Early Alterations in Gene Expression and Cell Morphology in a Mouse Model of Huntington's Disease

Early Alterations in Gene Expression and Cell Morphology in aMouse Model of Huntington’s Disease

C. Iannicola, *S. Moreno, S. Oliverio, *†R. Nardacci, *A. Ciofi-Luzzatto, and †‡M. Piacentini

Department of Biology, University of Rome “Tor Vergata”;*Department of Biology, University of Roma Tre;†Laboratory ofElectron Microscopy and Cell Biology, IRCCS “L. Spallanzani,” Rome; and‡Department of Environmental Sciences,

“La Tuscia” University, Viterbo, Italy

Abstract: Several mouse models for Huntington’s disease(HD) have been produced to date. Based on differences instrain, promoter, construct, and number of glutamines,these models have provided a broad spectrum of neurolog-ical symptoms, ranging from simple increases in aggres-siveness with no signs of neuropathology, to tremors andseizures in absence of degeneration, to neurological symp-toms in the presence of gliosis and TUNEL (terminal deoxy-nucleotidyl transferase-mediated dUTP nick end-labeling)positivity, and finally to selective striatal damage associatedwith electrophysiological and behavioral abnormalities. Wedecided to analyze the morphology of striatum and hip-pocampus from a mouse transgenic line obtained by mi-croinjection of exon 1 from the HD gene after introduction ofa very high number of CAG repeat units. We found a mas-sive darkening and compacting of striatal and hippocampalneurons in affected mice, associated with a lower degree ofmore classical apoptotic cell condensation. We then ex-plored whether this morphology could be explained withalterations in gene expression by hybridizing normal andaffected total brain RNA to a panel of 588 known mousecDNAs. We show that some genes are significantly andconsistently up-regulated and that others are down-regu-lated in the affected brains. Here we discuss the possiblesignificance of these alterations in neuronal morphologyand gene expression. Key Words: Huntington’s disease—Gene expression—Apoptosis—Neuronal inclusions—Neurodegeneration.J. Neurochem. 75, 830–839 (2000).

Huntington’s disease (HD) is one of a growing num-ber of neurodegenerative diseases caused by the expan-sion of a trinucleotide repeat. This results in an exceed-ingly high number of contiguous glutamine residues inthe translated protein, which gains a new function thatproves toxic for specific sets of neurons. HD patientssuffer from progressive dementia, choreic movements,rigidity, and, in juvenile cases, associated epileptic sei-zures. In postmortem brains from HD patients extensiveneurodegeneration of striatum and cortex is usually ob-served.

Several animal models have been produced to date toreproduce and dissect the pathways that lead to the

profound alterations in cell viability and function ob-served in HD patients (Mangiarini et al., 1996; Reddyet al., 1998; Hodgson et al., 1999; Schilling et al., 1999;Shelbourne et al., 1999). It is surprising that the spectrumof biological phenomena has proven to be very broaddepending on the promoter used, the amount of codingsequence included in the transgene, or the mouse strainused. In all animals that express mutated huntingtinproduced so far, an overt neurological phenotype is ob-served irrespective of the occurrence of cell death in thebrain, with one exception (Shelbourne et al., 1999). Acommon feature among the affected mice is the accumu-lation of ubiquitinated protein aggregates, invariably nu-clear but in some cases also cytoplasmic, in the animalsthat show neurological symptoms.

The supposed pathogenic role of the intranuclear in-clusions has been challenged by a series of recent find-ings. Klement et al. (1998) have shown that mutatedataxin 1, which causes spinocerebellar ataxia 1, remainspathogenic in transgenic mice expressing the proteindeprived of its self-assembly sequence and thereforeunable to form nuclear aggregates. When a nuclear lo-calization signal was mutated in the protein sequence, themice did not develop any cellular or neurological pathol-ogy, suggesting that the nuclear localization of the pro-tein is necessary and sufficient for the disease to occur.More recently, Hodgson et al. (1999) have produced amouse model for HD in which the mutated gene isintegrated in the host genome as a yeast artificial chro-mosome. In this model specific striatal degeneration is

Received November 12, 1999; revised manuscript received April 4,2000; accepted April 14, 2000.

Address correspondence and reprint requests to Dr. C. Iannicola atDepartment of Biology, Lab 358, University of Rome “Tor Vergata,”Via della Ricerca Scientifica, 00133, Roma, Italy. E-mail:[email protected]

Drs. A. Ciofi-Luzzatto and M. Piacentini contributed equally to thiswork.

Abbreviations used:HD, Huntington’s disease; LR White, LondonResin White; NII, neuronal intranuclear inclusion; PKC, protein kinaseC; SSC, saline–sodium citrate; Topo2, topoisomerase II; TTF, tran-scription termination factor.

830

Journal of NeurochemistryLippincott Williams & Wilkins, Inc., Philadelphia© 2000 International Society for Neurochemistry

observed, and the neurological symptoms do not corre-late with the presence of nuclear or cytoplasmic proteinaggregates. These experiments are suggestive of a mech-anism whereby biochemical changes in the affected neu-rons before the deposition of protein aggregates, or evenin the absence of such aggregation, must underlie theneurological dysfunction common to all these models.

