| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Neuroscience Research Group, University of Quebec at Trois-Rivieres, Trois-Rivieres, Quebec, Canada; and
Department of Cell Biology, Center for Models of Human Disease, IGSP, Duke University, Durham, North Carolina, USA
1Correspondence: M.G.C.: Department of Cell Biology, Center for Models of Human Disease, IGSP, Box 3287, Duke University, Durham, NC, 27710, USA; E-mail: caron002{at}mc.duke.edu or M.C.: Neuroscience Research Group, University of Quebec at Trois-Rivieres, C.P. 500, Trois-Rivieres, Quebec G9A 5H7, Canada. E-mail: cyrmi{at}uqtr.ca
ABSTRACT
An expansion in the CAG repeat of the IT15 (huntingtin) gene underlies the development of Huntington’s disease (HD), but the basis for the specific vulnerability of dopamine-receptive striatal neurons remains unclear. To examine the potential role of the dopamine system in the emergence of pathological conditions in HD, we generated a double mutant mouse strain with both enhanced dopamine transmission and endogenous expression of a mutant huntingtin gene. This strain was generated by crossing the dopamine transporter knock-out mouse, which exhibits a 5-fold elevation in extracellular dopamine levels in the striatum and locomotor hyperactivity, to a knock-in mouse model of HD containing 92 CAG repeats. These double mutant mice exhibited an increased stereotypic activity at 6 months of age, followed by a progressive decline of their locomotor hyperactivity. Expression of the mutated huntingtin did not alter dopamine or its metabolite levels in normal or dopamine transporter knock-out mice. However, the mutant huntingtin protein aggregated much earlier and to a greater extent in the striatum and other dopaminergic brain regions in the hyperdopaminergic mouse model of HD. Furthermore, the formation of neuropil aggregates in the striatum and other regions of hyperdopaminergic HD mice was observed at 4 months of age, well before similar events occurred in normal HD mice (12 months). These findings indicate that dopamine contributes to the deleterious effects of mutated huntingtin on striatal function, and this is accompanied by enhanced formation of huntingtin aggregates.—Cyr, M., Sotnikova, T. D., Gainetdinov, R. R., Caron, M. G. Dopamine enhances motor and neuropathological consequences of polyglutamine expanded huntingtin.
Key Words: polyglutamine disease • basal ganglia • chorea • hyperkinetic disorders • Huntington’s disease • medium spiny neurons
HUNTINGTON’S DISEASE (HD) is a fatal neurodegenerative disorder usually manifested around midlife with chorea, dementia, and a variety of neuropsychiatric conditions such as agitation, mood disorders, irritability, and even psychosis (1
2
3)
. This hyperkinetic disorder is directly related to the expansion of an unstable CAG repeat in exon 1 of the IT15 gene (4)
. CAG repeats beyond 36 predispose individuals to develop HD with a complete penetrance. The expanded HD allele is autosomal dominant, and symptoms will inevitably appear in a gene carrier (5)
. However, phenotypic expressions such as age of onset, disease duration, and the emergence of physical and mental symptoms are highly variable, even in individuals with the same length of CAG repeats in their IT15 gene (5
6
7
8)
.
The protein product of the IT15 gene, huntingtin, is a protein normally found in the cytoplasm, where it associates with vesicular structures and microtubules (9)
. The function of the huntingtin is not fully understood, and the critical processes determining deleterious effects of its expanded form have yet to be definitely identified. For example, despite ubiquitous expression of huntingtin throughout the brain and other tissues, GABAergic neurons expressing dopamine (DA) receptors in the striatum predominantly degenerate in HD patients, with layers of cerebral cortex being affected to a much lower extent and striatal interneurons (mostly NAPDH- and ChAT-positive) being spared (10
11
12)
. These observations are difficult to reconcile unless there are factors in addition to the mutation in the IT15 gene that enhance the vulnerability of the striatal projection neurons.
DA input from the ventral midbrain neurons has long been suspected to play a role in the pathophysiology of HD. Medium spiny GABA neurons in the striatum receive the densest dopaminergic innervations in the brain, and DA is recognized for its potential toxicity to striatal neurons (13
, 14)
. Clinical data support this contention, as DA-depleting agents reduce chorea in HD patients whereas administration of L-DOPA enhances dyskinetic symptoms (15
16
17
18
19
20
21)
. Recently, in vitro studies have documented that striatal primary culture containing a portion of the human huntingtin gene with expanded CAG repeats are more susceptible to DA-related reactive oxygen species (ROS) and degeneration induced by an excessive stimulation of DA receptors (22
, 23)
. However, whether changes in the efficacy of DA transmission can alter the consequences of mutated huntingtin expression in a physiologically relevant animal model of HD remain unclear.
