The pathogenic mechanism of TAR DNA-binding protein 43 (TDP-43) in amyotrophic lateral sclerosis

Abstract

The onset of amyotrophic lateral sclerosis is usually characterized by focal death of both upper and/or lower motor neurons occurring in the motor cortex, basal ganglia, brainstem, and spinal cord, and commonly involves the muscles of the upper and/or lower extremities, and the muscles of the bulbar and/or respiratory regions. However, as the disease progresses, it affects the adjacent body regions, leading to generalized muscle weakness, occasionally along with memory, cognitive, behavioral, and language impairments; respiratory dysfunction occurs at the final stage of the disease. The disease has a complicated pathophysiology and currently, only riluzole, edaravone, and phenylbutyrate/taurursodiol are licensed to treat amyotrophic lateral sclerosis in many industrialized countries. The TAR DNA-binding protein 43 inclusions are observed in 97% of those diagnosed with amyotrophic lateral sclerosis. This review provides a preliminary overview of the potential effects of TAR DNA-binding protein 43 in the pathogenesis of amyotrophic lateral sclerosis, including the abnormalities in nucleoplasmic transport, RNA function, post-translational modification, liquid-liquid phase separation, stress granules, mitochondrial dysfunction, oxidative stress, axonal transport, protein quality control system, and non-cellular autonomous functions (e.g., glial cell functions and prion-like propagation).

Introduction

Amyotrophic lateral sclerosis (ALS) is a lethal polygenic and complex neurodegenerative disease. The following are some of the risk factors for ALS: genetic mutations; history of exposure to heavy metals such as lead and mercury; airborne release of certain chemicals such as polychlorinated biphenyls (Andrew et al., 2022a), styrene, chromium, nickel, and dichloromethane (Andrew et al., 2022b); older age; male sex; smoking (although a weak risk factor for ALS in women); increased intake of glutamate (D’Amico et al., 2021), history of electrocution, and physical injury (including head trauma/injury) (Figure 1; Wang et al., 2017; Ralli et al., 2019). Among all ALS cases, only 5–10% cases have familial ALS (fALS), while the majority are sporadic ALS. An overwhelming proportion of fALS is related to genetic mutations: about 40% of fALS are caused by the repeat expansion of chromosome 9 open reading frame 72 (C9ORF72), roughly 20% are due to mutations in the copper-zinc superoxide dismutase 1 (SOD1) gene, and 4% are due to mutations in the TAR DNA binding protein (TARDPB) or RNA-binding protein fused in sarcoma (FUS) (Picher-Martel et al., 2016). These genes encode for multifunctional proteins. Thus, ALS is considered a polygenic and multifactorial disease. Research on the pathophysiology of ALS has revealed the involvement of several molecular pathways or factors leading to motor neuron degeneration, including oxidative injury, glutamate excitotoxicity, apoptosis, aberrant neurofilament function, protein misfolding and aggregation, RNA processing impairment, defective axonal transport, endosomal transport abnormalities, elevated inflammation, and mitochondrial dysfunction. Because of the involvement of multiple pathways or factors, the mechanisms underlying the pathogenesis of ALS are complex. Hence, no specific factor may sufficiently, by itself, induce neurodegeneration in ALS.

Figure 1.

The risk factors of amyotrophic lateral sclerosis (ALS).

Genetic influences, advanced age, dietary habits (increased intake of glutamate), lifestyle factors (such as smoking), overwork (repetitive/strenuous work), heavy metal exposure, and chemical exposure. Among them, genetics are understood to play a dominant role. Created with Figdraw (www.figdraw.com).

The TAR DNA-binding protein 43 (TDP-43) is an RNA-binding protein (RBP) mainly located in the nucleus and involved in many aspects of RNA metabolism. It is striking that the positive inclusion of TDP-43 has been found in a significant proportion of ALS patients (~97%) as well as in approximately 45% patients with frontotemporal lobar degeneration (FTLD) (Ling et al., 2013). In ALS/FTLD patients, the Homo sapiens TARDPB gene was discovered to encode the TDP-43 protein, demonstrating that TDP-43 plays a fundamental role in the pathogenesis of both ALS and FTLD (Kovacs et al., 2009). The TDP-43 mutation is one of the most prominent pathogenic factors associated with ALS-causing mutations of genes encoding RBP. TDP-43 binds to one-third of overall mice mRNAs and human mRNAs in the brain; however, such binding affects both the level and splicing pattern of > 20% of the above mRNAs (Da Cruz and Cleveland, 2011). One study showed that TDP-43 participates in several cytoplasmic processes besides nuclear function, which include mitochondrial functions, autophagy modulation, stress responses, translations, and stabilization and transportation of mRNA (Birsa et al., 2020). It was believed that the TDP-43 depletion in the neural nuclei of the brain and spinal cord neurons may be the characteristic pathological hallmark in the most prominent neurodegenerations including ALS (Gao et al., 2018). Pathologic TDP-43 redistribute and segregate in the nucleus of neurons, perikaryal, and neurites in the form of insoluble aggregates (Winton et al., 2008). Therefore, TDP-43 protein disorders are a group of neurodegenerative diseases that are caused by aggregation of aberrant TDP-43 (Keating et al., 2022). Research on the potential pathogenic mechanisms of TDP-43 in the pathogenesis of ALS have shown novel advances. The current potential pathogenic mechanisms of ALS associated with TDP-43 proteinopathy revealed in the models of cells and animals will be outlined and explained in this review.

