Journal of Neuropathology and Experimental Neurology
Copyright q 2003 by the American Association of Neuropathologists
Vol. 62, No. 10
October, 2003
pp. 1006 1018
Gene Expression Profiling in Ataxin-3 Expressing Cell Lines Reveals Distinct Effects of Normal
and Mutant Ataxin-3
BERND O. EVERT, PHD, INA R. VOGT, MD, ANA M. VIEIRA-SAECKER, LUCIA OZIMEK, PHD, ROB A. I.
EWOUT R. P. BRUNT, MD, THOMAS KLOCKGETHER, MD, AND ULLRICH WÜLLNER, MD
DE
VOS, MD,
Abstract. Spinocerebellar ataxia type 3 (SCA3) is a late-onset neurodegenerative disorder caused by the expansion of a
polyglutamine tract within the gene product, ataxin-3. We have previously shown that mutant ataxin-3 causes upregulation of
inflammatory genes in transgenic SCA3 cell lines and human SCA3 pontine neurons. We report here a complex pattern of
transcriptional changes by microarray gene expression profiling and Northern blot analysis in a SCA3 cell model. Twentythree differentially expressed genes involved in inflammatory reactions, nuclear transcription, and cell surface-associated
processes were identified. The identified corresponding proteins were analyzed by immunohistochemistry in human disease
and control brain tissue to evaluate their implication in SCA3 pathogenesis. In addition to several inflammatory mediators
upregulated in mutant ataxin-3 expressing cell lines and pontine neurons of SCA3 patients, we identified a profound repression
of genes encoding cell surface-associated proteins in cells overexpressing normal ataxin-3. Correspondingly, these genes were
upregulated in mutant ataxin-3 expressing cell lines and in pontine neurons of SCA3 patients. These findings identify for the
first time target genes transcriptionally regulated by normal ataxin-3 and support the hypothesis that both loss of normal
ataxin-3 and gain of function through protein-protein interacting properties of mutant ataxin-3 contribute to SCA3 pathogenesis.
Key Words: Ataxin-3; Differential gene expression; Extracellular matrix; Inflammation; Oligonucleotide microarray; SCA3;
Transcriptional regulation.
INTRODUCTION
Spinocerebellar ataxia type 3 (SCA3) or Machado-Joseph disease (MJD) is an autosomal dominantly inherited
neurodegenerative disorder with a wide range of clinical
manifestations, including ataxia, ophthalmoplegia, pyramidal signs, basal ganglia symptoms, and peripheral neuropathy. The mutation causing SCA3 is an unstable CAG
trinucleotide repeat expansion within exon 10 of a gene
encoding ataxin-3 (1). Ataxin-3 is ubiquitously expressed
in the central nervous system, while neuronal cell death
in SCA3 is limited to distinct subcortical brain regions
and is predominantly observed in the pontine nuclei, the
dentate nucleus, the pallidum, and the spinal cord (2).
The mutational mechanism places SCA3 in a group of
CAG repeat or polyglutamine (polyQ) disorders as the
expanded CAG repeat codes for elongated polyQ tracts
in the respective proteins. Despite their clinical variability, polyQ disorders share a number of molecular features, including the occurrence of nuclear and cytoplasmic protein aggregates, aberrant protein-protein
interactions, and profound transcriptional changes (3–5).
At present, 10 polyQ disorders have been identified:
From Department of Neurology (BOE, IRV, AMV-S, LO, TK, UW),
University of Bonn, Bonn, Germany; Laboratorium Pathologie Oost
Nederland (RAIdV), Enschede, The Netherlands; Department of Neurology (ERPB), University of Groningen, Groningen, The Netherlands.
Correspondence to: Bernd Evert, PhD, Department of Neurology,
University of Bonn, Sigmund-Freud-Str. 25, 53105 Bonn, Germany.
E-mail: [email protected]
This study was supported by the Deutsche Forschungsgemeinschaft
(DFG), a University of Bonn Center Grant (BONFOR) and the Nationalen Genomforschungsnetz (NGFN).
Huntington disease (HD), dentatorubropallidoluysian atrophy (DRPLA), spinal bulbar muscular atrophy
(SBMA), and several of the spinocerebellar ataxias
(SCA1, 2, 3, 6, 7, 12, and 17).
Transcriptional dysregulation appears to represent a
unifying pathogenic mechanism in polyQ disorders. A
recent comparison of gene expression changes in transgenic mice models of HD, DRPLA, SBMA, and SCA7
revealed a number of mRNA changes common to these
disorders (6). Sequestration of the transcriptional coactivator cAMP-response element (CREB)-binding protein
(CBP) by polyQ proteins has been suggested as a unifying
mechanism (7–12). Shimohata et al (13) found that expanded polyQ stretches preferentially bind to TAFII130, a
coactivator involved in CREB-dependent transcriptional
activation, thereby strongly suppressing CREB-dependent transcriptional activation. Recent experiments, however, revealed that transcriptional changes in polyQ diseases are not confined to CREB/CBP-mediated
transcription but rather display a disease-specific pattern.
In HD, in addition to genes dependent on CREB/CBPmediated transcription (14, 15), a number of genes involved in cholesterol and fatty acid metabolism were differentially expressed (16). In SCA1, genes involved in
neuronal calcium signaling were downregulated prior to
the development of any overt pathology, thus pointing to
a mechanism of action of ataxin-1 at the RNA level (17,
18). In SBMA, the observed changes in gene expression
revealed that the mutant androgen receptor undergoes a
partial loss of function and fails to regulate a subset of
genes that are dependent on ligand activation of the wild
type receptor (19). It is therefore conceivable that both
1006
GENE PROFILING IN ATAXIN-3 EXPRESSING CELL LINES
specific features of the mutant proteins and a loss of function of the normal proteins contribute to the molecular
mechanisms that underlie the selective neurodegeneration
observed in the various polyQ diseases.
We have previously identified upregulation of inflammatory genes in transgenic cell lines expressing mutant
ataxin-3 and in human SCA3 pontine neurons (20). With
respect to the putative pathogenetic mechanisms, the critical role of the nuclear localization of the expanded disease proteins points to the transcription machinery rather
than cytoplasmic protein-protein interactions. In human
SCA3 disease tissue and a Drosophila model of SCA3,
the transcription factor TATA-binding protein (TBP) and
the nuclear protein eyes absent protein (EYA) are recruited into nuclear inclusions, suggesting a direct interaction of mutant ataxin-3 with specific transcription factors (21). Further, it has been shown that mutant ataxin-3
colocalizes and completely immobilizes CBP into nuclear
inclusions, suggesting that CBP-mediated gene transcription is altered in SCA3 (9, 11, 12). In addition, a recent
study showed that normal ataxin-3 has histone binding
properties by itself and exerts transcriptional corepressor
activities in reporter gene studies (22). Thus a complex
pattern of transcriptional changes mediated by the mutant
protein and its aberrant interaction with important transcription factors on one hand, and disturbed transcriptional regulator activity of ataxin-3 on the other hand, is
to be expected. We therefore undertook a microarray
analysis in CSM14.1 clonal cell lines expressing normal
or mutant human full-length ataxin-3 to identify differentially expressed transcripts. Identified genes were further investigated using Northern blot analyses and finally
by immunohistochemical analyses of human brain tissue.
MATERIALS AND METHODS
Cell Lines and Culture
Gene expression profiling and Northern analysis were performed with the previously described inducible double stable
rat mesencephalic CSM14.1 clonal cell lines expressing nonexpanded (SCA3-Q23#1 and -#2) or expanded (SCA3-Q70#1 and
-#2) human full-length ataxin-3 and 2 double stable mock-transfected clonal control cell lines (Ctrl#1 and -#2) expressing only
the selection markers for hygromycin and puromycin (23). In
this SCA3 cell model, high level expression of the recombinant
human full-length ataxin-3 isoforms is induced after withdrawal
of tetracycline, while maintaining tetracycline (1 mg/ml) in media significantly reduces but does not eliminate recombinant
ataxin-3 expression. Therefore, the transgenic cell lines were
used in the induced condition for microarray and Northern analyses. Expression levels of the recombinant ataxin-3 isoforms in
the clonal cell lines were routinely screened by Western blot
analysis with the 1H9 antibody (kindly provided by Dr. Y. Trottier, France) as previously described (23).
