Triplet Repeat Diseases - Wiley-VCH · 2 Triplet Repeat Diseases 4.2 Fragile XE Mental Retardation 34 4.3 Friedreich’s Ataxia 36 4.4 Progressive Myoclonus Epilepsy Type 1 39 5 Type

1

Triplet Repeat Diseases

Stephan J. Guyenet and Albert R. La SpadaUniversity of Washington, Seattle, WA 98195

1 A Novel Mechanism of Genetic Mutation Emerges 31.1 Repeat Sequences of All Types and Sizes 31.2 Trinucleotide Repeat Expansion as a Cause of Disease: Unique

Features Explain Unusual Genetics 4

2 Repeat Diseases and Their Classification 52.1 Summary of Repeat Diseases 52.2 Differences in Repeat Sequence Composition and Location within

Gene 62.3 Classification Based upon Mechanism of Pathogenesis and

Nature of Mutation 6

3 Type 1: The CAG/Polyglutamine Repeat Diseases 93.1 Spinal and Bulbar Muscular Atrophy 93.2 Huntington’s Disease 123.3 Dentatorubral Pallidoluysian Atrophy 153.4 Spinocerebellar Ataxia Type 1 163.5 Spinocerebellar Ataxia Type 2 183.6 Spinocerebellar Ataxia Type 3/Machado–Joseph Disease 193.7 Spinocerebellar Ataxia Type 6 213.8 Spinocerebellar Ataxia Type 7 223.9 Spinocerebellar Ataxia Type 17 253.10 Role of Aggregation in Polyglutamine Disease Pathogenesis 253.11 Protein Context 293.12 Transcriptional Dysregulation 303.13 Proteolytic Cleavage 30

4 Type 2: the Loss-of-function Repeat Diseases 334.1 Fragile X Syndrome 33

Encyclopedia of Molecular Cell Biology and Molecular Medicine, 2nd Edition. Volume 15Edited by Robert A. Meyers.Copyright 2005 Wiley-VCH Verlag GmbH & Co. KGaA, WeinheimISBN: 3-527-30652-8

2 Triplet Repeat Diseases

4.2 Fragile XE Mental Retardation 344.3 Friedreich’s Ataxia 364.4 Progressive Myoclonus Epilepsy Type 1 39

5 Type 3: the RNA Gain-of-function Repeat Diseases 405.1 Myotonic Dystrophy Type 1 405.2 Myotonic Dystrophy Type 2 425.3 Spinocerebellar Ataxia Type 8 435.4 The Fragile X Tremor – Ataxia Syndrome (FXTAS) 44

6 Type 4: The Polyalanine Diseases 466.1 Overview 466.2 Oculopharyngeal Muscular dystrophy 466.3 Synpolydactyly (Syndactyly Type II) 486.4 Cleidocranial Dysplasia 486.5 Holoprosencephaly 486.6 Hand-foot-genital Syndrome 496.7 Blepharophimosis-ptosis-epicanthus Inversus Syndrome 496.8 Syndromic and Nonsyndromic X-linked Mental Retardation 496.9 Congenital Central Hypoventilation Syndrome 50

7 Unclassified Repeat Diseases Lacking Mechanistic Explanations 507.1 Spinocerebellar Ataxia Type 10 507.2 Spinocerebellar Ataxia Type 12 517.3 Huntington’s Disease Like 2 (HDL2) 52

Bibliography 54Books and Reviews 54Primary Literature 54

Keywords

AggregateAccumulation of proteinaceous and/or ribonucleic acid material into a structure that isvisible in a cell at the light microscope level.

AnticipationWorsening severity of a disease phenotype as the causal (typically dominant) geneticmutation is transmitted from one generation to the next in a family segregating thedisease of interest.

Triplet Repeat Diseases 3

Gain of functionRefers to a type of mutation that imparts a novel activity or action to the gene productcontaining the mutation.

Loss of functionRefers to a type of mutation that eliminates the action of the gene product encoded bythe gene within which the mutation resides.

Repeat ExpansionAn elongation of a repeat to a larger size that no longer falls within the size distributionrange typically seen in the normal population; this process is now recognized as amechanism of human genetic mutation.

TrinucleotideThree DNA base pairs of specific sequence composition (e.g. cytosine-adenine-guanine).

� The repeat expansion disorders are a group of human diseases that are caused bythe elongation of a DNA repeat sequence. In this chapter, we provide an overviewof the discovery of repeat expansion as an important cause of human disease, andwe summarize the molecular genetics and mechanistic basis of 27 microsatelliterepeat disorders. Comparison of the many repeat expansion disorders reveals distinctcategories of repeat diseases, allowing us to propose a classification of the repeatexpansion disorders based upon mutation sequence and pathogenic mechanism.The four types of repeat expansion disorders defined by this approach are theCAG/polyglutamine repeat diseases; the loss-of-function repeat diseases; the RNAgain-of-function repeat diseases; and the polyalanine diseases. Although the geneticbasis for most of these diseases was determined less than a decade or so ago,considerable advances have been made in our understanding of how ‘‘dynamicmutations’’ produce molecular pathology and human disease.

1A Novel Mechanism of Genetic MutationEmerges

1.1Repeat Sequences of All Types and Sizes

Long before the sequencing of the humangenome was undertaken, the discovery ofbacterial enzymes that recognize specificDNA sequences (‘‘recognition sites’’) and

cleaved them, yielded a new methodologyfor differentiating individuals (‘‘molec-ular fingerprinting’’). This methodologyalso found application in the mapping ofinherited genetic diseases, taking advan-tage of human variation in the form ofso-called restriction fragment length poly-morphisms (RFLP’s). In the search foreven more informative genetic markers,investigators uncovered a variety of very


short repeat sequences (‘‘short tandemrepeats’’ – STR’s; ‘‘simple sequence lengthpolymorphisms’’ – SSLP’s) that werehighly dispersed throughout the humangenome. These latter repeat sequencescame to be known as ‘‘microsatellites’’ todifferentiate them from the ‘‘minisatel-lites’’ that were being used for molec-ular fingerprinting. Minisatellites wereoriginally defined as tandem arrays of14–100 bp repeating sequences. In thecase of minisatellites, the repeat sequencewas in essence a ‘‘consensus’’ as devi-ations from the exact repeat sequencewere common. Microsatellites, however,were typically pure perfect repeats ofless than 13 bp, the most common mi-crosatellites being dinucleotide repeats,trinucleotide repeats, or tetranucleotide re-peats. The discovery of minisatellites andmicrosatellites yielded a virtual bonanzaof reagents for molecular fingerprintingand genetic linkage mapping, while alsoproviding evolutionary biologists and pop-ulation geneticists with intriguing materialto attempt to reconstruct evolutionaryrelationships and inter- or intraspecies re-lationships. Although human geneticistsand evolutionary biologists were applyingmicrosatellites in different ways for theirown studies, they shared the commonlyheld belief that such repeats were neutraland therefore unlikely to be of much func-tional consequence, let alone play a role incausing human disease. Of course, all thatwould soon change in the last decade ofthe twentieth century.

1.2Trinucleotide Repeat Expansion as aCause of Disease: Unique Features ExplainUnusual Genetics

In 1991, two groups working on seem-ingly unrelated inherited genetic diseases

independently made paradigm-shiftingdiscoveries. In one case, a CAG trinu-cleotide repeat expansion within the firstcoding exon of the androgen receptor (AR)gene was found to be the cause of anX-linked neuromuscular disorder knownas spinal and bulbar muscular atrophy(SBMA or Kennedy’s disease). CAG en-codes the amino acid glutamine; thus,elongation of a polyglutamine tract withinthe AR protein was hypothesized to bethe molecular basis for the motor neu-ron degeneration in SBMA. In the othercase, a disorder known as the fragile Xsyndrome of mental retardation (FRAXA),also X-linked but much more common,was reported to result from expansion ofa CGG repeat. In the latter case, althoughoriginally envisioned to encode a polyargi-nine tract, the CGG repeat turned outto be in the 5′ untranslated region of anovel gene, the so-called FMR-1 gene (for‘‘fragile X mental retardation-1’’).

The identification of triplet repeat ex-pansions as the cause of two inheriteddiseases was an exciting turning point inthe field of human molecular genetics notonly because of the novel nature of thesefindings, but also because of the unusualgenetic characteristics of this new type ofmutation. Analysis of families segregat-ing FRAXA revealed the existence of threedistinct allele categories defined by thelength of the pure CGG repeat: a normalsize range; a disease size range; and anintermediate, ‘‘premutation’’ size range.As discussed below, individuals carryingpremutation-sized CGG repeats never de-velop the mental retardation phenotype,but instead are at risk for passing on evenlarger CGG repeats to their children orgrandchildren who then display the men-tal retardation phenotype. An importanttenet that emerged from these early stud-ies was that expanded CGG repeats (larger


than the normal size range) displayed anexceptionally high rate of further mutation.This observation reversed a commonlyheld view of the genetic material – thenotion that any single nucleotide in the hu-man genome displayed a mutation rate of∼1 × 10−5. In the case of expanded CGGrepeats in premutation carriers, the rateapproached unity (100)! Besides displayingthis high mutation rate, there were otherunusual features: (1) the CGG repeatsshowed a marked tendency to further ex-pansion, suggesting that repeat mutationwas a polar process; and (2) the sex of theindividual transmitting the premutation-sized CGG allele determined whether alarge expansion into the disease rangewould be possible. For FRAXA, expansioninto the disease range could only occur ifthe premutation allele was transmitted by afemale carrier. All of these unusual aspectsof FRAXA CGG repeat genetics thoroughlyaccounted for the bizarre non-Mendelianinheritance patterns described in FRAXAfamilies as the ‘‘Sherman paradox.’’

The recognition of repeat expansion asthe cause of the neurodegenerative dis-order SBMA and as an explanation forthe puzzling genetics of FRAXA set thestage for further discoveries in the field ofneurogenetics (the study of inherited neu-rological disorders). One disorder knownas myotonic dystrophy (DM), the most com-mon of all muscular dystrophies, had beenthe center of a genetic controversy that hadgone on for nearly a century. The contro-versy involved the debate over whether theclinical phenomenon of anticipation trulyexisted or was simply an artifact of clin-ical study (Anticipation may be definedas a progressively earlier age of diseaseonset with increasing disease severity insuccessive generations of a family segre-gating an inherited disorder). Althoughanticipation was initially proposed as a

defining feature of DM in 1918, a num-ber of leading geneticists, among themPenrose, L.S., dismissed its authenticity,claiming that it was a product of ascertain-ment bias due to better clinically definingthe profound variable expressitivity in thisdisorder. This view persisted from its pro-mulgation in the 1950s, reinforced bythe concept that the genetic material isseldom subject to alteration or modifi-cation that could be heritable. However,with the discovery that expanded trinu-cleotide repeats could further expand, andthat indeed the expansion process was aprerequisite for the FRAXA disease pheno-type – accounting for maternal inheritanceand greater percentages of affected individ-uals in more recent generations – a rolefor repeats in diseases displaying antici-pation was entertained. Consequently, thethird repeat mutation disorder to be iden-tified – just one year after the SBMA andFRAXA discoveries – was DM. Studies ofthe causal CTG repeat expansion in DMdemonstrated a strong correlation betweendisease severity (i.e. age of onset and rateof progression) and the length of the CTGrepeat. In this way, it became clear thatanticipation was a genuine phenomenonand that expanded repeat mutational in-stability was its long awaited molecularexplanation. The mutational instabilitycharacteristics of expanded repeats haveled to their designation as so-called ‘‘dy-namic mutations.’’

2Repeat Diseases and Their Classification

2.1Summary of Repeat Diseases

The list of diseases caused by microsatel-lite repeat expansions (involving repeats of


3–12 bp) now includes more than 25 disor-ders (Table 1). Certain aspects of the list ofrepeat diseases deserve emphasis. Twentyof the repeat diseases either principallyor exclusively affect the neuraxis – that is,anywhere from the brain to the cerebel-lum/brainstem, to the spinal cord, to theperipheral nerve or muscle – and are de-generative disorders. Almost all of these in-herited neurological disorders are causedby large expanded repeats that displaythe property of pronounced genetic insta-bility (dynamic mutation). On the otherhand, the developmental malformationsyndromes for which repeat expansionmutations have been implicated all involvemodestly sized disease repeats by compar-ison, and these repeats do not exhibit suchdynamic mutation genetic instability.

2.2Differences in Repeat SequenceComposition and Location within Gene

When faced with the task of categorizingthe various repeat expansion diseases intodifferent classes, a number of approachesare possible. We have found that consid-eration of the sequence of the repeat andits location within the gene are the mostuseful characteristics to apply for groupingthe different repeat diseases. As shown inTable 1, there are many different types ofrepeats varying in length and sequencecomposition. However, among the recur-rent sequence types are CAG trinucleotiderepeats, CTG trinucleotide repeats, andGCG trinucleotide repeats. The rest of therepeat sequences are unique in composi-tion and differ widely in size, ranging from3 to 12 bp as noted above. Comparison ofthe location of a repeat within the geneit is affecting also yields different types ofrepeats. The largest single-repeat locationcategory is within the coding region of

a gene, which applies to both CAG (glu-tamine) and GCG (alanine) repeats, andhas been proposed but not yet demon-strated for CTG (leucine) repeats. The restof the locations defined for repeats varywidely, ranging from the gene’s promoterto its 5′ untranslated region to an intronto the 3′ untranslated region, and no morethan two to three repeat expansions canbe placed in each of these categories atthis time.

2.3Classification Based upon Mechanismof Pathogenesis and Nature of Mutation

To allow us to reconstruct how the dif-ferent repeat expansion mutations causemolecular pathology in the various dis-orders that they cause, we have chosento categorize the 25 repeat disorders intofour classes (Table 2). The first class ofdisorders, the Type 1 repeat diseases,are the ‘‘CAG-polyglutamine disorders.’’This class of repeat diseases includesnine inherited neurodegenerative disor-ders (SBMA, Huntington’s disease, den-tatorubral pallidoluysian atrophy, and sixforms of spinocerebellar ataxia) that allshare the common feature of being causedby a CAG repeat located within the codingregion of a gene. Upon CAG repeat expan-sion, a mutant protein with an extendedpolyglutamine tract is produced, mak-ing the protein then adopt an abnormalconformation and misfold to initiate thepathogenic cascade. The resultant pathol-ogy is believed to primarily stem from again of function of the mutant proteinimparted by the expanded polyglutaminetract. As discussed below, much effort hasgone into trying to define what the gain-of-function effect is for each polyglutaminedisease protein and into determining if


Tab.

1C

ompi

latio

nof

repe

atex

pans

ion

dise

ases

inhu

man

s.

Dis

ease

Gen

eR

epea

tN

orm

alPr

emut

atio

nD

isea

se

Ble

phar

ophi

mos

is-p

tosi

s-ep

ican

thus

inve

rsus

FOX

L2G

CG

14–

24C

leid

ocra

nial

dysp

lasi

aR

UN

X2

GC

G11

–17

–27

Con

geni

talc

entr

alhy

pove

ntila

tion

PHO

X-2

BG

CG

20–

25–

29D

enta

toru

bral

palli

dolu

ysia

nat

roph

y(H

awR

iver

synd

rom

e)A

trop

hin-

1C

AG

7–

35–

49–

88Fr

agile

Xsy

ndro

me

ofm

enta

lret

arda

tion

Fmr-

1C

GG

6–

5345

–20

0a20

0–>

2000

FRA

XE

men

talr

etar

datio

nFm

r-2

GC

C6

–40

61–

200

200

–10

00Fr

iedr

eich

’sat

axia

Frat

axin

GA

A7

–38

33–

6066

–17

00H

and-

foot

-gen

ital

HO

X-A

13G

CG

18–

24–

27H

untin

gton

’sdi

seas

eH

tt/I

T15

CA

G6

–35

27–

3639

–25

0H

untin

gton

’sdi

seas

elik

e2

JPH

3C

AG

7–

27–

50–

60H

olop

rose

ncep

haly

ZIC

2G

CG

15–

25M

yocl

onus

epile

psy

type

1C

ysta

tinB

CC

CC

GC

CC

CG

CG

2–

312

–17

30–

150

Myo

toni

cdy

stro

phy

type

1D

MPK

CTG

5–

3750

–80

50–

4000

Myo

toni

cdy

stro

phy

type

2Z

NF9

CC

TG10

4–

176

?17

5–

1100

0O

culo

phar

ynge

alm

uscu

lar

dyst

roph

yPA

BP2

GC

G6

–7

–8

–17

Spin

ocer

ebel

lar

atax

iaty

pe1

Ata

xin-

1C

AG

6–

44–

39–>

10Sp

inoc

ereb

ella

rat

axia

type

2A

taxi

n-2

CA

G13

–33

–31

–>

10Sp

inoc

ereb

ella

rat

axia

type

3(M

acha

do-J

osep

hdi

seas

e)A

taxi

n-3

CA

G12

–40

–55

–84

Spin

ocer

ebel

lar

atax

iaty

pe6

CA

CN

A1A

CA

G4

–18

–19

–33

Spin

ocer

ebel

lar

atax

iaty

pe7

Ata

xin-

7C

AG

4–

35–

37–

306

Spin

ocer

ebel

lar

atax

iaty

pe8

?C

TG16

–>

800

?71

–>

1000

Spin

ocer

ebel

lar

atax

iaty

pe10

Ata

xin-

10A

TTC

T10

–22

?80

0–

4500

Spin

ocer

ebel

lar

atax

iaty

pe12

PPP2

R2B

CA

G7

–32

–55

–78

Spin

ocer

ebel

lar

atax

iaty

pe17

TBP

CA

G25

–48

–43

–66

Synd

rom

ican

dno

nsyn

drom

icX

-link

edm

enta

lret

arda

tion

AR

XG

CG

10–

16–

17–

23Sy

npol

ydac

tyly

HO

X-D

13G

CG

9–

15–

16–

29X

-link

edsp

inal

&bu

lbar

mus

cula

rat

roph

yA

RC

AG

5–

34–

37–

70

a Mal

epr

emut

atio

nca

rrie

rsof

the

FMR

-1C

GG

repe

atde

velo

pth

eFr

agile

Xtr

emor

-ata

xia

synd

rom

e.–,

nopr

emut

atio

nal

lele

sid

entifi

ed,s

oth

eym

ayno

texi

st.

?,ex

iste

nce

ofpr

emut

atio

nal

lele

sis

unkn

own.


Tab.

2C

lass

ifica

tion

ofth

ere

peat

expa

nsio

ndi

seas

es.

CA

G/p

olyg

luta

min

edi

sord

ers

Loss

-of-f

unct

ion

diso

rder

sR

NA

gain

-of-f

unct

ion

diso

rder

sPo

lyal

anin

edi

sord

ers

Den

tato

rubr

alpa

llido

luys

ian

atro

phy

Frag

ileX

synd

rom

eof

MR

Myo

toni

cdy

stro

phy

type

1B

leph

arop

him

osis

-pto

sis-

epic

anth

usin

vers

usH

untin

gton

’sdi

seas

eFR

AX

EM

RM

yoto

nic

dyst

roph

yty

pe2

Spin

alan

dbu

lbar

mus

cula

rat

roph

yFr

iedr

eich

’sat

axia

Spin

ocer

ebel

lar

atax

iaty

pe8

Cle

idoc

rani

aldy

spla

sia

Spin

ocer

ebel

lar

atax

iaty

pe1

Myo

clon

usep

ileps

yty

pe1

Frag

ileX

trem

or-a

taxi

asy

ndro

me

Con

geni

talc

entr

alhy

pove

ntila

tion

Spin

ocer

ebel

lar

atax

iaty

pe2

Han

d-fo

ot-g

enita

lsyn

drom

eSp

inoc

ereb

ella

rat

axia

type

3H

olop

rose

ncep

haly

Spin

ocer

ebel

lar

atax

iaty

pe6

Ocu

loph

aryn

geal

mus

cula

rdy

stro

phy

Spin

ocer

ebel

lar

atax

iaty

pe7

Synp

olyd

acty

lySp

inoc

ereb

ella

rat

axia

type

17X

-link

edM

R(s

yndr

omic

&no

nsyn

drom

ic)

MR

=m

enta

lret

arda

tion.


shared pathways of toxicity are initiated inthe different diseases. At least one, and per-haps a greater number of these disordersmay principally involve a simultaneousdominant-negative partial loss of the nor-mal function of the disease protein. Thenext class of disorders, the Type 2 repeatdiseases, are a much more disparate groupof repeat disorders. The Type 2 repeat dis-eases are the ‘‘Loss-of-function repeat dis-orders.’’ These disorders include differentrepeats that vary in sequence composi-tion and gene location, but share a finalcommon pathway of disease pathogene-sis – a loss of function of the disease genewithin which they occur. This group in-cludes various classic trinucleotide repeatdisorders such as the two fragile X syn-dromes of mental retardation (FRAXA andFRAXE) and Friedreich’s ataxia – but alsoencompasses the dodecamer repeat ex-pansion in progressive myoclonic epilepsytype 1, and possibly the CAG repeat ex-pansion in Huntington’s disease like-2(HDL2) gene. Strong evidence for a loss-of-function pathway in the form of nonrepeatloss-of-function mutations supports manyof these classifications. The third group ofrepeat diseases, the Type 3 disorders, com-prise a shared class because all of themhave been proposed to involve the produc-tion of a toxic RNA species. This categoryof repeat diseases is thus called the RNAgain-of-function disorders. Included amongthese disorders are two closely relatedforms of DM, the common and classicmyotonic dystrophy type 1 (DM1) and itsuncommon phenocopy, myotonic dystro-phy type 2 (DM2). Another member ofthis group is the recently described frag-ile X tremor-ataxia syndrome (FXTAS) inmale premutation carriers – a fascinatingexample of two different disease path-ways operating upon the same expandedrepeat mutation based upon size range

differences. One form of spinocerebellarataxia (SCA8) with an unclear mechanismof pathogenesis has also been provision-ally placed into this category, based uponcurrent working models of how its re-peat causes disease. The last class ofrepeat disease, the Type 4 disorders, arethe ‘‘GCG-polyalanine disorders’’ that aregrouped together because all involve shortGCG repeat tracts falling within the codingregions of unrelated genes that become ex-panded to moderately sized GCG repeats.With the exception of oculopharyngealmuscular dystrophy, all are developmentalmalformation syndromes, and while gain-of-function polyalanine toxicity has beenproposed for a number of these disorders,loss of function due to the polyalanineexpansion seems more likely for others.Finally, a number of repeat disorders,spinocerebellar ataxia type 10 (SCA10),spinocerebellar ataxia type 12 (SCA12),and Huntington’s disease like 2 (HDL2),currently defy classification because verylittle is known about their molecular basis.These diseases will be considered in thefinal section of this chapter.

