|Home | About | Journals | Submit | Contact Us | Français|
Mutation in TDP-43, a DNA/RNA binding protein, causes an inherited form of Amyotrophic Lateral Sclerosis (ALS). Combination of its mislocalization in most incidences of sporadic ALS (as well as other neurodegenerative disorders) with discovery of ALS-causing mutations in FUS/TLS, another DNA/RNA binding protein, has initiated a paradigm shift in understanding ALS pathogenesis. TDP-43 and FUS/TLS have striking structural and functional similarities, implicating alterations in RNA processing as key in ALS and other neurodegenerative disorders.
ALS is an adult-onset neurodegenerative disorder in which premature loss of upper and lower motor neurons leads to fatal paralysis with a typical disease course of one to five years. Most incidences of ALS are sporadic but ~10% of patients have a familial history. The era of decoding pathogenesis of ALS was initiated with the landmark discovery of dominant mutations in the copper/zinc superoxide dismutase 1 (SOD1) gene as causative of ~20% of the familial ALS forms and ~1% of sporadic cases. Subsequently, other even rarer familial cases with atypical disease features were linked to mutations in multiple other genes.
Most efforts to understand ALS pathogenesis over the last 15 years have focused on mutations in the ubiquitously expressed SOD1. No consensus has yet emerged as to how the mutations lead to selective premature death of motor neurons, except that mutant damage within motor neurons themselves drives disease onset, while mutant damage within glial neighbors accelerates disease progression (Yamanaka et al, 2008). Multiple pathways for toxicity have been implicated, including misfolded mutant SOD1 triggering aberrant mitochondrial function, endoplasmic reticulum (ER) stress, axonal transport defects, excessive production of extracellular superoxide, or oxidative damage from aberrantly secreted mutant SOD1.
From its previous focus on SOD1, thinking about pathogenesis in ALS has seismically shifted in the last year with discovery of mutation in a pair of RNA/DNA binding proteins as causative of inherited ALS (Fig. 1).
The shifting paradigm in ALS actually began with identification (Arai et al., 2006; Neumann et al., 2006) of the 43 kD TAR DNA-binding protein (TDP-43) as a major component of ubiquitinated protein aggregates found in sporadic ALS and frontotemporal lobar degeneration with ubiquitinated inclusions (FTLD-U), the most common form of frontotemporal dementia. In ALS, TDP-43 immunoreactive inclusions are observed in the cytoplasm and/or nucleus of neurons, as well as glial cells. Affected brains and spinal cords present a biochemical signature characterized by abnormal hyperphosphorylation and ubiquitination of TDP-43 and the production of ~25 kD C-terminal fragments missing the nuclear targeting domain (Arai et al., 2006; Neumann et al., 2006). TDP-43 is partially cleared from nuclei in neurons containing cytoplasmic aggregations (Fig. 2A) (Neumann et al., 2006; Van Deerlin et al, 2008) supporting that pathogenesis is driven, at least in part, by a loss of TDP-43 nuclear function. Combined with a flurry of subsequent reports, TDP-43 inclusions are now recognized to be shared by the vast majority of ALS patients, with the striking exception of familial forms caused by SOD1 mutations.
While identification of TDP-43 aggregates was the initiating breakthrough, the pathology alone left unclear if TDP-43 aggregation had a primary role in pathogenesis rather than a by-product of the disease process. Accumulation of intracellular or extracellular misfolded or misprocessed proteins in the central nervous system had previously been found in most neurodegenerative conditions. Recognizing that rare mutations in the genes encoding the misfolded proteins had been reported in Alzheimer's and Parkinson's diseases, the tauopathies and prion diseases, TARDBP, the gene encoding TDP-43 on chromosome 1, became an excellent candidate for direct sequencing in cohorts of patients with motor neuron disease and/or frontotemporal dementia.
