Alternative splicing and muscular dystrophy

Abstract

Alternative splicing of pre-mRNAs is a major contributor to proteomic diversity and to the control of gene expression in higher eukaryotic cells. For this reasons, alternative splicing is tightly regulated in different tissues and developmental stages and its disruption can lead to a wide range of human disorders. The aim of this review is to focus on the relevance of alternative splicing for muscle function and muscle disease. We begin by giving a brief overview of alternative splicing, muscle-specific gene expression and muscular dystrophy. Next, to illustrate these concepts we focus on two muscular dystrophy, myotonic muscular dystrophy and facioscapulohumeral muscular dystrophy, both associated to disruption of splicing regulation in muscle.

Alternative Pre-mRNA Splicing: A Brief Overview

In eukaryotes, most protein-coding genes are split in multiple exons interrupted by introns. A typical human protein-coding gene contains 9 exons with an average size of 145 nt while the average intron size is 3,365 nt. Considering also the 5′ and 3′ untranslated regions (UTRs), the typical human gene spans close to 27 kbp. During the splicing reaction, intron sequences are removed from the primary RNA (pre-mRNA) transcript and the exons are fused together resulting in the formation of the mature messenger RNA (mRNA). Hence, more than 90% of the original pre-mRNA sequence is removed as introns. The joined exons, with some further modifications (5′ capping and 3′ polyadenylation), are then transported into the cytoplasm to be translated into proteins.

The splicing reaction is mediated by the spliceosome, a large molecular machine composed of five small nuclear RNAs (snRNPs U1, U2, U4, U5 and U6) and more than 200 polypeptides and RNA-binding proteins (Fig. 1).¹

Figure 1.

Schematic representation of the splicing reaction. Two exons (boxes) are separated by an intron (line). The consensus sequences in metazoans at the 5′ splice site, branch point and 3′ splice site are as indicated, where n is any nucleotide, r is a purine, and y is a pyrimidine. The polypyrimidine tract is a pyrimidine-rich stretch located between the branch site and the 3′ splice site. The cross-intron assembly and disassembly cycle of the major spliceosome is showed. The stepwise interaction of the spliceosomal snRNPs (colored structured RNAs) in the removal of the intron from the pre-mRNA is depicted.

To meet the physiological requirements of cells and tissues, splicing must be both rapid and precise, a very demanding task considering that introns are typically much larger than exons and are strewn with translation termination codons, which would result in truncated proteins if not accurately removed. To achieve this accuracy, the splicing machinery must efficiently recognize the intron-exon boundaries in the pre-mRNA since just a single nucleotide addition or deletion at the site of exon joining will shift the reading frame with serious consequences for the protein-coding potential of the mRNA. Splicing fidelity is achieved through several intercommunicating features of pre-mRNAs cis-acting elements that distinguish exons from introns, direct the spliceosome to the correct nucleotides for the successive exon-exon joining, and serve as binding sites for the factors that regulate splicing.¹ All together, these elements constitute what is known to be the ‘splicing code’, which appears to be present within and around exons^2,3 and, essentially, helps to distinguish an exon from a pseudo-exon.^4,5

In addition to the basal information of the ‘splicing code’, splicing regulatory proteins (hnRNPs, SR-proteins and other splicing factors) bind to pre-mRNAs cis-regulatory elements referred to as enhancers and silencers of splicing. These short elements, conventionally classified on the basis of their position as intronic (ISE, intronic splicing enhancer and ISS, intronic splicing silencer) or exonic (ESE, exonic splicing enhancer and ESS, exonic splicing silencer), can activate or suppress splice site recognition or the assembly of the spliceosome by several mechanisms.^6-8

Even by this short introduction, it should be clear that the metabolic energy cost of splicing and of the surveillance necessary to eliminate imprecisely spliced mRNAs is very high. Hence, it was not evident initially why splicing arose and has been conserved with the evolution of complex organisms. It is now clear that the real payoff of splicing comes from the potential for alternative splicing.⁸ Alternative splicing is a mechanism by which more than one mRNA is generated from the same pre-mRNA as a consequence of differential inclusion or exclusion of exons or introns. Alternative splicing can give rise to functionally different proteins or, through alternative splicing of UTRs, can also generate mRNAs with different localization, stability, as well as efficiency of translation (Fig. 2).

Figure 2.