Literature Review

Literature review was electronically performed using the PubMed database. The following combinations of key words were used to initially select the articles to be evaluated: pathogenic mechanism and TDP-43; TDP-43 and ALS; nucleoplasmic transport and TDP-43 and ALS; RNA function and TDP-43 and ALS; post-translational modification and TDP-43 and ALS; cleavage and TDP-43 and ALS; ubiquitination and TDP-43 and ALS; phosphorylation and TDP-43 and ALS; SUMOylation and TDP-43 and ALS; acetylation and TDP-43 and ALS; liquid-liquid phase separation and TDP-43 and ALS; stress granules and TDP-43 and ALS; mitochondrial dysfunction and TDP-43 and ALS; oxidative stress and TDP-43 and ALS; axonal transport and TDP-43 and ALS; protein quality control system and TDP-43 and ALS; non-cellular autonomous functions and TDP-43 and ALS; glial cell functions and TDP-43 and ALS; prion-like propagation and TDP-43 and ALS; Most of the selected studies (80% of all references) were published from 2008 to 2023.

Structure and Physiological Functions of TAR DNA-Binding Protein 43

According to two groundbreaking findings published in 2006, about 97% ALS patients had hyperphosphorylated and ubiquitin-positive cytoplasmic TDP-43 inclusions in contrast to approximately 1% of TDP-43 mutations (Arai et al., 2006; Neumann et al., 2006). TDP-43 is a highly conserved heterogeneous nuclear ribonucleoprotein (hnRNP). The TDP-43 structure determines its functional versatility, including mRNA stabilization, transcription, translation and splicing, axonal translocation, apoptosis, epigenetic modifications, and cryptic exon wrapping/repression (Jiang and Ngo, 2022). TDP-43 is located on chromosome 1, encodes the TDP-43 protein containing a total of 414 amino acids. The structure of TDP-43 comprises an N-terminal structural domain (NTD), two RNA recognition motifs (RRM1 and RRM2), and a C-terminal structural domain (CTD) (Figure 2; Kuo et al., 2014). The main function of NTD is the self-regulation of TDP-43 mRNA via homodimerization as well as assisting mRNA splicing (Wang et al., 2018). One of the main physiological functions of the glycine-rich CTD is to mediate the protein-protein interactions by binding to a few proteins of the hnRNP family with splicing inhibitory activity, in particular hnRNP A2/B1 and hnRNP A1 (Romano et al., 2014). The Q/N (asparagine-rich) region is referred to be the prion-like region (PrLD), which is often hyperphosphorylated and causes aggregation of aberrant TDP-43 in motor neurons. Most TDP-43 mutations related with ALS are located here. The C-terminal PrLD enables TDP-43 with the essential aggregation tendency (Monahan et al., 2016). The RRM structural domain is a distinctive character of hnRNP proteins, which directs the binding of RNA and DNA, enables splicing control and pre-mRNA regulation (Bose et al., 2008), and the physiological oligomer formation (Liu et al., 2021). Both nuclear localization signal and nuclear export signal structural domains (Darling and Shorter, 2021) control the TDP-43 transport between the nucleus and cytoplasm. Under pathological conditions, the cytoplasmic deposition of TDP-43 disrupts both the transporter protein and the nuclear pore complex (Chou et al., 2018), resulting in impaired nucleoplasmic transport (NCT). In this case, TDP-43 is unable to return to the nucleus to perform its normal function, eventually forming the TDP-43 inclusions in cytoplasm (Birsa et al., 2020). Undoubtedly, the downstream destabilizing consequences of TDP-43 pathological alterations could be caused by any change to the above structural domains or the specific function of TDP-43. It is important to note that TDP-43 protein diseases are caused by a complicated combination of the toxic gain and loss of functions. Based on the toxic gain theory, TDP-43 upregulation or mutations as well as mislocation can induce toxicity. For example, the proper expression of TDP-43 is significant in axonal growth regulation, while overall length or mutated TDP-43 overexpression leads to severely impaired axonal growth (Fallini et al., 2012). However, the function loss theory suggests that the normal function will be absent or decreased after gene deletion or mutation.

Figure 2.

TDP-43 domain structure.

Both RNA recognition motifs (RRM1 and RRM2) preferentially recognize the UG/TG-rich single-stranded or double-stranded DNA/RNA and play a variety of functions in the transcriptional inhibition, selective splicing of mRNA precursors, and translation regulation. The nuclear localization signal located in the N-terminal structural domain and the nuclear export signal located in the RRM2 structural domain regulate the localization and physiological shuttling of TDP-43 between the cytoplasm and nucleus. The C-terminal structural domain contains a PrLD with a carboxy-terminal glycine rich region. The numbers represent the positions of amino acids. Created using Microsoft PowerPoint. TDP-43: TAR DNA binding protein 43.

TAR DNA-Binding Protein 43 Pathogenic Mechanism

When the aberrant nucleocytoplasmic transport arises, the abnormal aggregation of TDP-43 occurs, leading to the dysregulation of RNA metabolism together with the loss of TDP-43 nuclear function. The TDP-43 inclusion aberrant aggregation and localization often play an important role in ALS development, and can be caused by several post-translational changes, such as cleavage, ubiquitination, phosphorylation, and small ubiquitin-related modifier (SUMO)ylation and acetylation, which promotes functional toxicity. A vicious cycle can develop between TDP-43 and its produced cytoplasmic toxicity. Eventually, the non-cell-autonomous processes might advance the ALS disease (Figure 3).

