the putamen, which is fairly compact, and not the entire striatum since the caudate has a more complex geometry. Moreover, data from the literature indicate that the degeneration of the putamen in HD is comparable to that of the caudate 4, 17, 18. In addition, we compared cell numbers in the caudate and putamen in two HD cases and one control (data not shown) and demonstrated that, at least in these cases, neuronal numbers are as severely reduced in the putamen as in the caudate. Thus, we believe that neuronal counts from the putamen are representative of the degeneration of the striatum as a whole.
Experimental and clinical studies of human and non-human primates indicate that lesions limited to the striatum are not sufficient to cause chorea, while lesions of the STN or its connections may lead to chorea or hemiballismus 19–22. Thus, we assessed the total number of neurons in the STN in HD.
The HD autopsy brains available for studies of the putamen included 6 males and 8 females, ages from 35 to 86 years (average 60.3±16.1 years), with duration of disease from 5 to 33 years (average 18.9 ± 8.4 years). The controls included 6 subjects, 4 males and 2 females, ages from 17 to 81 years (average 48.2 ±24.9years). The brains for the STN study included 10 HD cases, a subset of the 14 brains used for assessment of putamen, ranging from 35 to72years (average 54.5±14.3 years) and 7 age-matched controls, ages from 21–65 years (average 44.7 ±19.9 years).
HD brains were from patients followed by the Johns Hopkins University (JHU) Research Center Without Walls for Huntington’s Disease. Control brains from subjects without history of neurological disease were obtained from the Brain Research Center of the JHU under institutional review board-approved protocols. Demographic and clinical information on the subjects is presented in Tables 1 and and22.
All subjects in the HD group had their diagnosis genetically confirmed and were followed with periodic neurologic and psychiatric examinations and neuropsychological testing. The neurologic examination included a motor impairment (MIS) and chorea scores, and detailed description of clinical evaluation has been previously reported 9. The time interval between last clinical examination and death ranged from 5 to 147 months with average of 58.6±41.4 months. All clinical protocols used in the study were approved by the JHU Human Participants Ethics Committee and informed consent was obtained from all participants and families.
Autopsy and brain tissue dissection were performed at JHU Neuropathology. The diagnostic brain tissue blocks were processed and stained as described previously 23 and diagnosis and grading of severity followed Vonsattel 12. Tissue sections of the substantia nigra were stained with H&E and examined to exclude Lewy body disease. All HD cases in our study were free of Lewy bodies, although most showed mild pigment incontinence. One HD case, with the longest CAG repeat size (65) and t youngest age at death (35 years), had moderate neuronal degeneration and glosis of the substantia nigra.
The whole putamen was dissected from fixed coronal brain slabs and sliced into 1-cm thick tissue blocks for processing and paraffin-embedding. The putamen has well defined dorsal, lateral, and ventral boundaries. The medial boundary was arbitrarily defined as a line bisecting the internal capsule. The entire putamen was serially cut at 50 μm, and sections for stereology were selected with a random start and systematic interval (Q 60 sections) and immunostained with DARPP32 antibody. The entire STN was dissected in four 3-mm coronal blocks from the subthalamic region and rostral midbrain, and serial sections were cut at 50 μm; sections for stereology were selected with a random start and systematic interval (Q 20 sections) and stained with H&E-Luxol/Fast Blue.
A subset of 10–12 tissue sections encompassing the whole putamen was selected and immunostained with antibody against DARPP32 (Cell signaling, 1:50), using biotinylated goat anti-rabbit as secondary antibody, and developing immunoreactivity with Vectastain ABC peroxidase kit followed by staining with 3,3-diaminobenzidine (DAB). Detailed description of the immunohistochemistry method can be found in supplementary information.