We decided to define better the morphology and tostudy early alterations in gene expression in the first HDmouse model produced (Mangiarini et al., 1996). Thesemice show no apparent cell death in the presence of adramatic neurological phenotype, suggesting that bio-chemical alterations must underlie cell dysfunction andonset of symptoms. In this model, exon 1 of the mutatedHD gene is expressed ubiquitously, and neuronal aggre-gates are present in the majority of neurons in the af-fected brain. After describing cellular morphology, welooked for differences in gene expression by hybridizingRNA extracted from total brain to filters containing 588known mouse cDNAs.

MATERIALS AND METHODS

AnimalsAll animals were housed and handled according to guide-

lines proposed by the Society for Neuroscience and the ItalianNational Research Council. Mice were deeply anesthetizedwith tribromoethanol (Avertin; 0.2 ml of 1.25% solution/10 gof body weight, injected intraperitoneally) before being rapidlykilled by transcardial perfusion with the fixative solution. TheHD TgN6/1 line was used (Mangiarini et al., 1996).

Light and electron microscopyThree transgenic mice and three normal littermates of ages

varying between 5 and 7 months were used in these experi-ments. Mice were perfused at room temperature with phos-phate-buffered saline (pH 7.3), followed by 4% freshly depo-lymerized paraformaldehyde in 0.1M phosphate buffer, pH7.3. The brains were removed 1 h after perfusion. Sagittal brainsections were cut on a Vibratome at 40mm and collected inphosphate-buffered saline, pH 7.3. Slices were postfixed in 1%OsO4 in 0.1 M phosphate buffer, rinsed in water, dehydrated,and flat-embedded in Epon. Selected areas were then re-mounted on Epon blanks and sectioned on a Reichert UltracutS ultramicrotome. Semithin sections were stained with tolu-idine blue and photographed using an Axioplan 2 Zeiss lightmicroscope. Ultrathin sections were briefly contrasted withuranyl acetate and lead citrate and photographed in a PhilipsCM120 electron microscope. For London Resin White (LRWhite; Agar Scientific Ltd., Stanted, U.K.) embedding, corpusstriatum and hippocampus were isolated from brains and cut insmaller pieces. Samples were rinsed in buffer, partially dehy-drated (up to 95% alcohol), and embedded in LR White. Tissueblocks were cut on a Reichert Ultracut S ultramicrotome, andultrathin sections were contrasted with uranyl acetate and leadcitrate and photographed in a Philips CM120 electron micro-scope.

Statistical analysisSemithin sections from corpus striatum and hippocampus of

transgenic and normal mice were examined using an Axioplan2 Zeiss light microscope, at low to high magnifications. Dif-ferent embedding blocks were analyzed, and for each block a

minimum of 100 cells was observed. The percentage of abnor-mal cells versus the total cell number was evaluated consider-ing as altered those neurons showing high condensation andstrong nuclear basophilia. Cells showing mild shrinkage andnuclear staining were classified separately. Moreover, we eval-uated the percentage of cells containing intranuclear inclusions,and taking into account that these bodies are similar in size withnucleoli and that each cell contains two nucleoli, the number ofcells showing intranuclear inclusion was compared with thecells showing nucleolus.

In situ hybridizationThe PCR fragments used in the northern blot experiments

were cloned using the Perfectly Blunt cloning kit (Clontech).Clones were sequenced. At least one clone in either orientationwas obtained and linearized before RNA labeling. A digoxige-nin-labeled RNA was obtained using a DIG RNA labeling kit(Boehringer-Mannheim from Roche Diagnostic S.p.A.), andthe T7 RNA polymerase labeling reaction was performed for2 h at 37°C. Two transgenic mice, both 7 months old, were usedin two independent experiments. Animals were deeply anes-thetized with Avertin (0.2 ml of 1.25% solution/10 g of bodyweight, injected intraperitoneally) before being rapidly killedby cervical dislocation. Brains were quickly removed, sagittallycut in two parts, and frozen by immersion in isopentane at220°C. Specimens were cut on a cryostat, and 15-mm sectionswere collected on gelatin-coated slides and frozen at220°C.Slides were immersed in 4% formaldehyde in phosphate-buff-ered saline for 15 min at 4°C. Prehybridization and hybridiza-tion were performed according to the Roche Molecular Bio-chemicals manual (Nonradioactive In Situ Hybridization Ap-plication Manual, 1996) using the DIG Nucleic Acid DetectionKit (Boehringer-Mannheim from Roche Diagnostic S.p.A.).Slides were incubated overnight at 42°C with 50ml of hybrid-ization buffer containing 20 ng of digoxigenin-labeled RNAprobe. As control we used the sense labeled probe.

cDNA array screeningTotal RNA was extracted from whole brain of three normal

and three transgenic mice by Trizol reagent (LIFE Technolo-gies, U.K.). All mice were 7 months old. Poly(A)1 was isolatedfrom ;400 mg of total RNA by using the Oligotex mRNAsystem (Qiagen, Germany).32P-labeled cDNA was preparedfrom 1 mg of poly(A)1; cDNAs were hybridized to the AtlasMouse cDNA Expression Array (catalogue no. 7741-1; Clon-tech Laboratories, Palo Alto, CA, U.S.A.) according to the usermanual (catalogue no. PT3140; Clontech), and expression wasvisualized by autoradiography. Filters were also exposed on aphosphor screen and analyzed with a Storm 840 instrument(Molecular Dynamics), which produces digital images of la-beled samples. Spot intensity was quantified by Image QuaNTsoftware (Molecular Dynamics), which provides a volume re-port by integrating the area of the spot with its intensity profile.The values were normalized with the average intensities ofhousekeeping genes between the two filters.