Major forms of alternative splicing. Exons (boxes), introns (lines) alternative and splicing process (dotted lines). (A) Constitutive splicing. (B) Alternative use of an internal cassette exon. (C) Mutually exclusive exons. (D) Intron retention. (E) Alternative 5′ splice sites. (F) Alternative 3′ splice sites. (G) Alternative polyadenylation sites. (H) Alternative promoters. (I) Alternative last exons in which one exon contains the natural stop codon while the other exon contains a premature stop codon (PTC). In many cases, these common forms can be combined to generate more complicated alternative splicing events.

Deep-sequencing surveys of gene expression in different tissues and cell lines have revealed alternative splicing as a major contributor to the generation of the increasingly complex cellular, physiological and neurological systems of higher eukaryotes.^8-11 Indeed, it appears that more than 90% of human genes generate more than one transcriptional product by alternative splicing of their pre-mRNA,^10,11 and a significant number of genes present a staggering variety of splice variants and alternative transcription start and termination sites.⁹ On the contrary, alternative splicing is far less frequent in lower eukaryotes. For example, the current estimate of alternative splicing is of about 13% in C. elegans¹² and 40% in D. melanogaster.¹³ Collectively, these observations suggest that alternative splicing is one of the main sources of proteomic diversity in multicellular eukaryotes.

Alternative Splicing and Skeletal Muscle

Alternative splicing is often modulated in a tissue- and developmental stage-specific manner and the regulation of this process is essential for diverse cellular functions in both physiological and pathological situations.

Interestingly, splicing-sensitive microarrays and deep-sequencing analysis of alternative splicing in different human tissues reported that skeletal muscle is one of the tissues with the highest number of differentially expressed alternative exons,^10,11,14 revealing a previously unappreciated degree of alternative splicing complexity in muscle-specific transcripts.

Many proteins essential for striated muscle development exist in multiple isoforms generated by alternative splicing. These include myogenic transcription factors, metabolic enzymes, and components of the myofibril.¹⁵ For example, it is known that alternatively spliced isoforms of myofibrillar protein are differentially expressed in various muscles and muscle fiber types. In addition to this complexity, it is also known that their pattern of expression varies during muscle development and is affected by neural and hormonal influences. The variable expression of myofibrillar protein isoforms is a major determinant of the differential contractile properties of individual skeletal muscle fibers.¹⁶ For example, the Cardiac Troponin T (cTNT) gene undergoes a developmentally regulated alternative splicing: cTNT exon 5 is included in the mature mRNAs in embryonic striated muscle, whereas it is skipped in the adult.¹⁷ Interestingly, the two alternative cTNT splicing isoforms confer different levels of calcium sensitivity to the myofilament, thereby affecting the contractile properties of maturing muscles.^18,19

Different families of splicing factors have been discovered to regulate alternative splicing in skeletal muscle: the CELF (CUG-BP and ETR-3-like factors) family,²⁰ the muscleblind-like (MBNL) family,²¹ the Fox family²² and Polypyrimidinetract-binding protein (PTB).²³ In addition, different cis-acting elements that were found to be both necessary and sufficient to promote exon inclusion specifically in skeletal muscle have been also identified.^24-27 Among these sequences, the muscle-specific splicing enhancer 2 (MSE2), is conserved in sequence and position in chicken and human cTNT²⁶ and it is necessary and sufficient to promote muscle-specific cTNT exon 5 inclusion in embryonic skeletal muscle cultures when present in multiple copies.²⁵ In several genes, a conserved CUG motif within MSE2 is located near alternative exons that undergo alternative splicing in striated muscle and is required for enhancer activity, revealing its role of muscle-specific splicing regulator.^25,26 These CUG motifs are bound, in a sequence-specific manner, by CELF family members.²⁰ Another cis-element involved in muscle-specific alternative splicing is the hexanucleotide UGCAUG.^14,27 This element is specifically bounds by members of the Fox family of splicing factors: Fox-1/A2BP1, Fox-2/RBM9 and Fox-3.²² A third element enriched in exons selectively upregulated in muscle is UCUCU, which corresponds to the binding site of PTB.^14,27,28 Interestingly, while CELF and Fox proteins act as positive regulators, PTB acts as a suppressor of muscle-enriched exons splicing.

Splicing as a Cause of Disease

As alternative splicing affects the majority of human genes, it is not surprising that its deregulation is often causally associated with human diseases.^2,29,30

“Splicing diseases” can be broadly classified in two groups based on where the mutations are located (Fig. 3).

Figure 3.

