TDP-43 as a therapeutic target in neurodegenerative diseases: Focusing on motor neuron disease and frontotemporal dementia

A common feature of adult-onset neurodegenerative diseases is the presence of characteristic pathological accumulations of specific proteins. These pathological protein depositions can vary in their protein composition, cell-type distribution, and intracellular (or extracellular) location. For example, abnormal cytoplasmic protein deposits which consist of the TDP-43 protein are found within motor neurons in patients with amyotrophic lateral sclerosis (ALS, a common form of motor neuron disease) and frontotemporal dementia (FTD). The presence of these insoluble intracellular TDP-43 inclusions suggests that restoring TDP-43 homeostasis represents a potential therapeutical strategy, which has been demonstrated in alleviating neurodegenerative symptoms in cell and animal models of ALS/FTD. We have reviewed the mechanisms that lead to disrupted TDP-43 homeostasis and discussed how small molecule-based therapies could be applied in modulating these mechanisms. This review covers recent advancements and challenges in small molecule-based therapies that could be used to clear pathological forms of TDP-43 through various protein homeostasis mechanisms and advance the way towards finding effective therapeutical drug discoveries for neurodegenerative diseases characterized by TDP-43 proteinopathies, especially ALS and FTD. We also consider the wider insight of these therapeutic strategies for other neurodegenerative


Introduction
Motor Neuron Disease (MND) is the umbrella term used for a group of neurodegenerative conditions including amyotrophic lateral sclerosis (ALS), progressive bulbar palsy (PBP), primary lateral sclerosis (PLS) and progressive muscular atrophy (PMA), which lead to the progressive deterioration of the motor neurons that control voluntary muscles of the body.ALS is the most common form of MND, characterized by the loss of lower and upper motor neurons (LMN and UMN, respectively) (Li et al., 2015).Frontotemporal dementia (FTD) is a type of dementia with the loss of cortical neurons in the frontal and temporal cortices (Nguyen et al., 2018).ALS and FTD share a broad spectrum of overlap in genetic mutations, clinical symptoms, and pathological proteins (Ling et al., 2013), and the susceptibility of neurons and wider disease mechanisms in ALS/FTD are well documented in the literature (Braak et al., 2013;Ragagnin et al., 2019).The main pathological hallmark in ALS and almost half of the FTD cases are proteinopathies of transactive response DNA/RNA-binding protein with 43 kDa in molecular weight , which in some cases, carries one or more pathogenic ALS-relevant mutations.Additionally, proteolytic TDP-43 cleavage generates C-terminal fragments of 35 and 25 kDa size respectively), which have a higher propensity for aggregation (in comparison to wild type), and sequestration within cytoplasmic inclusions, collectively suggesting that these truncated forms of TDP-43 are linked with TDP-43 proteinopathies (Fig. 1) (Che et al., 2015;Cicardi et al., 2018;Smethurst et al., 2020).
TDP-43 was initially identified in 1995 as a protein that binds to pyrimidine-rich motifs of TAR DNA sequence in the HIV type 1 gene and modulates the HIV-1 gene expression (Ou et al., 1995).However, for the first time in 2006, it was reported that TDP-43 is the main protein component of ubiquitinated and insoluble protein inclusions (except SOD1 and FUS inclusions) in the brains and spinal cords of ALS and FTD patients (Arai et al., 2009;Neumann et al., 2006).In normal physiological conditions, TDP-43 is a predominantly nuclear protein, but has been demonstrated to shuttle between the cytoplasm and nucleus with important roles in the nucleus (i.e.transcription, translation, RNA splicing, mRNA stabilization, mRNA transport and nucleocytoplasmic shuttling of mRNA) through nuclear localization signal (NLS) in RNA-recognition motif 1 (RRM1) and nuclear export signal (NES) in RRM2, microRNA and long non-coding RNA (lncRNA) processing) and cytoplasm (i.e.ribonucleoprotein (RNP) transport granule formation, stress granule formation, and translation) (Fig. 2).
However, under pathological circumstances, TDP-43 proteinopathies are observed in both the cytoplasm and nucleus of diseased neurons.TDP-43 proteinopathies in ALS are characterized by (i) hyperphosphorylation, (ii) ubiquitination, (iii) protein aggregation, (iv) protein truncation forming prone-to-aggregation toxic C-terminal fragments of , and (v) nuclear depletion in neurons causing the nucleus to mislocalise to the cytoplasm.Collectively all of these disturbances reduce the level of functional and healthy TDP-43 in the cell nucleus which perturbs normal cellular function (a loss-offunction impact) and conversely, a high accumulation in cytoplasmic inclusions in the spinal cord and brain neurons which can have neurotoxic actions (a toxic gain-of-function impact, Fig. 3) (Ayala et al., 2008;Coyne et al., 2017b;Ling et al., 2013;Neumann et al., 2006).
Experimental studies using cell culture and animal models have provided supportive evidence that TDP-43 proteinopathies are highly involved in the onset and progression of MND (Huang et al., 2020;Ke et al., 2015;Walker et al., 2015).For example, transgenic mice over-expressing TDP-43 with either familial (fALS) or sporadic ALS (sALS) mutations develop a disease-like phenotype (Huang et al., 2020;Wu et al., 2012).Similarly, over-expression of a TDP-43 dNLS variant that lacks nuclear localisation signalling (dNLS), which is an experimental modelling of cytoplasmic TDP-43, develop a disease-like phenotype (Walker et al., 2015).Interestingly, turning off TDP-43 dNLS overexpression rescues neurons even after neurodegeneration onset and reversed motor deficits in multiple animal models of the ALS/FTD (Ke et al., 2015;Walker et al., 2015).All these findings are suggestive that TDP-43 is a signature protein involved in neurodegeneration processes in motor neurons.

