1

Triplet Repeat Diseases

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

1 A Novel Mechanism of Genetic Mutation Emerges 3 1.1 Repeat Sequences of All Types and Sizes 3 1.2 Trinucleotide Repeat Expansion as a Cause of Disease: Unique Features Explain Unusual Genetics 4

2 Repeat Diseases and Their Classification 5 2.1 Summary of Repeat Diseases 5 2.2 Differences in Repeat Sequence Composition and Location within Gene 6 2.3 Classification Based upon Mechanism of Pathogenesis and Nature of Mutation 6

3 Type 1: The CAG/Polyglutamine Repeat Diseases 9 3.1 Spinal and Bulbar Muscular Atrophy 9 3.2 Huntington’s Disease 12 3.3 Dentatorubral Pallidoluysian Atrophy 15 3.4 Spinocerebellar Ataxia Type 1 16 3.5 Spinocerebellar Ataxia Type 2 18 3.6 Spinocerebellar Ataxia Type 3/Machado–Joseph Disease 19 3.7 Spinocerebellar Ataxia Type 6 21 3.8 Spinocerebellar Ataxia Type 7 22 3.9 Spinocerebellar Ataxia Type 17 25 3.10 Role of Aggregation in Polyglutamine Disease Pathogenesis 25 3.11 Protein Context 29 3.12 Transcriptional Dysregulation 30 3.13 Proteolytic Cleavage 30

4 Type 2: the Loss-of-function Repeat Diseases 33 4.1 Fragile X Syndrome 33

Encyclopedia of Molecular Cell Biology and Molecular Medicine, 2nd Edition. Volume 15 Edited by Robert A. Meyers. Copyright  2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30652-8 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 3: the RNA Gain-of-function Repeat Diseases 40 5.1 Myotonic Dystrophy Type 1 40 5.2 Myotonic Dystrophy Type 2 42 5.3 Spinocerebellar Ataxia Type 8 43 5.4 The Fragile X Tremor – Ataxia Syndrome (FXTAS) 44

6 Type 4: The Polyalanine Diseases 46 6.1 Overview 46 6.2 Oculopharyngeal Muscular dystrophy 46 6.3 Synpolydactyly (Syndactyly Type II) 48 6.4 Cleidocranial Dysplasia 48 6.5 Holoprosencephaly 48 6.6 Hand-foot-genital Syndrome 49 6.7 Blepharophimosis-ptosis-epicanthus Inversus Syndrome 49 6.8 Syndromic and Nonsyndromic X-linked Mental Retardation 49 6.9 Congenital Central Hypoventilation Syndrome 50

7 Unclassified Repeat Diseases Lacking Mechanistic Explanations 50 7.1 Spinocerebellar Ataxia Type 10 50 7.2 Spinocerebellar Ataxia Type 12 51 7.3 Huntington’s Disease Like 2 (HDL2) 52

Bibliography 54 Books and Reviews 54 Primary Literature 54

Keywords

Aggregate Accumulation of proteinaceous and/or ribonucleic acid material into a structure that is visible in a cell at the light microscope level.

Anticipation Worsening severity of a disease phenotype as the causal (typically dominant) genetic mutation is transmitted from one generation to the next in a family segregating the disease of interest. Triplet Repeat Diseases 3

Gain of function Refers to a type of mutation that imparts a novel activity or action to the gene product containing the mutation.

Loss of function Refers to a type of mutation that eliminates the action of the gene product encoded by the gene within which the mutation resides.

Repeat Expansion An elongation of a repeat to a larger size that no longer falls within the size distribution range typically seen in the normal population; this process is now recognized as a mechanism of human genetic mutation.

Trinucleotide Three 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 by the elongation of a DNA repeat sequence. In this chapter, we provide an overview of the discovery of repeat expansion as an important cause of human disease, and we summarize the molecular genetics and mechanistic basis of 27 microsatellite repeat disorders. Comparison of the many repeat expansion disorders reveals distinct categories of repeat diseases, allowing us to propose a classification of the repeat expansion disorders based upon mutation sequence and pathogenic mechanism. The four types of repeat expansion disorders defined by this approach are the CAG/polyglutamine repeat diseases; the loss-of-function repeat diseases; the RNA gain-of-function repeat diseases; and the polyalanine diseases. Although the genetic basis for most of these diseases was determined less than a decade or so ago, considerable advances have been made in our understanding of how ‘‘dynamic mutations’’ produce molecular pathology and human disease.

1 cleaved them, yielded a new methodology A Novel Mechanism of Genetic Mutation for differentiating individuals (‘‘molec- Emerges ular fingerprinting’’). This methodology 1.1 also found application in the mapping of Repeat Sequences of All Types and Sizes inherited genetic diseases, taking advan- tage of human variation in the form of Long before the sequencing of the human so-called restriction fragment length poly- genome was undertaken, the discovery of morphisms (RFLP’s). In the search for bacterial enzymes that recognize specific even more informative genetic markers, DNA sequences (‘‘recognition sites’’) and investigators uncovered a variety of very 4 Triplet Repeat Diseases

short repeat sequences (‘‘short tandem independently made paradigm-shifting repeats’’ – STR’s; ‘‘simple sequence length discoveries. In one case, a CAG trinu- polymorphisms’’ – SSLP’s) that were cleotide repeat expansion within the first highly dispersed throughout the human coding exon of the androgen receptor (AR) genome. These latter repeat sequences gene was found to be the cause of an came to be known as ‘‘microsatellites’’ to X-linked neuromuscular disorder known differentiate them from the ‘‘minisatel- as spinal and bulbar muscular atrophy lites’’ that were being used for molec- (SBMA or Kennedy’s disease). CAG en- ular fingerprinting. Minisatellites were codes the amino acid glutamine; thus, originally defined as tandem arrays of elongation of a polyglutamine tract within 14–100 bp repeating sequences. In the the AR protein was hypothesized to be case of minisatellites, the repeat sequence the molecular basis for the motor neu- was in essence a ‘‘consensus’’ as devi- ron degeneration in SBMA. In the other ations from the exact repeat sequence case, a disorder known as the fragile X were common. Microsatellites, however, syndrome of mental retardation (FRAXA), were typically pure perfect repeats of also X-linked but much more common, less than 13 bp, the most common mi- was reported to result from expansion of crosatellites being dinucleotide repeats, a CGG repeat. In the latter case, although trinucleotide repeats, or tetranucleotide re- originally envisioned to encode a polyargi- peats. The discovery of minisatellites and nine tract, the CGG repeat turned out microsatellites yielded a virtual bonanza to be in the 5 untranslated region of a of reagents for molecular fingerprinting novel gene, the so-called FMR-1 gene (for and genetic linkage mapping, while also ‘‘fragile X mental retardation-1’’). providing evolutionary biologists and pop- The identification of triplet repeat ex- ulation geneticists with intriguing material pansions as the cause of two inherited to attempt to reconstruct evolutionary diseases was an exciting turning point in relationships and inter- or intraspecies re- the field of human molecular genetics not lationships. Although human geneticists only because of the novel nature of these and evolutionary biologists were applying findings, but also because of the unusual microsatellites in different ways for their genetic characteristics of this new type of own studies, they shared the commonly mutation. Analysis of families segregat- held belief that such repeats were neutral ing FRAXA revealed the existence of three and therefore unlikely to be of much func- distinct allele categories defined by the tional consequence, let alone play a role in length of the pure CGG repeat: a normal causing human disease. Of course, all that size range; a disease size range; and an would soon change in the last decade of intermediate, ‘‘premutation’’ size range. the twentieth century. As discussed below, individuals carrying premutation-sized CGG repeats never de- 1.2 velop the mental retardation phenotype, Trinucleotide Repeat Expansion as a but instead are at risk for passing on even Cause of Disease: Unique Features Explain larger CGG repeats to their children or Unusual Genetics grandchildren who then display the men- tal retardation phenotype. An important In 1991, two groups working on seem- tenet that emerged from these early stud- ingly unrelated inherited genetic diseases ies was that expanded CGG repeats (larger Triplet Repeat Diseases 5 than the normal size range) displayed an defining feature of DM in 1918, a num- exceptionally high rate of further mutation. ber of leading geneticists, among them This observation reversed a commonly Penrose, L.S., dismissed its authenticity, held view of the genetic material – the claiming that it was a product of ascertain- notion that any single nucleotide in the hu- ment bias due to better clinically defining man genome displayed a mutation rate of the profound variable expressitivity in this − ∼1 × 10 5. In the case of expanded CGG disorder. This view persisted from its pro- repeats in premutation carriers, the rate mulgation in the 1950s, reinforced by approached unity (100)! Besides displaying the concept that the genetic material is this high mutation rate, there were other seldom subject to alteration or modifi- unusual features: (1) the CGG repeats cation that could be heritable. However, showed a marked tendency to further ex- with the discovery that expanded trinu- pansion, suggesting that repeat mutation cleotide repeats could further expand, and was a polar process; and (2) the sex of the that indeed the expansion process was a individual transmitting the premutation- prerequisite for the FRAXA disease pheno- sized CGG allele determined whether a type – accounting for maternal inheritance large expansion into the disease range and greater percentages of affected individ- would be possible. For FRAXA, expansion uals in more recent generations – a role into the disease range could only occur if for repeats in diseases displaying antici- the premutation allele was transmitted by a pation was entertained. Consequently, the female carrier. All of these unusual aspects third repeat mutation disorder to be iden- of FRAXA CGG repeat genetics thoroughly tified – just one year after the SBMA and accounted for the bizarre non-Mendelian FRAXA discoveries – was DM. Studies of inheritance patterns described in FRAXA the causal CTG repeat expansion in DM families as the ‘‘Sherman paradox.’’ demonstrated a strong correlation between The recognition of repeat expansion as disease severity (i.e. age of onset and rate the cause of the neurodegenerative dis- of progression) and the length of the CTG order SBMA and as an explanation for repeat. In this way, it became clear that the puzzling genetics of FRAXA set the anticipation was a genuine phenomenon stage for further discoveries in the field of and that expanded repeat mutational in- neurogenetics (the study of inherited neu- stability was its long awaited molecular rological disorders). One disorder known explanation. The mutational instability as myotonic dystrophy (DM), the most com- characteristics of expanded repeats have mon of all muscular dystrophies, had been led to their designation as so-called ‘‘dy- the center of a genetic controversy that had namic mutations.’’ gone on for nearly a century. The contro- versy involved the debate over whether the clinical phenomenon of anticipation truly 2 existed or was simply an artifact of clin- Repeat Diseases and Their Classification ical study (Anticipation may be defined as a progressively earlier age of disease 2.1 onset with increasing disease severity in Summary of Repeat Diseases successive generations of a family segre- gating an inherited disorder). Although The list of diseases caused by microsatel- anticipation was initially proposed as a lite repeat expansions (involving repeats of 6 Triplet Repeat Diseases

3–12 bp) now includes more than 25 disor- a gene, which applies to both CAG (glu- ders (Table 1). Certain aspects of the list of tamine) and GCG (alanine) repeats, and repeat diseases deserve emphasis. Twenty has been proposed but not yet demon- of the repeat diseases either principally strated for CTG (leucine) repeats. The rest or exclusively affect the neuraxis – that is, of the locations defined for repeats vary anywhere from the brain to the cerebel- widely, ranging from the gene’s promoter lum/brainstem, to the spinal cord, to the to its 5 untranslated region to an intron peripheral nerve or muscle – and are de- to the 3 untranslated region, and no more generative disorders. Almost all of these in- than two to three repeat expansions can herited neurological disorders are caused be placed in each of these categories at by large expanded repeats that display this time. the property of pronounced genetic insta- bility (dynamic mutation). On the other 2.3 hand, the developmental malformation Classification Based upon Mechanism syndromes for which repeat expansion of Pathogenesis and Nature of Mutation mutations have been implicated all involve modestly sized disease repeats by compar- Toallowustoreconstructhowthedif- ison, and these repeats do not exhibit such ferent repeat expansion mutations cause dynamic mutation genetic instability. molecular pathology in the various dis- orders that they cause, we have chosen 2.2 to categorize the 25 repeat disorders into Differences in Repeat Sequence four classes (Table 2). The first class of Composition and Location within Gene disorders, the Type 1 repeat diseases, are the ‘‘CAG-polyglutamine disorders.’’ When faced with the task of categorizing This class of repeat diseases includes the various repeat expansion diseases into nine inherited neurodegenerative disor- different classes, a number of approaches ders (SBMA, Huntington’s disease, den- are possible. We have found that consid- tatorubral pallidoluysian atrophy, and six eration of the sequence of the repeat and forms of spinocerebellar ataxia) that all its location within the gene are the most share the common feature of being caused useful characteristics to apply for grouping by a CAG repeat located within the coding the different repeat diseases. As shown in Table 1, there are many different types of region of a gene. Upon CAG repeat expan- repeats varying in length and sequence sion, a mutant protein with an extended composition. However, among the recur- polyglutamine tract is produced, mak- rent sequence types are CAG trinucleotide ing the protein then adopt an abnormal repeats, CTG trinucleotide repeats, and conformation and misfold to initiate the GCG trinucleotide repeats. The rest of the pathogenic cascade. The resultant pathol- repeat sequences are unique in composi- ogy is believed to primarily stem from a tion and differ widely in size, ranging from gain of function of the mutant protein 3 to 12 bp as noted above. Comparison of imparted by the expanded polyglutamine the location of a repeat within the gene tract. As discussed below, much effort has it is affecting also yields different types of gone into trying to define what the gain-of- repeats. The largest single-repeat location function effect is for each polyglutamine category is within the coding region of disease protein and into determining if Triplet Repeat Diseases 7 10 10 2000 1000 > > > > 200– a 800 ? 71– > ATTCTCAGCAGGCGCAG 10–22 7–32 25–48 9–15 ? 5–34 – – 800–4500 – – 55–78 43–66 16–29 37–70 CAGCAG 4–18 4–35 – – 19–33 37–306 GCCGAAGCGCAGCAGGCGCCCCGCCCCGCGCTG 6–40CCTG 7–38 2–3GCG 18CAG 6–35CAG 61–200 7–27 33–60 15 12–17 27–36 104–176 5–37 200–1000 – 66–1700 – 6–7 30–150 6–44 39–250 13–33 – 50–80 ? 24–27 50–60 – – 50–4000 – 175–11 000 25 8–17 39– 31– CGG 6–53 45–200 GCG 10–16 – 17–23 CAG 12–40 – 55–84 CAG 7–35 – 49–88 GCGGCGGCG 14 11–17 20 – – – 27 24 25–29 Ataxin-10 PPP2R2B TBP HOX-D13 AR CACNA1A Ataxin-7 Fmr-2 Frataxin HOX-A13 Htt/IT15 JPH3 ZIC2 Cystatin B DMPK ZNF9 PABP2 Ataxin-1 Ataxin-2 Fmr-1 ARX Ataxin-3 Atrophin-1 FOXL2 RUNX2 PHOX-2B develop the Fragile X tremor-ataxia syndrome. n atrophy (Haw River syndrome) Compilation of repeat expansion diseases in humans. Male premutation carriers of the FMR-1 CGG repeat Spinocerebellar ataxia type 10 Spinocerebellar ataxia type 12 Spinocerebellar ataxia type 17 Syndromic and nonsyndromic X-linked mentalSynpolydactyly retardation X-linked spinal & bulbar muscular atrophy Spinocerebellar ataxia type 3 (Machado-JosephSpinocerebellar disease) ataxia type 6 Spinocerebellar ataxia type 7 Spinocerebellar ataxia type 8 ? CTG 16– FRAXE mental retardation Friedreich’s ataxia Hand-foot-genital Huntington’s disease Huntington’s disease like 2 Holoprosencephaly Myoclonus epilepsy type 1 Myotonic dystrophy type 1 Myotonic dystrophy type 2 Oculopharyngeal muscular dystrophy Spinocerebellar ataxia type 1 Spinocerebellar ataxia type 2 Fragile X syndrome of mental retardation –, no premutation alleles identified,?, so existence they of may premutation not alleles exist. is unknown. Tab. 1 DiseaseBlepharophimosis-ptosis-epicanthus inversus Cleidocranial dysplasia Congenital central hypoventilation Dentatorubralpallidoluysia Gene Repeata Normal Premutation Disease 8 Triplet Repeat Diseases tal central hypoventilation epicanthus inversus dystrophy nonsyndromic) Hand-foot-genital syndrome Holoprosencephaly Oculopharyngeal muscular Synpolydactyly X-linked MR (syndromic & phy type 1 Blepharophimosis-ptosis- tremor-ataxia syndrome Congeni Spinocerebellar ataxia type 8 Cleidocranial dysplasia s RNA gain-of-function disorders Polyalanine disorders Fragile X syndrome of MR Myotonic dystro Classification of the repeat expansion diseases. mental retardation. = Spinocerebellar ataxia type 2 Huntington’s diseaseSpinal and bulbar muscular atrophySpinocerebellar ataxia type 1Spinocerebellar ataxia type 3 Spinocerebellar ataxia type Friedreich’s 6 ataxia Spinocerebellar ataxia type 7 Spinocerebellar ataxia type 17 Myoclonus FRAXE epilepsy MR type 1 Fragile X Myotonic dystrophy type 2 Tab. 2 CAG/polyglutamine disordersDentatorubral pallidoluysian atrophy Loss-of-function disorder MR Triplet Repeat Diseases 9 shared pathways of toxicity are initiated in differences. One form of spinocerebellar the different diseases. At least one, and per- ataxia (SCA8) with an unclear mechanism haps a greater number of these disorders of pathogenesis has also been provision- may principally involve a simultaneous ally placed into this category, based upon dominant-negative partial loss of the nor- current working models of how its re- mal function of the disease protein. The peat causes disease. The last class of next class of disorders, the Type 2 repeat repeat disease, the Type 4 disorders, are diseases, are a much more disparate group the ‘‘GCG-polyalanine disorders’’ that are of repeat disorders. The Type 2 repeat dis- grouped together because all involve short eases are the ‘‘Loss-of-function repeat dis- GCG repeat tracts falling within the coding orders.’’ These disorders include different regions of unrelated genes that become ex- repeats that vary in sequence composi- panded to moderately sized GCG repeats. tion and gene location, but share a final With the exception of oculopharyngeal common pathway of disease pathogene- muscular dystrophy, all are developmental sis–alossoffunctionofthediseasegene malformation syndromes, and while gain- within which they occur. This group in- of-function polyalanine toxicity has been cludes various classic trinucleotide repeat proposed for a number of these disorders, disorders such as the two fragile X syn- loss of function due to the polyalanine dromes of mental retardation (FRAXA and expansion seems more likely for others. FRAXE) and Friedreich’s ataxia – but also Finally, a number of repeat disorders, encompasses the dodecamer repeat ex- spinocerebellar ataxia type 10 (SCA10), pansion in progressive myoclonic epilepsy spinocerebellar ataxia type 12 (SCA12), type 1, and possibly the CAG repeat ex- and Huntington’s disease like 2 (HDL2), pansion in Huntington’s disease like-2 currently defy classification because very (HDL2) gene. Strong evidence for a loss-of- little is known about their molecular basis. function pathway in the form of nonrepeat These diseases will be considered in the loss-of-function mutations supports many final section of this chapter. of these classifications. The third group of repeat diseases, the Type 3 disorders, com- prise a shared class because all of them 3 have been proposed to involve the produc- Type 1: The CAG/Polyglutamine Repeat tion of a toxic RNA species. This category Diseases of repeat diseases is thus called the RNA gain-of-function disorders. Included among 3.1 these disorders are two closely related Spinal and Bulbar Muscular Atrophy forms of DM, the common and classic myotonic dystrophy type 1 (DM1) and its Spinal and bulbar muscular atrophy uncommon phenocopy, myotonic dystro- (SBMA; Kennedy’s disease) is a late-onset phy type 2 (DM2). Another member of neurodegenerative disease with an inher- this group is the recently described frag- itance pattern resembling X-linked reces- ile X tremor-ataxia syndrome (FXTAS) in sive. It has a prevalence of about 1 in 50 000 male premutation carriers – a fascinating males. Patients suffer a late-onset, pro- example of two different disease path- gressive degeneration of primarily lower ways operating upon the same expanded motor neurons in the anterior horn of the repeat mutation based upon size range spinal cord and in the bulbar region of the 10 Triplet Repeat Diseases

brainstem; however, variable involvement heat-shock proteins (HSPs) 70 and 90, of sensory neurons in the dorsal root gan- and resides in the cytoplasm. Upon bind- glia also occurs. SBMA typically presents ing androgen, it dissociates from these with cramps, followed by proximal mus- HSP chaperone proteins and translocates cle weakness and atrophy. Patients often to the nucleus. Once in the nucleus, AR exhibit dysarthria, dysphagia, and fascic- dimers transactivate certain genes, many ulations of the tongue and lips. Affected of which are responsible for generating individuals have symptoms of mild andro- and maintaining male characteristics. Al- gen insensitivity, such as gynecomastia, though the glutamine expansion does not reduced fertility, and testicular atrophy. affect the binding of its ligand, androgen SBMA is caused by a polymorphic (testosterone), the glutamine tract is in the (CAG)n repeat in the first exon of the AR major transactivation domain, and may af- gene, which is expressed as a glutamine fect transactivation competence. However, tract. Unaffected individuals carry 5 to 34 theeffectofthepolyglutamineexpansion triplet repeats, while affected individuals upon AR transactivation competence re- carry 37 to 70 repeats. SBMA exhibits mains controversial. a paternal expansion bias. The disease Many lines of evidence suggest that AR does not appear to involve a simple loss- must translocate to the nucleus to exert of-function mechanism, as complete loss its toxicity. Nuclear inclusions (NIs) are of AR does not result in motor neuron present in motor neurons of the spinal degeneration. cord and brainstem in SBMA patients. AR is widely expressed in males and InaDrosophilamodelofSBMA,reti- females, and is a member of the steroid re- nal expression of mutant AR only yielded ceptor–thyroid receptor superfamily with a degenerative phenotype in the pres- a highly conserved DNA binding domain, ence of ligand. As in humans, androgen ligand binding domain and transactiva- binding causes the nuclear translocation tion domain (Fig. 1). In its inactive state, of AR in mice. Transgenic male SBMA it forms an apo-receptor complex with mice produce testosterone and will only

Normal = 5–34 SBMA = 37–70

...CAGCAGCAG...