Examples of mutations causing splicing diseases and their possible consequences. Exons (boxes), introns (lines) and mutations (stars). (A–E) Cis-acting mutations on a specific pre-mRNA could lead to complete loss of a protein, the production of a mutant protein lacking a domain or to an unbalance in the abundance of the protein isoforms encoded by the pre-mRNA. (A) Mutations disrupting either 5′ or 3′ splice sites. (B) Mutations creating cryptic 5′ or 3′ splice sites. (C) Mutations disrupting splicing regulatory sequences such as intronic (ISS) or exonic (ESS) splicing silencers or intronic (ISE) or exonic (ESE) splicing enhancers. (D) Mutations creating new ISS, ESS, ISE or ESE. (E) Mutations affecting the secondary structure of the pre-mRNA. (F) Trans-acting mutations of a particular splicing regulator could lead to aberrant splicing of several different pre-mRNAs.

The first group, representing as many as 50% of human genetic diseases caused by point mutations, contains disorders in which mutations are located within splicing regulatory sequences.³¹ Neurofibromatosis type 1, Fraiser’s syndrome and atypical cystic fibrosis are examples of splicing diseases caused by intronic mutations.^2,30,32,33 A number of other diseases display mutations in exons. The consequence of these mutations is generally assumed to result from their effect on the protein reading frame. However, for many exon mutations it is increasingly evident that the pathogenic effect is at the level of the splicing reaction.³⁴ For example, in the case of CFTR, the gene mutated in cystic fibrosis, one-quarter of even synonymous substitutions result in altered splicing.³⁵ Interestingly, secondary pre-mRNA structure contributes to exon recognition and thus mutations affecting the secondary structure of the pre-mRNA can also cause aberrant splicing and disease (Fig. 3).³⁶ In every case, the mechanisms causing altered splicing involve disruption of cis-acting elements within the affected gene. This type of mutations can be classified either as gain-of- or loss-of-splicing mutations. Gain-of-splicing mutations occur when a splicing regulatory element is enhanced or created, while loss-of-splicing mutations weaken or destroy splicing regulatory elements. As a result, these mutations alter the pre-existing ‘exon definition’ complex leading to the inappropriate skipping or inclusion of an exon, activation of cryptic splice site or intron retention (Fig. 3).³⁷

The second group of splicing diseases is due to mutations in genes encoding for splicing regulators, either the constitutive component of the spliceosome (snRNPs and proteins) or factors that regulate alternative splicing decisions as well as associated proteins such as mRNA-export factors and transcription factors (Fig. 3).^2,32,37 Importantly, these trans-acting splicing mutations can affect the expression of multiple genes at the same time (Fig. 3). Interestingly, neurodegenerative disorders are the principal clinical manifestation of RNA-binding proteins mutations. This is somewhat surprising since the mutated RNA-binding proteins are ubiquitously expressed and it is usually explained with the high prevalence of alternative splicing in the brain. Intriguingly, recent studies have shown that trans-regulatory mechanisms contribute also to several complex diseases, including cancer.^38,39

In this review we will focus on two peculiar forms of muscle disease associated to aberrant alternative splicing: myotonic muscular dystrophy and facioscapulohumeral muscular dystrophy.

Muscular Dystrophy

Muscular dystrophy covers a group of genetically determined disorders causing progressive weakness and wasting of skeletal muscles.⁴⁰ Although the main affected tissue in muscular dystrophies is the skeletal muscle, abnormalities can also be detected in other tissues such as cardiac muscle, brain and smooth muscle. A number of criteria are used to classify muscular dystrophies, including: age of onset, rate of disease progression, mode of inheritance, distribution of muscle weakness and genetic causes. The principal types of muscular dystrophies are: congenital, Duchenne and Becker, Emery-Dreifuss, distal, facioscapulohumeral, limb-girdle, oculopharyngeal and myotonic.

If we consider the possible molecular mechanisms, muscular dystrophies can be divided in two broad categories.

The first group, represented by congenital, Duchenne and Becker, distal and limb-girdle muscular dystrophies, contains diseases caused by mutations in genes encoding for protein components of the dystrophin-glycoprotein complex (DGC), its extracellular matrix (ECM) ligands, or enzymes involved in its post-translational modifications. The DGC is a multisubunit complex present at the sarcolemma, the muscle cell membrane, which connects the extracellular matrix to the intracellular actin cytoskeleton. A mutation in one protein component of the DGC disrupts the stability of the whole complex, which in turn makes the muscle cell membrane susceptible to contraction-induced injuries causing muscle degeneration.^41,42

