Multi-Omics for Biomarker Discovery and Target Validation in Biofluids for Amyotrophic Lateral Sclerosis Diagnosis

Abstract

Amyotrophic lateral sclerosis (ALS) is a rare but usually fatal neurodegenerative disease characterized by motor neuron degeneration in the brain and the spinal cord. Two forms are recognized, the familial that accounts for 5–10% and the sporadic that accounts for the rest. New studies suggest that ALS is a highly heterogeneous disease, and this diversity is a major reason for the lack of successful therapeutic treatments. Indeed, only two drugs (riluzole and edaravone) have been approved that provide a limited improvement in the quality of life. Presently, the diagnosis of ALS is based on clinical examination and lag period from the onset of symptoms to the final diagnosis is ∼12 months. Therefore, the discovery of robust molecular biomarkers that can assist in the diagnosis is of major importance. DNA sequencing to identify pathogenic gene variants can be applied in the cases of familial ALS. However, it is not a routinely used diagnostic procedure and most importantly, it cannot be applied in the diagnosis of sporadic ALS. In this expert review, the current approaches in identification of new ALS biomarkers are discussed. The advent of various multi-omics biotechnology platforms, including miRNomics, proteomics, metabolomics, metallomics, volatolomics, and viromics, has assisted in the identification of new biomarkers. The biofluids are the most preferable material for the analysis of potential biomarkers (such as proteins and cell-free miRNAs), since they are easily obtained. In the near future, the biofluid-based biomarkers will be indispensable to classify different ALS subtypes and understand the molecular heterogeneity of the disease.

Introduction

Amyotrophic Lateral Sclerosis (ALS) is a rare but progressive and usually fatal neurodegenerative disorder that involves the degradation of motor neurons in the brain and spinal cord. The disease starts insidiously with focal weakness, but it spreads to involve most of the muscles including the diaphragm. Typically, death occurs after 3–5 years due to respiratory paralysis, although extreme cases where patients are able to survive for >40 years have been reported (Brown and Al-Chalibi, 2017). The incidence of ALS is 2.1 per 100,000 persons per year, with an estimated prevalence of 5.4 cases per 100,000 populations (Chiò et al., 2013). There are, however, foci where the disease is more common, for example, in the island of Guam at the Pacific Ocean (Morris et al., 2001).

Two forms of ALS are recognized: the familial ALS (fALS) and the sporadic ALS (sALS). fALS accounts for 5–10% of ALS cases, and the rest are sporadic in nature. Pathogenic variants in the SOD1 gene were the first to be associated with ALS, and currently more than 220 genetic variants have been identified (http://alsod.iop.kcl.ac.uk/Overview/gene.aspx?gene_id=SOD1). SOD1 genomic variants account for approximately 20% of fALS cases. However, the most common ALS genomic variant is a hexanucleotide repeat expansion in the C9ORF72 gene that accounts for 10% of sALS and 45% of fALS cases (Brown and Al-Chalibi, 2017; Paez-Colasante et al., 2015).

Currently, ∼50 ALS-associated genes have been identified with TDP-43, FUS, and UBQLN2 (ubiquitin 2) being also common. Notably, the mutated genes in ALS encode for proteins with very distinct functions in the cell; SOD1 catalyzes the dismutation of hydrogen peroxide, TDP-43 and FUS are involved in DNA/RNA metabolism, ubiquitin 2 in protein degradation etc., thus making ALS a genetically heterogeneous disease (Brown and Al-Chalibi, 2017). The heterogeneity is further increased since environmental factors have been suggested as causative factors of a subset of sALS cases, as, for example, the exposure to toxic metals or other chemicals (such as the β-methylamino-l-alanine, a component of cycad seeds that are consumed in Guam). Also, endogenous retroviruses have been incriminated for certain sALS cases. Other genetic factors may affect the progression of the disease. Specifically, loss-of-function genomic variants in EPHA4 gene have been associated with longer survival in ALS (Van Hoecke et al., 2012).

Recently, a large number of ALS consortia have been created in the United States and Europe to facilitate larger research and clinical efforts (e.g., Northeast ALS Consortium, European Network for the Cure of ALS, Clinical Research in ALS, and related disorders for Therapeutic Development etc.).

Currently, the only ALS drugs approved by the FDA are the riluzole and edaravone. Both provide limited improvement in survival. Riluzole acts by suppressing excessive motor neuron firing and edaravone by suppressing oxidative stress (Brown and Al-Chalabi, 2017). Identification of ALS biomarkers is essential for precision medicine and early pharmacological intervention. Indeed, the most significant benefit of riluzole is observed after early intervention (Zoing et al., 2006); therefore, the quicker the diagnosis of ALS, the better the benefit. Current diagnosis of ALS relies on clinical symptoms, and the time from the first symptoms to diagnosis is ∼12 months, that is, a problematic delay for successful therapy (Brown and Al-Chalibi, 2017).

Further, due to the high heterogeneity of ALS, the development of new biomarkers that would assist in “ALS classification/stratification” will be indispensable for personalized treatment. The very recent report for the successful treatment of a sALS patient suspected to be due to mercury intoxication, with a combination of the chelating drugs 2,3-dimercaptopropanesulfate and α-lipoic acid (Mangelsdorf et al., 2017), highlights the importance to classify the different types of ALS to assist in more successful treatments. The purpose of the present expert review is to examine and synthesize the current efforts and directions taken to provide a rapid ALS diagnosis and classification based on molecular and biochemical markers available in easily obtained biofluids.

Biomarkers and ALS

Biomarkers aid in the rapid and more accurate disease diagnosis; they help to stratify the patient population (especially in ALS where a high heterogeneity is observed) and identify patients who will respond better to a particular drug, the essence of precision medicine. Biomarkers can also aid the identification of new drug targets and the preclinical drug development. They may provide a bridge between the preclinical disease models and the human patient population, with biomarkers common between the model system and patient population providing important mechanistic links between the two and highlighting potential therapeutic targets. Since tissue biopsies are difficult to obtain for diagnostic application in neurodegenerative disorders, biofluids including blood and cerebrospinal fluid (CSF) are the most highly investigated fluids to identify candidate ALS biomarkers.

Although CSF represents the fluid that is close to the affected cells in the central nervous system (CNS), it requires invasive lumbar puncture to obtain; thus, blood that requires a less invasive method is the most preferred fluid for molecular analysis (Jeromin and Bowser, 2017; Robelin and Gonzalez De Aguilar, 2014; van Es et al., 2017).

The biomarkers can be subdivided into (a) gene- (such as DNA variants, RNA, miRNA), (ii) protein-, and (iii) metabolite-based. Genomic biomarkers have been extensively reviewed previously (Keller et al., 2014; Renton et al., 2014). All the reported genomic variants represent biomarkers for diagnostic applications, and already the diagnostic testing facilities test a panel of the 17 most common genomic variants that are linked to fALS (Jeromin and Bowser, 2017). Here, we will focus on protein- and small-molecule-based biomarkers, although the miRNAs will be also discussed since they can be found free in serum or CSF. The various omics disciplines that assist in the identification of novel biomarkers are depicted in Figure 1.

FIG. 1.

Emerging omics technologies for ALS diagnostics. From the classical genomics that is the current trend, we tend to move to a multi-omics approach that is increasingly applied to biofluids from the standpoint of clinical feasibility. Such methodologies include the miRNomics, the proteomics, the metabolomics, the volatolomics, the metallomics, and the viromics, each of which can greatly assist in the near future diagnosis and classification of the various fALS and sALS forms. ALS, amyotrophic lateral sclerosis; fALS, familial amyotrophic lateral sclerosis; sALS, sporadic amyotrophic lateral sclerosis.

Biomarkers for ALS

Biochemical markers include proteins and small molecules found in biological fluids. OMIC technologies and especially proteomics and metabolomics can assist in the identification of new biomarkers that can be further validated in large cohorts of ALS patients. The protein-based biomarkers are the most extensively studied in biofluids, although in recent years the miRNAs released into the biofluids have also been utilized as biomarkers. Biofluids are also valuable for analysis of small molecules (metabolites) to identify new biomarkers. Further, other approaches, including volatolomics, metallomics, and viromics, will be discussed as they offer potential new strategies for the diagnosis and stratification of ALS.

miRNAs as Biomarkers for ALS

miRNAs are small 18–25 nt in length noncoding RNA molecules that regulate mRNAs at the post-transcriptional level. miRNAs are important circulating biomarkers since they do not require tissue removal by biopsy, but instead biofluid (CSF or blood) sampling. miRNAs have the advantage that although they are RNA molecules they exhibit high stability in serum and other biofluids. Importantly, alterations in miRNA profiling represent early changes of disease; therefore, the analysis of miRNAs could be used for early diagnosis and to boost presymptomatic treatment. Most of the studies examining miRNAs in ALS have applied miRNomics (miRNA microarrays) to identify alterations in miRNA profiles between ALS and healthy control samples, and the results have been validated in separated patient and healthy cohorts with RT-qPCR (reverse transcription quantitative PCR) assays.

