Part I: Myotonic dystrophy

Table des matières

Myotonic disorders are diseases in which symptomatic individuals have difficulty in relaxing a muscle group after contraction. One hallmark feature of these diseases is the difficulty of affected individuals of releasing grip on objects like tools, cups and door knobs for example (Harper, 2001). In 1876, came one of the first documented accounts of a myotonic disorder from Dr Julius Thomsen by studying his most famous patient, himself. The disease he described, known today as Thomsen’s myotonia congenita , was of hereditary nature, caused myotonia and muscle hypertrophy but was not progressive in nature. Since no muscle weakness is observed in this disease, myotonia congenita is not considered to be a muscular dystrophy. Some thirty years later, in the early 20th century, another myotonic disorder was described with much more pronounced and degenerative symptoms. Myotonic dystrophy or dystrophia myotonica was first described by a German physician by the name of Hans Steinert in 1904 (Harper, 2001). He was the first physician to publish a detailed account of a new myotonic disorder in which he noted symptoms such as muscle weakness and wasting, facial weakness, ptosis, selective atrophy of the sternomastoids in the neck and testicular atrophy which underlined an endocrine disturbance. Steinert also conducted pathological studies on an autopsied case where he noted extensive fibrosis and degeneration of skeletal muscle. While Steinert died in 1910 shortly after publication of his paper, others by the names of Batten and Gibb pursued the work and clearly established myotonic dystrophy as a distinct degenerative neurological disorder. They documented that the disease affected several family members and all had the characteristic symptoms of progressive difficulty in walking with weakness of the legs and grip and also a striking lack of expression due to weakness of the facial muscles. Over the past century, through progress in medicine and technology, several clinical and physio-pathological discoveries helped paint a more defined picture of the characteristic symptoms of myotonic dystrophy. It was not however until 1992, that the genetic defect was coined by several teams showing that an unstable CTG repeat expansion in the 3’untranslated region of a gene could be the molecular basis for the DM pathology (Aslanidis et al., 1992; Brook et al., 1992; Buxton et al., 1992b; Fu et al., 1992; Harley et al., 1992a; Mahadevan et al., 1992).

Myotonic dystrophy is the most frequent muscular dystrophy in adults with a generalized worldwide frequency of one affected individual in every 15 000 (Harper, 2001). Some regions of the world show however elevated frequencies of the disease. Of special relevance to Canadians, the Saguenay region in Quebec shows one of the most elevated frequencies in the world where 1 individual in every 500 is a carrier of the mutation. The population in the Saguenay region shows high frequencies of several genetically inherited disorders mainly caused by a phenomenon called the founder effect. A founder effect relates to the transmission of a genetic mutation by a single individual living in a secluded population to his progeny. The Saguenay region was colonized in the 17th century by a small group of French settlers. It is now known through genealogical studies that most cases of myotonic dystrophy originate from a single common ancestor (Harper, 2001). The Saguenay region is located in Northern Quebec and was generally inaccessible by common means of transportation for most of the winter period, until the avenue of roads and railways in the early 20th century. Because weak immigration to this remote region did not contribute to diversify the genetic pool, the population remained very homogenous and the mutation was widely distributed throughout the population.

What makes myotonic dystrophy such a devastating neuromuscular disease is that it is dominantly inherited and the symptoms occur at earlier ages with every generation, this phenomenon is known as anticipation. Also, the sex of the affected parent has an important role as to the severity of the disease since almost all congenital forms of the disease appear when the mother is the carrier of the mutation. These are some reasons why genetic testing within families known to carry the mutation is crucial.

There are three typical forms of myotonic dystrophy established according to the symptoms and the age of onset. Adult onset myotonic dystrophy is the most common and the less severe form of the disease with symptoms appearing in the early twenties to the late forties. Congenital myotonic dystrophy is the most severe form of the disease with symptoms present at birth. Childhood myotonic dystrophy is when the disease first appear early in life. Although less severe than the congenital cases, this form can be associated with mental retardation and retardation in muscle development. Childhood myotonic dystrophy will not be addressed here.

The severe form of the disease that is present at birth is called congenital myotonic dystrophy. This form is almost exclusively seen when the mother is the carrier of the disease (Harper, 2001). This devastating form of myotonic dystrophy is often regarded as a distinct neuromuscular disease altogether because of the likely involvement of developmental disturbances, the extreme severity of the symptoms and the heavy neurological component that is not present in adult onset cases. This form is associated with hypotonia and also cardiac and respiratory complications that very often lead to the death of the neonate (Harper, 2001). Neurological features include mental retardation, delayed motor development and difficulties in feeding, swallowing and speech. The condition of the few neonates that survive into childhood seems to improve at first, but rapidly regresses once the adult and degenerative features of the disease appear. Individuals with the congenital form are heavily disabled throughout their life and do not live beyond 30 years of age (Harper, 2001).

Table 2: Myotonic dystrophy clinical symptoms in adults

System / Organ

Defect

Endocrine

• Insulin resistance

• Testicular atrophy

• Increased FSH levels

Eyes

• Cataracts• Ptosis• Retinal degeneration

Heart

• Conduction defects

• Cardiac arrhythmias

Respiratory

• Weakness and myotonia of the diaphragm and respiratory muscles

• Abnormal pharyngeal and oesophageal contractions

• Bronchial aspiration

• Alveolar hypoventilation

Neurological / Brain

• Mental retardation in congenital forms

• Hypersomnia

• Variable cognitive impairment

• Personality : increased apathy, avoidance and passive-aggressive behaviour (Delaporte, 1998)