Northern blotNorthern blot analysis was performed by standard methods.

Equal amounts (10mg) of total RNA from three transgenic andthree control brains (mice 7 months old) were loaded on a 1%agarose gel containing 2.2M formaldehyde. DNA fragmentsfor three of the up-regulated genes were obtained by RT-PCR.DNA was radioactively labeled using the Megaprime kit(Amersham-Pharmacia, Rietech, U.K.). Each filter was thenstripped and reprobed with ab-actin probe to ensure that equal

J. Neurochem., Vol. 75, No. 2, 2000

831ALTERED MORPHOLOGY AND GENE EXPRESSION IN HD MICE

RNA amounts were loaded. Prehybridization and hybridizationwere performed using ExpressHyb solution (Clontech) at 68°C.Filters were washed several times at room temperature in 23saline–sodium citrate (SSC) and 0.5% sodium dodecyl sulfateand three times for 20 min each at 68°C in 0.1% SSC and0.01% sodium dodecyl sulfate as suggested by the manufac-turer.

RESULTS

Light microscopyMorphological analysis of semithin brain sections

from HD transgenic mice revealed signs of mild to

severe degeneration in the hippocampus and the corpusstriatum such as neuronal condensation and/or shrinkage.

In the hippocampal formation of transgenic mice (Fig.1a), pyramidal cells generally appear less closely juxta-posed than in the corresponding control (Fig. 1b). More-over, whereas in controls pyramidal neurons show clear,vesicular nuclei, in transgenic animals many neurons arecompacted, with darkly stained nuclei. The corpus stri-atum of transgenic mice (Fig. 2a) displays features sim-ilar to those described for the hippocampus. However,the neurodegenerative phenomena observed in this areaare more prominent than in any other part of the brainsanalyzed.

FIG. 1. Light micrographs of (a) normal and (b) transgenicmouse hippocampal CA3 field. Semithin sections were LRWhite-embedded and toluidine blue-stained. a: Large pyramidalneurons appear closely juxtaposed to each other and uniformlyoriented. Their nuclei are constantly euchromatic, with one ortwo nucleoli. b: Some relatively healthy pyramidal neurons showa clearly delimited nuclear compartment and evident nucleoli.Other neurons (arrowheads) appear highly condensed with astrongly basophilic nucleus.

FIG. 2. Light micrographs of (a) normal and (b) transgenicmouse striatum. Semithin sections were LR White-embeddedand toluidine blue-stained. a: Neuronal cells display basophiliccytoplasm and homogeneously euchromatic nucleus, with oneor more nucleoli. A few dark glial cells (arrows) are observed. b:Many neurons showing various degeneration degrees are rec-ognized (arrowheads). Darkly stained glial cells (arrows) are alsodetected. B, nerve bundles.


832 C. IANNICOLA ET AL.

Electron microscopyElectron microscopic examination of transgenic

mouse hippocampus revealed the presence of pyramidalneurons with different ultrastructures (Fig. 3). Numerousneurons show normal morphological features, such asregular plasma membrane profile, well-preserved or-ganelles, and large, euchromatic nucleus (Fig. 3a). Someneurons show clear degenerative signs, including shrink-age, plasma membrane indentations, vacuolation, andnuclear condensation (Fig. 3c). Specifically, chromatin isgenerally condensed in a scattered manner, thus notforming large heterochromatic masses (Fig. 3b and c).Few neurons appear as strongly electron-dense in theircytoplasm and nucleus, with the latter often hardly rec-ognizable (Fig. 3c). Many of the affected cells also showshrunken neurites (Fig. 3c). Several neurons show mor-phological features intermediate between the normal andthe degenerative phenotype. In these cells, mild nuclearcondensation, occasional vacuolation, but recognizablemembrane-limited compartments are seen (Fig. 3b).

Even in the corpus striatum, neurons show ultrastruc-tural features ranging from normal to profoundly alteredmorphology (Fig. 4). However, in this area, neurodegen-eration seems especially severe. Chromatin is often com-pacted to form a heterochromatic mass near the nucleo-lus (Fig. 4b and c).

Irrespective of their location and morphological ap-pearance, neurons often contain inside their nucleus afinely granular area, less electron dense than the nucle-oplasm, which is approximately of the same size as thenucleolus (Figs. 3 and 4). This aggregate closely resem-bles the typical neuronal intranuclear inclusion (NII)previously described (Davies et al., 1997).

Statistical analysisThe data represent mean6 SD values of three ani-

mals. In affected striata, the percentage of highly con-densed cells averaged 516 1.5%, whereas cells showinginitial apoptotic signs represented;40 6 2.2%. Bycontrast, normal striata contained no clearly apoptoticneurons and few, mildly altered cells (56 0.3%). Intransgenic hippocampus, dramatically degenerated cellsaveraged 126 0.7%, whereas mildly altered neuronswere;326 3.9%. In controls, as few as 26 0.2% of thetotal cell number showed some degeneration signs,whereas no apoptotic cells were observed. The percent-age of altered neurons is significantly (p , 0.01) higherin transgenic mice than in controls.