Eventually, it is hoped that the RT-qPCR determination of a small panel of miRNA in CSF or serum could provide valuable clinical applications. However, as will be outlined later, there is large inconsistency between the results obtained by different laboratories. The inconsistency could be due to the following reasons: detection methods including selection of appropriate normalization genes, sample size, disease course, fALS or sALS, and ethnic origin. An important factor that should be considered when analyzing miRNA profiles in the serum of ALS patients is the altered stability of ALS blood cells that results in hemolysis (Freischmidt et al., 2014; Ronnevi and Conradi, 1984) that, in turn, could lead to a false positive increase in serum miRNAs.

Searching for specific miRNAs that are downregulated in presymptomatic ALS patients carrying mapped pathogenic variants has been performed and led to the identification of 24 miRNAs that are specific for fALS and 22 characteristic for fALS-presymptomatic patients. A panel of the miR-4745-5p, miR-3665, miR1915-3p, and miR-4530 miRNAs in serum exhibiting the highest differences was selected for validation and found to be able to discriminate the presymptomic fALS from healthy controls (Freischmidt et al., 2014).

sALS is per se a reason for high heterogeneity compared with fALS (Freischmidt et al., 2015). Indeed, sALS shows significant high variability in miRNA expression; nonetheless, the downregulation of miR-1234-3p and miR-1825 appeared to be a common denominator for sALS after analyzing 18 sALS patients and 16 healthy control serum samples (Freischmidt et al., 2015). The downregulation of these two miRNAs was confirmed in an independent cohort composed of 20 sALS and 20 healthy controls, whereas miR-1825 was also found to be downregulated in fALS patients with pathogenic variants in C9ORF72 and SOD1 genes. In conclusion, the downregulation of miR-1234-3p appeared to be specific for sALS. Importantly, miR-1234-3p and miR-1825 did not exhibit change in Alzheimer's disease (AD) whereas they were found to be upregulated in Huntington's disease (HD) (Freischmidt et al., 2015). Thus, these miRNAs may have diagnostic applications.

The high variability in findings can be easily observed since the same group previously (Freischmidt et al., 2013) tested nine TDP-43-regulated miRNAs in 22 sALS patients and 24 healthy controls. They found that miR-132-5p, miR-132-3p, and miR-143-3p were downregulated and miR-143-5p and miR-574-5p were upregulated in CSF; whereas in the serum, downregulation of miR-132-5p, miR-132-3p, miR-143-5p, miR-143-3p, and let-7b was demonstrated. In the same study, the downregulation of miR-9-5p and miR-663a was found as a characteristic of patients with FUS genomic variants (Freischmidt et al., 2013). Dysregulation of miR-143-5p/3p seems to be a common feature of ALS pathology, downregulation of miR-132-5p/3p and miR-574-5p/3p was evident in sporadic, TARDBP, FUS, and C9ORF72, but not SOD1 mutant patients, supporting the existence of molecular markers that can identify certain ALS subpopulations (Freischmidt et al., 2013).

A very recent study performed miRNA profiling in sALS serum of 27 patients and compare it with disease mimics (seven patients with noninflammatory neuropathies, seven myopathies, eight inflammatory neuropathies including Guillain-Barré syndrome [GBS], six structural spinal disorders, and eight myasthenia gravis) and 25 controls (Waller et al. 2017). In total, 12 miRNAs were found to be differentially expressed and after validation in an independent cohort of ALS patients, miR-206, miR-143-3p, and miR-374-5p were found to have diagnostic significance with the first two increased whereas the latter decreased in ALS serum. Administration of riluzole did not affect the levels of these miRNAs. In addition, the levels of miR-143-3p were increased and the levels of miR-374-5p were decreased with disease progression (Waller et al., 2017).

Analysis of miRNA expression in leukocytes of Chinese sALS patients (five sALS, five healthy) showed that miR-183, miR-193b, and miR-451 were downregulated whereas miR-3935 was upregulated. This panel had a high diagnostic accuracy (Chen et al., 2016b). The findings were validated in 83 sALS patients, 24 Parkinson's Disease (PD) patients, and 61 healthy controls. Notably, the initial microarray profiling showed upregulation of miR-193b but in the validation process with RT-qPCR it was found to be downregulated. This could reflect differences in normalization. Only miR-183 was specific for sALS compared with PD (Chen et al., 2016a). miR-451 has also been found to be downregulated in leukocytes from Italian sALS patients (De Felice et al., 2012).

Studies on defining the role of individual miRNAs with RT-qPCR in ALS have also been performed. miR-338-3p has been found to be upregulated in blood leukocytes, CSF, serum, and in the spinal cord of a cohort of Italian and a small percentage of German sALS patients (De Felice et al., 2014).

In a study with Japanese patients (16 sALS and 10 healthy controls and a validation cohort of 48 sALS, 47 healthy controls, and 30 PD) to identify alterations in the expression of plasma-derived miRNAs, it was found that miR-4649-5p was upregulated and miR-4299 was downregulated (Takahashi et al., 2015). The authors did not find significant upregulation of miR-338-3p highlighting the heterogeneity in findings between different groups, although ethnicity could be a reason here. However, the author did confirm but did not study in detail the downregulation of miR-1234-3p and miR-1825, as reported by Freischmidt et al., 2015 (Takahashi et al., 2015). Another study has selected certain ratios between different miRNAs for diagnosis of ALS in CSF and specifically the miR-181a-5p/miR21-5p and miR-181-5p/miR-15b-5p ratios (Benigni et al., 2016). The findings from the aforementioned studies are summarized in Table 1.

Table 1.

List of miRNAs Found Deregulated in Amyotrophic Lateral Sclerosis

miRNA	Expression	Participants	Sample	Normalization	Reference
miR-4745-5p	↓	Discovery: 9 fALS, 10 controls	Serum	Cel-miR-39-3p (spiked)	Freischmidt et al. (2014)
miR-4745-5p	↓	Validation: 13 fALS, 13 controls	Serum	Cel-miR-39-3p (spiked)	Freischmidt et al. (2014)
miR-3665	↓	“	Serum	Cel-miR-39-3p
miR1915-3p	↓	“	Serum	Cel-miR-39-3p
miR-4530	↓	“	Serum	Cel-miR-39-3p
miR-132-5p	↓	24 controls, 22 sALS	CSF	Cel-miR-39-3p (spiked)	Freischmidt et al. (2013)
miR-132-3p	↓	“	CSF	Cel-miR-39-3p
miR-143-3p	↓	“	CSF	Cel-miR-39-3p
miR-143-5p	↑	“	CSF	Cel-miR-39-3p
miR-574-5p	↑	“	CSF	Cel-miR-39-3p
miR-1825	↓	Discovery: 18 sALS, 16 controls	Serum	Cel-miR-39-3p (spiked)	Freischmidt et al. (2015)
		Validation: 20 sALS, 20 controls
		13 fALS, 13 controls
		15 AD, 13 controls
		13 HD, 15 controls
miR-1234-3p	↓	“	Serum	Cel-miR-39-3p
miR-206	↑	27 sALS, 25 controls	Serum	Average of miR-17-5p, miR-223-3p, miR-24	Waller et al. (2017)
miR-206	↑	36 disease mimics	Serum	Average of miR-17-5p, miR-223-3p, miR-24	Waller et al. (2017)
miR-143-3p	↑	“	Serum	Average of miR-17-5p, miR-223-3p, miR-24
miR-374-5p	↓	“	Serum	Average of miR-17-5p, miR-223-3p, miR-24
miR-183	↓	Discovery: 5 sALS, 5 controls	Leukocytes	RNU6	Chen et al. (2016)
miR-183	↓	Validation: 84 sALS, 24 PD, 31 controls	Leukocytes	RNU6	Chen et al. (2016)
miR-193b	↓	“	Leukocytes	RNU6
miR-451	↓	“	Leukocytes	RNU6
miR-3935	↑	“	Leukocytes	RNU6
miR-451	↓ (not validated)	Discovery: 8 sALS, 12 controls	Leukocytes	U49	De Felice et al. (2012)
miR-451	↓ (not validated)	Validation: 14 sALS, 14 controls	Leukocytes	U49	De Felice et al. (2012)
miR-338-3p	↑		Leukocytes	U49
miR-338-3p	↑	89 sALS, 75 controls	CSF, blood leukocytes, serum, and spinal cord	RNU6B in blood leukocytes, hsa-let-7 in serum, miR-24 in CSF	De Felice et al. (2014)
miR-4649-5p	↑	Discovery: 16 sALS, 10 controls	Plasma	Cel-miR-39-3p (spiked)	Takahashi et al. (2015)
miR-4649-5p	↑	Validation: 48 sALS, 47 controls, 30 PD	Plasma	Cel-miR-39-3p (spiked)	Takahashi et al. (2015)
miR-4299	↓	“	Plasma	Cel-miR-39-3p (spiked)
miR-181a-5p	↑	24 sALS, 24 controls	CSF	Cel-miR-39-3p (spiked) and miR-608, miR-328-3p	Benigni et al. (2016)
miR-21-5p	↓	“	CSF	Cel-miR-39-3p (spiked) and miR-608, miR-328-3p
miR-15b-5p	↓	“	CSF	Cel-miR-39-3p (spiked) and miR-608, miR-328-3p

AD, Alzheimer's Disease; ALS, amyotrophic lateral sclerosis; CSF, cerebrospinal fluid; fALS, familial amyotrophic lateral sclerosis; HD, Huntington Disease; PD, Parkinson's Disease; sALS, sporadic amyotrophic lateral sclerosis.