Figure 3.

Pathogenic mechanism of TDP-43.

Various mechanisms lead to TDP-43 aggregation, mislocalization, and subsequent triggering of post-translational modifications, and cytoplasmic toxicity acquisition. These form a vicious cycle. Glial cell function and prion-like intercellular transmission may contribute to disease progression of amyotrophic lateral sclerosis by non-cellular autonomous functions. Created using Microsoft PowerPoint. ROS: Reactive oxygen species; TDP-43: TAR DNA binding protein 43.

NCT dysfunction

The defect in NCT is regarded as a common etiology in a number of neurodegenerative pathologies such as ALS. The following are the rough processes of NCT: TDP-43, karyopherin-α proteins, and karyopherin-β1 (KPNB1) form a complex and enter the cell nucleus. Ras-related nuclear protein guanosine triphosphate (RanGTP) dissociates the complex and releases TDP-43 into the cytoplasm from the cell nucleus. KPNB1 associates with RanGTP and forms KPNB1-RanGTP. KPNA associates with both RanGTP and cellular apoptosis susceptibility protein (CAS, a nuclear export factor) to form KPNA-RanGTP-CAS. After KPNB1-RanGTP and KPNA-RanGTP-CAS return to the cellular cytoplasm, RanGTP is hydrolyzed to RanGDP; KPNA and KPNB1 are then separated to start another import cycle of NCT. The nucleus export of TDP-43 can be mediated by the exportins that recognize its nuclear export signal. TDP-43 can also be exported by interacting with the mRNA export complex (Liao et al., 2022). It was shown that the insoluble TDP-43 aggregates are rich within the components of NCT machinery and the nuclear pore complex. TDP-43 aggregations triggered the segregation and mislocalization of both nucleotide and transport factors, and disrupted nucleoproteins import and RNA export within the mice primary cortical neurons, the human fibroblasts and the induced pluripotent stem cells-derived neurons (Chou et al., 2018). The adult neuroblastoma cells in the mitotic arrest are exposed to the reconstituted TDP-43 amyloidogenic fibers, resulting in the depletion of nuclear TDP-43 along with the generation of cytoplasmic TDP-43 aggregations. Subsequently, the cytoplasmic TDP-43 aggregates to segregate the nuclear pore protein Nup62, while RanGAP1, Ran, and Nup107 are induced to mislocalize, leading to NCT dysfunction or even cell death (Gasset-Rosa et al., 2019). A previous research study demonstrated that disrupting the typical nuclear import paths (TDP-43 binds to nuclear protein α) in neurons leads to the cytoplasmic TDP-43 aggregation. Furthermore, the expression of the important nuclear transporter CAS is significantly reduced in FTLD-TDP patients, indicating the pathogenic effects of CAS-related NCT defects (Nishimura et al., 2010). Nevertheless, many details of the TDP-43 pathology promoting NCT defects are currently unknown in the ALS.

Dysregulation of RNA metabolism because of loss of nuclear function

Numerous evidences suggest that the loss of TDP-43 nuclear function (nuclear RNA processing function) severely disrupts the stability of transcriptomes through different mechanisms, which include splice dysregulation, pre-mRNA splicing damage, cryptic exon inhibition failure, and reverse transcription transposons activation (Hayes and Kalab, 2022). The studies of TDP-43 depleted motor neuron-like cell lines (NSC34) and fibroblasts carrying mtTDP-43 have shown that nuclear TDP-43 protein depletion may be strongly correlated with splice dysregulation (Highley et al., 2014). Supported by the results of animal model studies in mice and Drosophila, it was confirmed that the splicing inhibition is the main role of TDP-43 within motor neurons, and strengthens the view that the TDP-43-mediated splicing fidelity loss is the critical pathogenic mechanism responsible for the motor-neuron deficiency in ALS (Donde et al., 2019). The dysregulation of TDP-43 self-regulation is actually also a splicing disruption of pre-mRNA. When TDP-43 is overexpressed, the selective splicing is encouraged to produce highly unstable mRNAs that cause the nonsense-mediated decay, or lowers mRNA levels by altering the nucleocytoplasmic distribution of these mRNA transcripts. This shows that TDP-43 autoregulates its levels by a negative feedback loop (Tziortzouda et al., 2021). Remarkably, the TDP-43 that binds to pre-mRNA may have a certain position-dependence (Rot et al., 2017).

Several cryptic exons will be spliced into mRNA once TDP-43 depletes within mice embryonic stem cells, usually impairing the translations and encouraging the nonsense-mediated decay. A study showed that the cryptic splicing is widely regulated by TDP-43 (Tan et al., 2016). TDP-43 inhibits the non-conserved cryptic exons splicing and maintains intron integrity, which prevents the death of TDP-43-deficient cells. However, in ALS, the cryptic exon impairs inhibition, indicating that the splicing defects may be the basis for TDP-43 protein disease (Ling et al., 2015). UNC13A (a key gene for synaptic function) genetic variants as one of the most potent inherited risk elements of ALS/FTD (frontotemporal dementia) have gained considerable attention. Recently, it has been found that depletion of TDP-43 can induce forceful incorporation of cryptic exons within UNC13A, leading to the nonsense-mediated decay and UNC13A protein loss (Brown et al., 2022); in addition to the aforementioned findings, it is also suggested that TDP-43 may compete with the additional splicing factors to bind to the cryptic exons (Humphrey et al., 2017), which further provides hopeful therapeutic targets for ALS and even TDP-43 protein disease. The RNA processing function of TDP-43 has also been associated with neuronal disorders, as massive amount of data links the retrotransposon activation to TDP-43 mislocalization. According to the transcriptome stratification of postmortem cortical samples from ALS patients (Tam et al., 2019), one of these groups showed a high level of retrotransposons and the malfunction of TDP-43, also demonstrating that TDP-43 binds a particular subset of retrotransposons transcripts and helps to silence them in vitro, while the TDP-43 aggregations may correlate with the retrotransposons de-silencing in vivo.

