Thermoregulatory and metabolic defects in Huntington’s
disease transgenic mice implicate PGC-1a in Huntington’s
Patrick Weydt,1,11Victor V. Pineda,1,11Anne E. Torrence,1,2Randell T. Libby,1Terrence F. Satterfield,1
Eduardo R. Lazarowski,8Merle L. Gilbert,3Gregory J. Morton,3Theodor K. Bammler,6Andrew D. Strand,9
Libin Cui,10Richard P. Beyer,6Courtney N. Easley,1Annette C. Smith,1Dimitri Krainc,10Serge Luquet,4,12
Ian R. Sweet,3Michael W. Schwartz,3and Albert R. La Spada1,3,5,7,*
1Department of Laboratory Medicine
2Department of Comparative Medicine
3Department of Medicine
4Department of Biochemistry
5Department of Neurology
6The Center for Ecogenetics and Environmental Health
7The Center for Neurogenetics and Neurotherapeutics
University of Washington, Seattle, Washington 98195
8University of North Carolina, Chapel Hill, North Carolina 27599
9Fred Hutchinson Cancer Research Center, Seattle, Washington 98109
10Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02129
11These authors contributed equally to this work.
12Present address: CNRS UMR, Paris, France.
Huntington’s disease (HD)is afatal, dominantly inherited disorder caused by polyglutamine repeatexpansion in the hunting-
tin (htt) gene. Here, we observe that HD mice develop hypothermia associated with impaired activation of brown adipose
tissue (BAT). Although sympathetic stimulation of PPARg coactivator 1a (PGC-1a) was intact in BAT of HD mice, uncoupling
protein1(UCP-1)induction wasblunted.Inculturedcells, expressionofmutant httsuppressed UCP-1promoteractivity;this
wasreversedbyPGC-1aexpression.HDmiceshowedreducedfoodintakeandincreased energyexpenditure, withdysfunc-
tional BAT mitochondria. PGC-1a is a known regulator of mitochondrial function; here, we document reduced expression of
PGC-1a target genes in HD patient and mouse striatum. Mitochondria of HD mouse brain show reduced oxygen consump-
tion rates. Finally, HD striatal neurons expressing exogenous PGC-1a were resistant to 3-nitropropionic acid treatment.
Altered PGC-1a function may thus link transcription dysregulation and mitochondrial dysfunction in HD.
Huntington’s disease (HD) is an autosomal dominant neuro-
degenerative disorder characterized by motor and cognitive
impairment, accompanied by a variable degree of personality
change and psychiatric illness (Nance, 1997). The motor abnor-
mality stems from dysfunction of the involuntary movement
control region of the midbrain known as the striatum and is
manifested as a hallmark feature of uncontrollable dance-like
movements (‘‘chorea’’). HD is relentlessly progressive, as pa-
tients succumb to the disease 10–25 years after disease onset.
glutamine tract in the huntingtin (htt) protein was determined to
be responsible for HD (Huntington’s Disease Collaborative
Research Group, 1993). HD is thus one of nine inherited neuro-
degenerative disorders all caused by CAG trinucleotide repeats
that expand to produce disease by encoding elongated poly-
glutamine (polyQ) tracts in their respective protein products
(Zoghbi and Orr, 2000). Although the mutant htt protein is widely
expressed, only certain populations of neurons degenerate and
only a subset of nonneuronal cell types are affected.
Neurons in the brain have enormous demands for continued
production of high-energy phosphate-bonded compounds
ofamitochondrial toxin, 3-nitropropionicacid(3-NP),resultedin
a selective loss of medium spiny neurons in the striatum—the
cell type whose degeneration has been linked to the HD pheno-
type (Beal et al., 1993). This finding suggested that mitochon-
drial dysfunction may underlie HD pathogenesis and account
for the cell-type specificity in this neurodegenerative disorder.
Follow-up studies performed upon HD patient material have
documented significant reductions in the enzymatic activities
of complexes II, III, and IV of the mitochondrial oxidative phos-
phorylation pathway in caudate and putamen (Browne et al.,
1997; Gu et al., 1996). PET scan analysis of HD patients also
strongly supports the hypothesis of defective energy metabo-
parent in the cortex and striatum (Stoessl et al., 1986). Magnetic
resonance spectroscopy corroborates such findings, revealing
elevated lactate levels in striata of HD patients (Harms et al.,
1997). In addition to chorea, cognitive decline, and personality
change, HD patients display clinical signs of disturbed energy
CELL METABOLISM 4, 349–362, NOVEMBER 2006 ª2006 ELSEVIER INC. DOI 10.1016/j.cmet.2006.10.004 349
A R T I C L E
metabolism, including weight loss (Lodi et al., 2000; Pratley
et al., 2000; Robbins et al., 2006). Despite the wealth of data im-
plicating mitochondrial dysfunction as a central feature of HD
pathogenesis, the molecular basis of the mitochondrial abnor-
mality has remained elusive. Indeed, since the discovery of the
HD gene, considerable evidence suggests that nuclear localiza-
tion of mutant htt resulting in abnormalities of gene expression
(‘‘transcriptional dysregulation’’) underlies HD pathogenesis
(Sugars and Rubinsztein, 2003). These findings collectively
implicate both transcription dysregulation and mitochondrial
dysfunction in HD pathogenesis.
ling et al., 1999) for metabolic abnormalities, we discovered that
HD transgenic mice develop profound hypothermia. In mam-
tone in the periphery ensues. In rodents, brown adipose tissue
(BAT) is the principal tissue that mediates the body’s response
to cold temperature (‘‘adaptive thermogenesis’’) (Cannon and
Nedergaard, 2004). The transcription factor coactivator peroxi-
some proliferator-activated receptor (PPAR)-g coactivator 1a
(PGC-1a) is expressed in BAT and is a key mediator of adaptive
thermogenesis (Puigserver et al., 1998). PGC-1a expression in
BAT is dramatically upregulated in response to b-adrenergic
stimulation. The key effector of adaptive thermogenesis in
BAT is uncoupling protein 1 (UCP-1), whose expression is
restricted to mitochondria of BAT (Lin et al., 2005). UCP-1
dissipates the proton gradient at the inner mitochondrial mem-
brane to prevent oxidative phosphorylation, instead yielding
a futile cycle that generates heat. PGC-1a coactivates expres-
sion of UCP-1 in BAT through its interaction with PPARg and
the retinoic acid receptor (RXRa), as the former has a binding
site in the UCP-1 promoter (Lin et al., 2005).
After characterizing the thermoregulatory defects in HDtrans-
genic mice, we assessed the function of PGC-1a by evaluating
HD BAT tissue. Based upon both in vitro and in vivo studies, we
determined that profound hypothermia in HD mice results from
PGC-1a transcription interference. As PGC-1a isa key regulator
of energy metabolism, we performed indirect calorimetry upon
HD mice and observed abnormalities in food intake and meta-
bolic activity during fasting. Studies of mitochondrial function
confirmed abnormalities in HD BAT. These findings led us to
evaluatethe striatum ofHD transgenic mice andhumanpatients
for evidence of impaired PGC-1a coactivator function. Micro-
array and RT-PCR analysis of murine and human striatal RNAs
revealed significant reductions in PGC-1a targets, while studies
of mitochondrial function revealed abnormalities in HD trans-
genic brain and knockin striatal neurons. Our results suggest
that mitochondrial dysfunction in HD may stem from polyQ-htt
interference with PGC-1a in the nucleus, linking transcription
dysregulation with mitochondrial pathology in HD. Furthermore,
our results identify a plausible explanation for the exquisite sus-
ceptibility of a highly metabolically active subset of neurons in
the striatum to degeneration and death in HD, and support a
paradigmatic view of HD as a metabolic disease.
HD transgenic mice display profound thermoregulatory
defects and abnormal BAT
As neurological deficits in HD are gradually progressive, consid-
erable emphasis has been placed upon the identification of ob-
jective and reproducible measures of disease onset and pro-
gression (i.e., ‘‘biomarkers’’) to improve the predictive value of
therapeutic trials. We therefore chose to evaluate a number of
metabolic parameters, including body temperature, in a com-
monly used model of HD, the N171-82Q transgenic mouse
taneously implanted a temperature-sensitive transponder, as
this system permits remote recording of body temperature
with minimal handling of the subject, obviating stress-induced
sympathetic stimulation. Using this system, which correlates
with rectal probe temperature (Figure S1 in the Supplemental
Data available with this article online), we monitored body tem-
perature and found that all HD mice developed progressive hy-
pothermia, beginning at 17 weeks of age (Figure 1A). With pro-
gression of motor symptoms and weight loss (Figure S1), some
HD N171-82Q mice displayed profoundly deranged thermoreg-
ulation (FigureS1),with bodytemperaturesdroppingto%27?C.