To explore the role of DA in the behavioral and cellular consequences of an expanded CAG repeats in the IT15 gene in vivo, we generated a mouse model of HD with a persistently elevated striatal dopaminergic tone. This strain was generated by mating the DA transporter knock-out (DAT–/–) mice, which have 5-fold elevation in striatal extracellular DA levels (24
, 25)
, to a knock-in mice with 92 CAG repeats (24
, 26)
. Characterization of this unique hyperdopaminergic mouse model of HD directly demonstrated that persistently enhanced DA transmission exacerbates locomotor abnormalities in this mouse model of HD and accelerates the formation of aggregates of mutant huntingtin in the striatal projection neurons without provoking neuronal death. This study implies that DA may contribute to the striatal deterioration induced by mutant huntingtin and provides a theoretical basis for the clinical utility of a strategy of suppressing DA neurotransmission in HD.
MATERIALS AND METHODS
Mice
The mouse strain with increased DA neurotransmission and endogenous expression of mutant IT15 gene (Hdh-Q92) has been generated by mating mice with heterozygous deletion of the DA transporter (24)
to mice with heterozygous targeted insertion of a chimeric human-mouse exon 1 with 92 CAG repeats (STOCK Hdhtm4Mem/J, The Jackson Laboratory, Bar Harbor, ME, USA). These mice were intercrossed for six generations. Mice were housed in an animal care facility at 23°C on a 12 h light/12 h dark cycle with food and water provided ad libitum. Animal care was in accordance with the Guide for Care and Use of Laboratory Animals (NIH publication 865–23, Bethesda, MD, USA) and approved by the Institutional Animal Care and Use Committee.
Behavioral assessments
Locomotor activity was evaluated using an automated Omnitech Digiscan apparatus (AccuScan Instruments, Columbus, OH, USA) under illuminated conditions as described previously (27)
. All behavioral experiments were performed between 10:00 AM and 5:00 PM. Mice were observed individually at 5 min intervals for 1 h. To ensure that the activity period was recorded rather than the reactivity period, we excluded the first 15 min of each session from the assay. Locomotor activity was measured in terms of the total distance covered, and stereotypy time refers to the total time animals exhibit stereotypic behavior (repetitive beam breaks of a given beam with an interval of less than 1 s).
Histological assessments
Mice were anesthetized with chloral hydrate (400 mg/kg, i.p.) and perfused transcardially with ice-cold saline (0.9% NaCl), followed by ice-cold 4% paraformaldehyde in 0.1 M borax buffer (PFA), pH 9.5. Brains were postfixed overnight in 4% PFA, immersed for a few hours in 10% (w/v) sucrose/4% PFA solution, frozen in isopentane over dry ice, and kept at –80°C. Twenty micron tissue sections were prepared using a cryostat and kept in PBS solution at 4°C until used. Free-floating sections were immunostained using the following primary antisera directed against: polyglutamine (dilution 1:10000, 1C2, MAB1574, Chemicon, El Segundo, CA, USA), GFAP (dilution 1:2000, MAB360, Chemicon), TH (dilution 1:1000, MAB5280, Chemicon), activated caspase-3 (dilution 1:20, AB3623, Chemicon), NeuN (dilution 1:200, MAB377, Chemicon), and aggregated mutant huntingtin [dilution 1:2000, electron microscopy (EM)-48, gift from X. J. Li]. The sections were incubated with the appropriate labeled secondary antibody (Ab) (Vector Laboratories, Burlingame, CA, USA). Note that some of the brain sections were also incubated for 10 min at room temperature with a marker of cell nucleus, Hoechst dye 33342 (Molecular Probes, Eugene, OR, USA), at a dilution of 1:10,000 in PBS. Sections were then slide-mounted using vectashieldTM (Vector Laboratories) and cover slipped. A Zeiss confocal microscope (LSM-510) was used for sections viewing and images capturing. The EM48 immunostaining revealed nuclear staining, nuclear aggregates, and neuropil aggregates. The categorizations were determined by counting the number of nuclear staining or nuclear aggregates as revealed by the EM-48 Ab for every 100 positive neurons labeled by the NeuN Ab. For each structure, positive neurons sampled in a 0.31 µm3 (0.125 mmx0.125 mmx20 µm; 40 x objective) area in both hemispheres of three different sections per animal were considered. Anatomical landmarks such as aspect, size, and situation of the anterior commissures, corpus callosum, and ventricles were used to ensure that levels studied were similar within and between groups. The category definitions follow, with percents representing the average number of nuclear staining or nuclear aggregates per neurons: –, no cells (0%); +, few cells (1–25%); ++, scattered cells (25–50%); +++, numerous cells (50–75%); ++++, almost all cells (75–90%). Categorization of the neuropil aggregates was performed similarly to nuclear staining or nuclear aggregates, except that the total numbers of neuropil aggregates were counted in a 0.05 µm3 area (0.05 mmx0.05 mmx20 µm; 100 x objective) in both hemispheres of three different sections per animal: –, no neuropil aggregates; +, few (1–10); ++, moderate (10–50); +++, dense (50–100); ++++ very dense (100–150). All sections were coded, and quantitative analyses of specimens were performed without the knowledge of genotypes.