Skeletal muscle miRNAs are potential sources for the identification of candidate biomarkers. Specifically, profiling of skeletal muscle miRNA of ALS patients could potentially identify miRNAs that can be validated in plasma. Using this strategy, miRNA-424 and miR-206 have been found to be overexpressed and validated as prognostic markers in spinal onset of ALS (De Andrade et al., 2016). Studies on miRNAs have also been facilitated by animal models, especially the SOD1^G93A transgenic mouse. Identification of miRNAs altered in SOD1^G93A mice could be validated in cohorts of ALS patients. miR-206 is found to be upregulated in SOD1^G93A mice, and its downregulation in SOD1^G93A mice accelerates disease progression since its expression is responsible for sensing motor neuron injury and promoting regeneration of neuromuscular synapses (Williams et al., 2009).

The upregulation of miR-206 coincidences with the onset of neurological symptoms as healthy SOD1^G93A mice have similar miR-206 levels (Williams et al., 2009). MiR-206 has been validated in other studies as well (Toivonen et al., 2014; Tasca et al., 2016). Inhibition of miR-155 prolongs survival in SOD1^G93A mice (Koval et al., 2013), and it is found to be increased fivefold and twofold in mice and human spinal cords, respectively (Koval et al., 2013). Also, miR-155 is upregulated in the presymptomatic stage of SOD1^G93A mice and the upregulation of miR-155 is a consistent marker for both presymptomatic and symptomatic stages (Cunha et al., 2017).

Biochemical Markers for ALS

Protein-based biomarkers

Proteins are the most preferable biomarkers for routine clinical analysis. The reasons for this lay in the fact that they are generally stable, secreted in biofluids such as blood, and easily to be measured with immunoassays such as enzyme-linked immunosorbent assay (ELISA) or, more recently, with the advent of mass spectrometry-based approaches. Therefore, protein-based biomarkers have been investigated for fALS and sALS diagnosis and stratification. The most important, and promising molecules will be discussed later. Further, proteomic studies focusing on the identification of proteins in CSF from ALS patients have been performed and are a subject of recent reviews (Barschke et al., 2017; Caballero-Hernandez et al., 2016). However, proteomic studies are characterized by poor reproducibility, probably due to differences in the MS methods used.

Neurofilaments

The most extensively studied proteins as biomarkers for ALS are the neurofilament (NF) proteins. NFs are a unique class of intermediate filaments in being heteropolymers consisting of four subunits: the NF heavy (NFH), medium (NFM), and light chains (NFL) and the alpha-internexin in the CNS that is replaced by peripherin in the peripheral nervous system (Yuan et al., 2017). NFs can be subjected to several post-translational modifications with much interest in the diagnosis of motor neuron diseases, such as the phosphorylation. Especially the phosphorylated NF heavy chain (pNFH) and the NF light chain (NFL) have been studied as biochemical markers for ALS.

A previous large study encompassing 220 ALS patients (mainly sALS), 316 disease controls, and 50 disease mimics showed that pNFH has an added value as a diagnostic biomarker for ALS (Poesen et al., 2017). Concentrations of pNFH and NFL in CSF were lower in slow disease progressors, although with a poor prognostic performance with respect to disease progression rate. Their concentrations increased significantly as a function of the number of regions with both upper and lower motor involvement. However, patients with chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) and GBS also show increased pNFH levels. NFL is less specific for ALS since patients with frontotemporal dementia (FTD) and a subset of patients with CIDP and GBS had increased levels. Overall, pNFH in CSF is a better biomarker to detect disease mimics and other motor neuron diseases from NFL (Poesen et al., 2017).

A study with 222 patients with sALS, 20 with fALS, and 11 patients with primary lateral sclerosis (PLS) also showed increased pNFH levels in CSF of ALS and PLS patients compared with mimicking disorders. In this study, NFL could not be detected in serum samples of ALS patients whereas pNFH levels, although being detected in a subset of patients, their levels were not significantly different between patients with motor neuron diseases and disease mimic sera (Steinacker et al., 2016). In other studies, however, higher levels of pNFH in plasma of patients with ALS have been found compared with controls. In a study of 136 ASL patients and 104 healthy and neurological controls, increased pNFH levels in plasma were found (Lu et al., 2015b).

In addition, fast progressing ALS patients had higher NFH levels at an early stage and lower at the end-stage disease (Lu et al., 2015b). High levels of pNFH in the serum of ALS patients have also been found in an Australian cohort of 223 patients and 61 controls. The high serum levels of pNFH were inversely related with survival (McCombe et al., 2015). In a Czech study encompassing 15 sALS patients and 16 controls, higher expression of pNFH and chromogranin A was also found in CSF of patients (Kaiserova et al., 2017).

Two studies have been performed in China to assess the expression of pNFH in ALS (Chen et al., 2016a; Li et al., 2016). In CSF samples obtained from 40 sALS and 40 healthy controls, pNFH and chitotriosidase (CHIT) were found in sALS whereas cystatin C was decreased (Chen et al., 2016a). In the other study encompassing analysis of pNFH levels in paired CSF and plasma samples from 51 sALS, 12 multiple system atrophy (MSA), and 30 healthy controls, increased levels of pNFH in sALS but not in MSA and healthy samples were found (Li et al., 2016).

Finally, one study has detected antibodies against NFH, NFH aggregates, and NFH cleavage products in plasma of ALS patients (Lu et al., 2015). The role of NFs as biomarkers has been a subject of a detailed recent review (Cost and de Carvalho, 2016). A meta-analysis of NF protein expression in CSF and blood has also been performed (Xu et al., 2016). Both pNFH and NFL were elevated in the CSF of ALS patients compared with healthy controls. However, pNFH showed a trend for increased concentration in blood that approached the levels of statistical significance (p = 0.057), whereas NFL concentration in blood is higher in ALS patients.

Also, the heavy chain of CSF negatively correlated with disease duration and ALS Function Rating Scale-Revised (ALSFRS-R). The NFL levels in CSF were negatively correlated with disease duration (Xu et al., 2016). Finally, NFL levels in CSF were higher in ALS patients compared with patients with other neurological diseases with CNS involvement (Xu et al., 2016; Tortelli et al., 2012; Lu et al., 2015a). A very recent study also confirmed the role of NFL as a biomarker in CSF for ALS in an Italian cohort of patients (Gaiani et al., 2017).

C90RF72-specific biomarkers

The hexanucleotide GGGGCC (G₄C₂) repeat expansion in the C9ORF72 gene is the very characteristic for these gene carriers. The carriers are referred to as C9FTD/ALS (FTD). C9ORF72 is important since it is the most common pathogenic variant, and, therefore, treatments based on antisense oligonucleotides (Donnelly et al., 2013) and/or small molecules that target G₄C₂ (Su et al., 2014) are currently being developed. The RNAs transcribed by a variant C9ORF72 gene are translated to dipeptides, called DPRs (dipeptide repeat proteins), by repeat-associated non-ATG translation. Five polypeptides, poly(GA), poly(GR), poly(GP), poly(PR), and poly(PA), are produced by using the open reading frame from both directions. Although the most toxic polypeptides are the poly(GR) and poly(PR), the poly(GP) has been chosen as a potential biomarker for C9FTD/ALS. The reasons were that polyGP is a very frequent DPR; it is more soluble such that poly(GA) is the most frequent, is very stable, and is produced by both sense and antisense C9ORF72 transcripts (Balendra et al. 2017; Gendron et al., 2017b; Lehmer et al., 2017).

Due to somatic instability of G₄C₂ repeats, it is likely that patients with no expansion in blood could carry the gene expansions in the CNS. Therefore, a combination of genetic analysis and CSF determination of poly(GP) proteins could substantially benefit the clinical diagnosis (Balendra et al. 2017). Further, the detection of poly(GP) in peripheral blood mononuclear cells could facilitate the molecular diagnosis of C9FTD/ALS by analyzing the more easily obtained blood sample (Gendron et al., 2017b). Significant levels of poly(GP) in CSF were found in asymptomatic carriers compared with healthy controls and patients with other neurodegenerative diseases (Gendron et al., 2017b; Lehmer et al., 2017). poly(GP) is also a pharmacodynamic marker that is a molecular indicator of drug effect.

Another characteristic, specific for C9FTD/ALS patients, is the observation that trimethylated histones H3K9me3 and H3K27me3 bind to the promoter of the variant C9ORF72 that carries G₄C₂ repeats and these complexes are detectable in the blood of C9FTD/ALS patients, raising another opportunity of future diagnosis. These binding events were not observed in ALS patients with the absence of hexanucleotide repeats in C9ORF72 gene (Belzil et al., 2013). However, since these assays require chromatic immunoprecipitation, they will not be the first choice for routine clinical diagnosis.