Post-translational modifications

Post-translational modifications (PTMs) may cause the aberrant aggregation of TDP-43, which further result in the gain of TDP-43 cytoplasmic functional toxicity.

Cleavage

The regulation of TDP-43 at the transcriptional level is a sophisticated process, in addition to its autoregulation through pre-mRNA splicing. Both ubiquitin-proteasome system (UPS) as well as autophagy-lysosome pathways are involved in the degradation of TDP-43 (Cascella et al., 2019). Experiments have shown that caspase 3, 4, and 7 could cleave TDP-43 (Li et al., 2015), forming different C-terminal fragments (CTFs) depending on the cleavage site, such as 15, 25, and 35 kDa. These CTFs gather in the brains of ALS/FTLD patients, while they are rarely observed in the spinal cord, seemingly indicating that the CTFs of TDP-43 are probably not a precondition for neurodegenerations (Berning and Walker, 2019). The 25-kDa TDP-43 fragments are formed when the C-terminus of TDP-43 is cleaved by cystathionase, and its heterotopic expression causes the development of the toxic, insoluble, ubiquitinated, and phosphorylation-positive cytosolic inclusion bodies. This fragment is more readily phosphorylated compared to the full-length TDP-43, but the inclusion body formation or toxicity does not require phosphorylation (Zhang et al., 2009). Similarly, one study noted that the stable expression of CTF25 induced remarkable oxidative stress by establishing the cell model of CTF25 with the stable expression of TDP-43, and its neurotoxicity was dependent on the proteasome activity (Liu et al., 2014).

Ubiquitination

Ubiquitination is an ATP-dependent cascade coupling reaction, which attaches ubiquitin to the target substrates through the successive working together of enzymes E1 (ubiquitin activation), E2 (ubiquitin coupling), and E3 (ubiquitin link). The ubiquitination of TDP-43 was first identified as one of the key modification features in ALS. The sequential mutagenesis of four lysine residues to arginine in the full-length TDP-43 revealed the ubiquitination sites located mainly in the nuclear localization sequences and RNA binding regions. The ubiquitination of Lys-408 seems to block the Ser-409/410 phosphorylation, and the ubiquitination event may be associated with the phosphorylation of TDP-43 (Hans et al., 2018). In 2020, it was also found that Praja 1 ring-finger E3 ubiquitin ligase (PJA1) has a significant inhibitory effect on the phosphorylation and aggregation of pathogenic cytosolic TDP-43 CTFs both in vitro and in vivo, although it is unknown whether PJA1 has the ability of ubiquitinating TDP-43. Furthermore, PJA1 was also identified to interact with the ubiquitin-conjugating enzyme E2 E3 (UBE2E3), an enzyme for ubiquitin conjugation, which implies that PJA1/UBE2E3 may trigger TDP-43 ubiquitination in response to the pathogenic events (Watabe et al., 2020). At present, it is reported that PJA1 is the prevalent sensor of multiple easily aggregated proteins, such as SOD1 and α-synuclein. PJA1 counteracts their tendency to aggregate, which provides ideas for possible treatment targets in neurodegenerative diseases such as ALS (Watabe et al., 2022). The cell biological analysis for the overall-length and cut-off forms of Ala315→Thr (A315T) mutants of ALS-related TDP-43 within the Neuro2a cell line suggested that the multiple ubiquitination of TDP-43, including the linear ubiquitin generated specially by the linear ubiquitin chain assembly complexes, affects protein aggregation and reaction to inflammation in vitro (Zhang et al., 2022).

Phosphorylation

In almost all ALS patients and a subset of FLTD patients, the insoluble inclusion bodies of phosphorylated TDP-43 are found in the lesioned neurons. These phosphorylated events occur most characteristically on serine 409 and 410 (S409/410), and most of the sites where TDP-43 is phosphorylated are located in the glycine-rich CTD (Kametani et al., 2016), but it has been suggested that the phosphomimetic substitution at S43 of the TDP-43 NTD damages the assemblies of TDP-43 polymers, prevents the liquid-liquid phase separation (LLPS), and disrupts the RNA splicing activity (Wang et al., 2018). Not coincidentally, MAPK/ERK kinase (MEK), as a kinase of mitogen-activated protein kinase/extracellular-signal-regulated kinase (MAPK/ERK) signal pathway, is activated by the heat shock response. This response significantly increases the TDP-43 bisphosphorylation by MEK at threonine 155 and tyrosine 153 (p-T155/Y153), and the activated MEK remarkably changes the splicing mediated by TDP-43. Moreover, the mimetic phosphosubstitution of these two residues reduces binding to the GU-rich RNA (Li et al., 2017). Apart from these, the TDP-43 phosphorylation affects the subcellular localization, increasing nucleolar (Li et al., 2017) and mitochondrial localization (Wang et al., 2016). Phosphorylation can form toxic TDP-43 oligomers and lead to increased toxicity of this oligomer (Choksi et al., 2014). However, a study published last year suggested that the hyperphosphorylation of TDP-43 mediated by casein kinase 1δ or C-terminal phosphomimic mutation reduced phase separation and TDP-43 aggregation (Gruijs da Silva et al., 2022). Notably, Eck et al. (2021) also suggested that the phosphorylation of TDP-43 is not related to localization, rather to changes in the dissolution and aggregation, and the phosphorylation of S409/410 occurs after the formation of aggregates.

