TDP43 and RNA instability in amyotrophic lateral sclerosis

Abstract

The nuclear RNA-binding protein TDP43 is integrally involved in RNA processing. In accord with this central function, TDP43 levels are tightly regulated through a negative feedback loop, in which TDP43 recognizes its own RNA transcript, destabilizes it, and reduces new TDP43 protein production. In the neurodegenerative disorder amyotrophic lateral sclerosis (ALS), cytoplasmic mislocalization and accumulation of TDP43 disrupt autoregulation; conversely, inefficient TDP43 autoregulation can lead to cytoplasmic TDP43 deposition and subsequent neurodegeneration. Because TDP43 plays a multifaceted role in maintaining RNA metabolism, its mislocalization and accumulation interrupt several RNA processing pathways that in turn affect RNA stability and gene expression. TDP43-mediated disruption of these pathways — including alternative mRNA splicing, non-coding RNA processing, and RNA granule dynamics — may directly or indirectly contribute to ALS pathogenesis. Therefore, strategies that restore effective TDP43 autoregulation may ultimately prevent neurodegeneration in ALS and related disorders.

1. Introduction

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder in which the progressive loss of motor neurons results in paralysis and respiratory failure (Bruijn, Miller, and Cleveland 2004). There is no effective disease-modifying therapy for ALS, and its heterogeneous biochemical, genetic, and clinical features complicate the identification of therapeutic targets. However, it is increasingly clear that RNA dysregulation is a key contributor to ALS pathogenesis. Over the past decade, disease-associated mutations have been identified in genes encoding multiple RNA-binding proteins participating in all aspects of RNA processing (Kapeli, Martinez, and Yeo 2017). Among these is TDP43, a nuclear protein integrally involved in RNA metabolism. Although mutations in the gene encoding TDP43 (TARDBP) account for only a small proportion of the disease burden (2–5%), cytoplasmic TDP43 mislocalization and accumulation are observed in >90% of individuals with ALS (Neumann 2009). Moreover, mutations in several other ALS-associated genes — including C9orf72 (Murray et al. 2011), ANG (Seilhean et al. 2009), TBK1 (Van Mossevelde et al. 2016), PFN1 (Smith et al. 2015), UBQLN2 (Deng et al. 2011), VCP (J. O. Johnson et al. 2010), and hnRNPA2/B1 (Kim et al. 2013) — result in TDP43 pathology. This convergence heavily implicates TDP43 and TDP43-dependent RNA processing in neurodegenerative disease. In this review, we examine how TDP43 dysregulation impacts RNA metabolism, in particular the maintenance of RNA stability, and how these downstream events may contribute to ALS pathogenesis.

2. Mechanisms of TDP43 Autoregulation

TDP43 is an essential protein involved in several RNA processing events, including splicing, transcription, and translation. Since TDP43 recognizes UG-rich sequences present within approximately one third of all transcribed genes (Polymenidou et al. 2011; Tollervey et al. 2011; Sephton et al. 2011), it is uniquely able to influence the processing of hundreds to thousands of transcripts. In keeping with these fundamental functions, the level and localization of TDP43 are tightly regulated and critical for cell health. TDP43 knockout mice die early in embryogenesis, and partial or conditional knockout animals exhibit neurodegeneration and behavioral deficits that correlate with the neuroanatomical pattern of TDP43 ablation (Kraemer et al. 2010; Wu, Cheng, and Shen 2012; Iguchi et al. 2013; Sephton et al. 2010). Additionally, sustained TDP43 overexpression results in neurodegeneration in primary neuron (Barmada et al. 2010), mouse (Swarup et al. 2011; Wils et al. 2010), rat (Dayton et al. 2013; Tatom et al. 2009), Drosophila (Voigt et al. 2010; Y. Li et al. 2010), zebrafish (Kabashi et al. 2010; Schmid et al. 2013), and primate models (Uchida et al. 2012; Kasey L. Jackson et al. 2015), providing convincing evidence that too little or too much TDP43 is lethal.