The second group, represented by Emery-Dreifuss, oculopharyngeal, myotonic and facioscapulohumeral muscular dystrophies, contains diseases caused by mutations affecting the activity or the expression of nuclear proteins. The molecular mechanism(s) underlying this class of muscular dystrophies are less clear. Emery-Dreifuss muscular dystrophy exists in two forms X-linked recessive (X-EMD) and autosomal dominant (AD-EMD). X-EMD is caused by mutations in the gene encoding for emerin, a protein localized to the inner nuclear membrane, whereas AD-EMD is caused by mutations in the lamin A/C gene that undergo alternative splicing to generate lamins A and C isoforms. It has been suggested that an emerin nuclear protein complex (composed of emerin, lamin A and B, and nuclear actin) exists at the nuclear envelope and that it has a dual role as (1) stabilizer of the nuclear membrane against the mechanical stresses that are generated in muscle cells during contraction and (2) in regulation of gene expression.^41,43 Oculopharyngeal muscular dystrophy (OPMD) is an autosomal dominant form of slowly progressive myopathy caused by an expansion of alanine tracts in the nuclear poly(A)-binding protein N1 (PABPN1), which, together with poly(A) polymerase, regulates the elongation of mRNA poly(A) tail length.⁴⁴ Expanded PABPN1 aggregates and forms intranuclear inclusions, which contain also mRNAs and other proteins. The mechanism through which expansions of PABPN1 polyalanine domain causes muscle demise is still unknown.

Myotonic muscular dystrophy (DM) and facioscapulohumeral muscular dystrophy (FSHD) are the two most common types of muscular dystrophy in adults. They are both characterized by progressive muscle weakness and wasting, cardiac conduction defects and, in the most severe cases, mental retardation.^40,45,46 Furthermore, cataract development has been described in DM and retinal vascular disease in FSHD.

Both diseases are inherited as dominant characters and are caused by gain-of-function mutations caused by alteration of non-coding repetitive sequences. In both cases, this is causing an epigenetic alteration of neighboring genes and aberrant alternative splicing. Finally, for both diseases the number of repeats influences the severity and the age of the onset.

In DM, a triplet expansion within non-coding region of RNAs alters the activity/expression of the splicing regulators MBNL1 and CUGBP1.⁴⁷ In FSHD, contraction of a DNA repeat alters the expression of the splicing regulator FRG1.⁴⁸

Myotonic Dystrophy

Clinical features

Myotonic dystrophy (DM) has a worldwide incidence of 1 in 8000.^32,46 Myotonic dystrophy type 1 (DM1) (OMIM 160900) result from a CTG repeat expansion in the 3′UTR of the DM protein kinase (DMPK) gene on chromosome 19q3.3.⁴⁹ It is a multisystemic disorder characterized by myotonia, progressive weakness, muscle wasting, insulin resistance, cardiac conduction defects, neuropsychiatric symptoms, gonadal atrophy, and cataracts. Beyond the adult form, which is the classic form of DM1, congenital and early-onset DM1 forms also exist.

A second form of myotonic dystrophy, DM type 2 (DM2) (OMIM 602668), results from expansion of a CCTG repeat in intron 1 of the zinc finger protein 9 (ZNF9) gene on chromosome 3q21.⁵⁰ DM2 shares with DM1 the autosomal dominant inheritance and the multisystem involvement. In contrast to DM1, however, proximal leg muscles are affected at onset. Moreover, while in DM1 muscle atrophy is present in type1 fibers, in DM2 type 2 fibers are affected.⁵¹

The rate of disease progression is slow. While longevity is not affected in many patients, overall life expectancy is reduced due to secondary respiratory disease, cardiovascular disease, and neoplasms. Unfortunately, no treatments are available to cure myotonic dystrophy: the existing therapy is limited to alleviate the symptoms but does nothing to stop or decrease disease progression.

Genetic mechanism

The pathophysiological mechanism for multisystem degeneration in DM is not fully understood. Several hypotheses have been proposed and the most accredited is derived from the evidence that CTG expansion in the 3′UTR of DMPK disrupts cellular processes at the chromatin, protein and RNA levels (Fig. 4).

Figure 4.

Model for the molecular consequences of triplet expansion in DM1. In healthy subjects, less than 40 CTG repeats are present in the 3′UTR of the DMPK gene and the DMPK and SIX5 genes are correctly expressed. In DM1 patients, CTG repeats are expanded to 80–1,000 repeats. This has a number of consequences. (A) Expanded repeats lead to epigenetic silencing of DMPK and SIX5 genes. (B) Long CUG repeats in the 3′UTR of the DMPK pre-mRNA titrate the RNA-binding protein MBNL1 leading to equivalent effects to MBNL1 loss-of-function. (C) Expanded CUG repeats cause phosphorylation and stabilization of the RNA-binding protein CUGBP1 leading to equivalent effects to CUGBP1 gain-of-function. (D) Altered activity/expression of MBNL1 and CUGBP1 is responsible for the alternative splicing changes observed in DM1. cTNT, IR and CLCN1 altered splicing is shown as example.