SUMOylation

SUMOylation was previously associated with the formation of SOD1 aggregation positivity (Chang, 2022). Its coupling is mediated by the enzymes E1, E2, and E3. Its de-coupling is catalyzed by the deSUMOylation enzyme. SUMOylation can regulate various processes such as targeting the lysine residue, which is similar to a reversible post-translational modification—ubiquitination. SUMOylation was previously thought to promote the aggregation of TDP-43 (Foran et al., 2013), and whether SUMOylation occurs in TDP-43 itself is yet unclear. The SUMOylation of TDP-43 regulates the splicing function and the nucleoplasmic distribution of TDP-43. SUMOylation was shown to modify the splicing activities of TDP-43 within the exon by using SUMO mutant TDP-43 K136R protein, which affects both the subcellular location and the emergency granule aggregation following oxidative stress. It was also found that the de-SUMOylation of cell-permeable SENP1 TS-1 peptide facilitates the cytoplasmic localization of TDP-43 and increases the C-terminal of TDP-43 fragment p35 in small cytoplasmic aggregates (Maraschi et al., 2021). Maurel et al. (2020) found that the mutations in the SUMO site modified the intracellular localization of TDP-43 aggregates by targeted mutagenesis. This change in localization resulted in improving the cell viability and the overall cell function. This study also demonstrated that SUMOylation inhibited by anacardic acid also significantly reduced these TDP-43 aggregations.

Acetylation

Acetylation is one of the critical PTMs involved in regulating some proteins. Many physiological activities such as DNA transcription, RNA modification, protein homeostasis, autophagy, and cytoskeletal structure regulation and metabolism are controlled by acetylation. The activities of histone acetyltransferase and histone deacetylase help to maintain the equilibrium state of acetylation (Kabir et al., 2022). Among 20 lysine residues of TDP-43, Lys-145 at RRM1 and Lys-192 at RRM2 are most likely acetylated (Cohen et al., 2015). One study found that the acetylation of the K84 site within the TDP-43 RNA-binding domain reduced the nuclear input, while K136 acetylation damaged the ability of TDP-43 to both bind and splice RNA, subsequently triggering the TDP-43 phase separations mediated by low-complexity CTD, resulting in the formation of insoluble aggregates with pathological hyperphosphorylated or ubiquitinated TDP-43. These above opinions were confirmed by introducing acetyl lysine into the defined sites through amber suppression, and the study generated the relevant selective antibodies and found that sirtuin-1 effectively deacetylated ACK136 TDP-43 and reduced its aggregation tendency (Garcia Morato et al., 2022). A study published in 2015 suggested that the TDP-43 acetylation enhanced the formation of aggregates, and acetylated mimics impaired the RNA regulating function of TDP-43. The finding of acetylated TDP-43 lesions in the spinal cords of ALS patients suggests that abnormal acetylation is a pathological feature associated with the full-length phosphorylated TDP-43 lesions in the spinal cord of ALS patients, and that deacetylation is also associated with the impairment of axonal transport, which is thought to be the critical pathomechanism in ALS (Cohen et al., 2015). Furthermore, the TDP-43 acetylation abnormalities and the RNA binding deficiencies may be related to the pathogenesis of ALS (Kabir et al., 2022).

Gain of cytoplasmic functional toxicity

LLPS

LLPS, the short-term and often reversible phase-shift, divides two compartments of liquids with different viscosities and/or compositions. Low-complexity domain induces LLPS, and it is in charge of forming membrane-less organelles such as P-particles, stress granules (SG), and Cajal bodies, and predisposes protein to prion-like behavior as they tend to shape the amyloid-like structure of proteopathy (Farina et al., 2021). Because the α-helix segment at the median section of CTD (spanning about 20 residues) tend to be self-assembled, TDP-43 LLPS may require only a few important residues, such as three tryptophan residues and four other aromatic residues. The number of motifs required to form the multivalent linkages is reduced (Li et al., 2018). Additionally, the dimerizations and oligomerizations of NTD are related to LLPS (Wang et al., 2018). The mutations in progranulin (PGRN) lead to haploinsufficiency and reduce the functions of PGRN, which is closely associated with ALS/FTLD. Researchers discovered that GRN-5 mediates LLPS in vitro, which is a particulate protein processed by the protein hydrolysis of PGRN (Bhopatkar et al., 2020). LLPS is likely the intermediate step between soluble and aggregated TDP-43, and it could further form SG. LLPS has a straightforward effect on promoting aggregations of TDP-43 low-complexity domain, possibly by offering a setting that increases the partial concentration of proteins (Babinchak et al., 2019). However, the faulty LLPS processes might inhibit nucleoplasmic transport and induce abnormal mislocalization of nuclear TDP-43 (Tamaki and Urushitani, 2022).