3Type 1: The CAG/Polyglutamine RepeatDiseases

3.1Spinal and Bulbar Muscular Atrophy

Spinal and bulbar muscular atrophy(SBMA; Kennedy’s disease) is a late-onsetneurodegenerative disease with an inher-itance pattern resembling X-linked reces-sive. It has a prevalence of about 1 in 50 000males. Patients suffer a late-onset, pro-gressive degeneration of primarily lowermotor neurons in the anterior horn of thespinal cord and in the bulbar region of the


brainstem; however, variable involvementof sensory neurons in the dorsal root gan-glia also occurs. SBMA typically presentswith cramps, followed by proximal mus-cle weakness and atrophy. Patients oftenexhibit dysarthria, dysphagia, and fascic-ulations of the tongue and lips. Affectedindividuals have symptoms of mild andro-gen insensitivity, such as gynecomastia,reduced fertility, and testicular atrophy.

SBMA is caused by a polymorphic(CAG)n repeat in the first exon of the ARgene, which is expressed as a glutaminetract. Unaffected individuals carry 5 to 34triplet repeats, while affected individualscarry 37 to 70 repeats. SBMA exhibitsa paternal expansion bias. The diseasedoes not appear to involve a simple loss-of-function mechanism, as complete lossof AR does not result in motor neurondegeneration.

AR is widely expressed in males andfemales, and is a member of the steroid re-ceptor–thyroid receptor superfamily witha highly conserved DNA binding domain,ligand binding domain and transactiva-tion domain (Fig. 1). In its inactive state,it forms an apo-receptor complex with

heat-shock proteins (HSPs) 70 and 90,and resides in the cytoplasm. Upon bind-ing androgen, it dissociates from theseHSP chaperone proteins and translocatesto the nucleus. Once in the nucleus, ARdimers transactivate certain genes, manyof which are responsible for generatingand maintaining male characteristics. Al-though the glutamine expansion does notaffect the binding of its ligand, androgen(testosterone), the glutamine tract is in themajor transactivation domain, and may af-fect transactivation competence. However,the effect of the polyglutamine expansionupon AR transactivation competence re-mains controversial.

Many lines of evidence suggest that ARmust translocate to the nucleus to exertits toxicity. Nuclear inclusions (NIs) arepresent in motor neurons of the spinalcord and brainstem in SBMA patients.In a Drosophila model of SBMA, reti-nal expression of mutant AR only yieldeda degenerative phenotype in the pres-ence of ligand. As in humans, androgenbinding causes the nuclear translocationof AR in mice. Transgenic male SBMAmice produce testosterone and will only

1 919

...CAGCAGCAG...

Normal = 5–34SBMA = 37–70

Fig. 1 Diagram of the androgen receptor. The androgen receptor is a member of thesteroid receptor-thyroid hormone receptor superfamily, and consequently displays astereotypical architecture. The CAG repeat – polyglutamine tract (striped box) resideswithin the amino-terminal domain, which mediates transcription activation throughan ‘‘activation function’’ domain (charcoal gray box). Additional conserved domainsinclude the DNA binding domain (light gray box), nuclear localization signal (blackbox), and the ligand binding domain (checkered box). Expansion of the CAG repeatto alleles of ≥37 triplets is the cause of spinal and bulbar muscular atrophy (SBMA).


develop motor neuron disease if express-ing polyglutamine-expanded full-lengthAR protein, yet when castrated, suchSBMA transgenic mice do not developa phenotype. In the same study, femaletransgenic mice hemizygous for expandedAR develop motor neuron degenerationwhen exposed to exogenous androgen. Im-portantly, even rare human females whoare homozygous for an AR CAG repeat ex-pansion mutation do not develop SBMA,despite widespread expression of mutantAR throughout the CNS (central nervoussystem). Thus, SBMA is not a true X-linkedrecessive disorder, but rather is classifiedas a sex-limited disorder, since expres-sion of the disease phenotype is dependentupon male levels of androgen.

A number of studies on SBMA patientsand mouse models have characterized theNIs seen in this disease. While NIs arewidely distributed in lower motor neu-rons of the spinal cord and brainstem,NIs also occur in a variety of nonneu-ronal tissues that appear to function nor-mally. NIs colocalize with componentsof the proteasome and with molecularchaperones. How this contributes to thephenotype is unknown, although HSP70overexpression attenuates toxicity in cellculture and in transgenic mouse. Thepresence of proteasome components andHSPs in NIs may thus be revealing aprotective intervention by the cell, or al-ternatively may be deleterious due todepletion of these important cellular pro-teins. Interestingly, only antibodies raisedto amino-terminal fragments of AR de-tect NIs, indicating that proteolysis mayplay a role in the disorder. Caspase-3 hasbeen shown to cleave AR in a polyglu-tamine tract length-dependent manner invitro, and this may have pathogenic sig-nificance, as truncated AR is more toxicthan full-length protein in cell culture

studies and transgenic mouse models.The phosphorylation of AR is modulatedby androgen, and this posttranslationalmodification appears to enhance caspase-3 cleavage.

One possible mechanism of expandedAR toxicity is through transcription in-terference. Polyglutamine-expanded ARinteracts with the transcription coactivatorCREB-binding protein (CBP) in a polyg-lutamine tract length-dependent manner,colocalizes with CBP in spinal cord NIsfrom patients, and can interfere withCBP-dependent transcription. CBP is atranscription cofactor that regulates theexpression of vascular endothelial growthfactor (VEGF), among other genes. VEGFis important in motor neuron health,as deletion of a portion of its promotercalled the hypoxia response element (HRE)causes motor neuron degeneration even innormoxic mice. Pathologically expandedAR reduces VEGF transcript expressionin males, with VEGF165 isoform ex-pression reduced at both the RNA andprotein levels. Adding VEGF165 to a mo-tor neuron-like cell line (MN-1) expressingpolyglutamine-expanded AR significantlyrescues its cell death, again supportingthe role of transcription interference inSBMA and suggesting that VEGF165 mayserve as a neurotrophic factor for mo-tor neurons.

While no effective treatment for SBMAhas yet been validated in human pa-tients, the role of testosterone in SBMAdisease progression has received consider-able attention. Interestingly, testosteronesupplementation was initially used as atreatment for SBMA. Rather than aggra-vating the disease as one might fear,it was reported to attenuate the pheno-type slightly, but did not significantlyretard disease progression. Such a ben-eficial effect of testosterone may be due


to anabolic effects on muscle strength orto a downregulation of AR in responseto elevated androgen levels. Still anotherpossibility is that AR-mediated transcrip-tion in motor neurons somehow performsa trophic function in the face of dam-age or injury. Very recent work, how-ever, strongly indicates that eliminationof ligand by surgical or pharmacologi-cal castration is a very effective treat-ment in SBMA transgenic mouse models,even showing efficacy when SBMA micedisplay an advanced phenotype. As lig-and binding is associated with nucleartranslocation of mutant AR and aber-rant effects upon nuclear transcriptionappear crucial for SBMA disease progres-sion, abrogation of nuclear localizationmay account for the success of castra-tion. Consistent with this hypothesis areresults of studies with flutamide, a drugthat appears to block AR-dependent trans-activation without preventing its nucleartranslocation. While pharmacological cas-tration with leuprolide is highly effectivetherapeutically, flutamide does not ame-liorate symptoms or disease progressionin an SBMA mouse model. Attemptedtranslation of these preclinical trial re-sults to human SBMA patients is cur-rently underway.

3.2Huntington’s Disease

Huntington’s disease (HD) is an autoso-mal dominant disorder with a prevalenceof 1 per 15 000 persons worldwide. It isa debilitating disease that often presentsclinically in the fourth or fifth decade oflife with chorea (i.e. spontaneous, invol-untary dancelike movements). Personalitychange and cognitive impairment mayprecede the clinical onset by years, andultimately culminate in dementia afteronset of the movement disorder. Choreagives way to bradykinesia and rigidity latein the disease. CNS atrophy occurs mostprominently in the striatum, which is re-duced to a fraction of its original size(Fig. 2). However, significant neurodegen-eration and neuron loss in the cortex is alsotypical, while cerebellum, brainstem, andspinal cord are relatively spared – except injuvenile-onset cases.

HD is caused by a (CAG)n repeat ex-pansion in the huntingtin gene (htt), whichencodes a 350-kDa protein containing 67exons. Unaffected individuals carry 6 to 35repeats, while affected individuals carry39 to 250 repeats. The largest repeatscause juvenile-onset HD and display a pa-ternal transmission bias. The htt proteinis ubiquitously expressed in brain tissue,

(a) (b)

Fig. 2 Huntington’s disease (HD)neuropathology. Hemi-coronal sectionsof postmortem brains from (a) a classic,adult-onset HD patient and (b) a normalcontrol reveal marked degeneration ofthe striatum (midmedial region) andconsiderable atrophy of cortical regions.(From Robataille et al. (1997) Brainpathol. 7, 901. Used with permission).


with highest levels in striatal interneu-rons and cortical pyramidal cells. Twohtt mRNA transcripts have been detected:one 10 kb and the other 13 kb in length,expressed most highly in CNS neurons.Ultrastructurally, the wild-type full-lengthprotein is predominantly cytoplasmic andis located in the pre- and postsynapticregions of dendrites and axons. Htt pro-tein associates with microtubules, vesicles,and organelles.

HD appears to have a predominantlydominant mechanism due to its inheri-tance pattern, evidence that homozygotesare no more severely affected than het-erozygotes, and the observation that het-erozygous deletion of huntingtin does notcause HD. The fact that no HD-causingloss-of-function mutations have been doc-umented in the huge htt gene furthersupports a gain-of-function mechanism.However, postnatal elimination of htt ex-pression in regions of the cortex can causestriatal degeneration in mice, so the no-tion that gain of function fully accountsfor the HD phenotype is being reexam-ined. As htt is a regulator of transcriptionactivation and/or mediator of vesicularbrain-derived neurotrophic factor (BDNF)transport up and down axons, many in-vestigators now envision the effects ofexpanded htt as twofold: simultaneouslycausing protein misfolding leading to gain-of-function toxicity, and inducing partialloss of an ill-defined normal function.

In 1997, it was first reported that mutanthtt forms dense amyloid-type aggregatesin the nucleus, perikarya, and neuropil ofneurons from a transgenic mouse modeland in human patients. The role of ag-gregation in the polyglutamine diseasesand in a wide range of neurodegenerativedisorders including Alzheimer’s disease,Parkinson’s disease, amyotrophic lateralsclerosis, and the prion diseases, thus

emerged as an important theme at theend of the last decade. While initially itwas thought that the formation of largeaggregates is the basis of polyglutamineneurotoxicity, the weight of evidence nowsuggests little correlation between visibleaggregate formation and disease pathol-ogy. However, the occurrence of aggre-gates along with neuropathology suggeststhat aggregation is inextricably linked tothe pathogenicity of polyglutamine diseaseproteins. We will address the role of poly-glutamine aggregation in neurotoxicity indetail in a separate section.

Many theories have been proposed toaccount for how polyglutamine-expandedhtt causes neurotoxicity. A thorough dis-cussion of this literature goes beyond thescope of this chapter, so the reader isreferred to the relevant books and re-views in the Bibliography for a moreintensive treatment of this topic. Ma-jor theories of htt neurotoxicity that willbe considered herein include transcrip-tion dysregulation, proteasome inhibition,mitochondrial dysfunction/excitotoxicity,and proteolytic cleavage. As multiple inde-pendent toxicity events may be occurringconcomitantly, these disease pathways arenot mutually exclusive.

Expanded htt protein may interferewith transcriptional processes. Many tran-scription factors, such as CBP, containglutamine tracts that mediate importantprotein–protein interactions. CBP inter-acts directly with htt, and mutant htttoxicity is ameliorated in striatal neurons invitro when CBP lacking the htt interactiondomain is overexpressed. CBP mediatesthe transcription of a number of neuronalsurvival factors, such as BDNF, and func-tions by acetylating histones, one aspect ofchromatin remodeling that permits tran-scription to occur. Studies of HD fruit flyand mouse models have highlighted the


potential importance of histone acetylationby demonstrating that histone deacety-lase inhibitors (HDAC Is) can successfullyrescue degenerative phenotypes in thesemodel organisms. Htt can affect transcrip-tion mediated by p53, Sp1, and RESTinteraction, thereby potentially altering theexpression of a large number of genes.Differential expression of survival factors,neurotransmitter receptors, and a numberof other genes may thus contribute to HDpathogenesis.

The ubiquitin-proteasome protein degra-dation pathway has also been implicatedin HD. Nuclear inclusions of htt colocalizewith ubiquitin, which indicates that the ex-panded protein has been identified as mis-folded and thus targeted for proteasomaldegradation. However, polyQ sequenceslonger than 9 glutamines are impossi-ble for eukaryotic proteasomes to cleave.The proteasomes of htt-transfected cellsare consequently less capable of degradingproteins other than htt, as demonstratedby the reduced degradation of GFP-taggedproteins in culture. If proteasome com-ponents are clogged with mutant proteinand/or sequestered into inclusions, theymay be unable to degrade other misfoldedor damaged proteins that carry out impor-tant functions. Alternatively, accumulationof improperly degraded proteins may inter-fere with other normal cellular processes,such as autophagy or mitochondrial oxida-tive phosphorylation.

Perhaps the longest and most thor-oughly studied htt toxicity pathway in-volves metabolic disturbances and excito-toxicity. Glucose metabolism and oxygenconsumption are reduced in HD brainsas measured by PET. Severe deficits inthe activity of complexes II/III and IV ofthe mitochondrial electron transport chain(ETC) are evident in HD brains. Inhibitorsof complex II of the mitochondrial electron

transport chain, such as 3-nitropropionicacid (3-NPA), can cause a selective degen-eration of the striatum in rat and primatemodels when injected systemically, sincethe striatum has among the highest energydemands of all neuronal regions. 3-NPAreduces levels of ATP, resulting in mito-chondrial and cellular depolarization withactivation of voltage-dependent NMDA re-ceptors. Excitotoxic damage may act inconcert with increased production of freeradicals to cause selective striatal degener-ation in HD. Some of the earliest animalmodels of HD were thus generated by ex-posing the striatum or entire brain of ratsor primates to metabolic or excitotoxic in-sults. In 1976, the first such model of HDwas created by injection of the glutamateanalog and excitotoxin kainic acid into thestriatum of rats. Such lesioned rats exhib-ited a selective degeneration of neuronsin the striatum reminiscent of HD. Re-cent studies have suggested that impairedCa++ flux due to aberrant interaction ofhtt with the inositol phosphate-3 receptor(IP3-R) may underlie excitotoxic pathology.Studies of mitochondria from HD patientsreveal abnormal mitochondrial Ca++ han-dling and decreased mitochondrial depo-larization thresholds, in support of this hy-pothesis. One very recent study has foundthat deletion of peroxisome proliferator-activated receptor (PPAR) gamma coacti-vator 1 alpha (PGC-1α), a key mediatorof mitochondrial biogenesis, yields HD-like striatal degeneration in mice. Thus,metabolic insults and excitotoxicity maybe key steps in HD disease pathogenesis.

Another important theory of htt neuro-toxicity posits that proteolytic cleavage ofhtt is a required step in neuronal dys-function and the degeneration process.As noted above, the htt protein is enor-mous, with full-length product consistingof >3100 amino acids. Analysis of human


HD material has indicated an absence ofmidprotein and C-terminal epitopes in httaggregates. Careful biochemical studies ofhtt have characterized a variety of putativecaspase and calpain cleavage sites. Variousstudies suggest that the more truncatedthe polyglutamine-expanded htt protein,the more toxic it is in cell culture and inanimal models, and the more likely it isto enter the nucleus and produce toxicitythere. In the case of htt, a series of pro-teolytic cleavage steps culminating withcleavage to an ∼100 amino acid peptidefragment by an aspartyl protease has beenproposed to yield a final ‘‘toxic fragment.’’As it turns out, the ‘‘toxic fragment hypoth-esis’’ (as it has also been called) may beapplicable to a number of polyQ diseases,which will be reviewed later in this section.

3.3Dentatorubral Pallidoluysian Atrophy

Dentatorubral pallidoluysian atrophy (DR-PLA) is a rare, autosomal dominantneurodegenerative disorder most preva-lent in Japan. Adult-onset DRPLA typi-cally involves progressive cerebellar ataxia,choreoatheosis, epilepsy, and dementia,while juvenile-onset cases also displaymyoclonus, epilepsy, and mental retar-dation. Neuropathological abnormalitiesinclude degeneration of the dentate nu-cleus of the cerebellum, rubral nucleus,and globus pallidus, as well as a moregeneralized degeneration and gliosis in-volving the brainstem, cerebellum, cortex,and pons. DRPLA is caused by a polymor-phic (CAG)n repeat in the carboxy-terminalcoding region of the atrophin-1 gene onchromosome 12. Normal individuals carry6–35 repeats, while affected individualscarry 49–88 repeats. Anticipation is promi-nent in DRPLA, as very large repeat

expansions can occur in a single gener-ation, typically via paternal transmission.

The DRPLA disease protein, atrophin-1, is widely expressed and appears tobe predominantly cytoplasmic. Atrophin-1contains both a putative nuclear localiza-tion signal (NLS) and nuclear export signal(NES), and nuclear localization of normalatrophin-1 is observed. Nuclear localiza-tion of polyglutamine-expanded atrophin-1 has been linked to increased toxicityin cell culture models. Ubiquitinated NIsare present in neurons and glia from pa-tient brains, and are also immunoreactivefor small ubiquitinlike modifier (SUMO)protein. Atrophin-1 is a substrate for c-Jun N-terminal kinase (JNK), with JNK’saffinity for atrophin-1 inversely propor-tional to the size of the polyglutaminetract expansion.

Although the relevance of JNK phos-phorylation to atrophin-1 function is un-known, other insights into the functionof atrophin-1 have been reported. UsingDrosophila melanogaster as a model sys-tem, Zhang et al. took advantage of theexistence of a fly ortholog of atrophin-1(Atro) and created lines of flies carry-ing mutations in the Atro gene. Theseflies demonstrated severe developmentalabnormalities due to complex pattern-ing defects, and subsequent experimentsrevealed that Atro is a transcription core-pressor whose activity diminishes withincreasing polyglutamine tract length. In-dependent studies of human atrophin-1 incell culture studies found evidence for tran-scription interference of polyglutamine-expanded atrophin-1 with CREB-mediatedgene activation. Transcription dysregula-tion may thus play a prominent rolein DRPLA. Another study suggests thatdisturbed carbohydrate metabolism maycontribute to the DRPLA neurodegenera-tive phenotype.


3.4Spinocerebellar Ataxia Type 1

Spinocerebellar ataxia type 1 (SCA1) is anautosomal dominant disorder character-ized primarily by a progressive cerebel-lar ataxia. It accounts for about 6% ofall autosomal dominant cerebellar ataxias(ADCAs) worldwide. SCA1 patients suf-fer from coordination difficulties includingdysarthria, dysphagia, and ophthalmople-gia. The cerebellum undergoes atrophy,gliosis, and severe loss of Purkinje cells, ac-companied by degeneration of the dentatenucleus, inferior olive, and some brain-stem nuclei. Disease onset typically occursin the third or fourth decade of life, butpresentation in childhood or adolescenceto late life may be seen. SCA1 is caused bythe expansion of a coding (CAG)n repeatin the amino-terminal coding region of theataxin-1 gene. The ataxin-1 CAG repeat ishighly polymorphic, ranging in size from6 to 44 triplets in unaffected individuals.The repeat length associated with diseaseranges from 39 to more than 100 CAGs.Unaffected individuals with more than20 repeats have CAT triplet interruptionswithin their (CAG)n repeat tracts. Such in-terruptions stabilize the repeat expansion,while absence of the CAT repeat intersper-sion is noted in SCA1 patient alleles. Asin many other polyglutamine disorders,

there is strong evidence supporting a gain-of-function mechanism in SCA1.

Ataxin-1 is widely expressed in the CNSand throughout the periphery, althoughexpression levels are several-fold higher innervous system tissues. The protein is pre-dominantly nuclear in the CNS, however,some cytoplasmic staining is apparent inPurkinje cells of the cerebellum and inbrainstem nuclei. Ataxin-1 knockout micedo not develop SCA1, although they doexhibit impairments in motor and spatiallearning. These mice also have decreasedpaired-pulse facilitation in the CA1 re-gion of the hippocampus, suggesting thatataxin-1 may normally function in synapticplasticity and learning.

Large ataxin-1 containing NI’s occurin the brainstem of affected individuals,and are immunoreactive for ubiquitin,the 20 S proteasome subunit, and HSPsHDJ2 and Hsc70. Work done on SCA1 intransgenic mice by the Orr and Zoghbi lab-oratories has been crucial in formulatingmodels of not only SCA1 disease patho-genesis but also for influencing viewsof the molecular basis of all polyglu-tamine diseases. Indeed, the first mousemodel for a polyglutamine disease wasgenerated by transgenic overexpression ofpolyglutamine-expanded ataxin-1 in Purk-inje cells (Fig. 3). This SCA1 transgenicmouse model has laid the foundation for

Fig. 3 The original spinocerebellar ataxia type 1 (SCA1) mouse model. (a) Diagram of thePcp2-SCA1 transgene construct. A Purkinje cell-specific expression cassette based upon inclusion ofthe promoter (straight line), first two noncoding exons (black boxes), and first intron (bent line) ofthe Purkinje cell protein 2 (Pcp2) gene and a SV40 polyadenylation sequence (open box) was thebasis for this landmark work. Ataxin-1 cDNAs (gray box) containing either 30 CAGs (control) or 82CAGs (expanded) were inserted into the Pcp2 expression cassette. Sites of various PCR primer setsare also shown. (b) Pcp2-SCA1 CAG-82 mice display ataxia. Still photographs of a 30-week-oldPcp2-SCA1 CAG-82 mouse from a transgenic line that greatly overexpresses the mutant ataxin-1transgene illustrates the inability of this mouse to maintain its balance when ambulating. Loss ofbalance when walking is consistent with the gait ataxia seen in human SCA1 patients. (From Burrightet al. (1995) Cell 82, 937, used with permission).


numerous follow-up studies of polyglu-tamine disease pathogenesis. For example,expression of mutant ataxin-1 lacking anintact self-association domain precludedaggregate formation, but permitted neu-rotoxicity, demonstrating that NIs are not

required for SCA1 in mice. In a later study,crossing of SCA1 transgenic mice withmice lacking a ubiquitin ligase enzymeyielded SCA1 mice incapable of aggregateformation. These mice were more severelyaffected than their transgenic counterparts

(a)

(b)


due to absence of the ubiquitin ligase.This work supported the view that visibleaggregate formation may represent a pro-tective cellular response for neutralizingmisfolded polyglutamine-containing pep-tides. In another study, mice expressingataxin-1 with a mutated NLS were foundnot to develop SCA1, suggesting that nu-clear localization is absolutely required forSCA1 molecular pathology. Although tar-geting ataxin-1 to Purkinje cells appearssufficient to recapitulate a dramatic SCA1-like disease in mice, a subsequent knockinmouse model of SCA1 indicated that ex-pression in other regions of the CNS yieldsa more representative disease phenotype.

