Amyotrophic lateral sclerosis (ALS) is a devastating neurodegenerative disorder characterized by the death of both upper motor neuron (UMN) and lower motor neuron (LMN) in the brain, brainstem, and spinal cord, ultimately leading to death from respiratory failure [1]. In the majority of cases (approximately 95%), excluding those with pathogenic variants in the FUS gene, the hallmark pathology is the nuclear-cytoplasmic mislocalization and aggregation of TAR DNA-binding protein 43 (TDP-43) [1,2]. Additionally, TDP-43 pathology is observed in patients with frontotemporal dementia (FTD), Alzheimer’s disease, limbic-predominant age-related TDP-43 encephalopathy (LATE), and inclusion body myositis, indicating that it is not specific to ALS [2-5]. ALS is diagnosed based on the revised El Escorial or Awaji criteria, which incorporate neurological examination findings and electrophysiological studies. The diagnosis requires confirmation of UMN or LMN signs in each anatomical segment (bulbar, cervical, thoracic, and lumbosacral) [6,7]. Importantly, while electromyography can provide evidence for the degeneration of LMNs, current data do not support the use of imaging or neurophysiological modalities to establish UMN dysfunction [8]. Furthermore, although TDP-43 positron emission tomography ligands are under investigation [9], they are not yet widely available for evaluating brain TDP-43 pathology. Therefore, a robust fluid biomarker that reflects pathological TDP-43 accumulation in the central nervous system (CNS) would be highly useful for confirming ALS diagnoses.

This review comprehensively summarizes the current status of biomarker development, focusing on TDP-43 pathology, and discusses the existing limitations and future avenues for ALS biomarker discovery.

TDP-43 is a 43 kDa heterogeneous nuclear ribonuclear protein (hnRNP) composed of 414 amino acids that is encoded by the TAR DNA binding protein (TARDBP) gene located on chromosome 1 [2]. It plays a crucial role in regulating gene expression and various RNA processing activities, such as RNA splicing, mRNA turnover, RNA trafficking, and microRNA biogenesis. TDP-43 targets more than 4,000 different mRNA transcripts, including those associated with diseases and its own mRNA transcript through autoregulation.

TDP-43 is composed of an N-terminal domain, a nuclear localization signal, two RNA-recognition motifs, a nuclear export signal, and a C-terminal glycine-rich domain (Fig. 1) [2]. The protein also has an amyloidogenic core region (residues 311-360) with two alpha-helices that convert into beta sheets in TDP-43 aggregates. The N-terminus primarily functions to regulate the homodimerization of TDP-43, which is crucial for proper protein folding and mRNA splicing. The C-terminus plays a significant role in mRNA splicing and hnRNP interactions and is also believed to contribute to the formation of TDP-43 inclusions.

TDP-43 pathology primarily develops in the CNS tissues of ALS patients, which initially directed research on TDP-43 fluid biomarkers toward the cerebrospinal fluid (CSF). In 2008, TDP-43 was first evaluated as a potential CSF biomarker for ALS [10]. Researchers found elevated levels of a 45 kDa band in the CSF of ALS patients compared to controls using a monoclonal antibody that targets amino acids 1 through 260 of recombinant TDP-43. Shortly afterward, another group measured TDP-43 levels in the CSF of ALS patients using a sandwich enzyme-linked immunosorbent assay (ELISA) with identical capture and detection antibodies [11]. Further analysis by immunoblotting revealed a 43 kDa band, indicating that the ELISA primarily detects full-length TDP-43. This study also demonstrated that CSF TDP-43 levels were significantly higher in individuals with sporadic ALS than in age-matched healthy or neurological disease controls. Subsequently, a different research group developed a TDP-43 sandwich ELISA using the same capture and detection antibodies [11]. In their study, CSF TDP-43 levels were significantly higher in ALS patients than in those with Guillain-Barré syndrome, showing a sensitivity of 84.6% and a specificity of 71.4% (cut-off, 1.16 ng/mL). These findings indicate that CSF TDP-43 measurements alone may not meet the standards for clinical application. In a later study using an ELISA, CSF TDP-43 levels were found to be elevated in ALS patients compared to those with neurodegenerative and inflammatory neurological diseases [12], aligning with previous findings [10,11]. The results showed a sensitivity of 59.3% and a specificity of 96% (cut-off, 27.9 ng/mL), suggesting that while levels below the cut-off may not exclude an ALS diagnosis, a positive result could aid in distinguishing ALS from other neurological conditions.

In a study involving 219 patients with ALS and 100 healthy controls, plasma TDP-43 levels measured by sandwich ELISA were significantly higher in the ALS cohort [13]. However, TDP-43 concentrations exceeded the assay's detection limit in only 28% of these patients, compared to 21% of controls, underscoring the need for a more sensitive assay. More recently, a commercial single-molecule array (SIMOA) assay, designed to detect both full-length and pathologically truncated TDP-43, was utilized [14]. Both CSF and plasma TDP-43 levels were significantly higher in patients with ALS than in controls within the discovery cohort. However, in the validation cohort, only CSF TDP-43 levels were found to be elevated in ALS patients compared to controls.