SG

SG quickly integrate into the cellular foci as the cell is exposed under stress. The mRNA-protein assembly forms by the untranslated mRNA, where it is dynamically assembled and disassembled by the phase separation depending on the stress conditions and helps cells respond to stress (Zhang et al., 2021). The phosphorylation of the alpha-subunit of eukaryotic initiation factor-2 inhibits the protein translation by reducing the level of complex eIF2-GTP-tRNAmet, whose phosphorylation leads to SG assembly (Kedersha and Anderson, 2002). Under pathological situations, when the LC structural domain of TDP-43 no longer mediates the liquid-liquid separation but the liquid-solid phase transition, it solidifies mRNP particles such as SG and transit particles, thus disrupting the mRNP particle dynamics, resulting in alterations to mRNA processing as well as RBP aggregate formation (Bowden and Dormann, 2016). SGs are also associated with NCT, as stress interferes with NCT by targeting the key transport factor to SG and inhibits SG assembly. For example, knocking down ataxin-2 (a polyglutamine protein essential for SGs assembly) inhibits the C9ORF72-mediated nucleocytoplasmic transport defects as well as neurodegeneration in ALS/FTD (Zhang et al., 2018), while the ablation of ataxin-2 has been shown to reduce TDP-43 proteopathy and neurotoxicity (Becker et al., 2017), supporting the idea that SGs play an important role in forming the TDP-43 inclusions. An alternative therapy of ALS that targets ataxin-2 has been proposed by using the ataxin-2 knockout mouse crossed with the TDP-43 transgenic mouse and the antisense oligonucleotide which targets ataxin-2 to treat neurodegeneration, both of which have significantly prolonged the survival of the ataxin-2 knockout mouse (Becker et al., 2017).

Mitochondrial dysfunction

Overwhelming evidence suggests that mitochondrial dysfunction has a significant impact on ALS development. Both TDP-43 overexpression and inhibition impair mitochondrial motility (Wang et al., 2013). As previously mentioned, mitochondrial cystathionase cleaves TDP-43 to 25 and 35 kDa fragments related to ALS/FTLD, thereby connecting the mitochondrial defect to TDP-43 proteinopathy. TDP-43 can regulate the mitochondrial mRNA transcripts by binding microRNAs (miRNAs); regulating the mitochondrial mRNA; and stabilizing the electron transfer chains, tRNA, and multiple cis-trans RNA transcripts of mitochondrial ribosomal RNA. When the mitochondrial proteins are imbalanced, the additional reactive oxygen species and TDP-43 cleavage will be generated, thus leading to subsequent TDP-43 aggregation (Wood et al., 2021). A previous study on brain samples, and both cellular and animal models of TDP-43 proteinopathy reported that the electron microscopy analysis showed a significant mitochondrial damage, including cristae abnormalities and cristae deletion, suggesting that the TDP-43 expression inhibits the activity of mitochondrial complex I, reduces ATP production in the mitochondria, and activates the mitochondrial unfolded protein response within cells and animals models (Wang et al., 2019). It has been revealed that TDP-43 can accumulate in the mitochondria of neurons from ALS/FTLD patients, and the suppressing TDP-43 mitochondrial localization eliminates the mitochondrial dysfunction and the neuron loss caused by both WT and mutant TDP-43, even improved the phenotype of transgenic TDP-43 mouse (Wang et al., 2016). Davis et al. (2018) suggested that TDP-43 increases the mitochondrial autophagy. Prohibition-2 is a key receptor for mitosis, and its deletion leads to mitochondrial instability. In the presence of the mitochondrial uncoupler and the autophagy inducer carbonyl cyanide m-chlorophenyl hydrazone, TDP-43 knockout caused 20% reduced expression of prohibition-2 protein; in addition, they suggested that TDP-43 might contribute to the age-dependent alterations within the mitochondrial dynamics by regulating mitofusin 2 protein (Davis et al., 2018).

Oxidative stress

Oxidative stress is present in all aerobic cells, and motor neurons are more susceptible to oxidative stress damage (Requejo-Aguilar, 2023; Wakatsuki and Araki, 2023). It is well known that reactive oxygen species are produced as a result of aerobic metabolism, which can injure cells. Oxidative stress is defined as a breakdown in the equilibrium between the production and clearance of reactive nitrogen species and reactive oxygen species. The TDP-43 aggregation in neurons from mice and humans may result in increasing the susceptibility to oxidative stress, which has been reported (Zuo et al., 2021). TDP-43 aggregation chelates the specialized miRNA and protein, causing dysfunction of mitochondrial proteins encoded by certain nuclear genes, leading to widespread mitochondrial imbalance, and thus increase oxidative stress. Studies have shown that DJ-1 overexpression reduces the paraquat-induced oxidative stress by inhibiting TDP-43 aggregation (Lei et al., 2018). Similarly, oxidative stress factors like H2O2, arsenic, and heat shock proteins can cause insoluble cross-linked TDP-43 variations (Cohen et al., 2015). Furthermore, it is possible that oxidative stress promotes ALS pathogenesis by triggering the generation of SG, which along with phosphorylated TDP-43 aggregates may reappear in patient-derived cells exposed to chronic oxidative stress (Ratti et al., 2020). Ethanoic acid can increase oxidative stress in cells through glutathione consumption, and leads to phosphorylation in the CTD of TDP-43, which suggests that oxidative stress may also be related to pathological TDP-43 modifications (Iguchi et al., 2012).