Despite the observed sensitivity of neurons and other cell types to long-term changes in TDP43 protein levels, TDP43 expression and localization are dynamically regulated in the short-term by physical injury and other cellular stressors (Moisse et al. 2009; Swarup et al. 2012; V. E. Johnson et al. 2011). This pattern of expression suggests that TDP43 may be important for orchestrating the response to acute injury and eventual recovery. However, even relatively minor (~2-fold) persistent changes in TDP43 levels are sufficient to drive neurodegeneration (Janssens et al. 2013; Wegorzewska and Baloh 2011; Barmada et al. 2010, 2014), indicative of a coping response that over time becomes ineffective and eventually detrimental to cell health.

Similar to systems employed by related RNA-binding proteins, TDP43 regulates its own expression through an intricate negative feedback loop. At high levels, TDP43 recognizes sequences within the 3′ untranslated region (UTR) of its own transcript (the TDP43 binding region, or TDPBR) (Bhardwaj et al. 2013; Ayala et al. 2011), triggering alternative splicing within the 3′ UTR (Tollervey et al. 2011; Polymenidou et al. 2011), mRNA destabilization, and reduced protein expression (Ayala et al. 2011; Polymenidou et al. 2011; Xu et al. 2010). Two separate mechanisms may account for this destabilization.

In the first, association of TDP43 with the TDPBR induces the removal of two alternative introns (6 and 7) within the last exon of the TARDBP mRNA transcript (Polymenidou et al. 2011; Koyama et al. 2016). These splicing events create perceived exon-exon junctions (EEJs) with subsequent deposition of exon-junction complexes (EJCs), structures composed of eukaryotic initiation factor 4A-III, Magoh, Y14, UPF2 and UPF3. During the process of translation, scanning ribosomes typically displace EJCs at EEJs upstream of a stop codon. Translation is stalled when the ribosome encounters a stop codon, allowing association of the SURF complex (SMG1, UPF1, and eRF1 and 2) with the ribosome. When an EJC is present >50 nt downstream of the stop codon, factors within the EJC (i.e. UPF2) may interact with UPF1 in the SURF complex, triggering UPF1 phosphorylation and nonsense-mediated mRNA decay (NMD) (Ivanov et al. 2008; Popp and Maquat 2013). In support of this model, knockdown of UPF1 — an essential NMD factor (Popp and Maquat 2013; Sun et al. 1998; Medghalchi et al. 2001) — increased the expression of constructs carrying the TARDBP 3′ UTR, while exogenous TDP43 reduced their expression (Polymenidou et al. 2011; Barmada et al. 2015).

This mechanism of autoregulation by RNA-binding proteins is not unique to TDP43, and forms the basis for a cascade labeled regulated unproductive splicing and transcription (RUST) that is also utilized by the splicing factors PTB and SC35 (Wollerton et al. 2004; Sureau et al. 2001; Ni et al. 2007; Lareau et al. 2007; Dredge et al. 2005). Like TDP43, these proteins recognize sequences present within the 3′ UTR of their respective transcripts, resulting in splicing and EJC deposition downstream of the canonical stop codon. This, in turn, causes RNA destabilization via NMD, and an overall reduction in protein levels. An analogous mechanism is responsible for the regulation of FUS, a nuclear RNA-binding protein whose cytoplasmic mislocalization and accumulation are implicated in ALS, much like TDP43 (Lagier-Tourenne and Cleveland 2009; Lagier-Tourenne et al. 2012; Kwiatkowski et al. 2009; Vance et al. 2009). FUS and TDP43 share basic structural and functional elements, including a glycine-rich low complexity domain that harbors ALS-associated mutations. FUS also binds its own transcript, resulting in exclusion of exon 7 and a shift in the reading frame (Zhou et al. 2013). This shift uncovers a premature stop codon in exon 8, leading to destabilization of the alternatively-spliced FUS mRNA via NMD. Furthermore, disease-associated mutations in FUS (Zhou et al. 2013) and TARDBP (Koyama et al. 2016) may impair effective autoregulation of these RNA-binding proteins, resulting in their accumulation and downstream toxicity (see below).