Skin

• Premature balding

Smooth muscle

• Widespread involvement in gastrointestinal tract

• Delayed gastric emptying

• Sphincter laxity

• Incoordinate contractions of the uterus during labour

Skeletal muscle

• Myotonia

• Muscle weakness and wasting

• Distal weakness of the limbs

• Weakness of facial, jaw, tongue and neck muscles

• Severe weakness and wasting of sternomastoids

The first attempts of positioning the DM mutation were carried out in the early 70’s when linkage was first established with the Lutheran blood group utilizing family studies in the USA and Britain (Harper et al., 1972). The presence of the Lutheran antigen a+ Lu(a+) was consistently passed down with the DM mutation in the studied families (Harper et al., 1972). In the following years, several other loci (ABH secretor locus, peptidase D, Lewis blood group, complement C3, ApoE) were all linked with transmission of the DM mutation (Harper et al., 1972; O'Brien et al., 1983; Schrott and Omenn, 1975; Sistonen, 1984; Whitehead et al., 1982). Assignment of the complement C3 loci to chromosome 19 in 1982, made it clear that DM mutation was also located on that chromosome (Whitehead et al., 1982). The DM mutation was then further narrowed down to the 19p13-19q13 region when linkage to its closest markers, the genes for apolipoprotein C2 (ApoC2) and the muscle-specific creatine kinase (CKMM) were established by studying French Canadian, British French and Dutch families (Brunner et al., 1989; Shaw et al., 1985). The final positioning of the DM gene was done in 1991 by Harley and co-workers by establishing that the DM gene lies in region 19q13.2-q13.3, telomeric to the ApoC2 and CKMM genes (Fig. 1) (Harley et al., 1991).

From then, several teams put much effort in isolating and constructing DNA libraries in lambda phages, yeast artificial chromosomes and cosmids (Aslanidis et al., 1992; Buxton et al., 1992a; Harley et al., 1992b). By Restriction Fragment Length Polymorphism (RFLP) studies, a genomic segment was identified that showed length variations in myotonic dystrophy patients (Aslanidis et al., 1992; Buxton et al., 1992b; Harley et al., 1992a). It was in early 1992, through positional cloning techniques, that the mutation responsible for myotonic dystrophy was first identified as an instable CTG expansion (Aslanidis et al., 1992; Brook et al., 1992; Buxton et al., 1992b; Fu et al., 1992; Harley et al., 1992a; Mahadevan et al., 1992). A very interesting observation came from Mahadevan and co-workers when they showed that the CTG expansion was located in the 3’ untranslated region (3’UTR) of dystrophia myotonica protein kinase gene (DMPK) in 98% of patients clinically diagnosed with myotonic dystrophy (Mahadevan et al., 1992). Although the function of DMPK is still to date unclear, it was challenging at the time to comprehend how this expansion both on the DNA and transcribed to the RNA was responsible for the cellular and molecular defects seen in dystrophic cells.

The left panel shows a schematic representation of chromosome 19 and the approximate location of the DM1 gene locus with or without the CTG expansion. The right panel shows flanking genes and probes used to locate the DM1 gene locus. (Illustration from. (Moxley, 1992) ).

The discovery of the mutation causing myotonic dystrophy came at a turning point in genetics when a number of diseases affecting the central nervous system (CNS) were being identified as trinucleotide repeat-related disorders. The previous year, in 1991, the CGG trinucleotide expansion causing fragile X syndrome had been identified (Verkerk et al., 1991) and the following year, in 1993, the CAG repeat expansion responsible for the toxic polyglutamine tract causing Huntington’s disease was also identified (Group, 1993). Several trinucleotide diseases identified to date share several key features, reviewed in (Brice, 1998; Martin, 1999; Paulson and Fischbeck, 1996):

1) Inheritance is autosomal dominant or x-linked (except for Friedrich’s ataxia, which is autosomal recessive)

2) Severity of the disease is variable but generally correlates with increased expansion length

3) Trinucleotide repeats are unstable when they reach a certain threshold and generally undergo expansion when transmitted, a premise to anticipation

4) Expanded repeats are polymorphic when inherited, but are transmitted stably when under the critical threshold

Trinucleotide diseases can be further broken-down within 2 categories, diseases where the repeats are translated into protein or not (Table 3). When the repeats are not translated, the diseases are characterized by much larger expansions and usually multiple systems are affected. It is fascinating to see how a similar type of mutation can generate on one hand common clinical symptoms which is CNS degeneration, especially for the group of translated expansion disorders coding for a polyglutamine tract and on the other hand disease-specific alterations. Both myotonic dystrophy and SCA8, for example, are caused by a transcribed but untranslated CTG repeat. Although SCA8 patients have generally much fewer CTG repeats than in myotonic dystrophy, very few clinical symptoms are shared by both diseases.

Table 3: Trinucleotide repeat diseases

Disease

Repeat

Translated

Threshold

• Fragile X types A-E

CGG

No

Normal: 6-52

Pre-mutation: 60-200

Disease: 200-2000

• Myotonic Dystrophy Type I

CTG

No

Normal: 3-49

Disease: 50-3000

• Spinocerebellar ataxia type 8 &12 (SCA8 & 12)

CTG

No

Normal: 16-37

Disease: 90-152

• Friedreich’s ataxia (spinal bulbar muscular atrophy)

GAA

No

Normal: 7-22

Disease: 100-2000

• Huntington’s disease

CAG

Yes

Normal: 11-34

Disease: 36-121

• Kennedy’s disease

CAG

Yes

Normal: 11-42

Disease: 40-62

• Spinocerebellar ataxia types (1-3,6, 7)

CAG

Yes

Normal: 7-36*

Disease: 37-130*

• Dentatorubral-pallidoluysian atrophy

CAG

Yes

Normal: 7-25

Disease: 49-88

*General approximation. Each SCA disease has its own defined clinical threshold (Martin, 1999)