RNA Preparation and Northern Analysis
Total RNA was isolated from individual CSM14.1 clonal cell
lines after 7 days of induced ataxin-3 expression at 338C. The
1007
procedures for RNA isolation and Northern analysis were identical to those previously described (20). The specific probes for
Northern analysis of the differentially expressed genes were
obtained by PCR amplification of a cDNA generated by reverse
transcription of 3 mg total CSM14.1 RNA using Oligod(T)30
primer and SuperScript (Gibco, Eggenstein, Germany) according to the manufacturer’s instructions. The primer pairs used for
generation of the specific probes are listed in Table 1. After
amplification, PCR products were purified using PCR purification spin columns (Qiagen, Hilden, Germany), 32P-radiolabeled with RadPrime labeling kit (Gibco) and hybridized in
ExpressHyb hybridization solution (Clontech, Palo Alto, CA)
according to the manufacturer’s instructions. Hybridization was
detected by autoradiography and intensifying screens after incubation at 2808C. Northern blot analyses were performed at
least in duplicate for all 43 genes identified by microarray analysis.
Sample Preparation, GeneChipT Array Hybridization,
and Data Analysis
Procedures followed the Affymetrix GeneChipt Expression
Analysis Manual (Affymetrix Inc., Santa Clara, CA). For gene
expression analysis, 4 RNA samples were isolated from separately cultured CSM14.1 clonal cell lines (SCA3-Q70#1 and
SCA3-Q23#2 in duplicate) expressing the respective ataxin-3
isoforms for 7 days at 338C. The RNA was cleaned up with the
RNeasy Mini Kit (Qiagen). Purity and integrity of the obtained
RNA were controlled by spectrophotometry and denaturing gel
electrophoresis. Twenty mg of total RNA were converted into
double-stranded cDNA using a T7-oligo-d(T) primer for firststrand synthesis with SSIIRT and self-priming for second strand
synthesis (Superscript Choice System, Invitrogen, Karlsruhe,
Germany). The cDNA was phenol-chloroform extracted, phases
were separated by the use of phase-lock-gels (Eppendorf, Hamburg, Germany) and subsequently ethanol precipitated. One mg
of double-stranded cDNA was used for biotinylated cRNA
preparation (Bioarray High Yield RNA Transcription Labeling
Kit; Enzo Diagnostics, Farmingdale, NY). The biotinylated
cRNA was purified (RNeasy Mini Kit, Qiagen), quantitated by
spectrophotometry, and fragmented. Fifteen mg of cRNA were
hybridized to Affymetrix Rat Genome U34A GeneChipt in the
Affymetrix Fluidics Station 400. Chips were washed with 63
SSPE at 258C, followed by stringent wash buffer (100 mM
MES pH 6.7; 0.1 M NaCl; 0.01% Tween 20) at 508C, stained
with streptavidin-phycoerythrin (Molecular Probes, Inc., Eugene, OR), washed again with 63 SSPE, stained with biotinylated anti-streptavidin IgG, followed by a second staining with
streptavidin-phycoerythrin and a final wash with 63 SSPE.
GeneChipt Arrays were scanned on an Affymetrix GeneArray
scanner. Data analysis was performed using the Affymetrix
Gene Chipt Expression analysis software and MicroArraySuite4.0 Algorithms according to the manufacturer’s recommendations, as described previously (24). The data sets obtained from mutant ataxin-3 and normal ataxin-3 expressing
clonal cell lines were compared giving a total of 4 pairwise
comparisons. For each comparison we used the default parameters in the GeneChip 4.0 software to identify genes that, at
minimum, met the p , 0.01 threshold as increased, decreased,
or not changed in the samples with respect to the baselines. The
J Neuropathol Exp Neurol, Vol 62, October, 2003
1008
EVERT ET AL
TABLE 1
Primer Pairs Used for Probe Generationa
Gene
Forward primer (59–39)
Reverse primer (59–39)
BDNF
CD9
C/EBPß
C/EBPd
C1s
EST190
EST195
EST212
EST106
Fit-1S
Hsp27
IL-1ra
IL-6
IRF-1
LANP
MKP-1
MMP-2
PA28ß
PPARg
TIMP1
TFP
UGT1A6
V3
ACTGCAGTGGACATGTCCGGTG
ATCGCAGTGCTTGCCATTGG
AGCACGAGCGCGCCATCGAC
CCACCCTAGAGTTGTGCCACG
TACTTTGAGTCCCCCCGAGG
TAGGGGATTTGGCAGGTAGC
ACATAGCTGCAAATCCCAGCTC
GTGGTAAGGGAACTGGGAAGG
GGGCACTCCTGAGCGTTTCG
TGAATCGATTAAGACCGGATC
GCGTGCCCTTCTCGCTACTGC
ATGGAAATCTGCAGGGGACC
CCTACTTCACAAGTCCGGAGAGG
AACAGTCTGAGTGGCAGCCGAC
CGAAAACAGAATCTCAGGGGAC
GGCAGTGCTTACCATGCTTCC
AACTTCTTTCCCCGCAA
AGCATGGCCAAGCCTTGTGG
TCTCAGTGGAGACCGCCCAG
CAGGCTTCAGCTTTTGCCAG
GGTCCTATCTGTCCCATGGAGG
GAAGCCTATGTCAACGCCTCC
TGAGACATGCGACTATGGCTGG
CCAGTCATCTGATTGGATTTTGC
AAGACCTCATCGATGGCATCC
CAGTGACCCGCCGAGGCCAG
GCTGCTCCACACGCTGATGC
TCCCACACTTTTTCCCCCAG
CTGCAGTCTGTGGCAAGGTGC
TGCTAAGAAGTGTGCTGGGTGC
TTGAGTACGTGGCCCCGGTG
TGCTGCTTCTCCCGCTGTGG
CCGTGTGAGTGAATCGACAT
CAGAGAGGGAGGGCTGGTGACG
TTTTGGTGTGTTGGTGAGGCTC
CATAGCACACTAGGTTTGCCGAG
TGGCTCTAGAACCATCCCAGG
AAATAAGTTTCGGGGGGCAG
AGCTCACATGAGCCTCTCCCAG
CCACCCATGGTAAACAA
CCAGCACCATGGCCCTAAGC
CCCTCAAAATAATAGTGCAATCG
CAGCTTTCTGCAACTCGGACC
TCTGTTCTTTCTGCGCTTGCAC
ACGGTCCTTGTGAAGGCTGG
TGGTATGCAGATGGGTTCATGC
a
Primer sequences are only shown for probes that have been used in Figure 1.
criteria used to define differences in gene expression levels
were a change of at least 1.8-fold and 3 identical difference
calls of increase or decrease out of 4 comparisons.
Immunohistochemistry
The SCA3 brain tissues were derived from 2 patients with
genetically confirmed diagnosis (1 female and 1 male; mean
age at death 59.5 6 1.5 years; disease duration 28 6 4 years;
postmortem delay 22 6 2 hours). Three individuals without a
history of neurological or inflammatory disease served as control (3 males; mean age at death 53 6 8.3 years; postmortem
delay 24.3 6 4.5 hours). The SCA3 patients came from 2 unrelated families from the Netherlands and all were diagnosed
and autopsied by one of the authors (ERB and RAIdV, respectively). Horizontal, serial, 6-mm-sections were cut from formalin-fixed and paraffin-embedded tissue blocks of pons and
cerebellum, including dentate nucleus. For immunohistochemistry, the 6-mm sections were processed and stained using the
peroxidase-DAB technique as previously described (25). The
antibodies and dilutions used in this study included mouse
monoclonal anti-human BDNF antibody (1:50) (R&D Systems,
Minneapolis, MN), mouse monoclonal anti-human CD9 antibody (1:50) (Santa Cruz Biotechnology, Santa Cruz, CA),
mouse monoclonal anti-human C/EBPß antibody (1:25) (Santa
Cruz Biotechnology), rabbit polyclonal anti-human C/EBPd antibody (1:500) (Santa Cruz Biotechnology), mouse monoclonal
anti-human Hsp27 (1:100) (Stressgen, Victoria, Canada), rabbit
polyclonal anti-human IL-6 antibody (1:125) (Santa Cruz Biotechnology), rabbit polyclonal anti-human IRF-1 antibody (1:
200) (Santa Cruz Biotechnology), rabbit polyclonal anti-bovine
LANP (1:150) (kindly provided by Dr. T. Isobe, Japan and Dr.