Nevertheless, alternatively-spliced TARDBP mRNA isoforms and predicted NMD substrates have been difficult to identify and measure, and additional studies suggest that TDP43 autoregulation operates by a separate mechanism. In the second model, TDP43-mediated splicing within the TARDBP 3′ UTR removes the primary mRNA polyadenylation site (pA1) present within intron 7 (Ayala et al. 2011; Avendaño-Vázquez et al. 2012). Transcripts that utilize the remaining polyadenylation sites pA2 and pA4 are preferentially retained in the nucleus (Avendaño-Vázquez et al. 2012) and degraded by the RNA exosome (Ayala et al. 2011). Genetic ablation of exosome components Rrp6 and Rrp44 is sufficient to increase exogenous TARDBP mRNA levels and protein production, implying that the RNA exosome is indeed responsible for degrading the overexpressed TARDBP minigene. However, more recent evidence suggests that differential polyadenylation cannot fully explain TDP43 autoregulation in this model (Bembich et al. 2014). Rather, TDP43-induced splicing of intron 7 within the TARDBP 3′ UTR destabilizes the transcript, reduces nuclear export, and decreases protein production. Artificial mutations that enhance intron 7 splicing promote TARDBP destabilization, while cDNA and other transcripts that intrinsically lack intron 7 escape autoregulation and are constitutively expressed at high levels (Bembich et al. 2014). This suggests that spliceosome assembly and intron 7 splicing is a key event in TDP43 autoregulation, but whether this process participates in the regulation of endogenous TDP43 levels is unclear, and further studies are required to fully elucidate its contribution.

3. Disruption of TDP43 Autoregulation in ALS

Regardless of whether TDP43 mRNA is destabilized by NMD or degraded by the exosome following nuclear retention, interruption of this autoregulatory process likely has severe consequences for cell health. Five disease-associated mutations have been identified within the TARDBP 3′ UTR (Pesiridis, Lee, and Trojanowski 2009), which may block binding of TDP43 to its own transcript and subsequent alternative splicing. At least one of these mutations is associated with a steady-state increase in TARDBP mRNA levels, supporting the notion of disrupted autoregulation as an underlying factor leading to TDP43 accumulation and disease (Gitcho et al. 2009). The majority of ALS-associated TARDBP mutations lie within the carboxy-terminal glycine rich domain of the protein (Barmada and Finkbeiner 2010), and although the precise mechanism remains unclear, several studies have suggested that these pathogenic mutations enhance cytoplasmic TDP43 mislocalization and aggregation (Barmada et al. 2010; W. Guo et al. 2011; Mutihac et al. 2015; Barmada et al. 2014) and stabilize cytoplasmic TDP43 (Austin et al. 2014; S.-C. Ling et al. 2010). By increasing the proportion of cytoplasmic TDP43, these changes would be expected to reduce autoregulation, resulting in elevated TDP43 production. Eventually, this vicious cycle may culminate in cytoplasmic TDP43 deposition, nuclear TDP43 clearance and neurodegeneration (E. B. Lee, Lee, and Trojanowski 2011).

Mutations in genes other than TARDBP may inhibit TDP43 autoregulation by affecting nucleocytoplasmic transport. The most prevalent mutation responsible for ALS, hexanucleotide (G₄C₂) expansions in the C9orf72 gene, may block nuclear protein import through one or more related mechanisms: repeat (G₄C₂)-containing RNA may sequester essential transport factors (i.e. RanGAP) (K. Zhang et al. 2015), or dipeptide repeat proteins produced by repeat-associated non-AUG (RAN) translation of the repeat RNA might directly clog the nuclear pore (Shi et al. 2017). Additional evidence suggests that cytoplasmic protein aggregates may universally impair nucleocytoplasmic transport in neurodegenerative conditions (Woerner et al. 2016), further facilitating the cytoplasmic deposition of proteins such as TDP43 and FUS that typically participate in nucleocytoplasmic shuttling. Because splicing takes place within the nucleus, cytoplasmic retention of FUS and TDP43 would interfere with the normal autoregulation process, ultimately increasing mRNA stability and protein production. Therefore, cytoplasmic sequestration of the proteins or inefficient nuclear import would be sufficient to inhibit autoregulation, accelerating the formation of TDP43 or FUS cytoplasmic inclusions that are characteristic of ALS (E. B. Lee, Lee, and Trojanowski 2011).

4. Downstream Consequences of Failed TDP43 Autoregulation: TDP43 Nuclear Exclusion and Cytoplasmic Accumulation

Disruption of TDP43 autoregulation influences both the protein level and localization of TDP43, resulting in cytoplasmic TDP43 deposition. Given TDP43’s crucial functions in RNA processing, its dysregulation leads to abnormalities in alternative mRNA splicing, non-coding RNAs, miRNA biogenesis, and the dynamics of RNA-rich granules.