Anticipation in myotonic dystrophy and also in other triplet repeat diseases is a hallmark clinical feature that has been documented for more than 90 years (Harper, 2001; Moxley, 1992). It was not until the early 90’s, once the mutation had been identified, that extensive studies were conducted to monitor several key components of anticipation in myotonic dystrophy. These studies were done on several families and looked at parameters such as the age at onset of the affected parent, the age at onset of offspring, the number of repeats in affected parents and offspring and the influence of the sex of the affected parent on the outcome of the disease in the offspring (Ashizawa et al., 1994; Ashizawa et al., 1992a; Ashizawa et al., 1992b; Harley et al., 1993). The results revealed that although contraction of the mutation was observed in 10% of cases when the father carried the mutation and in 3% of cases when the mother had the mutation, the general tendency was that offspring had an increase in the length of the repeat tract and severity of the disease (Ashizawa et al., 1992b). The increment in CTG repeats in children with myotonic dystrophy is closely related to the length of the repeats in the parents, especially if the mother is a carrier of the mutation (Barcelo et al., 1993; Harper, 2001). Observations on families with myotonic dystrophy allowed thresholds linking CTG repeat number amplification and disease severity to be established (table 4).

Table 4: Critical CTG expansion thresholds for myotonic dystrophy onset and symptoms

Number of CTG repeats

Phenotype

Comments

3 to 49

Normal individuals

Over 90% of the general population has less than 35 CTG repeats in the DM1 locus.

50 to 80

Very mild or no apparent symptoms except cataracts

Expansions in this range are incremental from generation to generation.

80 -1000

Full mutation, adult onset

80 CTG repeats is the threshold for saltatory amplification.

600 and more

Full mutation, adult onset

High risk of transmitting the congenital form if the mother is the carrier.

1000 to 3000

Generally congenital

Such expansions are most frequently seen in congenital cases.

In order to better understand how the severity of myotonic dystrophy increases within the lifespan of an affected individual, and why the children with myotonic dystrophy have generally longer repeat lengths than their parents, it is necessary to look at how the CTG repeat tract is replicated. To shed light on the molecular mechanisms of DNA replication, several teams have studied how these CTG repeats behave in Escherichia coli (Jaworski et al., 1995; Kang et al., 1995; Samadashwily et al., 1997; Sarkar et al., 1998) . Their first observation was that both expansions and deletions exist in E.coli depending on which strand lies the expansion (Iyer and Wells, 1999; Kang et al., 1995). A CTG repeat tract tends to expand when located on the leading template rather than on the lagging strand (Iyer and Wells, 1999; Kang et al., 1995). Figure 2 panels A and B displays a model of current beliefs on how this may occur. Although this model gave, in 1995, the first insight on how CTG repeats could expand from generation to generation, it could not explain how there is sometimes a dramatic amplification when the mutation is transmitted from mother to child, as in congenital cases for example.

In 1998, a paper by Sarkar and colleagues brought new insight to somatic instability of CTG repeat tracts. This work brought to light mechanisms that may be implicated in the bimodal pattern of CTG amplification (Sarkar et al., 1998) where small CTG repeats (n < 80) show incremental augmentation when inherited as compared to the large amplifications seen when the parent has relatively large expansion (>200). The main conclusions of this work proposed that expansions smaller than a single Okazaki fragment in humans ( < 100 – 200 bp) is susceptible to an incremental amplification but the much larger amplifications were caused when several CTG-encoding Okazaki fragments were necessary to complete the strand replication (Sarkar et al., 1998). These Okazaki fragments could undergo slippage from being unanchored at both ends by a non-repetitive sequence. The combined effect of having several Okazaki fragments containing long CTG hairpin expansions that could evade normal DNA repair during mitosis is the standing model for these large expansion amplifications seen when affected individuals reach the critical threshold of approximately 80 CTG repeats (240 bp). CTG amplification in this E.coli model necessitated however loss of SbcC, a protein that modulates cleavage of single-stranded DNA and degradation of duplex DNA from double-strand breaks. It is still not known today if disruption of a homologous protein to SbcC in humans or if the misregulation of other elements of the mismatch repair system like Msh3 and Msh6 are also implicated at some level for CTG amplification in myotonic dystrophy (Sarkar et al., 1998; van den Broek et al., 2002).

(A) A possible mechanism for expansion is that primer relocation and polymerase slippage when transcribing CAG repeats induces the formation of a CTG hairpin loop on newly synthesised DNA. The size of these loops will dictate the length of the expansion in the replicated strand. (B) A model for deletions (or contractions) where the replication fork skips a strong CTG hairpin secondary structure present in the template DNA. This skipping will result in the loss or reduction of the CTG tract in the synthesized DNA. (C) The predicted base-paring structure for CTG and CAG repeats. CTG repeat hairpin structures are far more stable because of the weak repulsion of T-T pairs compared to A-A pairs (Smith et al., 1995). (Illustration from (Wells, 1996) )

Because of the unstable nature of the repeat tract, it is expected that dividing cells are subject to increasing expansions with every division. This phenomenon of somatic instability underlines the very basis of anticipation in the disease and may explain why the severity of myotonic dystrophy increases in time. Patient studies focussed on somatic instability assessed two important points: variations in the mutation length between different tissues and the variation of the mutation length in a single tissue at different points in time.

Five important elements derived from patient studies highlight the unstable nature of the CTG repeats (Martorell et al., 1997):

1) The expanded allele appears as a smear on Southern blots

2) Identical twins affected by myotonic dystrophy have different expansion patterns

3) Immortalized lymphocytes from patients show increased expansions in culture

4) Repeat lengths in patients blood and tissue increase with time

5) Repeat lengths in DNA vary from tissue to tissue in a same patient

These important elements paint out myotonic dystrophy as a disease where repeat sizes increase at different rates in different tissues. Hence, this dynamic process may explain the age-dependent and tissue dependent phenotypic manifestations of the disease (Khajavi et al., 2001).