Finally, a very recent study bridges the levels of pNFH in CSF with ALS carriers of C9ORF72 pathogenic expansions (Gendron et al., 2017a). Specifically, pNFH levels were higher in the CSF of C9ALS or C9ALS/FTD relative to asymptomatic carriers or to C9FTD-only patients or to non-C9ALS/FTD patients and to controls. A strong association was observed between higher pNFH levels and shorter survival after disease onset for both C9 expansion carriers and no carriers. Interestingly, pNFH CSF levels were higher in C9ALS than in other ALS patients and consistently, there was strong evidence of shorter survival after disease onset for C9ALS carriers compared with other ALS patients (Gendron et al., 2017a).

Transactive response DNA-binding protein 43 (TDP-43)

TDP-43 is a protein mostly found in the nucleus. Under stress conditions or when mutated, TDP-43 translocates to the cytoplasm where it is hyperphosphorylated and forms insoluble ubiquitin-positive aggregates (Neumann et al., 2006, 2009). TDP-43 is implicated in miRNA biogenesis through interaction with both Drosha and Dicer (Kawahara and Mieda-Sato, 2012). Variations in TDP-43 have been identified in both fALS and sALS forms (Sreedharan et al., 2008). The importance of TDP-43 protein in ALS comes from the fact that aggregates of SOD1 and FUS are mainly limited in patients carrying variants in the corresponding genes, whereas aggregates of TDP-43 are also found in patients with pathogenic variants of C9ORF72 and in sALS (Chew et al., 2015; Mackenzie et al., 2007; Sreedharan et al., 2008; Van Blitterswijk et al., 2012).

In addition, C9ORF72 pathogenic expansion in mice induces TDP-43 pathology (Chew et al., 2015). Further, the importance of TDP-43 in ALS biology is demonstrated by the fact that aggregated forms of TDP-43 enhance human endogenous retrovirus K (HERV-K) viral protein accumulation (Manghera et al., 2016a). The role of this endogenous retrovirus in ALS and its applications in diagnosis will be discussed later in this review. TDP-43 is a marker for both ALS and FTD (Mackenzie and Rademakers, 2008). However, a recent study, using typing with highly specific single-chain Fvs antibodies, showed that sALS and FTD have different TDP-43 forms and this could be exploited for designing novel diagnostic procedures in the future that could discriminate these two diseases (Williams et al., 2017).

A potential issue that complicates the usage of TDP-43 as a biomarker for ALS and has to be solved in the future comes from the fact that TDP-43 is also found in AD (Higashi et al., 2007; McAleese et al., 2017; Wilson et al., 2011), PD (Nakashima-Yasuda et al., 2007), Dementia with Lewy Bodies (Higashi et al., 2007), and HD (Schwab et al., 2008) and also in a small percentage of the aged population (McAleese et al., 2017). The latter finding merits further investigation since it suggests that TDP-43 pathology could participate in age-associated neurodegeneration.

TDP-43 has been found to be significantly increased in ALS compared with GBS with ELISA (Hosokawa et al., 2014). A combination of TDP-43 with tau and phosphorylated T181 tau has been suggested as an alternative marker for ALS (Bourbouli et al., 2017). Finally, TDP-43 has been detected to be upregulated in horses with Equine motor neuron disease (EMND) with immunohistochemistry in neuronal tissues. EMND is an ALS closely related disorder in horses and it differs from ALS in that it does not involve alterations in any portion of the upper motor neurons (El-Assaad et al., 2012).

Other protein biomarkers

NF proteins and TDP-43 have been extensively examined by various groups as potential biomarkers for ALS. However, there are other studies that have identified novel protein markers, with emerging interest in the diagnosis of ALS. In a recent study encompassing a group of 20 ALS patients and 20 healthy controls, follistatin, interleukin-1 alpha, and KLK5 protein levels are reduced in the CSF of ALS patients (Lind et al., 2016). The urinary neurotrophin receptor p75 extracellular domain (p75^ECD) protein (protein per mg creatinine) is found to be increased in the urine of ALS patients in a study that enrolled 54 ALS patients (12 with fALS) and 45 healthy controls. In addition, the higher levels of p75^ECD correlated with the lower survival (Shepheard et al., 2017). Increased levels of blood hemoglobin A1c are characteristic of sALS patients who have higher risk of mortality (Wei et al., 2017).

A Portuguese study found increased deformability of erythrocytes and increased acetylcholinesterase activity in erythrocytes from 82 ALS patients compared with 40 controls (Lima et al., 2016). This study also showed decreased NO efflux from red blood cells, and decreased intracellular nitrite in ALS patients (Lima et al., 2016). In conclusion, the major findings in protein-based biomarkers that are likely to enter the clinics in ALS are summarized in Table 2.

Table 2.

The Major Findings in Protein Biomarker Discovery for Amyotrophic Lateral Sclerosis

Protein	Expression	Participants	Sample	Reference
Neurofilaments^a	↑ pNFH↑ pNFH↑ pNFH↑ pNFH	220316 controls50 disease mimics222 sALS20 fALS11 PLS85 disease mimics28 AD26 PD33 PNP30 facial palsy136 ALS104 controls223 ALS61 controls	CSFCSFPlasmaSerum	Poesen et al. (2017)Steinacker et al. (2016)Lu et al. (2015b)McCombe et al. (2015)
Polypeptides related to C9ORF72	↑ Poly-GP↑ Poly-GP	No C9 expansion group48 controls57 ALS15 other diseasesC9 expansion group27 asymptomatic83 ALS or ALS-FTD24 other diseases11 C9FTD9 C9ALS9 C9 asymptomatic11 sFTD18 sALS24 AD14 PD28 controls	CSFCSF	Gendron et al. (2017a)Lehmer et al. (2017)
Urinary neurotrophin receptor p75 extracellular domain	↑	54 ALS (12 with fALS), 45 controls	Urine	Shepheard et al. (2017)

Only the studies with the largest number of participants are shown for simplicity; for a detailed description, please refer to the text.

PLS, primary lateral sclerosis; pNFH, phosphorylated neurofilament heavy chain; PNP, polyneuropathy; sFTD, sporadic frontotemporal dementia.

Volatolomics

Volatile organic compounds (VOCs) are organic chemicals with high vapor pressure at room temperature. VOCs present in blood or exhaled breath offer an alternative route to identification of novel biomarkers. A recent study has suggested the analysis of VOCs present in blood with GC-MS as an alternative diagnostic methodology for the diagnosis of ALS (Jiang et al., 2015). However, this study focused on transgenic SOD1^G93A animals and the findings remain to be validated in human specimens. In total, it detected 12 metabolites at the early stage of disease that mainly belong to alkanes, esters, ketones, aldehydes, and substituted benzenes. Their presence is probably due to oxidative stress.

Exhaled breath of ALS and cervical spondylotic myelopathy (CSM, a disease with similar ALS symptoms) patients was analyzed with GC-MS, and the following four chemicals were identified: monoammonium salt of carbamic acid, (S)-l-alanine ethylamide, N,N-dimethyl-guanidine, and (p-hydroxyphenyl)-phosphonic acid, which are exhaled at lower concentrations in ALS compared with CSM (Wang et al., 2016). However, their metabolic origin has not been confirmed.

Given the fact that volatolomic profiling has been previously performed in AD and PD and shown that these diseases have unique chemical signatures that can discriminate them (Bach et al., 2015; Tisch et al., 2013), in the near future, comparisons in the chemical profiles of ALS exhaled breath could lead to a new diagnostic method. Finally, analysis of human breath with electronic nose (cyranose 320) has been performed and enabled the rapid analysis of VOCs profiles in a real-time manner. It showed that these profiles can correctly discriminate ALS patients from controls (Dragonieri et al. 2016).

Metabolomics

Metabolic profiling studies in ALS have been performed but it should be mentioned that metabolite alterations are usually not specific for ALS and can be shared in multiple disorders. However, few molecules could have potential diagnostic applications as described later.

Albumin and creatinine

A cohort of 638 patients was used for identification and 122 for validation of various serum-based clinical chemical markers, including albumin, creatinine, triglycerides, total cholesterol etc. It was found that serum albumin and creatinine are independent markers of outcome for both men and women (Chiò et al., 2014). Lower albumin and creatinine levels are strongly related to worse clinical function at diagnosis (ALSFRS-R score and forced vital capacity [FVC]). Creatinine reflects the muscle waste, and albumin is associated with inflammatory state. No other parameters were found to have validity as markers for progression of ALS.

Importantly, the sensitivity and specificity values for predicting 1-year mortality indicated that albumin and creatinine have similar properties to the established ALS prognostic factors FVC and ALSFRS-R (Chiò et al., 2014). Decreased albumin in plasma and increased albumin in CSF of ALS patients compared with controls has been also found by using artificial-gel antibodies (Ghasemzadeh et al., 2008). The increased albumin levels in CSF were suggested to be due to blood-brain barrier malfunctioning (Ghasemzadeh et al., 2008). Lower creatinine was also found in a study with Japanese patients that included 92 ALS patients and 92 controls (Ikeda et al., 2012).

