# Etiological Heterogeneity in Sporadic ALS: Current Hypotheses and Subgroup Proposals

**Date:** 2026-05-04  
**Scope:** This document reviews the strongest current hypotheses regarding etiological heterogeneity within sporadic amyotrophic lateral sclerosis (sALS), which accounts for ~90% of all ALS cases and lacks identified causal mutations. It evaluates proposed subgroups based on biomarkers, clinical phenotype, and environmental exposure, and identifies sub-populations that may be tractable for focused clinical trials.

---

## 1. Context: The Sporadic ALS Problem

Sporadic ALS is clinically and pathologically heterogeneous. No single mutation explains the majority of cases, and the 90% figure encompasses a likely mixture of multigenic susceptibility, epigenetic modification, environmental triggers, and stochastic events. The field increasingly operates under the hypothesis that "ALS" is a syndrome with multiple underlying etiologies rather than a unified disease. Heterogeneity is the dominant driver of clinical trial failure because homogeneous mechanisms are assumed when heterogeneous populations are enrolled. This document therefore emphasizes **subgroups with both biological plausibility and clinical tractability**.

---

## 2. Mechanistically Defined Molecular Subtypes

### 2.1 Transcriptomic/Molecular Subtypes (Marriott et al., 2023)

**Hypothesis:** Sporadic ALS patients cluster into three molecularly distinct pathogenic subtypes that reflect the major proposed ALS mechanisms.

**Methodology:** Unsupervised hierarchical clustering on the 5,000 most variably expressed autosomal genes from motor cortex post-mortem tissue (N=112 from KCL BrainBank).

**Three subgroups identified:**

| Subtype | Core Pathway | Cell-type Enrichment | Associated Clinical Features |
|---------|-------------|---------------------|------------------------------|
| **Cluster 1: Synaptic & Neuropeptide Signalling**| Synaptic transmission, neuropeptide signalling, glutamate metabolism | Oligodendrocytes, inhibitory neurons | Earlier disease onset, specific synaptic gene expression |
| **Cluster 2: Oxidative Stress & Apoptosis** | ROS response, apoptosis, mitochondrial stress | Astrocytes, excitatory neurons | Distinct age-at-onset profiles |
| **Cluster 3: Neuroinflammation** | Immune response, cytokine signalling, microglial activation | Microglia, astrocytes | Potentially more aggressive course |

**Evidence supporting:**
- **Cross-validation:** Linear discriminant analysis (LDA) models applied to independent TargetALS US motor cortex (N=93) and blood datasets (Italian N=15, Dutch N=397) achieved **80–90% assignment probability** for each subtype, with AUC 0.88 ± 0.10 for distinguishing ALS from controls.
- **Motor cortex specificity:** Expression signatures perfectly discriminated motor cortex from occipital cortex and cerebellum, confirming disease-relevance rather than tissue artefact.
- **Accessible tool:** A web portal is available (https://alsgeclustering.er.kcl.ac.uk) suggesting clinical adoption pathway.

**Evidence gaps:**
- Blood-based classification (more clinically practical) has lower sample sizes for validation (Italian N=15).
- No formal clinical trial has prospectively stratified by these clusters.
- The biological interpretation relies on bulk tissue inference; single-cell deconvolution could refine assignments.

**Trial tractability:** ⭐⭐⭐⭐ Very high. The three-cluster framework already has cross-cohort validation in both cortex and blood. A trial could prospectively assign treatment arms based on real-time blood transcriptomics (or approximated via plasma proteomics), enabling mechanism-matched therapy—e.g., anti-inflammatory agents for Cluster 3, antioxidants/metabolism modulators for Cluster 2, and synaptic stabilizers for Cluster 1.

### 2.2 Proteomic Biotypes (Pasternack, Paulsen & Nath, 2025)

**Hypothesis:** Machine learning on combined clinical and demographic data predicts three biologically distinct ALS biotypes.

**Evidence:** Published in *European Journal of Human Genetics* (August 2025). Three biotypes identified through supervised machine learning. Specific marker sets for each biotype remain to be fully characterized, but the study establishes that biotype assignment from clinical/demographic features alone is feasible.

**Trial tractability:** ⭐⭐⭐ Moderate to high. If biotype-specific treatments can be identified, clinical trial enrichment could occur at the bedside without requiring transcriptomic or proteomic assays.

---

## 3. Biomarker-Based Stratification: Neurofilament Light Chain (NfL)

### 3.1 Rapid vs. Slow Progressor Stratification

**Hypothesis:** NfL is a non-specific but powerful readout of axonal/neuronal damage load, and baseline NfL levels stratify sALS patients into prognostically distinct subgroups.