To evaluate the presence of NII in neurons of trans-genic mouse hippocampus and striatum, we comparedthe occurrence of this aggregate with that of nucleolus asthe two structures are similar in size. By considering that(a) most neurons contain two nucleoli, (b) only in 40–70% of cells the nucleolus is visible, and (c) the meanpercentage of cells showing the inclusion is 7%, wehypothesize that;25% of neurons contain an NII at thisstage.

In situ hybridizationIn situ hybridization performed on transgenic mouse

brain sections shows clear neuronal expression for pro-thymosina.

Signal is localized on pyramidal cells and on scatteredneurons in the hippocampus (Fig. 5a), on the medium-sized neurons of the corpus striatum (Fig. 6a), and onnumerous neurons of the cortical region. Control hybrid-ization, obtained by use of the “sense” probe, shows nosignal (Figs. 5b and 6b).

cDNA array screeningResults from hybridization of cDNA arrays with

cDNA probes prepared from poly(A)1 RNA isolatedfrom transgenic or normal mouse brains in three differentexperiments are averaged and summarized in Table 1.Figure 7 shows the hybridization pattern obtained in oneof these experiments. The degree of alteration in expres-sion levels of the genes listed in Table 1, as measured byphosphorimager quantification after normalization, washighly reproducible in all three experiments. Any genewhose signal did not show a similar variation in intensityin each of the three experiments was excluded. Thisapplies to only a few cDNAs as those listed in Table 1represent the vast majority of the genes that showed anysignal. The possibility exists that more genes could bedifferentially expressed, but our experiment did not showthese because their expression levels were below thepower of detection of this technique.

Northern blotThe results of our northern blot analysis are shown in

Fig. 8. Among the cDNAs that showed up-regulation inthe cDNA screening we selected three genes of particularinterest. The degree of up-regulation of these three genesin the array screening experiment varied from twofold(neuroleukin) to;10-fold (type Ib regulatory subunitcyclic AMP-dependent protein kinase and prothymosina). For all three genes analyzed the up-regulation wasconfirmed.

DISCUSSION

The corpus striatum and hippocampus of this trans-genic mouse model showed evident neurodegenerativefeatures, including nuclear and cytoplasmic abnormali-ties. Neuronal morphology ranged from apparently nor-mal, to mildly altered, to severely compromised. Al-though cytoplasmic and nuclear condensations areclosely reminiscent of apoptotic morphology, the ob-served phenomena cannot clearly be identified as classi-cal apoptosis. In fact, certain structural hallmarks of thiscell death process, e.g., cytoplasmic fragmentation andchromatin clumping, are lacking in this model. On theother hand, it should be pointed out that neurons arecharacterized by an almost totally euchromatic nucleus,in the absence of a proper nuclear lamina. In this model,therefore, chromatin remodeling, although involving for-mation of small heterochromatic clumps and occasion-ally larger heterochromatic masses, does not lead to



dramatic chromatin condensation. The large number ofaltered cells in the absence of cell fragments or phago-cytic phenomena raises more doubts about identificationof the observed phenotype as classical apoptosis. Ourobservations are rather suggestive of a slow or delayedneuronal death, in which degenerating events may beopposed by some unknown cell death-resistance features.A process of cell death distinct from classical apoptosishas been previously suggested in a model involvingtransfection of primary neurons with chimeric poly-glutamine-green fluorescent protein (Moulder et al.,1999).

In the attempt to clarify the molecular mechanismsunderlying this morphology, we studied the gene expres-sion patterns of affected versus normal total brains. Thegroup of genes that showed altered levels of expressionin the affected brains is somewhat heterogeneous, andthe function of the products of some of the altered genesis not fully known. It is intriguing to note that we did notdetect any variation in the expression levels of genes thatare directly related to apoptosis, i.e., caspases, tumornecrosis factor-like family of receptors,bcl-2 family,etc., although some of the genes that are up-regulated inthe affected brains have been shown to be involved inapoptosis or related processes, such as cell proliferation,cell survival, and chromatin structure, or in the activityof the ubiquitin/proteasome pathway.

Among the genes whose level of expression showed amore significant increase in affected mice, prothymosina has been shown to localize in the nucleus associatedwith condensed chromatin, and it can be phosphorylatedwhen the cells are stimulated to proliferate. Prothymosina has chromatin-remodeling properties (Gomez-Mar-quez and Rodriguez, 1998) and functions via a modula-tion of the interaction between chromatin and histone H1(Karetsou et al., 1998) in inducing nucleosome assembly(Dıaz-Jullien et al., 1996). More recently it has beenshown that prothymosina antisense oligonucleotidesinduce high levels of apoptotic death in HL60 cells(Rodrıguez et al., 1999).

The MHR23 gene, homologous to the yeast geneRad23, is among the ones that show up-regulation inaffected brains. Rad23 is a conserved protein involved innucleotide excision repair, but it may have also a regu-latory role. It has been shown to interact with the 26Sproteasome through a ubiquitin-like domain present at itsamino terminus (Schauber et al., 1998). Up-regulation ofMHR23could therefore be involved in the attempt madeby the cell at degrading the toxic protein coded for by thetransgene.