J Neuropathol Exp Neurol, Vol 62, October, 2003
O. Riess, Germany), rabbit polyclonal anti-human PA28ß antibody (1:100) (Biotrend, Köln, Germany), mouse monoclonal
anti-human PPARg (1:25) (Santa Cruz Biotechnology), rabbit
polyclonal anti-mouse PPARg (1:250, Dianova, Hamburg, Germany), rabbit polyclonal anti-human TIMP-1 (1:250) (Santa
Cruz Biotechnology), and mouse monoclonal anti-human Versican (1:200) (USBiological, Swampscott, MA).
For coimmunofluorescence studies of brain tissue, SCA3
pons sections were labeled with mouse monoclonal anti-ataxin3 antibody (1:1,500) (mab 1H9 kindly provided by Dr. Y. Trottier, France) together with rabbit polyclonal anti-human Hsp27
(1:100, Stressgen) and mouse monoclonal anti-ataxin-3 antibody (1:1,500) together with rabbit polyclonal anti-human
PA28ß antibody (1:100), followed by fluorescein (DATF) goat
anti-rabbit and Texas Red goat anti-mouse (Jackson
ImmunoResearch, West Grove, PA). For colocalization of ataxin-3 with LANP, double immunofluorescence was performed
using mouse monoclonal anti-ataxin-3 antibody (1:1,500) and
goat polyclonal anti-human LANP (1:400) (Santa Cruz Biotechnology). The sections were then exposed to Texas Red goat
anti-mouse (Jackson ImmunoResearch) and biotinylated rabbit
anti-goat immunoglobulin followed by fluorescein-avidin complex (Vector Laboratories, Burlingame, CA). Samples were observed with a Nikon Eclipse E800 fluorescence microscope (Nikon, Düsseldorf, Germany), digitized images were collected on
separate fluorescence channels using a Sony 3CCD digital camera and assembled with Adobe Photoshop. The colocalization
of ataxin-3 with Hsp27, PA28ß and LANP was confirmed by
confocal laser scanning microscopy (Leica TCS NT, Wetzlar,
Germany) and the dual channel imaging technique (krypton argon laser with emission of 488 nm for fluorescein and 568 nm
GENE PROFILING IN ATAXIN-3 EXPRESSING CELL LINES
for Texas Red), which permitted the simultaneous detection of
the 2 different dyes (data not shown). The number of ataxin-3-,
Hsp27-, PA28b-, and LANP-positive nuclear inclusions was determined on 2 pontine sections per case per antibody at the level
of the caudal locus ceruleus by a blinded observer (LO).
RESULTS
Analysis of Gene Expression in a SCA3 Cell Model
CSM14.1-inducible stable clonal cell lines (Tet-Off)
expressing human full-length ataxin-3 with 23 or 70 glutamines (called SCA3-Q23 and SCA3-Q70) were characterized previously. These cell lines provide a high level
expression of the recombinant ataxin-3 isoforms as well
as a time and polyglutamine length-dependent formation
of nuclear inclusions and induction of cell death (23).
Shifting cells from the permissive temperature (338C) to
the nonpermissive temperature (398C) results in differentiation to postmitotic neuronal cells after 7 days of culture. Using a PCR-based cDNA subtractive hybridization
strategy we identified upregulated inflammatory genes in
SCA3-Q70 cells, which were also increasingly expressed
in affected human SCA3 pontine neurons (20). To extend
the previous analysis we compared gene expression profiles by microarray analyses using the same CSM14.1
clonal cell lines (SCA3-Q23#2 and SCA3-Q70#1) and
the same cell culture conditions, that is, cells expressing
either normal or mutant ataxin-3 for 7 days at 338C. At
this time point and temperature, the transgenic cell lines
show normal viability and provide high levels of the recombinant ataxin-3 isoforms (23). For gene expression
analysis, duplicate RNA samples from each clonal cell
line were prepared and hybridized to Affymetrix U34A
oligonucleotide arrays that contain ;8,800 defined rat
genes and expressed sequence tags (ESTs). The hybridization results obtained from these duplicate experiments
allowed a 4-way comparison using the Affymetrix software to identify identically up- or downregulated genes;
threshold was set at p , 0.01. Forty-three genes showed
at least a 1.8-fold change in SCA3-Q70#1 compared to
SCA3-Q23#2 cells (Table 2).
Confirmation of Chip Data by Northern Analysis
and Immunohistochemistry
To verify the identified candidate genes, Northern blot
analyses were performed on total RNA from the original
and additional clonal cell lines expressing normal and
mutant ataxin-3 (SCA3-Q23#1, 2 and -Q70#1, 2) and 2
mock-transfected control cell lines (Ctrl#1, 2). Northern
blot confirmed differential expression in 23 of the 43
genes identified by the oligonucleotide array (Fig. 1),
while in the remaining altered gene expression was restricted to 1 clonal cell line (Table 2). The majority of
the identified upregulated or downregulated genes represent cell surface and extracellular matrix (ECM)-related
molecules, transcription factors, inflammatory cytokines,
1009
as well as proteins involved in the proteasomal pathway,
heat shock response, signal transduction, and neuronal
survival (Table 2). In agreement with our previous study
(20), the microarray data confirmed upregulation of
MMP-2 and Fit-1S, while APP with a 1.6-fold increase
did not meet the stringency criteria and SDF1a was not
represented on the U34A chip array. Confirmed differentially expressed genes were subsequently used to examine the expression of corresponding human proteins in
human postmortem tissue of SCA3 patients and healthy
controls to evaluate their possible implication in SCA3
pathogenesis. Suitable antibodies for immunostaining
studies in human brain sections were available for 14 of
the 23 differentially expressed corresponding rat genes.
Upregulated Genes in SCA3-Q70 Cell Lines and
SCA3 Brains
The group of upregulated genes that code for proteins
involved in inflammation (Fig. 1A) includes 3 interleukin
1-related cytokines: the Fos-inducible transcript (Fit-1S),
interleukin 1 receptor antagonist (IL-1ra), interleukin 6
(IL-6), and 2 cytokine-inducible transcription factors,
CCAAT/enhancer-binding proteins beta (C/EBPb) and
delta (C/EBPd). The upregulation of Fit-1S transcripts
confirmed our previous data (20) and was further extended by the finding that its related homolog IL-1ra is also
upregulated in SCA3-Q70 cell lines. In addition, we
found upregulation of mRNAs encoding the cell surface
tissue factor protein (TFP) triggering blood clotting in
response to inflammatory mediators (26) and a yet-unknown gene (EST190) exclusively in mutant ataxin-3 expressing cell lines.
Four differentially expressed genes could be further
assessed by immunohistochemical analysis in SCA3 pontine tissue sections. We previously showed increased cytoplasmic expression of IL-1ra and IL-1b in pontine neurons of SCA3 patients (20). The immunohistochemical
analysis of the newly identified inflammatory cytokine
IL-6 confirmed a significantly increased cytoplasmic
staining of IL-6 in pontine neurons of SCA3 patients
(Fig. 2B) compared to a much weaker cytoplasmic staining in control sections (Fig. 2A). In addition, immunostaining of C/EBPd revealed a strongly increased cytoplasmic expression in pontine neurons of SCA3 patients
(Fig. 2D) compared to a less intense staining in pontine
neurons of the respective control sections (Fig. 2C). In
contrast, immunostaining of C/EBPb did not show a significant difference between patient and control brains
(data not shown). Thus, mutant ataxin-3 mediates upregulation of several cytokines and cytokine-inducible transcription factors, confirming that SCA3 is associated with
inflammation.