HD mice with temperatures below 30?C were not within hours or
minutes of death, as they remained mobile and alive for at least
another 48 hr, often considerably longer. Since advanced HD is
characterized by muscle wasting (Beal and Ferrante, 2004), and
muscle mass is a key determinant of thermoregulation (Lowell
and Spiegelman, 2000), we wondered if profound hypothermia
in HD mice simply reflected muscle wasting and was therefore
nonspecific. To test this hypothesis, we charted body tempera-
ture regulation in SOD1-G93A amyotrophic lateral sclerosis
(ALS) mice, since they also develop a severe neurodegenerative
phenotype characterized by weight loss, muscle wasting, and
reduced lifespan (Gurney et al., 1994). In late stage ALS mice,
(Figure 1B). In light of the striking hypothermia phenotype, we
taining body temperature in the face of a 4?C cold challenge—a
process known as ‘‘adaptive thermogenesis’’ (Lowell and Spie-
gelman, 2000). After obtaining HD N171-18Q transgenic mice to
control for htt protein overexpression, we established three co-
horts of mice (HD 82Q; HD 18Q; and WT) for adaptive thermo-
genesis testing at 10 weeks and 20 weeks of age. Individual
mice were placed at 4?C for up to 9 hr, and body temperatures
were recorded at 1 hr intervals. Although control mice were able
to maintain normal thermoregulation, HD transgenic mice dis-
played significant reductions in body temperature during cold
(Figures 1C and 1D).
In rodents, BAT is the principal tissue that mediates adaptive
thermogenesis, and is distinguished from white fat by its high
degree of vascularization and mitochondrial density (Wang
abnormalities, including reductions in cell density and nuclei
number (Figures 1E and 1F). Indeed, the BAT of HD mice ap-
peared white fat-like in histology sections, suggesting that the
thermogenesis defect likely involves abnormalities in BAT com-
position and function. Importantly, RT-PCR analysis indicated
that the mutant htt transgene is expressed in BAT (Figure S2).
The PGC-1a - UCP-1 circuit is disrupted in the BAT of HD
In mammals, after cold is sensed in the hypothalamus, an in-
crease in sympathetic tone in the periphery ensues. In rodents,
BAT is the target of this increased sympathetic output. PGC-1a
is a transcription coactivator whose expression in BAT is
A R T I C L E
CELL METABOLISM : NOVEMBER 2006
dramatically upregulated in response to b-adrenergic stimula-
tion (Puigserver et al., 1998). The principal effector of adaptive
thermogenesis in BAT is uncoupling protein 1 (UCP-1), whose
expression isrestrictedtomitochondria ofbrownadiposetissue
(Puigserver and Spiegelman, 2003). To determine if PGC-1a
transactivation of UCP-1 is normal in HD mice, we dissected
infrascapular BAT after cold challenge, isolated total RNA,
and measured PGC-1a and UCP-1 transcripts. We observed
marked upregulation of PGC-1a in BAT of cold-challenged con-
trol and HD mice (Figure 2A). This result indicates that hypotha-
lamic sensing of temperature change, elevation of sympathetic
tone, and b-adrenergic stimulation of PGC-1a in BAT are intact
in HD mice. Detection of c-fos upregulation in the ventromedial
hypothalamic nucleus of cold-challenged N171-82Q HD mice
independently confirmed hypothalamus activation in response
to cold inHDmice (Figure S3).Despite preservationof hypothal-
amus-mediated b-adrenergic stimulation of PGC-1a in BAT,
cold-challenged HD transgenic mice failed to upregulate UCP-
1 mRNA (Figure 2B), and cold-challenged levels of UCP-1 pro-
tein were decreased in BAT from HD mice (Figure S4A). These
results suggest that interference with PGC-1a coactivation of
UCP-1 in BAT accounts for the adaptive thermogenesis defect
To further investigate this hypothesis, 3T3-L1 preadipocyte
cells were transfected with UCP1 promoter-reporter constructs
along with mutant or normal htt in the presence or absence of
PGC-1a. While baseline transactivation levels were similar,
polyQ-expanded htt repressed stimulation of UCP-1 promoter
activity; importantly, mutant htt repression of UCP-1 transcrip-
tion could be overcomeby coexpression of PGC-1a (Figure 2C).
As preadipocyte cells are not committed to BAT differentiation,
we established primary brown adipocyte cultures from HD
N171-82Q and control mice. Upon norepinephrine stimulation,
trols displayed comparable PGC-1a induction (data not shown);
however, UCP-1 induction was significantly blunted in adipo-
cytes expressing polyQ-expanded htt (Figure 2D). Failure of
As PGC-1a coordinates mitochondrial biogenesis and regu-
lates mitochondrial function, we chose to evaluate mitochon-
drial function in BAT from HD mice. We began by culturing
brown adipocytes from N171-82Q HD mice and age-matched
controls in the presence of Mitotracker Red, and then flow-
sorting BAT cells that had successfully taken up the dye. Com-
parison of fluorescent signal intensities for sets of cultured cells
revealed a significant reduction in Mitotracker Red uptake in
Figure 1. HD mice display a temperature regulation
defect and abnormal BAT
A) We measured body temperature in HD N171-82Q
male mice (HD; red line) and age-matched nontrans-
genic controls (WT; black line) for their entire lifespan
(n = 12 per group). Beginning at 120 days of age, HD
mice display a significant reduction in body temper-
ature (p = 0.027), and this reduction progresses with
time. By 150 days of age, the difference in body tem-
perature is quite marked (HD: w33?C; WT: w37?C;
p = 0.003). Body temperature in HD N171-18Q
mice was comparable to WT mice (data not shown).
B) Profound hypothermia is not a general feature
of late-stage neurodegeneration. HD mice show pro-
found hypothermia in comparison to controls (n = 6;
w180 days of age; p < 0.001 by t test), while SOD1
G93A ALS mice do not (n = 5; w135 days of age;
p = 0.45 by t test).
C) HD transgenic mice display an adaptive thermo-
genesis defect. Sets of 20-week-old mice were
placed individually at 4?C, and body temperatures
were recorded. By 1 hr into the cold challenge, HD
perature (p < 0.01). This progressively worsens with
time during the cold challenge (p < 0.001 at 2 and
D) Younger HD mice also display an adaptive ther-
mogenesis defect. While HD 82Q mice at this early
stage of disease are able to maintain body tempera-
ture for 5 hr, a thermoregulatory defect becomes ap-
parent by 7 hr into the cold challenge (p < 0.05) and
progressively worsens (p < 0.01 at 8 and 9 hr). For
all temperature comparisons, data points are dis-
played with SEM, and we used one-way ANOVA
with Bonferroni’s multiple comparisons test.
E–F) Sets of 20-week-old HD mice and their non-
transgenic (WT) littermates (n = 5 per group) were
euthanized, and their infrascapularBAT samples iso-
lated, sectioned, and H&E stained. While WT BAT
appears normal (E), BAT from HD 82Q mice is mark-
edly abnormal (F), as evidenced by decreased cellu-
lar content (note fewer nuclei) and marked accumu-
lation of large lipid droplets. The scale bar for each
panel represents 20 mm.
Thermometabolic defects implicate PGC-1a in HD
CELL METABOLISM : NOVEMBER 2006351
brown adipocytes from HD mice (Figure 2E), consistent with
a decrease in the number of functional mitochondria in HD
BAT. To independently confirm the mitochondrial dysfunction
in HD BAT, we performed HPLC analysis of adenine nucleotides
from extracts ofBAT tissues dissected from HD transgenic mice
and age-matched control mice. The ratio of ATP/ADP was sig-
nificantly reduced in BAT extracts from the HD mice (Figure 2F),
suggestingthat mitochondrial function ismarkedlydecreased in
brown adipocytes from HD mice. To assess the transactivation
status of PGC-1a in BAT cells from HD mice, we measured the
expression of PGC-1a target genes involved in mitochondrial
energy production. Using real-time RT-PCR, we quantified the
and found significant reductions in such targets (Figure S4C).
HD mice display profound metabolic abnormalities
along with impaired PGC-1a signaling
As PGC-1a is a key regulator of energy metabolism and ho-
meostasis, we wondered if the thermoregulatory defects in the
N171-82Q HD mice would be paralleled by abnormalities of
food intake, body composition, or energy expenditure. We
therefore performed indirect calorimetry studies on 10-week-
old N171-82Q mice, nontransgenic littermates, and N171-18Q
controls, using a sequential schedule of feeding, fasting, and
re-feeding (Figure S5). Studies were performed with presymp-
tomatic N171-82Q mice to avoid the confounding effects of
weight loss and neurological impairment that begin at approxi-
mately 13 wks of age (Figure S1). As expected, mean body
weight was comparable between HD mice and controls (Fig-
ure 3A), but HD mice had a significantly higher % body fat and
a correspondingly lower % lean mass (Figures 3B and 3C).
This abnormal ratio of fat to lean mass in HD mice was main-
tained during fasting, but unlike controls, HD mice did not
recover lost fat mass during refeeding, making the earlier differ-
ences in body fat content no longer detectable after the refeed-
ing period. This impairment of HD mice to recover lost weight
was due largely to a pronounced food intake deficit. Thus,
the modest decrease in food intake displayed by HD mice at
Figure 2. HD transgenic mice display PGC-1a tran-
scription interference and mitochondrial abnormali-
ties in BAT
A)HDtransgenicmiceare capable ofPGC-1a induc-
tion in the face of cold challenge. We performed real-
time RT-PCR analysis of BAT RNAs obtained from
sets of 20-week-old HD 82Q (red) and WT mice
(black) housed at room temperature (solid bars) or
at the completion of a 3 hr cold challenge (striped
bars). Upregulation of PGC-1a in HD 82Q BAT is
comparable to that of WT BAT (p = 0.77 by t test).