Oxidative stress-related small molecules

8-hydroxy-2′-deoxyguanosine (8OH2′dG) is a product of the oxidative injury to DNA. In fALS (n = 3) and sALS (n = 25) patients, higher concentrations of 8OH2′dG were found in CSF compared with controls (n = 16) (Ihara et al., 2005). 4-hydroxy-2,3-nonenal, a lipid peroxidation byproduct, is found to be increased in serum and CSF of sALS patients (n = 108) and increased in CSF in fALS patients (n = 14) compared with controls (n = 22) (Simpson et al., 2004).

Interestingly, in sALS, the levels of 4-hydroxy-2,3-nonenal in CSF were higher than in patients with other neurodegenerative diseases (including PD, AD etc., n = 19) (Simpson et al., 2004). Nitrotyrosine is another oxidative stress product produced by the reaction of peroxynitrite and tyrosine. Nitrotyrosine has been found to be increased in the CSF of ALS patients (Tohgi et al., 1999a) but also in AD patients (Tohgi et al., 1999b); therefore, its selectivity remains in question. In contrast, a recent study using a highly sensitive GC-MS method showed no change in the levels of nitrotyrosine in the CSF of ALS, AD patients, and controls (Ryberg et al., 2004). This finding indicates the necessity of assay optimization between different laboratories.

Metallomics

Metallomic studies using inductively coupled plasma mass spectrometry have been performed in ALS given the fact that environmental factors have been incriminated for certain sALS cases. The association of the metalloid Se with ALS has been extensively investigated in the past. In a group of 38 sALS patients and 38 controls, analysis of CSF showed higher concentrations of selenite and albumin-bound selenium in sALS patients (Vinceti et al., 2013). Peters et al. (2016) found inverse correlation in the levels of selenium in blood in a population of U.S. veterans between sALS (n = 163) and controls (n = 229).

Elevated selenium levels (in the form of selenite, selenate glutathione-peroxidase bound, and selenoprotein-P-bound) in the CSF were associated with patients carrying TUBA4A pathogenic variant (Mandrioli et al., 2017). In ALS patients carrying C9ORF72, SOD1, FUS, TARDBP, and ATXN2 pathogenic variants, increased levels of selenomethionine-bound selenium in CSF were found (Mandrioli et al., 2017). The main limitation of the study is that it only included one patient from each pathogenic variant except C9ORF72 and SOD1 that included three and two patients, respectively. However, it should be mentioned that TUBA4A is a very rare pathogenic variant linked to ALS with a prevalence of ∼1% in fALS (Smith et al., 2014) and 0.2% in sALS, making sample collection very difficult (Pensato et al., 2015; Smith et al., 2014).

A study of seven patients with sALS and five controls showed increased fivefold levels of Ni and twofold levels of Pb in the serum of sALS patients relative to controls and intriguingly lower As levels (De Benedetti et al., 2015). Further, increased serum Se was found in sALS, although not reaching the levels of statistical significance (De Benedetti et al., 2015). Decreased serum levels of As in a small group of geographically clustered sALS patients (six patients and five controls) have also been found in sALS patients, and the levels of As had a strong positive correlation with the duration of disease (De Benedetti et al., 2017).

The intriguing finding of lower As concentration in serum could be attributed to preferential accumulation in patients since As is known to cross the blood-brain barrier. Se concentration was inversely correlated with disease duration since higher concentrations were found in subjects with earliest onset. In whole-blood analysis, As and Cr concentrations positively correlated with disease duration (De Benedetti et al., 2017). The authors were not able to confirm the higher concentration of Ni and Pb previously described (De Benedetti et al., 2015), although they found that these higher levels did not reach the level of statistical significance. Maybe the small number of patients and controls is the main reason for this discrepancy. It should be noted that in the region of Genoa where the study was conducted, neural network analysis indicated association of tap water consumption with ALS (De Benedetti et al., 2017).

A more recent study that included 34 sALS patients from Sardinia, Italy, and 30 control individuals determined the levels of Ca, Cu, Fe, Mg, Se, and Zn in blood, hair, and urine. Ca and Cu in blood and Se and Zn in hair were significantly higher in ALS compared with controls, whereas the levels of Mg and Se in urine were decreased (Forte et al., 2017). Previously, the same group found increased blood concentrations of Al and Pb in ALS patients (n = 34) compared with controls (Bocca et al., 2015).

A Norwegian study found increased concentrations of Mn, Al, Cd, Co, Cu, Zn, Pb, V, and U in the CSF of ALS patients (n = 17) compared with controls (n = 10) (Roos et al., 2013). In addition, they demonstrated higher concentrations of these metals in the CSF of ALS patients in contrast to plasma concentrations, suggesting that mechanisms of accumulation may exist. Further, this group demonstrated changes in the levels of Se or As in the CSF between ALS patients and controls, whereas Ni and Fe were found to be increased in ALS without reaching statistically significant levels. Notably, uranium was found in the CSF of 47% of ALS cases whereas it was completely undetectable in controls. Any connection between uranium and ALS is currently missing and this may merit further investigation in the future.

Higher levels of Pb in the whole blood of ALS patients compared with controls have also been demonstrated in other studies (Garzillo et al., 2014). The role of Hg in inducing sALS has been recently highlighted with the successful treatment of an ALS patient as previously discussed (Mangelsdorf et al., 2017). Taking these studies together, it appears that Pb could be associated with ALS and may represent a marker although not alone but in combination with others. For the other metals, some studies have provided different results from others. These could be attributed to differences in sample preparations, speciation (e.g., in the case of selenium), different “ALS subtypes,” ethnicity etc. Therefore, more detailed studies are needed to address their roles or diagnostic potential in ALS. Another issue that has to be noted is the fact that metals have been shown to associate not only with ALS but also with PD and AD, such as Al and AD and Fe and PD (Cicero et al., 2017).

Viromics

It is known that HIV and HTLV-1-infected patients may occasionally develop ALS-like syndrome (Matsuzaki et al., 2000; Verma and Berger, 2006). Both serum and CSF from HIV-negative patients with ALS demonstrated reverse transcriptase activity at levels similar to HIV-infected individuals (MacGowan et al., 2007; McCornick et al., 2008). Further, increased RT activity has been found in the serum of first-degree relatives that may indicate that RT activity could be derived from human endogenous retroviruses (Steele et al., 2005). HERV-K sequences have been identified in autopsy brain material from sporadic and fALS patients (Douville et al., 2011; Li et al., 2015). HERV-K polymerase expression was higher in ALS compared with patients with chronic systemic illness whereas it was absent in PD patients and accidental death individuals.

Especially, a characteristic pattern of expression from loci HML-2 (7q34) and HML-3 (7q36.1) was identified for ALS that can discriminate it from systemic illness and can have potential diagnostic applications (Douville et al., 2011). It has been proven in vivo that the envelope protein of HERV-K drives the development of ALS symptoms (Li et al., 2015). Neuroinflammation appears to be the main mechanism to activate HERV-K expression; therefore, it is possible that certain chemicals, infections, inflammatory reactions, or cancer induce ALS through endogenous HERV-K activation (Manghera et al., 2016a,b). Importantly, HERV-K expression strongly correlated with TDP-43 expression (Douville et al., 2011). It was shown that expression of HERV-K is regulated by TDP-43 that binds to the long terminal repeats region of the virus (Li et al., 2015). The env gene of the retrovirus is responsible for the observed neurodegeneration (Li et al., 2015).

ALS associated with endogenous retrovirus could account for specific ALS cases, and, therefore, the inclusion of biomarkers for this subset of ALS patients in a greater ALS panel of biomarkers would improve the diagnosis and potentially the treatment of such patients. Viral infections could be treated with retroviral agents, although there is a need in the development of new antiviral agents since the treatment of HIV-negative ALS patients exhibiting increased reverse transcriptase activity with indinavir was not successful (MacGowan et al., 2007). On the other hand, in HIV-induced ALS-like complete recovery was observed after highly active antiretroviral therapy (MacGowan et al., 2001). Already the search for anti-retrovirals that could inhibit the HERV-K has been initiated (Tyagi et al., 2017). Table 3 summarizes the emerging omics-es applied for ALS.

Table 3.

A Brief Summary of the Emerging Omics-es Applied to Amyotrophic Lateral Sclerosis

Omics-es	Sample	Biomarker	References
Volatolomics	Exhaled air, blood	Monoammonium salt of carbamic acid, (S)-l-alanine ethylamide, N,N-dimethyl-guanidine, and (p-hydroxyphenyl)-phosphonic acid, othersOther small molecules (alkanes, aromatics, etc.)	Wang et al. (2016)Dragonieri et al. (2016)Jiang et al. (2015)
Metallomics	Blood, serum, plasma, CSF	MetalloidsSeleniumArsenicHeavy metalsMercury^aUranium	e.g., Vinceti et al. (2013); Peters et al. (2016)e.g., De Benedetti et al. (2017)Mangelsdorf et al. (2017)Roos et al. (2013) and other findings in the text
Viromics	Brain tissue but CSF has been shown to exert high reverse transcriptase activity and could potentially be used	HERV-K retrovirus sequences	Douville et al. (2011); McCornick et al. (2008); MacGowan et al. (2007) and other findings in the text

Indicates successful treatment based on chelation.