The initial theory was that the repeats expansion reduces the level of DMPK mRNA or protein resulting in DMPK haploinsufficiency in DM1 patients. Indeed, decreased expression of DMPK mRNA and protein in DM1 muscle was reported.⁵² To verify this hypothesis, Dmpk-knockout mice were generated but, surprisingly, they displayed only mild myopathy⁵³ and cardiac conduction defects.⁵⁴ Notably, the mice did not exhibit myotonia, which is one of the characteristic symptoms of DM1, suggesting that DMPK haploinsufficiency is not sufficient to cause DM1.

Next, it was hypothesized that the CTG repeats might influence expression of adjacent genes. The homoeodomain encoding gene SIX5 is located in this region and its mRNA level was found decreased in DM1 patients.⁵⁵ Interestingly, it has been shown that repeats expansion induces the formation of a condensed chromatin structure affecting SIX5 expression.⁵⁶ Nevertheless, it has to be noted that Six5-knockout mice develop only cataracts but no muscle pathology.⁵⁷ Therefore decreased SIX5 expression is not the primary cause of DM1.

Altogether, these evidences clearly indicate that the multisystemic features of DM1 cannot be explained solely by DMPK or SIX5 haploinsufficiency leading to the hypothesis that DM1 is mainly an RNA-mediated disease. In fact, a number of observations suggest that an RNA gain-of-function could be the underlying molecular mechanism of DM1. First, the RNA produced from the expanded CUG repeats (CUGexp) is retained in the nuclei of affected muscle tissues forming ribonuclear inclusions precluding its export to the cytoplasm for translation into DMPK protein.^58,59 Second, expression of just the DMPK 3′-UTR with expanded CTG repeats is sufficient to inhibit myogenesis in tissue culture.⁶⁰ Third, transgenic mice selectively expressing 250 CTG repeats in the 3′-UTR of the human skeletal actin (HSA^LR) develop typical features of DM1, such as myotonia and histological abnormalities,⁶¹ demonstrating that the expression of CUG repeats independently from the DMPK locus is sufficient to exhibit major features of DM. Forth, DM2 patients, resulting from an expanded CCTG repeats in the unrelated gene ZNF9, present symptoms similar to DM1. This strongly suggests that a common RNA gain-of-function is the pathogenic mechanism at the basis of both DM1 and DM2. In summary, the mutant RNA is the predominant cause of DM1 pathogenesis, although loss of DMPK and SIX5 may also play a minor role (Fig. 4).

Molecular mechanism

An important molecular feature of DM1 is the misregulation of alternative splicing. More than twenty splicing events are mis-regulated in DM1,⁴⁷ and some correlate directly with two common symptoms of the disease. For example, myotonia is due to the aberrant splicing of the skeletal muscle-specific Chloride Channel 1 (CLCN1). CLCN1 aberrant splicing produces an isoform containing premature stop codons (PTC) resulting in downregulation of CLCN1 mRNA and protein levels,^62,63 which is sufficient to cause myotonia (Fig. 4). Similarly, mis-regulated splicing of Insulin Receptor 1 (IR1) in skeletal muscle results in insulin resistance and diabetes (Fig. 4).⁶⁴ Interestingly, these pre-mRNAs are developmentally regulated and in DM1 the predominant alternative splicing isoform switches from the adult to the embryonic splicing pattern. For a review see ref. Lee, JE 2009.⁶⁵

Current evidence indicates that mutant DMPK and ZNF9 alleles alter the activities of at least two alternative splicing factors: MBNL1 and CUGBP1.^29,66

MBNL1

The expanded CUG turns into a non-correct hairpin structure⁶⁷ and this different fold resembles the binding site for different splicing factors. MBNL1 was identified in a screen for proteins that specifically bound to CUGexp repeats.⁶⁸ Interestingly, MBNL1 co-localizes to CUG-containing nuclear foci in DM1 cells and is depleted from the nucleoplasm raising the possibility that a MBNL1 loss-of-function could be at the basis of DM (Fig. 4).^69,70 Accordingly, Mbnl1 knockout mice develop myotonia, cataracts and alternative splicing defects characteristic of DM1.⁷¹ More importantly, Mbnl1 overexpression in skeletal muscle of HSA^LR mice is sufficient to rescue Clcn1 splicing and myotonia.⁷²