Starting in early 2008, dominant mutations in the TARDBP gene were reported by several groups as a primary cause of ALS (see (Banks et al., 2008) for review and (Corrado et al., 2009; Daoud et al., 2009; Del Bo et al., 2009; Kuhnlein et al., 2008; Lemmens et al., 2009; Rutherford et al., 2008), collectively providing persuasive evidence that aberrant TDP-43 can directly trigger neurodegeneration. A total of thirty different mutations are now known in 22 unrelated families (~3% of familial ALS cases) and in 29 sporadic cases (~1.5% of those cases) (Fig. 1A). Since linkage of familial ALS to chromosome 1 had not previously been identified, key among these genetic efforts was a retrospective analysis in a large family where the TDP-43M337V change had been identified through direct sequencing of TDP-43. This study (Sreedharan et al., 2008) identified linkage between the disease and only one genomic region -- an 8.2 Mb region on chromosome 1p36 that contains the TARDBP gene. Although unconventional, this approach provided a strong additional support for the pathogenic effect of the TDP-43M337V mutation.
Widely expressed and predominantly nuclear, TDP-43 is a 414 amino acid protein encoded by six exons and containing two RNA recognition motifs (RRM1 and 2) and a C-terminal glycine rich region that is proposed to mediate interactions with other proteins. All but one of the mutations identified so far are localized in the C-terminal region encoded by exon 6 of TARDBP (Fig. 1). All are dominantly inherited missense changes (Daoud et al., 2009). The pathogenicity of these missense changes is strongly supported by several lines of evidence. First, they affect amino acids highly conserved through evolution. Second, they have not been found in large cohorts of controls. Third, in the cases where the DNAs are available the mutations segregate with the disease and no mutations have been found in unaffected family members (except those below the typical age of onset). This indicates a high penetrance in these families, although further studies on this point are needed since TDP-43 mutants were also identified in apparently sporadic patients. Collectively, the evidence is now overwhelming that aberrant TDP-43 can trigger ALS.
Patients with TDP-43 mutations develop typical ALS with some variability in the site and age of onset within families. While up to 50% of ALS patients develop cognitive impairment of various severities, only one of the patients with TDP-43 mutations has been reported to develop cognitive deficits (Corrado et al., 2009), despite the presence of TDP-43 inclusions not only in neurons and glial cells within the spinal cord but also throughout the brain (Banks et al., 2008), pathologic abnormalities similar that described in sporadic cases (Banks et al., 2008). Diffuse granular cytoplasmic staining (which may represent an earlier stage in inclusion development) and nuclear clearing have also been described (Fig. 2A,B). It should be emphasized that it is not known whether TDP-43 mutations lead to neurodegeneration through a gain of one or more toxic properties and/or a loss of normal function arising from its sequestration into either nuclear or cytoplasmic aggregates and the corresponding disruption in its interactions with protein partners and/or RNA targets.
Identification of TDP-43 involvement in ALS rapidly fueled a breakthrough discovery by a pair of teams early in 2009 (Kwiatkowski et al., 2009; Vance et al., 2009) of an additional causative set of mutations in the gene encoding another RNA/DNA binding protein with a pair of names, FUS (for fused in sarcoma) or TLS (translocation in liposarcoma). Previous reports had identified linkage between chromosome 16 and a familial form of ALS, but the underlying mutation was unknown. Based on the proposed function(s) of TDP-43, Vance et al. (2009) prioritized sequencing of genes within the linked region identified in a large British family so as to focus on genes encoding DNA/RNA binding proteins. This lead to identification of a dominant missense mutation (R521C) in the FUS/TLS gene. A survey of 197 familial ALS cases identified the same R521C change in four other families, as well as two additional missense mutations in four further families.
Independently, Kwiatkowski et al. (2009) used a linkage study in an ALS family originating from Cape Verde in which the disease transmission was compatible with an autosomal recessive inheritance. A region of homozygosity by descent shared by all affected members of the family was identified on chromosome 16. This region overlapped with the previously reported ALS locus and contained the FUS/TLS gene. Focusing in again on genes coding for DNA/RNA binding proteins lead to mutation screening of the FUS/TLS locus, thereby identifying a homozygous missense mutation (H517Q) in all affected members. Although three healthy siblings were also homozygous for this mutation, they were younger than the typical age of disease onset. None of the individuals heterozygous for the mutation developed ALS, confirming autosomal recessive inheritance for this mutation. Subsequent screening in 292 familial ALS cases identified 12 dominant mutations in 16 families including two large families previously linked to chromosome 16 (Kwiatkowski et al., 2009). No mutation was found in a survey of 293 sporadic ALS patients.