We found up-regulation of a regulatory subunit (type-I-b) of a cyclic AMP-dependent kinase. These proteinFIG. 3. Electron micrographs of hippocampal pyramidal neu-

rons from transgenic mice. a and c: Paraformaldehyde-osmiumfixation and Epon embedding. b: Paraformaldehyde fixation andLR White embedding. a: Neuron with normal morphology. In thenucleus (N), note the highly decondensed chromatin, the nucle-olus (nu), and the typical inclusion (i). b: Neuron with mild mor-phological alterations, with scattered, condensed chromatin andintranuclear inclusion (i). N, nucleus. c: Strongly altered neuron,with highly condensed cytoplasm and nucleus (N), indented

plasma membrane, and vacuolated cytoplasm (open arrows).The chromatin appears condensed in a scattered manner anddoes not form a heterochromatic mass. A dystrophic neurite isobserved (heavy arrow). Around the affected cell two healthy cellnuclei (N) are visible.



kinases are involved in a large number of regulatorypathways within the cell, affecting processes that rangefrom synaptic plasticity to behavior (Brandon et al.,1997). The regulatory subunits of the cyclic AMP-de-pendent protein kinases are known to regulate the sub-cellular localization and the activity of the catalytic sub-units (Mellon et al., 1989; Wiley et al., 1999). Micelacking the type-I-b subunit, up-regulated in our exper-iment, show a severe deficit in hippocampal long-termdepression, suggesting that an attempt at compensating

FIG. 4. Electron micrographs of striatal neurons from transgenicmice. a and c: Paraformaldehyde-osmium fixation and Eponembedding. b: Paraformaldehyde fixation and LR White embed-ding. a: Neuron with normal morphology and an intranuclearinclusion (i). N, nucleus. b: Neuron with altered ultrastructuralfeatures, suggesting an apoptotic initial stage. In the nucleus (N),note the diffusely condensed chromatin, with a heterochromaticmass adjacent to the nucleolus (nu), and the intranuclear inclu-sion aggregate, which appears more evident than in normal

FIG. 5. In situ hybridization of transgenic mouse hippocampus.a: Antisense prothymosin a probe revealed by alkaline phospha-tase reaction. Pyramidal cells (Pyr) and scattered interneurons(arrows) show intense labeling. b: Sense prothymosin a-labeledprobe. Labeling is absent from neurons or other cell types. Bars5 200 mm.

neurons owing to the higher chromatin electron density. c: Neu-ron with strongly altered morphology. The strong cytoplasmicand nuclear (N) condensation almost prevents identification ofthe two compartments. The intranuclear inclusion (i) is seenclose to the nucleolus and to a heterochromatic mass. This cellis probably no longer viable.



for this defect is occurring in HD mice. It is of interestthat a cyclic AMP-dependent protein kinase has recentlybeen shown to phosphorylate the proapoptotic proteinBAD, making it suitable for binding to the 14-3-3 proteinand therefore unable to promote apoptosis (Harada et al.,1999). The up-regulation of the regulatory subunit of thiskinase might oppose this specific antiapoptotic effect.

An opposite effect has been demonstrated in thymo-cytes for protein kinase C (PKC)-u, one of the genes wefind up-regulated in affected brains. PKC-u is a specificactivator of the stress kinase JNK; the SEK1–JNK path-way is also activated in hippocampal cells that weretransfected with a protein containing 89 glutamines (Liu,1998). The activation and translocation from the cyto-solic fraction to the particulate fraction of this proteinkinase induce apoptosis, and it is regulated by cal-cineurin activation (Asada et al., 1998). Calcineurin hasbeen shown to be the phosphatase that selectively in-duces apoptosis by dephosphorylation of BAD (Wang

et al., 1999). The 14-3-3 protein and PKC-u have beenshown to interact directly in T cells; overexpression of14-3-3 inhibits PKC-u translocation and its ability toinduce interleukin-2 promoter activity (Meller et al.,1996). It is also known that PKC-u is proteolyticallycleaved by caspase-3 during apoptosis induced by vari-ous agents (Datta et al., 1997); it is interesting thatoverexpression of PKC-u in the cleaved, kinase-activeform, results in the induction of sub-G1-phase DNA,nuclear fragmentation, and lethality.

In glial cells, activators of PKC prevent apoptosisinduced by the c-erbA a protooncogene (Llanos et al.,1998). This protooncogene, which encodes the thyroidhormone receptora1, is among the cDNAs we foundup-regulated in the affected brains. c-erbAhas also beenshown to induce apoptosis in early erythrocytic progen-itor cells (Gandrillon et al., 1994).

Little is known about neuroleukin, another gene thatwe found up-regulated in affected brains; it has beenshown that neuroleukin in its monomeric form is a neu-rotrophic factor (Mizrachi, 1989) and promotes the sur-vival of peripheral and central neurons in vitro and motorneuron regeneration in vivo (Gurney et al., 1986; Gur-ney, 1987). Its up-regulation could be involved in theaffected neurons’ apparent resistance to undergo thecomplete apoptotic program. Neuroleukin is up-regu-lated in HL60 cells exposed to x-irradiation (Balcer-Kubiczek et al., 1998).