Over the past few years, numerousleads have emerged in the search for thepathogenic basis of SCA1. In one lineof investigation, phosphorylation of ser-ine 776 of the ataxin-1 protein was shownto affect pathogenesis, highlighting theimportance of this posttranslational mod-ification. SCA1 transgenic mice in whichserine 776 had been mutated to an ala-nine, exhibited a dramatically attenuatedphenotype and a complete lack of NI’s.In an accompanying study, interaction ofpolyglutamine-expanded ataxin-1, but notwild-type ataxin-1, with several isoforms ofthe phosphoserine/threonine binding pro-tein 14-3-3 was reported. This extremelyabundant peptide is thought to serve aregulatory function by binding proteinsand determining their subcellular localiza-tion, among other things. The interactionbetween ataxin-1 and 14-3-3 was shownto be dependent upon the Akt-mediatedphosphorylation of serine 776. A novelmechanism for ataxin-1 toxicity was thusproposed: upon Akt phosphorylation ofpolyglutamine-expanded ataxin-1, ataxin-1binds 14-3-3, is stabilized, and ultimatelyaccumulates in the nucleus. The down-stream effects of the nuclear accumulation

of mutant ataxin-1 on neuronal function,however, remain undefined.

Transcriptional dysregulation is also areasonable hypothesis for SCA1 patho-genesis. One study has demonstrated thatataxin-1 interacts with polyglutamine bind-ing protein 1 (PQBP-1), and that thisinteraction results in interference withRNA polymerase-dependent transcription.Independent studies have supported a rolefor ataxin-1 as a transcription corepressor.In pull-down assays and in Drosophila,ataxin-1 interacts with the proteins SMRT(silencing mediator of retinoid and thyroidhormone receptors) and HDAC3 (his-tone deacetylase 3), both transcriptionalrepressors. Aggregates of polyglutamine-expanded ataxin-1 sequester SMRTER, theDrosophila ortholog of SMRT. Transcrip-tion repressors and histone deacetylasesalso appeared in an earlier screen for mod-ulators of the Ataxin-1 phenotype in the fly.In addition, two separate studies of geneexpression alterations in presymptomaticSCA1 transgenic mice have uncoveredchanges in the levels of transcripts en-coding proteins involved in Ca++ fluxand metabolism. It thus appears that tran-scription dysregulation may be a key stepin SCA1 disease pathogenesis.


Spinocerebellar ataxia type 2 (SCA2) is anautosomal dominant, progressive cerebel-lar ataxia that accounts for about 13% ofall ADCA cases. Although patients dis-play problems with voluntary coordinatedmovements as in the rest of the SCAs,its main distinguishing clinical feature isextremely slow saccadic eye movements.Other symptoms may include hypore-flexia, myoclonus, and action tremor. Pa-tients suffer a gradual degeneration of


the cerebellum, inferior olive, pons, andspinal cord. Both Purkinje and granulecells degenerate in the cerebellum, andCNS atrophy in some patients can bewidespread. Interestingly, in certain cases,there is involvement of the substantianigra, with such patients displaying aprominent degree of parkinsonism. Thedisorder is caused by a coding (CAG)nexpansion in a novel gene of unknownfunction, named ataxin-2. The SCA2 CAGrepeat displays little polymorphism in thenormal population, with 95% of the popu-lation possessing 22 or 23 repeats in eachallele. The remaining 5% of alleles in un-affected individuals do nonetheless rangefrom 15–31 CAG repeats. Such unaffectedindividuals typically have two CAA inter-ruptions in their CAG repeat tract. AffectedSCA2 patients typically display CAG re-peats numbering from 32–63 triplets, andsuch disease alleles are always uninter-rupted CAG tracts. There is a nonlinearinverse correlation between expansion sizeand age of onset in SCA2, with some stud-ies documenting a paternal transmissionbias for larger expansions.

Both wild-type and expanded ataxin-2mRNA is widely expressed, with high-est levels in the substantia nigra andPurkinje cells of the cerebellum. The140-kDa protein is cytoplasmic and itsfunction remains uncertain. Since it con-tains Sm1 and Sm2 motifs common inproteins involved in RNA splicing andprotein–protein interactions, involvementin RNA processing has been proposed.A yeast two-hybrid screen indicated thatataxin-2 has a binding partner, ataxin-2 binding protein-1 (A2BP1). A2BP1 ishighly conserved throughout the animalkingdom and its expression pattern corre-sponds well with that of ataxin-2. A2BP1contains a domain that is also conservedin RNA-binding proteins, the RNP motif.

Thus, a complex including ataxin-2 andA2BP1 may be involved in RNA processingor metabolism.

The ataxin-2 gene is evolutionarily con-served. The murine ortholog of ataxin-2does not contain a polyglutamine repeattract, however, but instead possesses a sin-gle glutamine residue at the analogous lo-cation. (Absence of a substantial glutaminerepeat tract is typical for mouse orthologsof polyglutamine disease proteins.) In thecase of ataxin-2, study of the Drosophilaortholog (Datx2), which does show tworegions of marked amino acid similar-ity and does contain polyglutamine repeatregions, has yielded some potentially im-portant insights into ataxin-2s normalfunction. Modulation of Datx2 dosage re-sulted in mutant phenotypes whose causecould be traced to aberrant actin filamentformation (Fig. 4). This work suggeststhat alteration of ataxin-2-mediated regu-lation of cytoskeletal structure could affectdendrite formation or other aspects of neu-ronal function in SCA2. As SCA2 is oneof the few polyglutamine disorders to dis-play prominent cytosolic aggregates, sucha model of SCA2 pathogenesis seems plau-sible. In other experiments, eliminatingthe C. elegans ortholog of ataxin-2 (ATX-2) yielded a lethal phenotype, indicatingthat ATX-2 is required for early embryonicdevelopment of this nematode worm. Itremains to be seen how these observationsin worms and flies will apply to ataxin-2 function in mammals, especially sinceno simple or conditional knockout of themouse ataxin-2 gene has been performed.

3.6Spinocerebellar Ataxia Type3/Machado–Joseph Disease

Spinocerebellar ataxia type 3 (SCA3) isthe most common inherited dominant


(a)

(c)

(e) (f)

(d)

(b)

Fig. 4 Studies of the ataxin-2 ortholog in Drosophilamelanogaster reveals a role for Drosophila ataxin-2(Datx2) in actin filament formation duringoogenesis. Egg chambers from normal (Wild-Type(WT)) and mutant flies with reduced expression(Datx2) of Drosophila ataxin-2 were stained withDAPI (blue) and phalloidin (red) to indicate nucleiand filamentous actin respectively (a–d). The eggchambers of WT flies prior to cytoplasmic transport(a) display well demarcated, separated butinterconnected cells (blue) as expected, while theegg chambers of Datx2 flies (b) contain irregularlyarranged cells. After cytoplasmic transport, the eggchambers of WT flies (c) show one greatly enlargedoocyte (dashed line) with only a small section ofcompressed cells, while the egg chambers of Datx2flies (d) have failed to yield an enlarged oocyte,instead retaining dispersed and large adjacent cells.Confocal images of egg chambers prior tocytoplasmic transport stage reveal a prominent

actin filament network in WT flies (e), but a remarkably transparent actin filament network in Datx2flies (f). The decreased density of the actin filament network underlies the cytoplasmic ‘‘dumping’’defect in the Datx2 flies. (From Satterfield, T.F., Jackson, S.M., Pallanck, L.J. (2002) A Drosophilahomolog of the polyglutamine disease gene SCA2 is a dosage-sensitive regulator of actin filamentformation, Genetics 162, 1687–1702, used with permission of Genetics.) (See color plate p. xxiii).

spinocerebellar ataxia worldwide. SCA3 isalso known as Machado–Joseph disease(MJD) because of its initial description ina group of Portuguese residents of theAzores islands. It is a progressive, auto-somal dominant cerebellar ataxia whoseclinical features include ophthalmoplegia,dystonia, dysarthria, and signs of lowermotor neuron disease, such as tongueand facial fasciculations. Degeneration oc-curs in the spinocerebellar tracts, dentatenuclei, red nuclei, substantia nigra, andspinal cord. This disease is unique amongthe SCAs because the cerebellar cortexand inferior olive are largely spared. SCA3is caused by a (CAG)n repeat near the3′ end of the coding region of a novelgene (ataxin-3). Normal alleles range from12–40 CAG repeats while affected indi-viduals carry 55–84 repeats. Unlike manyother triplet repeat disorders, there is a sub-stantial gap between the largest normal re-peat allele and the smallest disease-causing

repeat allele. Some researchers have pro-posed that this could be due to a SCA3founder effect. The presence of SCA3 inevery major racial population worldwide,however, would require the founder to betruly ancient. There is the typical inversecorrelation between repeat number andage of onset in SCA3, and in this disease,paternal expansion bias is characteristic,as documented by repeat sizing of sperm.

Ataxin-3 is a ubiquitously expressed 42-kDa protein, making it the smallest of thepolyglutamine proteins. Ataxin-3 is highlyconserved in eukaryotes, with homologyto ENTH and VHS domain proteinsinvolved in regulatory adaptor functionsand membrane trafficking. Ataxin-3 hasseveral splice variants which reside inthe nucleus and the cytoplasm, and itappears developmentally regulated. Theexistence of NIs was first documented inthe brains of patients with SCA3, andthis feature of polyglutamine-expanded


ataxin-3 remains evident in SCA3 cellculture and mouse models. Ataxin-3 NIsare ubiquitinated and contain numeroustranscription factors.

A number of recent discoveries regard-ing the domain structure and normalfunction of ataxin-3 suggest pathwaysby which polyglutamine repeat expan-sion could result in disease pathogenesis.Comparison of ataxin-3 amino acid se-quences across a wide range of eukaryoticspecies revealed the presence of an ex-tremely highly conserved amino-terminalsequence that was named the josephin do-main. Study of this region indicates that itmay play a role in aggregate formationin concert with the expanded polyglu-tamine tract. A rather intriguing featureof ataxin-3 that was discovered indepen-dently is the presence of multiple ubiquitininteraction motifs (UIMs), and the demon-stration that ataxin-3 is a polyubiquitinbinding protein. This suggests a role forataxin-3 in mediating protein refolding anddegradation.

In addition to a possible normal rolein protein surveillance, other studieshave found ataxin-3 directly interacts withthe histone acetyltransferases CBP andp300, and can block histone acetyltrans-ferase activity by inhibiting access ofsuch coactivators to their histone sub-strates. This appears to be mediatedby the interaction of ataxin-3 with hi-stones. In vitro and in vivo studies ofataxin-3 further revealed an interactionwith histones and the chromatin remod-eling machinery that led to a repres-sion of transcription activation. Thus,polyglutamine-expanded ataxin-3 may alsoaffect transcriptional processes once itbegins to accumulate in the nuclear com-partments of the cell types where it isexpressed.


Spinocerebellar ataxia type 6 (SCA6) isan autosomal dominant, slowly progress-ing cerebellar ataxia that accounts for∼10–20% of ADCA worldwide. It ischaracterized predominantly by cerebel-lar dysfunction that may have an episodiccomponent. Other common features mayinclude dysarthria, nystagmus, loss of vi-bration sense and proprioception, andimbalance. Histopathological changes in-clude loss of Purkinje cells, cerebellargranule neurons, and neurons in the den-tate nucleus and inferior olive. SCA6 iscaused by a coding (CAG)n expansion inexon 47 of the gene CACNA1A, whichencodes the α1A subunit of the P/Q-typevoltage-gated calcium channel. This geneis located on chromosome 19 at bandp13. The SCA6 CAG repeat is small com-pared to other polyglutamine disorders,with a pathogenic range of only 19–33triplets. Unaffected individuals carry alle-les of 4–18 repeats. There is an inversecorrelation between repeat length and ageof disease onset, and minimal intergen-erational and somatic instability has beenreported for the SCA6 repeat expansion inaffected patients.

The 9.8 kb CACNA1A transcript impli-cated in SCA6 is expressed throughoutthe CNS, and most highly in cerebellarPurkinje cells and granule neurons. Theα1A subunit is the pore-forming compo-nent of the channel, which is importantin neurotransmitter release at the synapse.Polyglutamine expansions in this subunitcause variable changes in calcium trans-mission rates, depending on the systemand the β subunit coexpressed. Expansionsdo not cause a reduction in membranechannel density in HEK293 cells, sug-gesting that aggregation does not occur


at pathogenic repeat lengths. Ubiquitin-negative neuronal inclusions are visible inthe cytoplasm of SCA6 patient Purkinjecells, however.

In addition to displaying a disease al-lele range that does not overlap with theother polyQ diseases, there are severalother reasons to suspect that SCA6 is dif-ferent from the rest of the polyglutaminerepeat group. The first difference is theexistence of two disorders, episodic ataxiatype 2 (EA2) and familial hemiplegic mi-graine (FMH), that are both allelic to SCA6,as both are caused by point mutations inthe CACNA1A gene. EA2 resembles SCA6as affected EA2 individuals suffer from aslowly progressive form of episodic ataxia,accompanied by nystagmus, dysarthria,loss of balance and sometimes cerebellaratrophy. EA2 is usually caused by trun-cation mutations in CACNA1A, resultingin loss-of-function of the calcium chan-nel. CACNA1A knockout mice similarlydevelop ataxia and late-onset cerebellaratrophy, characteristic of SCA6 and EA2patient phenotypes. Another reason SCA6is different from other polyQ disorders isthat the CACNA1A protein probably doesnot undergo a structural change that con-verts it into an aggregate-prone, beta-sheetadopting, amyloid-like conformer. Indeed,even the largest SCA6 polyglutamine tractis below the threshold required for stableβ-pleated sheet formation in other polyg-lutamine disorders. Consistent with thisprediction, channel localization of mutantpolyglutamine-expanded CACNA1A pro-tein is not affected in cell culture models,and its channel function is not com-pletely abolished. Cytoplasmic aggregatesin patient material are ubiquitin-negative,suggesting that the protein may not begrossly misfolded. This evidence supportsa model whereby a dominant-negative lossof function of the P/Q-type voltage-gated

calcium channel due to association of themutant α1A subunit with other subunitscauses SCA6. The coincidence that this dis-order is caused by a polyQ tract expansionand causes a progressive cerebellar ataxiacannot be ignored, however, more data willbe required to soundly refute a possibleconcomitant toxic gain-of-function effect.


Spinocerebellar ataxia type 7 (SCA7) is anautosomal dominant, progressive cerebel-lar ataxia. It is unique among autosomaldominant SCAs, as patients typically de-velop visual impairment in addition totheir cerebellar ataxia. The visual impair-ment is due to a cone-rod dystrophy thatresults in retinal degeneration. Patientsfirst develop problems distinguishing col-ors, but ultimately go blind in this typeof retinal degeneration. SCA7 patientsmay present either with cerebellar ataxiaor visual impairment. The likelihood oftheir presentation is dictated by the sizeof their CAG repeat disease allele, withlarger repeats typically favoring presen-tation with visual impairment. Affectedindividuals display prominent dysarthria,and can develop increased reflexes, de-creased vibration sense, and oculomotordisturbances. Neuronal degeneration andreactive gliosis occur in the cerebellar cor-tex, dentate nucleus, inferior olive, pontinenuclei, and occasionally in the basal gan-glia. NIs are widespread. Infantile-onsetSCA7 has been documented, and in thisfatal form of the disease, nonneuronal tis-sues such as the heart and the kidneyare severely affected. SCA7 is caused bya highly polymorphic (CAG)n repeat inthe 5′ coding region of the ataxin-7 gene.Unaffected individuals carry 4–35 repeats,


while affected individuals carry 37–306 re-peats. The SCA7 trinucleotide repeat is oneof the most unstable of all polyglutaminedisease genes, sometimes expanding by asmuch as 250 repeats in a single generation.There is a pronounced paternal expansionbias, with large expansions occurring inmale germ cells, frequently causing em-bryonic lethality that results in reducedtransmission. Marked repeat instabilitycan occur in the brain, and produce largesomatic expansions on occasion.

Ataxin-7 is a ubiquitously expressedprotein of 892 amino acids. It is expressedmost highly in heart, skeletal muscle,and pancreas. Expression levels within theCNS are highest in the cerebellum andbrainstem. One splice variant, ataxin-7b, isexpressed predominantly in the CNS (32).Ataxin-7 contains a functional arrestindomain, a protein interaction domainthat is highly selective for phosphorylatedforms of its interacting protein(s). Thissuggests that it interacts with specificphospho-proteins, although none havebeen discovered to date. It also containstwo SH3 domains, which are proteininteraction domains that bind proline-richsequences and mediate a number of cellsignaling processes. Ataxin-7 also containsthree putative NLSs and one NES.

Ataxin-7 is conserved throughout eu-karyotes, and its yeast ortholog SGF73 ispart of a multisubunit histone acetyltrans-ferase complex called SAGA (Spt/Ada/Gcn5 acetyltransferase). The human or-thologs of SAGA comprise the so-calledSTAGA (SPT3/TAF9/ADA2/GCN5 acetyl-transferase) complex, and are essen-tial transcription coactivators requiredfor the transcription of certain genes.STAGA components immunoprecipitatewith ataxin-7. Although pathogenic expan-sion of ataxin-7 does not alter its ability tobe integrated into the STAGA complex, the

presence of the polyglutamine-expandedataxin-7 has a dominant-negative effectupon the GCN5 histone acetyltrans-ferase activity of the STAGA complex,resulting in transcription dysregulation.The transcription dysregulation causedby polyglutamine-expanded ataxin-7 likelycauses a disease phenotype by altering theability of certain transcription factors toactivate expression of their target genes.

The best characterized example of tran-scription dysregulation by polyglutamine-expanded ataxin-7 is its interference withthe cone-rod homeobox protein (CRX), aglutamine domain containing transcrip-tion factor expressed only in the retinaand the pineal gland. Ataxin-7 interactsdirectly and functionally with CRX, ac-cording to studies performed in vitroand in a mouse model of SCA7. Impor-tantly, the interaction between ataxin-7and CRX appears to involve the glu-tamine tract regions found in both pro-teins. Autosomal dominant mutations inCRX can cause a cone-rod dystrophyin humans, further supporting a modelin which CRX’s diminished transactiva-tion competence is central to the SCA7retinal degeneration phenotype. Severaltranscription factors, including CBP, canbe found in SCA7. CRX may be butone of a number of transcription factorswhose function is diminished by polyQ-expanded ataxin-7 interaction and dys-regulation of STAGA complex-mediatedgene expression.

One intriguing feature of SCA7 wasdiscovered upon generation of transgenicmice expressing ataxin-7 with the mouseprion protein promoter. This promoterdrives expression in every tissue, with theoccasional notable exception of the Purk-inje cells of the cerebellum. Despite lack ofPurkinje cell expression of ataxin-7, mice


Fig. 5 Noncell autonomous Purkinje celldegeneration in a mouse model of spinocerebellarataxia type 7 (SCA7). Confocal microscopy analysisof cerebellar sections from a SCA7 transgenicmouse (SCA7) created with an ataxin-7 CAG-92containing murine prion protein expression vectorand from an age- and sex-matched nontransgeniclittermate (Control). Staining with an anti-ataxin-7antibody (magenta), a calbindin antibody (green),and DAPI (blue) reveals a healthy, normal-appearing cerebellum characterized by properlyoriented Purkinje cells with extensive dendriticarborization in the ‘‘Control’’ mice. However, SCA7transgenic mice display pronounced Purkinje celldegeneration as evidenced by decreased dendriticarborization and displacement of Purkinje cellbodies. Interestingly, although numerous neuronsin the granule cell layer (GCL) and the molecularlayer (ML) display aggregates of ataxin-7, there

is no accumulation of mutant ataxin-7 in the degenerating Purkinje cells due to lack of appreciableexpression there. As the Purkinje cells degenerate without expressing the mutant protein, thedegeneration is described as noncell autonomous. (Adapted from Garden, G.A., Libby, R.T., Fu, Y.H.,Kinoshita, Y., Huang, J., Possin, D.E., Smith, A.C., Martinez, R.A., Fine, G.C., Grote, S.K., et al. (2002)Polyglutamine-expanded ataxin-7 promotes noncell-autonomous Purkinje cell degeneration anddisplays proteolytic cleavage in ataxic transgenic mice, J. Neurosci. 22, 4897–4905, used withpermission of the Journal of Neuroscience.) (See color plate p. xxiv).

developed a cerebellar ataxia accompa-nied by degeneration of the Purkinje cells(Fig. 5). This noncell-autonomous degen-eration may point to a disease mechanisminvolving withdrawal of trophic supportby communicating neurons (olivary, deepcerebellar, brainstem, or granule neurons)or dysfunction of glutamate transportersexpressed by surrounding glia. Damage toinferior olivary neurons or Bergmann gliacan indeed cause the degeneration of Purk-inje cells, lending credence to this theory.

Another interesting feature of SCA7 thatis common to other neurodegenerativedisorders is the prominence of morpholog-ical and functional degeneration withoutpronounced apoptosis. Some neurons inSCA7 humans and mouse models ex-hibit indentations in the nuclear enve-lope, reduced arborization, ectopy, andincreased autophagy. (Autophagy is thebulk degradation of cellular components

by a membrane-bound autophagosomeand is accelerated by a number of cellularinsults.) Caspase-3, a proteolytic enzymeclassically associated with apoptosis, is ac-tivated at abnormally high levels in SCA7patient brains. This may suggest thatcaspase activation, rather than causingapoptosis, is contributing to a nonapop-totic degenerative process.

In normal human neurons, ataxin-7 isvariably located in the nucleus or thecytoplasm. In SCA7 patients, ataxin-7 grad-ually undergoes a shift in localization intoNIs. This process has been replicated inSCA7 transgenic mice, in which ataxin-7immunoreactivity shifts from the cyto-plasm to the nucleus, and ultimately formsfoci there. In patient’s brains, these NIscolocalize with promyelocytic leukemia(PML) protein, which is an integral partof nuclear bodies (NBs). NBs are associ-ated with transcription regulation and the


ubiquitin-proteasome pathway; thus, accu-mulation of ataxin-7 in NBs may representan attempt by the cell to degrade misfoldedprotein.


Spinocerebellar ataxia type 17 (SCA17)is the most recently discovered polyglu-tamine repeat disease. It is an autosomaldominant, progressive disorder character-ized by dementia as well as cerebellar ataxiaand involuntary movement abnormalities.Its symptoms are diverse and heteroge-neous, but typically begin with behavioraldisturbances and cognitive impairment.This is followed by ataxia, rigidity, dysto-nia, hyperreflexia, and rarely parkinson-ism. Neuropathologically, patients maysuffer degeneration of the cortex, cere-bellum (including Purkinje cells), inferiorolive, caudate nucleus, and medial tha-lamic nuclei. SCA17 is caused by theexpansion of a coding (CAG)n repeat inthe TATA-binding protein (TBP) gene onchromosome 6q27. Unaffected individualscarry 25–48 repeats, while affected individ-uals carry 43–66 repeats. The area of over-lap between affected and unaffected allelesindicates incomplete penetrance at inter-mediate repeat lengths. The pathogenicthreshold is higher than for most otherpolyglutamine disorders, yet there is aninverse correlation between repeat lengthand age of onset.