HERV-K, human endogenous retrovirus K.

Diagnostic implications

ALS is a highly heterogeneous disease where various genes with apparently different functions and environmental factors have been incriminated. This huge diversity poses both diagnostic and therapeutic challenges that need to be urgently addressed. In this direction, the identification of novel biomarkers that will assist in the rapid diagnosis of the disease and its classification to certain subtypes is of great importance. Currently, NF proteins, especially pNFH, appear to hold great promise as ALS biomarkers based on numerous studies conducted throughout the globe.

It remains to be seen whether NF proteins are increased in presymptomatic patients. If this holds true, then it may boost the diagnostic filed by applying population screening programs in a similar manner to, for example, prostate-specific antigen (PSA/KLK3) used in screening for prostate cancer. However, due to the recognized heterogeneity of ALS, not a single biomarker could be applied. In this direction, assays for detection of dipeptide repeat proteins and TDP43 have a near future potential to be included in a panel for ALS diagnosis.

Conclusions and Future Perspectives

ALS is a highly heterogeneous disease with lack of satisfactory therapeutic strategies. Even the fALS is caused by a variety of genes that participate in completely different cellular processes. Due to the high heterogeneity, it appears that there cannot be a universal biomarker but instead a panel of biomarkers that can also assist in ALS stratification. Another issue that complicates the biomarker discovery efforts seems to be the lack of standardization and quantification of methods, the lack of a universal protocols for sample collection and storage, etc. that lead to inconsistent results. Therefore, the time to move to certain guidelines that will enable the usage of standard methods between all laboratories has come to achieve knowledge translation.

From the already conducted studies, it appears that NF proteins pNFH and NLF will have great impact on ALS diagnosis. Indispensable for stratification remains the detection of poly(GP) protein and the detection of HERV-K as has been extensively discussed previously in this review. The analysis for heavy metal intoxication, although not a primary choice, should be always kept in mind as a potential diagnostic/stratification procedure and selection of the appropriate therapy. What needs to be done is to identify easily measured biomarkers for the asymptomatic ALS patients. This will revolutionize ALS diagnosis and therapy. The road for this has already been opened by the demonstration that asymptomatic C9ALS patients have detectable poly(GP) protein in CSF.

Another characteristic of ALS is the presence of aggregated proteins, for example, SOD1, FUS, and TDP-43. The development of antibodies that can specifically recognize these aggregated forms would greatly facilitate the molecular diagnosis and stratification of ALS. On the other hand, multi-omics data integration will be required to obtain a more complete and accurate of the various ALS “subtypes” and to derive new potential biomarkers. Although integration of multi-omics-es data is not an easy task, currently a lot of effort has been put into this (Bersanelli et al., 2016; Huang et al., 2017).

Footnotes

Author Disclosure Statement

The authors declare that no conflicting financial interests exist.

Abbreviations Used

References

Bach

, Gold

, Mengel

, et al. (2015). Measuring compounds in exhaled air to detect Alzheimer's disease and Parkinson's disease. PLoS One, 10, e0132227.

Balendra

, Moens

, and Isaacs

. (2017). Specific biomarkers for C9orf72 FTD/ALS could expedite the journey towards effective therapies. EMBO Mol Med, 7, 853–855.

Barschke

, Oeckl

, Steinacker

, Ludolph

, and Otto

. (2017). Proteomic studies in the discovery of cerebrospinal fluid biomarkers for amyotrophic lateral sclerosis. Expert Rev Proteomics, 14, 769–777.

Belzil

, Bauer

, Prudencio

, et al. (2013). Reduced C9orf72 gene expression in C9FTD/ALS is caused by histone trimethylation, an epigenetic event detectable in blood. Acta Neuropathologica, 126, 895–905.

Benigni

, Ricci

, Jones

, Giannini

, Al-Chalabi

, and Battistini

. (2016). Identification of miRNAs as potential biomarkers in cerebrospinal fluid from amyotrophic lateral sclerosis patients. Neuromolecular Med, 18, 551–560.

Bersanelli

, Mosca

, Remondini

, et al. (2016). Methods for the integration of multi-omics data: Mathematical aspects BMC. Bioinformatics, 17(Suppl 2), S15.

Bocca

, Forte

, Oggiano

, et al. (2015). Level of neurotoxic metals in amyotrophic lateral sclerosis: A population-based case-control study. J Neurol Sci, 359, 11–17.

Bourbouli

, Rentzos

, Bougea

, et al. (2017). Cerebrospinal fluid TAR DNA-binding protein 43 combined with tau proteins as a candidate biomarker for amyotrophic lateral sclerosis and frontotemporal dementia spectrum disorders. Dement Geriatr Cogn Disord, 44, 144–152.

Brown

, and Al-Chalabi

. (2017). Amyotrophic lateral sclerosis. N Eng J Med, 377, 162–172.

10.

Caballero-Hernandez

, Toscano

, Cejudo-Guillen

, et al. (2016). The ‘omics’ of amyotrophic lateral sclerosis. Trends Mol Med, 22, 53–67.

11.

Cicero

, Mostile

, Vasta

, et al. (2017). Metals and neurodegenerative diseases. A systematic review. Environ Res, 159, 82–94.

12.

Chen

, Chen

, Wei

, et al. (2016a). Assessment of a multiple biomarker panel for diagnosis of amyotrophic lateral sclerosis. BMC Neurol, 16, 173.

13.

Chen

, Wei

, Chen

, et al. (2016b). Aberration of miRNAs expression in leukocytes from sporadic amytrophic lateral sclerosis. Front Mol Neurosci, 69, 1–11.

14.

Chew

, Gendron

, Prudencio

, et al. (2015). C9ORF72 repeat expansions in mice cause TDP-43 pathology, neuronal loss, and behavioral deficits. Science, 348, 1151–1154.

15.

Chiò

, Calvo

, Bovio

, et al. (2014). Amyotrophic lateral sclerosis outcome measures and the role of albumin and creatinine: A population-based study. JAMA Neurol, 71, 1134–1142.

16.

Chiò

, Logroscino

, Traynor

, et al. (2013). Global epidemiology of amyotrophic lateral sclerosis: A systematic review of the published literature. Neuroepidemiology, 41, 118–130.

17.

Cost

, and de Carvalho

. (2016). Emerging molecular biomarker targets for amyotrophic lateral sclerosis. Clin Chem Acta, 455, 7–14.

18.

Cunha

, Santos

, Gomes

, et al. (2017). Downregulated Glia interplay and increased miRNA-155 as promising markers to track ALS at an early stage. Mol Neurobiol (in press). DOI: 10.1007/s12035-017-0631-2

19.

De Andrade

, de Albuquerque

, Avansini

, et al. (2016). MicroRNAs-424 and 206 are potential prognostic markers in spinal onset amyotrophic lateral sclerosis. J Neurol Sci, 368, 19–24.

20.

De Benedetti

, Lucchini

, Del Bò

, et al. (2017). Blood trace metals in a sporadic amyotrophic lateral sclerosis geographical cluster. Biometals, 30, 355–365.

21.

De Benedetti

, Lucchini

, Marocchi

, et al. (2015). Serum metal evaluation in a small cohort of amytrophic lateral sclerosis patients reveals high levels of thiophylic species. Peptidomics, 2, 29–34.

22.

De Felice

, Annunziata

, Fiorentino

, et al. (2014). miR-338-3p is over-expressed in blood, CSF, serum and spinal cord from sporadic amyotrophic lateral sclerosis patients. Neurogenetics, 15, 243–253.

23.

De Felice

, Guida

, Coppola

, De Mieri

, and Cotrufo

. (2012). A miRNA signature in leukocytes from sporadic amyotrophic lateral sclerosis. Gene, 508, 35–40.

24.

Donnelly

, Zhang

, Pham

, et al. (2013). RNA toxicity from the ALS/FTD C9ORF72 expansion in mitigated by antisense intervention. Neuron, 80, 415–428.

25.

Douville

, Liu

, Rothstein

, and Nath

. (2011). Identification of active loci of a human endogenous retrovirus in neurons of patients with amyotrophic lateral sclerosis. Ann Neurol, 69, 141–151.

26.

Dragonieri

, Quaranta

, Carratu

, et al. (2016). An electronic nose may sniff out amyotrophic lateral sclerosis. Respir Physiol Neurobiol, 232, 22–25.

27.

El-Assaad

, Di Bari

, Yasuda

, et al. (2012). Differential expression of TAR DNA-binding protein (TDP-43) in the central nervous system of horses afflicted with equine motor neuron disease (EMND): A preliminary study of a potential pathologic marker. Vet Res Commun, 36, 221–226.

28.