1 919

Fig. 1 Diagram of the androgen receptor. The androgen receptor is a member of the steroid receptor-thyroid hormone receptor superfamily, and consequently displays a stereotypical architecture. The CAG repeat – polyglutamine tract (striped box) resides within the amino-terminal domain, which mediates transcription activation through an ‘‘activation function’’ domain (charcoal gray box). Additional conserved domains include the DNA binding domain (light gray box), nuclear localization signal (black box), and the ligand binding domain (checkered box). Expansion of the CAG repeat to alleles of ≥37 triplets is the cause of spinal and bulbar muscular atrophy (SBMA). Triplet Repeat Diseases 11 develop motor neuron disease if express- studies and transgenic mouse models. ing polyglutamine-expanded full-length The phosphorylation of AR is modulated AR protein, yet when castrated, such by androgen, and this posttranslational SBMA transgenic mice do not develop modification appears to enhance caspase- a phenotype. In the same study, female 3 cleavage. transgenic mice hemizygous for expanded One possible mechanism of expanded AR develop motor neuron degeneration AR toxicity is through transcription in- when exposed to exogenous androgen. Im- terference. Polyglutamine-expanded AR portantly, even rare human females who interacts with the transcription coactivator are homozygous for an AR CAG repeat ex- CREB-binding protein (CBP) in a polyg- pansion mutation do not develop SBMA, lutamine tract length-dependent manner, despite widespread expression of mutant colocalizes with CBP in spinal cord NIs AR throughout the CNS (central nervous from patients, and can interfere with system). Thus, SBMA is not a true X-linked CBP-dependent transcription. CBP is a recessive disorder, but rather is classified transcription cofactor that regulates the as a sex-limited disorder, since expres- expression of vascular endothelial growth sion of the disease phenotype is dependent factor (VEGF), among other genes. VEGF upon male levels of androgen. is important in motor neuron health, A number of studies on SBMA patients as deletion of a portion of its promoter and mouse models have characterized the called the hypoxia response element (HRE) NIs seen in this disease. While NIs are causes motor neuron degeneration even in widely distributed in lower motor neu- normoxic mice. Pathologically expanded rons of the spinal cord and brainstem, AR reduces VEGF transcript expression NIs also occur in a variety of nonneu- in males, with VEGF165 isoform ex- ronal tissues that appear to function nor- pression reduced at both the RNA and mally. NIs colocalize with components protein levels. Adding VEGF165 to a mo- of the proteasome and with molecular tor neuron-like cell line (MN-1) expressing chaperones. How this contributes to the polyglutamine-expanded AR significantly phenotype is unknown, although HSP70 rescues its cell death, again supporting overexpression attenuates toxicity in cell the role of transcription interference in culture and in transgenic mouse. The SBMA and suggesting that VEGF165 may presence of proteasome components and serve as a neurotrophic factor for mo- HSPs in NIs may thus be revealing a tor neurons. protective intervention by the cell, or al- While no effective treatment for SBMA ternatively may be deleterious due to has yet been validated in human pa- depletion of these important cellular pro- tients, the role of testosterone in SBMA teins. Interestingly, only antibodies raised disease progression has received consider- to amino-terminal fragments of AR de- able attention. Interestingly, testosterone tect NIs, indicating that proteolysis may supplementation was initially used as a play a role in the disorder. Caspase-3 has treatment for SBMA. Rather than aggra- beenshowntocleaveARinapolyglu- vating the disease as one might fear, tamine tract length-dependent manner in it was reported to attenuate the pheno- vitro, and this may have pathogenic sig- type slightly, but did not significantly nificance, as truncated AR is more toxic retard disease progression. Such a ben- than full-length protein in cell culture eficial effect of testosterone may be due 12 Triplet Repeat Diseases

to anabolic effects on muscle strength or 3.2 to a downregulation of AR in response Huntington’s Disease to elevated androgen levels. Still another possibility is that AR-mediated transcrip- Huntington’s disease (HD) is an autoso- tion in motor neurons somehow performs mal dominant disorder with a prevalence a trophic function in the face of dam- of 1 per 15 000 persons worldwide. It is age or injury. Very recent work, how- a debilitating disease that often presents ever, strongly indicates that elimination clinically in the fourth or fifth decade of of ligand by surgical or pharmacologi- life with chorea (i.e. spontaneous, invol- untary dancelike movements). Personality cal castration is a very effective treat- change and cognitive impairment may ment in SBMA transgenic mouse models, precede the clinical onset by years, and even showing efficacy when SBMA mice ultimately culminate in dementia after display an advanced phenotype. As lig- onset of the movement disorder. Chorea and binding is associated with nuclear gives way to bradykinesia and rigidity late translocation of mutant AR and aber- in the disease. CNS atrophy occurs most rant effects upon nuclear transcription prominently in the striatum, which is re- appear crucial for SBMA disease progres- duced to a fraction of its original size sion, abrogation of nuclear localization (Fig. 2). However, significant neurodegen- may account for the success of castra- eration and neuron loss in the cortex is also tion. Consistent with this hypothesis are typical, while cerebellum, brainstem, and results of studies with flutamide, a drug spinal cord are relatively spared – except in that appears to block AR-dependent trans- juvenile-onset cases. activation without preventing its nuclear HD is caused by a (CAG)n repeat ex- translocation. While pharmacological cas- pansioninthehuntingtin gene (htt), which tration with leuprolide is highly effective encodes a 350-kDa protein containing 67 therapeutically, flutamide does not ame- exons. Unaffected individuals carry 6 to 35 liorate symptoms or disease progression repeats, while affected individuals carry in an SBMA mouse model. Attempted 39 to 250 repeats. The largest repeats translation of these preclinical trial re- cause juvenile-onset HD and display a pa- sults to human SBMA patients is cur- ternal transmission bias. The htt protein rently underway. is ubiquitously expressed in brain tissue,

Fig. 2 Huntington’s disease (HD) neuropathology. Hemi-coronal sections of postmortem brains from (a) a classic, adult-onset HD patient and (b) a normal control reveal marked degeneration of the striatum (midmedial region) and considerable atrophy of cortical regions. (From Robataille et al. (1997) Brain (a) (b) pathol. 7, 901. Used with permission). Triplet Repeat Diseases 13 with highest levels in striatal interneu- emerged as an important theme at the rons and cortical pyramidal cells. Two end of the last decade. While initially it htt mRNA transcripts have been detected: was thought that the formation of large one 10 kb and the other 13 kb in length, aggregates is the basis of polyglutamine expressed most highly in CNS neurons. neurotoxicity, the weight of evidence now Ultrastructurally, the wild-type full-length suggests little correlation between visible protein is predominantly cytoplasmic and aggregate formation and disease pathol- is located in the pre- and postsynaptic ogy. However, the occurrence of aggre- regions of dendrites and axons. Htt pro- gates along with neuropathology suggests tein associates with microtubules, vesicles, that aggregation is inextricably linked to and organelles. the pathogenicity of polyglutamine disease HD appears to have a predominantly proteins. We will address the role of poly- dominant mechanism due to its inheri- glutamine aggregation in neurotoxicity in tance pattern, evidence that homozygotes detail in a separate section. are no more severely affected than het- Many theories have been proposed to erozygotes, and the observation that het- account for how polyglutamine-expanded erozygous deletion of huntingtin does not htt causes neurotoxicity. A thorough dis- cause HD. The fact that no HD-causing cussion of this literature goes beyond the loss-of-function mutations have been doc- scope of this chapter, so the reader is umented in the huge htt gene further referred to the relevant books and re- supports a gain-of-function mechanism. views in the Bibliography for a more However, postnatal elimination of htt ex- intensive treatment of this topic. Ma- pression in regions of the cortex can cause jor theories of htt neurotoxicity that will striatal degeneration in mice, so the no- be considered herein include transcrip- tion that gain of function fully accounts tion dysregulation, proteasome inhibition, for the HD phenotype is being reexam- mitochondrial dysfunction/excitotoxicity, ined. As htt is a regulator of transcription and proteolytic cleavage. As multiple inde- activation and/or mediator of vesicular pendent toxicity events may be occurring brain-derived neurotrophic factor (BDNF) concomitantly, these disease pathways are transport up and down axons, many in- not mutually exclusive. vestigators now envision the effects of Expanded htt protein may interfere expanded htt as twofold: simultaneously with transcriptional processes. Many tran- causing protein misfolding leading to gain- scription factors, such as CBP, contain of-function toxicity, and inducing partial glutamine tracts that mediate important loss of an ill-defined normal function. protein–protein interactions. CBP inter- In 1997, it was first reported that mutant acts directly with htt, and mutant htt htt forms dense amyloid-type aggregates toxicity is ameliorated in striatal neurons in in the nucleus, perikarya, and neuropil of vitro when CBP lacking the htt interaction neurons from a transgenic mouse model domain is overexpressed. CBP mediates and in human patients. The role of ag- the transcription of a number of neuronal gregation in the polyglutamine diseases survival factors, such as BDNF, and func- and in a wide range of neurodegenerative tions by acetylating histones, one aspect of disorders including Alzheimer’s disease, chromatin remodeling that permits tran- Parkinson’s disease, amyotrophic lateral scription to occur. Studies of HD fruit fly sclerosis, and the prion diseases, thus and mouse models have highlighted the 14 Triplet Repeat Diseases

potential importance of histone acetylation transport chain, such as 3-nitropropionic by demonstrating that histone deacety- acid (3-NPA), can cause a selective degen- lase inhibitors (HDAC Is) can successfully eration of the striatum in rat and primate rescue degenerative phenotypes in these models when injected systemically, since model organisms. Htt can affect transcrip- the striatum has among the highest energy tion mediated by p53, Sp1, and REST demands of all neuronal regions. 3-NPA interaction, thereby potentially altering the reduces levels of ATP, resulting in mito- expression of a large number of genes. chondrial and cellular depolarization with Differential expression of survival factors, activation of voltage-dependent NMDA re- neurotransmitter receptors, and a number ceptors. Excitotoxic damage may act in of other genes may thus contribute to HD concert with increased production of free pathogenesis. radicals to cause selective striatal degener- The ubiquitin-proteasome protein degra- ation in HD. Some of the earliest animal dation pathway has also been implicated models of HD were thus generated by ex- in HD. Nuclear inclusions of htt colocalize posing the striatum or entire brain of rats with ubiquitin, which indicates that the ex- or primates to metabolic or excitotoxic in- panded protein has been identified as mis- sults. In 1976, the first such model of HD folded and thus targeted for proteasomal was created by injection of the glutamate degradation. However, polyQ sequences analog and excitotoxin kainic acid into the longer than 9 glutamines are impossi- striatum of rats. Such lesioned rats exhib- ble for eukaryotic proteasomes to cleave. ited a selective degeneration of neurons The proteasomes of htt-transfected cells in the striatum reminiscent of HD. Re- are consequently less capable of degrading cent studies have suggested that impaired proteins other than htt, as demonstrated Ca++ flux due to aberrant interaction of by the reduced degradation of GFP-tagged htt with the inositol phosphate-3 receptor proteins in culture. If proteasome com- (IP3-R) may underlie excitotoxic pathology. ponents are clogged with mutant protein Studies of mitochondria from HD patients and/or sequestered into inclusions, they reveal abnormal mitochondrial Ca++ han- maybeunabletodegradeothermisfolded dling and decreased mitochondrial depo- or damaged proteins that carry out impor- larization thresholds, in support of this hy- tant functions. Alternatively, accumulation pothesis. One very recent study has found of improperly degraded proteins may inter- that deletion of peroxisome proliferator- fere with other normal cellular processes, activated receptor (PPAR) gamma coacti- such as autophagy or mitochondrial oxida- vator 1 alpha (PGC-1α), a key mediator tive phosphorylation. of mitochondrial biogenesis, yields HD- Perhaps the longest and most thor- like striatal degeneration in mice. Thus, oughly studied htt toxicity pathway in- metabolic insults and excitotoxicity may volves metabolic disturbances and excito- be key steps in HD disease pathogenesis. toxicity. Glucose metabolism and oxygen Another important theory of htt neuro- consumption are reduced in HD brains toxicity posits that proteolytic cleavage of as measured by PET. Severe deficits in htt is a required step in neuronal dys- the activity of complexes II/III and IV of function and the degeneration process. the mitochondrial electron transport chain As noted above, the htt protein is enor- (ETC) are evident in HD brains. Inhibitors mous, with full-length product consisting of complex II of the mitochondrial electron of >3100 amino acids. Analysis of human Triplet Repeat Diseases 15

HD material has indicated an absence of expansions can occur in a single gener- midprotein and C-terminal epitopes in htt ation, typically via paternal transmission. aggregates. Careful biochemical studies of The DRPLA disease protein, atrophin- htt have characterized a variety of putative 1, is widely expressed and appears to caspase and calpain cleavage sites. Various be predominantly cytoplasmic. Atrophin-1 studies suggest that the more truncated contains both a putative nuclear localiza- the polyglutamine-expanded htt protein, tion signal (NLS) and nuclear export signal the more toxic it is in cell culture and in (NES), and nuclear localization of normal animal models, and the more likely it is atrophin-1 is observed. Nuclear localiza- to enter the nucleus and produce toxicity tion of polyglutamine-expanded atrophin- there. In the case of htt, a series of pro- 1 has been linked to increased toxicity teolytic cleavage steps culminating with in cell culture models. Ubiquitinated NIs cleavage to an ∼100 amino acid peptide are present in neurons and glia from pa- fragment by an aspartyl protease has been tient brains, and are also immunoreactive for small ubiquitinlike modifier (SUMO) proposed to yield a final ‘‘toxic fragment.’’ protein. Atrophin-1 is a substrate for c- As it turns out, the ‘‘toxic fragment hypoth- Jun N-terminal kinase (JNK), with JNK’s esis’’ (as it has also been called) may be affinity for atrophin-1 inversely propor- applicable to a number of polyQ diseases, tional to the size of the polyglutamine which will be reviewed later in this section. tract expansion. Although the relevance of JNK phos- 3.3 Dentatorubral Pallidoluysian Atrophy phorylation to atrophin-1 function is un- known, other insights into the function Dentatorubral pallidoluysian atrophy (DR- of atrophin-1 have been reported. Using Drosophila melanogaster as a model sys- PLA) is a rare, autosomal dominant tem, Zhang et al. took advantage of the neurodegenerative disorder most preva- existence of a fly ortholog of atrophin-1 lent in Japan. Adult-onset DRPLA typi- (Atro) and created lines of flies carry- cally involves progressive cerebellar ataxia, ing mutations in the Atro gene. These choreoatheosis, epilepsy, and dementia, flies demonstrated severe developmental while juvenile-onset cases also display abnormalities due to complex pattern- myoclonus, epilepsy, and mental retar- ing defects, and subsequent experiments dation. Neuropathological abnormalities revealed that Atro is a transcription core- include degeneration of the dentate nu- pressor whose activity diminishes with cleus of the cerebellum, rubral nucleus, increasing polyglutamine tract length. In- and globus pallidus, as well as a more dependent studies of human atrophin-1 in generalized degeneration and gliosis in- cell culture studies found evidence for tran- volving the brainstem, cerebellum, cortex, scription interference of polyglutamine- and pons. DRPLA is caused by a polymor- expanded atrophin-1 with CREB-mediated phic (CAG)n repeat in the carboxy-terminal gene activation. Transcription dysregula- coding region of the atrophin-1 gene on tion may thus play a prominent role chromosome 12. Normal individuals carry in DRPLA. Another study suggests that 6–35 repeats, while affected individuals disturbed carbohydrate metabolism may carry 49–88 repeats. Anticipation is promi- contribute to the DRPLA neurodegenera- nent in DRPLA, as very large repeat tive phenotype. 16 Triplet Repeat Diseases

3.4 there is strong evidence supporting a gain- Spinocerebellar Ataxia Type 1 of-function mechanism in SCA1. Ataxin-1 is widely expressed in the CNS Spinocerebellar ataxia type 1 (SCA1) is an and throughout the periphery, although autosomal dominant disorder character- expression levels are several-fold higher in ized primarily by a progressive cerebel- nervous system tissues. The protein is pre- lar ataxia. It accounts for about 6% of dominantly nuclear in the CNS, however, all autosomal dominant cerebellar ataxias some cytoplasmic staining is apparent in (ADCAs) worldwide. SCA1 patients suf- Purkinje cells of the cerebellum and in fer from coordination difficulties including brainstem nuclei. Ataxin-1 knockout mice dysarthria, dysphagia, and ophthalmople- do not develop SCA1, although they do gia. The cerebellum undergoes atrophy, exhibit impairments in motor and spatial gliosis, and severe loss of Purkinje cells, ac- learning. These mice also have decreased companied by degeneration of the dentate paired-pulse facilitation in the CA1 re- nucleus, inferior olive, and some brain- gion of the hippocampus, suggesting that stem nuclei. Disease onset typically occurs ataxin-1 may normally function in synaptic in the third or fourth decade of life, but plasticity and learning. presentation in childhood or adolescence Large ataxin-1 containing NI’s occur to late life may be seen. SCA1 is caused by in the brainstem of affected individuals, the expansion of a coding (CAG)n repeat and are immunoreactive for ubiquitin, in the amino-terminal coding region of the the 20 S proteasome subunit, and HSPs ataxin-1 gene. The ataxin-1 CAG repeat is HDJ2 and Hsc70. Work done on SCA1 in highly polymorphic, ranging in size from transgenic mice by the Orr and Zoghbi lab- 6to44tripletsinunaffectedindividuals. oratories has been crucial in formulating The repeat length associated with disease models of not only SCA1 disease patho- ranges from 39 to more than 100 CAGs. genesis but also for influencing views Unaffected individuals with more than of the molecular basis of all polyglu- 20 repeats have CAT triplet interruptions tamine diseases. Indeed, the first mouse within their (CAG)n repeat tracts. Such in- model for a polyglutamine disease was terruptions stabilize the repeat expansion, generated by transgenic overexpression of while absence of the CAT repeat intersper- polyglutamine-expanded ataxin-1 in Purk- sion is noted in SCA1 patient alleles. As inje cells (Fig. 3). This SCA1 transgenic in many other polyglutamine disorders, mouse model has laid the foundation for

Fig. 3 The original spinocerebellar ataxia type 1 (SCA1) mouse model. (a) Diagram of the Pcp2-SCA1 transgene construct. A Purkinje cell-specific expression cassette based upon inclusion of the promoter (straight line), first two noncoding exons (black boxes), and first intron (bent line) of the Purkinje cell protein 2 (Pcp2) gene and a SV40 polyadenylation sequence (open box) was the basis for this landmark work. Ataxin-1 cDNAs (gray box) containing either 30 CAGs (control) or 82 CAGs (expanded) were inserted into the Pcp2 expression cassette. Sites of various PCR primer sets are also shown. (b) Pcp2-SCA1 CAG-82 mice display ataxia. Still photographs of a 30-week-old Pcp2-SCA1 CAG-82 mouse from a transgenic line that greatly overexpresses the mutant ataxin-1 transgene illustrates the inability of this mouse to maintain its balance when ambulating. Loss of balance when walking is consistent with the gait ataxia seen in human SCA1 patients. (From Burright et al. (1995) Cell 82, 937, used with permission). Triplet Repeat Diseases 17 numerous follow-up studies of polyglu- required for SCA1 in mice. In a later study, tamine disease pathogenesis. For example, crossing of SCA1 transgenic mice with expression of mutant ataxin-1 lacking an mice lacking a ubiquitin ligase enzyme intact self-association domain precluded yielded SCA1 mice incapable of aggregate aggregate formation, but permitted neu- formation. These mice were more severely rotoxicity, demonstrating that NIs are not affected than their transgenic counterparts

(a)