TBP is a ubiquitously expressed tran-scription initiation factor that is a corecomponent of the RNA polymerase IItranscription factor D (TFIID) complex.TBP possesses DNA binding activity in theTFIID complex, and is therefore requiredfor the transcription of numerous genes.SCA17 patients have NIs immunoreactivefor TBP, polyglutamine, and ubiquitin.

Given TBP’s central role in transcriptionregulation, the basis of SCA17 diseasepathogenesis is proposed to involve tran-scription dysregulation, but this is yet tobe proven.

3.10Role of Aggregation in PolyglutamineDisease Pathogenesis

In most polyglutamine disorders and inmany neurodegenerative diseases in gen-eral, protein aggregation is a prominentfeature. It occurs in almost all polyglu-tamine diseases, despite the lack of domainor structural similarity between the dif-ferent disease proteins. Aggregates havelong been considered reliable indicatorsof disease, although their pattern and on-set often do not correspond well with thecell-type specificity of disease pathology.As the role of aggregation in pathogenesishas been one of the most hotly debatedissues in the field of neurodegeneration,it may have implications for Alzheimer’sdisease, Parkinson’s disease, and amy-otrophic lateral sclerosis. Why? There areseveral characteristics of polyglutamine-mediated aggregation that are presumablycommon to all of the disorders that showaggregation. First of all, aggregates arerich in β-pleated sheets and have manyof the properties of amyloid. This is sup-ported by Congo red birefringence andimmunoreactivity with antiamyloid anti-bodies. Various studies have shown thatthe conformational change is associatedwith the production of visible aggregatesrather than in the soluble nonaggre-gated phase, where polyglutamine tractsare thought to remain in a random coilconformation.

In vitro, the kinetics of aggregation areconsistent with nucleated-growth polymer-ization, in which the rate-limiting step is


the formation of misfolded peptide as-semblies, often referred to as oligomers.A mechanism involving nucleated-growthpolymerization would predict that only one(or a few) visible aggregates would appearin each affected cell, and this appears tobe the case for most polyglutamine dis-orders. In the cellular milieu however,where membranes and intermolecular in-teractions compartmentalize proteins, thisprediction may not be valid. The dis-covery of polyglutamine microaggregatesconfirms this suspicion. The thresholdfor aggregation of polyglutamine tracts issimilar to the disease threshold for mostpolyglutamine diseases. In vitro, longerpolyglutamine tracts have a lower con-centration threshold for aggregation andnucleate more quickly than shorter tracts.This is one possible explanation for the cor-relation between tract length and diseaseonset and progression.

Some data support a role for aggregationin polyglutamine disease pathogenesis.Some proponents of this theory invokethe fact that aggregates sequester manyproteins besides the disease protein. Inter-molecular interactions between transcrip-tion factors often involve glutamine tractsor glutamine-rich regions. Some studieshave shown that normal proteins with glu-tamine tracts or glutamine-rich regionsare enriched in polyglutamine aggregatesin cell culture and animal models. Sev-eral transcription factors colocalize withsuch aggregates, including CBP, TBP, andnumerous TBP-associated factors (TAFs).CBP is responsible for the prosurvivaleffects of BDNF, and its soluble concen-tration is lowered in HD patient’s brains.Postnatal mice lacking CREB and its ho-molog CREM develop a progressive degen-eration of the hippocampus and striatum.

Titration of enzymes and factorsrequired for protein refolding and

degradation away from the soluble phaseand into aggregates has been proposed asa potential cause of cell toxicity in neuronswith aggregates. Caspase activation isanother way in which aggregation could belinked to pathology. Caspase recruitmentinto aggregates can lead to their activation,which could result in dysfunction or celldeath in neurons. Further supportingthe role of aggregates in polyglutamineprotein toxicity, injection of preformedpolyglutamine aggregates into the nucleiof cultured cells causes cell death,while the injection of nonpolyglutamineaggregates does not. At the sametime, there is strong evidence thatsoluble polyglutamine protein, ratherthan aggregates, is the primary sourceof toxicity. Importantly, the pattern ofaggregates observed in human patientsoften does not correlate with the patternof neuronal dysfunction. For example,in the striatum of HD patients, largeinterneurons contain aggregates morefrequently than medium spiny neurons,yet the former neurons are largely spared,while the latter are most vulnerable.HD transgenic mice expressing full-lengthmutant htt will develop inclusions inmany brain regions many of which areneuropathologically unaffected, while fewinclusions are detected in the striatum,the region of the brain displaying the mostprominent pathology.

Aggregation and toxicity have been di-rectly dissociated in other polyglutaminedisease model systems. Studies of theSCA1 knockin mouse model revealed thatneurons lacking aggregates were more sus-ceptible to dysfunction and demise, whilethose neurons displaying prominent ag-gregates were protected. Similarly, SCA7transgenic mice exhibit retinal pathologybefore the occurrence of visible aggregates.Finally, in a very provocative study, SCA1


transgenic mice lacking the ubiquitin lig-ase E6-AP were significantly less capableof forming large and numerous aggregatesin neurons in comparison to SCA1 trans-genic mice on a wild-type background,but displayed earlier onset of an ataxicphenotype and accelerated neurodegener-ation. This suggests that the aggregatesmay be protective, although the toxicityof compromising the proteasome mayhave contributed to the accelerated phe-notype. Thus ensued a contentious debateover the role of aggregate formation inpolyglutamine disease – with some work-ers espousing the view that aggregateswere responsible for disease pathology,others suggesting that the aggregationprocess was a protective coping mecha-nism of the cell and thereby beneficial,and still others insisting that aggregateswere incidental and irrelevant. This de-bate was complicated by the existence of‘‘microaggregates,’’ small clumps of aggre-gated protein visible only at the electronmicroscope level.

Over the last few years, studies decon-structing polyglutamine tract aggregationinto a multistep process suggest a par-simonious explanation for the divergentviews. Using a variety of biophysical ap-proaches, including transmission electronmicroscopy (TEM), Fourier transform in-frared spectroscopy (FTIR), and atomicforce microscopy (AFM), one study dis-sected the process of huntingtin (htt)exon 1 peptide aggregation and foundevidence for a number of sequential mor-phological and structural intermediates(Fig. 6). By showing that misfolded httexon 1–44Q adopted intermediate struc-tures, such work opened up the possibilitythat intermediates (not visible at the lightmicroscope level) are the toxic species andthat the ultimate visible aggregated formsof htt exon 1–44Q are neutralized versions

of mutant protein. To examine the role ofaggregate formation in polyglutamine tox-icity, another group then tracked survivaltime versus diffuse, soluble htt protein ex-pression levels in htt-transfected striatalneurons undergoing aggregate formation.Comparison of neurons that developed ag-gregates over time versus those that didnot confirmed that the level of solublenonaggregated mutant polyglutamine pro-tein was the more reliable predictor ofcellular toxicity. Such studies have shiftedour attention to the role of the intermedi-ates (oligomers, protofibrils, etc.) as thetoxic species instead of the final, visi-ble aggregates.

A reasonable model for polyglutaminetoxicity predicts that the process startswith a protein that misfolds because ofthe presence of an expanded polyglu-tamine tract. The misfolded protein isinitially detectable in the soluble phasedue to the cell’s ability to maintain the pro-tein in a properly folded state and directit to the degradation machinery. How-ever, the refolding capacity of the cell isultimately exceeded, and since the degra-dation machinery cannot turnover themisfolded mutant polyglutamine protein,accumulation occurs. Misfolded polyglu-tamine proteins can spontaneously changetheir structural properties and adopt ab-normal conformations. These abnormallyfolded proteins can then nucleate, formingoligomers. Oligomers then form protofib-rils that grow into fibrils. The transitionto the fibril stage is characterized by theattainment of a β-sheet structure, so atthis point the structures are amyloid-like.Fibrils then grow into fibers (also knownas microaggregates), which then assembleinto aggregates visible under a light micro-scope. According to such a model, blockingan intermediate step could be therapeu-tically effective. Consistent with this, one


(b)(a)

(d)(c)

(f)(e)

(h)(g)

Control 2 h

3 h

6 h5 h

7 h 8 h

4 h

Fig. 6 Polyglutamine-expandedhuntingtin protein undergoes a varietyof structural alterations on its way tobecoming a visible aggregate. In thisexperiment, transmission electronmicroscopy tracked the structure ofhuntingtin exon 1 peptide with 44glutamines after release from a linkedaffinity tag. (a) Prior to affinity tagrelease, nothing is observed. (b-c) Afteraffinity tag release, globular structuresbecome apparent. (d) This is followedby formation of short fibers (4–5 nm indiameter). (e-f) Fiber formationbecomes more prominent. However,during this time, large globularassemblies remain (arrowheads). (g-h)Eventually, only fibers are present, andthe fibers begin to adopt a uniformappearance, suggesting consolidationinto protofibrils. (From Poirier, M.A.,Li, H., Macosko, J., Cai, S., Amzel, M.,Ross, C.A. (2002) Huntingtin spheroidsand protofibrils as precursors inpolyglutamine fibrilization, J. Biol. Chem.277, 41032–41037, used withpermission of the Journal of BiologicalChemistry.).

group has nicely shown that Congo red canbind polyglutamine peptides once they areamyloid-like and prevent their conversioninto fibers, while an independent grouphas shown that Congo red delivery to anHD mouse model is a highly effective treat-ment intervention.

One satisfying aspect of this model isthat it allows us to simultaneously viewaggregates as harmful, protective, and in-nocuous. How? First, aggregates clearlymust be toxic as their creation is predicatedupon the production of earlier intermedi-ate toxic forms. At the same time, thefinal visible aggregates are less toxic thanthe earlier intermediates, so anything that

enhances their sequestration into visi-ble aggregates is beneficial. Finally, sinceoligomers are difficult to detect, it is of-ten not possible to know whether cells aresuccessfully sequestering the toxic precur-sors into aggregates or if high levels oftoxic intermediates are building up. So,aggregates are thus also incidental, sincetheir presence does not provide us withinsight into the steady state levels of thetoxic precursors. In conclusion, advancesin our understanding of the role of aggre-gate formation in polyglutamine diseaseprocesses suggest that aggregates may besimultaneously viewed as harmful, benefi-cial, and incidental.


3.11Protein Context

The importance of protein context in thepathogenesis of polyglutamine disordersis a subject of great interest to manyresearchers, because it is directly relatedto the mechanisms of toxicity in eachdisease. It is evident in the polyglutaminedisorders that protein context plays arole in cell-type specificity. For example,a polyglutamine expansion in the Httprotein causes a severe degenerationof the striatum and cortex, while thesame size glutamine tract in ataxin-1 causes a degeneration of structuresin the cerebellum and brainstem, whilesparing the striatum and cortex. Thiscannot be attributed to gross differencesin expression patterns, since both proteinsare pan-neural (Fig. 7).

One effect that protein context may haveon pathology is to affect subcellular local-ization. For example, the presence of afunctional NLS or NES dictates preferen-tial subcellular localization in the nucleusor cytoplasm. Mice expressing expanded

ataxin-1 with a mutated NLS do not developthe SCA1 phenotype, while those withoutmutated NLSs do. Interaction domainsalso affect polyglutamine protein toxic-ity. The ability of expanded polyglutamineproteins to interact with other moleculesis a promising avenue in the search formechanisms of cell-type specificity. Yeasttwo-hybrid screens and other techniqueshave yielded interaction partners for anumber of polyglutamine proteins. Forexample, Htt’s interactions with transcrip-tion factors may prove to be central to itspathological effects. At the same time, it isalso evident that polyglutamine tracts areinnately toxic. Pure polyglutamine tractscause toxicity in cell culture. Mice thatexpress a glutamine tract of 150 aminoacids in the Hprt protein, which does notnaturally contain any such tracts, exhibitprogressive neurological deficits and NIs.Thus, it appears that pathology in eachpolyglutamine disease is due to a gain offunction of the glutamine tract that is thenmodulated by the protein context in whichit resides.

Fig. 7 The conundrum of cell-typespecificity in the polyglutaminediseases. Although the differentpolyglutamine disease proteins areexpressed throughout the centralnervous system, only select populationsof neurons degenerate in the differentdisorders. The principal regions ofselective vulnerability are shown forcertain of the polyglutamine diseases.

Central nervous system:

Cerebralhemisphere

Brain stem

Spinal cord

Brain

Huntington's disease

Spinal muscular atrophy

Spinocerebellar ataxias


3.12Transcriptional Dysregulation

Transcriptional dysregulation is one the-ory of polyglutamine toxicity that hasgained support from many lines of evi-dence in several disorders, and it appearsincreasingly likely that it is one of thefundamental causes of pathogenesis. Mi-croarray technology and other genomictechniques are facilitating rapid advancesin this area of the polyglutamine field. Themajority of polyglutamine disorders arecaused by proteins that are either transcrip-tion factors/cofactors or interact closelywith transcription factors. TBP and AR aretranscription factors, atrophin-1 is a tran-scriptional corepressor, ataxin-7 is part ofa transcriptional coactivator complex, andataxin-1, htt, and ataxin-3 all interact withvarious transcription factors. It is thereforereasonable to suspect that an abnormallylong polyglutamine tract could interferewith the transcriptional activity of theseproteins and/or their interactors.

CBP is one of the most studied transcrip-tion factors in the polyglutamine field.It is a transcription activator that me-diates part of the cellular response tocAMP, and it can interact with numer-ous polyglutamine proteins in the solubleor insoluble phase, including huntingtin,ataxin-3, ataxin-7, and AR. Expanded httrepresses CBP-regulated genes, and over-expression of CBP causes a considerablerescue of the cell death phenotype in HDand SBMA cell culture models. Postna-tal mice lacking CBP’s upstream activatorCREB and its homolog CREM display aprofound degeneration of the striatumand hippocampus. Thus, interference withCBP appears to be a common theme inpolyglutamine pathology.

Another common theme in the tran-scription dysregulation equation is the

tendency for the polyglutamine dis-ease proteins to alter transcriptionfactor/coactivator-mediated histone acetyl-transferase (HAT) activity. The ability ofa gene to be transcribed depends uponits chromatin structure, and thus uponthe degree of histone acetylation in itsvicinity. This is because acetylated hi-stones cause chromatin to be in an‘‘open,’’ transcription-friendly conforma-tion. Acetylation status depends upon thebalance between the activity of HATsand their countervailing counterparts,the HDACs. CBP, p300, and p300/CBP-associated factor (PCAF) are all HATs thatare inhibited by polyglutamine-expandedhtt exon 1 peptide and ataxin-3. In cell cul-ture, expression of a mutant htt fragmentreduces global histone acetylation.

Finally, according to a very recent study,ataxin-7 directly interacts with GCN5 aspart of the STAGA complex, and uponpolyglutamine expansion, mutant ataxin-7 causes a dominant-negative effect uponGCN5 HAT activity. This results in CRXtranscription interference that may con-tribute to the SCA7 retinal degenerationphenotype. Inhibiting HDACs with drugsknown as HDAC inhibitors (HDAC Is) at-tenuates the phenotype of HD and SBMAin mouse and fly models, presumably byshifting the cells’ acetylation status. HDACIs such as sodium butyrate and especiallysuberoylanilide hydroxamic acid (SAHA)have thus emerged as possible candidatesfor therapeutic trials in human polyglu-tamine disease patients.

3.13Proteolytic Cleavage

The occurrence of proteolytic cleavage inpolyglutamine diseases first became ap-parent when it was shown in HD thathtt can be cleaved by extracts derived


Fig. 8 Proteolytic cleavage in the polyglutaminediseases. Production of amino-terminal truncatedproteins is observed in many of the polyglutaminediseases, as is shown here for SCA7. In thisexperiment, nuclear fractions of retinal lysatesfrom age- and sex-matched nontransgenic (nt),ataxin-7 CAG-24 (24Q), and ataxin-7 CAG-92 (92Q)mice were prepared and immunoblotted with anantibody directed against the amino-terminalregion of the protein ataxin-7. As shown here, inaddition to soluble full-length ataxin-7 (which istypically reduced in cells where it is aggregatinginto insoluble inclusions), an ∼50–60 kD

Ataxin-7 92QAtaxin-7 24Q

55-kD fragment

148

98

64

50

36

22

(kD) nt 24Q 92Q

fragment is detected. (Adapted from Garden, G.A., Libby, R.T., Fu, Y.H., Kinoshita, Y., Huang, J.,Possin, D.E., Smith, A.C., Martinez, R.A., Fine, G.C., Grote, S.K., et al. (2002) Polyglutamine-expanded ataxin-7 promotes noncell-autonomous Purkinje cell degeneration and displays proteolyticcleavage in ataxic transgenic mice, J. Neurosci. 22, 4897–4905. Used with permission of the Journal ofNeuroscience.).

from apoptotic cells. Subsequent publi-cations showed that only amino-terminalepitopes of htt protein are detectable in ag-gregates. SBMA, DRPLA, and SCA7 werethen added to the list of diseases in whichaggregates are only immunoreactive fora glutamine-containing fragment of thedisease protein. Since then, evidence ofproteolytic cleavage has been published fornearly all known polyglutamine disorders.Experimental studies revealed that inhibi-tion of caspase cleavage reduces aggregateformation and toxicity in a cell culturemodel of HD, underscoring the relevanceof proteolytic cleavage in polyglutaminedisease pathogenesis. Cleavage promotesaggregation and/or toxicity in SCA3 andSCA7 (Fig. 8). Polyglutamine tract con-taining htt fragments are more toxic in cellculture than full-length htt, and the mostwidely used HD mouse model expressesonly exon 1 of the htt gene.

There are several mechanisms by whichcleavage may modulate polyglutamineneurotoxicity. Abnormal proteolysis ofpolyglutamine proteins may lead to a toxicspecies that can cause damage in thesoluble or insoluble phase. Alternatively,

proteolysis may be a normal event in pro-tein turnover and the inability to clearcleaved protein may be the problem.Another possibility is that soluble polyglu-tamine proteins may cause the activationof proteases such as caspases, which thencleave the disease protein and send the cellon a path to degeneration and ultimatelyapoptosis. As we will see, each theory hassupport and not all are mutually exclusive.

Abnormal proteolysis of polyglutaminedisease proteins is one possible mecha-nism of toxicity. Some studies have shownthat polyglutamine repeat length modu-lates susceptibility to proteolytic cleavage,but results have been inconclusive. Fewstudies have been performed using hu-man brain tissue, however. Considerableevidence exists to show that polyglutaminetracts themselves are resistant to degrada-tion by mammalian proteasomes. The 20Sand 26S eukaryotic proteasomes are in-capable of cutting within polyglutaminetracts longer than 9 amino acids in vitro.The inability of the proteasome to di-gest glutamine tracts may contribute tothe toxicity and aggregation of expandedpolyglutamine peptide fragments.


Proteolysis may be a normal part of theturnover of wild-type and mutant htt. Onegroup published evidence that caspase-3cleaves both mutant and wild-type htt inthe cytoplasm of HD brains, suggestingthat this particular cleavage is a normalevent in its turnover. Since this fragmentwas located exclusively in the cytoplasm, italso implies that further cleavage occursbefore the nuclear translocation of htt.Another study suggested that pepstatin-sensitive aspartyl endopeptidases such asthe presenilins and cathepsins D and Emay be involved in the normal turnoverof htt. This process, which may normallyaid in maintaining the balance between httsynthesis and degradation, could generatea toxic, highly aggregation-prone (andhence intermediate oligomer/protofibril-prone) species when it cleaves mutant htt.

Another possible source of toxic cleav-age products is the ubiquitin-proteasomesystem, which is responsible for labelingand digesting misfolded proteins, amongother things. Polyglutamine tracts beyonda threshold of about 35 glutamines adoptan abnormal β-pleated sheet conforma-tion, which may cause the host proteinto be ubiquitinated and thus targeted fordegradation. In support of this theory,a common feature of polyglutamine dis-eases is the presence of ubiquitin-positiveinclusions. Arguing against this idea isan experiment showing that in a cell cul-ture model of SCA1, increased proteasomedegradation due to the introduction ofa degradation signal reduces aggregatesand toxicity. Another potential source oftoxic fragments is the autophagy pathway,which is responsible for the bulk degrada-tion of cellular components. According toone study, degradation of htt by autophagyand the autophagosome-associated cathep-sin D creates toxic htt fragments.

There are many proteases that are sus-pected to play a role in the proteolyticcleavage of polyglutamine proteins, themost studied of which are the caspases.Caspases are cysteine- dependent aspartylproteases that play a crucial role in apop-tosis, but are increasingly being inves-tigated for their nonapoptotic functions.In healthy cells, caspases are present pre-dominantly as inactive proenzymes, whichmust be cleaved for full activity. Caspase-3is the only protease that has been shownto cleave htt in HD patient’s material, al-though caspases 1 and 6 have also beenimplicated in cell culture and in vivo.Inhibiting the cleavage of htt by any ofthese three proteases is protective, and anHD mouse model expressing dominant-negative caspase-1 showed significantlydelayed disease onset. Caspase-3 has alsobeen implicated in the proteolytic cleavageof atrophin-1, ataxin-3 and AR. Caspasesare activated in response to a number ofcellular insults, and in postmitotic cellssuch as neurons where inhibitors of apop-tosis proteins (IAPs) are highly expressed,this may cause cellular damage that doesnot result in classical apoptosis.

Calpains are another family of pro-teases implicated in the cleavage ofpolyglutamine proteins. They are calcium-activated cysteine proteases that exist pre-dominantly as proenzymes. As mousemodels of HD and material from HD pa-tients exhibit abnormal mitochondrial cal-cium regulation, some investigators haveproposed that altered calcium flux is themechanism of calpain activation. Indeed,several cell culture studies support the roleof calpain proteases I, II, and ‘‘m’’ in httcleavage. As mentioned previously, otheraspartyl proteases such as the Alzheimer’sdisease-associated presenilins and auto-phagy-associated cathepsins have alsobeen implicated in htt cleavage.