J Neuropathol Exp Neurol, Vol 62, October, 2003
1010
EVERT ET AL
TABLE 2
Gene Expression Changes in SCA3-Q70 Cells Versus SCA3-Q23 Cells
Genbank
Description
Name
Changea,b
Northernc
Cell surface and ECM-related molecules
AI169327
EST215162 Homolog to tissue inhibitor of metalloproteinase 1
U65656
Matrix metalloproteinase type 2 or gelatinase A
X76489
Cell surface glycoprotein CD9
D88250
Serine protease (homolog to human complement component C1s)
U07619
Tissue factor protein
AF072892
Chondroitin sulfate proteoglycan versican V3 isoform precursor
X84039
Lumican
S77494
Lysyl oxidase
X65036
H36-alpha7 integrin alpha chain
U82612
Fibronectin
TIMP-1
MMP-2
CD9
C1s
TFP
V3
Lumican
Lox
H36-alpha7
Fn-1
1.9**
2.7**
2.7**
5.0**
5.4**
5.8**
2.1*
23.9**
24.7**
29.2**
C
C
C
C
C
C
R
R
R
R
Transcription factors and nuclear proteins
C/EBP-related transcription factor beta (C/EBPb or rNFIL-6)
S77528
C/EBP-related transcription factor delta (C/EBPd or CELF)
M65149
M34253
Interferon regulatory factor 1
D32209
Leucine-rich acidic nuclear protein
AB011365
PPAR-gamma protein
L03556
Hox1.3 protein
U17254
Immediate early gene transcription factor
cAMP response element binding protein 1
U38938
C/EBPb
C/EBPd
IRF-1
LANP
PPARg
Hox1.3
NGFI-B
CBP
2.4**
2.4**
24.4*
25.1**
25.5**
26.1*
1.8**
6.7*
C
C
C
C
C
R
R
R
Inflammatory cytokines
M63101
Interleukin 1 receptor antagonist
U04319
IL-1 receptor-related Fos-inducible transcript
Interleukin 6
M26744
IL-1ra
Fit-1S
IL-6
3.1*
3.4**
11.9**
C
C
C
Proteasome and heat shock molecules
M86389
Heat shock protein 27
D45250
Proteasome activator rPA28 subunit beta
M55534
Alpha-crystallin B chain
Hsp27
PA28b
aB crystallin
22.8**
23.6**
23.0**
C
C
R
Growth factors and signal transduction
Brain-derived neurotrophic factor
D10938
Protein tyrosine phosphatase
S74351
3-methylcholanthrene-inducible UDP-glucuronosyltransferase
S56937
cAMP phosphodiesterase
M25350
Phospholipase D
AB000778
Homocysteine respondent protein
AF036537
Insulin-like growth factor binding protein
M69055
BDNF
MKP-1
UGT1A6
PDE4
PLD
HCYP2
IGFBP6
24.5**
210.1**
4.3*
4.0*
3.7**
23.1*
24.4**
C
C
C
R
R
R
R
ESTs
AA800853
AA891527
AI102839
AA685152
AA891734
AA893022
EST190350
EST195330
EST212128
EST106386
EST195537
EST196825
EST190
EST195
EST212
EST106
EST1955
EST196
3.5**
24.1**
23.3**
212.9*
22.7**
22.4**
C
C
C
C
R
R
Others
S69874
AF014503
J04792
AF051561
D00688
AA899253
Cutaneous fatty acid-binding protein
p8
Ornithine decarboxylase
Na-K-Cl cotransporter
Monoamine oxidase A
Similar to mouse myristoylated alanine-rich C-kinase substrate
C-FABP
P8
ODC
Nkcc1
MAO-A
MARCKS
3.9**
22.8*
22.9**
23.5**
27.8**
222.2**
R
R
R
R
R
R
a
Mean fold changes in SCA3-Q70#1 compared to SCA3-Q23#2 as calculated from the array data by Affymetrix GeneChip
software 4.0.
b
P values were calculated from the array data with the 4 independent pairwise comparisons for each gene and are indicated
at the p , 0.01 (*) and p , 0.001 (**) levels.
c
Differential expression was verified by Northern analyses using independent clonal cell lines (see Materials and Methods and
Fig. 1); (R) restricted to one, or (C) confirmed for independent clonal cell lines.
J Neuropathol Exp Neurol, Vol 62, October, 2003
GENE PROFILING IN ATAXIN-3 EXPRESSING CELL LINES
1011
Fig. 1. Gene changes identified in the SCA3 cell model. Northern blot analyses of the identified differentially expressed
genes subdivided into groups of genes upregulated (A) or downregulated (B) in mutant ataxin-3 expressing cell lines (Q70-#1,
-#2) and genes upregulated (C) or downregulated (D) in normal ataxin-3 expressing cell lines (Q23-#1, -#2). The different groups
of genes were obtained through comparison to the corresponding mRNA levels in the mock-transfected control cell lines (Ctrl#1, -#2) regarded as the constitutive expression levels of the respective genes. Northern blot analyses were performed using total
RNA isolated from the different clonal cell lines after 7 days of induced expression and the indicated probes as described in
Materials and Methods. The mRNA sizes of the detected transcripts are indicated on the right side of each Northern blot image.
Equal loading of RNA was confirmed by methylene blue staining of 28S and 18S rRNA species; the respective amounts of 28S
rRNA are shown below each Northern blot.
Downregulated Genes in SCA3-Q70 Cell Lines and
SCA3 Brains
Two of the downregulated genes are known to be involved also in SCA1 (27) and HD (28). Figure 1B shows
a distinct downregulation of mRNAs encoding the leucine-rich acidic nuclear protein (LANP) and the 2 major
transcripts encoding brain-derived neurotrophic factor
(BDNF) in SCA3-Q70 clonal cell lines. Further, mRNAs
encoding the antiinflammatory peroxisome-proliferative
activated nuclear receptor gamma (PPARg) and 2 ESTs
(EST195 and EST106) were downregulated only in
SCA3-Q70 clonal cell lines.
The immunohistochemical analysis of the nuclear protein LANP revealed a significantly increased immunoreactivity and altered subcellular localization of LANP,
with the majority of pontine neurons showing strong nuclear but also occasional cytoplasmic staining in SCA3
tissue (Fig. 2F), while pontine neurons in controls
J Neuropathol Exp Neurol, Vol 62, October, 2003
1012
EVERT ET AL
Fig. 2. Immunohistochemical analyses of differentially expressed genes identified in mutant ataxin-3 expressing cells, in
human SCA3 disease, and control brain tissue. IL-6 immunostaining of control pons (A) compared to diseased pons (B) showed
several pontine neurons with increased cytoplasmic immunoreactivity (B). Immunostaining of C/EBPd in control (C) and disease
pons sections (D) revealed a significantly increased cytoplasmic staining of pontine neurons (D) in SCA3 patients. Comparison
of LANP-immunostained control (E) and disease pons sections (F) showed a significantly increased and altered subcellular staining
of pontine neurons (F, and inset) in SCA3 patients not present in controls. Coimmunofluorescence staining showed colocalization
of LANP to ataxin-3-positive NIs (G–I) in pontine neurons of SCA3 cases. Immunohistochemical analysis of BDNF in cerebellar
sections showed a strongly decreased staining of dentate neurons in SCA3 cases (K) compared to an intense cytoplasmic punctate
staining pattern of BDNF in dentate neurons of controls (J). Bars and inset bar 5 10 mm.
showed only moderate levels of cytoplasmic and nuclear
LANP immunoreactivity (Fig. 2E). Coimmunofluorescence staining of ataxin-3 and LANP in SCA3 pons sections showed that more than 50% of ataxin-3-positive
inclusions (n 5 72 6 6 per section) colocalized with nuclear aggregates of LANP (n 5 37 6 4 per section).
LANP mainly formed ring-like structures around NIs
(Fig. 2G–I). These findings indicate that altered proteinprotein interactions underlie the increased LANP immunoreactivity in SCA3 disease tissue, although LANP gene
transcription was found to be downregulated in SCA3Q70 clonal cell lines. Immunostaining of BDNF did not
reveal a decreased expression in pontine neurons (data
not shown) but a significantly decreased expression of
BDNF in neurons of the dentate nucleus of SCA3 patients (Fig. 2K) compared to an intense cytoplasmic
punctate staining pattern in controls (Fig. 2J). The antibodies tested for PPARg did not permit a specific immunohistochemical analysis in human brain tissue sections in our study (data not shown). Thus, mutant
J Neuropathol Exp Neurol, Vol 62, October, 2003
ataxin-3 is associated with an altered subcellular localization of LANP and a decreased expression of BDNF in
SCA3 disease tissue.