Pathological TDP-43 subtypes as therapeutic targets in MND
TDP-43 has many roles in healthy neurons, and extensive literature suggests that cells will no longer be functioning normally if (i) healthy TDP-43 levels become dysregulated (lower or higher) and (ii) any pathological forms of TDP-43 appear in the cells.ALS and frontotemporal lobar degeneration with TDP-43 pathology (FTLD-TDP) is neuropathologically characterized by intranuclear TDP-43 inclusions, TDP-43 oligomers, intracytoplasmic aggregates of TDP-43, and misfolded TDP-43.Toxic functions of TDP-43 might be probably thorough post-translational modifications (i.e.hyperphosphorylation, C-terminus fragmentation (mostly TDP-35 and TDP-25), ubiquitination, sumoylation, methylation, acetylation, and nitrosylation).The clearance of these pathological forms of TDP-43 will enhance the functional recovery in diseased cell and animal models (Afroz et al., 2017;Feneberg et al., 2018;Huang et al., 2020;Ke et al., 2015;Neumann et al., 2007;Walker et al., 2015).This complex array of biochemical forms of TDP-43 pose a challenge in terms of determining which are the pathogenic, disease-relevant forms of TDP-43 (Arai, 2014).In FTLD-TDP there are pathological inclusions of TDP-43 (visual classifications) found in dystrophic neurites (DNs), neuronal cytoplasmic inclusions (NCIs), neuronal intranuclear inclusions (NIIs), and glial cytoplasmic inclusions (GCIs), in which four pathological TDP-43 subtypes (types A-D) were detected in these TDP-43-positive structures (Mackenzie et al., 2011;Neumann et al., 2006).Type A is containing numerous short DNs and NCIs; type B has few DNs but more NCIs; type C has few NCIs but the majority of elongated and thicker DNs than type A; and type D is characterized by few NCIs but numerous DNs and NIIs (Fig. 4).In ALS, there are skein-like/ rounded TDP-43-positive inclusions in the LMN.These TDP-43-positive inclusions are similar to those that can be observed in UMN at the prefrontal gyrus section.Furthermore, near the UPN and LMN, other TDP-43-positive inclusions were founded in the glial cells (Tan et al., 2007).
Additionally, gene mutations can also be associated with the formation of these pathological TDP-43 subtypes, which gene mutations' overlaps make it more complicated to identify the exact gene mutations that are involved in the formation of each subtype.For example, mutations on progranulin gene (GRN) of the patients revealed high incidence of TDP-43 containing NIIs indicating that one possible NIIs formation mechanism might be associated with progranulin deficiency (Tanaka et al., 2014).Additionally, mutations in TARDBP, C9ORF72, and GRN cause overlaps in TDP-43 pathology between types A and B in FTLD-TDP and ALS and other neurological disorders such as nonfluent/agrammatic primary progressive aphasia (naPPA), semantic variant PPA (svPPA), and behavioral variant FTD (bvFTD) (Fig. 4) (Murray et al., 2011).According to these complicated different pathological forms of TDP-43, it is becoming a big challenge in developing and identifying the right therapeutical strategy to target the exact pathological forms without or less affecting the healthy forms of TDP-43, which the cells need for normal functioning.This becomes more complicated when considering pathological TDP-43 subtypes in different cell types.Therefore, there is an emerging need to understand the current and possible strategies targeting the pathological form of diseased or mis-functioned TDP-43.