4Type 2: the Loss-of-function RepeatDiseases

4.1Fragile X Syndrome

Fragile X syndrome (FRAXA) is an X-linked disorder with a prevalence of 1in 4000 in males and 1 in 8000 in fe-males, making it the most common formof inherited mental retardation. Neuro-logical presentation frequently includesmild to severe mental retardation, hy-peractivity, poor eye contact, high-pitchedspeech, and flapping or biting hand move-ments. Physical signs in males includelong, prominent ears, and jaws, macro-cephaly, postpubescent macroorchidism,and occasionally, connective tissue abnor-malities. The mutation responsible forFRAXA is typically an expansion of apolymorphic (CGG)n repeat found in the5′-untranslated region of the fragile X men-tal retardation-1 (FMR1) gene. Expandedchromosomes have a folate-sensitive frag-ile site at Xq27.3 that can be viewed undera light microscope under special cell cul-ture conditions. Normal alleles contain6–53 triplets punctuated by one or moreAGGs, which are considered to have astabilizing influence on the repeat. Dis-ease alleles contain expansions beyond 200and up to 2000 repeats, with no AGG in-terruptions. Pathogenic expansions resultexclusively from maternal transmission.FRAXA appears to be a loss-of-functiondisorder, since deletion of FMR1 andloss-of-function point mutations can alsocause FRAXA. This view is further sup-ported by the FMR1 knockout mouse,which reproduces certain aspects of thehuman disorder.

FRAXA patients have reduced levelsof the FMR1 gene product, FMRP, and

there is a linear correlation between re-duced protein levels and IQ test scores.Expansion of the disease allele resultsin hypermethylation of the (CGG)n tract,which spreads to a nearby CpG island inthe FMR1 promoter region. Some of theproposed secondary structures formed bythe FRAXA repeat contain C–C mispairs,which are good targets for human DNAmethyltransferase. Another theory invokesthe RNAi protein Dicer’s ability to cleaveCGG repeat RNA, postulating that the re-sulting siRNAs recruit DNA methyltrans-ferases to the FMR1 locus. Affected allelesalso display condensed chromatin, loss ofhistone acetylation and increased histonemethylation. These data support a processwhereby DNA methylation recruits tran-scription silencing machinery, which sub-sequently suppresses FMR1 transcriptionin the nearby promoter region. Interest-ingly, while premutation carriers had beenviewed as perfectly normal for decades,recent work indicates that some premuta-tion carriers (60–200 uninterrupted CGGrepeats) display a phenotype distinct fromFRAXA. Females have a predispositionto premature ovarian failure (POF), andmales may develop late-onset ataxia andtremor with the presence of neuronal in-tranuclear inclusions (NIs) that consistof RNA. (This new FXTAS is discussedin a separate section.) Expression stud-ies have shown that premutation carriersmay express up to seven times moreFMR1 mRNA than normal individuals.This upregulation is probably not due tocompensation for loss of function, be-cause individuals with point mutationsresulting in loss of function of FMRP donot have higher levels of the transcript.Abnormal transcript levels due to premu-tations or full mutations are thought tobe as a result of the expansion’s effect


on transcription initiation rather than onmRNA stability.

The FMR1 promoter lacks a functionalTATA box and initiator sequence. Fourfunctional transcription factor bindingsites have been identified in normal in-dividuals: two sites binding USF1/USF2and Nrf-1, and two GC boxes that bindmembers of the ‘‘Sp’’ family of transcrip-tion factors. Sp1 and Nrf-1 binding isdisrupted in cells from FRAXA patients.The FMR1 gene encodes the widely ex-pressed FMRP, which is an RNA-bindingprotein most highly expressed in the brainand testes. It is 60% homologous to twoother proteins, FXR1P and FXR2P, withwhich it interacts. The protein is thoughtto bind approximately 4% of all brainmRNA transcripts, through its RGG boxdomain. It selectively recognizes RNA-containing hairpin or tetraplex secondarystructures (‘‘G-quartet’’), and shuttles intoand out of the nucleus and associates withpolyribosomes in messenger ribonucleo-protein complexes (mRNPs). The FMRP-containing mRNP complexes also containPur α and mStaufen, proteins involvedin the transport of neuronal granules.These granules, which contain RNA andassociated proteins, are transported to den-dritic spines in a metabotropic glutamatereceptor 5 (mGluR5)-dependent manner.FMRP can suppress the translation of cer-tain transcripts in vitro and in vivo. FMRPmay thus regulate the transport, localiza-tion, and translation of certain mRNAs inan activity-dependent manner (Fig. 9). InDrosophila, FMRP interacts with compo-nents of the RNAi machinery, which isinvolved in gene silencing and thus trans-lational control. Although FMRP does notseem to affect the siRNA pathway, it mayregulate microRNAs (miRNAs), which arenoncoding RNAs thought to control thetranslation of mRNAs by binding to their

3′-untranslated region. FMRP associateswith miRNAs, and proteins in miRNA-containing complexes in mammals, inter-actions that may be relevant to FMRPsregulation of translation.

FMRP is important for the develop-ment of dendritic spines and synapticplasticity. It regulates the expression ofMAP1B, an important regulator of mi-crotubule stability. FMRP knockout micehave abnormally high levels of MAP1B,resulting in increased microtubule stabil-ity. This may affect the development ofdendrites and/or dendritic spines. The ab-sence of FMRP in hippocampal neuronsresults in immature dendritic spine mor-phology and delayed synaptic connections,perhaps contributing to the neurologicalphenotype observed in FRAXA. FMRPmay also affect mGluR5-dependent longterm depression (LTD). LTD is a process bywhich neuronal activity can cause a lastingdesensitization of neurons to depolariza-tion. LTD-associated protein synthesis atsynapses may be enhanced by mGluR5activation, while it appears to be sup-pressed by FMRP. This is supported byevidence of enhanced LTD in the FMR-1knockout mouse. FMRP suppression andmGluR5 activation of LTD-dependent localprotein synthesis may be opposing forcesthat are out of balance in FRAXA, per-haps accounting for part of the cognitiveabnormalities.

4.2Fragile XE Mental Retardation

Fragile XE mental retardation (FRAXE) isan X-linked disorder with a prevalence ofabout 1–4% that of FRAXA. It accountsfor approximately two-thirds of familieswith nonspecific X-linked mental retar-dation (MRX), a classification in whichmental retardation is the only consistent


Fig.

9A

dvan

ces

inth

eun

ders

tand

ing

ofFM

RP’

sno

rmal

func

tion.

FMR

Pbi

nds

tore

cent

lytr

ansc

ribe

dm

esse

nger

RN

As

inth

enu

cleu

sto

form

ari

bonu

cleo

prot

ein

com

plex

(mR

NP)

.The

FMR

P–

mR

NP

com

plex

mov

esou

toft

henu

cleu

san

dth

enei

ther

dire

ctly

asso

ciat

esw

ithri

boso

mes

orin

tera

cts

with

the

RN

A-in

duce

dsi

lenc

ing

com

plex

(RIS

C).

The

FMR

P–

mR

NP

com

plex

isal

sotr

ansp

orte

dto

dend

rite

s.W

heth

erin

the

peri

nucl

ear

cyto

solo

rin

dend

rite

s,FM

RP

isbe

lieve

dto

regu

late

prot

ein

tran

slat

ion.

(Fro

mJin

,P.,

Alis

ch,R

.S.,

War

ren,

S.T.

(200

4)R

NA

and

mic

roR

NA

sin

frag

ileX

men

talr

etar

datio

n,N

at.C

ellB

iol.

6,10

48–

1053

.Use

dw

ithpe

rmis

sion

ofN

atur

eC

ellB

iolo

gy.)

.


clinical feature. Other characteristics arevariable and may include hyperactivity,mild facial hypoplasia and nasal abnormal-ities. FRAXE is caused by the expansionof a (GCC)n repeat in the promoter re-gion of FMR2, a gene 600 kb downstreamof FMR-1. It also overlaps with the pu-tative promoter of a gene called FMR3,transcribed in the opposite direction. Bothmaternal and paternal transmission can re-sult in expansion. Normal individuals bear6–35 GCC repeats, premutation carriersbear 61–200 repeats, and full mutationcarriers bear >200 repeats. Pathogenic ex-pansion results in a folate-sensitive fragilesite at the disease locus.

In mice, FMR2 mRNA is expressed inadult and fetal brain, kidney, lung, and pla-centa, with highest brain levels in the hip-pocampus and amygdala. Transcripts ofboth FMR2 and FMR3 are reduced beyonddetection in FRAXE patients. This may bedue to a methylation-dependent silencingprocess similar to FRAXA. The presence ofthe repeat within the preinitiation regionof the FMR2 promoter may also suggesta more direct disruption of transcription.FMR2 mRNA is highly expressed in thehippocampus, an area critical for learn-ing and memory. Its paralogs AF4 andLAF4 are both transcription transactivatorproteins, and the FMR2 protein seems tobe a potent transcription activator itself.Furthermore, the FMR2 protein is nu-clear, consistent with its proposed role intranscription regulation. The Drosophilaortholog of FMR2/AR4, Lilliputian, is es-sential for proper organ development, andits loss of function is lethal. FMR2 knock-out mice display impairment in the condi-tioned fear test and enhanced LTP (long-term potentiation) in the hippocampus.The role of FMR3 in FRAXE, if any, is un-known. The mechanism of FRAXE patho-genesis may therefore involve silencing of

the human FMR2 gene resulting in alteredtranscription in many parts of the devel-oping nervous system and mature brain.The neurological deficits characteristic ofFRAXE may involve effects of altered tran-scription on the hippocampus, amygdala,and possibly other brain structures.

4.3Friedreich’s Ataxia

With a prevalence of about one per 50 000individuals in the Caucasian population,Friedreich’s ataxia (FRDA) is the mostcommon inherited ataxia in this ethnicgroup. It is a multisystem degenerativedisease that is unusual among the tripletrepeat disorders due to its autosomalrecessive inheritance. Neurological symp-toms include gait, limb and truncal ataxia,loss of position and vibration senses, di-minished tendon reflexes, and dysarthria.Neuropathology changes include degener-ation of the posterior columns of the spinalcord, loss of large primary sensory neuronsin the dorsal root ganglia (DRG), and mild,late-onset degeneration of the cerebellarcortex. Other common clinical features arecardiomyopathy, diabetes mellitus, scol-iosis, and other skeletal abnormalities.Patients often present with symptoms inchildhood, become wheelchair-bound bytheir late teens or early twenties, andhave reduced lifespans due to cardiac fail-ure. Adult presenting patients, however,can have nearly normal lifespans withmore protracted progression and less se-vere nonneuronal involvement. Indeed,until the identification of the causal muta-tion, many of these adult-onset cases wentundiagnosed.

FRDA results from the expansion of apolymorphic (GAA)n repeat in the firstintron of the gene X25, now known asfrataxin. FRDA patients have at least one


expanded allele. For the disease to occur,the second allele must contain either an ex-panded repeat, or rarely, a copy of frataxincontaining a loss-of-function mutation. Nopatients have been described with loss-of-function allele mutations in both FRDAgenes. This is probably due to prenatallethality, as frataxin-null mice die in utero.Disease severity and onset age are deter-mined by the size of the (GAA)n repeatof the smaller expansion allele. Unaffectedindividuals carry at least one allele between7–38 repeats, while affected individualscarry two alleles of 66–1700 repeats. In-terruptions of the (GAA)n repeat resultin later onset and attenuated presenta-tions. GAA expansion at the FRDA locuscan result in enormous repeats in a sin-gle generation. Maternal transmission canresult in expansion or contraction, whilepaternal transmission results primarily incontraction.

FRDA patients exhibit reduced frataxinRNA and protein levels, and evidence sug-gests a defect in transcription or RNAmaturation. Current models propose thatfrataxin RNA elongation may be disturbedby triplex structure formation betweenexpanded DNA strands during transcrip-tion. DNA triplexes are formed when onestrand of a double-stranded DNA moleculefolds back upon itself, and interacts withtwo previously annealed strands. This cre-ates a local structure comprising threeDNA strands held together by hydro-gen bonds. Certain sequences favor thisprocess, among them extended (GAA)nrepeats. Triplex formation in the first in-tron would presumably affect transcriptelongation but not initiation. In support ofthis model are in vitro transcription exper-iments demonstrating no effect of repeatlength on transcript initiation.

Frataxin is a 210 amino acid proteinthat is well conserved from prokaryotes

to mammals. It contains mitochondrialtargeting signal sequences and localizesto the mitochondrial matrix. Frataxinexpression occurs at the primary sitesof pathology: dorsal root ganglia, spinalcord, sensory nerves, heart, and pancreas.These are tissues that rely upon high levelsof oxidative metabolism and consequentlyare rich in mitochondria. In addition, suchcell types are often postmitotic, meaningthat most dividing cell types are spared inFRDA patients.

Studies with FRDA patient mate-rial have demonstrated an increasedheart iron content and a deficiency iniron–sulfur cluster-containing proteins,including aconitase, a protein involvedin iron homeostasis. In addition, fibrob-lasts derived from FRDA patients areabnormally sensitive to iron and hydro-gen peroxide induced stress. Ablationof the yeast frataxin homolog results inrespiratory dysfunction, abnormal accu-mulation of mitochondrial iron, impairedbiogenesis of iron–sulfur proteins, andincreased sensitivity to oxidative stress.Conditional knockout of frataxin in stri-ated muscle results in a heart-specificphenotype resembling the cardiac abnor-malities in human FRDA (Fig. 10). Thissupports a genetic mechanism involvingloss of function of the disease protein.Frataxin may be part of a complex that de-livers iron to Iron–sulfur clusters (ISCs),which are cofactors essential to the activ-ity of many important cellular proteins.The accumulation of iron outside the mi-tochondria, although probably not centralto FRDA pathogenesis, supports a defi-ciency in iron delivery to ISCs. AmongISC-dependent cofactors are proteins in-volved in mitochondrial electron trans-port and thus respiration. Disturbancesin oxidative metabolism are often associ-ated with increases in the production of


(a)

(b)

(c)

Fig. 10 Friedreich’s ataxia mice displaycardiac muscle pathology.(a) Transmission electron microscopyof mice lacking expression of frataxin inonly their muscle reveals the abnormalaccumulation of lipid droplets (L) incardiac muscle at 4 weeks of age. (b) By7 weeks of age, mitochondria appearabnormal and iron deposits are visible(arrows). (c) With further progression,large vacuoles emerge (inset-top) asmitochondria undergo continuedprominent degeneration. Ultimately, themitochondria become engorged withelectron-dense material consistent withiron deposits. (From Seznec, H.,Simon, D., Monassier, L.,Criqui-Filipe, P., Gansmuller, A.,Rustin, P., Koenig, M., Puccio, H. (2004)Idebenone delays the onset of cardiacfunctional alteration without correctionof Fe-S enzymes deficit in a mousemodel for Friedreich ataxia, Hum. Mol.Genet. 13, 1017–1024, used withpermission of HumanMolecular Genetics).

toxic reactive oxygen species (ROS). FRDAtherefore is thought to result from re-duced frataxin levels, leading to abnormaliron–sulfur metabolism, mitochondrialdysfunction, oxidative stress, and tissuedegeneration. FRDA is thus caused by thedynamic mutation of a nuclear-encodedmitochondrial protein. The disease sharesfeatures with classic mitochondrial dis-orders such as MELAS (mitochondrialmyopathy, encephalopathy, lactic acido-sis, and strokelike episodes syndrome) and

MERRF (myoclonus epilepsy with raggedred fibers), which are caused by stable mu-tations to mitochondrial-encoded proteins.FRDA appears to be a classic mitochon-drial disorder with an unusual geneticbasis – dynamic mutation.

Given the data suggesting a role for ox-idative metabolism and iron transport inFRDA pathogenesis, antioxidants and irontransport molecules are being evaluated astreatments for this disease. Several clini-cal trials have been conducted in FRDA


patients using idebenone, a synthetic ana-log of coenzyme Q10 and a potent antioxi-dant. Some studies have shown substantialimprovements in the heart function ofpatients, although the lack of random-ized placebo controls and the variability ofthe disorder render the results controver-sial. A recent conditional knockout mousemodel of FRDA also supports the use ofidebenone as a treatment for the disorder,making treatment of FRDA patients withthis antioxidant a distinct possibility.

4.4Progressive Myoclonus Epilepsy Type 1

Progressive myoclonus epilepsy type 1(EPM1) is a rare, autosomal recessiveneurological disorder most prevalent inFinland and parts of North Africa. Its mostprominent symptoms are progressive pho-tosensitive myoclonus and tonic-clonicepilepsy. Some patients also experience aprogressive cerebellar ataxia and cognitivedecline. Neurodegeneration occurs in thethalamus, spinal cord, and the Purkinjeand granular neurons of the cerebellum.Typical age of onset is 6 to 16 years of age.EPM1 is sometimes caused by missensemutations in the cystatin B gene (CSTB);however, analysis of affected patients lack-ing such mutations revealed a dodecamerrepeat expansion upstream of CSTB. Thedodecamer repeat is located between 66and 77 bp 5′ of two putative transcriptionstart sites, and has the sequence CCCCGC-CCCGCG. Unaffected individuals have arepeat number of 2–3, premutation car-riers have 12–17 repeats, and affectedindividuals carry 30–150 repeat alleles.

CSTB is a highly conserved gene inthe cystatin family of cysteine protease in-hibitors. CSTB mRNA is ubiquitous, withhigh transcript levels in the hippocampus.

The protein binds to and inhibits lysoso-mal proteases such as cathepsins B, H, L,and S. Pathogenic expansion of the repeatcauses a reduction in CSTB transcriptionin some cell types. Cstb knockout micedevelop symptoms similar to those seen inEPM1 patients, and missense mutationscausing the disease disrupt the ability ofCSTB to bind cysteine proteases. Together,these data suggest a loss-of-function mech-anism for EPM1. Lowered inhibition ofcysteine proteases may result in neuronaldamage, causing the phenotype observedin EPM1.

There are several ways in which therepeat might reduce CSTB expression.The first is that reduced gene expres-sion may be due to an increase in thedistance between promoter elements andthe transcription start site. AP-1 bind-ing sites have been located upstream ofthe repeat. One study demonstrated atwo- to fourfold reduction in promoter ac-tivity when the repeat is expanded. Anequivalent reduction also occurred whensimilarly sized heterologous DNA wasused in place of the expanded repeat. In-dependent studies have confirmed a largereduction in promoter activity due to re-peat expansion. Another way the repeatcould affect CSTB expression is throughalterations in chromatin structure. The re-peat region is G-C rich and might excludenucleosomes or promote the formation ofsecondary structures. Abnormal DNA sec-ondary structures have been observed inthe EPM1 repeat region. These alterationscould affect the expression of CSTB orother genes in the vicinity. A third possiblemechanism for expression changes lies inthe sequence of the repeat region. Each do-decamer repeat contains a GC box, whichcould act as an Sp1 binding site in vivo. Incases of abnormally large repeat numbers,


increased Sp1 binding could contribute tothe suppression of CSTB transcription.

5Type 3: the RNA Gain-of-function RepeatDiseases

5.1Myotonic Dystrophy Type 1

Myotonic dystrophy or DM is an au-tosomal dominant, multisystem disorderwith a prevalence of 1 in 100 000 world-wide, and 1 in 8000 in European andNorth American Caucasian populations.Patients typically present with proximalor distal muscle dysfunction includingweakness, pain, and myotonia (failure ofmuscle relaxation). DM exhibits a combi-nation of characteristic symptoms: cardiacconduction abnormalities, subcapsular iri-descent cataracts, and unusual endocrinechanges. Other features include testicularatrophy, type II diabetes, and late-onsetcognitive impairment. In its most severeform, congenital DM, mental retardation,craniofacial deformities, and other devel-opmental abnormalities are present. Itexhibits both paternal and maternal trans-mission, although there is an almost exclu-sive maternal transmission in congenitalDM. Myotonic dystrophy type 1 (DM1) iscaused by a (CTG)n expansion in the 3′-untranslated region of the gene, dystroph-ica myotonica protein kinase (DMPK), onchromosome 19. Unaffected individualscarry 5–37 repeats, while affected individ-uals carry 50–4000 repeats. DM exhibitspronounced anticipation and dramatic so-matic instability. Interestingly, the clinicalphenomenon of anticipation (worseningseverity as a disease gene is transmittedfrom one generation to the next) was firstdescribed nearly a century ago in a family

segregating DM1. Although the anticipa-tion phenomenon was dismissed as anartifact of ascertainment by geneticists ofthe mid twentieth century, anticipationis now known to be a genuine featureof dynamic mutation diseases. Anticipa-tion results from the tendency of diseaserepeats to expand and the inverse corre-lation between repeat length and age ofdisease onset.

Several theories have been advancedto explain the molecular basis of DM1,including haploinsufficiency of DMPK,local chromatin effects on neighboringgenes, and gain of function exerted byexpanded RNAs. DMPK is proposed tohave many functions, some of which couldrelate to the disease, such as modula-tion of skeletal muscle sodium channels,RNA metabolism, calcium homeostasis,and the cell stress response. Initial ex-pression studies reported a decrease inDMPK RNA and protein levels, but aDMPK knockout mouse designed to testthe haploinsufficiency theory exhibitedonly mild myopathy that was inconsis-tent with the DM phenotype. The (CTG)ntract is a strong nucleosome assemblysite, and it was therefore hypothesizedto have trans-effects on the expressionof neighboring genes. According to thismodel, the myriad symptoms of DM1 arecaused by disturbances in the expressionof multiple nearby genes due to the expan-sion – making DM1 a ‘‘contiguous genesyndrome.’’ There are several genes inclose proximity to the DMPK gene (i.e.<5 kb) whose roles in DM1 were con-sidered; the most studied of these geneshas been SIX5. The SIX5 homolog inDrosophila is required for normal eyedevelopment and its mouse homolog isinvolved in regulating distal limb muscledevelopment. Expression data on SIX5 in


DM1 patients has been inconsistent, how-ever. A Six5 knockout mouse developedcataracts, although they were not of thetype observed in DM1. Another gene im-plicated by this model is DMWD, whichis expressed in the testis and suspectedto be involved in male infertility. Studiesindicated that the expression of DMWD isnot altered.

The third theory for the molecular basisof DM proposes that an RNA gain-of-function mechanism is responsible. RNAfoci, which are accumulations of expandedtranscripts, accumulate in the nuclei ofpatient cells (Fig. 11). The RNA gain-of-function theory was buttressed by thediscovery that DM2, a disorder with symp-toms almost identical to DM1, is causedby a (CCTG)n expansion in intron 1 ofthe zinc finger 9 protein (ZNF9). The twogenes responsible for DM1 and DM2 areunrelated and reside on different chro-mosomes. Genes near the two loci bearno obvious resemblances. The only strik-ing parallels between the two expansionsare: (1) they both contain CTG triplets;and (2) they both occur in transcribed

but noncoding regions of the genome.More support for the RNA gain-of-functionmodel came from a mouse model gen-erated by Mankodi et al. in 2000, whichcontained a (CTG)n expansion in the 3′untranslated region of the skeletal actin(HSA) gene. This mouse exhibited myopa-thy typical of DM1, although the skeletalmuscle-restricted expression pattern ofHSA precludes broader conclusions. Thus,there are several lines of evidence suggest-ing that CUG expansion-containing RNAsare capable of causing DM.