**Evidence:**
- NfL serum/plasma concentrations at baseline **correlate strongly with survival time** and discriminate fast, intermediate, and slow progressors (Steinacker et al., 2017; Gille et al., 2019).
- In the RNS60 Phase 2 post-hoc analysis (Pupillo et al., 2024), patients with **low baseline NfL** showed the most pronounced survival benefit from therapy (median survival >4 years vs 3.3 years on placebo; 1.9 years in high NfL group regardless of treatment).
- Similarly, low baseline MCP-1 (monocyte chemoattractant protein-1) identified an immunologically "quieter" subgroup with better treatment response.

**Counter-evidence/considerations:**
- NfL is **not specific** to ALS; it rises in multiple sclerosis, traumatic brain injury, and other neurodegenerative diseases.
- High NfL may represent a later, more advanced disease state where neurodegeneration is already irreversible, rather than a purely biological subtype. It is partly a function of disease duration and stage.
- Baseline stratification alone may not capture evolving heterogeneity.

**Trial tractability:** ⭐⭐⭐⭐ Very high. NfL is already measureable in serum via commercially available assays (Simoa platform). It has been used to enrich trials for rapidly progressors (e.g., SOD1-antisense trials) and can be used to exclude non-responders. It is among the most immediately actionable stratifiers in sALS.

---

## 4. Proteomic Plasma Panels and Presymptomatic Risk Stratification

### 4.1 Multi-Protein Diagnostic/Prognostic Panel (2025 Nature Medicine)

**Hypothesis:** A plasma proteomic panel can distinguish ALS from mimics, track progression, and even identify presymptomatic risk years before onset.

**Evidence:**
- A high-throughput plasma proteomic analysis using Olink Explore of 516 serial samples from 33 phenoconverters, 35 ALS patients, 10 presymptomatic carriers, and 59 controls identified **81 proteins whose concentrations change prior to phenoconversion** (Ran et al., 2025, medRxiv; later associated with August 2025 Nature Medicine study by NIH/NIA).
- A core panel of 19 proteins predicted phenoconversion over 0.5–5 year horizons with **AUC 0.80–0.89** and estimated time-to-phenoconversion with mean absolute error 1.6 years.
- C9orf72 repeat expansion carriers showed 8 elevated proteins compared to non-carriers, suggesting **genetic subgroup-specific proteomic profiles**.

**Evidence gaps:**
- Replication in large, fully independent cohorts (e.g., sporadic cases from non-specialist centers) is still needed.
- Correlation with histopathological subtypes (TDP-43 type, C9orf72 status, SOD1) is incomplete.

**Trial tractability:** ⭐⭐⭐⭐ Very high. The presymptomatic window (~1.6 year prediction) creates opportunities for very early intervention trials (prevention trials) in enriched populations. C9orf72 vs. non-C9 proteomic divergence offers a rational path for mechanism-specific therapy.

---

## 5. TDP-43 Pathology-Based Subgroups

### 5.1 C9orf72 vs. Non-C9 Sporadic ALS

**Hypothesis:** C9orf72 hexanucleotide repeat expansion (HRE)—the most common genetic cause of ALS/FTD—produces a pathologically and biologically distinct TDP-43 proteinopathy compared to TARDBP/TDP-43-linked sporadic ALS, despite overlap at the syndromic level.

**Evidence:**
- C9orf72-ALS cases display **abundant p62-positive, TDP-43-negative inclusions** in cortex, hippocampus, and cerebellum (Troakes et al., 2012), a pattern distinct from sALS-FTLD-TDP.
- Western blots show high p62 and low TDP-43 with no high-molecular-weight smearing in C9orf72-ALS, versus typical TDP-43 proteinopathy in sporadic cases.
- Cross-syndrome overlap: some C9orf72 carriers with purely motor ALS never develop FTD; others remain presymptomatic into their 70s, suggesting interacting modifiers (Kortazar-Zubizarreta et al., 2023).

**Trial implications:**
- C9orf72-ALS trials can now use **antisense oligonucleotides** (ASOs) targeting C9orf72 repeat RNA and its toxic dipeptide repeat proteins.
- For sporadic, non-C9 cases, TDP-43 pathology remains the unifying feature, but the upstream trigger differs. C9orf72-negative sALS may respond differently to TDP-43-targeted ASOs (e.g., targeting TDP-43 itself is challenging due to essential nuclear function).
- **UNC13A cryptic exon peptides** (detectable in CSF and blood) specifically mark TDP-43 loss-of-function and have emerged as a **genetic stratification tool** distinguishing C9orf72 from sporadic ALS (Anjum et al., 2025).

**Trial tractability:** ⭐⭐⭐⭐ Very high for C9orf72 subpopulation (gene-specific ASOs). For non-C9, tractability depends on finding a druggable upstream lesion (e.g., cryptic splicing correction).