Forte

, Bocca

, Oggiano

, et al. (2017). Essential trace elements in amyotrophic lateral sclerosis (ALS): Results in a population of a risk area of Italy. Neurol Sci, 38, 1609–1615.

29.

Freischmidt

, Müller

, Ludolph

, and Weishaupt

. (2013). Systemic dysregulation of TDP-43 binding microRNAs in amyotrophic lateral sclerosis. Acta Neuropathol Commun, 1, 42.

30.

Freischmidt

, Müller

, Zondler

, et al. (2014). Serum microRNAs in patients with genetic amyotrophic lateral sclerosis and pre-manifest mutation carriers. Brain, 137, 2938–2950.

31.

Freischmidt

, Müller

, Zondler

, et al. (2015). Serum microRNAs in sporadic amytrophic lateral sclerosis. Neurobiol Aging, 132, 391–411.

32.

Gaiani

, Martinelli

, Bello

, et al. (2017). Diagnostic and prognostic biomarkers in amyotrophic lateral sclerosis: Neurofilament light chain levels in definite subtypes of disease. JAMA Neurol, 74, 525–532.

33.

Garzillo

, Lamberti

, Genovese

, et al. (2014). Blood lead, manganese, and aluminum levels in a regional Italian cohort of ALS patients: Does aluminum have an influence?. J Occup Environ Med, 56, 1062–1066.

34.

Gendron

, C9ORF72 Neurofilament Study Group, Daughrity

, et al. (2017a). Phosphorylated neurofilament heavy chain: A biomarker of survival for C9ORF72-associated amyotrophic lateral sclerosis. Ann Neurol, 82, 139–146.

35.

Gendron

, Chew

, Stankowski

, et al. (2017b). Poly(GP) proteins are a useful pharmacodynamic marker for C9ORF72-associated amyotrophic lateral sclerosis. Sci Transl Med, 9, eaai7866.

36.

Ghasemzadeh

, Nyberg

, and Hjertén

. (2008). Highly selective artificial gel antibodies for detection and quantification of biomarkers in clinical samples. II. Albumin in body fluids of patients with neurological disorders. J Sep Sci, 31, 3954–3958.

37.

Higashi

, Iseki

, Yamamoto

, et al. (2007). Concurrence of TDP-43, tau and alpha-synuclein pathology in brains of Alzheimer's disease and dementia with Lewy bodies. Brain Res, 1184, 284–294.

38.

Hosokawa

, Arai

, Yamashita

, et al. (2014). Differential diagnosis of amyotrophic lateral sclerosis from Guillain-Barré syndrome by quantitative determination of TDP-43 in cerebrospinal fluid. Int J Neurosci, 124, 344–349.

39.

Huang

, Chaudhary

, and Garmire

. (2017). More is better: Recent progress in multi-omics data integration methods. Front Genet, 8, 84.

40.

Ihara

, Nobukuni

, Takata

, and Hayabara

. (2005). Oxidative stress and metal content in blood and cerebrospinal fluid of amyotrophic lateral sclerosis patients with and without a Cu, Zn-superoxide dismutase mutation. Neurol Res, 27, 105–108.

41.

Ikeda

, Hirayama

, Takazawa

, Kawabe

, and Iwasaki

. (2012). Relathionships between disease progression and serum levels of lipid, urate, creatinine and ferritin in Japanese patients with amyotrophic lateral sclerosis: A cross-sectional study. Intern Med, 51, 1501–1508.

42.

Jeromin

, and Bowser

. (2017). Biomarkers in neurodegeneration diseases. Adv Neurobiol, 15, 491–528.

43.

Jiang

, Wang

, Ren

, et al. (2015). Blood volatile organic compounds as potential biomarkers for amyotrophic lateral sclerosis: An animal study in the SOD1 G93A mouse. J Mol Neurosci, 55, 167–173.

44.

Kaiserova

, Grambalova

, Otruba

, et al. (2017). Cerebrospinal fluid levels of chromogranin A and phosphorylated neurofilament heavy chain are elevated in amyotrophic lateral sclerosis. Acta Neurol Scand, 136, 360–364.

45.

Kawahara

, and Mieda-Sato

. (2012). TDP-43 promotes microRNA biogenesis as a component of the Drosha and Dicer complexes. Proc Natl Acad Sci USA, 109, 3347–3352.

46.

Keller

, Ferrucci

, Singleton

, et al. (2014). Genome-wide analysis of the heritability of amyotrophic lateral sclerosis. JAMA Neurol, 71, 1123–1134.

47.

Koval

, Shaner

, Zhang

, et al. (2013). Method for widespread microRNA-155 inhibition prolongs survival in ALS-model mice. Hum Mol Genet, 22, 4127–4135.

48.

Lehmer

, Oecld

, Weishaupt

, et al. (2017). Poly-GP in cerebrospinal fluid links C9orf72-associated dipeptide repeat expression to the asymptomatic phase of ALS/FTD. EMBO Mol Med, 9, 859–868.

49.

, Ren

, Zhu

, Yang

, Zhang

, and Huang

. (2016). Phosphorylated neurofilament heavy chain levels in paired plasma and CSF of amyotrophic lateral sclerosis. J Neurol Sci, 367, 269–274.

50.

, Lee

, Henderson

, et al. (2015). Human endogenous retrovirus-K contributes to motor neuron disease. Sci Transl Med, 7, 307ra153.

51.

Lima

, Pinto

, Napoleão

, et al. (2016). Identification of erythrocyte biomarkers in amyotrophic lateral sclerosis. Clin Hemorheol Microcirc, 63, 423–437.

52.

Lind

, Wu

, Freyhuit

, et al. (2016). A multiplex protein panel applied to cerebrospinal fluid reveals three new biomarker candidates in ALS but none in neuropathic pain patients. PLoS One, 11, e0149821.

53.

, Macdonald-Wallis

, Gray

, et al. (2015a). Neurofilament light chain: A prognostic biomarker in amyotrophic lateral sclerosis. Neurology, 84, 2247–2257.

54.

, Petzold

, Topping

, et al. (2015b). Plasma neurofilament heavy chain levels and disease progression in amyotrophic lateral sclerosis: Insights from a longitudinal study. J Neurol Neurosurg Psychiatry, 86, 565–573.

55.

MacGowan

, Scelsa

, Imperato

, Liu

, Baron

, and Polsky

. (2007). A controlled study of reverse transcriptase in serum and CSF of HIV-negative patients with ALS. Neurology, 68, 1944–1946.

56.

MacGowan

, Scelsa

, and Waldron

. (2001). A subacute ALS syndrome with new HIV infection and complete response to antiretroviral therapy. Neurology, 57, 1094–1097.

57.

Mackenzie

, Bigio

, Ince

, et al. (2007). Pathological TDP-43 distinguishes sporadic amyotrophic lateral sclerosis from amyotrophic lateral sclerosis with SOD1 mutations. Ann Neurol, 61, 427–434.

58.

Mackenzie

IRA

, and Rademakers

. (2008). The role of TDP-43 in amyotrophic lateral sclerosis and frontotemporal dementia. Curr Opin Neurol, 21, 693–700.

59.

Mandrioli

, Michalke

, Solovyev

, et al. (2017). Elevated levels of selenium species in cerebrospinal fluid of amyotrophic lateral sclerosis patients with disease-associated gene mutations. Neurodegener Dis, 17, 171–180.

60.

Mangelsdorf

, Walach

, and Mutter

. (2017). Healing of amyotrophic lateral sclerosis: A case report. Complement Med Res, 24, 175–181.

61.

Manghera

, Ferguson-Parry

, and Douville

. (2016a). TDP-43 regulates endogenous retrovirus-K viral protein accumulation. Neurobiol Dis, 94, 226–236.

62.

Manghera

, Ferguson-Parry

, Lin

, and Douville

. (2016b). NF-kB and IRF1 induce endogenous retrovirus K expression via interferon-stimulated response elements in its 5′ long terminal repeat. J Virol, 90, 9338–9349.

63.

Matsuzaki

, Nakagawa

, Nagai

, et al. (2000). HTLV-1-associated myelopathy (HAM)/tropical spastic paraparesis (TSP) with amyotrophic lateral sclerosis-like manifestations. J Neurovirol, 6, 544–548.

64.

McAleese

, Walker

, Erskine

, Thomas

, McKeith

, and Atterns

. (2017). TDP-43 pathology in Alzheimer's disease, dementia with Lewy bodies and ageing. Brain Pathol, 27, 472–479.

65.

McCombe

, Pfluger

, Singh

, Lim

CYH

, Airey

, and Henderson

. (2015). Serial measurements of phosphorylated neurofilament-heavy in the serum of subjects with amyotrophic lateral sclerosis. J Neurol Sci, 353, 122–129.

66.

McCornick

, Brown

Jr , Cudkowicz

, et al. (2008). Quantification of reverse transcriptase in ALS and elimination of a novel retroviral candidate. Neurology, 70, 278–283.

67.

Morris

, Al-Sarraj

, Schwab

, et al. (2001). A clinical and pathological study of motor neurone disease in Guam. Brain, 124, 2215–2222.

68.