SH-PTP2 is a mouse phosphotyrosine phosphatasecontaining two Src homology 2 (SH2) domains involvedin insulin, fibroblast growth factor, and possibly othersignaling pathways upstream of mitogen-activated pro-tein kinase.

Our gene expression analysis also showed that two ofthe 588 cDNAs screened are consistently down-regu-lated in affected mouse brains when compared withcontrol mice. The RNA polymerase I transcription ter-mination factor (TTF)-1 is a nucleolar protein responsi-ble for transcriptional termination by recognizing a sitecalled “Sal box” downstream of the 39 end of the pre-rRNA coding region (Grummt et al., 1986). In mice, italso recognizes a very similar region, called “T0 ele-ment,” which is located 150–200 bp upstream of thepromoter of the gene whose transcription is to be termi-nated. Binding of TTF-1 to this upstream sequence hasbeen shown to activate transcription from chromatintemplates by ATP-dependent nucleosome repositioning(Langst et al., 1997), demonstrating a new chromatin-remodeling activity by TTF-1. Its down-regulation couldalso be related to the changes in chromatin structureobserved in the affected neurons. The second gene thatshowed lower levels of expression in affected mice whencompared with controls is DNA topoisomerase II(Topo2). Topo2 is both an enzyme and a structuralcomponent of the nuclear matrix. It regulates the topo-logical states of DNA by transient cleavage, strand pass-ing, and religation of double-stranded DNA. Catalyticinhibition of Topo2 by merbarone induces apoptosis inan interleukin 1b-converting enzyme-dependent fashion

FIG. 6. In situ hybridization of transgenic mouse corpus stria-tum. a: Antisense prothymosin a probe. Medium-sized neuronsare markedly positive. b: Sense prothymosin a-labeled probe.No labeling is revealed by alkaline phosphatase. B, nerve bun-dles. Bars 5 200 mm.



in CEM cells (Khelifa and Beck, 1999), whereas inhibi-tion of Topo2 by fostriecin triggers apoptosis in HL60cells via changes in nuclear chromatin, affecting DNAsensitivity to denaturation (Hotz et al., 1992). The same

inhibitor causes necrosis in another cell type, MOLT-4.Topo2 is degraded during apoptosis (Sugimoto et al.,1998).

For one of the genes that showed a high degree ofup-regulation (prothymosina), we performed in situhybridization to establish whether the increase in expres-sion levels was due to glial contribution or by neuronalup-regulation only as our observations are made onwhole-brain RNA extracts, and the differences in geneexpression shown in Table 1 could correspond to the sumof variations occurring in different cell types or brainstructures rather than in each affected cell. Prothymosina remains expressed uniquely in neurons, as shown inFigs. 5 and 6. It will be important to establish thesignificance of the alterations in brain gene expressiondescribed by our work. Some of the altered genes areinvolved in chromatin and nucleosome remodeling, inagreement with the apparent nuclear condensation ob-served in the affected neurons, or in DNA repair, sug-gesting the possibility of DNA damage by either merenuclear translocation of the mutated portion of huntingtinor by accumulation of nuclear aggregates. Two of thegenes that are overexpressed in affected brains, neuro-leukin and prothymosina, seem to have a well-estab-lished role in promoting cell survival by inhibiting apo-

TABLE 1. Early modifications in brain gene expression in huntingtin transgenic mice

GenBank accession no. Gene/protein Fold increase

Up-regulatedD84372 Syp; SH-PTP2 (adaptor protein tyrosine phosphatase) .10M20473 Cyclic AMP-dependent protein kinase type-I-b regulatory chain .10X56135 Prothymosina .10X92411 MHR23B (RAD23 UV excision repair protein homologue) .6M14220 Neuroleukin .2D11091 PKC-u .2X51983 c-erbA oncogene (thyroid hormone receptor) .2

Down-regulatedX83974 RNA polymerase I termination factor TTF-1 .20D12513 DNA Topo2 .20

RNA from three different wild-type and mutant mice was used in three different screening hybridizationexperiments. The genes listed were found altered to similar degrees in all three sets of mice analyzed.

FIG. 7. Analysis of gene expression by cDNA array filter hybrid-ization. Two identical filters were hybridized with 32P-labeledcDNA prepared from poly(A)1 RNA extracted from (a) control or(b) transgenic mouse brains. The squared dots indicate thegenes that were found down-regulated in mutant brains: 1,TTF-1; and 2, Topo2. The circled dots show the genes that wereup-regulated in mutant brains: 1, c-erbA; 2, prothymosin a; 3,SYP; 4, type-I-b regulatory subunit of cyclic AMP-dependentprotein kinase; 5, PKC-u; 6, MHR23B; and 7, neuroleukin. Thesignal in the bottom part of the filter comes from two house-keeping genes used for normalization of signal intensity.

FIG. 8. Northern blot analysis of three of the genes that wefound up-regulated in the cDNA array screening experiment intransgenic mice (tg) versus control (wt). cAMP-dep PK, cyclicAMP-dependent protein kinase.



ptosis. This result might suggest that their up-regulationis a way sought by the affected neurons to retard oroppose the toxic effects of the mutated protein. Shouldthe use of these pathways be confirmed, it may provideclues to valuable options for a therapeutic delay or pre-vention of cell death in patients.