Impaired axonal transport

The transportation of essential cargoes over long distances between proximal and distal compartments of neurons is ensured through axonal transport, which is necessary for neuronal development and survival. A current hypothesis states that the selective degeneration of motor neurons in ALS patients is related to the axonal transport defect. It has been observed that the motor neuron with the long axon degenerates first, and the functional changes in the most distal sites occur at the initial stage of illness (Fischer et al., 2004). A previous study on the mitochondrial movement along neuropil found a remarkable reduction in the proportion and number of mobile mitochondria compared to controls within the mutant TDP-43 motor neurons. Notably, this study verified that the histone deacetylase 6 inhibitors may have the potential to repair the axonal transport deficits in motor nerve cells generated from patients with FUS-ALS (Guo et al., 2017), as well as restore the pathology of TDP-43 and the axonal transport defect in the motor neurons of patients with ALS carrying mutations in TDP-43 (Fazal et al., 2021). One study suggested that TDP-43 is likely to form cytoplasmic mRNP particles, participate in the bidirectional and microtubule-dependent transportation of neurons, both in vitro and in vivo, thereby facilitating target mRNA delivery to the distant neuronal compartments. The study on motor neurons generated from stem cells provides evidence that this axonal transportation is impaired by TDP-43 mutation (Alami et al., 2014). Furthermore, TDP-43 has been repeatedly identified to be related to axonal growth (Tripathi et al., 2014), and is also inextricably linked to axonal transport. The impaired axonal transport is more likely to be an upstream pathogenic mechanism of ALS.

Protein quality control system

The endoplasmic reticulum (ER) participates in protein syntheses, protein folding, and post-translational transportation in vivo (Bernard-Marissal et al., 2015) to ensure correct protein folding. The protein-folding ability is decreased by disruptions in ER-calcium homeostasis, irregularities in protein homeostasis, and the lack of oxygen, which results in ER stress (Lin et al., 2008). When ER stress is physiologically triggered, the unfolded protein response is initiated by the ER to reduce the misfolded protein and re-establish proteostasis. When ER stress is serious and sustained, the unfolded protein response will trigger the chain of reactions that leads to apoptosis (Liu et al., 2002). Thus, unfolded protein response is a key proteostasis pathway that responds to ER stress-caused protein aggregations by initiating the pro-adaptation or pro-apoptotic pathway. TDP-43 (A315T mutation) enhances the toxicities of neurons via triggering apoptosis and autophagy induced by ER stress in SH-SY5Y cell lines (Wang et al., 2015). SecinH3 was shown to effectively reduce the neurotoxicity caused by TDP-43 Q331K mutation through the suppression of ER stress-mediated apoptosis and the augmentation of autophagic flux (Hu et al., 2019). Animal model studies have shown that four compounds, namely methylene blue, salubrinal, guanabenz, and phenazine, are efficient mTDP-43 toxicity inhibitors, reducing palsy, nerve degeneration, and oxidative stress within the mTDP-43 model by reducing ER stress response, achieving a potential neuroprotective effect (Vaccaro et al., 2013). This suggests that ER stress is inextricably linked to TDP-43-induced ALS.

In addition to the role of ER, both UPS and autophagy (e.g., chaperone-mediated autophagy and macroautophagy) can degrade the majority of misfolded as well as accumulated proteins to maintain proteostasis. The ER cannot handle wrongly folded proteins, but these misfolded proteins will be transported to UPS for degradation. The very striking pathological feature of TDP-43 proteinopathy is the existence of intracellular ubiquitin-positive inclusion within neurons, suggesting that the cytosolic TDP-43 damages the UPS activities; this UPS damage might have a key influence on TDP-43 proteopathy. Lee et al. (2020) proposed that protein tyrosine kinase 2 may play a vital role in ubiquitinated aggregate deposition and UPS injury-induced neurotoxicity in TDP-43 proteinopathy by regulating the TANK binding kinase 1 (TBK1)-SQSTM1 pathway in their study of both a mammalian cell model and Drosophila TDP-43 proteopathy model (Lee et al., 2020). Dysfunctional autophagy is frequently observed in neurodegenerative diseases, and it has been suggested that TDP-43 toxicities in yeast occur partly because of autophagy suppression. The deletion of poly(A)-binding protein 1 and Tap42p-interacting protein could decrease the toxicities of TDP-43 through inhibiting TDP-43 to suppress the process of autophagy (Park et al., 2022). TBK1 ALS-related mutant proteins also relate to the deficient phosphorylation of p62 and reduced autophagic degradations of TDP-43, which indicates that TBK1 is likely one of the new regulators for TDP-43 level (Foster et al., 2020). Recently, it was found that thyroid transporter (TTR) cleared the cytoplasmic TDP-43 inclusion by the ATF4-induced autophagy up-regulation. The TTR expression in FTLD-TDP mouse neurons restores the function of autophagy, promotes the autophagy of initial soluble TDP-43 aggregates, and improves neuropathological and behavioral defects, which can provide certain hints for further study on the pathogenesis and treatment about ALS (Chu et al., 2023).