For TUNEL assays (Roche Diagnostics, Mannheim, Germany), brain sections were mounted on microscope slides, desiccated under vacuum overnight, and fixed in 4% PFA. Slides were then treated with proteinase K (50 µg/ml in 0.1 M Tris, pH 7.5, and 50 mM EDTA pH 8.0) and the protocol from the manufacturer (Roche Diagnostics) was followed. Note that a positive control was performed for every assay as recommended by the manufacturer.
Biochemical and neurochemical analyses
For filter trapping assays, the mouse striatum was rapidly dissected out, frozen in liquid nitrogen, and kept at –80°C prior to proteins extraction. Tissues were homogenized in a Tris/SDS buffer (50 mM Tris, pH 7.5, 50 mg/ml SDS) supplemented with a cocktail of protease inhibitors (Sigma, St. Louis, MO, USA). Protein concentrations were measured using a DC-protein assay (Bio-Rad, Hercules, CA, USA). Equal amounts of proteins (200 µg) were transferred to nitrocellulose membranes (0.2 µ, Schleicher and Schuell, Dassel, Germany) using a microfiltration apparatus (Bio-Dot, Bio-Rad). Proteins were detected using primary antibodies directed against polyglutamine (1C2) and mutant huntingtin (electron microscopy-48). Equal amounts of proteins (20 µg) were also separated on SDS/10% PAGE and transferred to nitrocellulose membranes in which the actin protein was detected as a loading control.
As described (28
, 29)
, microdialysis probes of 2 mm membrane length, 0.24 mm outer diameter (Cuprophane, 6 kDa cutoff; model CMA-11; CMA Microdialysis, Solna, Sweden) were implanted into the right striatum of anesthetized mice (stereotactic coordinates: antero-posterior, 0.0 mm; dorso-ventral, –4.4 mm; lateral, 2.5 mm relative to bregma; ref. 30
). Microdialysis experiments were performed at least 24 h after implantation of the probe. A quantitative "low perfusion rate" microdialysis approach (70 m/min) in freely moving mice was used to directly compare basal extracellular levels of DA (28)
. DA and its metabolite levels in dialysates or in striatal homogenates were measured by HPLC with electrochemical detection as described previously (28
, 29)
.
Statistical analysis
Student’s two-tailed t test was used for statistical comparisons between two groups and one-factor ANOVA followed by post hoc Fisher PLSD test, when appropriate, was applied for a comparison between more than two groups using GraphPad prism version 2.0 (Abacus concepts, Berkeley, CA, USA).
RESULTS
Behavioral manifestations in hyperdopaminergic mice expressing mutant huntingtin
To investigate the behavioral consequences of pairing the dopaminergic hyperactivity genotype (DAT–/–) with the mutant huntingtin protein genotype (HdhQ92/Q92), we followed the locomotor activity of DAT–/–,HdhQ92/Q92 mice in comparison to littermate control mice that included DAT+/+, Hdh+/+, DAT+/+,HdhQ92/Q92, and DAT–/–, Hdh+/+ mice. Two parameters of locomotor activity were analyzed over a period of 1 year: horizontal activity defined as the total distance covered per hour in the open field and the stereotypy time, the total time that mice exhibited a stereotypic behavior (repetitive beam breaks of a given beam).