(b) 18 Triplet Repeat Diseases

duetoabsenceoftheubiquitinligase. of mutant ataxin-1 on neuronal function, This work supported the view that visible however, remain undefined. aggregate formation may represent a pro- Transcriptional dysregulation is also a tective cellular response for neutralizing reasonable hypothesis for SCA1 patho- misfolded polyglutamine-containing pep- genesis. One study has demonstrated that tides. In another study, mice expressing ataxin-1 interacts with polyglutamine bind- ataxin-1 with a mutated NLS were found ing protein 1 (PQBP-1), and that this not to develop SCA1, suggesting that nu- interaction results in interference with clear localization is absolutely required for RNA polymerase-dependent transcription. SCA1 molecular pathology. Although tar- Independent studies have supported a role geting ataxin-1 to Purkinje cells appears for ataxin-1 as a transcription corepressor. sufficient to recapitulate a dramatic SCA1- In pull-down assays and in Drosophila, like disease in mice, a subsequent knockin ataxin-1 interacts with the proteins SMRT mouse model of SCA1 indicated that ex- (silencing mediator of retinoid and thyroid pression in other regions of the CNS yields hormone receptors) and HDAC3 (his- a more representative disease phenotype. tone deacetylase 3), both transcriptional Over the past few years, numerous repressors. Aggregates of polyglutamine- leads have emerged in the search for the expanded ataxin-1 sequester SMRTER, the pathogenic basis of SCA1. In one line Drosophila ortholog of SMRT. Transcrip- of investigation, phosphorylation of ser- tion repressors and histone deacetylases ine 776 of the ataxin-1 protein was shown also appeared in an earlier screen for mod- to affect pathogenesis, highlighting the ulators of the Ataxin-1 phenotype in the fly. importance of this posttranslational mod- In addition, two separate studies of gene ification. SCA1 transgenic mice in which expression alterations in presymptomatic serine 776 had been mutated to an ala- SCA1 transgenic mice have uncovered nine, exhibited a dramatically attenuated changes in the levels of transcripts en- phenotype and a complete lack of NI’s. coding proteins involved in Ca++ flux In an accompanying study, interaction of and metabolism. It thus appears that tran- polyglutamine-expanded ataxin-1, but not scription dysregulation may be a key step wild-type ataxin-1, with several isoforms of in SCA1 disease pathogenesis. the phosphoserine/threonine binding pro- tein 14-3-3 was reported. This extremely 3.5 abundant peptide is thought to serve a Spinocerebellar Ataxia Type 2 regulatory function by binding proteins and determining their subcellular localiza- Spinocerebellar ataxia type 2 (SCA2) is an tion, among other things. The interaction autosomal dominant, progressive cerebel- between ataxin-1 and 14-3-3 was shown lar ataxia that accounts for about 13% of to be dependent upon the Akt-mediated all ADCA cases. Although patients dis- phosphorylation of serine 776. A novel play problems with voluntary coordinated mechanism for ataxin-1 toxicity was thus movements as in the rest of the SCAs, proposed: upon Akt phosphorylation of its main distinguishing clinical feature is polyglutamine-expanded ataxin-1, ataxin-1 extremely slow saccadic eye movements. binds 14-3-3, is stabilized, and ultimately Other symptoms may include hypore- accumulates in the nucleus. The down- flexia, myoclonus, and action tremor. Pa- stream effects of the nuclear accumulation tients suffer a gradual degeneration of Triplet Repeat Diseases 19 the cerebellum, inferior olive, pons, and Thus, a complex including ataxin-2 and spinal cord. Both Purkinje and granule A2BP1 may be involved in RNA processing cells degenerate in the cerebellum, and or metabolism. CNS atrophy in some patients can be The ataxin-2 gene is evolutionarily con- widespread. Interestingly, in certain cases, served. The murine ortholog of ataxin-2 there is involvement of the substantia does not contain a polyglutamine repeat nigra, with such patients displaying a tract, however, but instead possesses a sin- prominent degree of parkinsonism. The gle glutamine residue at the analogous lo- disorder is caused by a coding (CAG)n cation. (Absence of a substantial glutamine expansion in a novel gene of unknown repeat tract is typical for mouse orthologs function, named ataxin-2. The SCA2 CAG of polyglutamine disease proteins.) In the repeat displays little polymorphism in the case of ataxin-2, study of the Drosophila normal population, with 95% of the popu- ortholog (Datx2), which does show two lation possessing 22 or 23 repeats in each regions of marked amino acid similar- allele. The remaining 5% of alleles in un- ity and does contain polyglutamine repeat affected individuals do nonetheless range regions, has yielded some potentially im- from 15–31 CAG repeats. Such unaffected portant insights into ataxin-2s normal individuals typically have two CAA inter- function. Modulation of Datx2 dosage re- ruptions in their CAG repeat tract. Affected sulted in mutant phenotypes whose cause SCA2 patients typically display CAG re- could be traced to aberrant actin filament peats numbering from 32–63 triplets, and formation (Fig. 4). This work suggests such disease alleles are always uninter- that alteration of ataxin-2-mediated regu- rupted CAG tracts. There is a nonlinear lation of cytoskeletal structure could affect inverse correlation between expansion size dendrite formation or other aspects of neu- and age of onset in SCA2, with some stud- ronal function in SCA2. As SCA2 is one ies documenting a paternal transmission ofthefewpolyglutaminedisorderstodis- bias for larger expansions. play prominent cytosolic aggregates, such Both wild-type and expanded ataxin-2 a model of SCA2 pathogenesis seems plau- mRNA is widely expressed, with high- sible. In other experiments, eliminating est levels in the substantia nigra and the C. elegans ortholog of ataxin-2 (ATX- Purkinje cells of the cerebellum. The 2) yielded a lethal phenotype, indicating 140-kDa protein is cytoplasmic and its that ATX-2 is required for early embryonic function remains uncertain. Since it con- development of this nematode worm. It tains Sm1 and Sm2 motifs common in remains to be seen how these observations proteins involved in RNA splicing and in worms and flies will apply to ataxin- protein–protein interactions, involvement 2 function in mammals, especially since in RNA processing has been proposed. no simple or conditional knockout of the A yeast two-hybrid screen indicated that mouse ataxin-2 gene has been performed. ataxin-2 has a binding partner, ataxin- 2 binding protein-1 (A2BP1). A2BP1 is 3.6 highly conserved throughout the animal Spinocerebellar Ataxia Type kingdom and its expression pattern corre- 3/Machado–Joseph Disease sponds well with that of ataxin-2. A2BP1 contains a domain that is also conserved Spinocerebellar ataxia type 3 (SCA3) is in RNA-binding proteins, the RNP motif. the most common inherited dominant 20 Triplet Repeat Diseases

Fig. 4 Studies of the ataxin-2 ortholog in Drosophila melanogaster reveals a role for Drosophila ataxin-2 (Datx2) in actin filament formation during oogenesis. Egg chambers from normal (Wild-Type (WT)) and mutant flies with reduced expression (a) (b) (Datx2) of Drosophila ataxin-2 were stained with DAPI (blue) and phalloidin (red) to indicate nuclei and filamentous actin respectively (a–d). The egg chambers of WT flies prior to cytoplasmic transport (a) display well demarcated, separated but (c) (d) interconnected cells (blue) as expected, while the egg chambers of Datx2 flies (b) contain irregularly arranged cells. After cytoplasmic transport, the egg chambers of WT flies (c) show one greatly enlarged oocyte (dashed line) with only a small section of compressed cells, while the egg chambers of Datx2 flies (d) have failed to yield an enlarged oocyte, instead retaining dispersed and large adjacent cells. Confocal images of egg chambers prior to (e) (f) cytoplasmic transport stage reveal a prominent actin filament network in WT flies (e), but a remarkably transparent actin filament network in Datx2 flies (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 Drosophila homolog of the polyglutamine disease gene SCA2 is a dosage-sensitive regulator of actin filament formation, Genetics 162, 1687–1702, used with permission of Genetics.) (See color plate p. xxiii).

spinocerebellar ataxia worldwide. SCA3 is repeat allele. Some researchers have pro- also known as Machado–Joseph disease posed that this could be due to a SCA3 (MJD) because of its initial description in founder effect. The presence of SCA3 in a group of Portuguese residents of the every major racial population worldwide, Azores islands. It is a progressive, auto- however, would require the founder to be somal dominant cerebellar ataxia whose truly ancient. There is the typical inverse clinical features include ophthalmoplegia, correlation between repeat number and dystonia, dysarthria, and signs of lower age of onset in SCA3, and in this disease, motor neuron disease, such as tongue paternal expansion bias is characteristic, and facial fasciculations. Degeneration oc- as documented by repeat sizing of sperm. curs in the spinocerebellar tracts, dentate Ataxin-3 is a ubiquitously expressed 42- nuclei, red nuclei, substantia nigra, and kDa protein, making it the smallest of the spinal cord. This disease is unique among polyglutamine proteins. Ataxin-3 is highly the SCAs because the cerebellar cortex conserved in eukaryotes, with homology and inferior olive are largely spared. SCA3 to ENTH and VHS domain proteins is caused by a (CAG)n repeat near the involved in regulatory adaptor functions 3 end of the coding region of a novel and membrane trafficking. Ataxin-3 has gene (ataxin-3). Normal alleles range from several splice variants which reside in 12–40 CAG repeats while affected indi- the nucleus and the cytoplasm, and it viduals carry 55–84 repeats. Unlike many appears developmentally regulated. The other triplet repeat disorders, there is a sub- existence of NIs was first documented in stantial gap between the largest normal re- the brains of patients with SCA3, and peat allele and the smallest disease-causing this feature of polyglutamine-expanded Triplet Repeat Diseases 21 ataxin-3 remains evident in SCA3 cell 3.7 culture and mouse models. Ataxin-3 NIs Spinocerebellar Ataxia Type 6 are ubiquitinated and contain numerous transcription factors. Spinocerebellar ataxia type 6 (SCA6) is A number of recent discoveries regard- an autosomal dominant, slowly progress- ing the domain structure and normal ing cerebellar ataxia that accounts for ∼ function of ataxin-3 suggest pathways 10–20% of ADCA worldwide. It is by which polyglutamine repeat expan- characterized predominantly by cerebel- sion could result in disease pathogenesis. lar dysfunction that may have an episodic Comparison of ataxin-3 amino acid se- component. Other common features may quences across a wide range of eukaryotic include dysarthria, nystagmus, loss of vi- species revealed the presence of an ex- bration sense and proprioception, and tremely highly conserved amino-terminal imbalance. Histopathological changes in- clude loss of Purkinje cells, cerebellar sequence that was named the josephin do- granule neurons, and neurons in the den- main. Study of this region indicates that it tate nucleus and inferior olive. SCA6 is may play a role in aggregate formation caused by a coding (CAG)n expansion in in concert with the expanded polyglu- exon 47 of the gene CACNA1A,which tamine tract. A rather intriguing feature encodes the α subunit of the P/Q-type of ataxin-3 that was discovered indepen- 1A voltage-gated calcium channel. This gene dently is the presence of multiple ubiquitin is located on chromosome 19 at band interaction motifs (UIMs), and the demon- p13. The SCA6 CAG repeat is small com- stration that ataxin-3 is a polyubiquitin pared to other polyglutamine disorders, binding protein. This suggests a role for with a pathogenic range of only 19–33 ataxin-3 in mediating protein refolding and triplets. Unaffected individuals carry alle- degradation. les of 4–18 repeats. There is an inverse In addition to a possible normal role correlation between repeat length and age in protein surveillance, other studies of disease onset, and minimal intergen- have found ataxin-3 directly interacts with erational and somatic instability has been the histone acetyltransferases CBP and reported for the SCA6 repeat expansion in p300, and can block histone acetyltrans- affected patients. ferase activity by inhibiting access of The 9.8 kb CACNA1A transcript impli- such coactivators to their histone sub- cated in SCA6 is expressed throughout strates. This appears to be mediated the CNS, and most highly in cerebellar by the interaction of ataxin-3 with hi- Purkinje cells and granule neurons. The stones. In vitro and in vivo studies of α1A subunit is the pore-forming compo- ataxin-3 further revealed an interaction nent of the channel, which is important with histones and the chromatin remod- in neurotransmitter release at the synapse. eling machinery that led to a repres- Polyglutamine expansions in this subunit sion of transcription activation. Thus, cause variable changes in calcium trans- polyglutamine-expanded ataxin-3 may also mission rates, depending on the system affect transcriptional processes once it and the β subunit coexpressed. Expansions begins to accumulate in the nuclear com- do not cause a reduction in membrane partments of the cell types where it is channel density in HEK293 cells, sug- expressed. gesting that aggregation does not occur 22 Triplet Repeat Diseases

at pathogenic repeat lengths. Ubiquitin- calcium channel due to association of the negative neuronal inclusions are visible in mutant α1A subunit with other subunits the cytoplasm of SCA6 patient Purkinje causes SCA6. The coincidence that this dis- cells, however. order is caused by a polyQ tract expansion In addition to displaying a disease al- and causes a progressive cerebellar ataxia lele range that does not overlap with the cannot be ignored, however, more data will other polyQ diseases, there are several be required to soundly refute a possible other reasons to suspect that SCA6 is dif- concomitant toxic gain-of-function effect. ferent from the rest of the polyglutamine repeat group. The first difference is the 3.8 existence of two disorders, episodic ataxia Spinocerebellar Ataxia Type 7 type 2 (EA2) and familial hemiplegic mi- graine (FMH), that are both allelic to SCA6, Spinocerebellar ataxia type 7 (SCA7) is an as both are caused by point mutations in autosomal dominant, progressive cerebel- the CACNA1A gene. EA2 resembles SCA6 lar ataxia. It is unique among autosomal as affected EA2 individuals suffer from a dominant SCAs, as patients typically de- slowly progressive form of episodic ataxia, velop visual impairment in addition to accompanied by nystagmus, dysarthria, their cerebellar ataxia. The visual impair- loss of balance and sometimes cerebellar ment is due to a cone-rod dystrophy that atrophy. EA2 is usually caused by trun- results in retinal degeneration. Patients cation mutations in CACNA1A,resulting first develop problems distinguishing col- in loss-of-function of the calcium chan- ors, but ultimately go blind in this type nel. CACNA1A knockout mice similarly of retinal degeneration. SCA7 patients develop ataxia and late-onset cerebellar may present either with cerebellar ataxia atrophy, characteristic of SCA6 and EA2 or visual impairment. The likelihood of patient phenotypes. Another reason SCA6 their presentation is dictated by the size is different from other polyQ disorders is of their CAG repeat disease allele, with that the CACNA1A protein probably does not undergo a structural change that con- larger repeats typically favoring presen- verts it into an aggregate-prone, beta-sheet tation with visual impairment. Affected adopting, amyloid-like conformer. Indeed, individuals display prominent dysarthria, even the largest SCA6 polyglutamine tract and can develop increased reflexes, de- is below the threshold required for stable creased vibration sense, and oculomotor β-pleated sheet formation in other polyg- disturbances. Neuronal degeneration and lutamine disorders. Consistent with this reactive gliosis occur in the cerebellar cor- prediction, channel localization of mutant tex, dentate nucleus, inferior olive, pontine polyglutamine-expanded CACNA1A pro- nuclei, and occasionally in the basal gan- tein is not affected in cell culture models, glia. NIs are widespread. Infantile-onset and its channel function is not com- SCA7 has been documented, and in this pletely abolished. Cytoplasmic aggregates fatal form of the disease, nonneuronal tis- in patient material are ubiquitin-negative, sues such as the heart and the kidney suggesting that the protein may not be are severely affected. SCA7 is caused by grossly misfolded. This evidence supports a highly polymorphic (CAG)n repeat in a model whereby a dominant-negative loss the 5 coding region of the ataxin-7 gene. of function of the P/Q-type voltage-gated Unaffected individuals carry 4–35 repeats, Triplet Repeat Diseases 23 while affected individuals carry 37–306 re- presence of the polyglutamine-expanded peats. The SCA7 trinucleotide repeat is one ataxin-7 has a dominant-negative effect of the most unstable of all polyglutamine upon the GCN5 histone acetyltrans- disease genes, sometimes expanding by as ferase activity of the STAGA complex, much as 250 repeats in a single generation. resulting in transcription dysregulation. There is a pronounced paternal expansion The transcription dysregulation caused bias, with large expansions occurring in by polyglutamine-expanded ataxin-7 likely male germ cells, frequently causing em- causes a disease phenotype by altering the bryonic lethality that results in reduced ability of certain transcription factors to transmission. Marked repeat instability activate expression of their target genes. can occur in the brain, and produce large The best characterized example of tran- somatic expansions on occasion. scription dysregulation by polyglutamine- Ataxin-7 is a ubiquitously expressed expanded ataxin-7 is its interference with protein of 892 amino acids. It is expressed the cone-rod homeobox protein (CRX), a most highly in heart, skeletal muscle, glutamine domain containing transcrip- and pancreas. Expression levels within the tion factor expressed only in the retina CNS are highest in the cerebellum and and the pineal gland. Ataxin-7 interacts brainstem. One splice variant, ataxin-7b, is directly and functionally with CRX, ac- expressed predominantly in the CNS (32). cording to studies performed in vitro Ataxin-7 contains a functional arrestin and in a mouse model of SCA7. Impor- domain, a protein interaction domain tantly, the interaction between ataxin-7 that is highly selective for phosphorylated and CRX appears to involve the glu- forms of its interacting protein(s). This tamine tract regions found in both pro- suggests that it interacts with specific teins. Autosomal dominant mutations in phospho-proteins, although none have CRX can cause a cone-rod dystrophy been discovered to date. It also contains in humans, further supporting a model two SH3 domains, which are protein interaction domains that bind proline-rich in which CRX’s diminished transactiva- sequences and mediate a number of cell tion competence is central to the SCA7 signaling processes. Ataxin-7 also contains retinal degeneration phenotype. Several threeputativeNLSsandoneNES. transcription factors, including CBP, can Ataxin-7 is conserved throughout eu- be found in SCA7. CRX may be but karyotes, and its yeast ortholog SGF73 is one of a number of transcription factors part of a multisubunit histone acetyltrans- whose function is diminished by polyQ- ferase complex called SAGA (Spt/Ada/ expanded ataxin-7 interaction and dys- Gcn5 acetyltransferase). The human or- regulation of STAGA complex-mediated thologs of SAGA comprise the so-called gene expression. STAGA (SPT3/TAF9/ADA2/GCN5 acetyl- One intriguing feature of SCA7 was transferase) complex, and are essen- discovered upon generation of transgenic tial transcription coactivators required mice expressing ataxin-7 with the mouse for the transcription of certain genes. prion protein promoter. This promoter STAGA components immunoprecipitate drives expression in every tissue, with the with ataxin-7. Although pathogenic expan- occasional notable exception of the Purk- sion of ataxin-7 does not alter its ability to inje cells of the cerebellum. Despite lack of be integrated into the STAGA complex, the Purkinje cell expression of ataxin-7, mice 24 Triplet Repeat Diseases

Fig. 5 Noncell autonomous Purkinje cell degeneration in a mouse model of spinocerebellar ataxia type 7 (SCA7). Confocal microscopy analysis of cerebellar sections from a SCA7 transgenic mouse (SCA7) created with an ataxin-7 CAG-92 containing murine prion protein expression vector and from an age- and sex-matched nontransgenic littermate (Control). Staining with an anti-ataxin-7 antibody (magenta), a calbindin antibody (green), and DAPI (blue) reveals a healthy, normal- appearing cerebellum characterized by properly oriented Purkinje cells with extensive dendritic arborization in the ‘‘Control’’ mice. However, SCA7 transgenic mice display pronounced Purkinje cell degeneration as evidenced by decreased dendritic arborization and displacement of Purkinje cell bodies. Interestingly, although numerous neurons in the granule cell layer (GCL) and the molecular layer (ML) display aggregates of ataxin-7, there is no accumulation of mutant ataxin-7 in the degenerating Purkinje cells due to lack of appreciable expression there. As the Purkinje cells degenerate without expressing the mutant protein, the degeneration 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 and displays proteolytic cleavage in ataxic transgenic mice, J. Neurosci. 22, 4897–4905, used with permission of the Journal of Neuroscience.) (See color plate p. xxiv).

developed a cerebellar ataxia accompa- by a membrane-bound autophagosome nied by degeneration of the Purkinje cells and is accelerated by a number of cellular (Fig. 5). This noncell-autonomous degen- insults.) Caspase-3, a proteolytic enzyme eration may point to a disease mechanism classically associated with apoptosis, is ac- involving withdrawal of trophic support tivated at abnormally high levels in SCA7 by communicating neurons (olivary, deep patient brains. This may suggest that cerebellar, brainstem, or granule neurons) caspase activation, rather than causing or dysfunction of glutamate transporters apoptosis, is contributing to a nonapop- expressed by surrounding glia. Damage to totic degenerative process. inferior olivary neurons or Bergmann glia In normal human neurons, ataxin-7 is can indeed cause the degeneration of Purk- variably located in the nucleus or the inje cells, lending credence to this theory. cytoplasm. In SCA7 patients, ataxin-7 grad- Another interesting feature of SCA7 that ually undergoes a shift in localization into is common to other neurodegenerative NIs. This process has been replicated in disorders is the prominence of morpholog- SCA7 transgenic mice, in which ataxin-7 ical and functional degeneration without immunoreactivity shifts from the cyto- pronounced apoptosis. Some neurons in plasm to the nucleus, and ultimately forms SCA7 humans and mouse models ex- foci there. In patient’s brains, these NIs hibit indentations in the nuclear enve- colocalize with promyelocytic leukemia lope, reduced arborization, ectopy, and (PML) protein, which is an integral part increased autophagy. (Autophagy is the of nuclear bodies (NBs). NBs are associ- bulk degradation of cellular components ated with transcription regulation and the Triplet Repeat Diseases 25 ubiquitin-proteasome pathway; thus, accu- Given TBP’s central role in transcription mulation of ataxin-7 in NBs may represent regulation, the basis of SCA17 disease an attempt by the cell to degrade misfolded pathogenesis is proposed to involve tran- protein. scription dysregulation, but this is yet to be proven. 3.9 Spinocerebellar Ataxia Type 17 3.10 Role of Aggregation in Polyglutamine Spinocerebellar ataxia type 17 (SCA17) Disease Pathogenesis is the most recently discovered polyglu- tamine repeat disease. It is an autosomal In most polyglutamine disorders and in dominant, progressive disorder character- many neurodegenerative diseases in gen- ized by dementia as well as cerebellar ataxia eral, protein aggregation is a prominent and involuntary movement abnormalities. feature. It occurs in almost all polyglu- Its symptoms are diverse and heteroge- tamine diseases, despite the lack of domain neous, but typically begin with behavioral or structural similarity between the dif- disturbances and cognitive impairment. ferent disease proteins. Aggregates have This is followed by ataxia, rigidity, dysto- long been considered reliable indicators nia, hyperreflexia, and rarely parkinson- of disease, although their pattern and on- ism. Neuropathologically, patients may setoftendonotcorrespondwellwiththe suffer degeneration of the cortex, cere- cell-type specificity of disease pathology. bellum (including Purkinje cells), inferior As the role of aggregation in pathogenesis olive, caudate nucleus, and medial tha- has been one of the most hotly debated lamic nuclei. SCA17 is caused by the issues in the field of neurodegeneration, expansion of a coding (CAG)n repeat in it may have implications for Alzheimer’s the TATA-binding protein (TBP) gene on disease, Parkinson’s disease, and amy- chromosome 6q27. Unaffected individuals otrophic lateral sclerosis. Why? There are carry 25–48 repeats, while affected individ- several characteristics of polyglutamine- uals carry 43–66 repeats. The area of over- mediated aggregation that are presumably lap between affected and unaffected alleles common to all of the disorders that show indicates incomplete penetrance at inter- aggregation. First of all, aggregates are mediate repeat lengths. The pathogenic rich in β-pleated sheets and have many threshold is higher than for most other of the properties of amyloid. This is sup- polyglutamine disorders, yet there is an ported by Congo red birefringence and inverse correlation between repeat length immunoreactivity with antiamyloid anti- and age of onset. bodies. Various studies have shown that TBP is a ubiquitously expressed tran- the conformational change is associated scription initiation factor that is a core with the production of visible aggregates component of the RNA polymerase II rather than in the soluble nonaggre- transcription factor D (TFIID) complex. gated phase, where polyglutamine tracts TBP possesses DNA binding activity in the are thought to remain in a random coil TFIID complex, and is therefore required conformation. for the transcription of numerous genes. In vitro, the kinetics of aggregation are SCA17 patients have NIs immunoreactive consistent with nucleated-growth polymer- for TBP, polyglutamine, and ubiquitin. ization, in which the rate-limiting step is 26 Triplet Repeat Diseases