4.1 Alternative Splicing

Alternative splicing is the differential inclusion or exclusion of exons within mature transcripts, enabling the expression of multiple RNA and protein isoforms from a single gene. Between 92 and 94% of all mRNAs in the human genome are alternatively spliced (Pan et al. 2008), and the brain expresses more alternatively spliced mRNAs than any other organ (M. B. Johnson et al. 2009; Yeo et al. 2004). Because changes in the splicing environment determine which isoforms are produced (Weg-Remers et al. 2001; van der Houven van Oordt et al. 2000), alternative splicing can regulate gene expression by creating transcripts that are more or less stable. In fact, an estimated 33% of alternatively-spliced transcripts contain premature termination codons that mark them as substrates for NMD (Lewis, Green, and Brenner 2003). Thus, NMD is not simply a mechanism for degrading abnormal or mutated transcripts, but also represents an active pathway regulating the stability of alternatively-spliced transcripts. Alternative splicing therefore represents an effective and rapid means of regulating gene expression via changes in RNA stability, without the need to revert to transcription. As mentioned above, RUST is an NMD-related mechanism utilized by RNA-binding proteins to dynamically and quickly modulate their own expression; signal transduction by inflammatory cytokines likewise affects gene expression via changes in splicing and RNA stability (Shakola, Suri, and Ruggiu 2015; Paulsen et al. 2013).

In addition to regulating the splicing of its own mRNA, TDP43 is crucial for the alternative splicing of hundreds of other transcripts (Polymenidou et al. 2011; Lagier-Tourenne et al. 2012; Arnold et al. 2013; Tollervey et al. 2011). It interacts strongly with several splicing factors (Freibaum et al. 2010), and loss of TDP43 causes widespread changes in alternative splicing (Polymenidou et al. 2011; Tollervey et al. 2011) including many transcripts that are critical for neuronal viability (Polymenidou et al. 2011; Lagier-Tourenne et al. 2012; Chang et al. 2014). ALS-associated TARDBP mutations can likewise alter alternative splicing and subsequent gene expression (Arnold et al. 2013; Highley et al. 2014).

Alternatively spliced transcripts can also be targeted for decay if they include mutations that create novel splice sites. This can lead to the inclusion of unannotated or “cryptic” exons and the production of faulty transcripts. Recent work suggests that TDP43 actively suppresses unannotated exon splicing events (J. P. Ling et al. 2015; Tan et al. 2016). Its depletion results in a widespread increase in cryptic exon splicing, and the inclusion of these exons typically leads to NMD (Humphrey et al. 2017). Many of these events are specific to neurons (Jeong et al. 2017), suggesting that the disruption of TDP43-mediated cryptic exon regulation may directly contribute to neurodegeneration. Furthermore, the unannotated exons affected by TDP43 are distinct in murine and human cells (J. P. Ling et al. 2015), indicating species-specific differences in TDP43 function that may predispose to mechanisms of neurodegeneration unique to humans. TDP43 is not the only RNA-binding protein that modulates exon inclusion — RBM17, PTBP1 and PTBP2 also repress the inclusion of unannotated exons (J. P. Ling et al. 2015; Tan et al. 2016). Like TDP43, each of these factors is essential for neuronal development and their loss results in neurodegeneration (Q. Li et al. 2014; Suckale et al. 2011; J. P. Ling et al. 2016; Tan et al. 2016), implying that neurons are particularly susceptible to the abnormal inclusion of unannotated exons.

4.2 Non-Coding RNAs

Though attention is often focused on the 1–2% of transcripts that encode protein, the vast majority of the genome is transcribed as non-protein-coding RNAs (ncRNAs) (Esteller 2011). These transcripts are loosely categorized as short or long non-coding RNAs (lncRNAs), and the latter act by regulating gene expression in a variety of ways. These include, but are not limited to: the sequestration (Hung et al. 2011), competition (Kino et al. 2010), or altered localization of transcription factors (Z. Liu et al. 2011); transcriptional coactivation (Hubé et al. 2011) or corepression (X. Wang et al. 2008); alternative mRNA splicing (Annilo, Kepp, and Laan 2009); mRNA transport and stability (Gong and Maquat 2011); and modulation of translation (Ebralidze et al. 2008). lncRNAs serve crucial functions in development and disease, and also help scaffold membraneless organelles such as nuclear speckles and paraspeckles (Kung, Colognori, and Lee 2013) that are important sites of RNA processing and modification (Galganski, Urbanek, and Krzyzosiak 2017).