Combining the efforts of both teams, FUS/TLS mutations were detected in ~4% of familial ALS (0.4% of all ALS). Like the TDP-43 mutations, all patients developed classical ALS with no cognitive deficit. Except for the recessive mutation in the family of Cape Verdean origin, the inheritance pattern is dominant (albeit with an incomplete penetrance reported for the mutation R521G) (Kwiatkowski et al., 2009).
Strikingly similar to TDP-43, FUS/TLS is a 526 amino acid protein encoded by 15 exons and characterized by a N-terminal domain enriched in serine, tyrosine, glutamine and glycine residues (SYQG region), a glycine-rich region, an RNA recognition motif (RRM), multiple RGG repeats implicated in RNA binding, a C-terminal zinc finger motif and a highly conserved extreme C-terminal region (Fig. 1B). The vast majority of the ALS-linked mutations clustered in the extreme C-terminus, with mutation in all five arginine residues in this region. All mutations are missense changes except for two, both of which are located in the glycine-rich region and correspond to an insertion or a deletion of two glycines to a tract of 10 glycines.
Again like TDP-43, FUS/TLS has an almost ubiquitous expression. It is mainly localized in the nucleus but various levels of cytoplasmic accumulation have been detected in most cell types. Analysis of brain and spinal cords from patients with FUS/TLS mutations revealed normal nuclear levels in the nuclei of many neurons and glial cells, but with FUS/TLS cytoplasmic aggregations in neurons (Kwiatkowski et al., 2009; Vance et al., 2009) (Fig. 2C). It has not been reported if FUS/TLS inclusions are also present in glial cells. Cell fractionation after expression of tagged wild type or mutant FUS/TLS by transient transfection confirmed an increased cytoplasmic accumulation of the ALS-linked mutants (Kwiatkowski et al., 2009; Vance et al., 2009).
A very curious aspect of TDP-43 pathology is a partial clearance from nuclei accompanying cytoplasmic aggregates. Moreover, in a minority of neurons from patients with FUS/TLS mutations (e.g, see Fig. 2C) or cells transfected to express GFP-tagged FUS/TLSR521G, a uniquely cytoplasmic pattern has been reported. Cytoplasmic inclusions of FUS/TLS protein are absent in controls, in patients with SOD1 mutation and in sporadic ALS cases that are presumably positive for TDP-43 aggregations. Importantly, TDP-43 positive inclusions are absent in FUS/TLS mutant patients, implying that neurodegenerative processes driven by FUS/TLS mutations are independent of TDP-43 aggregation (Vance et al., 2009). It will now be essential to assess FUS/TLS accumulation and localization in patients with TDP-43 mutations, as well as in other neurodegenerative diseases, especially those with mislocalized TDP-43.
The precise roles of TDP-43 and FUS/TLS are not fully elucidated but both are multifunctional proteins that have been associated with several steps of gene expression regulation including transcription, RNA splicing, RNA transport and translation (Buratti and Baralle, 2008; Janknecht, 2005). They might also be involved in microRNA processing and FUS/TLS may play a role in the maintenance of genomic integrity. Both proteins contain RNA-binding motifs and are structurally close to the family of heterogeneous ribonucleoproteins (hnRNP). (FUS/TLS is sometimes referred as hnRNP P2.) Consistently, both TDP-43 and FUS/TLS directly bind RNA, as well as single- and double-stranded DNA.
TDP-43 was initially proposed to repress transcription by binding the TAR DNA sequence of human immunodeficiency virus type-1 (HIV-1) and the mouse SP-10 gene promoter, but little is known about the mechanisms and selectivity of transcriptional repression. FUS/TLS normal function has been more extensively studied following its identification as a fusion protein generated by chromosomal translocations in human cancers. It is a member of the TET proteins family that also includes the Ewing's sarcoma protein (EWS) and the TATA-binding protein-associated factor (TAFII68). Wild-type FUS/TLS associates with both general and more specialized factors, presumptively influencing transcription initiation. Indeed, FUS/TLS interacts with several nuclear hormone receptors and with gene-specific transcription factors. It also associates with the general transcriptional machinery, interacting with RNA polymerase II and the TFIID complex.