Upregulated Genes in SCA3-Q23 Cell Lines and
Expression in SCA3 Brains
In clonal cell lines expressing normal ataxin-3 we
found a significant upregulation of mRNAs encoding the
small heat shock protein 27 (Hsp27), the proteasomal
subunit PA28 beta (PA28b), the interferon regulatory
transcription factor 1 (IRF-1), the tyrosine phosphatase
(MKP-1), and 1 EST (EST212) (Fig. 1C), while the corresponding transcript levels were normal in SCA3-Q70
cell lines compared to control cell lines.
Immunostaining of Hsp27 revealed an increased cytoplasmic Hsp27 immunoreactivity and several Hsp27positive nuclear aggregates in pontine neurons of SCA3
brains (Fig. 3B, and inset) not present in pontine neurons
from control sections (Fig. 3A). Coimmunofluorescence
staining of ataxin-3 and Hsp27 in SCA3 pons sections
GENE PROFILING IN ATAXIN-3 EXPRESSING CELL LINES
1013
Fig. 3. Immunohistochemical analyses of differentially expressed genes identified in normal ataxin-3 expressing cells, in
human SCA3 disease, and control brain tissue. Immunostaining of Hsp27 showed several pontine neurons with increased cytoplasmic staining and nuclear aggregates (B, and inset) in SCA3 cases compared to a less intense staining of pontine neurons in
controls (A). Coimmunofluorescence staining revealed colocalization of Hsp27 to ataxin-3-positive NIs (C–E) in pontine neurons
of SCA3 patients. Immunostaining of PA28b in control (F) and disease pons sections (G) revealed an enhanced cytoplasmic
staining and nuclear aggregates in pontine neurons of SCA3 patients (G). Coimmunofluorescence staining confirmed colocalization
of PA28b to ataxin-3-positive NIs (H–J) in pontine neurons of SCA3 patients. IRF-1 immunostaining revealed strong nuclear
IRF-1 immunoreactivity and only weak cytoplasmic staining of pontine neurons in control pons (K) whereas SCA3 pontine
neurons lacked intranuclear staining and displayed varying intensities of cytoplasmic immunoreactivity instead (L). Furthermore,
increased numbers of IRF-1-positive microglial (arrowhead) and oligodendroglial cells (asterisk) were present in healthy human
pons sections (K). Comparison of TIMP-1 immunostained control (M) and disease pons sections (N) showed a significantly
increased and altered subcellular staining of pontine neurons (N, and inset) in SCA3 patients that was not present in controls.
Bars and inset bars 5 10 mm.
showed that 17% of the ataxin-3-positive NIs (n 5 41 6
2 per section) colocalized to Hsp27-positive nuclear aggregates (n 5 7 6 1 per section) (Fig. 3C–E) in pontine
neurons of SCA3 patients. Likewise, immunostaining of
PA28b revealed an increased cytoplasmic and occasionally perinucleolar staining pattern reminiscent of nuclear
inclusions in pontine neurons from SCA3 patients (Fig.
3G) compared to a less intense staining in control sections (Fig. 3F). Coimmunofluorescence staining of ataxin-3 and PA28b in SCA3 pons sections revealed that
more than 18% of ataxin-3-positive NIs (n 5 49 6 2 per
section) colocalized to PA28b-positive aggregates (n 5 9
6 1 per section) (Fig. 3H–J). These findings suggest that
2 components involved in protein degradation are increasingly associated with NIs, presumably reflecting an
ongoing cellular stress response to remove the abnormal
protein, although transcription levels of these genes in
SCA3-Q70 cells were more or less unaffected (Fig. 1C).
In agreement with the decreased IRF-1 mRNA levels in
SCA3-Q70 cell lines, the immunohistochemical analysis
revealed a reduced and altered subcellular expression of
the transcription factor IRF-1 in pontine neurons of SCA3
brains (Fig. 3L). Control sections consistently showed
strong nuclear IRF-1 immunoreactivity and only weak
cytoplasmic staining (Fig. 3K), while SCA3 pontine neurons lacked intranuclear staining but displayed cytoplasmic staining instead with varying intensities (Fig. 3L). In
addition, increased numbers of IRF-1-positive microglial
and oligodendroglial cells appear to be present in
healthy human pons sections while fewer IRF-1-positive
glial cells were found in SCA3 cases (Fig. 3K, L). Thus,
pontine neurons of SCA3 patients displayed an altered
J Neuropathol Exp Neurol, Vol 62, October, 2003
1014
EVERT ET AL
subcellular IRF-1 immunoreactivity and, in addition,
showed increased Hsp27 and PA28b immunoreactivity
presumably due to enhanced interaction with mutant
ataxin-3.
Downregulated Genes in SCA3-Q23 Cell Lines and
Expression in SCA3 Brains
The majority of the identified downregulated genes in
cell lines expressing normal ataxin-3 (Fig. 1D) represent
cell surface molecules or components of the ECM, including the cell surface glycoprotein CD9 (CD9), a serine
protease (rat homolog to human complement component
C1s), the chondroitin sulfate proteoglycan versican V3
isoform (V3), a homolog of the tissue inhibitor of metalloproteinase 1 (TIMP-1) and, as shown previously, the
matrix metalloproteinase type 2 (MMP-2). Compared to
the constitutive transcript levels in control cell lines both
TIMP-1 and MMP-2 were downregulated by normal and
upregulated by mutant ataxin-3 (Fig. 1D). In addition, we
found a significant downregulation of an inducible UDPglucuronosyltransferase (UGT1A6) mainly involved in
detoxification (29).
We have already shown that differential expression of
MMP-2 mRNA correlates with an increased expression
of catalytically active MMP-2 isoforms in SCA3-Q70
clonal cell lines and increased cytoplasmic MMP-2 levels
in diseased pontine neurons of SCA3 patients (20). Immunostaining of TIMP-1 showed a significantly increased
cytoplasmic punctate staining pattern in pontine neurons
of SCA3 patients (Fig. 3N) not present in the corresponding control sections (Fig. 3M). In contrast, immunohistochemical analyses of V3 and CD9 did not show a differential expression pattern of control and disease pons
tissue in agreement with comparable transcript levels of
V3 and CD9 present in control and SCA3-Q70 cell lines
(Fig. 1D). Thus, in addition to MMP-2, a specific tissue
inhibitor of MMPs, TIMP-1, was overexpressed in pontine neurons of SCA3 patients.
DISCUSSION
In the present study we compared gene expression profiles of transgenic cell lines expressing normal (SCA3Q23) or mutant full-length ataxin-3 (SCA3-Q70) by oligonucleotide array hybridization. The identified
differentially expressed transcripts were re-evaluated by
Northern analyses of the transgenic cell lines and additional corresponding clonal and mock-transfected control
cell lines. Twenty-three differentially expressed genes involved in inflammatory reactions, nuclear transcription,
and cell surface-associated processes were identified.
Comparison of the differentially expressed genes to the
corresponding constitutive transcript levels in control cell
lines allows us to dissect putative physiological transcriptional effects of normal ataxin-3 and pathogenic transcriptional alterations induced by mutant ataxin-3. Thus,
J Neuropathol Exp Neurol, Vol 62, October, 2003
overexpression of normal ataxin-3 resulted in downregulation of specific genes, while the same genes were
upregulated in mutant ataxin-3 expressing cells and the
corresponding proteins were increased in human SCA3
pontine neurons. The data presented here strongly suggest
that both loss of normal ataxin-3 transcriptional function
and gain of function through protein-protein interacting
properties of mutant ataxin-3 contribute to SCA3 pathogenesis. It is important to note that these distinct effects
are not mutually exclusive and may even occur simultaneously in SCA3. Reduced mobility of normal ataxin-3
in the nucleus through interaction with nuclear inclusions
formed by mutant ataxin-3 (12, 30) presumably leads to
altered transcriptional repressing capacity and subsequent
dysregulation of a specific subset of target genes. In addition, mutant ataxin-3 recruits important transcriptional
coactivators such as CBP and PML (9, 11, 12, 31), which
may result in aberrant transcriptional changes of another
subset of genes.