Effects of the HdhQ92/Q92 genotype were subtle on the DAT+/+ genotype (Fig. 1
A, B). A between-genotype analysis at each age revealed that the distance traveled and stereotypy time were not statistically different between DAT+/+,HdhQ92/Q92 and DAT+/+, Hdh+/+ mice (Fig. 1A, B
). Nonetheless, the distance traveled was significantly lower in the DAT+/+,HdhQ92/Q92 mice at 12 months of age compared with those at 2 months of age (within-genotype analysis, P<0.05, Fisher PLSD; Fig. 1A
). The distance traveled and stereotypy time of DAT+/+, Hdh+/+ mice or the stereotypy time of DAT+/+,HdhQ92/Q92 mice were unchanged with aging as revealed by a within-genotype analysis (Fisher PLSD; Fig. 1A, B
).
|
On the other hand, remarkable changes were observed in the locomotor activity of DAT–/–,HdhQ92/Q92 mice. In these mice, the distance traveled was significantly lower at 6, 8, and 12 months of age than that of age-matched DAT–/–, Hdh+/+ mice (Fig. 1A
). Moreover, a within-genotype analysis revealed a significant decrease with aging in the distance traveled of DAT–/–,HdhQ92/Q92 mice at 6, 8, and 12 months of age compared with mice at 2 months of age (P<0.01, P<0.01, and P<0.005, respectively, Fisher PLSD; Fig. 1A
). No effect of aging was observed in DAT–/–, Hdh+/+ mice (Fisher PLSD; Fig. 1A
). Intriguingly, the stereotypy time did not follow this pattern of changes. In fact, it was significantly higher in DAT–/–,HdhQ92/Q92 mice at 6 and 8 months of age than in age-matched DAT–/–, Hdh+/+ mice (Fig. 1B
). The within-genotype analysis revealed that only 6-month-old DAT–/–,HdhQ92/Q92 mice showed higher stereotypy than 2-month-old mice (P<0.01, Fisher PLSD; Fig. 1B
). Furthermore, stereotypy time at 12 months of age returned to the values present at 2 months of age. No statistically significant effect of age was observed in DAT–/–, Hdh+/+ mice for either the distance traveled or stereotypy time (Fisher PLSD; Fig. 1A, B
).
Dopamine levels were unaltered in DAT–/–,HdhQ92/Q92 mice
To test whether the changes seen in locomotor activities in DAT–/–,HdhQ92/Q92 mice might be related to alterations in DA neurotransmission, the striatal levels of DA and its metabolites 3,4-dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA) were measured in mice of all genotypes at 8 months of age. No difference in DA, DOPAC, and HVA levels was observed between DAT+/+, Hdh+/+ mice and DAT+/+,HdhQ92/Q92 mice or between DAT–/–, Hdh+/+ and DAT–/–,HdhQ92/Q92 mice (Fig 2
A). However, there was a clear impact of DAT deletion on DA and HVA levels in both Hdh+/+ and HdhQ92/Q92 mice. In addition, basal extracellular levels of DA were verified by a quantitative in vivo microdialysis "low perfusion rate" approach (28)
in the striatum of freely moving mice (Fig. 2B
). In agreement with previous data, DAT deletion resulted in a significant increase in extracellular DA in both Hdh+/+ and HdhQ92/Q92 mice (25)
. However, no statistical difference was observed between DAT+/+, Hdh+/+ mice and DAT+/+,HdhQ92/Q92 mice or between DAT–/–,HdhQ92/Q92 and DAT–/–, Hdh+/+ mice (Fig 2B
). These data suggest that the changes observed in the motor behavior of DAT–/–,HdhQ92/Q92 mice cannot be explained by alterations in striatal DA levels.
|
Mutant huntingtin aggregated prematurely in the striatum of DAT–/–,HdhQ92/Q92 mice
The physical nature of the mutant huntingtin proteins was verified by immunofluorescence analysis in the striatum of DAT–/–,HdhQ92/Q92 mice, using the anti-aggregated huntingtin EM-48 Ab (31)
. As a control, EM-48 labeling was verified in the striatum of DAT+/+, Hdh+/+ mice and no positive signal was detected (Fig. 3
A). Similarly, no positive signal was observed in the striatum of DAT–/–, Hdh+/+ mice (Fig 3B
). In striatal neurons of DAT+/+,HdhQ92/Q92 mice, a positive nuclear staining known to be associated with the emergence of microaggregates of mutant huntingtin (26
, 32
, 33)
was observed at 8 months of age (Fig 3C
, Table 1
). This nuclear staining was observed more frequently at 12 months of age (Fig 3C
, Table 1
). Remarkably, the nuclear staining of mutant huntingtin was detected as early as 4 months of age (Fig. 3D
, Table 1
) in the striatal neurons of DAT–/–,HdhQ92/Q92 mice. By 8 months of age, in addition to nuclear microaggregates, nuclear aggregates of huntingtin were observed as well as neuropil aggregates (Fig. 3D
, Table 1
, and Table 2
). Nuclear aggregates were large, single, round, and similar to those first described for the R6/2 mouse model and in the brain of HD patients (34
, 35)
. The neuropil aggregates were more noticeable at 12 months of age, and nuclear staining as well as the number of nuclear aggregates was increased in the striatal neurons of DAT–/–, HdhQ92/Q92 mice (Fig. 3D
, Tables 1
, 2
).