Therapeutic approaches for restoring the homeostasis of TDP-43
Restoring TDP-43 homeostasis is well studied therapeutic strategy in the field.The approaches for restoring TDP-43 homeostasis can be categorized into three main approaches, i) indirect manipulation of TDP-43 expression levels (ie: knockdown or overexpression), ii) indirect manipulation of TDP-43 levels by regulating protein clearance pathways, iii) maintaining the correct intracellular localisation of TDP-43, and inhibiting the formation of pathogenic TDP-43.

Indirect modulation of TDP-43 levels via knockdown of other genes
The genetic basis of the ALS cases is almost unknown, and mostly known genes identified in ALS patients are SOD1, C9orf72, TARDP, Fig. 3. Gain and loss of functions of TDP-43.In healthy neurons, TDP-43 redistributes to cytoplasmic fraction participating in messenger ribonucleoprotein granules (i.e.neuronal transport granules) enhancing anterograde transport of these granules from soma to distal sections of the axon.In diseased neurons, (i) depletion of TDP-43 from the nucleus causes loss of nuclear function (i.e.transcription deficits), (ii) cytoplasmic TDP-43 aggregations cause gain of toxic functions (i.e. sequestering RNA binding proteins), and (iii) loss of cytoplasmic functions of TDP-43 can cause loss of axonal transport of mRNA-containing neuronal granules.
Antisense oligonucleotides (ASOs) are promising approaches for gene therapy in MND.ASOs are small single-stranded DNA sequences that can precisely target the specific gene expressions at the posttranslational stage and can consequently neutralise, modify, or remove disease-relevant proteins (Leavitt and Tabrizi, 2020).Currently, ASO therapy has been used for clinical trials for targeting genes of superoxide dismutase 1 (SOD1), fused in sarcoma (FUS), chromosome 9 open reading frame 72 (C9orf72), and ATXN2 (Boros et al., 2022).ASO-based gene knock-downs on these genes can indirectly affect the pathogenic forms of TDP-43.It was reported that TDP-43 has some key interactions with FUS (Ling et al., 2010), SOD1 (Volkening et al., 2009), Ataxin-2 (Elden et al., 2010), and C9orf72 (Cook et al., 2020) proteins.In this regard, it was reported that decreasing ataxin-2 levels by ATXN2 targeting-ASO reduced the TDP-43 aggregations in mice models of the TDP-43 (Becker et al., 2017).In another study, it was shown that the production of poly-glycine-arginine protein through G 4 C 2 -repeat expansion in C9orf72 gene (c9FTD/ALS) increased its co-aggregation with TDP-43 in GFP-(GR) 200 mice by altering nucleocytoplasmic transport and the use of specific G 4 C 2 -repeat-targeting ASO (c9ASO) reduced the TDP-43 aggregations level (Cook et al., 2020).There are no studies to date that have directly targeted TDP-43 using ASO therapy, while all the above-mentioned investigations, indirectly target disease-causing mutations on the TARDBP gene which would bring great effort towards finding treatments for TDP-43-causing diseases.A. Babazadeh et al.