On the other end, Mbnl1 knockout mice do not display muscle wasting.⁷¹ Accordingly, CUGexp has MBNL1-independent effects, particularly on mRNAs encoding for extracellular matrix proteins known to have roles in other forms of muscular dystrophy.⁷³ Moreover, while MBNL1 can be sequestered to both CUG- and CAG-containing foci with similar affinity, only CUG repeats induce abnormal splicing.⁷⁴ Thus, MBNL1 loss-of-function cannot fully explain DM symptoms.

CUGBP1

Another RNA-binding protein involved in DM1 is CUGBP1.⁷⁵ Unlike MBNL1, CUGBP1 binds to single-stranded CUG repeats⁷⁵ and its binding is not proportional to the length of the RNA repeats. Moreover, CUGBP1 does not colocalize with CUGexp RNA in the nuclear foci.^69,76,77 CUGBP1 steady-state protein levels are increased in DM1 patients^26,78 and in mouse models containing expanded repeats within the DMPK 3′-UTR gene.^79,80 Apparently, the mutant RNA activates the PKC signalling pathway through an unknown mechanism, that in turn induce CUGBP1 hyperphosphorylation and stabilization (Fig. 4).⁸¹ Notably, mice overexpressing CUGBP1 reproduce histopathological, functional and splicing changes observed in DM1 mouse models and DM1 patients.⁸²

While MBNL1 nuclear levels increase during development, CUGBP1 nuclear levels decrease.^70,83 Notably, a number of developmentally regulated alternative splicing events are modulated by MBNL1 and CUGBP1 in an antagonistic fashion,⁸³ supporting the hypothesis that altered level and localization of these splicing factors are responsible for the adult to fetal alternative splicing switch observed in DM1 tissues (Fig. 4).^64,84

Finally, MBNL1 and CUGBP1 are involved in a number of other regulatory pathways beside splicing, including mRNA turnover, localization and translation suggesting that triplet expanded RNAs in DM can also impact gene expression at several levels.^85,86

Facioscapolohumeral Muscular Dystrophy

Clinical features

FSHD (OMIM 158900) is characterized by wasting and atrophy of a selected group of skeletal muscles. As its name suggests, weakness first arises in facial mimic and shoulder girdle muscles.⁸⁷ Later, the disease may progress to abdominal muscles leading to abdominal protrusion and spine lordosis.⁸⁸ In more severe cases, muscle degeneration reaches the foot dorsiflexor muscle and the pelvic girdle impairing the ability to walk. The majority of FSHD patients show the first symptoms in the second decade of life, but age of onset can vary from childhood to fifty years old.⁸⁹ In general, the earlier is the age of onset the more severe is the pathology. For example, patients that display the first symptoms during childhood will probably need wheelchair in adulthood. Interestingly, several authors^90,91 noted a mild gender effect. Males are generally more severely affected than women: women tend to be more often asymptomatic, have a slightly later onset, and a somewhat milder course of the disease.⁹¹ Several extra-muscular manifestations are presents in FSHD patients. Subclinical high tone hearing loss and retinal telangiectasias have been described as part of the disease, occurring in 75% and 60% of affected individuals, respectively.^91-94 Mental impairment (mental retardation, autism and seizure behavior) has been described in severely affected FSHD infants.^40,87 Cardiac defects and respiratory insufficiency are observed,⁹⁵ together with muscle pain⁹⁶ and fatigue.⁹⁷

FSHD displays an extremely high degree of variability in terms of onset, progression and severity of the phenotype. Although the molecular basis of this heterogeneity is not fully understood, several evidences suggest that it derives from the interplay of complex genetic and epigenetic events.⁴⁸

Currently, no therapeutic treatments to either stop progression or reverse muscle weakness and atrophy is available for FSHD.

Genetic mechanism

FSHD is an autosomal dominant disorder and it is of considerable interest for the unique nature of the molecular mutation, which is a deletion within a large, complex DNA tandem array located in 4q35.⁹⁸ The locus involved in this rearrangement consists in a series of tandemly arranged 3.3 kbp repeat sequences, named D4Z4 (Fig. 5). In more than 95% of FSHD cases, the disease is described as the loss of an integral number of repeats.⁹¹ The number of D4Z4 repeats is highly polymorphic and varies between 11 and 150 in the normal population.⁹⁹ Instead, FSHD patients present only 1 to 10 D4Z4 copies (Fig. 5).^98,100 Interestingly, the number of D4Z4 repeats is inversely correlated to the age of onset and the severity of the disease.⁸⁷

Figure 5.