|
|
|
In addition to the immunofluorescence study, we performed a filter trapping assay on mouse striatal extracts followed by EM-48 immunoblotting. This assay allows the detection of nuclear aggregates of mutant huntingtin (SDS-resistant) selectively. Furthermore, to verify the gene-dose effect of deleting the DA transporter on the formation of nuclear aggregates, we investigated the consequences of the HdhQ92/Q92 mutation on DAT+/– background (DAT+/–,HdhQ92/Q92 mice), which normally have a 2-fold elevation in extracellular DA concentrations in the striatum (25)
. Striatal aggregates of huntingtin were detected in the DAT+/+,HdhQ92/Q92 at 12 months of age, in the DAT+/–,HdhQ92/Q92 at 8 and 12 months of age, and in the DAT–/–,HdhQ92/Q92 at 4, 8, and 12 months of age (Fig 4
A). As a loading control, equal amounts of striatal extracts were separated by Western blot and immunoblotted for actin protein (Fig 4B
). A densitometric analysis of the EM-48 immunoblot revealed a statistically significant effect of age and genotype on the aggregation of mutant huntingtin (Fig 4C
). Another Ab, the antipolyglutamine tract (1C2), which specifically recognizes large polyglutamine tracts (36)
, especially those that are expanded, confirmed that the aggregated form of the mutant huntingtin was trapped in the membrane (Fig 4D
). Note that as negative controls, filter trapping assays were also performed using the EM-48 and the 1C2 antibodies in striatal extracts of DAT+/+, Hdh+/+ and DAT–/–, and Hdh+/+ mice; no signal was detected. The presence of huntingtin aggregates was also verified by this technique in the cerebral cortex (Fig 1E
) and hippocampus (data not shown), and no signal was noticed at any age or genotype.
|
Distribution of aggregates of mutant huntingtin in the DAT–/–,HdhQ92/Q92 mice
To explore the brain regional specificity of aggregate formation, an immunofluorescence analysis was performed in the whole brain of mice. As observed in the striatum of the DAT–/–,HdhQ92/Q92 mice, nuclear staining as well as nuclear aggregates appeared much earlier in the olfactory tract, cerebral cortex, piriform cortex, striatum, hippocampus, dentate gyrus, and cerebellum in these mice compared with DAT+/+,HdhQ92/Q92 mice (Table 1
, Fig. 5
A–D). It is noteworthy that all of these brain regions receive dense dopaminergic input (37
, 38)
. No positive signal was observed in the brain of DAT+/+, Hdh+/+ and DAT–/–, Hdh+/+ mice.
|
One fascinating result was the emergence of neuropil aggregates in the striatum, olfactory tract, external segment of globus pallidus and subtantia nigra pars reticulata of DAT–/–,HdhQ92/Q92 mice (Fig. 5E, F
; Table 2
). These aggregates were not observed in the brain of DAT+/+,HdhQ92/Q92 (Fig. 5E, F
; Table 2
) or in DAT+/+,Hdh+/+ and DAT–/–,Hdh+/+ mice. In the striatum and olfactory tract, neuropil aggregates became visible together with the emergence of nuclear staining and before the appearance of nuclear aggregates (Tables 1
and 2)
. It is noteworthy that whereas few neuropil aggregates were seen in the subtantia nigra at 4 months of age in the DAT–/–,HdhQ92/Q92 mice, they were already observed in moderate number in the external segment of globus pallidus (Table 2)
.