B) HD transgenic mice fail to upregulate UCP-1 RNA
of the BAT samples shown in (A) was performed for
UCP-1. WT mice increase UCP-1 expression by
w3-fold, but HD 82Q mice display no induction of
UCP-1 expression (p < 0.001 by t test). HD 18Q con-
trol transgenic mice appropriately upregulate UCP-1
during cold challenge (data not shown).
C) Htt polyQ-length-dependent transcription inter-
ference of PGC-1a regulated gene expression.
3T3-L1 preadipocyte cells were cotransfected with
PPARg, RXRa, the UCP-1 promoter reporter con-
struct, pRL-CMV, exon1/2 htt 24Q or 104Q expres-
sion constructs, and PGC-1a expression construct
or empty vector. While no significant differences
are observed in unstimulated cotransfected cells,
htt 104Q markedly suppresses UCP-1 promoter re-
porter expression in cells treated with 20 mM troglita-
zone (TGZ) (p = 0.011 versus htt 24Q + TGZ; p =
0.557.versushtt104Q unstimulated). Cotransfection
of the PGC-1a expression vector rescues htt 104Q
repression(p < 0.005). Experiments were done indu-
plicate, corrected for transfection efficiency, and
of htt 24Q.
D) Primary brown adipocytes from HD mice are
resistant to norepinephrine stimulation. After estab-
lishing cultures of primary brown adipocytes (n = 3
per group), RNAs were isolated from unstimulated
adipocytes and norepinephrine-treated adipocytes
for measurement of UCP-1 expression levels by
real-time RT-PCR. Unstimulated primary brown adipocytes display negligible UCP-1 expression. Norepinephrine yields marked upregulation of UCP-1 RNA expression
in WT primary brown adipocytes, but such UCP-1 induction is significantly blunted in primary brown adipocytes expressing htt 82Q (p = 0.0002 by t test).
E) HD BAT contains fewer functional mitochondria. After establishing cultures of primary brown adipocytes from HD 82Q mice and control mice (n = 3 per group), BAT
cultures were treated with Mitotracker Red and then flow-sorted. HD 82Q BAT displayed significantly lower mean fluorescence (p = 0.02 by t test), indicative of fewer
F) ATP/ADP ratios are decreased in HD BAT. BAT samples from HD 82Q and control mice (n = 9–12 per group) were harvested, and after TCA extraction, HPLC measure-
ments of adenine nucleotides were performed. Mean ATP/ADP ratios for HD 82Q mice were significantly reduced (p = 0.03 by t test).
A R T I C L E
CELL METABOLISM : NOVEMBER 2006
baseline wasexaggerated after fasting, with re-feeding HDmice
consuming less than one-third of the amount of food eaten by
age-matched controls (Figure 3D).
Togauge energyexpenditure inthe10-week-oldHDmice, we
Nonetheless, oxygen consumption was slightly increased in HD
mice at baseline (Figure 3F). Upon fasting, control mice dis-
penditure, but both of these responses were attenuated in HD
mice (Figures 3E and 3F). Indeed, HD mice had significantly
higher oxygen consumption rates vis-a `-vis controls during fast-
ing (Figure 3F). Therefore, prior to onset of weight loss, HD mice
displayed a hypophagic, hypermetabolic phenotype, and they
mounted neither the decrease of metabolic rate northe increase
of food intake induced by fasting in normal mice. HD mice and
control mice exhibited similar blood glucose levels, both prior
to and at the end of an 18 hr fast (data not shown), and leptin
levels dropped significantly in both HD and control mice during
fasting (data not shown), as expected. As fasting normally re-
duces BAT UCP-1 levels (Scarpace et al., 1998), presumably
through reduced PGC-1a coactivation (Lowell and Spiegelman,
2000), we quantified PGC-1a and UCP-1 mRNA levels in the
BAT of young, presymptomatic, fasted HD mice and controls.
In control mice, we observed marked reductions in UCP-1
expression in BAT during fasting; however, HD mice failed to re-
duce BAT UCP-1 levels, and actually displayed a paradoxical
upregulation of PGC-1a levels in BAT during fasting (Figures
3G and 3H).
Evidence for PGC-1a transcription interference
in mouse striatum
1a transcription interference in HD transgenic mice raised the
possibility that impaired PGC-1a function might contribute
to the striatal degeneration, since the striatum is among the
most metabolically active regions of the brain (Beal et al.,
Figure 3. HD mice display abnormalities in body
composition, food intake, and metabolic control
A–C) We weighed and determined body composi-
tions for sets of 10-week-old HD N171-82Q mice
(red), 18Q mice (black dashed), and WT littermates
(black solid) during a feeding-fasting-refeeding chal-
lenge (n = 6–8 per group). While total body weights
were comparable (p = 0.38, 0.98, 0.11) (A), HD 82Q
mice had significantly higher % fat mass and lower
% lean mass at baseline and during fasting (p =
0.0002, 0.0005, fat mass; p = 0.002, 0.0001, lean
mass) (B–C). After refeeding, HD mice did not in-
crease fat mass, as did control mice (B–C).
D) Food intake measurements revealed a trend to-
ward hypophagia at baseline in HD 82Q mice (red)
in comparison to 18Q (black stippled) and WT (solid
black) controls (p = 0.034; posthoc for WT versus
82Q = 0.14, posthoc for 18Q versus 82Q = 0.04).
However, after fasting, hypophagia in HD 82Q mice
was dramatically manifested (p < .0001).
E) Locomotor activity was measured during this pro-
mice (red) displayed a trend toward decreased activ-
ity in comparison to 18Q (black stippled) and WT
(solid black) controls (p = 0.15). However, during
fasting, HD 82Q mice display a trend toward in-
creased activity (p = 0.25).
F) Oxygen consumption was determined by indirect
calorimetry, and was increased in HD 82Q mice (red)
at baseline, though not significantly (p = 0.102). Dur-
ing fasting, however, HD 82Q mice exhibited signifi-
to 18Q (black stippled) and WT (solid black) controls
(p = < .0001).
ing fasting. While PGC-1a levels did not change in
WT controls during fasting (p = 0.402 by t test),
PGC-1a levels markedly increased in HD 82Q mice
(p = 0.037 by t test). Group sizes were 5 or 6 mice,
and error bars are SEM.
H) Real-time RT-PCR analysis of UCP-1 in BAT dur-
ing fasting. As expected, UCP-1 levels dropped sig-
nificantly in WT controls (p = 0.039 by t test), UCP-1
failed to change in HD 82Q mice (p = 0.25 by t test).
Group sizes were 5 or 6 mice, and error bars are
Thermometabolic defects implicate PGC-1a in HD
CELL METABOLISM : NOVEMBER 2006353
1993; Brouillet et al., 1995). To determine if PGC-1a function
was compromised in the striatum of HD N171-82Q transgenic
mice, we isolated striatal RNA’s and measured the expression
level of PGC-1a target genes whose protein products mediate
oxidative metabolism in the mitochondria (Leone et al., 2005;
Mootha et al., 2003). In 20-week-old HD N171-82Q mice, there
was a significant reduction in the expression of such mitochon-
drial genes (Figure 4A). We also determined the expression level
of PGC-1a, and observed a significant decrease in PGC-1a in
striatal RNA’s from HD N171-82Q transgenic mice. These find-
ings support a role for PGC-1a transcription interference in the
degeneration of the striatum in HD.
Human HD patients display PGC-1a transcription
interference in the striatum
PGC-1atranscription abnormalities inthe brain andperiphery of
the N171-82Q HD model led us to ask: Do HD patients display
PGC-1a transcription interference in the striatum? To address
this question, we analyzed caudate nucleus microarray expres-
sion data obtained from a large cohort of human HD patients
and matched controls (Hodges et al., 2006). We selected 26
genes known to rely upon PGC-1a coactivator function for their
expression (Leone et al., 2005; Mootha et al., 2003), and using
the gcRMA application from the BioConductor software pro-
nificant reductions in 24 of these 26 PGC-1a target genes
(Figure 4B; Table 1). The presence of significant expression re-
ductions in 35 of 46 probes (corresponding to the 26 PGC-1a
target genes) from the Affymetrix HG-U133A/B 45,000 probe
set is highly unlikely to occur by chance (p < .0001; c2), and
thus strongly supports the existence of PGC-1a transcription in-
terference inthestriatumofpresymptomatic andearlystagehu-
RNA’s from a subset of these cases (Table S1), and performed
real-time RT-PCR analysis. We confirmed significant reductions
for mitochondrial PGC-1a target genes in the human HD sample
Figure 4. PGC-1a transcription interference in the
striatum of HD N171-82Q mice and human HD
A) Real-time RT-PCR analysis of striatal RNAs, ob-
tained from sets (n = 5 per group) of 20-week-old
HD 82Q mice (red), 18Q mice (gray), and WT mice
(black) indicate that RNA levels for PGC-1a and six
of its mitochondrial target genes are reduced in HD
brain (p < 0.05 for CYCS, NDUFS3, LDH-B,
COX6A1, and TFAM; p = 0.17 for ACADM).