Somatic instability in myotonic dystrophy seems to occur in a time-dependent fashion. Studies conducted on congenitally affected fetuses and neonates revealed that repeat heterogeneity between organs and tissues appears after 16 weeks of development (Martorell et al., 1997). This timing coincides with the second trimester of development and the second wave of myogenic development (Martorell et al., 1997). Following the 16th week of development, expansion heterogeneity sets-in and continues to expand throughout the life of the affected individual. Progression of the CTG tract instability does not seem to coincide solely with the mitotic activity of the tissue since mean expansion length in lymphocytes and sperm are generally much lower that those seen in muscle, heart or kidney (Anvret et al., 1993; Martorell et al., 1997; Martorell et al., 1998; Monckton et al., 1995; Zatz et al., 1995). It is of interest to note that although the mean repeat length in sperm is lower than that of other tissues, there is extensive variability in repeat length in the male germline including both contractions and even reversions to normal length (Monckton et al., 1995). This variability may be responsible for some observed cases of contractions seen when the father is the carrier of the disease and may also add an alternative explanation to the intergenerational amplification seen between affected fathers and their offspring (Harper, 2001; Monckton et al., 1995).

Although the clinical and physiological perturbations of myotonic dystrophy have long been described, the more subtle molecular and cellular defects are much more elusive. It is still not clear today, twelve years after the initial identification of the mutation responsible for myotonic dystrophy, which are the molecular mechanisms involved in causing the vast range of symptoms seen in the disease.

Retention of the mutant DMPK mRNA causes a haplo-insufficiency of DMPK protein since only the normal transcripts are translated. Accordingly, it was demonstrated using specific anti-DMPK antibodies that DMPK protein levels were reduced to 50-57% of the levels in normal myoblasts (Furling et al., 2001a; Furling et al., 2001b). DMPK is a cyclic AMP-dependent serine/threonine protein kinase belonging to the Rho family. Fifteen exons predicts a 70.6 kDa protein with over 7 different tissue-specific splicing isoforms (Fig. 4a) (Groenen et al., 2000; Tiscornia and Mahadevan, 2000; Wansink et al., 2003). However, through strong post-translational modifications, certain isoforms of the protein can achieve an apparent molecular mass of up to ~ 80-86 kDa in skeletal muscle (Bush et al., 2000; Lam et al., 2000; Mahadevan et al., 1993; Wansink et al., 2003). These numerous DMPK isoforms exhibit cell-type and location-dependent substrate specificities possibly conferring to DMPK different physiological roles (Wansink et al., 2003). Figure 4b depicts the DMPK protein with its various structural domains. The function of these domains still needs to be defined, but one study has suggested that DMPK must be cleaved from it’s membrane associated domain in order to adopt an active conformation in the cytoplasm (Bush et al., 2000). Another study from the same group, suggests that phospholemman, a membrane-bound substrate for protein kinase A and C involved in ion transport, is also a substrate for DMPK phosporylation (Mounsey et al., 2000a). Modification of muscle Ca2+ and Na+ ion channels homeostasis could lead to an alteration of muscle excitability, as seen in myotonic dystrophy (Benders et al., 1997; Mounsey et al., 2000b). At this time, there are but speculations of the true biochemical role of DMPK in healthy individuals or in myotonic dystrophy (Wansink et al., 2003).

(A) The DMPK gene consists of 15 exons. The CTG expansion is located in exon 15. Exons are depicted as grey boxes, black boxes represent alternatively spliced exons and small white boxes represent cryptic intron segments. (B) The DMPK protein is comprised of a N-terminal leucine rich region (LR) followed by the serine/threonine kinase domain. Region II encodes a VSGGG motif responsible for substrate specificity, followed by a possible Rho-binding domain (RBD), a coiled-coil domain (CC) and a subcellular localization tail domain (Tail) allowing targeting of the protein to mitochondria, endoplasmic reticulum or to the cytosol (Wansink et al., 2003).

The development of specialized tools like mouse models and cell lines were necessary in order to examine the effects of the myotonic dystrophy mutation in live cells and tissues. In 1996, appeared the first mouse models to help understand myotonic dystrophy (Jansen et al., 1996). In these models, Jansen and colleagues disrupted the endogenous murine DMPK gene and/or overexpressed a normal human DMPK transgene. Their goal was to determine the influence of DMPK misregulation in the disease. Interestingly, the nullizygous mice they generated showed only mild atrophy in head and neck muscles, while the mice overexpressing human DMPK showed copy number-dependent cardiomyopathy (Jansen et al., 1996). In another study, the authors showed that these knockout mice showed cardiac conduction defects (Berul et al., 1999), which is in accordance with the previously published results and decreased muscular strength implicating a role for DMPK in maintenance of muscle structure (Reddy et al., 1996). However, none of these mouse models have exhibited the most prominent features of myotonic dystrophy such as the myotonia, cataracts and severe muscle wasting. These results clearly point out that the DMPK protein itself has but a subtle role in myotonic dystrophy pathogenesis in mice. Since murine and human DMPK share much similarity, it is not expected that DMPK plays any critical role in humans. Also, patients homozygous for the DM mutation do not exhibit symptoms more severe than heterozygous patients with a similar expansion length (Harper, 2001). This implies that the complete lack of DMPK protein in humans does not account for any visible phenotype at first glance.