To verify whether neuropil and nuclear aggregates were located in DA neurons, we performed double immunofluorescence in the brain of DAT–/–,HdhQ92/Q92 mice using EM-48 and a tyrosine hydroxylase Ab as a marker of dopaminergic neurons (Fig 5G
). The neuropil aggregates observed in the subtantia nigra and the globus pallidus were not found in the dopaminergic neurons (Fig. 5G
). Moreover, no nuclear aggregates were detected in the DA neurons of the subtantia nigra pars compacta (Fig 5G
, left panel). The presence of neuronal death was investigated on brain sections of 12-month-old mice by immunofluorescence using the activated-caspase 3 Ab and the TUNEL assay. No positive signal was detected using both techniques in all brain regions investigated (striatum, olfactory tract, cerebellum, hippocampus, and cerebral cortex) in the DAT+/+,Hdh+/+, DAT+/+,HdhQ92/Q92 and DAT–/–,HdhQ92/Q92 mice (data not shown).
DISCUSSION
Although the predisposition to develop HD is linked to genetic factors, mechanisms underlying the emergence of specific neuropathological processes and development of symptoms in carriers of mutant huntingtin are still unidentified. Here we show, using a mouse model of HD with enhanced dopaminergic transmission, that the neuronal context plays a key role in influencing the effects of mutant huntingtin. We describe that the striatum’s dense dopaminergic innervation may contribute significantly to the vulnerability of this region in HD through its ability to trigger the formation of nuclear and neuropil aggregates of mutant huntingtin.
We have observed that the locomotor activity of the DAT+/+,HdhQ92/Q92 mice is not significantly altered, although biphasic subtle motor alterations can be discerned with an increase in locomotion and stereotypy at 6 months of age, followed by a decrease at 12 months. However, in DAT–/–,HdhQ92/Q92 mice there is a progressive decrease in activity and a marked increase in stereotypic behavior that becomes apparent at 6 months of age and returns to control values at 12 months. This suggests that an enhanced DA transmission in the brain of HD mouse model exacerbates the emergence of behavioral symptoms without affecting the biphasic pattern of locomotor alterations with aging. An increase in repetitive movements at an early age followed by a decrease is documented in a knock-in mouse model of HD with 94 CAG repeats, when locomotor activity is recorded during the dark phase of the diurnal cycle (33)
. In general, along with our data, no major signs of motor dysfunction have been reported in knock-in mouse models of HD with 92 CAG repeats or with a similar number of CAG repeats in relation to aging (26
, 32
, 39)
. On the other hand, robust biphasic motor alterations with aging are reported in knock-in mouse models with a high number of CAG repeats (>94 CAG repeats) in YAC and transgenic mouse models as well as in a rat model of HD (34
, 40
41
42
43
44)
. It thus appears that the motor behavior of the double mutant mice is to some extent comparable with one of the other mouse models of HD with more severe phenotypes. These observations indicate that enhanced DA transmission in the brain of HD mice does not influence the pattern of appearance of motor alterations, but rather the severity of its manifestation.
Formation of nuclear aggregates appears much earlier and to a greater extent in striatal neurons of HD mice with 5-fold enhanced DA levels in the DAT–/– background. Even a relatively moderate 2-fold elevation in extracellular DA in DAT+/– mice significantly affects the formation of huntingtin aggregates. Normally, nuclear aggregates can be observed at 15 to 18 months of age in HdhQ92/Q92 mice alone or in HdhQ94/Q94 knock-in mice (26
, 33)
. However, inclusion formation is believed to be the end stage of a process beginning with huntingtin translocation to the nucleus and continuing with the formation of microaggregates (33
, 34)
. The brain areas showing early nuclear staining or microaggregates in the DAT–/–,HdhQ92/Q92 mice (i.e., the striatum and olfactory tract) all receive a dense dopaminergic innervation (37
, 38)
. As the mice aged, nuclear aggregates became larger and were seen in other areas of the brain such as the piriform cortex, hippocampus, cerebellum, and some layers of the cerebral cortex. These regions also receive significant dopaminergic input (38)
. The observation that brain regions with early formation of nuclear aggregates of the mutant huntingtin proteins all receive dense dopaminergic inputs has also been noted in knock-in mouse models of HD with 140 and 150 CAG repeats (42
, 45)
. Altogether, these observations demonstrate that in addition to the dose effect of extracellular DA levels on the formation of aggregates in the HdhQ92/Q92 knock-in mice, the kinetics of aggregation of mutant huntingtin is to some extent under the influence of dopaminergic transmission.