A recent study utilized sandwich ELISAs to measure both total TDP-43 and phosphorylated TDP-43 (pTDP-43) in CSF and plasma [15]. The results showed that plasma levels of TDP-43 and pTDP-43 were significantly elevated in ALS patients compared to healthy controls, whereas CSF levels did not exhibit a similar increase. Furthermore, ALS patients displayed significantly lower plasma pTDP-43/TDP-43 ratios, while their CSF pTDP-43/TDP-43 ratios did not significantly differ from those of the controls. In terms of diagnostic performance, plasma TDP-43 achieved a sensitivity of 91.3% and a specificity of 91.5%. In contrast, plasma pTDP-43 demonstrated a sensitivity of 82.6% and a specificity of 67.8%.

Interestingly, while pathological TDP-43 in biofluids is believed to originate from damaged motor neurons in the intrathecal compartment of ALS patients, it is important to note that the concentrations of total TDP-43 and pTDP-43 are significantly higher in plasma than in CSF [15]. This discrepancy could be attributed to active transport between the CSF and plasma or to the release from non-neuronal peripheral tissues or cells, which merits further investigation. In this context, platelet-rich plasma might be a significant source, as TDP-43 levels in platelets from ALS patients are markedly elevated compared to those in healthy controls [16]. Although no direct comparison studies have been conducted, TDP-43 levels in platelets appear to be substantially higher than those in plasma.

The real-time quaking-induced conversion reaction (RT-QuIC) has proven to be a robust technique for prion amplification in diseases such as Creutzfeldt-Jakob disease [17]. Recently, this technique was adapted to use TDP-43 as a substrate, exploiting its ability to amplify minute amounts of misfolded proteins for the detection of pathological TDP-43 species in the CSF of ALS and FTD patients [18]. The TDP-43 RT-QuIC method was able to detect as little as 15 pg of TDP-43 aggregates, successfully discriminating between patients affected by ALS or FTD harboring pathogenic variants in the C9orf72, granulin precursor (GRN), and TARDBP genes from age-matched controls, with a sensitivity of 94% and a specificity of 85%. These findings require further validation in larger cohorts of sporadic ALS.

Evidence indicates that cryptic exons arising from the loss of TDP-43 function play a crucial role in the pathogenesis of ALS [19]. Recent research utilizing a novel monoclonal antibody that targets a TDP-43-dependent cryptic epitope has shown that splicing repression by TDP-43 is compromised in ALS-FTD, including in presymptomatic carriers of C9orf72 mutation [20]. This cryptic hepatoma-derived growth factor-related protein 2 (HDGFL2) accumulates in the CSF at significantly higher levels in familial ALS-FTD and sporadic ALS compared to controls, and it is elevated earlier than neurofilament light and phosphorylated neurofilament heavy chain levels in familial cases. These findings suggest that the loss of TDP-43 cryptic splicing repression occurs early in disease progression and that detecting the HDGFL2 cryptic neoepitope could serve as a potential diagnostic biomarker for ALS.

Although CSF or plasma TDP-43 is considered a potential diagnostic biomarker for ALS, several limitations should be acknowledged. First, TDP-43 proteinopathy is not specific to ALS and is observed in other neurodegenerative diseases, including FTD, Alzheimer’s disease, LATE, and inclusion body myositis. Second, the use of antibodies with varying sensitivities and specificities for immunoassays to detect different forms of TDP-43 has led to inconsistent results and challenges in comparing findings across studies. Third, TDP-43 is expressed ubiquitously in both the CNS and peripheral tissues, making it unclear whether soluble TDP-43 detected in biofluids specifically originates from the damaged CNS. Fourth, increases in TDP-43 levels in CSF or plasma might reflect neurodegeneration and the release of TDP-43 from damaged neurons, while decreases could be due to its sequestration in cytoplasmic aggregates. Although several studies have reported elevated TDP-43 levels in these fluids in ALS patients compared to controls, these findings have yet to demonstrate clear clinical utility at the individual patient level.

Numerous studies have investigated the diagnostic utility of CSF or plasma TDP-43 in ALS, frequently alongside other biomarkers like tau and neurofilament. However, there is currently no conclusive evidence to support the use of TDP-43 alone as a diagnostic biomarker for TDP-43 proteinopathy in ALS. Future research should focus on creating a comprehensive diagnostic strategy that integrates clinical data, genetic information, and a range of biomarkers—including fluid-based, electrophysiological, and imaging assessments—to enhance the accuracy of ALS diagnosis.