Pathological TDP-43 clearing pathways
Understanding the endogenous protein homeostasis responses to cytoplasmic inclusions of TDP-43 and which pathway/s based on what mechanisms can effectively target the clearance of pathological TDP-43, might be helpful towards finding protein homeostasis targeting therapeutics to find the treatment for especially ALS and FTD.Protein quality control phenomena in the cells ensures that nascently translated proteins are folded/refolded to the right configurations.Folding/refolding process occurs in the endoplasmic reticulum (ER) via chaperones (i.e.heat shock proteins), where unfolded protein response (UPR) and heatshock response (HSR) involved for protein homeostasis (Ciechanover and Kwon, 2017;Kostova and Wolf, 2003).Any misfolded/aggregated proteins will be directed to the protein degradation pathways in order to maintain the protein balance in the cells.The endogenous pathways to restore protein degradation are (i) ubiquitin-proteasome system (UPS) and (ii) autophagy including macro-autophagy or autophagy-lysosomal pathway (ALP) and chaperone-mediated autophagy (CMA) (Boland et al., 2018).
Degradation and removal of abnormal protein inclusions occur through these pathways individually or in combination together (Fig. 5).Multiple protein clearance pathways are involved in TDP-43 homeostasis to restore native TDP-43, however, pathological forms of TDP-43 showed more complicated clearance behaviour.The UPS is mostly responsible for the clearance of proteins through 26 S proteasome, however, UPS-dysfunction occurs in LMN and spinal cord homogenates of ALS patients (Kabashi et al., 2012).At the cell level, blocking proteasome activity with MG132 showed increased levels of disease-relevant cytoplasmic inclusions of TDP-43, TDP-35, and TDP-25 in neuronal cell cultures (Neuro2A, NSC34 motor-neuron like cells, and rat primary neurons) (Scotter et al., 2014;Uchida et al., 2016;van Eersel et al., 2011;Walker et al., 2013).Additionally, the knock-down of Rtp3, Fig. 5. Schematic representation of endogenous protein clearance pathways and TDP-43 fates in neurons.TDP-43 undergoes various inclusions in healthy and diseased neurons which impact the recruitment of protein homeostasis pathways of the ubiquitin-proteasome system (UPS), autophagy-lysosomal pathway (ALP), chaperon-medicated autophagy (CMA), and heat-shock response (HSR).In the healthy neurons, aggregation-prone TDP-43 (orange arrows) can be cleared by protein degradation systems inhibiting these forms from becoming pathological forms.Green arrows show functional protein degradation pathways occurring in healthy neurons.UPS acts through the ubiquitination process and degrades different pools of TDP-43 through the proteasome.Normal TDP-43 inclusions (TDP-43 aggregations at healthy conditions) undergo autophagy-mediated degradations (ALP and CMA).CMA involves protein chaperones that are needed for shuttling target proteins to degradation pathways.In healthy neurons, HSR could also act as a refolding system in stressed situations, where heat shock factor 1 (HSF1) induces the production of heat shock proteins (HSPs) such as DNAJB2a and HSP70, in which DNAJB2a could detect misfolded and oligomeric TDP-43 directing them to the HSP70 for refolding process and the resultant functional TDP-43 could normally shuttles between cytoplasm and the nucleus.In stressed situations, liquid stress granules (SG) occur in the nucleus and cytoplasm (through liquid-liquid phase separation, LLPS) to protect the healthy TDP-43, and upon removal of stress conditions, these structures dissipate to restore TDP-43 to its functional form.However, in disease neurons, liquid SGs non-reversibly change to a solid state which then can form insoluble pathological aggregations.In diseased neurons, TDP-43 undergoes post-translational modifications causing different pathological forms resulting in excessive levels of C-terminus fractions of TDP-43 (CTF), oligomeric TDP-43, and finally insoluble abnormal TDP-43 inclusions, in which impairing protein degradation systems.
A. Babazadeh et al. which is an essential protein subunit associated with 26 S proteasome activity, caused TDP-43 aggregation and consequently motor neuron death and ALS-like motor neuron deficits in mice (Tashiro et al., 2012).These studies indicate that the UPS dysfunction may lead to aggregation of misfolded TDP-43 within the cytoplasm and probably UPS is mostly involved in the clearance of misfolded and cytoplasmic mis-localised fractions of TDP-43.On the other hand, the ALP is known as a self-digestion process occurring within cells and provides protein clearance by engulfing abnormal proteins through making autophagosomes and subsequently fusing with lysosomes and finally digesting abnormal proteins in autophagolysosomal vesicles (Codogno and Meijer, 2005).It has been suggested that these vesicles become non-functional in ALS and FTD patients indicating that TDP-43 pathology is also associated with ALP-dysfunction because of the co-aggregation of TDP-43 with ALP machinery components (Sasaki, 2011).Blocking ALP through chloroquine, bafilomycin A1, and 3-methyladenin also showed the accumulation of TDP-43 and its C-terminal fractions and oligomers (Scotter et al., 2014;Wang et al., 2010).Moreover, the use of UPS and ALP-blocking chemicals at the same time increased the insoluble TDP-43 (Scotter et al., 2014).The K48-linked polyubiquitin chains appearance on cytoplasmic aggregations of TDP-43 indicates a failure on UPS and further K63-linked chains ubiquitination moves substrates towards the ALP (Lee et al., 2019;Pohl and Dikic, 2019;Scotter et al., 2014).Evidence from genetic studies also implicates the importance of protein clearance pathways in TDP-43 pathology and disease pathogenesis.For example, causative mutations have been identified in UBQLN2 and p62/SQSTM1, which have dual effects on both pathways and gene knock-out and or disease-associated mutations and consequently loss of functions of each gene would result in aggregations of TDP-43 (Osaka et al., 2016).It was shown that the loss of functional UBQLN2, adversely affected the degradation capability of ALP through the reduction in acidification of autophagosomes and mutations in TMEM106B, GRN, TBK1, OPTN, p62/SQSTM1, and UBQLN2 will cause ALP dysfunctions (and in some cases UPS dysfunctions) and resulting in TDP-43 aggregations (Feng et al., 2020;Mao et al., 2021;Osaka et al., 2016;Shen et al., 2015;Taylor et al., 2018;Wu et al., 2020).The CMA is another type of autophagy that does not require autophagosomes and directly translocate some specific soluble cytosolic proteins (i.e.proteins with KFERQ-like motif) into lysosomal proteolysis.In this type of autophagy, chaperone of HSC70 (heats-hock cognate protein-70) and cochaperones of HSP40 (heat sock protein 40), CHIP (carboxyl terminus of HSC70-interacting protein), and HOP (HDP70-HSP90 organising protein) binds to the proteins and send them into the lysosome via LAMP2A (lysosome-associated membrane protein type 2A) receptor existing on the lysosomal membrane (Kaushik and Cuervo, 2018).CMA is involved in the clearance of wild-type TDP-43 through binding of HSC70 to Q 134 VKKD 138 motif in the RRM1 domain, and the CMA was not involved in the clearance of mutant TDP-43 with A 134 AKKD 138 mutation, declaring specificity of CMA in the clearance of TDP-43 (Huang et al., 2014).Moreover, silencing expression of HSC70 in human neuroblastoma cells causes an increment in TDP-43 levels and sALS cases diagnosed with TDP-43 pathology showed reduced levels of HSC70 expression (Arosio et al., 2020).Misfolded TDP-43 and C-terminal fractions of TDP35 and TDP25 might also be affected by the CMA, however, impairment of LAMP2A-positive lysosomes and consequently CMA dysfunctionalities might occur upon long-time exposure to TDP-43 aggregates illustrating that CMA might be limited to clearance of some specific pools (i.e. of soluble cytoplasmic TDP-43 with the CMA-recognition motifs) and is not a main pathway in the clearance of TDP-43 aggregates (Ormeño et al., 2020).Another protein homeostasis pathway is HSR, which acts indirectly in terms of TDP-43 clearance by stimulating fast expression of HSP chaperones such as HSP40, HSP70, and HSP90, in which these proteins could be involved in the UPS, ALP, and or CMA-mediated clearance of TDP-43 (San Gil et al., 2017) and any mutations in genes encoding these chaperons or long-term existence of TDP-43 aggregates may cause HSR dysfunctionalities due to reduced levels of HSPs expressions and refolding capabilities (Coyne et al., 2017a;Wang et al., 2017).