Transcriptional Alterations Induced by Mutant Ataxin-3
The genes identified herein support the hypothesis that
the pathogenic effect of mutant ataxin-3 is associated
with the induction of an inflammatory cascade, resulting
in enhanced transcription of 3 cytokines (Fit-1S, IL-1ra
and IL-6) and 2 cytokine-inducible transcription factors
(C/EBPb and C/EBBPd). IL-6 may play a particular important role. Transgenic mice overexpressing IL-6 develop a severe neurological phenotype and massive neurodegeneration (32), suggesting a pathogenic role of IL-6
in neurodegenerative diseases. In Alzheimer disease, IL6 has been detected in diffuse plaques, indicating local
activation of inflammatory mechanisms (33, 34). Inflammatory cytokines such as IL-1 and IL-6 induce the expression of the transcription factors C/EBPb and C/EBPd
(35, 36). Further, IL-6 may induce IL-1ra gene transcription as part of a negative feedback loop to terminate the
inflammatory response (37). Moreover, the release of IL6 may also be responsible for the upregulation of TFP
mRNA in the SCA3-Q70 clonal cell lines as recently
reported for activated endothelial cells (38).
Among the downregulated genes, LANP and BDNF
are already known to be involved in polyQ diseases. In
SCA1, mutant ataxin-1 was shown to interact with LANP
and to alter its subnuclear distribution (27). LANP is a
nuclear protein with a leucine-rich domain presumably
mediating protein-protein interactions and is mainly expressed in Purkinje cells, the primary site of SCA1 pathology (27, 39). Similarly, we found an altered subcellular localization of LANP and colocalization to NIs
formed by ataxin-3 in affected pontine neurons of SCA3
patients, suggesting that ataxin-3 may also interact with
LANP. The discrepancy between the downregulation of
LANP transcripts in SCA3-Q70 cell lines and the altered
GENE PROFILING IN ATAXIN-3 EXPRESSING CELL LINES
immunostaining pattern of LANP in SCA3 patients indicates dual effects of mutant ataxin-3. Recruitment of
transcription factors required for LANP gene transcription could be responsible for the distinct but not complete
downregulation of LANP mRNA observed in SCA3-Q70
cell lines (Fig. 1B), while aberrant or altered protein-protein interactions may have resulted in an increased recruitment of LANP protein into NIs formed by mutant
ataxin-3 (Fig. 2G–I).
BDNF functions as a survival factor for growth and
survivability of nerve and glial cells and is widely distributed throughout the central nervous system (40). The
downregulation of BDNF-specific transcripts in mutant
ataxin-3 expressing cell lines was paralleled by a reduced
immunostaining of BDNF protein in dentate but not in
pontine neurons of SCA3 patients. The dentate nucleus
as well as the pontine nuclei are both severely affected
in SCA3, and the majority of dentate neurons is lost during the course of the disease (2, 41). Also, SCA3 transgenic mice have a significant loss of dentate neurons (42).
Similarly, in HD, reduced BDNF mRNA and protein levels were found in the cortex of HD patients and in transgenic HD mice (28). Moreover, mutant huntingtin reduced while normal huntingtin enhanced transcription of
the BDNF gene in striatal cells (28). Similarly, reduced
levels of BDNF mRNA in mutant ataxin-3 expressing
cell lines support a repressive transcriptional effect mediated by the mutant protein. Interestingly, an important
component of BDNF gene expression (43–46), the transcriptional coactivator CBP, is recruited by mutant ataxin3 (11, 12) and other polyglutamine disease proteins (7–
10). Thus, a decrease of BDNF mRNA expression may
turn out to represent a common phenomenon in polyQ
disorders.
Transcriptional Alterations Induced by Normal Ataxin-3
In cells overexpressing normal ataxin-3 we found a
significant upregulation of the chaperone Hsp27 and proteasomal subunit PA28b transcripts. The small heat shock
protein Hsp27 acts as a molecular chaperone preventing
unfolded proteins from irreversible aggregation and facilitates refolding of unfolded proteins (47). PA28 is a
heteromeric protein complex of 2 alternating subunits,
PA28a and PA28b, and regulates the specificity of the
20S proteasome in protein degradation (48). However,
PA28 is also required for Hsp90-dependent protein refolding and thus acts as a molecular link between degradation and refolding of unfolded proteins (49). The immunohistochemical findings suggest that components
involved in protein degradation interact and associate
with mutant ataxin-3 in order to refold and/or degrade
the mutant protein. Increased transcription of these genes
in SCA3-Q23 cells on the other hand indicate that normal
ataxin-3 is associated with an increased expression of
these genes. In contrast, transcript levels of Hsp27 and
1015
PA28b in SCA3-Q70 cell lines were normal compared to
those in control cell lines (Fig. 1C), suggesting that mutant ataxin-3 does not affect the transcriptional regulation
of both genes. The increased immunostaining and colocalization of Hsp27 (Fig. 3B–E) and PA28b (Fig. 3G–J)
to NIs in pontine neurons of SCA3 patients may then
rather reflect an altered capability to execute an adequate
cellular stress response. Recently, Cowan et al (50)
showed that cells expressing an polyQ-expanded androgen receptor protein formed nuclear aggregates that preferentially sequestered the heat shock protein 72 (Hsp72).
Moreover, these cells exhibited a delayed expression of
Hsp72 after heat shock and failed to degrade Hsp72 during the recovery period. Thus, it is conceivable that
chronic exposure to an expanded polyQ protein contributes to disease progression by initiating cellular stress
responses that cannot be properly executed and/or attenuated. The involvement of both Hsp27 and PA28b in
SCA3 pathogenesis is in agreement with previous findings and supports the hypothesis that a stress response
via chaperones and proteasomal constituents is induced
in SCA3 pathogenesis (51, 52). Several other recent studies have shown that different members of heat shock proteins colocalize to intranuclear inclusions formed by
polyQ disease proteins and strategies increasing the expression and activation of these chaperones were protective against polyQ-mediated cytotoxicity (51, 53–61). A
putative regulatory effect of normal ataxin-3 on Hsp27
and PA28b may be lost when incorporated into nuclear
inclusions.
Most strikingly, genes encoding cell surface and ECMassociated constituents appear to be significantly repressed in response to normal ataxin-3. As recently
shown, an important role of endogenous ataxin-3 is associated with a transcriptional repressor activity by histone binding and inhibition of the intrinsic histone acetyltransferase activities of the transcriptional coactivators
CBP, p300, and PCAF (22). Therefore, it is likely that at
least a part of the identified repressed genes in SCA3Q23 clonal cell lines represent target genes specifically
regulated by ataxin-3. Interestingly, TIMP-1 and MMP2 were both downregulated by normal and upregulated
by mutant ataxin-3, possibly as a consequence of normal
ataxin-3 functions lost in the disease process. Accordingly, immunohistochemical analysis confirmed an increased
and altered immunostaining of TIMP-1 as well as an increased cytoplasmic MMP-2 immunoreactivity in affected
pontine neurons of SCA3 patients (20). MMP-2 belongs
to a family of Zn21-dependent matrix metalloproteinases
(MMPs) that are involved in degradation of ECM in physiological and pathological conditions (62). MMPs are subject to 3 levels of regulation, including transcriptional induction, proenzyme activation, and inhibition by TIMPs.
TIMPs are the major endogenous regulators of MMP activities in the tissue. TIMP-1 and TIMP-2 are capable of
J Neuropathol Exp Neurol, Vol 62, October, 2003
1016
EVERT ET AL
inhibiting the activities of all known MMPs and play a
key role in maintaining the balance between ECM deposition and degradation in different physiological processes (63). While the increased expression of TIMP-1
in SCA3-Q70 cell lines and SCA3 pons may reflect a
compensatory reaction to counteract the increased MMP2 expression, alternatively, a specific ability of normal
ataxin-3 to repress transcription of MMP-2 and/or TIMP1 might be lost in the disease state. For MMP-2 transcription, an enhancer region has been identified in the
human MMP-2 promoter that is subject to p53 and AP2 upregulation and E1A repression (64, 65). The E1A
repression seems to be mediated by interaction and inhibition of the transcriptional coactivator p300, a wellknown member of the CBP family (65). Thus, endogenous ataxin-3 may suppress MMP-2 transcription similar
to E1A through inhibition of p300 acetylation. A reduced
transcriptional repressor capacity through recruitment of
ataxin-3 and p300 into nuclear inclusions may then be
responsible for the increased MMP-2 expression in SCA3
brains.