Two other mouse models were later generated to look at CTG tract stability. The first group used a fragment of DMPK 3’UTR containing 162 CTG repeats (Monckton et al., 1997), the other inserted the whole DM locus (45 kb) containing 55 CTG repeats and adjacent SIX5 and DMWD genes (Gourdon et al., 1997). Both groups arrived at similar conclusions as they observed modest, but significant meiotic and mitotic CTG repeat instability. The best myotonic dystrophy mouse model to date came later when Gourdon’s group inserted a 300 CTG repeat expansion in the 45 kb genomic fragment instead of the 55 CTG repeat sequence. (Seznec et al., 2000). These mice show a very similar pathological phenotype as seen in humans:

A high level of instability increasing with age in tissues and sperm

Histological muscle abnormalities (atrophy of slow muscle fibers, centronucleated muscle fibers, fibrosis and heterogeneity in fiber diameters)

Myotonia

Formation of foci revealed by a DMPK riboprobe

A major step in understanding the true nature of myotonic dystrophy came with the development of a mouse model carrying a large CTG repeat expansion (~250) in 3’UTR of the human skeletal actin gene (Mankodi et al., 2000). The aim of that study was to determine if the pathogenic effect of the mutation comes from the RNA or the DNA. What they observed was that mice carrying the expanded repeat developed myotonia, several histological features of myotonic dystrophy (ring fibers, sarcoplasmic masses, centronucleation) and CUG-related foci formation in muscle cells. However, no muscle weakness was reported in these mice. The fact that these mice exhibited the cardinal features of human myotonic dystrophy strongly supported a toxic gain-of-function role for RNA containing the large CUG repeats.

The initial idea of the gain-of-function effect of the RNA in myotonic dystrophy came in 1996 from Timchenko and co-workers when they identified CUG-BP, a heterogeneous nuclear ribonucleoprotein (hnRNP), that could bind to single-stranded CUG repeats and also accumulates in the nucleus of cells from myotonic dystrophy patients (Roberts et al., 1997; Timchenko et al., 1996a; Timchenko et al., 1996b; Timchenko et al., 2001a). CUG-BP is a phosphorylation substrate for the DMPK protein and exists as two phosphorylation isoforms that are ubiquitously expressed in both the cytoplasm and the nucleus (Roberts et al., 1997; Timchenko et al., 1996b). It is the hypophosphorylated isoform that accumulates in the nucleus of cells form myotonic dystrophy patients. The role hnRNPs, such as CUG-BP, is to perform multiple posttranscriptional regulatory functions such as splicing. CUG-BP is a member of the CELF family of RNA-processing factors that regulate alternative splicing (Ladd et al., 2001). Aberrant accumulation of the hypophosphorylated form of CUG-BP in the nucleus has been associated with abnormal splicing of several of its substrate mRNAs, ultimately lending support for a trans-dominant effect of expanded CUG repeats on RNA processing in myotonic dystrophy type 1 (Savkur et al., 2001). However, this accumulation is unlikely to be a titration effect from binding to the CUG repeat tract itself since accumulation of CUG-BP in the nucleus was not shown to be proportional to it’s length (Michalowski et al., 1999).

Five pre-mRNAs have been identified to be aberrantly spliced in myotonic dystrophy type 1 tissues and mouse models: tau (Sergeant et al., 2001), myotubularin-related protein 1 (MTMR1) (Buj-Bello et al., 2002), the insulin receptor (IR) (Savkur et al., 2001), cardiac troponine T (cTNT) (Philips et al., 1998) and that of chloride channel 1 (Clc-1) (Charlet et al., 2002; Mankodi et al., 2002). While the role of MTMR1 is still elusive in myotonic dystrophy, disruption of appropriate splicing of cTNT, IR and CLC-1 mRNAs can be associated with symptoms of the disease. It has not yet been demonstrated that the splicing defect of tau is associated with particular symptoms of myotonic dystrophy. However, tau is a microtubule-associated protein implicated in axon and neurite development in the brain (Shahani and Brandt, 2002). Alterations in the balance of tau isoforms may well be responsible for the abnormal personality traits of myotonic dystrophy patients and could explain the severe involvement of the CNS in congenital cases.

Human cTNT exists as two splicing isoforms with either inclusion or exclusion of exon 5. The isoform that includes exon 5 is normally only expressed during embryonic development in the heart and skeletal muscle, while in adults exon 5 is excluded (Cooper and Ordahl, 1984). Homozygous and heterozygous myotonic dystrophy patients show alternate splicing patterns with a significant increase in cTNT mRNAs with exon 5 inclusion (Philips et al., 1998).

The insulin receptor is composed of two extracellular α and two intracellular β subunits. Alternative splicing of exon 11 generates two isoforms of the α subunit: isoform A (IRA) lacking exon 11 and isoform B (IRB) that includes exon 11 (Fig. 5a) (Seino and Bell, 1989). Expression of the two isoforms is tissue specific depending on glucose requirement of each particular tissue since both isoforms do not have the same signalling capabilities (Kellerer et al., 1992; Moller et al., 1989). The B isoform, that is expressed dominantly in skeletal muscle, adipose tissue and the liver, has been reported to signal more efficiently in response to insulin binding despite having a 2-fold lower affinity for insulin (McClain, 1991; Moller et al., 1989; Mosthaf et al., 1991; Yamaguchi et al., 1993). In myotonic dystrophy, the IR is aberrantly spliced thus promoting the IRA isoform is skeletal muscle (Savkur et al., 2001).