the formation of misfolded peptide as- degradation away from the soluble phase semblies, often referred to as oligomers. and into aggregates has been proposed as A mechanism involving nucleated-growth a potential cause of cell toxicity in neurons polymerization would predict that only one with aggregates. Caspase activation is (or a few) visible aggregates would appear another way in which aggregation could be in each affected cell, and this appears to linked to pathology. Caspase recruitment bethecaseformostpolyglutaminedis- into aggregates can lead to their activation, orders. In the cellular milieu however, which could result in dysfunction or cell where membranes and intermolecular in- death in neurons. Further supporting teractions compartmentalize proteins, this the role of aggregates in polyglutamine prediction may not be valid. The dis- protein toxicity, injection of preformed covery of polyglutamine microaggregates polyglutamine aggregates into the nuclei confirmsthissuspicion.Thethreshold of cultured cells causes cell death, for aggregation of polyglutamine tracts is while the injection of nonpolyglutamine similar to the disease threshold for most aggregates does not. At the same polyglutamine diseases. In vitro,longer time, there is strong evidence that polyglutamine tracts have a lower con- soluble polyglutamine protein, rather centration threshold for aggregation and than aggregates, is the primary source nucleate more quickly than shorter tracts. of toxicity. Importantly, the pattern of This is one possible explanation for the cor- aggregates observed in human patients relation between tract length and disease often does not correlate with the pattern onset and progression. of neuronal dysfunction. For example, Some data support a role for aggregation in the striatum of HD patients, large in polyglutamine disease pathogenesis. interneurons contain aggregates more Some proponents of this theory invoke frequently than medium spiny neurons, the fact that aggregates sequester many yet the former neurons are largely spared, proteins besides the disease protein. Inter- while the latter are most vulnerable. molecular interactions between transcrip- HD transgenic mice expressing full-length tion factors often involve glutamine tracts mutant htt will develop inclusions in or glutamine-rich regions. Some studies many brain regions many of which are have shown that normal proteins with glu- neuropathologically unaffected, while few tamine tracts or glutamine-rich regions inclusions are detected in the striatum, are enriched in polyglutamine aggregates the region of the brain displaying the most in cell culture and animal models. Sev- prominent pathology. eral transcription factors colocalize with Aggregation and toxicity have been di- such aggregates, including CBP, TBP, and rectly dissociated in other polyglutamine numerous TBP-associated factors (TAFs). disease model systems. Studies of the CBP is responsible for the prosurvival SCA1 knockin mouse model revealed that effects of BDNF, and its soluble concen- neurons lacking aggregates were more sus- tration is lowered in HD patient’s brains. ceptible to dysfunction and demise, while Postnatal mice lacking CREB and its ho- those neurons displaying prominent ag- molog CREM develop a progressive degen- gregates were protected. Similarly, SCA7 eration of the hippocampus and striatum. transgenic mice exhibit retinal pathology Titration of enzymes and factors before the occurrence of visible aggregates. required for protein refolding and Finally, in a very provocative study, SCA1 Triplet Repeat Diseases 27 transgenic mice lacking the ubiquitin lig- of mutant protein. To examine the role of ase E6-AP were significantly less capable aggregate formation in polyglutamine tox- of forming large and numerous aggregates icity, another group then tracked survival in neurons in comparison to SCA1 trans- time versus diffuse, soluble htt protein ex- genic mice on a wild-type background, pression levels in htt-transfected striatal but displayed earlier onset of an ataxic neurons undergoing aggregate formation. phenotype and accelerated neurodegener- Comparison of neurons that developed ag- ation. This suggests that the aggregates gregates over time versus those that did may be protective, although the toxicity not confirmed that the level of soluble of compromising the proteasome may nonaggregated mutant polyglutamine pro- have contributed to the accelerated phe- tein was the more reliable predictor of notype. Thus ensued a contentious debate cellular toxicity. Such studies have shifted over the role of aggregate formation in our attention to the role of the intermedi- polyglutamine disease – with some work- ates (oligomers, protofibrils, etc.) as the ers espousing the view that aggregates toxic species instead of the final, visi- were responsible for disease pathology, ble aggregates. others suggesting that the aggregation A reasonable model for polyglutamine process was a protective coping mecha- toxicity predicts that the process starts nism of the cell and thereby beneficial, with a protein that misfolds because of and still others insisting that aggregates the presence of an expanded polyglu- were incidental and irrelevant. This de- tamine tract. The misfolded protein is bate was complicated by the existence of initially detectable in the soluble phase ‘‘microaggregates,’’ small clumps of aggre- due to the cell’s ability to maintain the pro- gated protein visible only at the electron tein in a properly folded state and direct microscope level. it to the degradation machinery. How- Over the last few years, studies decon- ever, the refolding capacity of the cell is structing polyglutamine tract aggregation ultimately exceeded, and since the degra- into a multistep process suggest a par- dation machinery cannot turnover the simonious explanation for the divergent misfolded mutant polyglutamine protein, views. Using a variety of biophysical ap- accumulation occurs. Misfolded polyglu- proaches, including transmission electron tamine proteins can spontaneously change microscopy (TEM), Fourier transform in- their structural properties and adopt ab- frared spectroscopy (FTIR), and atomic normal conformations. These abnormally force microscopy (AFM), one study dis- folded proteins can then nucleate, forming sected the process of huntingtin (htt) oligomers. Oligomers then form protofib- exon 1 peptide aggregation and found rils that grow into fibrils. The transition evidence for a number of sequential mor- to the fibril stage is characterized by the phological and structural intermediates attainment of a β-sheet structure, so at (Fig. 6). By showing that misfolded htt this point the structures are amyloid-like. exon 1–44Q adopted intermediate struc- Fibrils then grow into fibers (also known tures, such work opened up the possibility as microaggregates), which then assemble that intermediates (not visible at the light into aggregates visible under a light micro- microscope level) are the toxic species and scope. According to such a model, blocking that the ultimate visible aggregated forms an intermediate step could be therapeu- of htt exon 1–44Q are neutralized versions tically effective. Consistent with this, one 28 Triplet Repeat Diseases

Fig. 6 Polyglutamine-expanded huntingtin protein undergoes a variety of structural alterations on its way to becoming a visible aggregate. In this experiment, transmission electron microscopy tracked the structure of huntingtin exon 1 peptide with 44 glutamines after release from a linked (a) Control (b) 2 h affinity tag. (a) Prior to affinity tag release, nothing is observed. (b-c) After affinity tag release, globular structures become apparent. (d) This is followed by formation of short fibers (4–5 nm in diameter). (e-f) Fiber formation becomes more prominent. However, during this time, large globular (c) 3 h (d) 4 h assemblies remain (arrowheads). (g-h) Eventually, only fibers are present, and the fibers begin to adopt a uniform appearance, suggesting consolidation into protofibrils. (From Poirier, M.A., Li, H., Macosko, J., Cai, S., Amzel, M., Ross, C.A. (2002) Huntingtin spheroids and protofibrils as precursors in (e) 5 h (f) 6 h polyglutamine fibrilization, J. Biol. Chem. 277, 41032–41037, used with permission of the Journal of Biological Chemistry.).

(g) 7 h(h) 8 h

group has nicely shown that Congo red can enhances their sequestration into visi- bind polyglutamine peptides once they are ble aggregates is beneficial. Finally, since amyloid-like and prevent their conversion oligomers are difficult to detect, it is of- into fibers, while an independent group ten not possible to know whether cells are has shown that Congo red delivery to an successfully sequestering the toxic precur- HD mouse model is a highly effective treat- sors into aggregates or if high levels of ment intervention. toxic intermediates are building up. So, One satisfying aspect of this model is aggregates are thus also incidental, since that it allows us to simultaneously view their presence does not provide us with aggregates as harmful, protective, and in- insight into the steady state levels of the nocuous. How? First, aggregates clearly toxic precursors. In conclusion, advances must be toxic as their creation is predicated in our understanding of the role of aggre- upon the production of earlier intermedi- gateformationinpolyglutaminedisease ate toxic forms. At the same time, the processes suggest that aggregates may be final visible aggregates are less toxic than simultaneously viewed as harmful, benefi- the earlier intermediates, so anything that cial, and incidental. Triplet Repeat Diseases 29

3.11 ataxin-1 with a mutated NLS do not develop Protein Context the SCA1 phenotype, while those without mutated NLSs do. Interaction domains Theimportanceofproteincontextinthe also affect polyglutamine protein toxic- pathogenesis of polyglutamine disorders ity. The ability of expanded polyglutamine is a subject of great interest to many proteins to interact with other molecules researchers, because it is directly related is a promising avenue in the search for to the mechanisms of toxicity in each mechanisms of cell-type specificity. Yeast disease. It is evident in the polyglutamine two-hybrid screens and other techniques disorders that protein context plays a have yielded interaction partners for a role in cell-type specificity. For example, number of polyglutamine proteins. For a polyglutamine expansion in the Htt example, Htt’s interactions with transcrip- protein causes a severe degeneration tion factors may prove to be central to its of the striatum and cortex, while the pathological effects. At the same time, it is same size glutamine tract in ataxin- 1 causes a degeneration of structures also evident that polyglutamine tracts are in the cerebellum and brainstem, while innately toxic. Pure polyglutamine tracts sparing the striatum and cortex. This cause toxicity in cell culture. Mice that cannot be attributed to gross differences express a glutamine tract of 150 amino in expression patterns, since both proteins acids in the Hprt protein, which does not are pan-neural (Fig. 7). naturally contain any such tracts, exhibit Oneeffectthatproteincontextmayhave progressive neurological deficits and NIs. on pathology is to affect subcellular local- Thus,itappearsthatpathologyineach ization. For example, the presence of a polyglutamine disease is due to a gain of functional NLS or NES dictates preferen- function of the glutamine tract that is then tial subcellular localization in the nucleus modulated by the protein context in which or cytoplasm. Mice expressing expanded it resides.

Central nervous system:

Cerebral hemisphere Brain Brain stem

Spinal cord

Fig. 7 The conundrum of cell-type specificity in the polyglutamine diseases. Although the different polyglutamine disease proteins are Huntington's disease expressed throughout the central Spinal muscular atrophy nervous system, only select populations of neurons degenerate in the different Spinocerebellar ataxias disorders. The principal regions of selective vulnerability are shown for certain of the polyglutamine diseases. 30 Triplet Repeat Diseases

3.12 tendency for the polyglutamine dis- Transcriptional Dysregulation ease proteins to alter transcription factor/coactivator-mediated histone acetyl- Transcriptional dysregulation is one the- transferase (HAT) activity. The ability of ory of polyglutamine toxicity that has agenetobetranscribeddependsupon gained support from many lines of evi- its chromatin structure, and thus upon dence in several disorders, and it appears the degree of histone acetylation in its increasingly likely that it is one of the vicinity. This is because acetylated hi- fundamental causes of pathogenesis. Mi- stones cause chromatin to be in an croarray technology and other genomic ‘‘open,’’ transcription-friendly conforma- techniques are facilitating rapid advances tion. Acetylation status depends upon the in this area of the polyglutamine field. The balance between the activity of HATs majority of polyglutamine disorders are and their countervailing counterparts, caused by proteins that are either transcrip- the HDACs. CBP, p300, and p300/CBP- tion factors/cofactors or interact closely associated factor (PCAF) are all HATs that with transcription factors. TBP and AR are are inhibited by polyglutamine-expanded transcription factors, atrophin-1 is a tran- htt exon 1 peptide and ataxin-3. In cell cul- scriptional corepressor, ataxin-7 is part of ture, expression of a mutant htt fragment a transcriptional coactivator complex, and reduces global histone acetylation. ataxin-1, htt, and ataxin-3 all interact with Finally, according to a very recent study, various transcription factors. It is therefore ataxin-7 directly interacts with GCN5 as reasonable to suspect that an abnormally part of the STAGA complex, and upon long polyglutamine tract could interfere polyglutamine expansion, mutant ataxin- with the transcriptional activity of these 7 causes a dominant-negative effect upon proteins and/or their interactors. GCN5 HAT activity. This results in CRX CBP is one of the most studied transcrip- transcription interference that may con- tion factors in the polyglutamine field. tribute to the SCA7 retinal degeneration It is a transcription activator that me- phenotype. Inhibiting HDACs with drugs diates part of the cellular response to known as HDAC inhibitors (HDAC Is) at- cAMP, and it can interact with numer- tenuates the phenotype of HD and SBMA ous polyglutamine proteins in the soluble in mouse and fly models, presumably by or insoluble phase, including huntingtin, shifting the cells’ acetylation status. HDAC ataxin-3, ataxin-7, and AR. Expanded htt Is such as sodium butyrate and especially represses CBP-regulated genes, and over- suberoylanilide hydroxamic acid (SAHA) expression of CBP causes a considerable have thus emerged as possible candidates rescue of the cell death phenotype in HD for therapeutic trials in human polyglu- and SBMA cell culture models. Postna- tamine disease patients. tal mice lacking CBP’s upstream activator CREB and its homolog CREM display a 3.13 profound degeneration of the striatum Proteolytic Cleavage and hippocampus. Thus, interference with CBP appears to be a common theme in The occurrence of proteolytic cleavage in polyglutamine pathology. polyglutamine diseases first became ap- Another common theme in the tran- parent when it was shown in HD that scription dysregulation equation is the htt can be cleaved by extracts derived Triplet Repeat Diseases 31

Fig. 8 Proteolytic cleavage in the polyglutamine (kD) nt 24Q 92Q diseases. Production of amino-terminal truncated proteins is observed in many of the polyglutamine 148 Ataxin-7 92Q diseases, as is shown here for SCA7. In this Ataxin-7 24Q experiment, nuclear fractions of retinal lysates 98 from age- and sex-matched nontransgenic (nt), ataxin-7 CAG-24 (24Q), and ataxin-7 CAG-92 (92Q) 64 mice were prepared and immunoblotted with an 55-kD fragment antibody directed against the amino-terminal 50 region of the protein ataxin-7. As shown here, in addition to soluble full-length ataxin-7 (which is 36 typically reduced in cells where it is aggregating 22 into insoluble inclusions), an ∼50–60kD 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 proteolytic cleavage in ataxic transgenic mice, J. Neurosci. 22, 4897–4905. Used with permission of the Journal of Neuroscience.). from apoptotic cells. Subsequent publi- proteolysis may be a normal event in pro- cations showed that only amino-terminal tein turnover and the inability to clear epitopes of htt protein are detectable in ag- cleaved protein may be the problem. gregates. SBMA, DRPLA, and SCA7 were Another possibility is that soluble polyglu- then added to the list of diseases in which tamine proteins may cause the activation aggregates are only immunoreactive for of proteases such as caspases, which then a glutamine-containing fragment of the cleave the disease protein and send the cell disease protein. Since then, evidence of on a path to degeneration and ultimately proteolytic cleavage has been published for apoptosis. As we will see, each theory has nearly all known polyglutamine disorders. support and not all are mutually exclusive. Experimental studies revealed that inhibi- Abnormal proteolysis of polyglutamine tion of caspase cleavage reduces aggregate disease proteins is one possible mecha- formation and toxicity in a cell culture nism of toxicity. Some studies have shown model of HD, underscoring the relevance that polyglutamine repeat length modu- of proteolytic cleavage in polyglutamine lates susceptibility to proteolytic cleavage, disease pathogenesis. Cleavage promotes but results have been inconclusive. Few aggregation and/or toxicity in SCA3 and studies have been performed using hu- SCA7 (Fig. 8). Polyglutamine tract con- man brain tissue, however. Considerable taining htt fragments are more toxic in cell evidence exists to show that polyglutamine culture than full-length htt, and the most tracts themselves are resistant to degrada- widely used HD mouse model expresses tion by mammalian proteasomes. The 20S only exon 1 of the htt gene. and 26S eukaryotic proteasomes are in- There are several mechanisms by which capable of cutting within polyglutamine cleavage may modulate polyglutamine tracts longer than 9 amino acids in vitro. neurotoxicity. Abnormal proteolysis of The inability of the proteasome to di- polyglutamine proteins may lead to a toxic gest glutamine tracts may contribute to species that can cause damage in the the toxicity and aggregation of expanded soluble or insoluble phase. Alternatively, polyglutamine peptide fragments. 32 Triplet Repeat Diseases

Proteolysis may be a normal part of the There are many proteases that are sus- turnover of wild-type and mutant htt. One pected to play a role in the proteolytic group published evidence that caspase-3 cleavage of polyglutamine proteins, the cleaves both mutant and wild-type htt in most studied of which are the caspases. the cytoplasm of HD brains, suggesting Caspases are cysteine- dependent aspartyl that this particular cleavage is a normal proteases that play a crucial role in apop- event in its turnover. Since this fragment tosis, but are increasingly being inves- was located exclusively in the cytoplasm, it tigated for their nonapoptotic functions. also implies that further cleavage occurs In healthy cells, caspases are present pre- before the nuclear translocation of htt. dominantly as inactive proenzymes, which Another study suggested that pepstatin- must be cleaved for full activity. Caspase-3 sensitive aspartyl endopeptidases such as is the only protease that has been shown the presenilins and cathepsins D and E to cleave htt in HD patient’s material, al- may be involved in the normal turnover though caspases 1 and 6 have also been of htt. This process, which may normally implicated in cell culture and in vivo. aid in maintaining the balance between htt Inhibiting the cleavage of htt by any of synthesis and degradation, could generate these three proteases is protective, and an a toxic, highly aggregation-prone (and HD mouse model expressing dominant- negative caspase-1 showed significantly hence intermediate oligomer/protofibril- delayed disease onset. Caspase-3 has also prone) species when it cleaves mutant htt. been implicated in the proteolytic cleavage Another possible source of toxic cleav- of atrophin-1, ataxin-3 and AR. Caspases age products is the ubiquitin-proteasome are activated in response to a number of system, which is responsible for labeling cellular insults, and in postmitotic cells anddigestingmisfoldedproteins,among such as neurons where inhibitors of apop- other things. Polyglutamine tracts beyond tosis proteins (IAPs) are highly expressed, a threshold of about 35 glutamines adopt this may cause cellular damage that does β an abnormal -pleated sheet conforma- notresultinclassicalapoptosis. tion, which may cause the host protein Calpains are another family of pro- to be ubiquitinated and thus targeted for teases implicated in the cleavage of degradation. In support of this theory, polyglutamine proteins. They are calcium- a common feature of polyglutamine dis- activated cysteine proteases that exist pre- eases is the presence of ubiquitin-positive dominantly as proenzymes. As mouse inclusions. Arguing against this idea is models of HD and material from HD pa- an experiment showing that in a cell cul- tients exhibit abnormal mitochondrial cal- ture model of SCA1, increased proteasome cium regulation, some investigators have degradation due to the introduction of proposed that altered calcium flux is the a degradation signal reduces aggregates mechanism of calpain activation. Indeed, and toxicity. Another potential source of several cell culture studies support the role toxic fragments is the autophagy pathway, of calpain proteases I, II, and ‘‘m’’ in htt which is responsible for the bulk degrada- cleavage. As mentioned previously, other tion of cellular components. According to aspartyl proteases such as the Alzheimer’s one study, degradation of htt by autophagy disease-associated presenilins and auto- and the autophagosome-associated cathep- phagy-associated cathepsins have also sin D creates toxic htt fragments. been implicated in htt cleavage. Triplet Repeat Diseases 33