B) Microarray expression analysis of PGC-1a-regu-
lated genes in human caudate. Here, we see a heat
map comparing the caudate nucleus expression of
26 PGC-1a target genes for 32 Grade 0–2 HD pa-
tients (adjacent to gold bar) and 32 matched controls
corresponds to degree of relative upregulation, while
intensity of green signal corresponds to degree of
relative downregulation. Most PGC-1a target genes
are downregulated in HD patients. Please see Table
1 for statistical values for the displayed probe data.
C) Confirmation of microarray data by real-time RT-
PCR. We obtained striatal RNA samples for a subset
of HD patients and matched controls (see Table S2),
and measured RNA expression levels for six PGC-1a
targets (NDUFS3; CYCS; COX7C; NDUFB5; ACADM;
and LDHB), PGC-1a, and two control genes (GFAP
and DRD2). In this way, we confirmed significant
reductions in the expression of PGC-1a targets and
detected reduced PGC-1a in human HD striatum
from early grade patients. Statistical comparisons
were performed with t test (*p < 0.05; **p < 0.005;
***p < 0.0005).
A R T I C L E
CELL METABOLISM : NOVEMBER 2006
set in comparison to human control samples (Figure 4C). To
control for the validation analysis, we included the glial fibrillary
acidic protein (GFAP) and dopamine D2 receptor (DRD2) genes
in this experiment, as GFAP is known to be significantly upregu-
lated and DRD2 is known to be significantly downregulated in
expression studies of HD caudate (Hodges et al., 2006; Luthi-
Carter et al., 2000). When we interrogated the expression level
of PGC-1a itself, the striatal microarray data did not reveal a sig-
nificant change in the PGC-1a level; however, real-time RT-PCR
analysis showed a reduction of PGC-1a in the striatum (Fig-
ure 4C). We then interrogated the striatal microarray data for ex-
pression of nuclear hormone receptors and transcription factors
known to rely upon PGC-1a for target gene transactivation.
Significant increases were noted for two or more oligonucleo-
tides from the PPAR-a, RXR-a, and NRF-1 genes (Figure S6,
Table S2), suggesting possible compensatory upregulation of
PGC-1a-dependent transcription factors in human HD caudate.
Interestingly, involvement of the RXR signaling pathway in HD
neurodegeneration was previously reported in microarray stud-
ies performed upon R6/2 striatum (Luthi-Carter et al., 2000).
Table 1. Summary of human caudate microarray data for PGC-1a pathway
Gene symbolProbeset ID fold changep valueFunction
electron carrier protein
The analyzed list of probes for PGC-1a target genes and pathway factors is shown with fold-change in the human HD caudate sample set given relative to the human
caudate control sample set. Methods for the calculation of individual p values are described in the Experimental Procedures section.
Thermometabolic defects implicate PGC-1a in HD
CELL METABOLISM : NOVEMBER 2006355
Functional abnormalities of mitochondrial energy
production in the CNS of HD mice
corresponded to abnormalities in mitochondrial function in the
CNS of HD mice, we performed two independent sets of exper-
iments. We began by measuring the oxygen consumption rate
(OCR) in brain slices from HD mice with a recently developed
flow culture system for continuous, noninvasive metabolic mon-
itoring of living cells (Sweet et al., 2004). This ex vivo approach
permitted us to record OCR’s in brain tissues in real time, and
to do so in response to different substrates (Figures 5A and
5B). When we measured OCR’s in brain slices from 11-week-
trols, OCR’s for 20 mM glucose were comparable (Figure 5C).
However, when brain slices were perfused with 5 mM lactate
trols (Figure 5D). In contrast, OCR response to succinate was
normal (Figure 5E), indicating that electron transport capacity
had not been compromised, and that the lesion in HD was local-
ized to a factor associated with the metabolism of the lactate-
pyruvate substrate. These results, using a novel ex vivo assay
system to measure brain OCR, document a specific mitochon-
drial abnormality in presymptomatic HD mice.
To further assess the role ofPGC-1ain HD mitochondrial dys-
model (Trettel et al., 2000). ST-HdhQ111striatal neurons have
been shown to display low mitochondrial ATP production
(Seong et al., 2005), and therefore should be exquisitely sensi-
tiveto mitochondrial toxins, suchas3-NP, acomplex IIinhibitor.
When ST-HdhQ111striatal neurons were treated with 3-NP in the
mitochondrial membrane potential, minimal orange fluores-
chondria, was noted (Figure 5F). However, when we analyzed
JC-1 staining in 3-NP treated ST-HdhQ111striatal neurons
expressing stably transfected PGC-1a, we observed mainte-
nance of mitochondrial membrane potential (Figure 5G). Thus,
expression of PGC-1a rescued the mitochondrial membrane
depolarization produced by 3-NP in HD striatal neurons in vitro,
presumably by bolstering mitochondrial function.
R6/2 mice display thermoregulatory defects,
but longer-lived HD mouse models do not
The presence of a thermoregulatory defect in the HD N171
transgenic mouse model led us to ask: Will other mouse models
for HD display similar abnormalities of thermoregulation? We
Figure 5. Functional mitochondrial abnormalities in
HD brain and striatal neurons
A–B) Representative traces from the brain slice OCR
experiment. WT and HD brain slices exhibit compa-
rable OCR’s when perfused with 20 mM glucose
‘‘C.’’ Addition of 5 mM lactate and 1 mM pyruvate
‘‘D’’ yields an increased OCR in WT brain, but has lit-
tle effect upon HD brain OCR. After rotenone treat-
ment to inhibit mitochondrial complex I, brain slices
were perfused with succinate ‘‘E,’’ and comparable
OCR’s were observed for WT and HD. To verify
that tissue hypoxia in these experiments was not
a factor, we quantified cytochrome c redox state
and confirmed that in the presence of glucose, lac-
tate, and pyruvate, it was highly oxidized, and capa-
ble of responding to succinate (data not shown).
C–E) Measurement of oxygen consumption rate
(OCR) in brain slices from 11-week-old HD and WT
mice using a flow-through perfusion system. For
20 mM glucose (C), HD and WT OCR’s were similar
(p = 0.694 by t test). However, for lactate/pyruvate
(D), the HD OCR was markedly reduced compared
to WT (p = 0.002 by t test). Subsequent perfusion
of succinate (E), however, then yielded comparable
OCR’s for HD and WT brain slices (p = 0.351 by
F–G) Resistance of ST-HdhQ111neurons expressing
PGC-1a to 3-NP. When ST-HdhQ111neurons were
treated with 100 mM 3-NP and stained with JC-1
(F), little orange fluorescence is seen, indicating
mitochondrial membrane depolarization. However,
when we treated ST-HdhQ111neurons stably ex-
pressing PGC-1a with 3-NP and stained them with
rescence, indicating normal mitochondrial mem-
A R T I C L E
CELL METABOLISM : NOVEMBER 2006
thus obtained HD R6/2 (Mangiarini et al., 1996), YAC72 (Hodg-
son et al., 1999), and htt knockin mice (Menalled et al., 2003),
as these mouse models are among the most commonly used
and comprise a representative sampling of available models.
We began by testing adaptive thermogenesis, and noted that
YAC72 and htt knockin mice could maintain normal body tem-
perature in the face of a cold challenge, but R6/2 mice could
not (Figures 6A–6C). We then charted body temperature in R6/
2 mice housed at ambient temperatures and recorded a signifi-
cant depression in body temperature in R6/2 mice beginning at
10 weeks of age (Figure 6D). Thus, of the four tested HD mouse
models, only those models characterized by decreased survival
and shortened lifespan exhibited thermoregulatory defects,
suggesting a possible link between the presence of thermoreg-
ulatory defects and death in HD mice.
Cold precipitates motor defects in HD mice, while higher
temperatures prolong survival
While evaluating R6/2 mice for a thermoregulatory defect, we
observed a correlation between failed adaptive thermogenesis
and accentuation of motor abnormalities. Many of the tested
R6/2 mice displayed a visible tremor phenotype with ataxia,
but otherwise were active and ambulatory, with normal explor-
atory behavior (Movie S1). However, upon cold challenge, all
R6/2 mice, that failed to maintain normal body temperature, de-
veloped a striking motor phenotype of shaking and tremors and
no longer explored their surroundings, but instead remained im-
mobile (Movie S2). Thus, exposure to cold in R6/2 mice with
a moderate phenotype made such mice appear as if they
were end-stage. As heat generation in BAT relies upon intact
PGC-1a coactivation, and HD transgenic mice display PGC-
1a transcription interference, HD transgenic mice that were un-
able to maintain body temperature likely resorted to shivering to
produce heat. Consequently, onset of shiveringin late-stage HD
mice may account for the worsening of their motor phenotype.
Since the cold-challenge appeared to accelerate the phenotype
temperatures might delay their premature death. To test this hy-
pothesis, we divided HD N171-82Q littermates into two groups:
the first group was housed at room temperature, while the sec-
ond group was housed at an ambient temperature of 30?C. HD
mice housed at the higher ambient temperature lived signifi-
cantly longer (Figure 6E), as average lifespan was extended 24
days (15%; p = .04).