(A) Insulin receptor (IR) α-subunit alternative splicing. IR-A is produced with inclusion of exon 11, while the IR-B isoform is produced by exclusion of exon 11. IR-A is the high-signalling isoform normally found in insulin-responsive tissue like skeletal muscle and the liver. (B) Alternative splicing of the chloride channel-1 (Clc-1) mRNA. Aberrant splicing generates 3 isoforms coding for inactive proteins with premature stop codons. (Illustration reproduced from: (A) (Savkur et al., 2001) (B) (Charlet et al., 2002))

Myotonia is one of the major cardinal features of myotonic dystrophy. The inability to relax muscles after contractions is due to hyperpolarisation of the muscle fibers after ion entry (Lehmann-Horn and Jurkat-Rott, 1999). Several channelopathies such as mutations in the muscle-specific sodium ( SCN4A ) and chloride ( CLCN1 ) channel genes are known to cause myotonia in humans and animals (Lehmann-Horn and Jurkat-Rott, 1999). The Clc-1 channel is the product of the CLCN1 gene and is the major skeletal muscle chloride channel in humans (Lehmann-Horn and Jurkat-Rott, 1999). Since defects in both sodium and chloride conductance have been documented in myotonic dystrophy, Charlet and co-workers investigated whether SCN4A and CLCN1 gene products are misregulated in the disease (Charlet et al., 2002; Franke et al., 1990; Koty et al., 1996). Their study showed, that similar to cTNT and the IR, the Clc-1 mRNA was aberrantly spliced resulting in 3 defective isoforms containing a premature termination codon (Fig. 5b) (Charlet et al., 2002; Mankodi et al., 2002). Using a myotonic dystrophy mouse model it was determined that loss of Clc-1 function was sufficient to induce myotonia (Mankodi et al., 2002).

A common mechanism binds the processing of the IR, cTNT and Clc-1: they are all natural targets for the alternative splicing regulator CUG-BP (Fig. 6) (Ladd et al., 2001). CUG-BP has been shown to bind U/G-rich motifs in introns adjacent to splice sites in all three of these pre-mRNAs (Charlet et al., 2002; Philips et al., 1998; Savkur et al., 2001). Furthermore, overexpression of CUG-BP in normal cells induces the same defective alternative splicing as that seen in myotonic dystrophy patients (Faustino and Cooper, 2003). The exact mechanism by which CUG-BP accumulates in the nucleus of myotonic dystrophy cells is unknown. Two possibilities have been put forward to explain this observation: the first is the haplo-insufficiency or complete lack of DMPK favours the hypophosphorylated form of CUG-BP that is shown to accumulate in the nucleus (Philips et al., 1998; Roberts et al., 1997). The second explanation suggests that double strand RNA binding proteins (dsRNA-BP) attach themselves to the CUG expansion hairpin loop thereby forming the characteristic foci and consequently inducing an elevation in CUG-BP steady-state levels in myotonic dystrophy cells (Charlet et al., 2002; Faustino and Cooper, 2003; Miller et al., 2000; Timchenko et al., 2001a).

Binding of MBNL on hairpin structures of the repeat tract induces an elevation in the nuclear levels of CUG-BP. The splicing factor, CUG-BP, then binds to specific sequences in target genes responsible for alternative splicing. CUG-BP can then either induce exon inclusion (cTNT) or exclusion (IR and CLC-1). Defective alternative splicing will result in several of the hallmark features of DM1.Illustration from (Faustino and Cooper, 2003)

Myoblast differentiation is regulated by a sequential cascade of gene activation following the initial mitogen withdrawal (Fig. 7). Myoblasts are held in proliferation by presence of growth factors and of proteins Notch, Msx-1 and the bone morphogenic protein 4 (BMP-4), that block MyoD and Myf5 expression (Bailey et al., 2001). Transcription factors MyoD and Myf5 are the first genes to be activated since they are responsible for all downstream activation events in myogenesis. They bind the promoters of muscle-specific genes and activate their transcription (Rudnicki et al., 1993). Myogenin expression is the next key factor to be expressed and coincides with commitment of myoblasts to terminal differentiation (Sabourin and Rudnicki, 2000). Myogenin expression is crucial for fusion of myoblasts into multinucleated myotubes (Wright et al., 1989). Shortly following myogenin, the inhibitor p21 is expressed and marks the irreversible withdrawal from the cell cycle followed by expression terminal differentiation markers like myosin heavy chain (MyHC) and muscle creatine kinase (CK) (Sabourin and Rudnicki, 2000).

Myoblasts are maintained in proliferation by presence of growth factors and through positive regulation from proliferation factors BMP-4, Msx-1 and Notch. Once growth factors are withdrawn, MyoD and Myf5 are expressed and antagonists Noggin and Ssh inhibit expression of BMP-4, Msx-1 and Notch. Master genes MyoD and Myf5 activate muscle-specific genes needed for entering the differentiation program. Myogenin is activated and triggers terminal differentiation followed by p21 that irreversibly withdraws the myoblasts from the cell cycle. Muscle-specific structural genes and markers are then expressed to allow cell fusion of myotubes formation.

Patients with myotonic dystrophy display progressive muscle weakness and wasting but in congenital cases, skeletal muscle biopsies display defects in muscle maturation indicating that alterations in myogenesis occurred during development (Harper, 2001). Myogenic satellite cells from myotonic dystrophy patients and murine C2C12 myogenic cells transfected with a plasmid expression 200 CTG repeats both show severe alterations in differentiation capabilities (Amack et al., 1999; Bhagwati et al., 1999; Furling et al., 2001a). Although it is clear that the repeat tract in the RNA has a role to play in altering myogenesis, the exact mechanism is unclear and still strongly debated. Three different groups propose three different pathways to explain the disruption of myoblast differentiation in myotonic dystrophy. The first group used a C2C12 cell model to express various constructs containing either a CTG expansion with or without DMPK 3’UTR (Amack and Mahadevan, 2001; Amack et al., 1999). What they observed was that although the RNA containing the CUG expansion did indeed form foci in this cell model, the repeat tract alone was not sufficient to alter differentiation (Amack and Mahadevan, 2001). When the repeat tract was expressed with DMPK 3’UTR, the authors noted that myoblasts differentiation was halted in parallel to diminished expression of both myogenin and p21 mRNAs. While they did not measure any reduction in MyoD or Myf5 mRNAs in this initial study, three years later, they showed that MyoD protein levels were reduced in myoblasts expressing DMPK 3’UTR and the CTG repeat tract (Amack et al., 2002). Reduction of MyoD levels is post translational and the authors speculate that reduced protein stability may be caused by a trans effect of RNA foci (Amack et al., 2002). Interestingly, they also found that MyoD or myogenin overexpression in these cells contributed to restore myoblast differentiation. This involves that differentiation defects lies at or before myogenin expression and that events downstream are intact and functional (Amack et al., 2002).