4 there is a linear correlation between re- Type 2: the Loss-of-function Repeat duced protein levels and IQ test scores. Diseases Expansion of the disease allele results in hypermethylation of the (CGG)n tract, 4.1 Fragile X Syndrome which spreads to a nearby CpG island in the FMR1 promoter region. Some of the Fragile X syndrome (FRAXA) is an X- proposed secondary structures formed by linked disorder with a prevalence of 1 the FRAXA repeat contain C–C mispairs, in 4000 in males and 1 in 8000 in fe- which are good targets for human DNA males, making it the most common form methyltransferase. Another theory invokes of inherited mental retardation. Neuro- the RNAi protein Dicer’s ability to cleave logical presentation frequently includes CGG repeat RNA, postulating that the re- mild to severe mental retardation, hy- sulting siRNAs recruit DNA methyltrans- peractivity, poor eye contact, high-pitched ferases to the FMR1 locus. Affected alleles speech, and flapping or biting hand move- also display condensed chromatin, loss of ments. Physical signs in males include histone acetylation and increased histone long, prominent ears, and jaws, macro- methylation. These data support a process cephaly, postpubescent macroorchidism, whereby DNA methylation recruits tran- and occasionally, connective tissue abnor- scription silencing machinery, which sub- malities. The mutation responsible for sequently suppresses FMR1 transcription FRAXA is typically an expansion of a in the nearby promoter region. Interest- polymorphic (CGG)n repeat found in the ingly, while premutation carriers had been 5 -untranslated region of the fragile X men- viewed as perfectly normal for decades, tal retardation-1 (FMR1) gene. Expanded recent work indicates that some premuta- chromosomes have a folate-sensitive frag- tion carriers (60–200 uninterrupted CGG ile site at Xq27.3 that can be viewed under repeats) display a phenotype distinct from a light microscope under special cell cul- FRAXA. Females have a predisposition ture conditions. Normal alleles contain to premature ovarian failure (POF), and 6–53 triplets punctuated by one or more males may develop late-onset ataxia and AGGs, which are considered to have a tremor with the presence of neuronal in- stabilizing influence on the repeat. Dis- tranuclear inclusions (NIs) that consist ease alleles contain expansions beyond 200 ofRNA.(ThisnewFXTASisdiscussed and up to 2000 repeats, with no AGG in- in a separate section.) Expression stud- terruptions. Pathogenic expansions result ies have shown that premutation carriers exclusively from maternal transmission. may express up to seven times more FRAXA appears to be a loss-of-function FMR1 mRNA than normal individuals. disorder, since deletion of FMR1 and This upregulation is probably not due to loss-of-function point mutations can also compensation for loss of function, be- causeFRAXA.Thisviewisfurthersup- cause individuals with point mutations ported by the FMR1 knockout mouse, resulting in loss of function of FMRP do which reproduces certain aspects of the not have higher levels of the transcript. human disorder. Abnormal transcript levels due to premu- FRAXA patients have reduced levels tations or full mutations are thought to of the FMR1 gene product, FMRP, and be as a result of the expansion’s effect 34 Triplet Repeat Diseases

on transcription initiation rather than on 3 -untranslated region. FMRP associates mRNA stability. with miRNAs, and proteins in miRNA- The FMR1 promoter lacks a functional containing complexes in mammals, inter- TATA box and initiator sequence. Four actions that may be relevant to FMRPs functional transcription factor binding regulation of translation. sites have been identified in normal in- FMRP is important for the develop- dividuals: two sites binding USF1/USF2 ment of dendritic spines and synaptic and Nrf-1, and two GC boxes that bind plasticity. It regulates the expression of members of the ‘‘Sp’’ family of transcrip- MAP1B, an important regulator of mi- tion factors. Sp1 and Nrf-1 binding is crotubule stability. FMRP knockout mice disrupted in cells from FRAXA patients. have abnormally high levels of MAP1B, The FMR1 gene encodes the widely ex- resulting in increased microtubule stabil- pressed FMRP, which is an RNA-binding ity. This may affect the development of protein most highly expressed in the brain dendrites and/or dendritic spines. The ab- and testes. It is 60% homologous to two sence of FMRP in hippocampal neurons other proteins, FXR1P and FXR2P, with results in immature dendritic spine mor- which it interacts. The protein is thought phology and delayed synaptic connections, to bind approximately 4% of all brain perhaps contributing to the neurological mRNA transcripts, through its RGG box phenotype observed in FRAXA. FMRP domain. It selectively recognizes RNA- may also affect mGluR5-dependent long containing hairpin or tetraplex secondary term depression (LTD). LTD is a process by structures (‘‘G-quartet’’), and shuttles into which neuronal activity can cause a lasting and out of the nucleus and associates with desensitization of neurons to depolariza- polyribosomes in messenger ribonucleo- tion. LTD-associated protein synthesis at protein complexes (mRNPs). The FMRP- synapses may be enhanced by mGluR5 containing mRNP complexes also contain activation, while it appears to be sup- Pur α and mStaufen, proteins involved pressed by FMRP. This is supported by in the transport of neuronal granules. evidence of enhanced LTD in the FMR-1 These granules, which contain RNA and knockout mouse. FMRP suppression and associated proteins, are transported to den- mGluR5 activation of LTD-dependent local dritic spines in a metabotropic glutamate protein synthesis may be opposing forces receptor 5 (mGluR5)-dependent manner. that are out of balance in FRAXA, per- FMRP can suppress the translation of cer- haps accounting for part of the cognitive tain transcripts in vitro and in vivo. FMRP abnormalities. may thus regulate the transport, localiza- tion, and translation of certain mRNAs in 4.2 an activity-dependent manner (Fig. 9). In Fragile XE Mental Retardation Drosophila, FMRP interacts with compo- nents of the RNAi machinery, which is Fragile XE mental retardation (FRAXE) is involved in gene silencing and thus trans- an X-linked disorder with a prevalence of lational control. Although FMRP does not about 1–4% that of FRAXA. It accounts seem to affect the siRNA pathway, it may for approximately two-thirds of families regulate microRNAs (miRNAs), which are with nonspecific X-linked mental retar- noncoding RNAs thought to control the dation(MRX),aclassificationinwhich translation of mRNAs by binding to their mental retardation is the only consistent Triplet Repeat Diseases 35 le X or rted to dendrites. Whether in the perinuclear n either directly associates with ribosomes .). ds to recently transcribed messenger RNAs in the nucleus to form a Nature Cell Biology n. (From Jin, P., Alisch, R.S., Warren, S.T. (2004) RNA and microRNAs in fragi x (RISC). The FMRP–mRNP complex is also transpo –mRNP complex moves out of the nucleus and the , 1048–1053. Used with permission of 6 Nat. Cell Biol. Advances in the understanding of FMRP’s normal function. FMRP bin ribonucleoprotein complex (mRNP). The FMRP mental retardation, interacts with the RNA-induced silencingcytosol comple or in dendrites, FMRP is believed to regulate protein translatio Fig. 9 36 Triplet Repeat Diseases

clinical feature. Other characteristics are the human FMR2 gene resulting in altered variable and may include hyperactivity, transcription in many parts of the devel- mild facial hypoplasia and nasal abnormal- oping nervous system and mature brain. ities. FRAXE is caused by the expansion The neurological deficits characteristic of of a (GCC)n repeat in the promoter re- FRAXE may involve effects of altered tran- gion of FMR2, a gene 600 kb downstream scription on the hippocampus, amygdala, of FMR-1.Italsooverlapswiththepu- and possibly other brain structures. tative promoter of a gene called FMR3, transcribed in the opposite direction. Both 4.3 maternal and paternal transmission can re- Friedreich’s Ataxia sult in expansion. Normal individuals bear 6–35 GCC repeats, premutation carriers With a prevalence of about one per 50 000 bear 61–200 repeats, and full mutation individuals in the Caucasian population, carriers bear >200 repeats. Pathogenic ex- Friedreich’s ataxia (FRDA) is the most pansion results in a folate-sensitive fragile common inherited ataxia in this ethnic site at the disease locus. group. It is a multisystem degenerative In mice, FMR2 mRNA is expressed in disease that is unusual among the triplet adult and fetal brain, kidney, lung, and pla- repeat disorders due to its autosomal centa, with highest brain levels in the hip- recessive inheritance. Neurological symp- pocampus and amygdala. Transcripts of toms include gait, limb and truncal ataxia, both FMR2 and FMR3 are reduced beyond loss of position and vibration senses, di- detection in FRAXE patients. This may be minished tendon reflexes, and dysarthria. due to a methylation-dependent silencing Neuropathology changes include degener- process similar to FRAXA. The presence of ation of the posterior columns of the spinal the repeat within the preinitiation region cord, loss of large primary sensory neurons of the FMR2 promoter may also suggest in the dorsal root ganglia (DRG), and mild, a more direct disruption of transcription. late-onset degeneration of the cerebellar FMR2 mRNA is highly expressed in the cortex. Other common clinical features are hippocampus, an area critical for learn- cardiomyopathy, diabetes mellitus, scol- ing and memory. Its paralogs AF4 and iosis, and other skeletal abnormalities. LAF4 are both transcription transactivator Patients often present with symptoms in proteins, and the FMR2 protein seems to childhood, become wheelchair-bound by be a potent transcription activator itself. their late teens or early twenties, and Furthermore, the FMR2 protein is nu- have reduced lifespans due to cardiac fail- clear, consistent with its proposed role in ure. Adult presenting patients, however, transcription regulation. The Drosophila can have nearly normal lifespans with ortholog of FMR2/AR4, Lilliputian, is es- more protracted progression and less se- sential for proper organ development, and vere nonneuronal involvement. Indeed, itslossoffunctionislethal.FMR2knock- until the identification of the causal muta- out mice display impairment in the condi- tion, many of these adult-onset cases went tioned fear test and enhanced LTP (long- undiagnosed. term potentiation) in the hippocampus. FRDA results from the expansion of a The role of FMR3 in FRAXE, if any, is un- polymorphic (GAA)n repeat in the first known. The mechanism of FRAXE patho- intron of the gene X25, now known as genesis may therefore involve silencing of frataxin. FRDA patients have at least one Triplet Repeat Diseases 37 expanded allele. For the disease to occur, to mammals. It contains mitochondrial the second allele must contain either an ex- targeting signal sequences and localizes panded repeat, or rarely, a copy of frataxin to the mitochondrial matrix. Frataxin containing a loss-of-function mutation. No expression occurs at the primary sites patients have been described with loss-of- of pathology: dorsal root ganglia, spinal function allele mutations in both FRDA cord, sensory nerves, heart, and pancreas. genes. This is probably due to prenatal These are tissues that rely upon high levels lethality, as frataxin-null mice die in utero. of oxidative metabolism and consequently Disease severity and onset age are deter- are rich in mitochondria. In addition, such mined by the size of the (GAA)n repeat cell types are often postmitotic, meaning of the smaller expansion allele. Unaffected that most dividing cell types are spared in individuals carry at least one allele between FRDA patients. 7–38 repeats, while affected individuals Studies with FRDA patient mate- carry two alleles of 66–1700 repeats. In- rial have demonstrated an increased terruptions of the (GAA)n repeat result heart iron content and a deficiency in in later onset and attenuated presenta- iron–sulfur cluster-containing proteins, tions. GAA expansion at the FRDA locus including aconitase, a protein involved can result in enormous repeats in a sin- in iron homeostasis. In addition, fibrob- gle generation. Maternal transmission can lasts derived from FRDA patients are result in expansion or contraction, while abnormally sensitive to iron and hydro- paternal transmission results primarily in gen peroxide induced stress. Ablation contraction. of the yeast frataxin homolog results in FRDA patients exhibit reduced frataxin respiratory dysfunction, abnormal accu- RNA and protein levels, and evidence sug- mulation of mitochondrial iron, impaired gests a defect in transcription or RNA biogenesis of iron–sulfur proteins, and maturation. Current models propose that increased sensitivity to oxidative stress. frataxin RNA elongation may be disturbed Conditional knockout of frataxin in stri- by triplex structure formation between ated muscle results in a heart-specific expanded DNA strands during transcrip- phenotype resembling the cardiac abnor- tion. DNA triplexes are formed when one malities in human FRDA (Fig. 10). This strand of a double-stranded DNA molecule supports a genetic mechanism involving folds back upon itself, and interacts with loss of function of the disease protein. two previously annealed strands. This cre- Frataxin may be part of a complex that de- ates a local structure comprising three livers iron to Iron–sulfur clusters (ISCs), DNA strands held together by hydro- which are cofactors essential to the activ- gen bonds. Certain sequences favor this ity of many important cellular proteins. process, among them extended (GAA)n The accumulation of iron outside the mi- repeats. Triplex formation in the first in- tochondria, although probably not central tron would presumably affect transcript to FRDA pathogenesis, supports a defi- elongation but not initiation. In support of ciency in iron delivery to ISCs. Among this model are in vitro transcription exper- ISC-dependent cofactors are proteins in- iments demonstrating no effect of repeat volved in mitochondrial electron trans- length on transcript initiation. port and thus respiration. Disturbances Frataxin is a 210 amino acid protein in oxidative metabolism are often associ- that is well conserved from prokaryotes ated with increases in the production of 38 Triplet Repeat Diseases

Fig. 10 Friedreich’s ataxia mice display cardiac muscle pathology. (a) Transmission electron microscopy of mice lacking expression of frataxin in only their muscle reveals the abnormal accumulation of lipid droplets (L) in cardiac muscle at 4 weeks of age. (b) By 7 weeks of age, mitochondria appear (a) abnormal and iron deposits are visible (arrows). (c) With further progression, large vacuoles emerge (inset-top) as mitochondria undergo continued prominent degeneration. Ultimately, the mitochondria become engorged with electron-dense material consistent with iron deposits. (From Seznec, H., Simon,D.,Monassier,L., Criqui-Filipe, P., Gansmuller, A., Rustin, P., Koenig, M., Puccio, H. (2004) (b) Idebenone delays the onset of cardiac functional alteration without correction of Fe-S enzymes deficit in a mouse model for Friedreich ataxia, Hum. Mol. Genet. 13, 1017–1024, used with permission of Human Molecular Genetics).

(c)

toxic reactive oxygen species (ROS). FRDA MERRF (myoclonus epilepsy with ragged therefore is thought to result from re- red fibers), which are caused by stable mu- duced frataxin levels, leading to abnormal tations to mitochondrial-encoded proteins. iron–sulfur metabolism, mitochondrial FRDA appears to be a classic mitochon- dysfunction, oxidative stress, and tissue drial disorder with an unusual genetic degeneration. FRDA is thus caused by the basis – dynamic mutation. dynamic mutation of a nuclear-encoded Given the data suggesting a role for ox- mitochondrial protein. The disease shares idative metabolism and iron transport in features with classic mitochondrial dis- FRDA pathogenesis, antioxidants and iron orders such as MELAS (mitochondrial transport molecules are being evaluated as myopathy, encephalopathy, lactic acido- treatments for this disease. Several clini- sis, and strokelike episodes syndrome) and cal trials have been conducted in FRDA Triplet Repeat Diseases 39 patients using idebenone, a synthetic ana- The protein binds to and inhibits lysoso- log of coenzyme Q10 and a potent antioxi- mal proteases such as cathepsins B, H, L, dant. Some studies have shown substantial and S. Pathogenic expansion of the repeat improvements in the heart function of causes a reduction in CSTB transcription patients, although the lack of random- in some cell types. Cstb knockout mice ized placebo controls and the variability of develop symptoms similar to those seen in the disorder render the results controver- EPM1 patients, and missense mutations sial. A recent conditional knockout mouse causing the disease disrupt the ability of model of FRDA also supports the use of CSTB to bind cysteine proteases. Together, idebenone as a treatment for the disorder, these data suggest a loss-of-function mech- making treatment of FRDA patients with anism for EPM1. Lowered inhibition of this antioxidant a distinct possibility. cysteine proteases may result in neuronal damage, causing the phenotype observed 4.4 in EPM1. Progressive Myoclonus Epilepsy Type 1 There are several ways in which the repeat might reduce CSTB expression. Progressive myoclonus epilepsy type 1 The first is that reduced gene expres- (EPM1) is a rare, autosomal recessive sion may be due to an increase in the neurological disorder most prevalent in distance between promoter elements and Finland and parts of North Africa. Its most the transcription start site. AP-1 bind- prominent symptoms are progressive pho- ing sites have been located upstream of tosensitive myoclonus and tonic-clonic the repeat. One study demonstrated a epilepsy. Some patients also experience a two- to fourfold reduction in promoter ac- progressive cerebellar ataxia and cognitive tivity when the repeat is expanded. An decline. Neurodegeneration occurs in the equivalent reduction also occurred when thalamus, spinal cord, and the Purkinje similarly sized heterologous DNA was and granular neurons of the cerebellum. used in place of the expanded repeat. In- Typical age of onset is 6 to 16 years of age. dependent studies have confirmed a large EPM1 is sometimes caused by missense reduction in promoter activity due to re- mutations in the cystatin B gene (CSTB); peat expansion. Another way the repeat however, analysis of affected patients lack- could affect CSTB expression is through ing such mutations revealed a dodecamer alterations in chromatin structure. The re- repeat expansion upstream of CSTB. The peat region is G-C rich and might exclude dodecamer repeat is located between 66 nucleosomes or promote the formation of and 77 bp 5 of two putative transcription secondary structures. Abnormal DNA sec- start sites, and has the sequence CCCCGC- ondary structures have been observed in CCCGCG. Unaffected individuals have a the EPM1 repeat region. These alterations repeat number of 2–3, premutation car- could affect the expression of CSTB or riers have 12–17 repeats, and affected other genes in the vicinity. A third possible individuals carry 30–150 repeat alleles. mechanism for expression changes lies in CSTB is a highly conserved gene in thesequenceoftherepeatregion.Eachdo- the cystatin family of cysteine protease in- decamer repeat contains a GC box, which hibitors. CSTB mRNA is ubiquitous, with could act as an Sp1 binding site in vivo.In high transcript levels in the hippocampus. cases of abnormally large repeat numbers, 40 Triplet Repeat Diseases

increased Sp1 binding could contribute to segregating DM1. Although the anticipa- the suppression of CSTB transcription. tion phenomenon was dismissed as an artifact of ascertainment by geneticists of the mid twentieth century, anticipation 5 is now known to be a genuine feature Type 3: the RNA Gain-of-function Repeat of dynamic mutation diseases. Anticipa- Diseases tion results from the tendency of disease repeats to expand and the inverse corre- 5.1 lation between repeat length and age of Myotonic Dystrophy Type 1 disease onset. Several theories have been advanced Myotonic dystrophy or DM is an au- to explain the molecular basis of DM1, tosomal dominant, multisystem disorder including haploinsufficiency of DMPK, with a prevalence of 1 in 100 000 world- local chromatin effects on neighboring wide, and 1 in 8000 in European and North American Caucasian populations. genes, and gain of function exerted by Patients typically present with proximal expanded RNAs. DMPK is proposed to or distal muscle dysfunction including have many functions, some of which could weakness, pain, and myotonia (failure of relate to the disease, such as modula- muscle relaxation). DM exhibits a combi- tion of skeletal muscle sodium channels, nation of characteristic symptoms: cardiac RNA metabolism, calcium homeostasis, conduction abnormalities, subcapsular iri- and the cell stress response. Initial ex- descent cataracts, and unusual endocrine pression studies reported a decrease in changes. Other features include testicular DMPK RNA and protein levels, but a atrophy, type II diabetes, and late-onset DMPK knockout mouse designed to test cognitive impairment. In its most severe the haploinsufficiency theory exhibited form, congenital DM, mental retardation, only mild myopathy that was inconsis- craniofacial deformities, and other devel- tent with the DM phenotype. The (CTG)n opmental abnormalities are present. It tract is a strong nucleosome assembly exhibits both paternal and maternal trans- site, and it was therefore hypothesized mission, although there is an almost exclu- to have trans-effects on the expression sive maternal transmission in congenital of neighboring genes. According to this DM. Myotonic dystrophy type 1 (DM1) is model, the myriad symptoms of DM1 are caused by a (CTG)n expansioninthe3- caused by disturbances in the expression untranslated region of the gene, dystroph- of multiple nearby genes due to the expan- ica myotonica protein kinase (DMPK), on sion – making DM1 a ‘‘contiguous gene chromosome 19. Unaffected individuals syndrome.’’ There are several genes in carry 5–37 repeats, while affected individ- close proximity to the DMPK gene (i.e. uals carry 50–4000 repeats. DM exhibits <5 kb) whose roles in DM1 were con- pronounced anticipation and dramatic so- sidered; the most studied of these genes matic instability. Interestingly, the clinical has been SIX5. The SIX5 homolog in phenomenon of anticipation (worsening Drosophila is required for normal eye severity as a disease gene is transmitted development and its mouse homolog is from one generation to the next) was first involved in regulating distal limb muscle described nearly a century ago in a family development. Expression data on SIX5 in Triplet Repeat Diseases 41

DM1 patients has been inconsistent, how- but noncoding regions of the genome. ever. A Six5 knockout mouse developed More support for the RNA gain-of-function cataracts, although they were not of the model came from a mouse model gen- type observed in DM1. Another gene im- erated by Mankodi et al. in 2000, which plicated by this model is DMWD,which contained a (CTG)n expansioninthe3 is expressed in the testis and suspected untranslated region of the skeletal actin to be involved in male infertility. Studies (HSA) gene. This mouse exhibited myopa- indicated that the expression of DMWD is thy typical of DM1, although the skeletal not altered. muscle-restricted expression pattern of The third theory for the molecular basis HSA precludes broader conclusions. Thus, of DM proposes that an RNA gain-of- there are several lines of evidence suggest- function mechanism is responsible. RNA ing that CUG expansion-containing RNAs foci, which are accumulations of expanded are capable of causing DM. transcripts, accumulate in the nuclei of RNA-binding proteins and transcription patient cells (Fig. 11). The RNA gain-of- factors colocalize with the RNA foci found function theory was buttressed by the in DM, potentially altering nuclear pro- discovery that DM2, a disorder with symp- cesses. Elevated levels of CUG-containing toms almost identical to DM1, is caused RNA has been shown to alter gene splic- by a (CCTG)n expansion in intron 1 of ing in specific transcripts which could the zinc finger 9 protein (ZNF9). The two be relevant to DM: cardiac troponin T genes responsible for DM1 and DM2 are (cTNT), involved in cardiomyopathy; In- unrelated and reside on different chro- sulin Receptor (IR), involved in diabetes; mosomes. Genes near the two loci bear and Clc-1, the main chloride channel in no obvious resemblances. The only strik- muscle. Abnormal splicing of cTNT, IR, ing parallels between the two expansions and ClC-1 are proposed to account for are: (1) they both contain CTG triplets; the cardiac abnormalities, insulin insen- and (2) they both occur in transcribed sitivity and myotonia observed in DM.