HD is a neurodegenerative disorder characterized by selective
vulnerability of medium spiny neurons of the striatum and
of innervating corticostriatal projection neurons. Numerous
Figure 6. Role of adapative thermogenesis defects
in shortening lifespan in HD mice
mogenesis defect. Sets of 11-week-old R6/2 mice
(R6/2) and littermate controls (WT) were placed indi-
vidually at 4?C and body temperatures were re-
corded (n = 4 per group). By 1 hr into the cold chal-
lenge, HD R6/2 mice display a significant reduction
in body temperature (p < .05).
B) Sets of 17 month-old HD YAC-72Q mice (YAC
72Q) (Hodgson et al., 1999) and nontransgenic litter-
mate controls (WT) were placed individually at 4?C
and body temperatures were recorded (n = 3 per
group). YAC-72Q mice are capable of normal adap-
tive thermogenesis (p = 0.80).
C) Sets of 1 year-old HD knockin mice homozygous
for a 140 CAG repeat (140/140) (Menalled et al.,
2003) and nontransgenic littermate controls (WT)
were placed at 4?C and body temperatures were re-
corded (n = 3 per group). HD knockin mice maintain
body temperatures comparable to controls for the
duration of the cold challenge (p = 0.75).
D) Wemeasured body temperature weekly inHDR6/
2 male mice (red line) and age-matched WT controls
(black line) for their entire lifespan (n = 4 per group).
Beginning at 70 days of age, R6/2 mice display a sig-
nificant reduction in body temperature (p < .05).
E) HD N171-82Q mice live longer at higher ambient
temperature. We established cohorts of N171-82Q
mice by dividing littermates into two groups (n =
5–6 per group), and housing each group at either
room temperature (20?C; blue) or an elevated tem-
perature (30?C; orange). Mice housed at 30?C dis-
played a significant prolongation in median survival
(p = 0.046 by log-rank test).
Thermometabolic defects implicate PGC-1a in HD
CELL METABOLISM : NOVEMBER 2006 357
independent lines of evidence have implicated mitochondrial
dysfunction and impaired energy metabolism in HD (Grunewald
and Beal, 1999). The molecular basis of disordered energy me-
tabolism in HD, however, remains unknown. Herein, we report
that HD transgenic mice display profound thermoregulatory
and metabolic defects. Our discovery of deranged thermoregu-
lation in HD mice led us to evaluate this pathway, which in
thalamus degeneration does occur in HD patients and trans-
genic mice (Li et al., 2003), N171-82Q mice appropriately upre-
gulated PGC-1a in BAT in response to hypothalamic-mediated
sympathetic stimulation and activated c-fos in the ventromedial
nucleus of the hypothalamus in response to cold, indicating that
the thermoregulatory defect does not originate in the CNS.
Rather, our data suggest that impaired thermogenesis in HD
mice stems from PGC-1a transcription interference in BAT.
As weight loss is a prominent feature in both HD patients
(Pratley et al., 2000; Robbins et al., 2006) and N171-82Q HD
mice, we quantified body composition, food intake, and energy
ical disease. During ad libitum feeding, the ratio of fat to lean
eating less food than controls. This combination of abnormali-
ties suggests that at 10 weeks of age, the disturbance of energy
mice, and this defect was greatly exaggerated by fasting. Unlike
the hyperphagia and rapid recovery of lost weight observed in
controls, food intake was markedly reduced during refeeding
in HD mice, and this defect, combined with inappropriately ele-
vated levels of physical activity and oxygen consumption, re-
sulted in a failure to replenish depleted fat stores. A key mech-
anism whereby energy expenditure is reduced during fasting
is via reduced UCP-1 expression in BAT, apparently triggered
in part by reduced leptin levels and decreased SNS outflow
(Scarpace et al., 1998; Sivitz et al., 1999). While we noted a sig-
nificant reduction in UCP-1 mRNA levels in fasted control mice,
we observed little change in BAT UCP-1 expression in HD mice,
despite the expected drop in serum leptin levels. The failure of
BAT UCP-1 from HD mice to decrease during fasting was asso-
ciated with an unexpected increase of PGC-1a expression, an
effect that was not seen in BAT from fasted control mice. The
basis for this paradoxical upregulation of PGC-1a in HD BAT
is unknown, and the hypothesis that it played a role in the mal-
adaptive, hypermetabolic response of HD mice to food depriva-
tion merits future study—as does the question of whether re-
duced food intake, increased metabolic rate, or both defects
contribute to weight loss in HD patients. Additional studies are
also warranted to determine whether the mechanism underlying
hypophagia in HD mice involves huntingtin-associated protein 1
(HAP1), a possibility raised by work indicating that HAP1 down-
regulation inthe hypothalamus decreasesfood intakein rodents
(Sheng et al., 2006), and that HAP1 is bound more avidly by
mutant than WT htt protein (Li et al., 1995).
Our findings support a model in which defective PGC-1a ac-
tivity links mitochondrial dysfunction in neurodegeneration to
thermoregulatory and metabolic defects in HD mice. Our finding
ATP/ADP ratio in BAT from HD mice, combined with reduced
expression of PGC-1a target genes involved in energy produc-
tion in BAT, suggests that reduced PGC-1a activity may cause
a global defect in mitochondrial function in HD mice. To deter-
mine if a similar defect occurs in the CNS, we measured the ex-
pression of PGC-1a target genes whose protein products par-
ticipate directly or indirectly in the mitochondrial respiratory
chain, and documented significant reductions in HD striatum.
To investigate whether reduced neuronal expression of PGC-
1a target genes was linked to impaired mitochondrial function
in HD brain (Beal, 2005; Grunewald and Beal, 1999), we mea-
sured OCR’s in an ex vivo brain slice preparation (Sweet et al.,
2004). These studies revealed a substrate-specific defect in mi-
tochondrial function when HD brain was perfused with lactate.
Unlike normal brain, the ability of lactate to be converted to py-
ruvate and subsequently undergo oxidative metabolism in the
citric acid cycle, a key step in fueling mitochondrial energy pro-
duction, was impaired in HD mouse brain. PGC-1a has a role in
driving this pathway, as it coactivates expression of lactate
dehydrogenase B (LDH-B) (Lin et al., 2004), and thereby favors
interconversion of lactate to pyruvate. Interestingly, although
expression of many PGC-1a gene targets was reduced in HD
striatum, levels of LDH-B were among the lowest. Indeed, mag-
netic resonance spectroscopy studies of human patients have
documented elevated lactate levels in HD brain (Harms et al.,
1997). These findings identify reduced LDH-B expression as
a potential contributor to the OCR reduction detected in HD
In experiments using a striatal (ST-HdhQ111) neuron cell cul-
ture system, we showed that the deleterious effect of the toxin
3-NP on mitochondrial membrane potential was prevented by
overexpression of PGC-1a. Thus, HD neurodegeneration is
characterized by both reduced PGC-1a activity and mitochon-
drial dysfunction while, conversely, increased PGC-1a activity
rescues toxin-induced mitochondrial dysfunction in normal
striatal neurons. These findings are consistent with previous ev-
idence that PGC-1a knockout mice are hyperactive and have
pronounced vacuolar degeneration in the striatum (Leone
et al., 2005; Lin et al., 2004). More significantly, HD neurodegen-
eration is enhanced when HD knockin mice are crossed with
mice lacking PGC-1a, whereas striatal overexpression of
PGC-1a attenuates neurodegeneration in R6/2 HD mice (Cui
et al., 2006).
Although our results provide strong functional, physiological,
and molecular evidence that PGC-1a dysfunction contributes to
HD, the mechanistic basis of the PGC-1a abnormality will re-
quire further investigation. Indeed, in addition to demonstrating
altered expression of PGC-1a gene targets in striatum and BAT,
RT-PCR analysis indicated that decreased expression of PGC-
1a itself may contribute to PGC-1a transcription interference in
striatum. Another study similarly documented a reduction in the
expression of PGC-1a and its target genes in HD striatum, and
has attributed this reduction to repression of PGC-1a gene ex-
pression bymutant htt (Cui etal., 2006). Ouranalysis ofadaptive
thermogenesis in HD BAT indicated that mutant htt also inter-
feres with transcription of genes downstream of PGC-1a.
Consistent with this notion, a yeast two-hybrid screen recently
identified PPARg as an htt interactor, and then validated the bi-
ological significance of the interaction by demonstrating an ef-
fect of PPARg dosage upon HD neurodegeneration in the fly
eye (J. Botas & R.E. Hughes, personal communication), sug-
gesting that PPARg could also be a target of mutant htt. Taken
together, all these studies suggest that reduced expression of
PGC-1a and its targets contributes to HD striatal degeneration
and support a role for mutant htt-mediated transcription
A R T I C L E
CELL METABOLISM : NOVEMBER 2006
interference upon PGC-1a. While future studies will sort out
the relative contributions of upstream and downstream effects,
PGC-1a dysfunction may be central to HD pathogenesis.