Model for the alternative splicing consequences of D4Z4 repeats deletion in FSHD. In healthy subjects, more than 10 D4Z4 repeats are present at the 4q35 sub-telomere and the FRG1 gene is correctly expressed. In FSHD patients, D4Z4 repeats are deleted leaving 1-10. This leads to the epigenetic overexpression of FRG1. Increased FRG1 expression is responsible for the alternative splicing changes observed in FSHD.

D4Z4-like repeats are present on several other subtelomeric regions throughout the genome. In particular, D4Z4 repeats in chromosome 10 share 99% identity with 4q35 repeats and are equally polymorphic.¹⁰¹ Nevertheless, only contractions on chromosome 4 are pathogenic. Moreover, hybrid alleles (a mix of 4q35- and 10q26-type D4Z4 repeats) can be observed both in 4q35 and in 10q26.¹⁰² Importantly, while the contraction of 4q35-type repeats on chromosome 10 does not cause FSHD,¹⁰² contraction of 10q26-type repeats on chromosome 4 can lead to FSHD.⁹¹ These observations suggest that FSHD is not caused by alteration of the specific sequence composition of the D4Z4 repeats but rather by their structural organization in the 4q region and that 4q35 regions outside the D4Z4 repeats are relevant for the pathogenesis of FSHD.¹⁰³

Molecular mechanism

A number of evidences strongly indicate that the contraction of D4Z4 repeats causes a drastic alteration in the chromatin organization of the 4q35 region.⁴⁸ As a result, in normal individuals the FSHD locus displays a repressed chromatin conformation while in FSHD patients the region became more open and this might favor gene expression (Fig. 5). In fact, overexpression of several 4q35 genes has been reported in FSHD patients^104-107 although with some controversy.^108-110

For the purpose of this review, we will uniquely focus on the 4q35 gene FRG1.

FRG1 (FSHD Region Gene 1)

FRG1 was the first candidate gene for FSHD identified in the region nearby the D4Z4 repeats.¹¹¹ FRG1 is an interesting candidate for the pathogenesis of FSHD. FRG1 was found overexpressed in FSHD patients.¹⁰⁴ Intriguingly, D4Z4 deletion at 4q35 is associated with alteration in chromatin loops involving the FRG1 promoter.^112,113 As a result, an enhancer located in the 4q subtelomer strongly interacts with the FRG1 promoter specifically in FSHD¹¹³ possibly explaining the reported overexpression of FRG1 in FSHD.^113,114 Significantly, individuals carrying deletions of D4Z4 and FRG1 are normal, suggesting that FRG1 is required for FSHD.¹¹⁵ Accordingly, FRG1 overexpression in transgenic mice causes the development of an FSHD-phenotype.¹¹⁶ Notably, recent Xenopus results supported an FRG1 role in muscular and extra-muscular manifestations of FSHD.^117,118

While FRG1 is highly conserved from non-vertebrate to vertebrate,¹¹⁹ its aminoacid sequence does not display homology and similarities with known proteins.¹¹¹

FRG1 is a nuclear protein localized in the nucleolus, Cajal bodies and in nuclear speckles.^119,120 Interestingly, FRG1 has been repeatedly co-purified with the C-complex of the human spliceosome.^121-123 Moreover, FRG1 was identified in C. elegans and Zebrafish clusters of coregulated genes involved in ribosomal and mRNA biogenesis.^124,125 Finally, SMN and PABPN1, two proteins involved in RNA metabolism, were identified as binding partners of FRG1.¹²⁰ Intriguingly, SMN and PABPN1 are themselves involved in neuromuscular disorders.¹²⁶ Indeed, OPMD is caused by alanine expansions in the PABPN1 protein leading it to form intranuclear inclusions,^127,128 whereas Spinal Muscular Atrophy (SMA) is caused by mutations in SMN, which is involved in spliceosome biogenesis.¹²⁹

Collectively, these evidences are highly suggestive for an active role of FRG1 in alternative splicing regulation. Accordingly, aberrant alternative splicing of several genes has been described in mice and cells overexpressing FRG1, and in FSHD patients (Fig. 5).^116,130