RNA-binding proteins and transcriptionfactors colocalize with the RNA foci foundin DM, potentially altering nuclear pro-cesses. Elevated levels of CUG-containingRNA has been shown to alter gene splic-ing in specific transcripts which couldbe relevant to DM: cardiac troponin T(cTNT), involved in cardiomyopathy; In-sulin Receptor (IR), involved in diabetes;and Clc-1, the main chloride channel inmuscle. Abnormal splicing of cTNT, IR,and ClC-1 are proposed to account forthe cardiac abnormalities, insulin insen-sitivity and myotonia observed in DM.

(a) (b) (c)

Fig. 11 In situ hybridization of muscle sections with fluorescently labeled antisenseoligonucleotide probes reveals accumulation of mutant RNA in DM1 and DM2.(a) Probing of DM2 muscle with a CAGG probe indicates that multiple RNA foci arepresent. (b) Probing of normal muscle with a CAGG probe demonstrates absence ofRNA foci. (c) Probing of DM1 muscle with a CAG probe yields prominent RNA foci.(From Liquori, C.L., Ricker, K., Moseley, M.L., Jacobsen, J.F., Kress, W., Naylor, S.L.,Day, J.W., Ranum, L.P. (2001) Myotonic dystrophy type 2 caused by a CCTG expansionin intron 1 of ZNF9, Science 293, 864–867, used with permission of Science).


Splicing abnormalities cause a reductionin the membrane concentration of ClC-1 and reduce chloride conductance tolevels consistent with myotonia. Thesespicing alterations are thought to be dueto the repeat-expanded RNA’s effects ontwo families of RNA-binding proteins:‘‘CUG-BP1 and ETR-3-like factors’’ (CELF)and ‘‘muscleblind-like proteins’’ (MBNL).CELF proteins regulate pre-mRNA splic-ing in cTNT, IR, and ClC-1. In patienttissue and cell culture, CUG-BP1 levelsand activity are increased in response toelevated levels of CUG-containing RNA.This may be due to a lengthening of theprotein’s half-life. MBNL proteins werenamed after their Drosophila orthologmuscleblind, which is required for pho-toreceptor and muscle differentiation inflies. The three known proteins in thisfamily (MBNL1 (MBNL), MBNL2 (MBLL),and MBNL3 (MBXL)) are splicing reg-ulators that are thought to act antago-nistically to CELF family proteins. Theycolocalize with RNA foci in vivo, and amouse model lacking specific isoformsof MBNL1 recapitulates the myotonia,cataracts, and splicing dysregulation ob-served in DM.

The splicing alterations seen in DMpatients are consistent with loss of func-tion of MBNL proteins or an increase inCELF protein activity in muscle and brain.Because of the colocalization of MBNL pro-teins with RNA foci, it has been proposedthat sequestration and subsequent deple-tion of these proteins from the cellularmilieu is responsible for the symptoms ofDM. Although this process probably playsa critical role in the disease, recent evi-dence suggests it is not solely responsible.The most plausible current theory explain-ing DM pathogenesis proposes an im-balance between the antagonistic MBNLand CELF proteins, resulting in specific

splicing abnormalities that cause DM’s di-verse range of symptoms.

5.2Myotonic Dystrophy Type 2

Myotonic dystrophy type 2 (DM2) is anautosomal dominant, multisystem disor-der very similar to DM1. The majority ofits symptoms resemble DM1: progressiveweakness, myotonia, cardiac disturbances,iridescent cataracts, and insulin insensi-tivity. There are some notable differences,however. DM2, unlike DM1, predomi-nantly affects proximal muscles at itsonset, which is why many cases wereoriginally classified as proximal myotonicmyopathy (PROMM). Other interestingdifferences are that mental retardation isnot observed in DM2, DM2 patients showincreased sweating, and DM2 congenitalforms have not been observed. DM2 iscaused by a (CCTG)n tetranucleotide ex-pansion in the first intron of the zinc fingerprotein 9 (ZNF9) gene on chromosome3q21.3. The tetranucleotide repeat can ex-pand to stunning lengths, with the longestcases comprising 44 kb of DNA, makingthem the longest tracts observed in the re-peat expansion disorders. The DM2 locusexhibits marked somatic instability. Overthe course of a patient’s lifetime, the aver-age repeat length increases substantially,as judged by blood drawings.

The similarities between DM1 and DM2are not restricted to the clinical presen-tation. The expansions are both large,CTG-containing tracts that are transcribed,but not translated. Both disorders causenuclear RNA foci that sequester specificRNA-binding proteins (Fig. 11), includingCELF and MBNL family members. Similarsplicing abnormalities are also observed inDM2. Since the genes associated with therepeat expansions in DM1 and DM2 are


unrelated, it is likely that the cause of bothdisorders is a toxic gain of function of longtracts of CUG-containing RNA. A mousemodel generated by Mankodi et al. in 2000,which contained a (CTG)n expansion inthe 3′ untranslated region of the skeletalactin (HSA) gene, exhibited myopathy typ-ical of DM. Thus, even outside the contextof the DM1 and DM2 loci, CTG-containingexpansions can cause DM-like symptoms.This supports the RNA gain-of-functiontheory, as transcribed CTG-containing ex-pansions are inherently capable of causingdisease. Differences between DM1 andDM2 may be due to regional or tempo-ral expression patterns or differences inthe affinity of RNA-binding proteins forCTG or CCTG tracts.


Spinocerebellar ataxia type 8 (SCA8) is adominantly inherited cerebellar ataxia. Af-fected individuals suffer from late-onset,slowly progressing gait ataxia as wellas dysarthria, oculomotor incoordination,spasticity, and decreased vibration sense.The cerebellar cortices and vermis un-dergo a slowly progressing but dramaticatrophy, while the brainstem exhibits lit-tle evidence of degeneration. Patients maybecome wheelchair-bound as early as theirfourth decade of life. Using a direct methodfor cloning expanded triplet repeats (i.e.RAPID cloning), a gene containing anexpanded CTG tract was identified in alarge family (MN-A) segregating the SCA8phenotype. Interestingly, this CTG is con-tained within a gene on the long armof chromosome 13 that is transcribed,but apparently not translated into a pro-tein product. Thus, it was proposed thatthe production of an RNA transcript con-taining this expanded CTG repeat tract is

the cause of SCA8. Numerous subsequentstudies, however, have indicated that pos-session of the expanded CTG repeat tractappears necessary, but not sufficient, forthe production of the SCA8 phenotype.Thus, reduced penetrance is viewed as akey feature of SCA8 at this time. On thebasis of the available genetic data, nor-mal individuals always carry fewer than 70CTG repeats, while affected SCA8 patientscan carry anywhere from 71 to >1000 CTGrepeats. As extremely large CTG repeat al-leles (>800 triplets) were shown to notcause disease in early reports, this was ini-tially attributed to lack of stability of themutant RNA product. However, furtherwork has revealed considerable overlapbetween disease-causing CTG repeat ex-pansions and nonpenetrant CTG repeatalleles, indicating that a secondary fac-tor – either a trans-acting genetic factor oran environmental factor – must be presentto yield the SCA8 disease phenotype. In thecase of the original MN-A family, evidencefor a cis modifier appears responsible forthe extremely high penetrance in this largepedigree. Thus, while the causality of theCTG repeat expansion in SCA8 has beensomewhat controversial, review of the cur-rent literature suggests that the CTG repeatexpansion is directly involved in SCA8,but may not alone be sufficient to producethe disorder.

Another unique feature of the SCA8CTG repeat is its extreme and unusualgenetic instability. Paternal transmissionsgenerally result in contractions, while ma-ternal transmissions generally result in ex-pansions. Expansions of up to 600 repeatshave occurred in one generation throughmaternal transmission. Large deletions inexpanded alleles often occur in sperm cells,offering an explanation for the paternalcontraction bias. Pathogenic expansions


often have 5′ triplet interruptions, the roleof which is unclear.

The molecular basis of how the ex-panded SCA8 RNA causes disease remainsunclear. The transcript has been detectedexclusively in the brain, and appears to betranscribed in the CTG orientation. Trans-lation of a polyglutamine tract from theCAG-containing transcript in the oppo-site direction is precluded by the existenceof stop codons flanking the repeat. Thelongest transcript identified to date con-tains 6 exons, and many alternativelyspliced forms of the gene have been de-tected, although all these isoforms areexpressed at very low levels. None con-tain significant open reading frames. In1999, Koob et al. reported that a gene par-tially overlaps the 5′ end of the SCA8 genelocus on the antisense strand. This gene,called Kelch-like 1 (KLHL1), for its ho-mology to the Drosophila KELCH gene,is highly conserved and predicted to en-code an actin-binding protein of 748 aminoacids. Its expression overlaps with that ofthe SCA8 transcript, suggesting that theSCA8 gene may produce an antisense RNAwhose function is to regulate KLHL1 ex-pression. Whether or how this is occurringremains uncertain at this time.

With the discovery of the DM RNA toxicgain-of-function pathway, an emergingtheory for SCA8 pathology holds that theproduction of a CUG-expanded transcriptresults in RNA gain-of-function toxicitywithin the restricted neuronal populationswhere the SCA8 gene is expressed. In sup-port of this hypothesis, transgenic micederived with a human bacterial artificialchromosome containing the entire SCA8gene with a 118 CTG repeat expansion de-velop neurological disease. The severity oftheir phenotype depends on the expressionlevel of the transgene. In the more moder-ate expressing lines, the SCA8 mice display

a slowly progressive gait ataxia reminiscentof the human disease. While no proteinproduct has been detected, 1C2 and ubiq-uitin antibody staining reveal intranuclearinclusions in cerebellar neurons, consis-tent with the RNA-containing inclusionsobserved in FXTAS brains. An alternativeinterpretation is that translation of CTGinto a polyleucine tract is occurring. Fur-ther work will be needed to distinguishthese two possibilities.

Studies of the SCA8 CTG repeat expan-sion in D. melanogaster have also beeninformative. Whether directing expressionof the SCA8 gene to fly retina with a nor-mal CTG repeat tract or an expanded CTGrepeat tract, a neurodegenerative eye phe-notype results. Modifier screens using flystocks carrying mutations in genes encod-ing RNA-binding protein yielded a numberof genes that could either enhance orsuppress this retinal degeneration pheno-type. Interestingly, Drosophila muscleblind,whose mammalian counterpart has beenimplicated in the DM RNA toxicity path-way, modified the retinal degenerationcaused by expanded SCA8 CTG repeatexpression more so than the retinal de-generation caused by normal SCA8 CTGrepeat expression. Such data supports thehypothesis that the SCA8 CTG repeatexpansion is producing neurotoxicity byaltering the function of RNA-binding pro-teins as in DM. Although it would bepremature to definitively conclude thatSCA8 is an RNA gain-of-function repeatdisease, provisional classification in thiscategory is appropriate.

5.4The Fragile X Tremor – Ataxia Syndrome(FXTAS)

Fragile X tremor-ataxia syndrome (FX-TAS) is a rare and unusual disorder


that is associated with premutation ofthe FRAXA locus in males. Patients de-velop a progressive ataxia and intentiontremor. This is sometimes accompanied bydementia, parkinsonism, and autonomicdysfunction. Neuropathological changesinclude degeneration of the cerebellumand ubiquitin-positive intranuclear inclu-sions in neurons and glia. FXTAS is asso-ciated with the expansion of a (CGG)n tractin the 5′-untranslated region of the FMR1gene, found on chromosome Xq27.3. Ex-pansion of this tract beyond 200 repeatscauses FRAXA, a disorder resulting froma reduction in the expression of FMR1.

Males with tracts between 55 and 200 CGGrepeats are considered to be in the ‘‘pre-mutation range,’’ and are now consideredto be at risk for FXTAS beyond middle age.

In contrast to FRAXA, FXTAS patientsshow increased expression of the FMR1transcript, with longer repeats correspond-ing to higher transcript levels. Premutationcarriers with more than 100 repeats havean average of five times more FMR1mRNA than individuals in the normalrepeat range. Despite the increase in tran-script level, the protein product of theFMR1 gene, FMRP, is slightly reduced inFXTAS patients. Since elevated transcript

Fig. 12 Intranuclear inclusions arepresent in the brains of FXTASpatients. (a) Hematoxylin & eosinstaining of cerebral neurons revealsa refractile, eosiniphilic nuclearinclusion of ∼5 µM in diameter(white arrowhead). (b) Hematoxylin& eosin staining of cerebralastrocytes reveals refractile,eosiniphilic nuclear inclusions of∼2 µM in diameter (whitearrowhead). (c) Antineurofilamentantibody staining of cerebellumdemonstrates presence ofdystrophic neurites, consistent withongoing Purkinje cell degeneration.(d) Antiubiquitin antibody stainingof cerebral neurons labelsintranuclear inclusions.(e) Antiubiquitin positiveintranuclear inclusions form in bothneurons (white arrowhead; largercell) and astrocytes (whitearrowhead; smaller cell). (FromGreco et al. (2002) Brain 125, 1760,used with permission).

(a) (b)

(c) (d)

(e)

10 µm

50 µm

10 µm

10 µm

10 µm


levels are present without great alterationin protein levels, this disorder may bedue to an RNA gain of function of theFMR1 transcript. Indeed, immunostain-ing of FXTAS patients’ brain sectionsreveal intranuclear inclusions in neuronsand glia, believed to be comprised ofaccumulated Fmr1 RNA transcripts andvarious proteins (Fig. 12). To determinethe molecular basis of FXTAS, a knockinmouse model of FXTAS was generatedusing a human 98 CGG repeat, and wasnoted to produce elevated levels of theFmr1 transcript and display intranuclearribonucleoprotein inclusions, containingubiquitin, Hsp40 and the 20S proteasomesubunit. The inclusions may be a responseto RNA toxicity or a result of the aggre-gation of CGG-binding proteins. Whilethe mechanism of FXTAS may be dueto a dominant gain of function of triplet-expanded RNA on RNA-binding proteinsas in DM, there is currently no direct evi-dence supporting this theory.

6Type 4: The Polyalanine Diseases

6.1Overview

Polyalanine disorders are characterized bysmall expansions in trinucleotide repeatsencoding alanine tracts. Most are rare,autosomal dominant developmental mal-formations. Although the normal functionof the alanine tract is unknown, they tendto be found in transcription factors. Themajority of alanine tract expansions asso-ciated with disease loci are also locatedin transcription factors, most of whichplay a role in development. In contrastto polyglutamine expansions, polyalanineexpansions are small and stable, not

exceeding 30 triplets. Most are composedof ‘‘imperfect’’ alanine tracts, includingcombinations of GCG, GCA, GCC, andGCT codons. This composition arguesagainst an expansion mechanism involv-ing single-strand, hairpin-like secondarystructures as in other repeat disorders.A more plausible mechanism for polyala-nine expansion is unequal crossing-over,in which alleles mispair during meioticcrossing-over, resulting in one expandedand one contracted tract.

6.2Oculopharyngeal Muscular dystrophy

Oculopharyngeal muscular dystrophy(OPMD) is a predominantly autosomaldominant, late-onset disorder character-ized by progressive drooping eyelids, dys-phagia, and proximal limb weakness. Cer-tain skeletal muscles in affected patientsdegenerate and contain nuclear inclusionsand rimmed vacuoles. This effect is partic-ularly striking in the levator palpebra andpharyngeal muscles, responsible for lift-ing the eyelids and swallowing. OPMD iscaused by an alanine expansion in the genepolyadenine-binding protein 2 (PABP2).PABP2 normally contains a 10-alaninetract encoded by (GCG)6(GCA)3GCG. Af-fected individuals carry 12–17 alanines,probably resulting from unequal crossing-over of the two alleles. Repeat lengthappears to correlate with disease severity.

PABP2 is an abundant, ubiquitously ex-pressed pre-mRNA-binding protein thatplays a role in controlling the formationand length of mRNA polyA tails. NIsimmunoreactive for PABP2 are presentin the skeletal muscle cells of OPMDpatients and sequester polyA-containingtranscripts. The aggregates are filamen-tous and contain PABP2, ubiquitin and


proteasome subunits, leading investiga-tors to conclude that expansion of thealanine tract probably causes PABP2 tomisfold and aggregate. Aggregation inOPMD has been linked to toxicity inseveral experimental systems. Polyalanine-expanded PABP2 aggregates and causescell death in cultured cells and trans-genic mice. Indeed, widespread expressionof the human PABP2 gene in trans-genic mice yielded polyalanine length-dependent muscle pathology, includingrimmed vacuoles, central nuclei, andnumerous dystrophic changes (Fig. 13).Deleting the C-terminal oligomerizationdomain, overexpressing chaperones, or ex-posing cells to aggregation inhibitors such

as Congo red and doxycycline, reducesaggregate formation and toxicity in cellculture. A polyalanine-expanded peptide,similar to the amino-terminal region ofPABP2 was shown to adopt a β-sheet con-formation, whereas the same peptide with7 alanines adopted an α-helical conforma-tion. This is reminiscent of polyglutamineproteins, which also adopt a β-sheet con-formation when the polyglutamine tractexceeds a threshold of about 35 glu-tamines. Expanded polyalanine proteinsalso activate caspase-3 and -8 in culturedcells, another feature reminiscent of polyg-lutamine toxicity.

In 2004, Wirtschafter et al. proposed amodel for the selective vulnerability of

(a) (c) (e)

(b) (d) (f)

Fig. 13 Mice expressing polyalanine-expanded PABP2 protein display prominent musclepathology consistent with the OPMD phenotype. Transgenic mice were generated using thepCAGGS expression vector that consists of the chicken beta-actin promoter and a CMVenhancer, and thus ubiquitously express either human PABP2 protein with six alanines (c, e)or nine alanines (a, b, d, f) at roughly comparable levels. Hematoxylin & eosin staining ofsections from the soleus muscle (a, b), the pharynx muscle (c, d), and the eyelid muscle (e, f)reveal normal muscle histology in the six-alanine PABP2 expressing mice (c, e). However,mice expressing PABP2 with nine alanines display nonuniform muscle fiber size andprominent connective tissue in pharynx and eyelid musculature (d, f). At high power,cytoplasmic vacuoles, reminiscent of OPMD ‘‘rimmed vacuoles’’ are apparent in soleusmuscle from the nine-alanine PABP2 expressing mice (a). Low power examination furtherindicates frequent central nuclei in the soleus muscle from such mice (b). (From Hino et al.(2004) Hum. Mol. Genet. 13, 181, used with permission).


extraocular muscles in OPMD. Unlikeother skeletal muscles, extraocular mus-cles are not postmitotic and they contin-ually undergo remodeling. This requiresthe frequent upregulation of genes in-volved in cell cycling and protein synthesis.Failure of correct mRNA polyadenylationor transport in these cells may have acumulative toxic effect resulting in the pro-gressive degeneration of these muscles.The mechanism of OPMD could be dueto an interference with polyadenylation,disturbances in intracellular trafficking ofmRNA, or toxicity due to the aggregationof misfolded and/or aggregated PABP2species. Further research will be requiredto distinguish between these possibilities.

6.3Synpolydactyly (Syndactyly Type II)

Synpolydactyly (SPD) is a rare, autosomaldominant developmental disorder charac-terized by fused and extra digits (syndactylyand polydactyly, respectively). It is causedby the expansion of a polyalanine tract inthe amino-terminal region of the transcrip-tion factor HOX-D13. HOX genes act inconcert to coordinate axial patterning inanimals. Tracts of 7–14 alanines in HOX-D13 have been linked to SPD, and diseaseseverity is proportional to repeat length.Multiple studies support a ‘‘dominant-negative’’ role for expanded HOX-D13protein in SPD. Mice null for Hox-d13have a phenotype less severe than SPD,while mice with alanine tract expansionshave a form that more closely resemblesthe disorder. Mice lacking Hox11, Hox12,and Hox13 have a phenotype similar toSPD, suggesting that the alanine expan-sion in SPD antagonizes the function ofother HOX genes. This is supported by agenetic complementation study. In further

support of this theory, humans with sus-pected loss-of-function mutations in theHOX-D13 gene do not have a phenotypeconsistent with SPD.

6.4Cleidocranial Dysplasia

Cleidocranial dysplasia (CCD) is a rare, au-tosomal dominant developmental disordercharacterized by holes in the skull, den-tal malformations, absent or hypoplasticclavicles and maxillae, and other skeletalmalformations. The primary cause of CCDis thought to be loss-of-function mutationsin the gene RUNX2. This gene affects os-teoblast differentiation and is a memberof the Runt family of transcription fac-tors. In one family, phenotypically distinctfrom classic CCD, an expansion from 17to 27 alanines in RUNX2 has been de-tected. This family exhibited brachydactylyand a mild CCD phenotype. The differ-ence in phenotype between the polyalanineexpansion mutation family and typical pre-sumed haploinsufficient, loss-of-functionCCD patients supports a gain-of-functioneffect of the expanded alanine tract in thisatypical family. Contraction of the tract inRUNX2 is common and does not cause adetectable phenotype.

6.5Holoprosencephaly

Holoprosencephaly (HPE) is a commondevelopmental malformation resulting inpartial or full cyclopia, failure to developmidline structures in the ventral fore-brain, and prenatal lethality. In rare cases,HPE is caused by the heterozygous ex-pansion of a 15-amino acid alanine tractin the protein ZIC2. This protein is onemember of a family of zinc finger pro-teins believed to regulate neurulation,


left–right axis formation and other de-velopmental processes. Other individualswith heterozygous loss-of-function muta-tions have a phenotype indistinguishablefrom those with alanine tract expansions.Partial loss of function of Zic2 in micesimilarly causes developmental abnormal-ities similar to HPE. Therefore, expansionof the alanine tract likely causes loss offunction of ZIC2 in the case of HPE.

6.6Hand-foot-genital Syndrome

Hand-foot-genital syndrome (HFGS) is arare, dominantly inherited developmentalabnormality characterized by malforma-tion of the distal limbs and lower urogen-ital tract. Short thumbs, short great toes,and abnormal carpals and tarsals are someof its salient features. Some HFGS patientshave alanine tract expansions in the pro-tein HOX-A13, which is in the same familyas the SPD-associated protein HOX-D13.HOX-A13 contains three alanine tracts,the most C-terminal of which is the mostcommonly mutated. Expansions enlargethe second or third tract by 6 to 9 ala-nines. In humans, deletion of HOX-A13causes a phenotype that is mild in com-parison to HFGS caused by alanine tractmutations. Hoxa13-null mice also have amilder phenotype than mice carrying aframeshift deletion suspected to confergain of function. This evidence suggests adominant-negative mechanism for HFGScaused by alanine tract expansions.