The nuclear aggregates of mutant huntingtin in the DAT–/–,HdhQ92/Q92 mice are observed without the emergence of signs of degeneration. It is well established that although there is formation of nuclear aggregates in several HD mouse models, none reproduce the massive selective neuronal death that can be seen in the striatum of advanced HD patients (46)
. The reasons for this are unknown, but may include the relatively short life span of the mouse. It is noteworthy that in these mouse models, nuclear aggregates are observed at a time point when both behavioral and neuropathological changes, including neuronal loss, are present, suggesting that aggregates are not involved in the initiation of neuronal loss (33
, 43
, 46
, 47)
. In fact, although aggregates are the common hallmark of numerous polyglutamine disorders, studies in HD patients showed little overlap between those cells exhibiting nuclear aggregates and cells that undergo neurodegeneration, further supporting the lack of correlation between aggregates and neuronal loss (31
, 48)
. Some studies even propose the idea that aggregates might be a protective mechanism, as inclusion body formation seems to reduce the risk of neuronal death (47
, 49)
. This is an interesting issue; we reported earlier that a subpopulation of DAT–/– mice exhibited evidence of neuronal degeneration in the striatum (14)
. Whether aggregates of mutant huntingtin play a role in the absence of neuronal degeneration in the DAT–/–,HdhQ92/Q92 clearly requires further investigation. On the other hand, ROS, induced by an excess of DA, is documented to make striatal neurons expressing the mutant huntingtin gene in primary culture more vulnerable to neuronal death (22
, 23)
. However, despite the increased DA levels in DAT–/– mice, indirect observations suggest that an increase in ROS is unlikely in these mice, which might explain why neuronal death is not observed in DAT–/–,HdhQ92/Q92 (14
, 25
, 50
51
52)
.
There is considerable evidence that aberrant DA metabolism and transmission might underlie the clinical manifestations of HD. This idea was first proposed when asymptomatic relatives of individuals with HD developed dyskinesias after being given L-DOPA (13
, 15
, 16)
. Later it was reported that DA receptor antagonists and release inhibitors were effective in treating chorea, a major symptom in HD (17
18
19
20
21)
. Moreover, psychiatric manifestations of cortical and striatal dysfunction can be observed in carriers of the HD mutation prior to the onset of overt clinical movement disorder, and these manifestations can be reversed by DA receptor antagonists (53)
. These clinical observations suggest that lowering the DA transmission could be beneficial in HD and indicate that "hyperactive" DA transmission might contribute to development of the disease. In contrast, some biochemical studies in HD brain tissues and mouse models of HD suggest a "hypoactive" DA transmission. For instance, several PET scanning studies show a reduction in levels of striatal D1 and D2 receptors in asymptomatic patients carrying the mutation of the gene for huntingtin before the onset of overt clinical movement disorder (7
, 54
55
56
57
58)
. Whether these lower numbers of D1 and D2 receptors are paralleled by an alteration in presynaptic components at an early stage or simply reflect degeneration of striatal neuronal population is unclear. Nonetheless, despite the decrease in D1 and D2 receptors, D1 agonists elicited comparable induction of the immediate early genes Zif268 and N10 and enhanced c-Fos and Jun B in a mouse model of HD, suggesting supersensitivity of this receptor (59)
. There is also an increase in the number of protein G
and D5 DA receptors, enhanced levels of Ca2+, and a neuronal hypersensitivity to application of NMDA in many presymptomatic HD mouse models (39
, 60
61
62
63)
. Altogether, these data suggest that changes in DA transmission emerge early during disease progression and might include at least for a period of time an overactivation of the striatal neuronal elements.
In DAT–/– mice, the persistently enhanced DA transmission in the striatum is associated with an early decrease in D1 and D2 receptors, followed by an increase in the level of markers of DA receptors activation (14
, 24
, 64)
. The sustained overactivation of DA receptors in these mice has been documented to be associated with subtle degeneration of striatal GABAergic neurons and motor alterations, in addition to an enhanced sensitivity to 3-nitropropionic acid, a mitochondrial toxin known to induce striatal degeneration (14
, 52
, 65
, 66)
. Whether the sustained DA receptor activation in DAT–/– mice are responsible for the formation of aggregates of mutant huntingtin in the striatal projection system of DAT–/–,Hdh Q92/Q92 mice requires further investigation. However, in accord with our contention, a recent study using striatal primary cultures transiently transfected with CAG repeat-expanded fragments of huntingtin shows that DA-induced aggregates formation can be reversed by a selective D2 receptor antagonist and reproduced by direct D2 DA receptor agonist (23)
. Other potential mechanisms for the early aggregate phenotype in the DAT–/–,HdhQ92/Q92 mice may include alterations in cdk5 and Akt activity, which have been described in an HD mouse model and in brains of patients. Cdk5 and Akt activities are known to be altered in the striatum of DAT–/– mice (14
, 67)
, and a down-regulation in these proteins’ activity has been shown to enhance the aggregation of mutant huntingtin (68
69
70)
.