Both TDP43 and FUS recognize lncRNAs (Lagier-Tourenne et al. 2012; Tollervey et al. 2011; Hoell et al. 2011), including gadd7 (X. Liu et al. 2012), MALAT1 (F. Guo et al. 2015), and NEAT1_2 (Nishimoto et al. 2013), via UG-rich binding sites (Tollervey et al. 2011; Lagier-Tourenne et al. 2012). The abundance of many lncRNAs is altered in response to TDP43 knockdown in murine models of ALS (Polymenidou et al. 2011) and in human post-mortem tissue (Tollervey et al. 2011). Thus, TDP43 deposition in ALS likely has profound consequences for lncRNA expression and function. However, further studies are required to determine how TDP43 pathology influences lncRNA-related processes, and whether TDP43-mediated impairment of lncRNA contributes significantly to neurodegeneration in ALS.

4.3 miRNA Biogenesis

MicroRNAs (miRNAs) are small, non-coding RNAs that base-pair with complementary sequences within mRNA transcripts to trigger their decay and/or translational repression. These 20–25 nt RNAs are produced from an RNA precursor (pri-miRNA) that forms a hairpin loop shortly after transcription (Cai, Hagedorn, and Cullen 2004; Y. Lee et al. 2004). The enzyme Drosha then cleaves the hairpin from the rest of the transcript (Y. Lee et al. 2003; Gregory, Chendrimada, and Shiekhattar 2006), and the resulting molecule (pre-miRNA) is exported to the cytoplasm (Murchison and Hannon 2004). There, the enzyme Dicer cuts away the looped end (Lund and Dahlberg 2006), leaving a duplex of two short, complementary RNA strands. The two strands dissociate and the mature miRNA associates with the RNA-induced silencing complex (RISC), which assists in orienting the miRNA with its mRNA target, repressing translation of the target transcript and triggering its degradation.

TDP43 promotes miRNA biogenesis through a direct association with pri-miRNA, pre-miRNA, and both Drosha and Dicer (Kawahara and Mieda-Sato 2012). In so doing, TDP43 regulates the formation of key miRNAs that are essential for neuronal development, activity and survival (Fan, Chen, and Chen 2014; Buratti et al. 2010; Kawahara and Mieda-Sato 2012; Gascon and Gao 2014; Z. Zhang et al. 2013). FUS also interacts with Drosha and pri-miRNA in neurons, suggesting that it plays a similar role in neuronal miRNA biogenesis (Morlando et al. 2012).

Concordant with TDP43 pathology, the expression of several TDP43-associated miRNAs were altered in the CSF of sporadic ALS patients, compared to healthy controls (Freischmidt et al. 2013; Figueroa-Romero et al. 2016). Similar changes in miRNA levels were detected in transgenic mutant SOD1 mouse spinal cord and human ALS monocytes, but not fibroblasts from ALS patients (Butovsky et al. 2012; Koval et al. 2013). Human neurons carrying TARDBP mutations exhibited reduced levels of miR-9 and the immature pri-miRNA precursor pri-miR-9-2 (Z. Zhang et al. 2013). Knockdown of endogenous TARDBP in control neurons reproduced these deficits, suggesting that TDP43 actively participates in miR-9 biogenesis, and that disease-associated TARDBP mutations inhibit this function. One of the predicted targets of miRNAs disrupted in ALS tissues is EIF2/AGO4 (Figueroa-Romero et al. 2016), a component of RISC that participates in miRNA-mediated RNA degradation (Sasaki et al. 2003). Thus, abnormal miRNA biogenesis triggered by TDP43 dysfunction in ALS may have direct and indirect consequences for the maintenance of RNA stability. Further, since each individual miRNA can regulate the stability and translation of many downstream mRNA targets, the potential implications of even minor abnormalities in miRNA biogenesis are considerable (Paez-Colasante et al. 2015).