6.7Blepharophimosis-ptosis-epicanthusInversus Syndrome

Blepharophimosis-ptosis-epicanthus in-versus syndrome (BPES) is a rare, auto-somal dominant developmental disorder

resulting in malformation of the uppereyelids and forehead, and occasionally pre-mature ovarian failure in women. It iscaused by a number of different muta-tions in the gene Forkhead L2 (Foxl2), themost common of which is expansion ofits carboxy-terminal 14-amino acid alaninetract. FOXL2 is a highly conserved tran-scription factor whose role in the ovaryhas been studied most thoroughly. It isexpressed in the ovaries and eyelids dur-ing development and adulthood. Studiesin which Foxl2 have been ablated inmice show that it is important for thedifferentiation of ovarian granulosa cells,and its absence causes accelerated folliclecell depletion leading to POF. Knockoutmice also display craniofacial abnormali-ties consistent with BPES, suggesting thata dominant-negative mechanism may beresponsible in humans.

6.8Syndromic and Nonsyndromic X-linkedMental Retardation

Several loosely related disorders are as-sociated with alanine tract expansions inthe Aristaless related homeobox (ARX)protein on chromosome Xp22.13. Oneis nonsyndromic X-linked mental retarda-tion (XLMR), a heterogeneous condition inwhich mental retardation is the main con-sistent feature. Several syndromic XLMRdisorders linked to alanine expansionin ARX include West syndrome (WS)and Partington syndrome (PRTS). WScauses progressive mental retardation withabnormal EEG and infantile seizures,while PRTS causes mental retardation,dysarthria, and dystonic movements of thehands. Brain anatomy appears normal inthese disorders.

ARX is a paired-class homeodomainprotein that is expressed in the ventricular


and marginal zones of the developingmouse brain and continues to be expressedin adult cortex. It is suspected to play a rolein neuroepithelial cell differentiation andmaintenance of neuronal subtypes in theadult cortex. ARX-null mice have smallbrains and neuronal migration deficits.Expansions of two different alanine tractsin ARX cause XLMR: a 12-alanine tractat amino acids 144–155, and a secondtract at amino acids 100–115. Both sitescause highly variable forms of XLMR, withexpansions in the former alanine tractresulting in nonsyndromic XLMR, WS orPRTS. Alanine tract expansions in ARXprobably result in partial loss of function ofthe protein. Heterozygous female carriersof expanded ARX do not exhibit XLMR,suggesting that one normal copy of theprotein is sufficient for normal cognitiveability. Also, humans with null mutationsof ARX suffer from X-linked lissencephalywith abnormal genitalia (XLAG), a muchmore severe condition causing majordevelopmental abnormalities in the brainand genitalia. Thus, XLMR due to alanineexpansions in ARX may be because ofpartial loss of function leading to subtledevelopmental defects and/or failure tomaintain specific neuronal populations.

6.9Congenital Central HypoventilationSyndrome

Congenital central hypoventilation syn-drome (CCHS) is a rare, autosomal dom-inant disorder causing a failure of auto-nomic control of breathing. It attenuatesor abolishes responses to hypercarbia andhypoxemia. In the majority of cases, it iscaused by the expansion of one of twoalanine tracts in the protein PHOX-2B.Mutations expand the 20-residue tract to25–29 alanines. PHOX-2B is a paired-class

transcription factor containing a home-odomain. A loss-of-function mutation inmurine Phox2b is homozygous lethal,and specifically prevents the developmentof parasympathetic ganglia. Heterozygousmice show chronic pupil dilation butno parasympathetic or respiratory distur-bances. Also, a patient hemizygous fora 5-Mb deletion including Phox2b doesnot have CCHS. No cases of CCHS havebeen reported in which PHOX2B is trun-cated before the homeobox domain. Thisevidence suggests that the expansion ofalanine tracts in PHOX2B may cause asubtle, dominant-negative effect on the de-velopment of respiratory control pathways.

7Unclassified Repeat Diseases LackingMechanistic Explanations


Spinocerebellar ataxia type 10 (SCA10)is an autosomal dominant, progressiveataxia that exhibits nearly pure cerebellarsigns. It appears restricted to individualsof Mexican ethnicity. SCA10 patients typ-ically present with gait ataxia, followed bydysarthria, dysphagia, and ocular dysme-tria, and most patients also experience re-current motor seizures. Cerebellar atrophyis the most prominent neuropathologicalchange. SCA10 is caused by the expansionof a highly polymorphic pentanucleotiderepeat, (ATTCT)n, on chromosome 22. Un-affected individuals carry 10–22 repeats.While expanded alleles have not been suc-cessfully PCR amplified, transcript sizeson Northern blots and fragment sizes onSouthern blots indicate that disease allelescan range from ∼800 to 4500 repeats. An-ticipation occurs in SCA10, independently


supporting the causal nature of the repeatexpansion. The ATTCT repeat is located inintron 9 of the SCA10 gene, a previouslyunrecognized 66-kb gene that encodes anovel putative 475 amino acid protein ofunknown function with few recognizablemotifs or domains. The 2-kb SCA10 tran-script is ubiquitously expressed though thehighest levels of expression occur in thebrain, testis, and adrenal glands. Withinthe brain, it is most highly expressed in thecerebellum and associated structures. Thecarboxy-terminal portion of the proteinappears to contain ‘‘armadillo repeats,’’which are responsible for the membraneassociation of β-catenins. Ataxin-10 doesnot seem to associate with membranes,however. The entire protein is highly con-served between humans and rodents, andpotential orthologs exist in Arabidopsisand Drosophila.

There are several proposed mechanismsfor SCA10 pathogenesis. The most obvi-ous hypothesis is that the large intronicexpansion affects ataxin-10 expression,perhaps by altering local chromatin struc-ture. siRNA knockdown of ataxin-10 in cellculture experiments yields higher ratesof cell death in cerebellar neurons thanin cortical neurons. However, the expres-sion levels of the SCA10 transcript are notreduced in patient’s lymphoblast cells, ar-guing against simple haploinsufficiency.Consequently, RNA gain-of-function tox-icity has been proposed as the potentialmechanism; however, there are no fur-ther data at this time to support such ahypothesis. Additional work will need tobe done to distinguish between these andother possibilities, and will need to ac-count for the cell-type specific pattern ofneurodegeneration that occurs in the faceof apparently widespread expression of theSCA10 gene mutation.


Among the more recent additions to theunstable repeat disease group is spinocere-bellar ataxia type 12 (SCA12), a rare, auto-somal dominant disorder that may be mostprevalent in Indian populations. Its symp-toms are distinct from the other SCAs,typically beginning with an action tremorof the upper extremities and progressing toinclude hyperreflexia, mild cerebellar dys-function, bradykinesia, increased muscletone, psychiatric symptoms, and demen-tia. The brains of SCA12 patients likelyundergo a slow, generalized atrophy thatis most prominent in the cortex, but alsoresults in loss of Purkinje cells in the cere-bellum. Disease onset typically occurs inthe third or fourth decade, and a graduallyprogressive disease course is typical.

SCA12 is caused by a (CAG)n expansionin chromosome 5q31 – q33. The expan-sion is 5′ to the PPP2R2B gene, encodinga regulatory subunit of the protein phos-phatase 2A enzyme (PP2A). This genehas many transcription start sites, someof which include the repeat (but many ofwhich do not). GENSCAN predicts an exonincluding the expansion that would encodea polyserine tract, but this prediction is oflow probability and considered unlikely.Other evidence suggests that the expan-sion is located in the promoter region,and this is currently the most widely ac-cepted view. Unaffected individuals carry7–32 CAG repeats while affected individ-uals carry 55–78 CAG triplets. The mostcommon repeat size in unaffected individ-uals is 10 CAGs. The repeat is fairly stable,with only modest expansions and contrac-tions resulting equally from maternal andpaternal transmission. A significant corre-lation between repeat length and age ofonset has not been documented.


The protein PP2A is an essential ser-ine/threonine phosphatase expressed inall known eukaryotic cells. It is involvedin diverse cellular functions, including cellgrowth, differentiation, DNA replication,neurotransmitter release, and apoptosis.PPP2R2B is a brain-specific regulatorysubunit of PP2A. The class of regulatorysubunits including PPP2R2B may affectPP2A’s phosphatase activity for certainsubstrates, including histone-1, vimentin,and tau. It may also affect PP2A’s subcel-lular localization.

There are several possible explanationsfor SCA12 pathogenesis. The first isthat expanded PPP2R2B may generate apolyamino acid tract-containing protein,resulting in toxicity. Northern blots prob-ing for sequence flanking the CAG repeatdid not detect a PPP2R2B transcript, in-dicating that if a repeat-containing exonexists and is transcribed, it is not presentat appreciable levels. Nevertheless, thepossibility of polyglutamine, polyserine,or polyalanine toxicity, though unlikely,cannot be completely ruled out at thistime. A second possible cause of SCA12pathogenesis is RNA gain of function, asoccurs in DM and FXTAS. The repeatexpansions are smaller in SCA12 thanin DM, however, and more importantly,CAG tract-containing transcripts appearto be rather scarce. A third possibilityis altered splicing of the PPP2R2B tran-script. Several different amino-termini arepossible, the ratio of which may affectPP2A’s subcellular localization. Yet an-other theory of SCA12 pathogenesis isthat the expansion affects PPP2R2B tran-script levels. Repeat expansion causes asubstantial increase in PPP2R2B expres-sion as measured in reporter assays usinga neuroblastoma cell line. Altered lev-els of the protein could affect PP2A’sspecificity or subcellular localization. This

has the potential to disturb a multi-tude of processes in the CNS. At thistime, all of the above theories of SCA12CAG repeat expansion neurotoxicity re-main plausible.

7.3Huntington’s Disease Like 2 (HDL2)

Perhaps the most exciting and enigmaticrecent discovery in the repeat expansionfield is that of the mutational basis ofa disorder known as Huntington’s dis-ease like 2 or HDL2. HDL2 is so namedbecause it is in essence a genocopy ofclassical HD, as the original HDL2 pedi-gree was labeled with a diagnosis of HDuntil HD CAG repeat testing indicatedthat this family’s HD-like disease did notresult from a CAG repeat expansion inthe htt gene. HDL2 patients present withweight loss and diminished coordination,and then develop tremors, dysarthria, hy-perreflexia, and rigidity. Patients displaypsychiatric involvement, chorea, and dys-tonia, and ultimately become demented.Death occurs 15–25 years after onset,when patients become bedridden as in typ-ical HD cases. MRI findings reveal markedatrophy of the caudate and of the cere-bral cortex, making the neuropathologyindistinguishable from classic HD. HDL2patients do not show cerebellar signs orneuropathology.

As soon as it was found that HDL2patients do not have the HD CAG re-peat expansion, direct cloning methods fortriplet repeat expansions of the CAG/CTGtype were applied to patient samples andan expanded CAG/CTG repeat (n = 55)was isolated. Sequence flanking this repeatindicated that HDL2 is caused by CTGrepeat expansions in the junctophilin-3(JPH3) gene, one of a family of struc-tural proteins whose function is to link


Ca++ channels on the ER with voltagesensors on the plasma membrane. Anal-ysis of JPH3 CTG repeat indicates thatnormal individuals typically carry 7–27CTG repeats while affected HDL2 patientsusually have expansions of 50–60 CTGrepeats. Widespread screening of HD-likepatients from around the globe suggeststhat the HDL2 JPH3 CTG repeat expan-sions are most common in individuals ofAfrican ethnicity.

The question of how the JPH3 CTGrepeat expansion causes HDL2 remainsunknown, but there are at least four possi-ble explanations. These alternative (but notmutually exclusive) theories stem from thedocumented alternative processing of theJPH3 gene, which permits the predictionof the CTG repeat tract as: (1) part of intron1; (2) as part of the 3′ untranslated region;or (3) as encoding either a polyleucineor polyalanine tract. One theory is that

haploinsufficiency of the Ca++ regulatingbrain- and testes-specific junctophilin-3protein is responsible for HDL2. WhileJPH3 knockout mice display motor inco-ordination, no histological abnormalitiesare found in their brains. Further workwith these mice is ongoing to evaluatethis hypothesis. Evidence against simplehaploinsufficiency has come from study ofHDL2 patient’s brain material, however.1C2 and ubiquitin antibody immunos-taining reveal intranuclear inclusions inneurons throughout the brains of thesepatients, with dramatic similarity in dis-tribution to patients with classic HD(Fig. 14). At this time, the molecular ba-sis of HDL2 is unknown; however, theincredible overlap between HD and HDL2in terms of clinical phenotype and neu-ropathology strongly suggests that solvingHDL2 should have profound implicationsfor our mechanistic understanding of HD.

Fig. 14 HDL2 patients have1C2-positive intranuclear inclusions.1C2 staining of cerebral cortex (frontallobe) from an HDL2 patient reveals aprominent intranuclear inclusion thatresembles the intranuclear inclusionsseen in classic HD patients. As the 1C2antibody is directed against expanded,misfolded polyglutamine tracts, whichare not predicted to be expressed fromthe causal HDL2 gene, the explanationfor the presence of such nuclearinclusions in HDL2 patients remainsunknown. (From Margolis, R.L.,O’Hearn, E., Rosenblatt, A., Willour, V.,Holmes, S.E., Franz, M.L., Callahan, C.,Hwang, H.S., Troncoso, J.C., Ross, C.A.(2001) A disorder similar toHuntington’s disease is associated witha novel CAG repeat expansion, Ann.Neurol. 50, 373–380, used withpermission of Annals of Neurology, andJohn Wiley & Sons, publisher).


See also Genetics, Molecular Ba-sis of; Motor Neuron Diseases:Cellular and Animal Models; Mo-tor Neuron Diseases: MolecularMechanism, Pathophysiology, andTreatments; Noncoding TandemlyRepeated DNA Sequences.

Bibliography

Books and Reviews

Brown, L.Y., Brown, S.A. (2004) Alanine tracts:the expanding story of human illness andtrinucleotide repeats, Trends Genet. 20, 51–58.

Harper, P.S. (2001) Myotonic Dystrophy,Saunders, London, UK.

Harper, P.S., Perutz, M. (2001) GlutamineRepeats and Neurodegenerative Diseases:Molecular Aspects, Oxford University Press,New York.

Jin, P., Alisch, R.S., Warren, S.T. (2004) RNAand microRNAs in fragile X mentalretardation, Nat. Cell Biol. 6, 1048–1053.

Ranum, L.P., Day, J.W. (2004) Pathogenic RNArepeats: an expanding role in genetic disease,Trends Genet. 20, 506–512.

Wells, R.D., Warren, S.T. (1998) GeneticInstabilities and Hereditary NeurologicalDiseases, Academic Press, San Diego, CA.

Primary Literature

Albrecht, M., Golatta, M., Wullner, U.,Lengauer, T. (2004) Structural and functionalanalysis of ataxin-2 and ataxin-3. Eur. J.Biochem. 271, 3155–3170.

Amiel, J., Laudier, B., Attie-Bitach, T., Trang, H.,de Pontual, L., Gener, B., Trochet, D.,Etchevers, H., Ray, P., Simonneau, M. et al.(2003) Polyalanine expansion and frameshiftmutations of the paired-like homeobox genePHOX2B in congenital central hypoventilationsyndrome, Nat. Genet. 33, 459–461.

Arrasate, M., Mitra, S., Schweitzer, E.S., Se-gal, M.R., Finkbeiner, S. (2004) Inclusion bodyformation reduces levels of mutant huntingtin

and the risk of neuronal death, Nature. 431,805–810.

Banfi, S., Servadio, A., Chung, M., Capozzoli, F.,Duvick, L.A., Elde, R., Zoghbi, H.Y., Orr, H.T.(1996) Cloning and developmental expressionanalysis of the murine homolog of thespinocerebellar ataxia type 1 gene (Sca1), Hum.Mol. Genet. 5, 33–40.

Brais, B., Bouchard, J.P., Xie, Y.G., Rochefort,D.L., Chretien, N., Tome, F.M., Lafre-niere, R.G., Rommens, J.M., Uyama, E., No-hira, O. et al. (1998) Short GCG expansionsin the PABP2 gene cause oculopharyngealmuscular dystrophy, Nat. Genet. 18, 164–167.

Brook, J.D., McCurrach, M.E., Harley, H.G.,Buckler, A.J., Church, D., Aburatani, H.,Hunter, K., Stanton, V.P., Thirion, J.P., Hud-son, T. et al. (1992) Molecular basis of my-otonic dystrophy: expansion of a trinucleotide(CTG) repeat at the 3′ end of a transcript en-coding a protein kinase family member, Cell.69, 385.

Brown, V., Jin, P., Ceman, S., Darnell, J.C.,O’Donnell, W.T., Tenenbaum, S.A., Jin, X.,Feng, Y., Wilkinson, K.D., Keene, J.D. et al.(2001) Microarray identification of FMRP-associated brain mRNAs and altered mRNAtranslational profiles in fragile X syndrome,Cell. 107, 477–487.

Brown, L.Y., Odent, S., David, V., Blayau, M.,Dubourg, C., Apacik, C., Delgado, M.A.,Hall, B.D., Reynolds, J.F., Sommer, A. et al.(2001) Holoprosencephaly due to mutations inZIC2: alanine tract expansion mutations maybe caused by parental somatic recombination,Hum. Mol. Genet. 10, 791–796.

Burnett, B., Li, F., Pittman, R.N. (2003) Thepolyglutamine neurodegenerative proteinataxin-3 binds polyubiquitynated proteins andhas ubiquitin protease activity, Hum. Mol.Genet. 12, 3195–3205.

Calado, A., Tome, F.M., Brais, B., Rouleau, G.A.,Kuhn, U., Wahle, E., Carmo-Fonseca, M.(2000) Nuclear inclusions in oculopharyngealmuscular dystrophy consist of poly(A) bindingprotein 2 aggregates which sequesterpoly(A) RNA, Hum. Mol. Genet. 9, 2321–2328.

Chen, S., Peng, G.H., Wang, X., Smith, A.C.,Grote, S.K., Sopher, B.L., La Spada, A.R.(2004) Interference of Crx-dependenttranscription by ataxin-7 involves interactionbetween the glutamine regions and requiresthe ataxin-7 carboxy-terminal region for


nuclear localization, Hum. Mol. Genet. 13,53–67.

Chen, H.K., Fernandez-Funez, P., Acevedo, S.F.,Lam, Y.C., Kaytor, M.D., Fernandez, M.H.,Aitken, A., Skoulakis, E.M., Orr, H.T.,Botas, J. et al. (2003) Interaction of Akt-phosphorylated ataxin-1 with 14-3-3 mediatesneurodegeneration in spinocerebellar ataxiatype 1, Cell. 113, 457–468.

Cummings, C.J., Reinstein, E., Sun, Y., An-talffy, B., Jiang, Y., Ciechanover, A., Orr, H.T.,Beaudet, A.L., Zoghbi, H.Y. (1999) Mutationof the E6-AP ubiquitin ligase reduces nu-clear inclusion frequency while acceleratingpolyglutamine-induced pathology in SCA1mice, Neuron. 24, 879–892.

David, G., Abbas, N., Stevanin, G., Durr, A.,Yvert, G., Cancel, G., Weber, C., Imbert, G.,Saudou, F., Antoniou, E. et al. (1997) Cloningof the SCA7 gene reveals a highly unstableCAG repeat expansion, Nat. Genet. 17, 65–70.

Davies, S.W., Turmaine, M., Cozens, B.A.,DiFiglia, M., Sharp, A.H., Ross, C.A.,Scherzinger, E., Wanker, E.E., Mangiarini, L.,Bates, G.P. (1997) Formation of neuronalintranuclear inclusions underlies theneurological dysfunction in mice transgenicfor the HD mutation, Cell. 90, 537–548.

Day, J.W., Ricker, K., Jacobsen, J.F., Ras-mussen, L.J., Dick, K.A., Kress, W., Schnei-der, C., Koch, M.C., Beilman, G.J., Harri-son, A.R. et al. (2003) Myotonic dystrophy type2: molecular, diagnostic and clinical spectrum,Neurology. 60, 657–664.

De Baere, E., Beysen, D., Oley, C., Lorenz, B.,Cocquet, J., De Sutter, P., Devriendt, K.,Dixon, M., Fellous, M., Fryns, J.P. et al. (2003)FOXL2 and BPES: mutational hotspots,phenotypic variability, and revision of thegenotype-phenotype correlation, Am. J. Hum.Genet. 72, 478–487.

DiFiglia, M., Sapp, E., Chase, K.O., Davies, S.W.,Bates, G.P., Vonsattel, J.P., Aronin, N. (1997)Aggregation of huntingtin in neuronalintranuclear inclusions and dystrophicneurites in brain, Science. 277, 1990–1993.

Dunah, A.W., Jeong, H., Griffin, A., Kim, Y.M.,Standaert, D.G., Hersch, S.M., Mouradian,M.M., Young, A.B., Tanese, N., Krainc, D.(2002) Sp1 and TAFII130 transcriptionalactivity disrupted in early Huntington’sdisease, Science. 296, 2238–2243.

Ellerby, L.M., Hackam, A.S., Propp, S.S., Ellerby,H.M., Rabizadeh, S., Cashman, N.R.,

Trifiro, M.A., Pinsky, L., Wellington, C.L.,Salvesen, G.S. et al. (1999) Kennedy’s disease:caspase cleavage of the androgen receptor is acrucial event in cytotoxicity, J. Neurochem. 72,185–195.

Fu, Y.H., Kuhl, D.P., Pizzuti, A., Pieretti, M.,Sutcliffe, J.S., Richards, S., Verkerk, A.J.,Holden, J.J., Fenwick, R.G., Jr., Warren, S.T.,et al. (1991) Variation of the CGG repeat atthe fragile X site results in genetic instability:resolution of the Sherman paradox, Cell. 67,1047–1058.

Garden, G.A., Libby, R.T., Fu, Y.H., Kino-shita, Y., Huang, J., Possin, D.E., Smith,A.C., Martinez, R.A., Fine, G.C., Grote, S.K.,et al. (2002) Polyglutamine-expanded ataxin-7promotes noncell-autonomous Purkinje celldegeneration and displays proteolytic cleavagein ataxic transgenic mice, J. Neurosci. 22,4897–4905.

Gecz, J. (2000) The FMR2 gene, FRAXE and non-specific X-linked mental retardation: clinicaland molecular aspects, Ann. Hum. Genet. 64,95–106.

Goodman, F.R., Scambler, P.J. (2001) HumanHOX gene mutations, Clin. Genet. 59, 1–11.

Grabczyk, E., Usdin, K. (2000) The GAA*TTCtriplet repeat expanded in Friedreich’s ataxiaimpedes transcription elongation by T7RNA polymerase in a length and supercoildependent manner, Nucleic Acids Res. 28,2815–2822.

Greene, E., Handa, V., Kumari, D., Usdin, K.(2003) Transcription defects induced by repeatexpansion: fragile X syndrome, FRAXE mentalretardation, progressive myoclonus epilepsytype 1, and Friedreich ataxia, Cytogenet.Genome Res. 100, 65–76.