PGC-1a is an appealing candidate, as it would provide a link
between two established aspects of HD molecular pathol-
ogy—transcription dysregulation and mitochondrial dysfunc-
tion. If altered PGC-1a function does cause HD mitochondrial
dysfunction, then PGC-1a deserves consideration as a prime
One item worthy of consideration in any mouse model study
of a human disease is the relevance of the murine findings to
the human disorder. The present study raises a number of note-
and therefore do not regulate body temperature as rodents do,
there is little reason to expect that HD patients will display hypo-
thermia.Toanticipate suchaparallel, orto concludethatthe rel-
evance of our findings hinges upon the existence of this pheno-
type in human HD patients would be incorrect. As countless
studies in C. elegans, Drosophila, and yeast model systems
have demonstrated, dissection of interesting phenotypes in
model organisms can provide clues to mechanistic pathways
that underlie human disease processes, even when the model
organism phenotype cannot be directly extrapolated to the hu-
man. Nonetheless, to gauge the relevance of our PGC-1a find-
ing in HD mice to striatal neurodegeneration in human HD, we
interrogated caudate microarray data from a large cohort of
Grade 0–2 HD patients for PGC-1a targets. The results of this
analysis were compelling, as 24 of 26 PGC-1a target genes
were coordinately downregulated, and strongly supported a
role for PGC-1a in human HD pathogenesis.
While some clarification of the relevance of our findings to
human HD may be necessary, there can be little doubt that dis-
covery of impaired thermogenesis in HD mice has immediate,
important consequences for preclinical trial design and our un-
derstanding of why HD mice die. Our studies indicate that hypo-
thermia is a reliable end-stage feature in N171-82Q mice, and
therefore may be used as a humane surrogate marker of death
when survival studies are not ethically permitted. To determine
the broader utility of this finding, we recorded body tempera-
tures and assessed adaptive thermogenesis in R6/2 (Mangiarini
et al.,1996),YAC72 (Hodgson et al.,1999),andhtt knockinmice
(Menalled et al., 2003), but observed thermoregulatory defects
only in R6/2 mice. Furthermore, we noted that cold challenge,
resulting in hypothermia in mid-stage R6/2 mice, elicited a
severe motor phenotype that resembled end-stage. We hypoth-
esized that abnormal thermoregulation may play a role in the
death of such HD mice, and found that HD mice raised at
30?C significantly outlived littermate HD mice maintained at
room temperature. That the immediate cause of death in HD
mice involves dysfunction outside the CNS was expected,
since: (1) In a heat shock protein rescue of the R6/2 HD mouse
model, transgenic overexpression of heat shock factor 1 re-
stricted to nonneural tissues significantly prolonged R6/2 life-
span (Fujimoto et al., 2005); and (2) peripheral delivery of Congo
matic result at the time—as Congo red can not cross the blood-
essary for a complete therapeutic response in preclinical trials.
and aging is the role of mitochondria in the process of neuronal
dysfunction and death. Numerous mutations in the mitochon-
drial genome have been characterized in neurological and neu-
romuscular disorders, clearly establishing that postmitotic neu-
rons and muscle cells are exquisitely sensitive to impaired
energy metabolism (DiMauro and Schon, 2003). Friedreich’s
ataxia, a disorder of nerve, muscle, and heart, is caused by
loss of function of frataxin, a protein involved in the production
of iron-sulfur containing enzymes in the mitochondrial respira-
tory chain (Bulteau et al., 2004). Dominant optic neuropathy,
hereditary spastic paraplegia, neurodegeneration with brain
iron accumulation (formerly Hallervorden–Spatz disease), and
Charcot-Marie-Tooth disease type 2a all result from mutations
in nuclear genes whose protein products localize to the mito-
chondria (Beal, 2005). In each case, evidence for impaired en-
ergy production and/or impaired responses to oxidative stress
has been demonstrated. Although considerable evidence for
CNS metabolic and mitochondrial abnormalities exists in HD,
their mechanistic basis has remained elusive. Our study indi-
cates that an evaluation of metabolic processes occurring in
nonneuronal tissues in the periphery can yield crucial factors
and pathways that contribute to neurodegenerative disease.
Consequently, we propose that careful consideration of periph-
eral metabolic processes in neurological diseases could shed
light on fundamental abnormalities occurring in muscle, nerve,
and glial cells. The utility of this paradigm for deconstructing
neurodegenerative disease is yet to be fully tested.
Body temperature and behavioral analysis
All mice were housed at the University of Washington transgenic animal facil-
ity. All experiments and animal care were performed in accordance with the
University of Washington IACUC guidelines. Mice were checked at least 33
per week for survival, weight loss, and motor abnormalities. Body tempera-
ture was monitored every day at noon with a telemetry system using subcu-
taneously implanted transponders placed in the interscapular space (Bio
Medic Data Systems, Seaford, DE). For cold challenges, mice were housed
individually and exposed to 4?C for 3–24 hr in a cardboard box with bedding
and ample food. Rotarod analysis was performed as previously described
(Garden et al., 2002). For the warm-room survival study, 13-week-old sex
and weight-matched mice were housed in environmental chambers in the
decentralized animal facility of the University of Washington at either 30?C
Real-time RT-PCR and Western blot analysis
Total RNAs were isolated with the Qiagen RNeasy Lipid tissue extraction kit
(Qiagen,Valencia,CA).GenomicDNAwasremoved using RNasefreeDNase
(Ambion; Austin, TX). Quantification of mRNA was performed using an
Applied Biosystems 7500 Real Time Sequence Detection System with ABI
Assays-on-Demand primers and TaqMan based probes (Livak et al., 1995).
Selected ABI TaqMan primer and probe set designations are available
upon request. 18S RNA (human) and b-actin RNA (mouse) were used as in-
ternal controls to normalize results. Relative expression levels were calcu-
lated via the standard curve method (La Spada et al., 2001). UCP-1 antibody
was used for immunoblotting as previously described (Luquet et al., 2003).
Cell culture experiments
3T3-L1 preadipocyte cells were transiently transfected with the UCP-1 pro-
moter-reporter, PPARg, RXRa, PGC-1a, and htt expression constructs
(Puigserver et al., 1998; Rim and Kozak, 2002; Vega et al., 2000), and then
treated with troglitazone (Sigma) for 24 hr. Primary cultures of brown adipo-
cytes were obtained from adult mice, and stimulated with norepinephrine for
3 hr (Chernogubova et al., 2005). For quantification of functional mitochon-
dria, we trypsinized primary cultured brown adipocytes, washed them, and
treated them with 100 nM Mitotracker Red CMX-Ros (Invitrogen) in complete
eter (Cytopeia), flow sorted, and their fluorescence measured.
Thermometabolic defects implicate PGC-1a in HD
CELL METABOLISM : NOVEMBER 2006359
Body composition, food intake, locomotor activity, and indirect
Age- and sex-matched control and HD mice were individually housed and
acclimated to metabolic cages for 3 days. Food intake, physical activity,
and calorimetric measurements were continuously recorded over a 24 hr
period during which food was available ad libitum. Subsequently, food was
removed for 18 hr beginning at the onset of the dark cycle. Food was then
replaced and measurements made for an additional 24 hr. Determinations
of body lean mass, fat mass and water content were made in conscious
body composition analyzer; Echo Medical Systems, Houston, TX). Locomo-
tor activity was assessed by the infrared beam break method using an Opto-
Varimetrix-3 sensor system,whilefood and waterintakewere measured with
the Feed-Scale System (Columbus Instruments, Columbus, OH). Indirect
calorimetry was performed with a computer-controlled open circuit calorim-
etry system (Oxymax; Columbus Instruments Co., Columbus, OH). Rates of
malized to lean body mass. Serum leptin levels were determined according
to the manufacturer’s instruction using the mouse/rat leptin assay kit (Crys-
talchem; Downer’s Grove, IL). Blood glucose levels were determined with
a commercial Accu-Chek Advantage glucometer (Roche).
Microarray expression analysis and bioinformatics
Using a recently published Affymetrix GeneChip data set (that is available
from the Gene Expression Omnibus http://www.ncbi.nlm.nih.gov/geo/GEO
with GEO Accession Number GSE3790) (Hodges et al., 2006), we compared
caudate array data from 32 Grade 0–2 human HD patients with 32 age-and
sex-matched controls for the expression levels of 26 PGC-1a target genes
(Leone et al., 2005; Mootha et al., 2003). For this analysis, we used the
gcRMA package available through the BioConductor project http://www.
bioconductor.org (Bolstad et al., 2003;Gentleman et al., 2004). After correct-
ing for nonspecific hybridization background noise by considering individual
probe sequences, we performed within and between group comparisons
with the ‘‘limma’’ package in BioConductor. The limma application employs
a modified t test to calculate p values using an empirical Bayes method to
moderate the standard errors of the estimated log-fold changes, and also
takes into account the variance information from all the genes on the array
to determine per gene variance for the t test calculations (Smyth et al.,
2005). p values were then adjusted for multiplicity with the q value program
(Storey and Tibshirani, 2003), as this application allows selection of statisti-
cally significant genes with simultaneous consideration of the expected
‘‘false discovery rate.’’ A heat map illustrating the results of the microarray
analysis was generated by clustering probe sets based upon sequence
and functional relationships.
For measurement of adenine nucleotides, ATP, ADP, AMP, and adenosine
present in TCA, extracts from brown adipose tissues were quantitatively
converted to fluorescent 1,N6-ethenoadenine derivatives as previously
described (Lazarowski et al., 2004).