(a) (b) (c) Fig. 11 In situ hybridization of muscle sections with fluorescently labeled antisense oligonucleotide probes reveals accumulation of mutant RNA in DM1 and DM2. (a) Probing of DM2 muscle with a CAGG probe indicates that multiple RNA foci are present. (b) Probing of normal muscle with a CAGG probe demonstrates absence of RNA 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 expansion in intron 1 of ZNF9, Science 293, 864–867, used with permission of Science). 42 Triplet Repeat Diseases

Splicing abnormalities cause a reduction splicing abnormalities that cause DM’s di- in the membrane concentration of ClC- verse range of symptoms. 1 and reduce chloride conductance to levels consistent with myotonia. These 5.2 spicing alterations are thought to be due Myotonic Dystrophy Type 2 to the repeat-expanded RNA’s effects on two families of RNA-binding proteins: Myotonic dystrophy type 2 (DM2) is an ‘‘CUG-BP1 and ETR-3-like factors’’ (CELF) autosomal dominant, multisystem disor- and ‘‘muscleblind-like proteins’’ (MBNL). der very similar to DM1. The majority of CELF proteins regulate pre-mRNA splic- its symptoms resemble DM1: progressive ing in cTNT, IR, and ClC-1. In patient weakness, myotonia, cardiac disturbances, tissue and cell culture, CUG-BP1 levels iridescent cataracts, and insulin insensi- and activity are increased in response to tivity. There are some notable differences, elevated levels of CUG-containing RNA. however. DM2, unlike DM1, predomi- This may be due to a lengthening of the nantly affects proximal muscles at its protein’s half-life. MBNL proteins were onset, which is why many cases were named after their Drosophila ortholog originally classified as proximal myotonic muscleblind, which is required for pho- myopathy (PROMM). Other interesting toreceptor and muscle differentiation in differences are that mental retardation is flies. The three known proteins in this not observed in DM2, DM2 patients show family (MBNL1 (MBNL), MBNL2 (MBLL), increased sweating, and DM2 congenital and MBNL3 (MBXL)) are splicing reg- forms have not been observed. DM2 is ulators that are thought to act antago- caused by a (CCTG)n tetranucleotide ex- nistically to CELF family proteins. They pansion in the first intron of the zinc finger colocalize with RNA foci in vivo,anda protein 9 (ZNF9) gene on chromosome mouse model lacking specific isoforms 3q21.3. The tetranucleotide repeat can ex- of MBNL1 recapitulates the myotonia, pand to stunning lengths, with the longest cataracts, and splicing dysregulation ob- cases comprising 44 kb of DNA, making served in DM. them the longest tracts observed in the re- The splicing alterations seen in DM peat expansion disorders. The DM2 locus patients are consistent with loss of func- exhibits marked somatic instability. Over tion of MBNL proteins or an increase in the course of a patient’s lifetime, the aver- CELF protein activity in muscle and brain. age repeat length increases substantially, Because of the colocalization of MBNL pro- as judged by blood drawings. teins with RNA foci, it has been proposed The similarities between DM1 and DM2 that sequestration and subsequent deple- are not restricted to the clinical presen- tion of these proteins from the cellular tation. The expansions are both large, milieu is responsible for the symptoms of CTG-containing tracts that are transcribed, DM. Although this process probably plays but not translated. Both disorders cause a critical role in the disease, recent evi- nuclear RNA foci that sequester specific dence suggests it is not solely responsible. RNA-binding proteins (Fig. 11), including The most plausible current theory explain- CELF and MBNL family members. Similar ing DM pathogenesis proposes an im- splicing abnormalities are also observed in balance between the antagonistic MBNL DM2. Since the genes associated with the and CELF proteins, resulting in specific repeat expansions in DM1 and DM2 are Triplet Repeat Diseases 43 unrelated, it is likely that the cause of both the cause of SCA8. Numerous subsequent disordersisatoxicgainoffunctionoflong studies, however, have indicated that pos- tracts of CUG-containing RNA. A mouse session of the expanded CTG repeat tract model generated by Mankodi et al. in 2000, appears necessary, but not sufficient, for which contained a (CTG)n expansion in the production of the SCA8 phenotype. the 3 untranslated region of the skeletal Thus, reduced penetrance is viewed as a actin (HSA) gene, exhibited myopathy typ- key feature of SCA8 at this time. On the ical of DM. Thus, even outside the context basis of the available genetic data, nor- of the DM1 and DM2 loci, CTG-containing mal individuals always carry fewer than 70 expansions can cause DM-like symptoms. CTG repeats, while affected SCA8 patients This supports the RNA gain-of-function can carry anywhere from 71 to >1000 CTG theory, as transcribed CTG-containing ex- repeats. As extremely large CTG repeat al- pansions are inherently capable of causing leles (>800 triplets) were shown to not disease. Differences between DM1 and cause disease in early reports, this was ini- DM2 may be due to regional or tempo- tially attributed to lack of stability of the ral expression patterns or differences in mutant RNA product. However, further the affinity of RNA-binding proteins for work has revealed considerable overlap CTG or CCTG tracts. between disease-causing CTG repeat ex- pansions and nonpenetrant CTG repeat 5.3 alleles, indicating that a secondary fac- Spinocerebellar Ataxia Type 8 tor – either a trans-acting genetic factor or an environmental factor – must be present Spinocerebellar ataxia type 8 (SCA8) is a dominantly inherited cerebellar ataxia. Af- to yield the SCA8 disease phenotype. In the fected individuals suffer from late-onset, case of the original MN-A family, evidence slowly progressing gait ataxia as well for a cis modifier appears responsible for as dysarthria, oculomotor incoordination, the extremely high penetrance in this large spasticity, and decreased vibration sense. pedigree. Thus, while the causality of the The cerebellar cortices and vermis un- CTG repeat expansion in SCA8 has been dergo a slowly progressing but dramatic somewhat controversial, review of the cur- atrophy, while the brainstem exhibits lit- rent literature suggests that the CTG repeat tle evidence of degeneration. Patients may expansion is directly involved in SCA8, become wheelchair-bound as early as their but may not alone be sufficient to produce fourth decade of life. Using a direct method the disorder. for cloning expanded triplet repeats (i.e. Another unique feature of the SCA8 RAPID cloning), a gene containing an CTG repeat is its extreme and unusual expanded CTG tract was identified in a genetic instability. Paternal transmissions large family (MN-A) segregating the SCA8 generally result in contractions, while ma- phenotype. Interestingly, this CTG is con- ternal transmissions generally result in ex- tained within a gene on the long arm pansions. Expansions of up to 600 repeats of chromosome 13 that is transcribed, have occurred in one generation through but apparently not translated into a pro- maternal transmission. Large deletions in tein product. Thus, it was proposed that expanded alleles often occur in sperm cells, the production of an RNA transcript con- offering an explanation for the paternal taining this expanded CTG repeat tract is contraction bias. Pathogenic expansions 44 Triplet Repeat Diseases

often have 5 triplet interruptions, the role a slowly progressive gait ataxia reminiscent of which is unclear. of the human disease. While no protein The molecular basis of how the ex- product has been detected, 1C2 and ubiq- panded SCA8 RNA causes disease remains uitin antibody staining reveal intranuclear unclear. The transcript has been detected inclusions in cerebellar neurons, consis- exclusively in the brain, and appears to be tent with the RNA-containing inclusions transcribed in the CTG orientation. Trans- observed in FXTAS brains. An alternative lation of a polyglutamine tract from the interpretation is that translation of CTG CAG-containing transcript in the oppo- into a polyleucine tract is occurring. Fur- site direction is precluded by the existence ther work will be needed to distinguish of stop codons flanking the repeat. The these two possibilities. longest transcript identified to date con- Studies of the SCA8 CTG repeat expan- tains 6 exons, and many alternatively sion in D. melanogaster have also been spliced forms of the gene have been de- informative. Whether directing expression tected, although all these isoforms are of the SCA8 gene to fly retina with a nor- expressed at very low levels. None con- mal CTG repeat tract or an expanded CTG tain significant open reading frames. In repeat tract, a neurodegenerative eye phe- 1999, Koob et al. reported that a gene par- notype results. Modifier screens using fly tially overlaps the 5 end of the SCA8 gene stocks carrying mutations in genes encod- locus on the antisense strand. This gene, ing RNA-binding protein yielded a number called Kelch-like 1 (KLHL1), for its ho- of genes that could either enhance or mology to the Drosophila KELCH gene, suppress this retinal degeneration pheno- is highly conserved and predicted to en- type. Interestingly, Drosophila muscleblind, code an actin-binding protein of 748 amino whose mammalian counterpart has been acids. Its expression overlaps with that of implicated in the DM RNA toxicity path- the SCA8 transcript, suggesting that the way, modified the retinal degeneration SCA8 gene may produce an antisense RNA caused by expanded SCA8 CTG repeat whose function is to regulate KLHL1 ex- expression more so than the retinal de- pression. Whether or how this is occurring generation caused by normal SCA8 CTG remains uncertain at this time. repeat expression. Such data supports the With the discovery of the DM RNA toxic hypothesis that the SCA8 CTG repeat gain-of-function pathway, an emerging expansion is producing neurotoxicity by theory for SCA8 pathology holds that the altering the function of RNA-binding pro- production of a CUG-expanded transcript teins as in DM. Although it would be results in RNA gain-of-function toxicity premature to definitively conclude that within the restricted neuronal populations SCA8 is an RNA gain-of-function repeat where the SCA8 gene is expressed. In sup- disease, provisional classification in this port of this hypothesis, transgenic mice category is appropriate. derived with a human bacterial artificial chromosome containing the entire SCA8 5.4 gene with a 118 CTG repeat expansion de- The Fragile X Tremor – Ataxia Syndrome velop neurological disease. The severity of (FXTAS) their phenotype depends on the expression level of the transgene. In the more moder- Fragile X tremor-ataxia syndrome (FX- ate expressing lines, the SCA8 mice display TAS) is a rare and unusual disorder Triplet Repeat Diseases 45 that is associated with premutation of Males with tracts between 55 and 200 CGG the FRAXA locus in males. Patients de- repeats are considered to be in the ‘‘pre- velop a progressive ataxia and intention mutation range,’’ and are now considered tremor. This is sometimes accompanied by to be at risk for FXTAS beyond middle age. dementia, parkinsonism, and autonomic In contrast to FRAXA, FXTAS patients dysfunction. Neuropathological changes show increased expression of the FMR1 include degeneration of the cerebellum transcript, with longer repeats correspond- and ubiquitin-positive intranuclear inclu- ing to higher transcript levels. Premutation sions in neurons and glia. FXTAS is asso- carriers with more than 100 repeats have ciated with the expansion of a (CGG)n tract an average of five times more FMR1 in the 5 -untranslated region of the FMR1 mRNA than individuals in the normal gene, found on chromosome Xq27.3. Ex- repeat range. Despite the increase in tran- pansion of this tract beyond 200 repeats script level, the protein product of the causes FRAXA, a disorder resulting from FMR1 gene,FMRP,isslightlyreducedin a reduction in the expression of FMR1. FXTAS patients. Since elevated transcript

Fig. 12 Intranuclear inclusions are present in the brains of FXTAS patients. (a) Hematoxylin & eosin staining of cerebral neurons reveals a refractile, eosiniphilic nuclear inclusion of ∼5 µMindiameter (white arrowhead). (b) Hematoxylin & eosin staining of cerebral astrocytes reveals refractile, eosiniphilic nuclear inclusions of ∼2 µMindiameter(white µ arrowhead). (c) Antineurofilament (a)10 m (b) 10 µm antibody staining of cerebellum demonstrates presence of dystrophic neurites, consistent with ongoing Purkinje cell degeneration. (d) Antiubiquitin antibody staining of cerebral neurons labels intranuclear inclusions. (e) Antiubiquitin positive intranuclear inclusions form in both neurons (white arrowhead; larger (c)µ (d) µ cell) and astrocytes (white 50 m 10 m arrowhead; smaller cell). (From Greco et al. (2002) Brain 125, 1760, used with permission).

(e) 10 µm 46 Triplet Repeat Diseases

levels are present without great alteration exceeding 30 triplets. Most are composed in protein levels, this disorder may be of ‘‘imperfect’’ alanine tracts, including due to an RNA gain of function of the combinations of GCG, GCA, GCC, and FMR1 transcript. Indeed, immunostain- GCT codons. This composition argues ing of FXTAS patients’ brain sections against an expansion mechanism involv- reveal intranuclear inclusions in neurons ing single-strand, hairpin-like secondary and glia, believed to be comprised of structures as in other repeat disorders. accumulated Fmr1 RNA transcripts and A more plausible mechanism for polyala- various proteins (Fig. 12). To determine nine expansion is unequal crossing-over, the molecular basis of FXTAS, a knockin in which alleles mispair during meiotic mouse model of FXTAS was generated crossing-over, resulting in one expanded using a human 98 CGG repeat, and was and one contracted tract. noted to produce elevated levels of the Fmr1 transcript and display intranuclear 6.2 ribonucleoprotein inclusions, containing Oculopharyngeal Muscular dystrophy ubiquitin, Hsp40 and the 20S proteasome subunit. The inclusions may be a response Oculopharyngeal muscular dystrophy to RNA toxicity or a result of the aggre- (OPMD) is a predominantly autosomal gation of CGG-binding proteins. While dominant, late-onset disorder character- the mechanism of FXTAS may be due ized by progressive drooping eyelids, dys- to a dominant gain of function of triplet- phagia, and proximal limb weakness. Cer- expanded RNA on RNA-binding proteins tain skeletal muscles in affected patients as in DM, there is currently no direct evi- degenerate and contain nuclear inclusions dence supporting this theory. and rimmed vacuoles. This effect is partic- ularly striking in the levator palpebra and pharyngeal muscles, responsible for lift- 6 ing the eyelids and swallowing. OPMD is Type 4: The Polyalanine Diseases caused by an alanine expansion in the gene polyadenine-binding protein 2 (PABP2). 6.1 Overview PABP2 normally contains a 10-alanine tract encoded by (GCG)6(GCA)3GCG. Af- Polyalanine disorders are characterized by fected individuals carry 12–17 alanines, small expansions in trinucleotide repeats probably resulting from unequal crossing- encoding alanine tracts. Most are rare, over of the two alleles. Repeat length autosomal dominant developmental mal- appears to correlate with disease severity. formations. Although the normal function PABP2 is an abundant, ubiquitously ex- of the alanine tract is unknown, they tend pressed pre-mRNA-binding protein that to be found in transcription factors. The plays a role in controlling the formation majority of alanine tract expansions asso- and length of mRNA polyA tails. NIs ciated with disease loci are also located immunoreactive for PABP2 are present in transcription factors, most of which in the skeletal muscle cells of OPMD play a role in development. In contrast patients and sequester polyA-containing to polyglutamine expansions, polyalanine transcripts. The aggregates are filamen- expansions are small and stable, not tous and contain PABP2, ubiquitin and Triplet Repeat Diseases 47 proteasome subunits, leading investiga- as Congo red and doxycycline, reduces tors to conclude that expansion of the aggregate formation and toxicity in cell alanine tract probably causes PABP2 to culture. A polyalanine-expanded peptide, misfold and aggregate. Aggregation in similar to the amino-terminal region of OPMD has been linked to toxicity in PABP2 was shown to adopt a β-sheet con- several experimental systems. Polyalanine- formation, whereas the same peptide with expanded PABP2 aggregates and causes 7 alanines adopted an α-helical conforma- cell death in cultured cells and trans- tion. This is reminiscent of polyglutamine genic mice. Indeed, widespread expression proteins, which also adopt a β-sheet con- of the human PABP2 gene in trans- formationwhenthepolyglutaminetract genic mice yielded polyalanine length- exceeds a threshold of about 35 glu- dependent muscle pathology, including tamines. Expanded polyalanine proteins rimmed vacuoles, central nuclei, and also activate caspase-3 and -8 in cultured numerous dystrophic changes (Fig. 13). cells, another feature reminiscent of polyg- Deleting the C-terminal oligomerization lutamine toxicity. domain, overexpressing chaperones, or ex- In 2004, Wirtschafter et al. proposed a posing cells to aggregation inhibitors such model for the selective vulnerability of

(a) (c) (e)

(b) (d) (f) Fig. 13 Mice expressing polyalanine-expanded PABP2 protein display prominent muscle pathology consistent with the OPMD phenotype. Transgenic mice were generated using the pCAGGS expression vector that consists of the chicken beta-actin promoter and a CMV enhancer, and thus ubiquitously express either human PABP2 protein with six alanines (c, e) orninealanines(a,b,d,f)atroughlycomparable levels. Hematoxylin & eosin staining of sections 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 and prominent connective tissue in pharynx and eyelid musculature (d, f). At high power, cytoplasmic vacuoles, reminiscent of OPMD ‘‘rimmed vacuoles’’ are apparent in soleus muscle from the nine-alanine PABP2 expressing mice (a). Low power examination further indicates frequent central nuclei in the soleus muscle from such mice (b). (From Hino et al. (2004) Hum. Mol. Genet. 13, 181, used with permission). 48 Triplet Repeat Diseases

extraocular muscles in OPMD. Unlike support of this theory, humans with sus- other skeletal muscles, extraocular mus- pected loss-of-function mutations in the cles are not postmitotic and they contin- HOX-D13 gene do not have a phenotype ually undergo remodeling. This requires consistent with SPD. the frequent upregulation of genes in- volved in cell cycling and protein synthesis. 6.4 Failure of correct mRNA polyadenylation Cleidocranial Dysplasia or transport in these cells may have a cumulative toxic effect resulting in the pro- Cleidocranial dysplasia (CCD) is a rare, au- gressive degeneration of these muscles. tosomal dominant developmental disorder The mechanism of OPMD could be due characterized by holes in the skull, den- tal malformations, absent or hypoplastic to an interference with polyadenylation, clavicles and maxillae, and other skeletal disturbances in intracellular trafficking of malformations. The primary cause of CCD mRNA, or toxicity due to the aggregation is thought to be loss-of-function mutations of misfolded and/or aggregated PABP2 in the gene RUNX2. This gene affects os- species. Further research will be required teoblast differentiation and is a member to distinguish between these possibilities. of the Runt family of transcription fac- tors. In one family, phenotypically distinct 6.3 from classic CCD, an expansion from 17 Synpolydactyly (Syndactyly Type II) to 27 alanines in RUNX2 has been de- tected. This family exhibited brachydactyly Synpolydactyly (SPD) is a rare, autosomal and a mild CCD phenotype. The differ- dominant developmental disorder charac- ence in phenotype between the polyalanine terizedbyfusedandextradigits(syndactyly expansion mutation family and typical pre- and polydactyly, respectively). It is caused sumed haploinsufficient, loss-of-function by the expansion of a polyalanine tract in CCD patients supports a gain-of-function the amino-terminal region of the transcrip- effect of the expanded alanine tract in this tion factor HOX-D13. HOX genes act in atypical family. Contraction of the tract in concert to coordinate axial patterning in RUNX2 is common and does not cause a animals. Tracts of 7–14 alanines in HOX- detectable phenotype. D13 have been linked to SPD, and disease severity is proportional to repeat length. 6.5 Multiple studies support a ‘‘dominant- Holoprosencephaly negative’’ role for expanded HOX-D13 protein in SPD. Mice null for Hox-d13 Holoprosencephaly (HPE) is a common have a phenotype less severe than SPD, developmental malformation resulting in while mice with alanine tract expansions partial or full cyclopia, failure to develop have a form that more closely resembles midline structures in the ventral fore- the disorder. Mice lacking Hox11, Hox12, brain, and prenatal lethality. In rare cases, and Hox13 have a phenotype similar to HPE is caused by the heterozygous ex- SPD, suggesting that the alanine expan- pansion of a 15-amino acid alanine tract sion in SPD antagonizes the function of in the protein ZIC2. This protein is one other HOX genes. This is supported by a member of a family of zinc finger pro- genetic complementation study. In further teins believed to regulate neurulation, Triplet Repeat Diseases 49 left–right axis formation and other de- resulting in malformation of the upper velopmental processes. Other individuals eyelids and forehead, and occasionally pre- with heterozygous loss-of-function muta- mature ovarian failure in women. It is tions have a phenotype indistinguishable caused by a number of different muta- from those with alanine tract expansions. tions in the gene Forkhead L2 (Foxl2), the Partial loss of function of Zic2 in mice most common of which is expansion of similarly causes developmental abnormal- its carboxy-terminal 14-amino acid alanine ities similar to HPE. Therefore, expansion tract. FOXL2 is a highly conserved tran- of the alanine tract likely causes loss of scription factor whose role in the ovary function of ZIC2 in the case of HPE. has been studied most thoroughly. It is expressed in the ovaries and eyelids dur- 6.6 ing development and adulthood. Studies Hand-foot-genital Syndrome in which Foxl2 have been ablated in mice show that it is important for the Hand-foot-genital syndrome (HFGS) is a differentiation of ovarian granulosa cells, rare, dominantly inherited developmental and its absence causes accelerated follicle abnormality characterized by malforma- cell depletion leading to POF. Knockout tion of the distal limbs and lower urogen- mice also display craniofacial abnormali- ital tract. Short thumbs, short great toes, ties consistent with BPES, suggesting that and abnormal carpals and tarsals are some a dominant-negative mechanism may be of its salient features. Some HFGS patients responsible in humans. have alanine tract expansions in the pro- tein HOX-A13, which is in the same family 6.8 as the SPD-associated protein HOX-D13. Syndromic and Nonsyndromic X-linked HOX-A13 contains three alanine tracts, Mental Retardation the most C-terminal of which is the most commonly mutated. Expansions enlarge Several loosely related disorders are as- the second or third tract by 6 to 9 ala- sociated with alanine tract expansions in nines. In humans, deletion of HOX-A13 the Aristaless related homeobox (ARX) causes a phenotype that is mild in com- protein on chromosome Xp22.13. One parison to HFGS caused by alanine tract is nonsyndromic X-linked mental retarda- mutations. Hoxa13-null mice also have a tion (XLMR), a heterogeneous condition in milder phenotype than mice carrying a which mental retardation is the main con- frameshift deletion suspected to confer sistent feature. Several syndromic XLMR gain of function. This evidence suggests a disorders linked to alanine expansion dominant-negative mechanism for HFGS in ARX include West syndrome (WS) caused by alanine tract expansions. and Partington syndrome (PRTS). WS causes progressive mental retardation with 6.7 abnormal EEG and infantile seizures, Blepharophimosis-ptosis-epicanthus while PRTS causes mental retardation, Inversus Syndrome dysarthria, and dystonic movements of the hands. Brain anatomy appears normal in Blepharophimosis-ptosis-epicanthus in- these disorders. versus syndrome (BPES) is a rare, auto- ARX is a paired-class homeodomain somal dominant developmental disorder protein that is expressed in the ventricular 50 Triplet Repeat Diseases