Transgenic mice selectively overexpressing FRG1 in skeletal muscle develop a pathology with physiological, histological and molecular features analogous to FSHD patients.¹¹⁶ While all the different transgenic lines analyzed displayed myopathic changes, the severity of the muscular dystrophy was proportional to the FRG1 expression level.¹¹⁶ Aberrant alternative splicing patterns of Myotubularin Related protein 1 (MTMR1) and Troponin T type 3 (TNNT3) pre-mRNAs was present in muscle of FRG1 mice and in FSHD muscle cells and these alterations precede the onset of the muscular disease.¹¹⁶ Moreover, the extent of aberrant alternative splicing correlated with the level of FRG1 overexpression.¹¹⁶ Significantly, aberrant alternative splicing of the same genes was not detected in mice overexpressing other two 4q35 genes, such as FRG2 and ANT1.¹¹⁶ Interestingly, the alternative splicing observed in FRG1 mice is not a secondary effect of the dystrophic process, as the splicing alterations were present in 4-week old FRG1 mice (an age at which they do not display any histological and ultrastructural sign of dystrophic alterations) and more importantly they were absent in mdx mice (the mouse model for Duchenne muscular dystrophy).¹¹⁶ Moreover, the alteration of TNNT3 and MTMR1 alternative splicing was a direct and specific effect of FRG1 overexpression since it was detected in C2C12 muscle cells overexpressing FRG1.¹¹⁶ Interestingly, TNNT and MTMR1 aberrant splicing has been also described in myotonic muscular dystrophy.^63,71

Recently, Davidovic and its colleagues reported altered expression of alternative splicing isoforms derived from Fragile X Related Protein 1 (FXR1) gene in FSHD patients.¹³⁰ Moreover, reduced stability of FXR1 mRNA transcripts, leading to a reduced expression of muscle specific FXR1 isoforms, was also described. Based on this, the authors suggested that FRG1 could regulate different levels of mRNA metabolism, including splicing and stability of RNAs.¹³⁰

FXR1 belongs to the fragile X related (FXR) family of RNA-binding proteins and Fxr1 null mice display a disruption of the skeletal muscle cellular architecture.¹³¹ Moreover, complete or partial inactivation of fxr1 in Xenopus perturbs normal myogenesis.¹³² Finally, fxr1 reduction in Zebrafish causes muscular dystrophy.¹³³ Interestingly, altered FXR1 splicing was reported also in DM.⁸⁰ Hence, FXR1 splicing deregulation might be relevant to both DM and FSHD pathogenesis and could contribute to the molecular mechanisms causing these related muscular dystrophies.

Conclusion

Myotonic muscular dystrophy and facioscapulohumeral muscular dystrophy are the two most common types of muscular dystrophy in adults. They have a number of similarities. They are autosomal dominant disorders that, unlike the majority of genetic diseases, are not caused by mutations in protein coding genes but are caused by gain-of-function mutations caused by alterations of non-coding repetitive sequences. These alterations are associated to both an epigenetic alteration in the expression of the genes mapping in the disease locus and to aberrant alternative splicing of several pre-mRNAs.

While the molecular mechanism responsible for aberrant splicing in myotonic muscular dystrophy is relatively clear, the precise mechanism responsible for splicing alterations in FSHD is unknown. The current data suggest that altered splicing in FSHD is due to altered expression of FRG1 but the mechanism of action of FRG1 is largely unknown. Further work is required to investigate if FRG1 binds RNA and change directly splicing dynamics or if it regulates the activity or the expression of splicing factors.

Abbreviations

ANT1

adenine nucleotide translocator 1

A2BP1

ataxin 2-binding protein 1

CFTR

cystic fibrosis transmembrane conductance regulator

Clcn1

chloride channel 1

CUGBP1

CUG-binding protein 1

cTNT

cardiac troponin T

DGC

dystrophin-glycoprotein complex

DM

myotonic dystrophy

DMPK

DM protein kinase

ECM

extracellular matrix

EMD

emery-dreifuss muscular dystrophy

ESE

exonic splicing enhancer

ESS

exonic splicing silencer

FRG1

FSHD region gene 1

FRG2

FSHD region gene 2

FSHD

facioscapulohumeral muscular dystrophy

FXR1

fragile x related protein 1

HSA

human skeletal actin

IR1

insulin receptor 1

ISE

intronic splicing enhancer

ISS

intronic splicing silencer

MBNL1

muscleblind-like 1

MTMR1

myotubularin related protein 1

OPMD

oculopharyngeal muscular dystrophy

PABPN1

poly(A)-binding protein N1

SMA

spinal muscular atrophy

SMN

survival motor neuron

TNNT3

troponin T type 3