Oxygen consumption rate measurements
Using a recently developed flow culture system for continuous, noninvasive
monitoring of living cells (Sweet et al., 2004), we performed measurements of
was loaded into each of 4 chambers and sandwiched between two layers of
Cytopore beads (Amersham Biosciences Corp.), as described previously
(Sweet et al., 2004; Sweet et al., 2002). Flow rate was set to approximately
40 ml/min for all perfusion experiments; chamber volume was w250 ml.
Data was corrected for delays in the flow system by referencing the OCR
and cytochrome c reduction data to the response to antimycin A, which
was given at the end of each flow culture experiment as described (Sweet
etal.,2004).Methodsforthecalculation ofOCRandcytochrome Creduction
have been described (Sweet et al., 2002, 2004), and will be furnished upon
Mitochondrial membrane depolarization studies
ST-HdhQ111striatal neurons expressing stably transfected PGC-1a were de-
rived with the pcDNA3- PGC-1a construct that drives expression of PGC-1a
with the CMV promoter. ST-HdhQ111striatal neurons were incubated with
100 mM 3-NP for 1 hr, and then stained with JC-1 (5, 50, 6, 60–tetrachloro-
1, 10, 3, 30tetraethylbenzimidazol-carbocyanine iodide) for 10 min at 37?C.
AftertwowasheswithPBS,stained neuronswere visualized underaFluores-
All data were prepared for analysis with standard spread sheet software (Mi-
crosoft Excel). Statistical analysis was done using Prism 4.0 (Graph Pad) or
the VassarStats website http://faculty.vassar.edu/lowry/VassarStats.html.
We performed ANOVA unless indicated otherwise. If statistical significance
(p < 0.05) was achieved, we performed posttest analysis to account for
Histology & Immunohistochemistry
Protocols in Supplemental Experimental Procedures.
mental References, three figures, two tables, and two movies and can be
found with this article online at http://www.cellmetabolism.org/cgi/content/
The authorswishtothankJ.Olsonforhelpfuladviceandfor providingthehu-
man DNA samples, P. Calses, K. Ogimoto, and J. Choi for technical assis-
tance, S. Finkbeiner for the htt exon 1-2 expression constructs, D. Kelly for
the PPARg and RXRa expression constructs, L. Kozak for the UCP-1 pro-
moter-reporter construct, and C. Ross for HD N171-18Q mice. This work
was supported by funding from Hereditary Disease Foundation, High Q,
and grants from the National Institutes of Health (NIH) (DK17047 and
DK063986 to I.R.S.; NS050352 to D.K.). Body composition and energy
expenditure measurements were performed with support from the Clinical
Nutrition Research Unit at the University of Washington. A.R.L. is the recipi-
ent of a Paul Beeson Physician Faculty Scholar in Aging Research award
from the American Foundation for Aging Research (AFAR), and V.V.P. is an
NIH Genetics of Aging postdoctoral fellow (AG00057).
Received: January 11, 2006
Revised: September 20, 2006
Accepted: October 9, 2006
Published online: October 19, 2006
Beal, M.F. (2005). Mitochondria take center stage in aging and neurodegen-
eration. Ann. Neurol. 58, 495–505.
Beal, M.F., Brouillet, E., Jenkins, B.G., Ferrante, R.J., Kowall, N.W., Miller,
J.M., Storey, E., Srivastava, R., Rosen, B.R., and Hyman, B.T. (1993). Neuro-
chemical and histologic characterization of striatal excitotoxic lesions pro-
duced by the mitochondrial toxin 3-nitropropionic acid. J. Neurosci. 13,
mouse models of Huntington’s disease. Nat. Rev. Neurosci. 5, 373–384.
Bolstad, B.M., Irizarry, R.A., Astrand, M., and Speed, T.P. (2003). A compar-
ison of normalization methods for high density oligonucleotide array data
based on variance and bias. Bioinformatics 19, 185–193.
Brouillet, E., Hantraye, P., Ferrante, R.J., Dolan, R., Leroy-Willig, A., Kowall,
N.W., and Beal, M.F. (1995). Chronic mitochondrial energy impairment pro-
duces selective striatal degeneration and abnormal choreiform movements
in primates. Proc. Natl. Acad. Sci. USA 92, 7105–7109.
Browne, S.E., Bowling, A.C., MacGarvey, U., Baik, M.J., Berger, S.C., Muqit,
M.M., Bird, E.D., and Beal, M.F. (1997). Oxidative damage and metabolic
dysfunction in Huntington’s disease: selective vulnerability of the basal gan-
glia. Ann. Neurol. 41, 646–653.
A R T I C L E
CELL METABOLISM : NOVEMBER 2006
Bulteau, A.L., O’Neill, H.A., Kennedy, M.C., Ikeda-Saito, M., Isaya, G., and
Szweda, L.I. (2004). Frataxin acts as an iron chaperone protein to modulate
mitochondrial aconitase activity. Science 305, 242–245.
Cannon, B., and Nedergaard, J. (2004). Brown adipose tissue: function and
physiological significance. Physiol. Rev. 84, 277–359.
Chernogubova, E., Hutchinson, D.S., Nedergaard, J., and Bengtsson, T.
(2005). Alpha1- and beta1-adrenoceptor signaling fully compensates for
beta3-adrenoceptor deficiency in brown adipocyte norepinephrine-stimu-
lated glucose uptake. Endocrinology 146, 2271–2284.
Cui, L., Jeong, H., Borovecki, F., Parkhurst, C.N., Tanese, N., and Krainc, D.
(2006). Repression of PGC-1a gene by mutant huntingtin leads to mitochon-
drial dysfunction in Huntington’s disease. Cell 127, 59–69.
DiMauro, S., and Schon, E.A. (2003). Mitochondrial respiratory-chain dis-
eases. N. Engl. J. Med. 348, 2656–2668.
Fujimoto, M., Takaki, E., Hayashi, T., Kitaura, Y., Tanaka, Y., Inouye, S., and
Nakai, A. (2005). Active HSF1 significantly suppresses polyglutamine aggre-
gate formation in cellular and mouse models. J. Biol. Chem. 280, 34908–
Garden, G.A., Libby, R.T., Fu, Y.H., Kinoshita, Y., Huang, J., Possin, D.E.,
Smith, A.C., Martinez, R.A., Fine, G.C., Grote, S.K., et al. (2002). Polyglut-
amine-expanded ataxin-7 promotes non-cell-autonomous purkinje cell
degeneration and displays proteolytic cleavage in ataxic transgenic mice.
J. Neurosci. 22, 4897–4905.
Gentleman, R.C., Carey, V.J., Bates, D.M., Bolstad, B., Dettling, M., Dudoit,
S., Ellis, B., Gautier, L., Ge, Y., Gentry, J., et al. (2004). Bioconductor: open
software development for computational biology and bioinformatics. Ge-
nome Biol. 5, R80.
Grunewald, T., and Beal, M.F. (1999). Bioenergetics in Huntington’s disease.
Ann. N Y Acad. Sci. 893, 203–213.
Gu, M., Gash, M.T., Mann, V.M., Javoy-Agid, F., Cooper, J.M., and Schapira,
A.H. (1996). Mitochondrial defect in Huntington’s disease caudate nucleus.
Ann. Neurol. 39, 385–389.
Gurney, M.E., Pu, H., Chiu, A.Y., Dal Canto, M.C., Polchow, C.Y., Alexander,
D.D., Caliendo, J., Hentati, A., Kwon, Y.W., Deng, H.X., et al. (1994). Motor
neuron degenerationin mice that express a human Cu,Zn superoxide dismu-
tase mutation. Science 264, 1772–1775.
Harms,L.,Meierkord, H.,Timm,G.,Pfeiffer, L.,andLudolph,A.C.(1997). De-
creased N-acetyl-aspartate/choline ratio and increased lactate in the frontal
lobe of patients with Huntington’s disease: a proton magnetic resonance
spectroscopy study. J. Neurol. Neurosurg. Psychiatry 62, 27–30.
Hodges, A., Strand, A.D., Aragaki, A.K., Kuhn, A., Sengstag, T., Hughes, G.,
Elliston, L.A., Hartog, C., Goldstein, D.R., Thu, D., et al. (2006). Regional and
Mol. Genet. 15, 965–977.
Hodgson, J.G., Agopyan, N., Gutekunst, C.A., Leavitt, B.R., LePiane, F., Sin-
garaja, R., Smith, D.J., Bissada, N., McCutcheon, K., Nasir, J., et al. (1999). A
YAC mouse model for Huntington’s disease with full-length mutant hunting-
tin, cytoplasmic toxicity, and selective striatal neurodegeneration. Neuron
Huntington’s Disease Collaborative Research Group (1993). A novel gene
containing a trinucleotide repeat that is expanded and unstable on Hunting-
ton’s disease chromosomes. Cell 72, 971–983.
La Spada, A.R., Fu, Y., Sopher, B.L., Libby, R.T., Wang, X., Li, L.Y., Einum,
D.D., Huang, J., Possin, D.E., Smith, A.C., et al. (2001). Polyglutamine-
expanded ataxin-7 antagonizes CRX function and induces cone-rod dystro-
phy in a mouse model of SCA7. Neuron 31, 913–927.
Lazarowski,E.R.,Tarran,R., Grubb,B.R., vanHeusden, C.A.,Okada, S., and
Boucher, R.C. (2004). Nucleotide release provides a mechanism for airway
surface liquid homeostasis. J. Biol. Chem. 279, 36855–36864.