Inhibiting/ blocking the formation of pathological inclusions of TDP-43
Presumably, TDP-43 synthesis starts healthy (if there are no diseaselinking mutations on TARDBP) and in some disease conditions, healthy TDP-43 forms diseased-linked aggregation types being a causative factor in ALS/FTD.Therefore, preventing the formation of these aggregated forms of TDP-43 could be another therapeutical strategy in ALS/FTD.In order to figure out the feasibility of this therapeutic strategy, understanding the roles of each domain in TDP-43 structure (Fig. 1) and how they would interact with other proteins, enzymes, etc, and also understanding the pathways and mutations involved in the normal functioning of TDP-43 would be useful (Fig. 6).Two RRM domains in TDP-43 structures are responsible for binding to nucleic acids (protein-nucleic acid interactions) and the C-terminus domain (CTD) is involved in protein-protein interactions (Kuo et al., 2009).CTD mediates liquid-liquid separations (LLPS) of TDP-43 and consequently the formation of RNA and protein-rich membrane-less organelles.LLPS is a reversible process, which facilitates normal TDP-43 functions (and other DNA/RNA binding proteins) through membrane-less organelles (Malik and Barmada, 2020).There are several membrane-less organelles such as nuclear bodies (NB) and nuclear stress granules (nSG) in the nucleus A. Babazadeh et al. and cytosolic stress granules (cSG) in the cytoplasmic compartment.In a normal situation, NB and SG both appear in response to cellular stress (i.e. oxidative stress by arsenite) and they reversibly disappear upon removal of stress-inducing conditions, declaring the capabilities of cells in protecting healthy TDP-43 levels in the nucleus and cytoplasm from abnormal aggregations (Wang et al., 2020).Any abnormalities in LLPS may be involved in the transition from liquid droplets of membrane-less organelles to solid-phase aggregations causing pathological aggregations of TDP-43 (Chew et al., 2019).This transition is reported to be associated with the NEAT1 lncRNA (nuclear-enriched abundant transcript 1 long-noncoding RNA), in which stressed conditions upregulate NEAT1 and normal TDP-43 binds to NEAT1f TDP-43 NBs, which inhibits mislocalization of TDP-43 into the cytoplasm.Therefore, any abnormalities in NEAT1 functions or any mutations on RRMs domains of TDP-43 (i.e.D169G) that prevents TDP-43 from binding to NEAT1 will cause solid-state SG and excessive pathogenic TDP-43 aggregates in the cytoplasm (Li et al., 2013;Wang et al., 2020).Therefore, new therapeutical strategies could take the benefits of the LLPS pathway and its roles in TDP-43 NBs formation to maintain the TDP-43 in the nucleus and avoid nuclear clearance of TDP-43 and excessive translocating to the cytoplasm via regulating NEAT1 activities.