Hagerman, R.J., Leehey, M., Heinrichs, W.,Tassone, F., Wilson, R., Hills, J., Grigsby, J.,Gage, B., Hagerman, P.J. (2001) Intentiontremor, parkinsonism, and generalized brainatrophy in male carriers of fragile X, Neurology.57, 127–130.

Helmlinger, D., Hardy, S., Sasorith, S., Klein, F.,Robert, F., Weber, C., Miguet, L., Potier, N.,Van-Dorsselaer, A., Wurtz, J.M., et al. (2004)Ataxin-7 is a subunit of GCN5 histoneacetyltransferase-containing complexes, Hum.Mol. Genet. 13, 1257–1265.

Hillman, M.A., Gecz, J. (2001) Fragile XE-associated familial mental retardation protein2 (FMR2) acts as a potent transcriptionactivator, J. Hum. Genet. 46, 251–259.


Hodgson, J.G., Agopyan, N., Gutekunst, C.A.,Leavitt, B.R., LePiane, F., Singaraja, R., Smith,D.J., Bissada, N., McCutcheon, K., Nasir, J.,et al. (1999) A YAC mouse model forHuntington’s disease with full-length mutanthuntingtin, cytoplasmic toxicity, and selectivestriatal neurodegeneration, Neuron. 23,181–192.

Holmes, S.E., O’Hearn, E., McInnis, M.G.,Gorelick-Feldman, D.A., Kleiderlein, J.J.,Callahan, C.A., Ingersoll-Ashworth, R.G.,Sherr, M., Sumner, A.J. (1999) Expansion ofa novel CAG repeat in the 5′ region of a geneencoding a subunit of protein phosphatase 2Ais associated with spinocerebellar ataxia type12, Am. J. Hum. Genet. 65, A4.

Holmes, S.E., O’Hearn, E., Rosenblatt, A., Calla-han, C., Hwang, H.S., Ingersoll-Ashworth,R.G., Fleisher, A., Stevanin, G., Brice, A., Pot-ter, N.T., et al. (2001) A repeat expansion inthe gene encoding junctophilin-3 is associatedwith Huntington disease-like 2, Nat. Genet. 29,377–378.

Huntington’s Disease, Collaborative, Researchand Group (1993) A novel gene containinga trinucleotide repeat that is expandedand unstable on Huntington’s diseasechromosomes, Cell 72, 971–983.

Huynh, D.P., Figueroa, K., Hoang, N., Pulst,S.M. (2000) Nuclear localization or inclusionbody formation of ataxin-2 are not necessaryfor SCA2 pathogenesis in mouse or human,Nat. Genet. 26, 44–50.

Ishikawa, K., Fujigasaki, H., Saegusa, H.,Ohwada, K., Fujita, T., Iwamoto, H., Komat-suzaki, Y., Toru, S., Toriyama, H., Watan-abe, M., et al. (1999) Abundant expression andcytoplasmic aggregations of [alpha]1A voltage-dependent calcium channel protein associatedwith neurodegeneration in spinocerebellarataxia type 6, Hum. Mol. Genet. 8, 1185–1193.

Jiang, H., Mankodi, A., Swanson, M.S., Mox-ley, R.T., Thornton, C.A. (2004) Myotonic dys-trophy type 1 is associated with nuclear fociof mutant RNA, sequestration of muscleblindproteins and deregulated alternative splicingin neurons, Hum. Mol. Genet. 13, 3079–3088.

Katsuno, M., Adachi, H., Kume, A., Li, M., Nak-agomi, Y., Niwa, H., Sang, C., Kobayashi, Y.,Doyu, M., Sobue, G. (2002) Testosterone re-duction prevents phenotypic expression in atransgenic mouse model of spinal and bulbarmuscular atrophy, Neuron. 35, 843–854.

Kitamura, K., Yanazawa, M., Sugiyama, N.,Miura, H., Iizuka-Kogo, A., Kusaka, M.,Omichi, K., Suzuki, R., Kato-Fukui, Y., Kami-irisa, K., et al. (2002) Mutation of ARX causesabnormal development of forebrain and testesin mice and X-linked lissencephaly with ab-normal genitalia in humans, Nat. Genet. 32,359–369.

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

Koob, M.D., Moseley, M.L., Schut, L.J., Ben-zow, K.A., Bird, T.D., Day, J.W., Ranum, L.P.(1999) An untranslated CTG expansion causesa novel form of spinocerebellar ataxia (SCA8),Nat. Genet. 21, 379–384.

Kremer, E.J., Pritchard, M., Lynch, M., Yu, S.,Holman, K., Baker, E., Warren, S.T., Sch-lessinger, D., Sutherland, G.R., Richards, R.I.(1991) Mapping of DNA instability at thefragile X to a trinucleotide repeat sequencep(CCG)n, Science. 252, 1711–1714.

La Spada, A.R., Wilson, E.M., Lubahn, D.B.,Harding, A.E., Fischbeck, K.H. (1991) Andro-gen receptor gene mutations in X-linked spinaland bulbar muscular atrophy, Nature. 352,77–79.

La Spada, A.R., Roling, D.B., Harding, A.E.,Warner, C.L., Spiegel, R., HausmanowaPetrusewicz, I., Yee, W.C., Fischbeck, K.H.(1992) Meiotic stability and genotype-phenotype correlation of the trinucleotiderepeat in X-linked spinal and bulbar muscularatrophy, Nat. Genet. 2, 301–304.

La Spada, A.R., Fu, Y., Sopher, B.L., Libby, R.T.,Wang, X., Li, L.Y., Einum, D.D., Huang, J.,Possin, D.E., Smith, A.C., et al. (2001)Polyglutamine-expanded ataxin-7 antagonizesCRX function and induces cone-rod dystrophyin a mouse model of SCA7, Neuron. 31,913–927.

Lafreniere, R.G., Rochefort, D.L., Chretien, N.,Rommens, J.M., Cochius, J.I., Kalviainen, R.,Nousiainen, U., Patry, G., Farrell, K., Soder-feldt, B. et al. (1997) Unstable insertion in the5′ flanking region of the cystatin B gene is themost common mutation in progressive my-oclonus epilepsy type 1, EPM1, Nat. Genet. 15,298–302.

Lalioti, M.D., Antonarakis, S.E., Scott, H.S.(2003) The epilepsy, the protease inhibitor


and the dodecamer: progressive myoclonusepilepsy, cystatin b and a 12-mer repeatexpansion, Cytogenet. Genome Res. 100,213–223.

Lalioti, M.D., Scott, H.S., Antonarakis, S.E.(1999) Altered spacing of promoter elementsdue to the dodecamer repeat expansioncontributes to reduced expression of thecystatin B gene in EPM1, Hum. Mol. Genet.8, 1791–1798.

Lin, X., Antalffy, B., Kang, D., Orr, H.T.,Zoghbi, H.Y. (2000) Polyglutamine expansiondown-regulates specific neuronal genes beforepathologic changes in SCA1, Nat. Neurosci. 3,157–163.

Liquori, C.L., Ricker, K., Moseley, M.L., Jacob-sen, J.F., Kress, W., Naylor, S.L., Day, J.W.,Ranum, L.P. (2001) Myotonic dystrophy type2 caused by a CCTG expansion in intron 1 ofZNF9, Science. 293, 864–867.

Loesch, D.Z., Huggins, R.M., Bui, Q.M., Ep-stein, J.L., Taylor, A.K., Hagerman, R.J. (2002)Effect of the deficits of fragile X mental retar-dation protein on cognitive status of fragile xmales and females assessed by robust pedigreeanalysis, J. Dev. Behav. Pediatr. 23, 416–423.

Lunkes, A., Lindenberg, K.S., Ben-Haiem, L.,Weber, C., Devys, D., Landwehrmeyer, G.B.,Mandel, J.L., Trottier, Y. (2002) Proteasesacting on mutant huntingtin generatecleaved products that differentially build upcytoplasmic and nuclear inclusions, Mol. Cell.10, 259–269.

Mangiarini, L., Sathasivam, K., Seller, M., Coz-ens, B., Harper, A., Hetherington, C., Law-ton, M., Trottier, Y., Lehrach, H., Davies, S.W.,et al. (1996) Exon 1 of the HD gene with anexpanded CAG repeat is sufficient to cause aprogressive neurological phenotype in trans-genic mice, Cell. 87, 493–506.

Mankodi, A., Logigian, E., Callahan, L., Mc-Clain, C., White, R., Henderson, D., Krym, M.,Thornton, C.A. (2000) Myotonic dystrophy intransgenic mice expressing an expanded CUGrepeat, Science. 289, 1769–1773.

Margolis, R.L., O’Hearn, E., Rosenblatt, A.,Willour, V., Holmes, S.E., Franz, M.L., Calla-han, C., Hwang, H.S., Troncoso, J.C., Ross,C.A. (2001) A disorder similar to Hunting-ton’s disease is associated with a novel CAGrepeat expansion, Ann. Neurol. 50, 373–380.

Matsuura, T., Fang, P., Lin, X., Khajavi, M.,Tsuji, K., Rasmussen, A., Grewal, R.P.,Achari, M., Alonso, M.E., Pulst, S.M., et al.

(2004) Somatic and germline instability of theATTCT repeat in spinocerebellar ataxia type10, Am. J. Hum. Genet. 74, 1216–1224.

Matsuura, T., Yamagata, T., Burgess, D.L.,Rasmussen, A., Grewal, R.P., Watase, K.,Khajavi, M., McCall, A.E., Davis, C.F., Zu, L.,et al. (2000) Large expansion of the ATTCTpentanucleotide repeat in spinocerebellarataxia type 10, Nat. Genet. 26, 191–194.

McCampbell, A., Taylor, J.P., Taye, A.A., Ro-bitschek, J., Li, M., Walcott, J., Merry, D.,Chai, Y., Paulson, H., Sobue, G., et al. (2000)CREB-binding protein sequestration by ex-panded polyglutamine, Hum. Mol. Genet. 9,2197–2202.

Monckton, D.G., Cayuela, M.L., Gould, F.K.,Brock, G.J., Silva, R., Ashizawa, T. (1999) Verylarge (CAG)(n) DNA repeat expansions in thesperm of two spinocerebellar ataxia type 7males, Hum. Mol. Genet. 8, 2473–2478.

Mortlock, D.P., Innis, J.W. (1997) Mutation ofHOXA13 in hand-foot-genital syndrome, Nat.Genet. 15, 179–180.

Mosemiller, A.K., Dalton, J.C., Day, J.W.,Ranum, L.P. (2003) Molecular genetics ofspinocerebellar ataxia type 8 (SCA8), Cytogenet.Genome Res. 100, 175–183.

Mundlos, S., Otto, F., Mundlos, C., Mul-liken, J.B., Aylsworth, A.S., Albright, S., Lind-hout, D., Cole, W.G., Henn, W., Knoll, J.H.,et al. (1997) Mutations involving the tran-scription factor CBFA1 cause cleidocranialdysplasia, Cell. 89, 773–779.

Muragaki, Y., Mundlos, S., Upton, J., Olsen, B.R.(1996) Altered growth and branching pat-terns in synpolydactyly caused by mutationsin HOXD13, Science. 272, 548–551.

Mutsuddi, M., Marshall, C.M., Benzow, K.A.,Koob, M.D., Rebay, I. (2004) The spinocere-bellar ataxia 8 noncoding RNA causes neu-rodegeneration and associates with staufen inDrosophila, Curr. Biol. 14, 302–308.

Nagafuchi, S., Yanagisawa, H., Sato, K., Shi-rayama, T., Ohsaki, E., Bundo, M., Takeda, T.,Tadokoro, K., Kondo, I., Murayama, N., et al.(1994) Dentatorubral and pallidoluysian at-rophy expansion of an unstable CAG trinu-cleotide on chromosome 12p, Nat. Genet. 6,14–18.

Nakamura, K., Jeong, S.Y., Uchihara, T.,Anno, M., Nagashima, K., Nagashima, T.,Ikeda, S., Tsuji, S., Kanazawa, I. (2001)SCA17, a novel autosomal dominant cerebellarataxia caused by an expanded polyglutamine


in TATA-binding protein, Hum. Mol. Genet.10, 1441–1448.

Okazawa, H., Rich, T., Chang, A., Lin, X.,Waragai, M., Kajikawa, M., Enokido, Y.,Komuro, A., Kato, S., Shibata, M., et al. (2002)Interaction between mutant ataxin-1 andPQBP-1 affects transcription and cell death,Neuron. 34, 701–713.

Oostra, B.A., Willemsen, R. (2003) A fragilebalance: FMR1 expression levels, Hum. Mol.Genet. 12, Spec No 2, R249–R257.

Orr, H.T., Chung, M.Y., Banfi, S., Kwiatkowski,Jr., T.J., Servadio, A., Beaudet, A.L., Mc-Call, A.E., Duvick, L.A., Ranum, L.P., Zoghbi,H.Y. (1993) Expansion of an unstable trinu-cleotide CAG repeat in spinocerebellar ataxiatype 1, Nat. Genet. 4, 221–226.

Paulson, H.L., Perez, M.K., Trottier, Y., Tro-janowski, J.Q., Subramony, S.H., Das, S.S.,Vig, P., Mandel, J.L., Fischbeck, K.H., Pitt-man, R.N. (1997) Intranuclear inclusions ofexpanded polyglutamine protein in spinocere-bellar ataxia type 3, Neuron. 19, 333–344.

Penrose, L.S. (1948) The problem of anticipationin pedigrees of dystrophica myotonica, Ann.Eugen. 14, 125–132.

Perutz, M.F., Johnson, T., Suzuki, M., Finch, J.T.(1994) Glutamine repeats as polar zippers:their possible role in inherited neurodegener-ative diseases, Proc. Natl. Acad. Sci. U. S. A.91, 5355–5358.

Poirier, M.A., Li, H., Macosko, J., Cai, S.,Amzel, M., Ross, C.A. (2002) Huntingtinspheroids and protofibrils as precursors inpolyglutamine fibrilization, J. Biol. Chem. 277,41032–41037.

Pulst, S.M., Nechiporuk, A., Nechiporuk, T.,Gispert, S., Chen, X.N., Lopes-Cendes, I.,Pearlman, S., Starkman, S., Orozco-Diaz, G.,Lunkes, A., et al. (1996) Moderate expansionof a normally biallelic trinucleotide repeat inspinocerebellar ataxia type 2, Nat. Genet. 14,269–276.

Rasmussen, A., Matsuura, T., Ruano, L.,Yescas, P., Ochoa, A., Ashizawa, T., Alonso, E.(2001) Clinical and genetic analysis of fourMexican families with spinocerebellar ataxiatype 10, Ann. Neurol. 50, 234–239.

Sanchez, I., Mahlke, C., Yuan, J. (2003) Pivotalrole of oligomerization in expandedpolyglutamine neurodegenerative disorders,Nature. 421, 373–379.

Sanpei, K., Takano, H., Igarashi, S., Sato, T., Oy-ake, M., Sasaki, H., Wakisaka, A., Tashiro, K.,

Ishida, Y., Ikeuchi, T., et al. (1996) Identifica-tion of the spinocerebellar ataxia type 2 geneusing a direct identification of repeat expan-sion and cloning technique, DIRECT, Nat.Genet. 14, 277–284.

Satterfield, T.F., Jackson, S.M., Pallanck, L.J.(2002) A Drosophila homolog of thepolyglutamine disease gene SCA2 is a dosage-sensitive regulator of actin filament formation,Genetics. 162, 1687–1702.

Scherzinger, E., Lurz, R., Turmaine, M., Man-giarini, L., Hollenbach, B., Hasenbank, R.,Bates, G.P., Davies, S.W., Lehrach, H.,Wanker, E.E. (1997) Huntingtin-encodedpolyglutamine expansions form amyloid-likeprotein aggregates in vitro and in vivo, Cell. 90,549–558.

Schilling, G., Wood, J.D., Duan, K., Slunt, H.H.,Gonzales, V., Yamada, M., Cooper, J.K.,Margolis, R.L., Jenkins, N.A., Copeland, N.G.,et al. (1999) Nuclear accumulation of truncatedatrophin-1 fragments in a transgenic mousemodel of DRPLA, Neuron. 24, 275–286.

Seznec, H., Simon, D., Monassier, L., Criqui-Filipe, P., Gansmuller, A., Rustin, P.,Koenig, M., Puccio, H. (2004) Idebenonedelays the onset of cardiac functional alterationwithout correction of Fe-S enzymes deficit in amouse model for Friedreich ataxia, Hum. Mol.Genet. 13, 1017–1024.

Sopher, B.L., Thomas, P.S., Jr., LaFevre-Bernt,M.A., Holm, I.E., Wilke, S.A., Ware, C.B., Jin,L.W., Libby, R.T., Ellerby, L.M., La Spada, A.R.(2004) Androgen receptor YAC transgenicmice recapitulate SBMA motor neuronopathyand implicate VEGF164 in the motor neurondegeneration, Neuron. 41, 687–699.

Stevanin, G., Camuzat, A., Holmes, S.E.,Julien, C., Sahloul, R., Dode, C., Hahn-Barma, V., Ross, C.A., Margolis, R.L.,Durr, A., et al. (2002) CAG/CTG repeatexpansions at the Huntington’s disease-like2 locus are rare in Huntington’s diseasepatients, Neurology. 58, 965–967.

Stromme, P., Mangelsdorf, M.E., Shaw, M.A.,Lower, K.M., Lewis, S.M., Bruyere, H.,Lutcherath, V., Gedeon, A.K., Wallace, R.H.,Scheffer, I.E., et al. (2002) Mutations in thehuman ortholog of Aristaless cause X-linkedmental retardation and epilepsy, Nat. Genet.30, 441–445.

Sutherland, G.R., Baker, E. (1992) Characterisa-tion of a new rare fragile site easily confused


with the fragile X, Hum. Mol. Genet. 1,111–113.

Uda, M., Ottolenghi, C., Crisponi, L., Garcia,J.E., Deiana, M., Kimber, W., Forabosco, A.,Cao, A., Schlessinger, D., Pilia, G. (2004)Foxl2 disruption causes mouse ovarian failureby pervasive blockage of follicle development,Hum. Mol. Genet. 13, 1171–1181.

Ueno, S., Kondoh, K., Kotani, Y., Komure, O.,Kuno, S., Kawai, J., Hazama, F., Sano, A.(1995) Somatic mosaicism of CAGrepeat in dentatorubral-pallidoluysian atrophy(DRPLA), Hum. Mol. Genet. 4, 663–666.

Wallis, J., Williamson, R., Chamberlain, S.(1990) Identification of a hypervariablemicrosatellite polymorphism within D9S15tightly linked to Friedreich’s ataxia, Hum.Genet. 85, 98–100.

Warren, S.T. (1997) Polyalanine expansion insynpolydactyly might result from unequalcrossing-over of HOXD13, Science. 275,408–409.

Watase, K., Weeber, E.J., Xu, B., Antalffy, B.,Yuva-Paylor, L., Hashimoto, K., Kano, M.,Atkinson, R., Sun, Y., Armstrong, D.L., et al.(2002) A long CAG repeat in the mouse Sca1locus replicates SCA1 features and revealsthe impact of protein solubility on selectiveneurodegeneration, Neuron. 34, 905–919.

Wellington, C.L., Ellerby, L.M., Gutekunst, C.A.,Rogers, D., Warby, S., Graham, R.K., Loub-ser, O., van Raamsdonk, J., Singaraja, R.,Yang, Y.Z., et al. (2002) Caspase cleavage ofmutant huntingtin precedes neurodegenera-tion in Huntington’s disease, J. Neurosci. 22,7862–7872.

Wellington, C.L., Ellerby, L.M., Hackam, A.S.,Margolis, R.L., Trifiro, M.A., Singaraja, R.,McCutcheon, K., Salvesen, G.S., Propp, S.S.,Bromm, M., et al. (1998) Caspase cleavageof gene products associated with tripletexpansion disorders generates truncatedfragments containing the polyglutamine tract,J. Biol. Chem. 273, 9158–9167.

Willemsen, R., Hoogeveen-Westerveld, M., Reis,S., Holstege, J., Severijnen, L.A., Nieuwen-huizen, I.M., Schrier, M., van Unen, L.,

Tassone, F., Hoogeveen, A.T., et al. (2003) TheFMR1 CGG repeat mouse displays ubiquitin-positive intranuclear neuronal inclusions;implications for the cerebellar tremor/ataxiasyndrome, Hum. Mol. Genet. 12, 949–959.

Wirtschafter, J.D., Ferrington, D.A., McLoon,L.K. (2004) Continuous remodeling ofadult extraocular muscles as an explanationfor selective craniofacial vulnerability inoculopharyngeal muscular dystrophy, J.Neuroophthalmol. 24, 62–67.

Yoo, S.Y., Pennesi, M.E., Weeber, E.J., Xu, B.,Atkinson, R., Chen, S., Armstrong, D.L.,Wu, S.M., Sweatt, J.D., Zoghbi, H.Y. (2003)SCA7 knockin mice model human SCA7 andreveal gradual accumulation of mutant ataxin-7 in neurons and abnormalities in short-termplasticity, Neuron. 37, 383–401.

Yvert, G., Lindenberg, K.S., Picaud, S., Landwe-hrmeyer, G.B., Sahel, J.A., Mandel, J.L. (2000)Expanded polyglutamines induce neurode-generation and trans-neuronal alterations incerebellum and retina of SCA7 transgenicmice, Hum. Mol. Genet. 9, 2491–2506.

Zakany, J., Duboule, D. (1996) Synpolydactyly inmice with a targeted deficiency in the HoxDcomplex, Nature. 384, 69–71.

Zhang, S., Xu, L., Lee, J., Xu, T. (2002)Drosophila atrophin homolog functionsas a transcriptional corepressor inmultiple developmental processes, Cell 108,45–46.

Zhuchenko, O., Bailey, J., Bonnen, P., Ashizawa,T., Stockton, D.W., Amos, C., Dobyns, W.B.,Subramony, S.H., Zoghbi, H.Y., Lee, C.C.(1997) Autosomal dominant cerebellar ataxia(SCA6) associated with small polyglutamineexpansions in the alpha 1A-voltage-dependent calcium channel, Nat. Genet. 15,62–69.

Zuccato, C., Tartari, M., Crotti, A., Goffredo, D.,Valenza, M., Conti, L., Cataudella, T., Leav-itt, B.R., Hayden, M.R., Timmusk, T., et al.(2003) Huntingtin interacts with REST/NRSFto modulate the transcription of NRSE-controlled neuronal genes, Nat. Genet. 35,76–83.


Tumor Immunology: seeImmunology of Cancer

Tumorigenesis, EpigeneticMechanisms: see EpigeneticMechanisms in Tumorigenesis

Documents

Triplet Repeat Diseases - Wiley-VCH · 2 Triplet Repeat Diseases 4.2 Fragile XE Mental Retardation 34 4.3 Friedreich’s Ataxia 36 4.4 Progressive Myoclonus Epilepsy Type 1 39 5 Type