Leone, T.C., Lehman, J.J., Finck, B.N., Schaeffer, P.J., Wende, A.R., Bou-
dina, S., Courtois, M., Wozniak, D.F., Sambandam, N., Bernal-Mizrachi, C.,
et al. (2005). PGC-1alpha deficiency causes multi-system energy metabolic
derangements: muscle dysfunction, abnormal weight control and hepatic
steatosis. PLoS Biol. 3, e101.
Li, S.H., Yu, Z.X., Li, C.L., Nguyen, H.P., Zhou, Y.X., Deng, C., and Li, X.J.
(2003). Lack of huntingtin-associated protein-1 causes neuronal death re-
sembling hypothalamic degeneration in Huntington’s disease. J. Neurosci.
Li, X.J., Li, S.H., Sharp, A.H., Nucifora, F.C., Jr., Schilling, G., Lanahan, A.,
Worley, P., Snyder, S.H.,and Ross, C.A.(1995). A huntingtin-associated pro-
tein enriched in brain with implications for pathology. Nature 378, 398–402.
Lin, J., Handschin, C., and Spiegelman, B.M. (2005). Metabolic control
through the PGC-1 family of transcription coactivators. Cell Metab. 1, 361–
tha, V.K., Jager, S., Vianna, C.R., Reznick, R.M., et al. (2004). Defects in
adaptive energy metabolism with CNS-linked hyperactivity in PGC-1alpha
null mice. Cell 119, 121–135.
Livak, K.J., Flood, S.J., Marmaro, J., Giusti, W., and Deetz, K. (1995). Oligo-
nucleotides with fluorescent dyes at opposite ends provide a quenched
probe system useful for detecting PCR product and nucleic acid hybridiza-
tion. PCR Methods Appl. 4, 357–362.
Lodi, R., Schapira, A.H., Manners, D., Styles, P., Wood, N.W., Taylor, D.J.,
in Huntington’s disease and dentatorubropallidoluysian atrophy. Ann. Neu-
rol. 48, 72–76.
Lowell, B.B., and Spiegelman, B.M. (2000). Towards a molecular under-
standing of adaptive thermogenesis. Nature 404, 652–660.
Luquet, S., Lopez-Soriano, J., Holst, D., Fredenrich, A., Melki, J., Rassoulza-
degan, M., and Grimaldi, P.A. (2003). Peroxisome proliferator-activated re-
ceptor delta controls muscle development and oxidative capability. FASEB
J. 17, 2299–2301.
Luthi-Carter, R., Strand, A., Peters, N.L., Solano, S.M., Hollingsworth, Z.R.,
Menon, A.S., Frey, A.S., Spektor, B.S., Penney, E.B., Schilling, G., et al.
(2000). Decreased expression of striatal signaling genes in a mouse model
of Huntington’s disease. Hum. Mol. Genet. 9, 1259–1271.
Mangiarini, L., Sathasivam, K., Seller, M., Cozens, B., Harper, A., Hethering-
ton, C., Lawton, M., Trottier, Y., Lehrach, H., Davies, S.W., and Bates, G.P.
(1996). Exon 1 of the HD gene with an expanded CAG repeat is sufficient
to cause a progressive neurological phenotype in transgenic mice. Cell 87,
Menalled, L.B., Sison, J.D., Dragatsis, I., Zeitlin, S., and Chesselet, M.F.
(2003). Time course of early motor and neuropathological anomalies in
a knock-in mouse model of Huntington’s disease with 140 CAG repeats.
J. Comp. Neurol. 465, 11–26.
Mootha, V.K., Lindgren, C.M., Eriksson, K.F., Subramanian, A., Sihag, S.,
Lehar, J., Puigserver, P., Carlsson, E., Ridderstrale, M., Laurila, E., et al.
(2003). PGC-1alpha-responsive genes involved in oxidative phosphorylation
Nance, M.A. (1997). Clinical aspects ofCAGrepeatdiseases.Brain Pathol. 7,
Pratley, R.E., Salbe, A.D., Ravussin, E., and Caviness, J.N. (2000). Higher
sedentary energy expenditure in patients with Huntington’s disease. Ann.
Neurol. 47, 64–70.
Puigserver, P., and Spiegelman, B.M. (2003). Peroxisome proliferator-
coactivator and metabolic regulator. Endocr. Rev. 24, 78–90.
Puigserver, P., Wu, Z., Park, C.W., Graves, R., Wright, M., and Spiegelman,
tive thermogenesis. Cell 92, 829–839.
Rim, J.S., and Kozak, L.P. (2002). Regulatory motifs for CREB-binding pro-
tein and Nfe2l2 transcription factors in the upstream enhancer of the mito-
chondrial uncoupling protein 1 gene. J. Biol. Chem. 277, 34589–34600.
Thermometabolic defects implicate PGC-1a in HD
CELL METABOLISM : NOVEMBER 2006361
Robbins, A.O., Ho, A.K.,and Barker, R.A. (2006). Weight changesin Hunting- Download full-text
ton’s disease. Eur. J. Neurol. 13, e7.
Sanchez, I., Mahlke, C., and Yuan, J. (2003). Pivotal role of oligomerization in
expanded polyglutamine neurodegenerative disorders. Nature 421, 373–
Scarpace, P.J., Nicolson, M., and Matheny, M. (1998). UCP2, UCP3 and lep-
tin gene expression: modulation by food restriction and leptin. J. Endocrinol.
Schilling, G., Becher, M.W., Sharp, A.H., Jinnah, H.A., Duan, K., Kotzuk, J.A.,
Slunt, H.H., Ratovitski, T., Cooper, J.K., Jenkins, N.A., et al. (1999). Intra-
nuclear inclusions and neuritic aggregates in transgenic mice expressing a
mutant N-terminal fragment of huntingtin. Hum. Mol. Genet. 8, 397–407.
Seong, I.S., Ivanova, E., Lee, J.M., Choo, Y.S., Fossale, E., Anderson, M.,
Gusella, J.F., Laramie, J.M., Myers, R.H., Lesort, M., and MacDonald, M.E.
(2005). HD CAG repeat implicates a dominant property of huntingtin in mito-
chondrial energy metabolism. Hum. Mol. Genet. 14, 2871–2880.
Sheng, G., Chang, G.Q., Lin, J.Y., Yu, Z.X., Fang, Z.H., Rong, J., Lipton, S.A.,
Li, S.H., Tong, G., Leibowitz, S.F., and Li, X.J. (2006). Hypothalamic hunting-
tin-associated protein 1 as a mediator of feeding behavior. Nat. Med. 12,
Sivitz, W.I., Fink, B.D., and Donohoue, P.A. (1999). Fasting and leptin modu-
late adipose and muscle uncoupling protein: divergent effects between mes-
senger ribonucleic acid and protein expression. Endocrinology 140, 1511–
Smyth, G.K., Michaud, J., and Scott, H.S. (2005). Use of within-array repli-
cate spots for assessing differential expression in microarray experiments.
Bioinformatics 21, 2067–2075.
Stoessl, A.J., Martin, W.R., Clark, C., Adam, M.J., Ammann, W., Beckman,
J.H., Bergstrom, M., Harrop, R., Rogers, J.G., Ruth, T.J., et al. (1986). PET
studies of cerebral glucose metabolism in idiopathic torticollis. Neurology
Storey, J.D., and Tibshirani, R. (2003). Statistical significance for genome-
wide studies. Proc. Natl. Acad. Sci. USA 100, 9440–9445.
Sugars, K.L., and Rubinsztein, D.C. (2003). Transcriptional abnormalities in
Huntington disease. Trends Genet. 19, 233–238.
Sweet, I.R., Cook, D.L., DeJulio, E., Wallen, A.R., Khalil, G., Callis, J., and
Reems, J. (2004). Regulation of ATP/ADP in pancreatic islets. Diabetes 53,
Sweet, I.R., Khalil, G., Wallen, A.R., Steedman, M., Schenkman, K.A.,
Reems, J.A., Kahn, S.E., and Callis, J.B. (2002). Continuous measurement
of oxygen consumption by pancreatic islets. Diabetes Technol. Ther. 4,
Trettel, F., Rigamonti, D., Hilditch-Maguire, P., Wheeler, V.C., Sharp, A.H.,
Persichetti, F., Cattaneo, E., and MacDonald, M.E. (2000). Dominant pheno-
types produced by the HD mutation in STHdh(Q111) striatal cells. Hum. Mol.
Genet. 9, 2799–2809.
Vega, R.B., Huss, J.M., and Kelly, D.P. (2000). The coactivator PGC-1 coop-
erates with peroxisome proliferator-activated receptor alpha in transcrip-
tional control of nuclear genes encoding mitochondrial fatty acid oxidation
enzymes. Mol. Cell. Biol. 20, 1868–1876.
Wang, L., Liu, J., Saha, P., Huang, J., Chan, L., Spiegelman, B., and Moore,
sion and energy production in brown adipocytes. Cell Metab. 2, 227–238.
Zoghbi, H.Y., and Orr, H.T. (2000). Glutamine repeats and neurodegenera-
tion. Annu. Rev. Neurosci. 23, 217–247.
A R T I C L E
CELL METABOLISM : NOVEMBER 2006