The putamen was outlined in each of the 10–12 immunostained sections laterally by the external capsule and medially by the globus pallidus and the internal capsule (Fig 1A&B) 12. We examined a subset of 10–12 slides from the STN stained with H&E-Luxol/Fast Blue, and the boundaries of the STN were outlined according to Hamani et al 24. Using a Stereo-Investigator system (MBF Bioscience, Williston VT), we estimated neuronal numbers and measured the volumes of neuronal cell bodies and nuclei described in previous studies 25, 26, Detailed description of the stereology method can be found in supplementary information.
All statistical analyses were performed using SigmaStat 3.1 (Systat, Software, Inc). Detailed description of statistical analyses can be found in supplementary information.
DARPP32 immunostaining in controls revealed abundant putaminal neurons which displayed intense perikaryal staining and extensive neuritic arborization; axons showed their typical varicosities, and dendrites studded with abundant spines. By contrast, remaining DARPP32 neurons in HD patients were sparse and displayed dystrophic neurites, loss of dendritic spines, and cell body atrophy. In cases of HD Vonsattel grades III and IV, DARPP32 neurons showed scalloping and irregularities of their cellular membranes, fainter immunostaining intensity, axonal fragmentation, extensive disruption of neurites, and attenuation of the neuropil. Similar changes, but of lesser magnitude were present in cases of HD Vonsattel grade II (Fig. 1C &D).
We found significant loss of DARPP32 neurons in the putamen of HD patients across all VS grades: 58% loss in grade II, 73% in grade III, and 76% in grade IV (Fig. 1E). We also observed a significant reduction in the overall volume of the putamen in HDs compared to controls: 42% decrease in grades II, 53% in grade III, and 56% in grade IV (Fig 1F). In the HD group, there was no significant difference of DARPP 32 neuronal number in putamen or the overall volume of the putamen among VS grades II through IV.
The loss of DARPP32 neurons in the putamen of HD patients was accompanied by a 40% (p<0.01) reduction in the volume of the remaining neurons (Fig. 1G). However, we found no difference in nuclear size of DARPP32 neurons between HD patients and controls (Fig. 1H). This resulted in sharp differences in the ratio of nuclear volume to neuronal volume between the two groups: 0.27±0.05 in controls and 0.41±0.07 in HD.
We found a small but significant neuronal loss in the STN in HD patients compared to controls (20%, p< 0.05) (Fig. 2A); we also detected a trend for reduction in overall STN volume in HD (Fig 2B). In addition, four HD patients (1 VS II and 3 VS III) had significant loss of neurons in the putamen, but not in the STN. This observation suggests that the pathological changes in STN trail the changes in putamen.
We next investigated whether the loss of DARPP32 neurons correlated with motor impairment score (MIS). We found a significant inverse correlation between the number of DARPP32 neurons and MIS, i.e., HD patients who displayed more neuronal loss had a more severe motor impairment and higher MIS, (Fig. 3A). In contrast, there was no correlation between DARPP32 neuronal loss and the severity of chorea (Fig. 3B). There was a trend for direct correlation between DARPP32 neuronal volume and chorea score (r=0.527, p=0.053), indicating the possibility that neurons with larger remaining volume were associated with higher chorea score. There was, however, no correlation between neuronal volumes and MIS (r= −0.34, p=0.49). Multiple linear regression analysis showed that only DARPP 32 neuronal loss had a significant power to predict the change of MIS (p=0.049), whereas neuronal atrophy and CAG repeat size failed to show a significant effect on prediction of MIS.
We found that the overall reduction in volume of the putamen in HD correlated strongly and significantly with MIS (p=0.004) (Fig. 3C). There was a positive trend for correlations between the volume of the putamen and the number (r=0.471, p=0.089, Fig. 3D) and sizes (r= 0.382, p=0.178) of DARPP 32 neurons. These results suggest that loss of neurons and their decrease in size contribute, at least in part, to the atrophy of the putamen.