Glial cell functions

Although glial cells are not neurons that participate in signal transmission, their involvement in neurological diseases cannot be ignored, as they provide the structural and nutritional support for neurons. Glial cells are divided into macroglias (e.g., astrocyte, ependymal cell, and oligodendrocyte) and phagocytic microglias. Misfolded TDP-43 can be identified and cleared through von Hippel-Lindau protein/cullin-2 (Uchida et al., 2016). Moreover, especially in the spinal cord of ALS, the cytoplasmic inclusion bodies of oligodendrocytes showed immune reactivity of both TDP-43 phosphorylation and von Hippel-Lindau protein, suggesting that oligodendrocyte dysfunction is the potential cause of ALS (Uchida et al., 2016). Studies using virus-mediated and transgenic mice models have shown that the deletion of triggering receptor expressed on myeloid cells-2 (TREM2) affects the microglial phagocytosis of pathologic TDP-43, which may explain the higher inclusion levels of glial cells in TDP-43 protein disease, causing increased neuronal damage. Combined with the analysis of mass spectrometry and surface plasma resonance, the data emphasize that TDP-43 may be the ligand for TREM2 and such associations might mediate neurological protection of microglias in TDP-43-linked neurodegenerative diseases such as ALS (Xie et al., 2022). Moreover, the TREM2-apolipoprotein E pathway restores the homeostastic characteristics of microglias in the mice models of ALS and prevents neuron loss (Krasemann et al., 2017). In vitro studies on the induced pluripotent stem cell-derived astrocytes from TDP-43 mutant patients and mice have found that the excessive astro-cell-derived polyps are the key factors of non-cell-autonomous motor neuron degeneration (Arredondo et al., 2022).

Prion-like propagation

A series of extremely contagious neurodegeneration disorders known as prion diseases are brought on by the misfolding, aggregating, and spreading of prion protein structure. Furukawa et al. (2011) were the first to show that TDP-43 had the pathogenic prion-like features in cultured cells by inoculating them with fibrils produced from the recombinant TDP-43 (Furukawa et al., 2011). It has been proposed that ALS may be a type of prion-like disorder. After TDP-43 aggregation in vitro, TDP-43 showed the classical characteristics of insolubility, phosphorylation, self-continuous reproduction, and aggregation spread among cells in a prion-like manner (Smethurst et al., 2016). Nonaka et al. (2013) not only provided strong evidence that the pathologically aggregated TDP-43 can multiply in a prion-like manner, but they were also further identified in the exosomal transport as a possible pathway for the intercellular transmission of pathologically prion-like TDP-43. Based on the existence of CTFs in the neuron inclusion described above, TDP-43 apparently shows high potential for prion-like behavior (Grad et al., 2015). The TDP-43-positive helicoid inclusion with amyloidosis features have been detected within the ALS spinal cord sample (Robinson et al., 2013). Recently, a study used induced pluripotent stem cells to generate brain-like organs, after injecting the postmortem spinal cord protein extracts from sporadic ALS individuals into the brain-like organs and showed the generation of TDP-43 pathologies and prion-like propagations in human central nervous system tissues (Tamaki et al., 2023). However, there are some questions regarding the significance of prion-like mechanisms of transmission and spread in ALS. For example, no transmission of disease between individuals has been documented in ALS cases; whereas, such observations have been reported in Alzheimer’s disease, which also has prion-like transmission characteristics (Gosset et al., 2022).

Limitations

This review has some limitations. First, our review only covers the main pathogenic mechanisms of TDP-43 in ALS; other poorly studied mechanisms were not analyzed. Second, specific signaling pathways in the pathogenesis were not demonstrated, and more detailed mechanisms need to be explored further. Third, the article is limited to English-language publications or translations, and thus may lack relevant international data published in other languages.

Conclusions

In this review, new advances in the TDP-43 pathogenic mechanism associated with ALS over the last few years are described, which are primarily associated with the mislocalization of TDP-43 and involve various aspects of cellular processes. In general, nuclear dysfunction, cytoplasmic toxicities, and disturbance of cytoplasmic transport lead to abnormal TDP-43 aggregations within the cytoplasm, further causing disease progression by the prion-like transmission. TDP-43 unquestionably mediates, at least in part, the pathogenesis of various neurodegenerations including ALS. Although the treatment targeting TDP-43 aggregates for ALS progression to achieve significant therapeutic benefits likely requires further research (Suk and Rousseaux, 2020), this approach can be potentially beneficial as it can slow or prevent disease progression in ALS. Stathmin-2 (STMN2) is a protein that regulates the development of axons. TDP-43 regulates its mRNA splicing. One study concluded that the decrease of STMN2 caused by the TDP-43 pathology is one of the etiological factors for ALS. It has been noted that increasing STMN2 levels can be a therapeutic approach for ALS; therefore, antisense oligonucleotide targeting STMN2 splicing is currently being developed. It is also critical to determine whether STMN2 homologs can be employed therapeutically to clinically treat ALS (Krus et al., 2022). At present, numerous potentially valuable biomarkers has been evaluated: PTMs such as the ubiquitination and phosphorylation of TDP-43 can be used as biomarkers of ALS (Raghunathan et al., 2022). Additionally, neurofilaments (Benatar et al., 2020), polyphosphate (Arredondo et al., 2022), and chemokine (C-X-C motif) ligand 13 in the cerebrospinal fluid (Trolese et al., 2020) also have biomarker potential. Moreover, the impact of the environment on ALS is also crucial, as it remains a risk factor that can be easily changed (Al-Chalabi and Hardiman, 2013). Early diagnosis, early control of high-risk factors, and early treatment with medications to delay the course of ALS are clinically useful approaches towards ALS treatment. Further clinical trials are needed to assess the clinical application prospective of therapeutic targets of TDP-43 to transform the discovered fundamental research outcomes into the clinical practice.

Funding Statement

Funding: This work was in part supported by the National Natural Science Foundation of China, Nos. 30560042, 81160161, 81360198, and 82160255; Education Department of Jiangxi Province, Nos. GJJ13198 and GJJ170021, Jiangxi Provincial Department of Science and Technology, No. 20192BAB205043; and Health and Family Planning Commission of Jiangxi Province, Nos. 20181019 and 202210002 (all to RX).