Acknowledgment: The authors would like to thank E.Signer for support and advice, Angelo Merante for invaluabletechnical assistance, and G. Bates, A. M. Tata, M. G. Farrace,P. Mastroberardino, and F. Capolunghi for their help. We thankDr. Gennaro Citro of SAFU at the IRE in Rome for use of theanimal facility. The electron microscopy data were obtainedusing the Interdepartmental Laboratory of Electron Microscopyfacilities (University of Roma Tre, Rome, Italy). This work waspartially supported by grants from AIRC and Cofin 97–98 toM.P. C.I. is the recipient of a “Cure HD Initiative” Fellowshipfrom the Hereditary Disease Foundation.

REFERENCES

Asada A., Zhao Y., Kondo S., and Iwata M. (1998) Induction ofthymocyte apoptosis by Ca21-independent protein kinase C(nPKC) activation and its regulation by calcineurin activation.J. Biol. Chem.273,28392–28398.

Balcer-Kubiczek E. K., Zhang X. F., Han L. H., Harrison G. H., DavisC. C., Zhou X. J., Ioffe V., McCready W. A., Abraham J. M., andMeltzer S. J. (1998) BIGEL analysis of gene expression in HL60cells exposed to X rays or 60 Hz magnetic fields.Radiat. Res.150,663–672.

Brandon E. P., Idzerda R. L., and McKnight G. S. (1997) PKAisoforms, neural pathways, and behaviour: making the connection.Curr. Opin. Neurobiol.3, 397–403.

Datta R., Kojima H., Yoshida K., and Kufe D. (1997) Caspase-3-mediated cleavage of protein kinase C theta in induction of apo-ptosis.J. Biol. Chem.272,20317–20320.

Davies S. W., Turmaine M., Cozens B. A., DiFiglia M., Sharp A. H.,Ross C. A., Scherzinger E., Wanker E. E., Mangiarini L., andBates G. P. (1997) Formation of neuronal intranuclear inclusionsunderlies the neurological dysfunction in mice transgenic for theHD mutation.Cell 90, 537–548.

Dıaz-Jullien C., Pe´rez-Este´vez A., Covelo G., and Freire M. (1996)Prothymosin alpha binds histones in vitro and shows activity innucleosome assembly assay.Biochim. Biophys. Acta1296,219–227.

Gandrillon O., Ferrand N., Michaille J. J., Roze L., Zile M. H., andSamarut J. (1994) c-erbA alpha/T3R and RARs control commit-ment of hematopoietic self-renewing progenitor cells to apoptosisor differentiation and are antagonized by the v-erbA oncogene.Oncogene9, 749–758.

Gomez-Marquez J. and Rodrı´guez P. (1998) Prothymosin alpha is achromatin-remodelling protein in mammalian cells.Biochem. J.333,1–3.

Grummt I., Rosenbauer H., Niedermeyer I., Maier U., and Ohrlein A.(1986) A repeated 18 bp sequence motif in the mouse rDNAspacer mediates binding of a nuclear factor and transcriptiontermination.Cell 45, 837–846.

Gurney M. E. (1987) Neuroleukin: basic biology and functional inter-action with human immunodeficiency virus.Immunol. Rev.100,203–223.

Gurney M. E., Heinrich S. P., Lee M. R., and Yin H. S. (1986)Molecular cloning and expression of neuroleukin, a neurotrophicfactor for spinal and sensory neurons.Science234,566–574.

Harada H., Becknell B., Wilm M., Mann M., Huang L. J., Taylor S. S.,Scott J. D., and Korsmeyer S. J. (1999) Phosphorylation andinactivation of BAD by mitochondria-anchored protein kinase A.Mol. Cell 3, 413–422.

Hodgson J. G., Agopyan N., Gutekunst C. A., Leavitt B. R., LePiane F.,Singaraja R., Smith D. J., Bissada N., McCutcheon K., Nasir J.,

Jamot L., Li X. J., Stevens M. E., Rosemond E., Roder J. C.,Phillips A. G., Rubin E. M., Hersch S. M., and Hayden M. R.(1999) A YAC mouse model for Huntington’s disease with full-length mutant huntingtin, cytoplasmic toxicity, and selective stri-atal neurodegeneration.Neuron23, 181–192.

Hotz M. A., Traganos F., and Darzynkiewicz Z. (1992) Changes innuclear chromatin related to apoptosis or necrosis induced by theDNA topoisomerase II inhibitor fostriecin in MOLT-4 and HL-60cells are revealed by altered DNA sensitivity to denaturation.Exp.Cell Res.201,184–191.

Karetsou Z., Sandaltzopoulos R., Frangou-Lazaridis M., Lai C. Y.,Tsolas O., Becker P. B., and Papamarcaki T. (1998) Prothymosinalpha modulates the interaction of histone H1 with chromatin.Nucleic Acids Res.26, 3111–3118.

Khelifa T. and Beck W. T. (1999) Merbarone, a catalytic inhibitor ofDNA topoisomerase II, induces apoptosis in CEM cells throughactivation of ICE/CED-3-like protease.Mol. Pharmacol.55,548–556.