We have identified differentially expressed genes and
proteins in SCA3 cell lines and SCA3 brains. The diversity of changes induced by mutant ataxin-3 does not allow identification of a single causative pathogenetic
mechanism. Rather, it appears that diverse processes involving transcriptional changes and compensatory reactions combine to constitute the overall pathogenetic
mechanism of SCA3. The genes identified in this study
suggest that mutant ataxin-3 is associated with 1) induction of a cascade of inflammatory mediators, 2) significant alteration of the nuclear environment, and 3) loss of
important neurotrophic support in disease-affected neurons. Further, normal ataxin-3 seems to be involved in
the modulation of a stress response and transcriptional
repression of cell surface- and ECM-associated components. In future studies, it will be important to differentiate between those genes transcriptionally regulated by
normal ataxin-3 and genes altered as a result of sequestered transcriptional factors and coactivators by mutant
ataxin-3. This approach may be helpful to elucidate the
pathogenic mechanisms of SCA3 and possible therapeutic strategies.
ACKNOWLEDGMENTS
We thank Dr. T. Isobe, Japan and Dr. O. Riess, Germany for kindly
providing the LANP antibody and Dr. Y. Trottier, France for the 1H9
antibody against ataxin-3. We are thankful to Dr. J. Reimann confirming
the colocalization studies by confocal laser scanning microscopy and
K. Nathani for statistical analyses of the microarray data.
REFERENCES
1. Kawaguchi Y, Okamoto T, Taniwaki M, et al. CAG expansions in
a novel gene for Machado-Joseph disease at chromosome 14q32.1.
Nat Genet 1994;8:221–28
J Neuropathol Exp Neurol, Vol 62, October, 2003
2. Schmidt T, Landwehrmeyer GB, Schmitt I, et al. An isoform of
ataxin-3 accumulates in the nucleus of neuronal cells in affected
brain regions of SCA3 patients. Brain Pathol 1998;8:669–79
3. Paulson HL. Toward an understanding of polyglutamine neurodegeneration. Brain Pathol 2000;10:293–99
4. Zoghbi HY, Orr HT. Glutamine repeats and neurodegeneration.
Annu Rev Neurosci 2000;23:217–47
5. Margolis RL, Ross CA. Expansion explosion: New clues to the
pathogenesis of repeat expansion neurodegenerative diseases.
Trends Mol Med 2001;7:479–82
6. Luthi-Carter R, Strand AD, Hanson SA, et al. Polyglutamine and
transcription: Gene expression changes shared by DRPLA and
Huntington’s disease mouse models reveal context-independent effects. Hum Mol Genet 2002;11:1927–37
7. Kazantsev A, Preisinger E, Dranovsky A, Goldgaber D, Housman
D. Insoluble detergent-resistant aggregates form between pathological and nonpathological lengths of polyglutamine in mammalian
cells. Proc Natl Acad Sci USA 1999;96:11404–9
8. Steffan JS, Kazantsev A, Spasic-Boskovic O, et al. The Huntington’s
disease protein interacts with p53 and CREB-binding protein and
represses transcription. Proc Natl Acad Sci USA 2000;97:6763–68
9. McCampbell A, Taylor JP, Taye AA, et al. CREB-binding protein
sequestration by expanded polyglutamine. Hum Mol Genet 2000;
9:2197–2202
10. Nucifora FC, Jr, Sasaki M, Peters MF, et al. Interference by huntingtin and atrophin-1 with cbp-mediated transcription leading to
cellular toxicity. Science 2001;291:2423–28
11. Chai Y, Wu L, Griffin JD, Paulson HL. The role of protein composition in specifying nuclear inclusion formation in polyglutamine
disease. J Biol Chem 2001;276:44889–97
12. Chai Y, Shao J, Miller VM, Williams A, Paulson HL. Live-cell
imaging reveals divergent intracellular dynamics of polyglutamine
disease proteins and supports a sequestration model of pathogenesis. Proc Natl Acad Sci USA 2002;99:9310–15
13. Shimohata T, Nakajima T, Yamada M, et al. Expanded polyglutamine stretches interact with TAFII130, interfering with CREB-dependent transcription. Nat Genet 2000;26:29–36
14. Luthi-Carter R, Strand A, Peters NL, et al. Decreased expression
of striatal signaling genes in a mouse model of Huntington’s disease. Hum Mol Genet 2000;9:1259–71
15. Wyttenbach A, Swartz J, Kita H, et al. Polyglutamine expansions
cause decreased CRE-mediated transcription and early gene expression changes prior to cell death in an inducible cell model of
Huntington’s disease. Hum Mol Genet 2001;10:1829–45
16. Sipione S, Rigamonti D, Valenza M, et al. Early transcriptional
profiles in huntingtin-inducible striatal cells by microarray analyses.
Hum Mol Genet 2002;11:1953–65
17. Lin X, Antalffy B, Kang D, Orr HT, Zoghbi HY. Polyglutamine
expansion down-regulates specific neuronal genes before pathologic
changes in SCA1. Nat Neurosci 2000;3:157–63
18. Yue S, Serra HG, Zoghbi HY, Orr HT. The spinocerebellar ataxia
type 1 protein, ataxin-1, has RNA-binding activity that is inversely
affected by the length of its polyglutamine tract. Hum Mol Genet
2001;10:25–30
19. Lieberman AP, Harmison G, Strand AD, Olson JM, Fischbeck KH.
Altered transcriptional regulation in cells expressing the expanded
polyglutamine androgen receptor. Hum Mol Genet 2002;11:1967–76
20. Evert BO, Vogt IR, Kindermann C, et al. Inflammatory genes are
upregulated in expanded ataxin-3-expressing cell lines and spinocerebellar ataxia type 3 brains. J Neurosci 2001;21:5389–96
21. Perez MK, Paulson HL, Pendse SJ, Saionz SJ, Bonini NM, Pittman
RN. Recruitment and the role of nuclear localization in polyglutamine- mediated aggregation. J Cell Biol 1998;143:1457–70
22. Li F, Macfarlan T, Pittman RN, Chakravarti D. Ataxin-3 is a histone-binding protein with two independent transcriptional corepressor activities. J Biol Chem 2002;277:45004–12
GENE PROFILING IN ATAXIN-3 EXPRESSING CELL LINES
23. Evert BO, Wullner U, Schulz JB, et al. High level expression of
expanded full-length ataxin-3 in vitro causes cell death and formation of intranuclear inclusions in neuronal cells. Hum Mol Genet
1999;8:1169–76
24. Lipshutz RJ, Fodor SP, Gingeras TR, Lockhart DJ. High density
synthetic oligonucleotide arrays. Nat Genet 1999;21:20–24
25. Wullner U, Kornhuber J, Weller M, et al. Cell death and apoptosis
regulating proteins in Parkinson’s disease—A cautionary note. Acta
Neuropathol 1999;97:408–12
26. Lorenzet R, Napoleone E, Celi A, Pellegrini G, Di Santo A. Cellcell interaction and tissue factor expression. Blood Coagul Fibrinolysis 1998;9:49–59
27. Matilla A, Koshy BT, Cummings CJ, Isobe T, Orr HT, Zoghbi HY.
The cerebellar leucine-rich acidic nuclear protein interacts with
ataxin-1. Nature 1997;389:974–78
28. Zuccato C, Ciammola A, Rigamonti D, et al. Loss of huntingtinmediated BDNF gene transcription in Huntington’s disease. Science
2001;293:493–98
29. Roy-Chowdhury J, Huang TJ, Kesari K, Lederstein M, Arias IM,
Roy-Chowdhury N. Molecular basis for the lack of bilirubin-specific and 3-methylcholanthrene-inducible UDP-glucuronosyltransferase activities in Gunn rats. The two isoforms are encoded by
distinct mRNA species that share an identical single base deletion.