There was significant neuronal loss (20%, p< 0.05) (Fig. 2A) in the STN in HD patients compared to those in controls. Interestingly, we found a significant inverse correlation between neuronal numbers and MIS (p=0.032) (Fig. 3E), but no correlation with chorea (Fig. 3F), suggesting that neuronal loss in STN may contribute to motor impairment but not to chorea in HD. In addition, four HD patients (1 VS II and 3 VS III) had significant loss of neurons in the putamen but not in the STN
Our stereological study of post-mortem HD brains reveals substantial neuronal loss in the putamen and mild neuronal loss in STN. The loss of neurons correlates significantly with motor impairment but not with chorea. This observation supports the hypothesis that motor impairment in HD is caused by neuronal loss, and is consistent with the hypothesis that chorea may be a manifestation of neuronal dysfunction rather than of neuronal loss in early stages of HD 3.
Recently, MRI studies have reported significant reduction of striatal volume in HD including premanifest and moderate HD brains 27, 28. The PREDICT-HD and TRACK-HD longitudinal analyses of striatal atrophy demonstrate steady changes from premanifest gene carrier to severe HD29, 30. Neuroimaging studies in patients with HD have also suggested an early atrophy in striatum and extrastriatal regions 31. The correlation between motor manifestations and striatal neuronal degeneration in HD, demonstrated by our investigation, therefore provides a novel perspective that complements imaging studies relevant to the pathogenesis of motor dysfunction in HD.
In our study, the reduction in overall volume of the putamen appears strongly correlated with motor impairment (r= −0.708, p=0.004), and both the loss and atrophy of DARPP 32 neurons contribute to putaminal atrophy in HD (See results). The pattern of DARPP 32 neuronal loss paralleled the pattern of putaminal atrophy (Fig. 1E & 1F). In addition to our quantitative study, qualitative examination of DARPP32 immunolabeling revealed a spectrum of morphologic abnormalities in the remaining putaminal neurons. These changes were characterized by perikaryal shrinkage and loss of stained dendritic processes in early stages of HD, whereas axonal-dendritic processes were almost completely destroyed in VS III and IV cases (Fig 1C & 1D). The remaining DARPP 32 neurons in later stages of the disease appeared isolated as they lost their connection with their neighboring nerve cells. Together, the massive loss of both DARPP 32 neurons and their neurites could result in disruption of local circuitry within the striatum and also in its efferent connections and thereby lead to motor dysfunction in HD patients. Our observations are consistent with the Golgi cytoarchitectonic study by Graveland et al 32. Given the better visualization of dendrites and spine structures with Golgi, it could be helpful to combine Golgi impregnation with DARPP32 immunohistochemistry to study MSNs in HD.
The mechanisms underlying chorea are not fully understood, but both the striatum and STN have been implicated. It has been proposed that chorea might reflect a progressive period of neuronal dysfunction before cell death or an imbalance between different neuronal populations such as enkephalin vs substance P 15, 33–35. Indeed, our study documents striking and massive degeneration of DARPP 32 neurons in the putamen, and the remaining atrophic DARPP32 neurons with variable cell sizes in the putamen of HD patients may represent neurons in different stages of neuronal degeneration. Interestingly, we observed a strong trend for correlation between larger volume of remaining DARRP32 positive neurons and a higher chorea score (r=0.527, p=0.053), consistent with the notion that chorea is perhaps the manifestation of neurons that are dysfunctional or in the early stages of degeneration.
Studies from both human and non-human primates indicate that chorea can be produced by destruction of neurons in STN 36, 37. Depending on their size and location, lesions in STN can induce hypokinetic or hyperkinetic movement disorders (hemiballismus and choreoathetosis) 38, 39. There are no previous stereological studies of the STN in HD, but Lange et al reported a 25% loss of STN neurons using a non-stereological method 40. Using stereology, we identified a mild (20%, p<0.05) but significant neuronal loss in STN in HDs that correlated significantly with motor impairment score but not with chorea (Fig. 3F). However, four HD patients (1 VS II and 3 VS III) had significant loss of neurons in the putamen but not in the STN. This observation suggests that the pathological changes in STN trail the changes in putamen. Furthermore, three of these four patients did have marked motor impairment. Thus, striatal atrophy appears sufficient to cause motor impairment even in the absence of neuronal loss in the STN. Our results further indicate that the loss of STN neurons appears in relatively later stage of HD, compared to striatal MSNs, which are more vulnerable in HDs 3, 4.