Small molecule-based therapies for MND and targeting TDP-43
Small molecule drug discovery could be an efficient way of targeting each of the discussed pathological TDP-43 removing strategies (Fig. 7), however, endogenous protein clearance pathways (UPS, ALP, CMA, and HSR) were the main pathways that have attracted most of the currently reported small molecule therapies.In terms of targeting UPS, deubiquitinating enzymes cleave the peptide or iso-peptide bonds between ubiquitin and proteins and inhibit protein degradation through 26 S proteasome and playing important role in protein homeostasis.However, this might act adversely in terms of clearing ubiquitinated TDP-43 and causing abnormal cytoplasmic aggregations of TDP-43.Therefore, small molecules inhibiting these enzymes might enhance pathological TDP-43 removal, in which the small molecule IU1 (1-[1-(4-fluorophenyl)− 2,5-dimethylpyrrol-3-yl]− 2-pyrrolidin-1-ylethanone) exhibited an enhanced proteasome activity through inhibiting one of the deubiquitinating enzymes, USP14 (ubiquitin-specific protease 14) and consequently improved clearance of overexpressed TDP-43 (Lee et al., 2010).Forskolin is another small molecule that could improve the clearance of diseased variant TDP-43 M337V through enhancing cAMP-dependant protein kinase activity, which in turn acts on the phosphorylation of 26 S proteasome in the UPS (Lokireddy et al., 2015).Hydrophobic tagging (HyT) small molecules are another group of small molecules that contain hydrophobic motif, protein of interest recognition motif, and cell-penetrating peptide targeting proteins of interest stimulating denatured protein state, in which these proteins could be degraded through UPS (Neklesa et al., 2011).A series of TDP-43 targeting HyT small molecules were designed based on two core aggregation-prone motifs of TDP-43, 246 EDLIIKGISV 255 and 311 MNFGAFSINP 320 , in which D4 KGSGS-EDLIIKGISV-GRKKRRQRRR showed improvements in TDP-43 degradation in vitro and in vivo Drosophila model (Gao et al., 2019).HyT method of targeting TDP-43 clearance is an interesting area and could be coupled by PROTAC (proteolysis targeting chimera) small molecules, in which is in its early stages for TDP-43 clearance purposes.PROTAC small molecules are composed of two domains in which one domain binds to a target protein with high selectivity and another domain engages to E3-ligase complex works through E3-ligase complex which ubiquitinates the proteins meant for degradation through proteasome (Zeng et al., 2021).TDP-43 targeting HyT small molecules could be involved in the protein targeting domain of PROTAC for designing TDP-43 PROTACs.
Small molecules could target different pathways in ALP and mTOR is one of the most important pathways that have regulatory roles in ALP, in which Rapamycin-induced inhibition of mTOR showed enhanced ALP activity regarding the degradation of cytoplasmic TDP-43 and TDP25 Fig. 7. Potent small-molecule mechanisms of actions targeting TDP-43 proteinopathies.Small molecules mechanisms of actions (green arrows) could potentially modulate TDP-43 recruitment to stress granules (SGs) and prevent abnormal accumulations of TDP-43 in ALS/FTD cases.Reducing the formation of SGs and reducing the size of SGs are the potent mechanisms that small molecules could be involved regarding reducing permanent TDP-43 accumulations.Another mechanism might be preventing TDP-43 from getting inside the SGs via various pathways.Enhancing protein degradation pathways such as ubiquitin-proteasome system (UPS) and autophagy-lysosome system (ALP) is considered as another mechanism of action for small molecules.