In vitro studies by another group found that p21 was reduced in relation to defective expression of cytoplasmic CUG-BP (Timchenko et al., 2001b). They state that cytoplasmic CUG-BP induces p21 levels during myoblast differentiation by regulating its translation (Timchenko et al., 2001b; Timchenko et al., 1999). In myotonic dystrophy, CUG-BP aberrantly accumulates in the nucleus whereby being unavailable for binding to p21. Lending support to this model, Khajavi and colleagues noted that lymphoblasts with large CTG expansions had a growth advantage over cells with smaller expansions (Khajavi et al., 2001). They found that this event was due to increased Erk1,2 activation. Erk1,2, kinases involved in the regulation of meiosis, mitosis and postmitotic functions in differentiation, are negatively regulated by p21, itself downregulated in myotonic dystrophy (Johnson and Lapadat, 2002; Khajavi et al., 2001).

While all of these reports describe downstream events that may help explain why differentiation is compromised, the cause of the disturbances seem to funnel to the RNA retention and sequestration of binding proteins. While Amack and colleagues confer to these foci the capacity of inducing by some unknown mechanism the degradation of MyoD, Timchenko and co-workers attribute accumulation of the nuclear form of CUG-BP to interaction with CUG-binding proteins (Amack et al., 2002; Timchenko et al., 2001b). One recent report described that constitutive expression the muscleblind-related protein CHCR shown to accumulate in myotonic dystrophy foci can cause downregulation of myogenin and MyHC (Fardaei et al., 2002; Squillace et al., 2002). This adds new credit to the toxic gain-of-function theory through which aberrant nuclear accumulation and stabilisation of CUG-binding proteins involved in terminal differentiation of muscle may indirectly be responsible for myopathy in myotonic dystrophy.

Trinucleotide CTG repeats in DNA form stable hairpin structure in vivo (Freudenreich et al., 1997; Mariappan et al., 1996). The chromatin structure in proximity of the repeat tract is highly perturbed and anomalously condensed from tight nucleosome assembly (Wang et al., 1994). This highly condensed region of the chromatin could result in transcription repression (Wang et al., 1994). In fact, the DMPK locus is flanked by two genes: upstream by DMWD and downstream by SIX5 (formerly known as DMAHP) (Fig. 8) (Alwazzan et al., 1999; Klesert et al., 1997). Both these genes have been shown to be downregulated in presence of large CTG expansions (Alwazzan et al., 1999; Klesert et al., 1997; Thornton et al., 1997). While the role of the DMWD gene product is suspected to be involved in brain-related symptoms in myotonic dystrophy, deficiency of SIX5 causes cataracts in knockout mouse models (Klesert et al., 2000; Sarkar et al., 2000; Westerlaken et al., 2003). While these results support a cis effect for the DM mutation, these mouse models did not present any muscle-related abnormalities (Klesert et al., 2000; Sarkar et al., 2000).

The DMPK gene is flanked by DMWD and SIX5 genes. The DMPK polyadenylation site is less than 300 bp from the SIX5 initiation codon. In black are represented CpG islands that are methylated in congenital myotonic dystrophy (Image from (Tapscott, 2000).

While the triplet repeats may constitute themselves a regulatory element for adjacent genes, the CpG island at the 3’ end of the DMPK gene may also have a potential role in gene regulation (Boucher et al., 1995; Steinbach et al., 1998). In fragile X syndrome, hypermethylation of the CpG island in the promoter of the FMR1 gene causes transcriptional repression (Knight et al., 1993; Schwemmle et al., 1997). A study on the methylation status of this CpG island revealed that although there is no allele-specific methylation or parental imprinting (Jansen et al., 1993; Shaw et al., 1993), hypermethylation was present in cells from congenital myotonic dystrophy patients (Steinbach et al., 1998). Hypermethylation of the CpG island in the DMPK gene constitutes one of the only molecular alterations to clearly distinguish adult from congenital forms of myotonic dystrophy in primary culture cells (Filippova et al., 2001; Steinbach et al., 1998).

Unlike its role in fragile X syndrome, hypermethylation of the CpG island in 3’ of the DMPK gene has been shown to disrupt the function of an insulator allowing delimitation between DMPK and SIX5 cis -acting elements (Filippova et al., 2001). An insulator is a gene boundary element that blocks an enhancer from acting on a promoter when placed between them or by protecting a transgene from chromatin silencing when placed at either end (Zhan et al., 2001). Essentially, their role is to prevent positioning effects thereby allowing independent function and regulations of the genes within their boundaries (Zhan et al., 2001). CTCF is the zinc-finger protein that can bind insulator sequences to confer functionality (Bell et al., 1999). Two CTCF binding sites flank the CTG tract in the DMPK gene to form a functional insulator (Fig. 9) (Filippova et al., 2001). CTCF proteins cannot bind the insulators when CpG islands are hypermethylated. This causes the loss of insulator activity and may allow crosstalk between the Six5 enhancer and the DMPK promoter.

CTG repeats are flanked by 2 CTCF binding sites. All three elements are needed for enhancer-blocking activity. CTCF binding sites 1 and 2 are within CpG islands that are hypermethylated in congenital myotonic dystrophy. Hypermethylation of the CpG islands prevent binding of CTCF, thus abolishing insulator activity thereby exposing the DMPK promoter to cis effects from the Six5 enhancer (E).