4.4 Stress Granule Dynamics

Cells undergo a wide range of molecular changes in response to environmental stressors, including the inhibition of conventional translation (Koritzinsky et al. 2007; Spriggs, Bushell, and Willis 2010) and the formation of stress granules (SGs), cytoplasmic ribonucleoprotein particles rich in mRNA, RNA-binding proteins, and stalled translation initiation complexes (Kedersha et al. 1999; Harding et al. 2000; Mazroui et al. 2006). TDP43 is one of several RNA-binding proteins that localize to SGs in response to various conditions (Colombrita et al. 2009; Dewey et al. 2011; Liu-Yesucevitz et al. 2010; McDonald et al. 2011). Although it is not essential for SG formation per se, changes in TDP43 levels or localization affect SG dynamics. For instance, TARDBP knockdown slows SG formation (Colombrita et al. 2009; McDonald et al. 2011), while expression of ALS-associated mutant TDP43 accelerates SG formation and results in larger SGs than wild-type TDP43 overexpression (Dewey et al. 2011; Liu-Yesucevitz et al. 2010). Based on its ability to recognize thousands of GU-rich transcripts, it is possible that excess TDP43 inclusion within SGs enables broad mRNA sequestration, shifting transcripts from actively translating polysomes to the relatively inert SGs. Conversely, SG-localized TDP43 may also bind to and prevent the degradation of RNAs that would have otherwise been degraded through association with components of processing (P)-bodies, including decapping proteins and exonucleases. Therefore, cytoplasmic TDP43 accumulation within normal or abnormal SGs in ALS might effectively increase mRNA stability without causing a reciprocal increase in mRNA translation. Nevertheless, any potential RNA stabilizing effect of TDP43 deposition is likely to be outweighed by the substantive TDP43-dependent changes in alternative splicing, unannotated exon inclusion, and miRNA biogenesis that collectively act to destabilize RNA.

4.5 RNA Transport Granules

Localized translation of mRNA is a common mechanism for regulating protein expression in specific regions of the cell. This is of particular importance in highly compartmentalized cells such as neurons, in which local translation is essential for synaptic plasticity (Klann and Dever 2004), neurotransmitter production (Mohr, Fehr, and Richter 1991), axon guidance, and recovery from injury (Willis et al. 2005). mRNAs are transported in granules comprised primarily of RNA-binding proteins (Ainger et al. 1993) that stabilize and translationally repress (Huang et al. 2003; Krichevsky and Kosik 2001) their cargo. TDP43 colocalizes with mRNA and related RNA-binding proteins in transport granules that undergo bidirectional, microtubule-dependent transport (Alami et al. 2014; Fallini, Bassell, and Rossoll 2012), suggesting that TDP43 acts as a neuronal mRNA transport factor (I.-F. Wang et al. 2008). Disease-associated TARDBP mutations impair the motility of TDP43-positive axonal granules (Alami et al. 2014), and the overexpression of TDP43 C-terminal fragments sequester components of transport granules such as HuD (Fallini, Bassell, and Rossoll 2012). Taken together with evidence showing that wild-type or mutant TDP43 overexpression impairs axon outgrowth (Fallini, Bassell, and Rossoll 2012), these observations imply that TDP43-dependent dysregulation of mRNA transport and local protein synthesis may contribute to axon degeneration in ALS.

5. Therapeutic Implications

Consistent with a potential contribution of NMD to ALS pathogenesis, modifications to this pathway have consistently demonstrated protective effects in ALS models. Components of the NMD pathway (UPF1 and UPF2) were originally identified in an unbiased screen for genes that rescue FUS-mediated toxicity in yeast (Ju et al. 2011). Subsequent studies demonstrated that UPF1 overexpression significantly prevented TDP43- and FUS-dependent cell death in rodent primary neuron ALS models (Barmada et al. 2015). As in yeast, this protective effect was also observed in response to UPF2. Moreover, the rescue was blocked by inhibition of NMD, suggesting that UPF1 and UPF2 act through NMD to improve survival. Neuroprotection by UPF1 was also confirmed in an in vivo model of ALS. Here, the expression of UPF1 in rats effectively prevented TDP43-mediated forelimb paralysis (K. L. Jackson et al. 2015). Although the mechanism by which UPF1 overexpression extends neuronal survival and improves outcomes in these models is unknown, one possibility is that UPF1 overexpression enables the cells to cope with the influx of NMD substrates elicited by TDP43 dysfunction and unannotated alternative splicing. Alternatively, UPF1 may enhance NMD-dependent degradation of the TARDBP transcript itself, facilitating TDP43 autoregulation and preventing further TDP43 deposition.