Nakashima-Yasuda

, Uryu

, Robinson

, et al. (2007). Co-morbidity of TDP-43 proteinopathy in Lewy body related diseases. Acta Neuropathol, 114, 221–229.

69.

Neumann

, Kwong

, Lee

, et al. (2009). Phosphorylation of S409/410 of TDP-43 is a consistent feature in all sporadic and familial forms of TDP-43 proteinopathies. Acta Neuropathol, 117, 137–149.

70.

Neumann

, Sampathu

, Kwong

, et al. (2006). Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science, 314, 130–134.

71.

Paez-Colasante

, Figueroa-Romero

, Sakowski

, Goutman

, and Feldman

. (2015). Amyotrophic lateral sclerosis: Mechanism and therapeutics in the epigenomic era. Nat Rev Neurol, 11, 266–279.

72.

Pensato

, Tiloca

, Corrado

, et al. (2015). TUBA4A gene analysis in sporadic amyotrophic lateral sclerosis: Identification of novel mutations. J Neurol, 262, 1376–1378.

73.

Peters

, Beard

, Umbach

, et al. (2016). Blood levels of trace metals and amyotrophic lateral sclerosis. Neurotoxicology, 54, 119–126.

74.

Poesen

, De Schaepdryer

, Stubendorff

, et al. (2017). Neurofilament markers for ALS correlate with extent of upper and lower motor neuron disease. Neurology, 88, 1–8.

75.

Renton

, Chiò

, and Traynor

. (2014). State of play in amyotrophic lateral sclerosis genetics. Nat Neurosci, 17, 17–23.

76.

Robelin

, Gonzalez De Aguilar

. (2014). Blood biomarkers for amyotrophic lateral sclerosis: Myth or reality?. Biomed Res Int, 2014, 525097.

77.

Ronnevi

, and Conradi

. (1984). Increased fragility of erythrocytes from amyotrophic lateral sclerosis (ALS) patients provoked by mechanical stress. Acta Neurol Scad, 69, 20–26.

78.

Roos

, Vesterberg

, Syversen

, Flaten

, and Nordberg

. (2013). Metal concentrations in cerebrospinal fluid and blood plasma from patients with amyotrophic lateral sclerosis. Biol Trace Elem Res, 151, 159–170.

79.

Ryberg

, Söderling

, Davidsson

, Blennow

, Caidahl

, and Persson

. (2004). Cerebrospinal fluid levels of free 3-nitrotyrosine are not elevated in the majority of patients with amyotrophic lateral sclerosis or Alzheimer's disease. Neurochem Int, 45, 57–62.

80.

Schwab

, Arai

, Hasegawa

, Yu

, and McGeer

. (2008). Colocalization of transactivation-responsive DNA-binding protein 43 and Huntingtin in inclusions of Huntington disease. J Neuropathol Exp Neurol, 67, 1159–1165.

81.

Shepheard

, Wuu

, Cardoso

, et al. (2017). Urinary p75ECD: A prognostic, disease progression, and pharmacodynamic biomarker in ALS. Neurology, 88, 1–7.

82.

Simpson

, Henry

, Henkel

, Smith

, and Appel

. (2004). Increased lipid peroxidation in sera of ALS patients: A potential biomarker of disease burden. Neurology, 62, 1758–1765.

83.

Smith

, Ticozzi

, Fallini

, et al. (2014). Exome-wide rare variant analysis identifies TUBA4A mutations associated with familial ALS. Neuron, 84, 324–331.

84.

Sreedharan

, Blair

, Tripathi

, et al. (2008). TDP-43 mutations in familial and sporadic amyotrophic lateral sclerosis. Science, 319, 1668–1672.

85.

Steele

, Al-Chalabi

, Ferrante

, et al. (2005). Detection of serum reverse transcriptase activity in patients with ALS and unaffected blood relatives. Neurology, 64, 454–458.

86.

Steinacker

, Feneberg

, Weishaupt

, et al. (2016) Neurofilaments in the diagnosis of motoneuron diseases: A prospective study on 455 patients. J Neurol Neurosurg Psychiatry, 87, 12–20.

87.

, Zhang

, Gendron

, et al. (2014). Discovery of a biomarker and lead small molecules to target r(GGGGCC)-associated defects in c9FTD/ALS. Neuron, 83, 1043–1050.

88.

Takahashi

, Hama

, Matsushima

, et al. (2015). Identification of plasma microRNAs as a biomarker of sporadic Amyotrophic Lateral Sclerosis. Mol Brain, 8, 67.

89.

Tasca

, Pegoraro

, Merico

, and Angelini

. (2016). Circulating microRNAs as biomarkers of muscle differentiation and atrophy in ALS (2016). Clin Neuropathol, 35, 22–30.

90.

Tisch

, Schlesinger

, Ionescu

, et al. (2013). Detection of Alzheimer's and Parkinson's disease from exhaled breath using nanomaterial-based sensor. Nanomedicine, 8, 43–56.

91.

Tohgi

, Abe

, Yamazaki

, Murata

, Ishizaki

, and Isobe

. (1999a). Remarkable increase in cerebrospinal fluid 3-nitrotyrosine in patients with sporadic amyotrophic lateral sclerosis. Ann Neurol, 46, 129–131.

92.

Tohgi

, Abe

, Yamazaki

, Murata

, Ishizaki

, and Isobe

. (1999b). Alterations of 3-nitrotyrosine concentration in the cerebrospinal fluid during aging and in patients with Alzheimer's disease. Neurosci Lett, 269, 52–54.

93.

Toivonen

, Manzaro

, Oliván

, Zaragoza

, García-Redondo

, and Osta

. (2014). MicroRNA-206: A potential circulating biomarker candidate for amyotrophic lateral sclerosis. PLoS One, 9, e89065.

94.

Tortelli

, Ruggieri

, Cortese

, et al. (2012) Elevated cerebrospinal fluid neurofilament light levels in patients with amyotrophic lateral sclerosis: A possible marker of disease severity and progression. Eur J Neurol, 19, 1561–1567.

95.

Tyagi

, Li

, Parades

, Bianchet

, and Nath

. (2017). Inhibition of human endogenous retrovirus-K by antiretroviral drugs. Retrovirology, 14, 21.

96.

Van Blitterswijk , DeJesus-Hernandez

, and Rademakers

. (2012). How do C9ORF72 repeat expansions cause ALS and FTD: Can we learn from other non-coding repeat expansion disorders?. Curr Opin Neurol, 25, 689–700.

97.

van Es

, Hardiman

, Chio

, et al. (2017). Amyotrophic lateral sclerosis. Lancet, 390, 2084–2098.

98.

Van Hoecke

, Schoonaert

, Lemmens

, et al. (2012). EPHA4 is a disease modifier of amyotrophic lateral sclerosis in animal models and in humans. Nat Medicine, 18, 1418–1422.

99.

Verma

, and Berger

. (2006). ALS syndrome in patients with HIV-1 infection. J Neurol Sci, 40, 59–64.

100.

Vinceti

, Solovyev

, Mandrioli

, et al. (2013). Cerebrospinal fluid of newly diagnosed amyotrophic lateral sclerosis patients exhibits abnormal levels of selenium species including elevated selenite. Neurotoxicology, 38, 25–32.

101.

Waller

, Goodall

, Milo

, et al. (2017). Serum miRNAs miR-206, 143-3p and 374b-5p as potential biomarkers for amyotrophic lateral sclerosis (ALS). Neurobiol Aging, 55, 123–131.

102.

Wang

, Li

, Jiang

, et al. (2016). Comparative analysis of VOCs in exhaled breath of amyotrophic lateral sclerosis and cervical spondylotic myelopathy patients. Sci Rep, 6, 26120.

103.

Wei

, Chen

, Cao

, et al. (2017). Blood hemoglobin A1c levels and amyotrophic lateral sclerosis survival. Mol Neurodegener, 12, 69.

104.

Williams

, Valdez

, Moresi

, et al. (2009). MicroRNA-206 delays ALS progression and promotes regeneration of neuromuscular synapses in mice. Science, 326, 1549–1554.

105.

Williams

, Khan

, Harris

, Ravits

, and Sierks

. (2017). TDP-43 protein variants as biomarkers in amyotrophic lateral sclerosis. BMC Neurosci, 18, 20.

106.

Wilson

, Dugger

, Dickson

, and Wang

D-S

. (2011). TDP-43 in aging and Alzheimer's disease-a review. Int J Clin Exp Pathol, 4, 147–155.

107.

, Henderson

, David

, and McCombe

. (2016). Neurofilaments as biomarkers for amyotrophic lateral sclerosis: A systematic review and meta-analysis. PLoS One, 11, e0164625.

108.

Yuan

, Rao

, Veeranna, and Nixon

. (2017). Neurofilaments and neurofilament proteins in health and disease. Cold Spring Harb Perspect Biol, 9, pii: a018309.

109.

Zoing

, Burke

, Pamphlett

, et al. (2006). Riluzole therapy for motor neurone disease: An early Australian experience (1996–2002). J Clin Neurosci, 13, 78–83.