In our current study, we did not evaluate neuronal loss in the globus pallidus or degeneration of subtypes of striatal projection neurons, (enkephalin- and substance P-containing neurons), and striatal interneurons in HD 33–35. Therefore, it may be interesting for future studies to identify the neurochemical nature of the remaining neurons in the putamen.
An important caveat in the interpretation of our results is the long average time interval between the last clinical evaluation and death. In a previous study of 100 postmortem cases, we found no correlation between Vonsattel grades of motor impairment score and interval between last neurological exam and death. We interpret this to indicate that clinical features of HD are fairly stable at end stage of disease, making clinical-pathological correlations feasible even given relatively infrequent exams in the period preceding death9. Therefore, in the current study, even though there was a substantial interval between the last neurological examination and autopsy in some cases, we were still able to detect the significant correlation between DARPP 32 neuronal loss and MIS.
In this study in which the 14 HD cases exhibited a Vonsattel neuropathological severity ranging from grade II to grade IV (Table 2), we only observed a trend for inverse correlation between DARPP32 neuronal loss and Vonsattel grades (r= −0.464, p=0.091). HD patients who had longer CAG repeats had a tendency to exhibit more severe DARPP32 neuronal loss (r= −0.283, p=0.37); but there was no correlation between age of onset (motor symptoms onset) and DARPP32 neuronal loss (r=0.012, p=0.973).
In conclusion, our findings indicate that neuronal loss and atrophy of the putamen correlate significantly with motor impairment but not with chorea in HD. In advanced HD, there is also a modest loss of neurons in the STN, but this loss does not appear as necessary for the development of motor impairment. Our observations support the hypothesis that motor impairment in HD is caused by neuronal loss, and are consistent with the notion that chorea may be a manifestation of neuronal dysfunction rather than of neuronal loss in early stages of HD.
We gratefully acknowledge Ms. Wanda Stirling for her dedicated technical assistance, Ms. Karen Wall for administrative support, Dr. Lee Martin for providing research reagents, Dr. Mark West for helpful discussion on stereology methods, and the NICHD Brain & Tissue Bank for Developmental Disorders at the University of Maryland for providing control post mortem human brain samples. This research was supported by the Johns Hopkins University (JHU) Research Center Without Walls for Huntington’s Disease (NINDS NS016375), and the JHU Alzheimer’s Disease Research Center (NIA P50AG05146.)
Author’s RolesZhihong Guo contributed to the conception, organization and execution of the project; statistical analysis, and wrote the first draft of the manuscript and implemented multiple reviews. Gay Rudow contributed to the stereological analyses. Olga Pletnikova contributed to the organization and execution of the project. Barbara Crain, Kari-Elise Codispoti and Brent A. Orr conducted neuropathological autopsies and formulated diagnoses. Wenzhen Duan contributed to the manuscript writing and review. Russell Margolis and Adam Rosenblatt were the clinicians who followed the patients in the Huntington’s disease clinic. Christopher A. Ross participated in the clinical assessments, conception and initiation of the project, manuscript writing and review. Juan C. Troncoso contributed to the conception, organization, and execution of the project, and participated in the preparation and review of the manuscript. All authors reviewed the manuscript.
Juan C. Troncoso and Barbara Crain received grant support from NINDS (Research Center Without Walls for Huntington’s Disease- PO1NS16375) and NIA (Johns Hopkins University Alzheimer’s Disease Research Center -P50AG05146). Christopher A. Ross received grant support from NINDS (PO1NS16375). Wenzhen Duan received grants from the Hereditary Disease Foundation and CHDI Foundation. All other authors report no disclosures. Adam Rosenblatt currently is a full time employee at Virginia Commonwealth University School of Medicine. All other authors are currently full time employees of the Johns Hopkins University School of Medicine.