A second therapeutic approach seeks to compensate for the loss of functional TDP43 within the nucleus of affected neurons in ALS. As previously discussed, TDP43 is a repressor of cryptic exon splicing, and its depletion results in a widespread excess of unannotated splicing events. Fusion of the amino-terminus of TDP43 to the splicing repressor domain from the ribonucleoprotein Raver1 results in a chimeric protein that retains much of the function of full-length TDP43. In cells depleted of TDP43, the TDP43-Raver1 fusion is localized to the nucleus, is capable of recognizing UG repeats, and effectively blocked cryptic exon inclusion elicited by TDP43 knockout (J. P. Ling et al. 2015). Further, TDP43-Raver1 expression prevented apoptosis in these cells, suggesting that the chimera can functionally substitute for TDP43. These results further imply that cryptic exon inclusion contributes to cell death, and the introduction of a splicing repressor capable of inhibiting these events may be therapeutically advantageous in the absence of native TDP43 function.

Figure 1. TDP43 autoregulation.

TDP43 may destabilize its own mRNA transcript through two distinct mechanisms. In the first (gray arrows), TDP43 protein recognizes the TDP43 binding region (TDPBR) within the 3′ UTR of its own transcript, stimulating the removal of alternative intron 7 and the primary polyadenylation site (pA1) contained within the intron. Spliced transcripts are preferentially retained in the nucleus and targeted for exosome-mediated decay. In the second mechanism (black arrows), the removal of introns 6 and/or 7 creates exon-exon junctions (EEJs) and the assembly of exon junction complexes (EJCs). The transcript is then exported to the cytoplasm. During the first or pioneer round of translation, the ribosome pauses at the stop codon, allowing the association of the SURF complex with the ribosome. Factors within the downstream EJC interact with UPF1 in the SURF complex, triggering UPF1 phosphorylation and nonsense-mediated mRNA decay.

Figure 2. TDP43 deposition impacts RNA stability through several pathways.

(A) Alternative splicing. Mutations that introduce novel splice sites can lead to the inclusion of unannotated or “cryptic” exons (pink box). These faulty transcripts are often targeted by NMD. Typically, TDP43 is a strong repressor of these unannotated splicing events, but nuclear exclusion prevents TDP43 from performing this function, and abnormal transcripts accumulate. (B) miRNA biogenesis. TDP43 promotes several steps of miRNA biogenesis, and regulates the formation of key miRNAs that, in turn, control the stability and translation of mRNAs that are essential for neuronal survival, growth and development. (C) Stress granule dynamics. TDP43 is one of several RNA-binding proteins (blue circles) that localize to SGs (light green) in response to various conditions. Because TDP43 recognizes thousands of GU-rich transcripts, cytoplasmic TDP43 deposition within SGs forces mRNA recruitment to SGs, shifting transcripts from actively translating polysomes to inert, though stable, SGs.

Highlights.

• TDP43 is an important regulator of RNA processing whose expression is regulated through an intricate negative feedback loop.

• Cytoplasmic mislocalization and accumulation of TDP43 are characteristic features of ALS.

• TDP43 pathology influences RNA stability and gene expression via alternative mRNA splicing, non-coding RNA processing, and RNA granule dynamics.

• Targeting these pathways may enable the development of novel neuroprotective therapies for ALS.

Abbreviations

ALS

amyotrophic lateral sclerosis

3′ UTR

3′ untranslated region

TDPBR

TDP43 binding region

EEJ

exon-exon junction

EJC

exon junction complex

NMD

nonsense-mediated decay

RUST

regulated unproductive splicing and translation

RAN translation

repeat-associated non-AUG translation

ncRNA

non-coding RNA

lncRNA

long non-coding RNA

miRNA

microRNA

RISC

RNA-induced silencing complex

SG

stress granule

P-body

processing body