In addition to nuclear aggregates of mutant huntingtin, neuropil aggregates were observed in the striatum, the external segment of globus pallidus, and substantia nigra pars reticulata of DAT–/–,HdhQ92/Q92 mice. That both neuropil and nuclear aggregates were observed simultaneously in the striatum of these mice is intriguing. Although further studies are needed to determine whether striatal neuropil aggregates are found in dendrites of striatal neurons or in the terminals of the corticostriatal axons, these neuropil aggregates may potentially affect the corticostriatal circuitry. For instance, a deleterious effect of these aggregates on synaptic transmission and axonal transport has been documented in mouse models of HD (32
, 60
, 61
, 71)
. On the other hand, the fact that the globus pallidus and substantia nigra pars reticulata contain neuropil but not nuclear aggregates suggests that neuropil aggregates are found in the terminals of striato-pallidal (GPe) and -nigral axons. A higher density of neuropil aggregates are found in the external segment of globus pallidus (GPe), suggesting that neurons of the striatopallidal pathway are affected before the striatonigral. Striatal deterioration is known to occur in the striatal projection system that terminates in the external segment of the globus pallidus (72)
, followed by losses in the striatonigral pathway (73)
, whereas striatal projections to the internal segment of the globus pallidus (GPi) remain relatively spared in HD (74
, 75)
. In fact, neostriatal wiring models predict that the direct striato-nigral/-GPi pathway activation generates hyperkinesis (76
, 77)
, a major behavioral symptom associated with HD. In DAT–/–,HdhQ92/Q92 mice, the formation of robust nuclear microaggregates and neuropil aggregates parallels the increase in stereotypy at 6 months of age, whereas formation of nuclear aggregates is noticed when stereotypy and locomotion are significantly decreased. Whether aggregates of huntingtin are directly responsible for the behavioral changes in DAT–/–,HdhQ92/Q92 mice requires further study. However, these data raise the interesting possibility that enhanced DA transmission could accelerate the formation of striatal neuropil aggregates, which in turn may exert deleterious effects on synaptic transmission and exacerbate locomotor alterations in HD mouse models.
GABAergic/enkephalinergic striatal neurons that project to the external segment of the globus pallidus and the subtantia nigra are known to express high levels of D2 receptors (76)
. Moreover, a higher density of neuropil aggregates is found in the external segment of globus pallidus (GPe) and the substantia nigra, two brain regions where D2 receptors are concentrated. Since neuropil aggregates are found in the external segment of globus pallidus (GPe) and in the subtantia nigra of DAT–/–,HdhQ92/Q92 mice at an early age, one could speculate that the D2 receptor might play a role in the early formation of aggregates. In accord with this, recent in vitro findings demonstrate that a D2 receptor agonist increases aggregate formation of mutated huntingtin in all cellular compartments of striatal primary cultures, including neurites, soma, and nuclei (23)
. Altogether, these findings argue for a role of D2 receptor stimulation in aggregate formation.
DA modulation of striatal activity is a primary physiological event in basal ganglia nuclei. Therefore, changes in postsynaptic phenotype expression, cytoarchitectural properties of dendrites and their spines, and losses in the DA receptors associated with these neuronal elements will have serious consequences for synaptic communication in the striatum and for the output of the basal ganglia as a whole (32
, 60
, 61
, 71
, 76
, 77)
. Here we directly demonstrate that enhanced DA transmission potentiates locomotor alterations and increases the occurrence of nuclear and neuropil aggregates in the striatum and other brain regions receiving the dopaminergic input in a mouse model of HD. Until now there was no clear indication that decreasing efficacy of DA transmission (by using DA antagonists or depleters) could affect the rate of illness progression in HD (78
79
80)
. Our study unravels a role for DA in the striatal deterioration induced by mutant huntingtin in vivo and suggests a molecular basis for clinical strategy of reducing activity of the DA system in HD.
ACKNOWLEDGMENTS
This work was supported in part by National Institutes of Health grant NS19576 (to M.G.C.) and a Hereditary Disease Foundation research grant (to M.C.). During part of this work, M.C. was the holder of a Huntington’s Disease Society of America fellowship. We thank S. Suter for excellent technical assistance. EM-48 Ab was a gift from Dr. X. J. Li (School of Medicine, Emory University, Atlanta, GA, USA).
Received for publication May 22, 2006. Accepted for publication July 17, 2006.
REFERENCES
P < 0.05, 