and marginal zones of the developing transcription factor containing a home- mouse brain and continues to be expressed odomain. A loss-of-function mutation in in adult cortex. It is suspected to play a role murine Phox2b is homozygous lethal, in neuroepithelial cell differentiation and and specifically prevents the development maintenance of neuronal subtypes in the of parasympathetic ganglia. Heterozygous adult cortex. ARX-null mice have small mice show chronic pupil dilation but brains and neuronal migration deficits. no parasympathetic or respiratory distur- Expansions of two different alanine tracts bances. Also, a patient hemizygous for in ARX cause XLMR: a 12-alanine tract a 5-Mb deletion including Phox2b does at amino acids 144–155, and a second not have CCHS. No cases of CCHS have tract at amino acids 100–115. Both sites been reported in which PHOX2B is trun- cause highly variable forms of XLMR, with cated before the homeobox domain. This expansions in the former alanine tract evidence suggests that the expansion of resulting in nonsyndromic XLMR, WS or alanine tracts in PHOX2B may cause a PRTS. Alanine tract expansions in ARX subtle, dominant-negative effect on the de- probably result in partial loss of function of velopment of respiratory control pathways. the protein. Heterozygous female carriers of expanded ARX do not exhibit XLMR, suggesting that one normal copy of the 7 protein is sufficient for normal cognitive Unclassified Repeat Diseases Lacking ability. Also, humans with null mutations Mechanistic Explanations of ARX suffer from X-linked lissencephaly with abnormal genitalia (XLAG), a much 7.1 more severe condition causing major Spinocerebellar Ataxia Type 10 developmental abnormalities in the brain and genitalia. Thus, XLMR due to alanine Spinocerebellar ataxia type 10 (SCA10) expansions in ARX may be because of is an autosomal dominant, progressive partial loss of function leading to subtle ataxia that exhibits nearly pure cerebellar developmental defects and/or failure to signs. It appears restricted to individuals maintain specific neuronal populations. of Mexican ethnicity. SCA10 patients typ- ically present with gait ataxia, followed by 6.9 dysarthria, dysphagia, and ocular dysme- Congenital Central Hypoventilation tria, and most patients also experience re- Syndrome current motor seizures. Cerebellar atrophy is the most prominent neuropathological Congenital central hypoventilation syn- change. SCA10 is caused by the expansion drome (CCHS) is a rare, autosomal dom- of a highly polymorphic pentanucleotide inant disorder causing a failure of auto- repeat, (ATTCT)n, on chromosome 22. Un- nomic control of breathing. It attenuates affected individuals carry 10–22 repeats. or abolishes responses to hypercarbia and While expanded alleles have not been suc- hypoxemia. In the majority of cases, it is cessfully PCR amplified, transcript sizes caused by the expansion of one of two on Northern blots and fragment sizes on alanine tracts in the protein PHOX-2B. Southern blots indicate that disease alleles Mutations expand the 20-residue tract to can range from ∼800 to 4500 repeats. An- 25–29 alanines. PHOX-2B is a paired-class ticipation occurs in SCA10, independently Triplet Repeat Diseases 51 supporting the causal nature of the repeat 7.2 expansion. The ATTCT repeat is located in Spinocerebellar Ataxia Type 12 intron 9 of the SCA10 gene, a previously unrecognized 66-kb gene that encodes a Among the more recent additions to the novel putative 475 amino acid protein of unstable repeat disease group is spinocere- unknown function with few recognizable bellar ataxia type 12 (SCA12), a rare, auto- motifs or domains. The 2-kb SCA10 tran- somal dominant disorder that may be most script is ubiquitously expressed though the prevalent in Indian populations. Its symp- highest levels of expression occur in the toms are distinct from the other SCAs, brain, testis, and adrenal glands. Within typically beginning with an action tremor the brain, it is most highly expressed in the of the upper extremities and progressing to cerebellum and associated structures. The include hyperreflexia, mild cerebellar dys- carboxy-terminal portion of the protein function, bradykinesia, increased muscle tone, psychiatric symptoms, and demen- appears to contain ‘‘armadillo repeats,’’ tia. The brains of SCA12 patients likely which are responsible for the membrane undergo a slow, generalized atrophy that association of β-catenins. Ataxin-10 does is most prominent in the cortex, but also not seem to associate with membranes, results in loss of Purkinje cells in the cere- however. The entire protein is highly con- bellum. Disease onset typically occurs in served between humans and rodents, and the third or fourth decade, and a gradually potential orthologs exist in Arabidopsis progressive disease course is typical. and Drosophila. SCA12 is caused by a (CAG)n expansion There are several proposed mechanisms in chromosome 5q31 – q33. The expan- for SCA10 pathogenesis. The most obvi- sion is 5 to the PPP2R2B gene, encoding ous hypothesis is that the large intronic a regulatory subunit of the protein phos- expansion affects ataxin-10 expression, phatase 2A enzyme (PP2A). This gene perhaps by altering local chromatin struc- has many transcription start sites, some ture. siRNA knockdown of ataxin-10 in cell of which include the repeat (but many of culture experiments yields higher rates which do not). GENSCAN predicts an exon of cell death in cerebellar neurons than including the expansion that would encode in cortical neurons. However, the expres- a polyserine tract, but this prediction is of sion levels of the SCA10 transcript are not low probability and considered unlikely. reduced in patient’s lymphoblast cells, ar- Other evidence suggests that the expan- guing against simple haploinsufficiency. sion is located in the promoter region, Consequently, RNA gain-of-function tox- andthisiscurrentlythemostwidelyac- icity has been proposed as the potential cepted view. Unaffected individuals carry mechanism; however, there are no fur- 7–32 CAG repeats while affected individ- ther data at this time to support such a uals carry 55–78 CAG triplets. The most hypothesis. Additional work will need to common repeat size in unaffected individ- be done to distinguish between these and uals is 10 CAGs. The repeat is fairly stable, other possibilities, and will need to ac- with only modest expansions and contrac- count for the cell-type specific pattern of tions resulting equally from maternal and neurodegeneration that occurs in the face paternal transmission. A significant corre- of apparently widespread expression of the lation between repeat length and age of SCA10 gene mutation. onset has not been documented. 52 Triplet Repeat Diseases

The protein PP2A is an essential ser- has the potential to disturb a multi- ine/threonine phosphatase expressed in tude of processes in the CNS. At this all known eukaryotic cells. It is involved time, all of the above theories of SCA12 in diverse cellular functions, including cell CAG repeat expansion neurotoxicity re- growth, differentiation, DNA replication, main plausible. neurotransmitter release, and apoptosis. PPP2R2B is a brain-specific regulatory 7.3 subunit of PP2A. The class of regulatory Huntington’s Disease Like 2 (HDL2) subunits including PPP2R2B may affect PP2A’s phosphatase activity for certain Perhaps the most exciting and enigmatic substrates, including histone-1, vimentin, recent discovery in the repeat expansion and tau. It may also affect PP2A’s subcel- field is that of the mutational basis of lular localization. a disorder known as Huntington’s dis- There are several possible explanations ease like 2 or HDL2. HDL2 is so named for SCA12 pathogenesis. The first is becauseitisinessenceagenocopyof that expanded PPP2R2B may generate a classical HD, as the original HDL2 pedi- polyamino acid tract-containing protein, gree was labeled with a diagnosis of HD resulting in toxicity. Northern blots prob- until HD CAG repeat testing indicated ing for sequence flanking the CAG repeat that this family’s HD-like disease did not did not detect a PPP2R2B transcript, in- result from a CAG repeat expansion in dicating that if a repeat-containing exon the htt gene. HDL2 patients present with exists and is transcribed, it is not present weight loss and diminished coordination, at appreciable levels. Nevertheless, the and then develop tremors, dysarthria, hy- possibility of polyglutamine, polyserine, perreflexia, and rigidity. Patients display or polyalanine toxicity, though unlikely, psychiatric involvement, chorea, and dys- cannot be completely ruled out at this tonia, and ultimately become demented. time. A second possible cause of SCA12 Death occurs 15–25 years after onset, pathogenesis is RNA gain of function, as when patients become bedridden as in typ- occurs in DM and FXTAS. The repeat ical HD cases. MRI findings reveal marked expansions are smaller in SCA12 than atrophy of the caudate and of the cere- in DM, however, and more importantly, bral cortex, making the neuropathology CAG tract-containing transcripts appear indistinguishable from classic HD. HDL2 to be rather scarce. A third possibility patients do not show cerebellar signs or is altered splicing of the PPP2R2B tran- neuropathology. script. Several different amino-termini are As soon as it was found that HDL2 possible, the ratio of which may affect patients do not have the HD CAG re- PP2A’s subcellular localization. Yet an- peat expansion, direct cloning methods for other theory of SCA12 pathogenesis is triplet repeat expansions of the CAG/CTG that the expansion affects PPP2R2B tran- type were applied to patient samples and script levels. Repeat expansion causes a an expanded CAG/CTG repeat (n = 55) substantial increase in PPP2R2B expres- was isolated. Sequence flanking this repeat sion as measured in reporter assays using indicated that HDL2 is caused by CTG a neuroblastoma cell line. Altered lev- repeat expansions in the junctophilin-3 els of the protein could affect PP2A’s (JPH3) gene, one of a family of struc- specificity or subcellular localization. This tural proteins whose function is to link Triplet Repeat Diseases 53

Ca++ channels on the ER with voltage haploinsufficiency of the Ca++ regulating sensors on the plasma membrane. Anal- brain- and testes-specific junctophilin-3 ysis of JPH3 CTG repeat indicates that protein is responsible for HDL2. While normal individuals typically carry 7–27 JPH3 knockout mice display motor inco- CTG repeats while affected HDL2 patients ordination, no histological abnormalities usually have expansions of 50–60 CTG are found in their brains. Further work repeats. Widespread screening of HD-like with these mice is ongoing to evaluate patients from around the globe suggests this hypothesis. Evidence against simple that the HDL2 JPH3 CTG repeat expan- haploinsufficiency has come from study of sions are most common in individuals of HDL2 patient’s brain material, however. African ethnicity. 1C2 and ubiquitin antibody immunos- The question of how the JPH3 CTG taining reveal intranuclear inclusions in repeat expansion causes HDL2 remains neurons throughout the brains of these unknown, but there are at least four possi- patients, with dramatic similarity in dis- ble explanations. These alternative (but not tribution to patients with classic HD mutually exclusive) theories stem from the (Fig. 14). At this time, the molecular ba- documented alternative processing of the sisofHDL2isunknown;however,the JPH3 gene, which permits the prediction incredible overlap between HD and HDL2 of the CTG repeat tract as: (1) part of intron in terms of clinical phenotype and neu- 1; (2) as part of the 3 untranslated region; ropathology strongly suggests that solving or (3) as encoding either a polyleucine HDL2 should have profound implications or polyalanine tract. One theory is that for our mechanistic understanding of HD.

Fig. 14 HDL2 patients have 1C2-positive intranuclear inclusions. 1C2 staining of cerebral cortex (frontal lobe) from an HDL2 patient reveals a prominent intranuclear inclusion that resembles the intranuclear inclusions seen in classic HD patients. As the 1C2 antibody is directed against expanded, misfolded polyglutamine tracts, which arenotpredictedtobeexpressedfrom the causal HDL2 gene, the explanation for the presence of such nuclear inclusions in HDL2 patients remains unknown. (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 to Huntington’s disease is associated with a novel CAG repeat expansion, Ann. Neurol. 50, 373–380, used with permission of Annals of Neurology,and John Wiley & Sons, publisher). 54 Triplet Repeat Diseases See also Genetics, Molecular Ba- and the risk of neuronal death, Nature. 431, 805–810. sis of; Motor Neuron Diseases: Banfi, S., Servadio, A., Chung, M., Capozzoli, F., Cellular and Animal Models; Mo- Duvick, L.A., Elde, R., Zoghbi, H.Y., Orr, H.T. (1996) Cloning and developmental expression tor Neuron Diseases: Molecular analysis of the murine homolog of the Mechanism, Pathophysiology, and spinocerebellar ataxia type 1 gene (Sca1), Hum. Mol. Genet. 5, 33–40. Treatments; Noncoding Tandemly Brais, B., Bouchard, J.P., Xie, Y.G., Rochefort, Repeated DNA Sequences. D.L., Chretien, N., Tome, F.M., Lafre- niere, R.G., Rommens, J.M., Uyama, E., No- hira, O. et al. (1998) Short GCG expansions in the PABP2 gene cause oculopharyngeal Bibliography muscular dystrophy, Nat. Genet. 18, 164–167. Brook, J.D., McCurrach, M.E., Harley, H.G., Books and Reviews Buckler, A.J., Church, D., Aburatani, H., Hunter, K., Stanton, V.P., Thirion, J.P., Hud- Brown, L.Y., Brown, S.A. (2004) Alanine tracts: son, T. et al. (1992) Molecular basis of my- otonic dystrophy: expansion of a trinucleotide the expanding story of human illness and trinucleotide repeats, Trends Genet. 20, 51–58. (CTG) repeat at the 3 end of a transcript en- Harper, P.S. (2001) Myotonic Dystrophy, coding a protein kinase family member, Cell. Saunders, London, UK. 69, 385. Harper, P.S., Perutz, M. (2001) Glutamine Brown, V., Jin, P., Ceman, S., Darnell, J.C., Repeats and Neurodegenerative Diseases: O’Donnell, W.T., Tenenbaum, S.A., Jin, X., Molecular Aspects, Oxford University Press, Feng, Y., Wilkinson, K.D., Keene, J.D. et al. New York. (2001) Microarray identification of FMRP- Jin, P., Alisch, R.S., Warren, S.T. (2004) RNA associated brain mRNAs and altered mRNA and microRNAs in fragile X mental translational profiles in fragile X syndrome, Cell. 107, 477–487. retardation, Nat. Cell Biol. 6, 1048–1053. Brown, L.Y., Odent, S., David, V., Blayau, M., Ranum, L.P., Day, J.W. (2004) Pathogenic RNA Dubourg, C., Apacik, C., Delgado, M.A., repeats: an expanding role in genetic disease, Hall, B.D., Reynolds, J.F., Sommer, A. et al. Trends Genet. 20, 506–512. (2001) Holoprosencephaly due to mutations in Wells, R.D., Warren, S.T. (1998) Genetic ZIC2: alanine tract expansion mutations may Instabilities and Hereditary Neurological be caused by parental somatic recombination, Diseases, Academic Press, San Diego, CA. Hum. Mol. Genet. 10, 791–796. Burnett, B., Li, F., Pittman, R.N. (2003) The Primary Literature polyglutamine neurodegenerative protein ataxin-3 binds polyubiquitynated proteins and Albrecht, M., Golatta, M., Wullner, U., has ubiquitin protease activity, Hum. Mol. Lengauer, T. (2004) Structural and functional Genet. 12, 3195–3205. analysis of ataxin-2 and ataxin-3. Eur. J. Calado, A., Tome, F.M., Brais, B., Rouleau, G.A., Biochem. 271, 3155–3170. Kuhn, U., Wahle, E., Carmo-Fonseca, M. Amiel, J., Laudier, B., Attie-Bitach, T., Trang, H., (2000) Nuclear inclusions in oculopharyngeal de Pontual, L., Gener, B., Trochet, D., muscular dystrophy consist of poly(A) binding Etchevers, H., Ray, P., Simonneau, M. et al. protein 2 aggregates which sequester (2003) Polyalanine expansion and frameshift poly(A) RNA, Hum. Mol. Genet. 9, 2321–2328. mutations of the paired-like homeobox gene Chen, S., Peng, G.H., Wang, X., Smith, A.C., PHOX2B in congenital central hypoventilation Grote, S.K., Sopher, B.L., La Spada, A.R. syndrome, Nat. Genet. 33, 459–461. (2004) Interference of Crx-dependent Arrasate, M., Mitra, S., Schweitzer, E.S., Se- transcription by ataxin-7 involves interaction gal, M.R., Finkbeiner, S. (2004) Inclusion body between the glutamine regions and requires formation reduces levels of mutant huntingtin the ataxin-7 carboxy-terminal region for Triplet Repeat Diseases 55

nuclear localization, Hum. Mol. Genet. 13, Trifiro, M.A., Pinsky, L., Wellington, C.L., 53–67. Salvesen, G.S. et al. (1999) Kennedy’s disease: Chen, H.K., Fernandez-Funez, P., Acevedo, S.F., caspase cleavage of the androgen receptor is a Lam, Y.C., Kaytor, M.D., Fernandez, M.H., crucial event in cytotoxicity, J. Neurochem. 72, Aitken, A., Skoulakis, E.M., Orr, H.T., 185–195. Botas, J. et al. (2003) Interaction of Akt- Fu, Y.H., Kuhl, D.P., Pizzuti, A., Pieretti, M., phosphorylated ataxin-1 with 14-3-3 mediates Sutcliffe, J.S., Richards, S., Verkerk, A.J., neurodegeneration in spinocerebellar ataxia Holden, J.J., Fenwick, R.G., Jr., Warren, S.T., type 1, Cell. 113, 457–468. et al. (1991) Variation of the CGG repeat at Cummings, C.J., Reinstein, E., Sun, Y., An- the fragile X site results in genetic instability: talffy, B., Jiang, Y., Ciechanover, A., Orr, H.T., resolution of the Sherman paradox, Cell. 67, Beaudet, A.L., Zoghbi, H.Y. (1999) Mutation 1047–1058. of the E6-AP ubiquitin ligase reduces nu- Garden, G.A., Libby, R.T., Fu, Y.H., Kino- clear inclusion frequency while accelerating shita, Y., Huang, J., Possin, D.E., Smith, polyglutamine-induced pathology in SCA1 A.C., Martinez, R.A., Fine, G.C., Grote, S.K., mice, Neuron. 24, 879–892. et al. (2002) Polyglutamine-expanded ataxin-7 David, G., Abbas, N., Stevanin, G., Durr, A., promotes noncell-autonomous Purkinje cell Yvert, G., Cancel, G., Weber, C., Imbert, G., degeneration and displays proteolytic cleavage Saudou, F., Antoniou, E. et al. (1997) Cloning in ataxic transgenic mice, J. Neurosci. 22, of the SCA7 gene reveals a highly unstable 4897–4905. CAG repeat expansion, Nat. Genet. 17, 65–70. Gecz, J. (2000) The FMR2 gene, FRAXE and non- Davies, S.W., Turmaine, M., Cozens, B.A., specific X-linked mental retardation: clinical DiFiglia, M., Sharp, A.H., Ross, C.A., and molecular aspects, Ann. Hum. Genet. 64, Scherzinger, E., Wanker, E.E., Mangiarini, L., 95–106. Bates, G.P. (1997) Formation of neuronal Goodman, F.R., Scambler, P.J. (2001) Human intranuclear inclusions underlies the HOX gene mutations, Clin. Genet. 59, 1–11. neurological dysfunction in mice transgenic Grabczyk, E., Usdin, K. (2000) The GAA*TTC for the HD mutation, Cell. 90, 537–548. triplet repeat expanded in Friedreich’s ataxia Day,J.W.,Ricker,K.,Jacobsen,J.F.,Ras- impedes transcription elongation by T7 mussen, L.J., Dick, K.A., Kress, W., Schnei- RNA polymerase in a length and supercoil der, C., Koch, M.C., Beilman, G.J., Harri- dependent manner, Nucleic Acids Res. 28, son, A.R. et al. (2003) Myotonic dystrophy type 2815–2822. 2: molecular, diagnostic and clinical spectrum, Greene, E., Handa, V., Kumari, D., Usdin, K. Neurology. 60, 657–664. (2003) Transcription defects induced by repeat De Baere, E., Beysen, D., Oley, C., Lorenz, B., expansion: fragile X syndrome, FRAXE mental Cocquet, J., De Sutter, P., Devriendt, K., retardation, progressive myoclonus epilepsy Dixon, M., Fellous, M., Fryns, J.P. et al. (2003) type 1, and Friedreich ataxia, Cytogenet. FOXL2 and BPES: mutational hotspots, Genome Res. 100, 65–76. phenotypic variability, and revision of the Hagerman, R.J., Leehey, M., Heinrichs, W., genotype-phenotype correlation, Am.J.Hum. Tassone, F., Wilson, R., Hills, J., Grigsby, J., Genet. 72, 478–487. Gage, B., Hagerman, P.J. (2001) Intention DiFiglia, M., Sapp, E., Chase, K.O., Davies, S.W., tremor, parkinsonism, and generalized brain Bates, G.P., Vonsattel, J.P., Aronin, N. (1997) atrophy in male carriers of fragile X, Neurology. Aggregation of huntingtin in neuronal 57, 127–130. intranuclear inclusions and dystrophic Helmlinger, D., Hardy, S., Sasorith, S., Klein, F., neurites in brain, Science. 277, 1990–1993. Robert, F., Weber, C., Miguet, L., Potier, N., Dunah, A.W., Jeong, H., Griffin, A., Kim, Y.M., Van-Dorsselaer, A., Wurtz, J.M., et al. (2004) Standaert, D.G., Hersch, S.M., Mouradian, Ataxin-7 is a subunit of GCN5 histone M.M., Young, A.B., Tanese, N., Krainc, D. acetyltransferase-containing complexes, Hum. (2002) Sp1 and TAFII130 transcriptional Mol. Genet. 13, 1257–1265. activity disrupted in early Huntington’s Hillman, M.A., Gecz, J. (2001) Fragile XE- disease, Science. 296, 2238–2243. associated familial mental retardation protein Ellerby, L.M., Hackam, A.S., Propp, S.S., Ellerby, 2 (FMR2) acts as a potent transcription H.M., Rabizadeh, S., Cashman, N.R., activator, J. Hum. Genet. 46, 251–259. 56 Triplet Repeat Diseases