Klement I. A., Skinner P. J., Kaytor M. D., Yi H., Hersch S. M., ClarkH. B., Zoghbi H. Y., and Orr H. T. (1998) Ataxin-1 nuclearlocalization and aggregation: role in polyglutamine-induced dis-ease in SCA1 transgenic mice.Cell 95, 41–53.

Langst G., Blank T. A., Becker P. B., and Grummt I. (1997) RNApolymerase I transcription on nucleosomal templates: the tran-scription termination factor TTF-I induces chromatin remodel-ing and relieves transcriptional repression.EMBO J.16, 760 –768.

Liu Y. F. (1998) Expression of polyglutamine-expanded huntingtinactivates the SEK1–JNK pathway and induces apoptosis in ahippocampal neuronal cell line.J. Biol. Chem.273, 28873–28877.

Llanos S., Caelles C., Azorıń I., Renau-Piqueras J., Fernańdez-LunaJ. L., Bosca´ L., and Munoz A. (1998) The c-erbAa protooncogeneinduces apoptosis in glial cells via a protein kinase C- andbcl-2-suppressible mechanism.J. Neurochem.70, 2315–2326.

Mangiarini L., Sathasivam K., Seller M., Cozens B., Harper A., Heth-erington C., Lawton M., Trottier Y., Lehrach H., Davies S. W.,and Bates G. P. (1996) Exon 1 of the HD gene with an expandedCAG repeat is sufficient to cause a progressive neurologicalphenotype in transgenic mice.Cell 87, 493–506.

Meller N., Liu Y. C., Collins T. L., Bonnefoy-Be´rard N., Baier G.,Isakov N., and Altman A. (1996) Direct interaction betweenprotein kinase C theta (PKC theta) and 14-3-3 tau in T cells:14-3-3 overexpression results in inhibition of PKC theta translo-cation and function.Mol. Cell. Biol. 10, 5782–5791.

Mellon P. L., Clegg C. H., Correll L. A., and McKnight G. S. (1989)Regulation of transcription by cyclic AMP-dependent protein ki-nase.Proc. Natl. Acad. Sci. USA86, 4887–4891.

Mizrachi Y. (1989) Neurotrophic activity of monomeric glucophos-phoisomerase was blocked by human immunodeficiency virus(HIV-1) and peptides from HIV-1 envelope glycoprotein.J. Neu-rosci. Res.23, 217–224.

Moulder K. L., Onodera O., Burke J. R., Strittmatter W. J., and JohnsonE. M. Jr. (1999) Generation of neuronal intranuclear inclusions bypolyglutamine-GFP: analysis of inclusion clearance and toxicityas a function of polyglutamine length.J. Neurosci.19, 705–715.

Reddy P. H., Williams M., Charles V., Garrett L., Pike-Buchanan L.,Whetsell W. O. Jr., Miller G., and Tagle D. A. (1998) Behaviouralabnormalities and selective neuronal loss in HD transgenic miceexpressing mutated full-length HD cDNA.Nat. Genet.20, 198–202.

Rodrıguez P., Vin˜uela J. E., Alvarez-Fernańdez L., and Go´mez-Mar-quez J. (1999) Prothymosin alpha antisense oligonucleotides in-duce apoptosis in HL-60 cells.Cell Death Differ.1, 3–5.

Schauber C., Chen L., Tongaonkar P., Vega I., Lambertson D., PottsW., and Madura K. (1998) Rad23 links DNA repair to the ubiq-uitin/proteasome pathway.Nature391,715–718.

Schilling G., Becher M. W., Sharp A. H., Jinnah H. A., Duan K.,Kotzuk J. A., Slunt H. H., Ratovitski T., Cooper J. K., JenkinsN. A., Copeland N. G., Price D. L., Ross C. A., and Borchelt D. R.(1999) Intranuclear inclusions and neuritic aggregates in trans-



genic mice expressing a mutant N-terminal fragment of hunting-tin. Hum. Mol. Genet.8, 397–407.

Shelbourne P. F., Killeen N., Hevner R. F., Johnston H. M., TecottL., Lewandoski M., Ennis M., Ramirez L., Li Z., Iannicola C.,Littman D. R., and Myers R. M. (1999) A Huntington’s diseaseCAG expansion at the murinehdh locus is unstable and asso-ciated with behavioural abnormalities in mice.Hum. Mol.Genet.8, 763–774.

Sugimoto K., Yamada K., Egashira M., Yazaki Y., Hirai H., KikuchiA., and Oshimi K. (1998) Temporal and spatial distribution of

DNA topoisomerase II alters during proliferation, differentiation,and apoptosis in HL-60 cells.Blood 91, 1407–1417.

Wang H. G., Pathan N., Ethell I. M., Krajewski S., Yamaguchi Y.,Shibasaki F., McKeon F., Bobo T., Franke T. F., and Reed J. C.(1999) Ca21-induced apoptosis through calcineurin dephosphor-ylation of BAD. Science284,339–343.

Wiley J. C., Wailes L. A., Idzerda R. L., and McKnight G. S. (1999)Role of regulatory subunits and protein kinase inhibitor (PKI) indetermining nuclear localization and activity of the catalytic sub-unit of protein kinase A.J. Biol. Chem.274,6381–6387.



Early Alterations in Gene Expression and Cell Morphology in a Mouse Model of Huntington's Disease

Documents