Challenges of therapeutical approaches targeting pathological forms of TDP-43
The functional form of TDP-43 plays many essential roles in cellular activities such as RNA metabolisms (i.e.RNA splicing, Fig. 2).This is critically an important step towards designing or exploring therapeutical strategies that need to consider possible consequences or implications of indiscriminate targeting of TDP-43, while maintaining its functional form.Therefore, in this section, we discuss small molecule-based TDP-43 targeted therapies and address the possible challenges of targeting only the pathological/toxic forms of TDP-43.
Aggregated form of TDP-43 is one of the major neurotoxic hallmarks in ALS/FTD, which designing small molecules for targeting these toxic forms would be feasible by understanding the crystal structure of these toxic forms (Cohen et al., 2015;Fang et al., 2014;Kuo et al., 2014).Uncovering the crystal structure of toxic forms of TDP-43 will enable in silico docking studies to develop small molecules which can specifically bind to the toxic forms of TDP-43 while minimising targeting its functional forms.In silico docking of small molecules using the crystal structure of the N-terminal domain of wild type TDP-43 introduced a promising compound (nTRD22), which showed therapeutic effects in targeting TDP-43 and mitigating motor impairments in Drosophila model of ALS (Mollasalehi et al., 2020).Furthermore, similar studies need to be conducted considering in silico molecular docking to develop small molecules for targeting mutant TDP-43 and as well as other toxic forms.Therefore, there would be some other challenges that should be addressed to (i) avoid or minimise indiscriminate targeting of TDP-43 and (ii) understand potential complementing strategies to restore or maintain functional TDP-43, which currently might limit the efficiency of small molecules in in vitro/vivo models of ALS/FTD.First, protein clearance systems are a complex interconnected network, and administration of small molecules that might target one specific pathway may have adverse effects on other pathways, illustrating how it would be important to investigate all protein degradation pathways together for a better understanding of each pathway's role.Secondly, there are concerns about off-target activities of small molecules as they might adversely impact other cell mechanisms (i.e.metabolic processes).For example, mTOR is a regulator of multiple metabolic pathways and mTOR inhibitors enhancing ALP activities may cause impairments in these pathways (Dossou and Basu, 2019).Another concern is about small molecule effects on the endogenous TDP-43 aggregations and their fates as the majority of small molecule mediated TDP-43 clearance investigations were based on overexpression of fluorescently labelled TDP-43 in cells and animal models, which might not reflect the different endogenous human TDP-43 pools in human neurons and their interactive role with other proteins, which might adversely affect other proteins level that is vital for normal cell functions.Therefore, investigating small molecules' effects on the possible proteins interacting with TDP-43 would be useful when studying TDP-43 clearance pathways.Additionally, most of the small molecules that have been investigated in cell/animal models rely upon the normal function of protein clearance pathways, but in ALS/FTD cases these pathways might be impaired or adversely impacted by age and other diseases causing factors.In sum, the above-mentioned concerns are challenges that associated with small molecules assuming that 100% of these molecules will be present in the cell/animal models, which is practically not feasible.Additionally, without control over how much TDP-43 is degraded, it is possible that functional TDP-43 may also be removed which would have deleterious effects on the cell.More importantly, the major concern is that small molecules need to be delivered into the target site of action, diseased motor neurons, to avoid any unwanted effects to other neurons, which is more complicated in in vivo models even if small molecules are directly injected into the brain and / or spinal cord.Additionally, another major remaining concern is the blood-brain barrier (BBB) crossing abilities of the small molecules, where brain delivery of these molecules via nano-drug delivery systems could be an advanced way of small-molecule-based therapies (Babazadeh et al., 2020).Functionalized nanodelivery of small molecules will make it feasible that they could cross the BBB and target diseased motor neurons, and finally affect protein clearance pathways (which the associated concerns were highlighted above).

Current clinical trials for targeting TDP-43
Current therapeutical advances toward finding treatment in MND specifically targeting clearance of TDP-43 include (i) genetic discoveries and viral gene delivery, which showed some persuasive therapeutical effects on disease-linked genes affecting the levels of TDP-43, and (ii) small molecules affecting cellular pathways and improving clearance of pathological TDP-43.Among these advances, some are already tested and approved, and some are under investigation.Riluzole is an FDAapproved drug for ALS with mysterious mechanisms of action, in which it could act by inhibiting protein kinase CK1δ and preventing hyper-phosphorylation of TDP-43 and C-terminal fragments of TDP-43 (Bissaro and Moro, 2019).Edaravone is another FDA-approved treatment for ALS, which could enhance pathological TDP-43 clearance through Nrf2-oxidative stress response and autophagy and could also prevent TDP-43 from misfolding through unfold protein response (Soejima-Kusunoki et al., 2022).
A number of currently non-FDA-approved drugs have entered clinical trials.Phase I dose dependant (100, 200, 300 or 400 mg/day) study of a small molecule, bosutinib, were conducted on two 12-week treatment period followed by 1 week transition period and four weeks followup periods.ALS patients were observed for ALSFRS-R score (ALS Functional Rating Scale-Revised score), which is a questionnaire-based scale used to track the progression of disability in patients with ALS (Cedarbaum et al., 1999;Imamura et al., 2019).As bosutinib is a drug with improving impaired autophagy effects, it could potentially target the clearance of pathological TDP-43.Tamoxifen (40 mg/day) is another small molecule that underwent 12-month phase II clinical trial studies and ALS patients were monitored by ALSFRS-R score and pulmonary function measured by FVC (forced vital capacity).Tamoxifen mildly attenuated disease progress.Tamoxifen is an autophagy enhancer through (i) inhibiting AKT/PKB in mTOR dependent pathway and (ii) dissociating autophagy proteins Beclin-1 and BCL-2, in which it could enhance pathological TDP-43 clearance (Chen et al., 2020).Another 24 weeks of phase II clinical trial of combinational drug treatment with sodium phenylbutyrate (3 g) and taurursodiol (1 g) was conducted on 137 ALS patients, in which patients were monitored with ALSFRS-R score.This combinational treatment significantly slowed the functional decline in ALS patients by alleviating the effects of sodium phenylbutyrate and taurursodiol on endoplasmic reticulum (ER) stress and mitochondrial dysfunction, respectively (Paganoni et al., 2021)  The drug was introduced as the third FDA approved drug for ALS, which the final decision will be made on 2024 to retain the drug as the 3rd approved drug for ALS or not.The drug showed prolonged survival rate and slower rate of functional decline (ALSFRS-R) in ALS cases.