Hodgson, J.G., Agopyan, N., Gutekunst, C.A., Kitamura, K., Yanazawa, M., Sugiyama, N., Leavitt, B.R., LePiane, F., Singaraja, R., Smith, Miura, H., Iizuka-Kogo, A., Kusaka, M., D.J., Bissada, N., McCutcheon, K., Nasir, J., Omichi, K., Suzuki, R., Kato-Fukui, Y., Kami- et al. (1999) A YAC mouse model for irisa, K., et al. (2002) Mutation of ARX causes Huntington’s disease with full-length mutant abnormal development of forebrain and testes huntingtin, cytoplasmic toxicity, and selective in mice and X-linked lissencephaly with ab- striatal neurodegeneration, Neuron. 23, normal genitalia in humans, Nat. Genet. 32, 181–192. 359–369. Holmes, S.E., O’Hearn, E., McInnis, M.G., Klement,I.A.,Skinner,P.J.,Kaytor,M.D.,Yi,H., Gorelick-Feldman, D.A., Kleiderlein, J.J., Hersch, S.M., Clark, H.B., Zoghbi, H.Y., Callahan, C.A., Ingersoll-Ashworth, R.G., Orr, H.T. (1998) Ataxin-1 nuclear localiza- Sherr, M., Sumner, A.J. (1999) Expansion of tion and aggregation: role in polyglutamine- a novel CAG repeat in the 5 region of a gene induced disease in SCA1 transgenic mice, Cell. encoding a subunit of protein phosphatase 2A 95, 41–53. is associated with spinocerebellar ataxia type Koob, M.D., Moseley, M.L., Schut, L.J., Ben- 12, Am.J.Hum.Genet.65,A4. zow,K.A.,Bird,T.D.,Day,J.W.,Ranum,L.P. Holmes, S.E., O’Hearn, E., Rosenblatt, A., Calla- (1999) An untranslated CTG expansion causes han, C., Hwang, H.S., Ingersoll-Ashworth, a novel form of spinocerebellar ataxia (SCA8), R.G., Fleisher, A., Stevanin, G., Brice, A., Pot- Nat. Genet. 21, 379–384. ter, N.T., et al. (2001) A repeat expansion in Kremer, E.J., Pritchard, M., Lynch, M., Yu, S., the gene encoding junctophilin-3 is associated Holman, K., Baker, E., Warren, S.T., Sch- lessinger, D., Sutherland, G.R., Richards, R.I. with Huntington disease-like 2, Nat. Genet. 29, (1991) Mapping of DNA instability at the 377–378. fragile X to a trinucleotide repeat sequence Huntington’s Disease, Collaborative, Research p(CCG)n, Science. 252, 1711–1714. and Group (1993) A novel gene containing La Spada, A.R., Wilson, E.M., Lubahn, D.B., a trinucleotide repeat that is expanded Harding, A.E., Fischbeck, K.H. (1991) Andro- and unstable on Huntington’s disease gen receptor gene mutations in X-linked spinal chromosomes, Cell 72, 971–983. and bulbar muscular atrophy, Nature. 352, Huynh, D.P., Figueroa, K., Hoang, N., Pulst, 77–79. S.M. (2000) Nuclear localization or inclusion La Spada, A.R., Roling, D.B., Harding, A.E., body formation of ataxin-2 are not necessary Warner, C.L., Spiegel, R., Hausmanowa for SCA2 pathogenesis in mouse or human, Petrusewicz, I., Yee, W.C., Fischbeck, K.H. Nat. Genet. 26, 44–50. (1992) Meiotic stability and genotype- Ishikawa, K., Fujigasaki, H., Saegusa, H., phenotype correlation of the trinucleotide Ohwada, K., Fujita, T., Iwamoto, H., Komat- repeat in X-linked spinal and bulbar muscular suzaki,Y., Toru,S., Toriyama,H., Watan- atrophy, Nat. Genet. 2, 301–304. abe, M., et al. (1999) Abundant expression and LaSpada,A.R.,Fu,Y.,Sopher,B.L.,Libby,R.T., cytoplasmic aggregations of [alpha]1A voltage- Wang, X., Li, L.Y., Einum, D.D., Huang, J., dependent calcium channel protein associated Possin, D.E., Smith, A.C., et al. (2001) with neurodegeneration in spinocerebellar Polyglutamine-expanded ataxin-7 antagonizes ataxia type 6, Hum. Mol. Genet. 8, 1185–1193. CRX function and induces cone-rod dystrophy Jiang, H., Mankodi, A., Swanson, M.S., Mox- in a mouse model of SCA7, Neuron. 31, ley, R.T., Thornton, C.A. (2004) Myotonic dys- 913–927. trophy type 1 is associated with nuclear foci Lafreniere, R.G., Rochefort, D.L., Chretien, N., of mutant RNA, sequestration of muscleblind Rommens, J.M., Cochius, J.I., Kalviainen, R., proteins and deregulated alternative splicing Nousiainen, U., Patry, G., Farrell, K., Soder- in neurons, Hum. Mol. Genet. 13, 3079–3088. feldt, B. et al. (1997) Unstable insertion in the Katsuno, M., Adachi, H., Kume, A., Li, M., Nak- 5 flanking region of the cystatin B gene is the agomi, Y., Niwa, H., Sang, C., Kobayashi, Y., most common mutation in progressive my- Doyu, M., Sobue, G. (2002) Testosterone re- oclonus epilepsy type 1, EPM1, Nat. Genet. 15, duction prevents phenotypic expression in a 298–302. transgenic mouse model of spinal and bulbar Lalioti, M.D., Antonarakis, S.E., Scott, H.S. muscular atrophy, Neuron. 35, 843–854. (2003) The epilepsy, the protease inhibitor Triplet Repeat Diseases 57

and the dodecamer: progressive myoclonus (2004) Somatic and germline instability of the epilepsy, cystatin b and a 12-mer repeat ATTCT repeat in spinocerebellar ataxia type expansion, Cytogenet. Genome Res. 100, 10, Am.J.Hum.Genet.74, 1216–1224. 213–223. Matsuura, T., Yamagata, T., Burgess, D.L., Lalioti, M.D., Scott, H.S., Antonarakis, S.E. Rasmussen, A., Grewal, R.P., Watase, K., (1999) Altered spacing of promoter elements Khajavi, M., McCall, A.E., Davis, C.F., Zu, L., due to the dodecamer repeat expansion et al. (2000) Large expansion of the ATTCT contributes to reduced expression of the pentanucleotide repeat in spinocerebellar cystatin B gene in EPM1, Hum. Mol. Genet. ataxia type 10, Nat. Genet. 26, 191–194. 8, 1791–1798. McCampbell, A., Taylor, J.P., Taye, A.A., Ro- Lin, X., Antalffy, B., Kang, D., Orr, H.T., bitschek,J., Li,M., Walcott,J., Merry,D., Zoghbi, H.Y. (2000) Polyglutamine expansion Chai, Y., Paulson, H., Sobue, G., et al. (2000) down-regulates specific neuronal genes before CREB-binding protein sequestration by ex- pathologic changes in SCA1, Nat. Neurosci. 3, panded polyglutamine, Hum. Mol. Genet. 9, 157–163. 2197–2202. Liquori, C.L., Ricker, K., Moseley, M.L., Jacob- Monckton, D.G., Cayuela, M.L., Gould, F.K., sen, J.F., Kress, W., Naylor, S.L., Day, J.W., Brock, G.J., Silva, R., Ashizawa, T. (1999) Very Ranum, L.P. (2001) Myotonic dystrophy type large (CAG)(n) DNA repeat expansions in the 2 caused by a CCTG expansion in intron 1 of sperm of two spinocerebellar ataxia type 7 ZNF9, Science. 293, 864–867. males, Hum. Mol. Genet. 8, 2473–2478. Loesch, D.Z., Huggins, R.M., Bui, Q.M., Ep- Mortlock, D.P., Innis, J.W. (1997) Mutation of stein, J.L., Taylor, A.K., Hagerman, R.J. (2002) HOXA13 in hand-foot-genital syndrome, Nat. Effect of the deficits of fragile X mental retar- Genet. 15, 179–180. dation protein on cognitive status of fragile x Mosemiller, A.K., Dalton, J.C., Day, J.W., males and females assessed by robust pedigree Ranum, L.P. (2003) Molecular genetics of analysis, J. Dev. Behav. Pediatr. 23, 416–423. spinocerebellar ataxia type 8 (SCA8), Cytogenet. Lunkes, A., Lindenberg, K.S., Ben-Haiem, L., Genome Res. 100, 175–183. Weber, C., Devys, D., Landwehrmeyer, G.B., Mundlos, S., Otto, F., Mundlos, C., Mul- Mandel, J.L., Trottier, Y. (2002) Proteases liken, J.B., Aylsworth, A.S., Albright, S., Lind- acting on mutant huntingtin generate hout, D., Cole, W.G., Henn, W., Knoll, J.H., cleaved products that differentially build up et al. (1997) Mutations involving the tran- cytoplasmic and nuclear inclusions, Mol. Cell. scription factor CBFA1 cause cleidocranial 10, 259–269. dysplasia, Cell. 89, 773–779. Mangiarini, L., Sathasivam, K., Seller, M., Coz- Muragaki, Y., Mundlos, S., Upton, J., Olsen, B.R. ens, B., Harper, A., Hetherington, C., Law- (1996) Altered growth and branching pat- ton, M., Trottier, Y., Lehrach, H., Davies, S.W., terns in synpolydactyly caused by mutations et al. (1996) Exon 1 of the HD gene with an in HOXD13, Science. 272, 548–551. expanded CAG repeat is sufficient to cause a Mutsuddi, M., Marshall, C.M., Benzow, K.A., progressive neurological phenotype in trans- Koob, M.D., Rebay, I. (2004) The spinocere- genic mice, Cell. 87, 493–506. bellar ataxia 8 noncoding RNA causes neu- Mankodi, A., Logigian, E., Callahan, L., Mc- rodegeneration and associates with staufen in Clain, C., White, R., Henderson, D., Krym, M., Drosophila, Curr. Biol. 14, 302–308. Thornton, C.A. (2000) Myotonic dystrophy in Nagafuchi, S., Yanagisawa, H., Sato, K., Shi- transgenic mice expressing an expanded CUG rayama,T.,Ohsaki,E.,Bundo,M.,Takeda,T., repeat, Science. 289, 1769–1773. Tadokoro, K., Kondo, I., Murayama, N., et al. Margolis, R.L., O’Hearn, E., Rosenblatt, A., (1994) Dentatorubral and pallidoluysian at- Willour, V., Holmes, S.E., Franz, M.L., Calla- rophy expansion of an unstable CAG trinu- han, C., Hwang, H.S., Troncoso, J.C., Ross, cleotide on chromosome 12p, Nat. Genet. 6, C.A. (2001) A disorder similar to Hunting- 14–18. ton’s disease is associated with a novel CAG Nakamura, K., Jeong, S.Y., Uchihara, T., repeat expansion, Ann. Neurol. 50, 373–380. Anno, M., Nagashima, K., Nagashima, T., Matsuura, T., Fang, P., Lin, X., Khajavi, M., Ikeda, S., Tsuji, S., Kanazawa, I. (2001) Tsuji,K.,Rasmussen,A.,Grewal,R.P., SCA17, a novel autosomal dominant cerebellar Achari, M., Alonso, M.E., Pulst, S.M., et al. ataxia caused by an expanded polyglutamine 58 Triplet Repeat Diseases

in TATA-binding protein, Hum. Mol. Genet. Ishida, Y., Ikeuchi, T., et al. (1996) Identifica- 10, 1441–1448. tion of the spinocerebellar ataxia type 2 gene Okazawa, H., Rich, T., Chang, A., Lin, X., using a direct identification of repeat expan- Waragai, M., Kajikawa, M., Enokido, Y., sion and cloning technique, DIRECT, Nat. Komuro, A., Kato, S., Shibata, M., et al. (2002) Genet. 14, 277–284. Interaction between mutant ataxin-1 and Satterfield, T.F., Jackson, S.M., Pallanck, L.J. PQBP-1 affects transcription and cell death, (2002) A Drosophila homolog of the Neuron. 34, 701–713. polyglutamine disease gene SCA2 is a dosage- Oostra, B.A., Willemsen, R. (2003) A fragile sensitive regulator of actin filament formation, balance: FMR1 expression levels, Hum. Mol. Genetics. 162, 1687–1702. Genet. 12, Spec No 2, R249–R257. Scherzinger, E., Lurz, R., Turmaine, M., Man- Orr, H.T., Chung, M.Y., Banfi, S., Kwiatkowski, giarini, L., Hollenbach, B., Hasenbank, R., Jr., T.J., Servadio, A., Beaudet, A.L., Mc- Bates, G.P., Davies, S.W., Lehrach, H., Call, A.E., Duvick, L.A., Ranum, L.P., Zoghbi, Wanker, E.E. (1997) Huntingtin-encoded H.Y. (1993) Expansion of an unstable trinu- polyglutamine expansions form amyloid-like cleotide CAG repeat in spinocerebellar ataxia protein aggregates in vitro and in vivo, Cell. 90, type 1, Nat. Genet. 4, 221–226. 549–558. Paulson, H.L., Perez, M.K., Trottier, Y., Tro- Schilling, G., Wood, J.D., Duan, K., Slunt, H.H., janowski, J.Q., Subramony, S.H., Das, S.S., Gonzales, V., Yamada, M., Cooper, J.K., Vig, P., Mandel, J.L., Fischbeck, K.H., Pitt- Margolis, R.L., Jenkins, N.A., Copeland, N.G., man, R.N. (1997) Intranuclear inclusions of et al. (1999) Nuclear accumulation of truncated expanded polyglutamine protein in spinocere- atrophin-1 fragments in a transgenic mouse bellar ataxia type 3, Neuron. 19, 333–344. model of DRPLA, Neuron. 24, 275–286. Penrose, L.S. (1948) The problem of anticipation Seznec, H., Simon, D., Monassier, L., Criqui- in pedigrees of dystrophica myotonica, Ann. Filipe, P., Gansmuller, A., Rustin, P., Eugen. 14, 125–132. Koenig, M., Puccio, H. (2004) Idebenone Perutz, M.F., Johnson, T., Suzuki, M., Finch, J.T. delays the onset of cardiac functional alteration (1994) Glutamine repeats as polar zippers: without correction of Fe-S enzymes deficit in a their possible role in inherited neurodegener- mouse model for Friedreich ataxia, Hum. Mol. ative diseases, Proc. Natl. Acad. Sci. U. S. A. Genet. 13, 1017–1024. 91, 5355–5358. Sopher, B.L., Thomas, P.S., Jr., LaFevre-Bernt, Poirier, M.A., Li, H., Macosko, J., Cai, S., Amzel, M., Ross, C.A. (2002) Huntingtin M.A., Holm, I.E., Wilke, S.A., Ware, C.B., Jin, spheroidsandprotofibrilsasprecursorsin L.W., Libby, R.T., Ellerby, L.M., La Spada, A.R. polyglutamine fibrilization, J. Biol. Chem. 277, (2004) Androgen receptor YAC transgenic 41032–41037. mice recapitulate SBMA motor neuronopathy Pulst, S.M., Nechiporuk, A., Nechiporuk, T., and implicate VEGF164 in the motor neuron Gispert, S., Chen, X.N., Lopes-Cendes, I., degeneration, Neuron. 41, 687–699. Pearlman, S., Starkman, S., Orozco-Diaz, G., Stevanin, G., Camuzat, A., Holmes, S.E., Lunkes, A., et al. (1996) Moderate expansion Julien, C., Sahloul, R., Dode, C., Hahn- of a normally biallelic trinucleotide repeat in Barma, V., Ross, C.A., Margolis, R.L., spinocerebellar ataxia type 2, Nat. Genet. 14, Durr, A., et al. (2002) CAG/CTG repeat 269–276. expansions at the Huntington’s disease-like Rasmussen, A., Matsuura, T., Ruano, L., 2 locus are rare in Huntington’s disease Yescas, P., Ochoa, A., Ashizawa, T., Alonso, E. patients, Neurology. 58, 965–967. (2001) Clinical and genetic analysis of four Stromme, P., Mangelsdorf, M.E., Shaw, M.A., Mexican families with spinocerebellar ataxia Lower, K.M., Lewis, S.M., Bruyere, H., type 10, Ann. Neurol. 50, 234–239. Lutcherath, V., Gedeon, A.K., Wallace, R.H., Sanchez, I., Mahlke, C., Yuan, J. (2003) Pivotal Scheffer, I.E., et al. (2002) Mutations in the role of oligomerization in expanded human ortholog of Aristaless cause X-linked polyglutamine neurodegenerative disorders, mental retardation and epilepsy, Nat. Genet. Nature. 421, 373–379. 30, 441–445. Sanpei, K., Takano, H., Igarashi, S., Sato, T., Oy- Sutherland, G.R., Baker, E. (1992) Characterisa- ake, M., Sasaki, H., Wakisaka, A., Tashiro, K., tion of a new rare fragile site easily confused Triplet Repeat Diseases 59

with the fragile X, Hum. Mol. Genet. 1, Tassone, F., Hoogeveen, A.T., et al. (2003) The 111–113. FMR1 CGG repeat mouse displays ubiquitin- Uda, M., Ottolenghi, C., Crisponi, L., Garcia, positive intranuclear neuronal inclusions; J.E., Deiana, M., Kimber, W., Forabosco, A., implications for the cerebellar tremor/ataxia Cao, A., Schlessinger, D., Pilia, G. (2004) syndrome, Hum. Mol. Genet. 12, 949–959. Foxl2 disruption causes mouse ovarian failure Wirtschafter, J.D., Ferrington, D.A., McLoon, by pervasive blockage of follicle development, L.K. (2004) Continuous remodeling of Hum. Mol. Genet. 13, 1171–1181. adult extraocular muscles as an explanation Ueno, S., Kondoh, K., Kotani, Y., Komure, O., for selective craniofacial vulnerability in Kuno, S., Kawai, J., Hazama, F., Sano, A. oculopharyngeal muscular dystrophy, J. (1995) Somatic mosaicism of CAG Neuroophthalmol. 24, 62–67. repeat in dentatorubral-pallidoluysian atrophy Yoo, S.Y., Pennesi, M.E., Weeber, E.J., Xu, B., (DRPLA), Hum. Mol. Genet. 4, 663–666. Atkinson,R., Chen,S., Armstrong,D.L., Wallis, J., Williamson, R., Chamberlain, S. Wu, S.M., Sweatt, J.D., Zoghbi, H.Y. (2003) (1990) Identification of a hypervariable SCA7 knockin mice model human SCA7 and microsatellite polymorphism within D9S15 reveal gradual accumulation of mutant ataxin- tightly linked to Friedreich’s ataxia, Hum. 7 in neurons and abnormalities in short-term Genet. 85, 98–100. plasticity, Neuron. 37, 383–401. Warren, S.T. (1997) Polyalanine expansion in Yvert, G., Lindenberg, K.S., Picaud, S., Landwe- synpolydactyly might result from unequal hrmeyer, G.B., Sahel, J.A., Mandel, J.L. (2000) crossing-over of HOXD13, Science. 275, Expanded polyglutamines induce neurode- 408–409. generation and trans-neuronal alterations in Watase, K., Weeber, E.J., Xu, B., Antalffy, B., cerebellum and retina of SCA7 transgenic Yuva-Paylor, L., Hashimoto, K., Kano, M., mice, Hum. Mol. Genet. 9, 2491–2506. Atkinson, R., Sun, Y., Armstrong, D.L., et al. Zakany, J., Duboule, D. (1996) Synpolydactyly in (2002) A long CAG repeat in the mouse Sca1 mice with a targeted deficiency in the HoxD locus replicates SCA1 features and reveals complex, Nature. 384, 69–71. the impact of protein solubility on selective Zhang, S., Xu, L., Lee, J., Xu, T. (2002) neurodegeneration, Neuron. 34, 905–919. Drosophila atrophin homolog functions Wellington, C.L., Ellerby, L.M., Gutekunst, C.A., as a transcriptional corepressor in Rogers, D., Warby, S., Graham, R.K., Loub- multiple developmental processes, Cell 108, ser, O., van Raamsdonk, J., Singaraja, R., 45–46. Yang, Y.Z., et al. (2002) Caspase cleavage of Zhuchenko, O., Bailey, J., Bonnen, P., Ashizawa, mutant huntingtin precedes neurodegenera- T., Stockton, D.W., Amos, C., Dobyns, W.B., tion in Huntington’s disease, J. Neurosci. 22, Subramony, S.H., Zoghbi, H.Y., Lee, C.C. 7862–7872. (1997) Autosomal dominant cerebellar ataxia Wellington, C.L., Ellerby, L.M., Hackam, A.S., (SCA6) associated with small polyglutamine Margolis, R.L., Trifiro, M.A., Singaraja, R., expansions in the alpha 1A-voltage- McCutcheon, K., Salvesen, G.S., Propp, S.S., dependent calcium channel, Nat. Genet. 15, Bromm, M., et al. (1998) Caspase cleavage 62–69. of gene products associated with triplet Zuccato, C., Tartari, M., Crotti, A., Goffredo, D., expansion disorders generates truncated Valenza, M., Conti, L., Cataudella, T., Leav- fragments containing the polyglutamine tract, itt, B.R., Hayden, M.R., Timmusk, T., et al. J. Biol. Chem. 273, 9158–9167. (2003) Huntingtin interacts with REST/NRSF Willemsen, R., Hoogeveen-Westerveld, M., Reis, to modulate the transcription of NRSE- S., Holstege, J., Severijnen, L.A., Nieuwen- controlled neuronal genes, Nat. Genet. 35, huizen, I.M., Schrier, M., van Unen, L., 76–83. 60 Triplet Repeat Diseases Tumor Immunology: see Immunology of Cancer

Tumorigenesis, Epigenetic Mechanisms: see Epigenetic Mechanisms in Tumorigenesis