J Biol Chem 1991;266:18294–98
30. Paulson HL, Perez MK, Trottier Y, et al. Intranuclear inclusions of
expanded polyglutamine protein in spinocerebellar ataxia type 3.
Neuron 1997;19:333–44
31. Yamada M, Sato T, Shimohata T, et al. Interaction between neuronal
intranuclear inclusions and promyelocytic leukemia protein nuclear
and coiled bodies in CAG repeat diseases. Am J Pathol 2001;159:
1785–95
32. Campbell IL, Abraham CR, Masliah E, et al. Neurologic disease
induced in transgenic mice by cerebral overexpression of interleukin 6. Proc Natl Acad Sci USA 1993;90:10061–65
33. Huell M, Strauss S, Volk B, Berger M, Bauer J. Interleukin-6 is
present in early stages of plaque formation and is restricted to the
brains of Alzheimer’s disease patients. Acta Neuropathol 1995;89:
544–51
34. Gadient RA, Patterson PH. Leukemia inhibitory factor, Interleukin
6, and other cytokines using the GP130 transducing receptor: Roles
in inflammation and injury. Stem Cells 1999;17:127–37
35. Tenen DG, Hromas R, Licht JD, Zhang DE. Transcription factors,
normal myeloid development, and leukemia. Blood 1997;90:489–519
36. Lekstrom-Himes J, Xanthopoulos KG. Biological role of the
CCAAT/enhancer-binding protein family of transcription factors. J
Biol Chem 1998;273:28545–48
37. Jordan M, Otterness IG, Ng R, Gessner A, Rollinghoff M, Beuscher
HU. Neutralization of endogenous IL-6 suppresses induction of IL1 receptor antagonist. J Immunol 1995;154:4081–90
38. Mesri M, Altieri DC. Leukocyte microparticles stimulate endothelial cell cytokine release and tissue factor induction in a JNK1
signaling pathway. J Biol Chem 1999;274:23111–18
39. Matsuoka K, Taoka M, Satozawa N, et al. A nuclear factor containing the leucine-rich repeats expressed in murine cerebellar neurons. Proc Natl Acad Sci USA 1994;91:9670–74
40. Murer MG, Yan Q, Raisman-Vozari R. Brain-derived neurotrophic
factor in the control human brain, and in Alzheimer’s disease and
Parkinson’s disease. Prog Neurobiol 2001;63:71–124
41. Takiyama Y, Oyanagi S, Kawashima S, et al. A clinical and pathologic study of a large Japanese family with Machado-Joseph disease tightly linked to the DNA markers on chromosome 14q. Neurology 1994;44:1302–8
42. Cemal CK, Carroll CJ, Lawrence L, et al. YAC transgenic mice
carrying pathological alleles of the MJD1 locus exhibit a mild and
slowly progressive cerebellar deficit. Hum Mol Genet 2002;11:
1075–94
1017
43. West AE, Chen WG, Dalva MB, et al. Calcium regulation of neuronal gene expression. Proc Natl Acad Sci USA 2001;98:11024–31
44. Shieh PB, Ghosh A. Molecular mechanisms underlying activitydependent regulation of BDNF expression. J Neurobiol 1999;41:
127–34
45. Shieh PB, Hu SC, Bobb K, Timmusk T, Ghosh A. Identification of
a signaling pathway involved in calcium regulation of BDNF expression. Neuron 1998;20:727–40
46. Tao X, Finkbeiner S, Arnold DB, Shaywitz AJ, Greenberg ME.
Ca21 influx regulates BDNF transcription by a CREB family transcription factor-dependent mechanism. Neuron 1998;20:709–26
47. Rogalla T, Ehrnsperger M, Preville X, et al. Regulation of Hsp27
oligomerization, chaperone function, and protective activity against
oxidative stress/tumor necrosis factor alpha by phosphorylation. J
Biol Chem 1999;274:18947–56
48. Song X, Mott JD, von Kampen J, et al. A model for the quaternary
structure of the proteasome activator PA28. J Biol Chem 1996;271:
26410–17
49. Minami Y, Kawasaki H, Minami M, Tanahashi N, Tanaka K,
Yahara I. A critical role for the proteasome activator PA28 in
the Hsp90-dependent protein refolding. J Biol Chem 2000;275:
9055–61
50. Cowan KJ, Diamond MI, Welch WJ. Polyglutamine protein aggregation and toxicity are linked to the cellular stress response. Hum
Mol Genet 2003;12:1377–91
51. Chai Y, Koppenhafer SL, Bonini NM, Paulson HL. Analysis of the
role of heat shock protein (Hsp) molecular chaperones in polyglutamine disease. J Neurosci 1999;19:10338–47
52. Schmidt T, Lindenberg KS, Krebs A, et al. Protein surveillance
machinery in brains with spinocerebellar ataxia type 3: Redistribution and differential recruitment of 26S proteasome subunits and
chaperones to neuronal intranuclear inclusions. Ann Neurol 2002;
51:302–10
53. Cummings CJ, Mancini MA, Antalffy B, DeFranco DB, Orr HT,
Zoghbi HY. Chaperone suppression of aggregation and altered subcellular proteasome localization imply protein misfolding in SCA1.
Nat Genet 1998;19:148–54
54. Cummings CJ, Sun Y, Opal P, et al. Over-expression of inducible
HSP70 chaperone suppresses neuropathology and improves motor
function in SCA1 mice. Hum Mol Genet 2001;10:1511–18
55. Stenoien DL, Cummings CJ, Adams HP, et al. Polyglutamine-expanded androgen receptors form aggregates that sequester heat
shock proteins, proteasome components and SRC-1, and are suppressed by the HDJ-2 chaperone. Hum Mol Genet 1999;8:731–41
56. Jana NR, Tanaka M, Wang G, Nukina N. Polyglutamine lengthdependent interaction of Hsp40 and Hsp70 family chaperones with
truncated N-terminal huntingtin: Their role in suppression of aggregation and cellular toxicity. Hum Mol Genet 2000;9:2009–18
57. Kobayashi Y, Kume A, Li M, et al. Chaperones Hsp70 and Hsp40
suppress aggregate formation and apoptosis in cultured neuronal
cells expressing truncated androgen receptor protein with expanded
polyglutamine tract. J Biol Chem 2000;275:8772–78
58. Zhou H, Li SH, Li XJ. Chaperone suppression of cellular toxicity
of huntingtin is independent of polyglutamine aggregation. J Biol
Chem 2001;276:48417–24
59. Warrick JM, Chan HY, Gray-Board GL, Chai Y, Paulson HL, Bonini NM. Suppression of polyglutamine-mediated neurodegeneration
in Drosophila by the molecular chaperone HSP70. Nat Genet 1999;
23:425–28
60. Wyttenbach A, Sauvageot O, Carmichael J, Diaz-Latoud C, Arrigo
AP, Rubinsztein DC. Heat shock protein 27 prevents cellular polyglutamine toxicity and suppresses the increase of reactive oxygen
species caused by huntingtin. Hum Mol Genet 2002;11:1137–51
61. Wyttenbach A, Carmichael J, Swartz J, et al. Effects of heat shock,
heat shock protein 40 (HDJ-2), and proteasome inhibition on protein aggregation in cellular models of Huntington’s disease. Proc
Natl Acad Sci USA 2000;97:2898–2903
J Neuropathol Exp Neurol, Vol 62, October, 2003
1018
EVERT ET AL
62. Woessner JF, Jr. Matrix metalloproteinases and their inhibitors in
connective tissue remodeling. FASEB J 1991;5:2145–54
63. Nagase H, Woessner JF, Jr. Matrix metalloproteinases. J Biol Chem
1999;274:21491–94
64. Frisch SM, Morisaki JH. Positive and negative transcriptional elements of the human type IV collagenase gene. Mol Cell Biol 1990;
10:6524–32
J Neuropathol Exp Neurol, Vol 62, October, 2003
65. Somasundaram K, Jayaraman G, Williams T, Moran E, Frisch S,
Thimmapaya B. Repression of a matrix metalloprotease gene by
E1A correlates with its ability to bind to cell type-specific transcription factor AP-2. Proc Natl Acad Sci USA 1996;93:3088–93
Received April 18, 2003
Revision received June 23, 2003
Accepted June 25, 2003