Rotigotine (NCT04937452)
A 24-week study on 75 patients that were newly diagnosed with behavioural Frontotemporal Dementia (bvFTD)  The drug enhances the expression of heat shock protein B8 and of some autophagy players while preventing accumulation of TDP-43 in neurons.Additionally, the well-known anti-inflammatory effects of Colchicine may be involved in cell homeostasis.(Mandrioli et al., 2019) N/A (continued on next page) A. Babazadeh et al. 12-month clinical study about lithium carbonate effects on 518 ALS patients showed promising survival improvements in ALS patients.The mechanism of action might be associated with autophagy induction, which in turn could affect TDP-43 levels (van Eijk et al., 2017).In sum, these small molecule-mediated therapies in ALS patients at clinical trial levels declare promising therapeutical advances that could possibly regulate TDP-43 homeostasis protecting motor neurons from further TDP-43 proteinopathies, however, it needs further investigations to determine the effectiveness of small molecules on pathological forms of TDP-43, in which ongoing clinical trials for TDP-43 pathology is summarized in Table 1.

Conclusion
TDP-43 is a nuclear-functioning protein, and alterations in its nucleocytoplasmic distribution and consequent formation of pathological aggregations of TDP-43 could revert this balance causing a progressive neurodegeneration state.Small molecule-based therapies for restoring TDP-43 homeostasis in TDP-43-associated neurodegenerative diseases especially ALS (as a common form of MND) and FTD is one important ongoing area of investigation towards finding treatment in ALS/FTD.Small molecules could therapeutically prevent cytoplasmic aggregations of TDP-43 by post-translational modifications through (i) protease kinase inhibitors and inhibiting TDP-43 hyperphosphorylation, (ii) inhibiting solid state stress granule formations, and (iii) regulating protein homeostasis pathways UPS, ALP, HSR, and CMA.Additionally, small molecules could potentially ameliorate impairments in protein homeostasis systems and enhance the clearance of pathological TDP-43.A prospective therapeutical approach for ALS/FTD would benefit from small molecule-based therapy, which could be enhanced when coupled with gene therapy and drug delivery systems that could protect small molecules from unwanted physiological conditions and boost their effects in the target site of the action.

Fig. 1 .
Fig. 1.TDP-43 structure and the mutation regions causing ALS.TDP-43, a 414 amino acid length protein, consists of an N-terminal domain (NTD), two RNA recognition motifs (RRM1 and RRM2), and a C-terminal glycine-rich region.Identified mutations in TDP-43 identified from sporadic and familial ALS patients are highlighted in red.There are several known caspase cleavage sites (shown by scissors) which generate the C-terminal fragments of 35 kDa (TDP-35) and 25 kDa (TDP-25) size.Nuclear localization signal (NLS) and nuclear export signal (NES) sequences are highlighted in blue and pink, respectively.

Fig. 4 .
Fig. 4. Schematic representation of morphology and anatomical distribution of pathological subtypes of TDP-43 and associated gen mutations in each subtype in different neurodegenerative diseases.Four pathological TDP-43 subtypes that were detected in TDP-43-positive structures mainly contain different levels of dystrophic neurites (DNs), neuronal cytoplasmic inclusions (NCIs), and neuronal intranuclear inclusions (NIIs) in different cortical layers.Type A contains numerous short DNs and NCIs, Type B contains few DNs but more NCIs, Type C has few NCIs and more elongated and thicker DNs than Type A, and Type D has few NCIs but numerous DNs and NIIs.Amyotrophic lateral sclerosis (ALS), nonfluent/agrammatic primary progressive aphasia (naPPA), semantic variant PPA (svPPA), frontotemporal dementia (FTD), behavioral variant FTD (bvFTD), Granulin (GRN), Vasolin containing protein (VCP).
. A

Table 1
Ongoing small molecule-based clinical trials for neurodegenerative diseases associated with TDP-43 proteinopathies reported by U.S. National Library of Medicine (Provided the national clinical trial number-NTC number).