Recently, a very interesting and unexpected mechanism of transcription regulation was described for FUS/TLS (Wang et al., 2008). In response to DNA damage, FUS/TLS is recruited by sense and antisense non-coding RNAs transcribed in the 5′ regulatory region of the cyclin D1 gene. There it binds and inhibits CREB-binding protein and p300 histone acetyltransferase activities leading to the repression of cyclin D1 transcription (Wang et al., 2008). These properties provide a direct link between FUS/TLS RNA binding properties and a role in transcription regulation. Moreover, this kind of regulation might be more general in light of four back-to-back reports published recently in Science demonstrating that production of short sense and antisense non-coding RNAs upstream of the active transcription start sites occurs in other contexts (e.g., (Core et al., 2008)).
Beyond transcription, TDP-43 and FUS/TLS have been implicated in RNA maturation and splicing. Only a few of their respective RNA targets have yet been identified and a comprehensive map of their RNA targets is a crucial next goal. Recent technologies using high-throughput sequencing have demonstrated that a single RNA binding protein can affect many alternatively spliced transcripts (Licatalosi et al., 2008). Such approaches will be necessary to understand the role of TDP-43 and FUS/TLS in neurodegeneration. Indeed, observation of a widespread mRNA splicing defect in TDP-43 and/or FUS/TLS proteinopathies would reinforce the crucial role of splicing regulation for neuronal integrity and potentially identify candidate genes whose altered splicing is central to ALS pathogenesis. It should not be overlooked that TDP-43 and FUS/TLS may also have a role in microRNA processing -- both have been found by a mass spectrometry to associate with Drosha (Gregory et al., 2004).
Despite their nuclear enrichment and potential roles in nuclear RNA maturation, TDP-43 and FUS/TLS shuttle between nucleus and cytosol. Both have also been found in granules associated with RNA transport in neurons, with translocation of both to dendritic spines following different neuronal stimuli (e.g., (Fujii et al., 2005)). Moreover, abnormal spine morphology is observed in cultured neurons from FUS/TLS knock-out mice (Fujii et al., 2005). These results suggest that both proteins could play a role in the modulation of neuronal plasticity and/or other properties by altering mRNA transport and local translation in neurons.
The identification of causative TDP-43 and FUS/TLS mutations, along with TDP-43 pathology in most ALS cases, represent what is likely to be the beginning of a new paradigm in ALS. Needed now are to define the normal roles of TDP-43 and FUS/TLS and to determine if mutants and/or abnormal aggregation of these proteins lead to general or specific alterations of gene expression. Both cellular and animal models are now essential to define the link between TDP-43 and FUS/TLS mutations and disease.
The lessons to be learned for TDP-43 and FUS/TLS in provoking neurodegenerative disease will likely not be unique to ALS. TDP-43 proteinopathy is present in most sporadic and familial FTLD-U cases including patients with progranulin or valosin-containing protein mutations. Moreover, abnormal TDP-43 inclusions have been reported for several other neurodegenerative conditions, including ~30% of Alzheimer's patients (Banks et al., 2008). Wild-type FUS/TLS was recently identified as a major component of polyQ aggregates in cellular models for spinal cerebellar ataxia 3 and Huntington (Doi et al., 2008). The latter observation was confirmed in intranuclear inclusions in Huntington patient neurons, provoking the proposal that the protein directly binds to polyQ aggregates at an early stage (Doi et al., 2008).
Thus, discovery of involvement of TDP-43 and FUS/TLS in ALS and these other neurodegenerative diseases reinforces the role of altered RNA processing in neurodegeneration. Earlier known examples include errors in RNA metabolism from loss of survival of motor neurons (SMN) in spinal muscular atrophy and FMRP in fragile-X mental retardation. Additionally, an RNA gain of function mechanism has been implicated in a set of diseases, including the myotonic dystrophies, where a transcript with an abnormal repeat expansion alters the function and localization of alternative splicing regulators. The emerging TDP-43 and FUS/TLS stories add considerable support to the proposal that defects in RNA processing play a central role in neurodegeneration.