Since the discovery of the DM mutation, genetic testing has been used to validate clinical diagnosis for myotonic dystrophy. These tests revealed that 98% of patients who displayed clinical symptoms for myotonic dystrophy had in fact the DM mutation (Mahadevan et al., 1992). However, between 1-2 % of patients that displayed most of the classical symptoms of myotonic dystrophy did not have the characteristic CTG expansion in 3’UTR of the DMPK gene (Mahadevan et al., 1992).

The first clinical evidence that there might exist a second form of myotonic dystrophy came from patient studies in 1994 (Ricker et al., 1994). Patients in these studies displayed most clinical symptoms of myotonic dystrophy without the CTG expansion in the DM1 locus. They also displayed proximal muscle weakness as opposed to distal as in classical myotonic dystrophy. This distinct form of myotonic dystrophy was from then on named proximal myotonic myopathy (PROMM), until the molecular identification of the mutation in 2001 by Liquori and colleagues (Harper, 2001). There are five main clinical symptoms that distinguish myotonic dystrophy and PROMM (Harper, 2001; Vihola et al., 2003):

1. The predominant distributions of muscle weakness is proximal in PROMM

2. PROMM is associated with muscle pain

3. PROMM patients display type II fiber atrophy as opposed to type I

4. Facial weakness is absent in PROMM

5. No CNS involvement has been documented for PROMM

These findings quickly led research teams in a rally to identify a second DM locus. In 1998, a second myotonic dystrophy locus was mapped to a 10-cM region on chromosome 3 by studying a five-generation family showing symptoms of a ‘myotonic dystrophy-like’ disease (Ranum et al., 1998). Three years later, the mutation responsible for the unaccounted cases of classical myotonic dystrophy was coined (Liquori et al., 2001). PROMM or myotonic dystrophy type 2 (DM2) is caused by a CCTG expansion in intron 1 of the zinc finger protein 9 (ZNF9) gene (Liquori et al., 2001). A normal individual has less than 70 repeats, while severely affected individuals can have more than 11000 of the tetranucleotide repeats (Liquori et al., 2001). As seen in myotonic dystrophy type 1, i n situ hybridization with a fluorescent CAG probe revealed that RNA harbouring large CCUG expansions can also form discrete foci in the nucleus of DM2 cells and bind muscleblind proteins (Fardaei et al., 2002; Liquori et al., 2001; Mankodi et al., 2001). However, there is some preliminary evidence indicating that the CCUG expansion is spliced out from the mutant ZNF9 pre-mRNA and exists in the nucleus as a stable lariat (L. Ranum, unpublished data, Glasgow 2003). In whole, these results support that DM2 is caused by the same gain-of-function theory as proposed for myotonic dystrophy type I, where an RNA exerts toxicity by forming nuclear foci through the recruitment of CUG-binding proteins. Differences between these two forms of myotonic dystrophy may lie within cis or trans effects caused by the expansions in the DNA and not the RNA (L. Ranum, unpublished data, Glasgow 2003).

The great heterogeneity of symptoms and defects at the molecular, cellular and metabolical level make myotonic dystrophy an extremely complicated disease to understand. Identification of the mutation in 1992 only confirmed that understanding the basis of the disease was still a long way down the road. There is now a general consensus that multiple factors contribute to the disease. Three theories have been put forward over the past 5 years to give some explanation on how triplet expansions in the untranslated region of a gene can cause myotonic dystrophy (Tapscott, 2000; Tapscott and Thornton, 2001):

1) The expanded CTG repeats cause haplo-insufficiency of the DMPK protein by nuclear retention of its mRNA

2) Modification of the regional chromatin alters adjacent gene expression

3) Aberrant sequestration of CUG-binding proteins leads to toxic foci formation and aberrant mRNA splicing

Much of this three-level theory is based upon results generated with the three major mouse models:

1) DMPK knock-in/out mice (Jansen et al., 1996)

2) Transgenic mice expressing CTG repeats in 3’UTR of the human actin or DMPK gene (Mankodi et al., 2000; Seznec et al., 2000)

3) SIX5 knock-out mice (Klesert et al., 2000; Sarkar et al., 2000)

However, some new reports in the literature have been breaking-down these long-standing beliefs. For instance, how can the DM2 mutation, located on a completely different chromosome, generate most of the same clinical symptoms as myotonic dystrophy type I? Involvement of the CTG repeats in the disruption of neighbouring gene expression (DMPK, Six5 and DMWD) is now though to be irrelevant, or at very best of meek influence in disease physiopathology. Second, spinocerebellar ataxia type 8 (SCA8) which is caused by an expanded CTG repeat that is transcribed, generates a completely different clinical portrait as compared to myotonic dystrophy types 1 or 2 (Koob et al., 1999; Nemes et al., 2000). The answer to this riddle may reside in tissue-specific gene expression, as SCA8 is most predominantly expressed in the brain (Koob et al., 1999). Further credibility may be provided by the fact that DMPK and not ZNF9 is expressed in the brain, and only myotonic dystrophy type I patients, especially those suffering from the congenital from, exhibit strong CNS involvement and mental retardation, as seen in SCA8 (Harper, 2001; Nemes et al., 2000).

So as of late 2003, it is believed that the main driving force in generating myotonic dystrophy clinical symptoms is the toxic RNA gain-of-function theory based on sequestration and aberrant function of critical CUG-binding proteins (Mankodi and Thornton, 2002). This theory must now also be fine-tuned in specifying that the RNA containing the large CUG or CCUG repeats be expressed in smooth, skeletal and cardiac muscle to provoke the necessary disruptions to generate the typical clinical symptoms. Myotonic dystrophy is the first example of a dominant disease caused by a toxic gain-of-function mechanism of a mutant RNA.