### 5.2 TDP-43 Pathological Staging (pTDP-43)

**Hypothesis:** Post-mortem TDP-43 pathology stage correlates with molecular subtype and disease duration, enabling retrospective biological subgrouping.

**Evidence:**
- Stage 4 pTDP-43 pathology shows distinct cerebellar gene and protein expression changes (Grima et al., 2025).
- STMN2 cryptic exon expression is pTDP-43 pathology-specific, confirmed in 22 sporadic ALS cases with staged pathology.

**Limitations:**
- Staging is post-mortem; equivalent in vivo biomarkers are needed for trial stratification.
- Not all sALS cases have TDP-43 pathology (e.g., SOD1-ALS and FUS-ALS are TDP-43-negative).

---

## 6. Clinical Phenotype Subgroups

### 6.1 The Eight Classic Phenotypes

Clinicians have long subdivided ALS into distinct motor phenotypes:
- **Classic** (limb-onset, mixed UMN/LMN)
- **Bulbar** (onset in speech/swallowing)
- **Flail arm** (proximal arm LMN-predominant)
- **Flail leg** (distal leg LMN-predominant)
- **Pyramidal/PLS** (UMN-predominant)
- **Progressive muscular atrophy (PMA)** (pure LMN)
- **Respiratory-onset**

**Prognostic differences are robust:**
| Phenotype | Survival | Evidence |
|-----------|----------|----------|
| Flail arm / flail leg | **Better** (more indolent, longer survival) | Wolf et al. (2014); Wei et al. (2018) |
| Bulbar / Classic / Respiratory | **Worse** (shorter survival, faster FVC decline) | Wolf et al. (2014) |
| Pyramidal / PLS | Most indolent (decades in some cases) | Meyer et al. (2025) |

**Molecular underpinnings:**
- Transcriptomic analysis of PBMCs from 48 sALS patients across 5 phenotypes (Classic, Bulbar, Flail Arm, Flail Leg, Pyramidal) found that **only one gene (Y3_RNA, a misc_RNA component of the Ro60 ribonucleoprotein)** was universally upregulated; most differentially expressed genes were **phenotype-specific** (Dragoni et al., 2025).

**OPM consensus classification (2025):**
A new three-determinant anatomical classification (Onset × Propagation × Motor-neuron dysfunction) standardizes phenotype assignment for clinical trials.

**Trial tractability:** ⭐⭐⭐ Moderate. Phenotypes are easy to classify clinically but represent downstream manifestations, not upstream etiologic splits. Trials have historically shown benefits only when phenotype interacts with mechanism (e.g., NIV has greatest survival benefit in bulbar-onset ALS; Berlowitz et al., 2016).

### 6.2 Cortical Excitability Stratification (MEP:CMAP Ratio)

**Hypothesis:** The ratio of motor evoked potential (MEP) to compound muscle action potential (CMAP) reflects corticospinal excitability and varies systematically across phenotypes.

**Evidence (Ranieri et al., 2025, *Annals of Neurology*):**
- In 743 multicenter ALS patients, the MEP:CMAP ratio categorized patients as hyperexcitable, normal, or hypoexcitable.
- **Hyperexcitability** predominated in LMN-predominant (flail, classic, bulbar) forms; **hypoexcitability** in UMN-predominant (pyramidal, PLS) forms.
- **Hyperexcitable patients had significantly shorter survival** (HR 1.84, 95% CI 1.12–3.03, p=0.016) even when tested within 1 year of onset.

**Trial tractability:** ⭐⭐⭐ Moderate. TMS is scalable but requires specialist equipment and expertise. The MEP:CMAP ratio provides a real-time, non-invasive biological stratifier at the point of care and is a strong candidate for enriching trials with early, active cortical hyperexcitability.

---

## 7. Environmental Exposure Subgroups

### 7.1 Heavy Metal and Solvent Exposure

**Hypothesis:** A subset of sALS reflects heavy-metal–induced genotoxicity, particularly in individuals with occupational or environmental exposure.

**Evidence:**
- A US National ALS Registry matched case-control study (N=267 pairs; Wu et al., 2024) found:
  - **Vinyl chloride** (Q4 vs Q1: aOR 6.00, 95% CI 1.87–19.25)
  - **2,4-dinitrotoluene** (aOR 5.45, 95% CI 1.53–19.36)
  - **Cyanide** (aOR 4.34, 95% CI 1.52–12.43)
  - **Cadmium** (aOR 3.30, 95% CI 1.11–9.77)
  - **Organic/chlorinated solvents** (aOR 2.62, 95% CI 1.003–6.85)
- Residential air selenium showed **inverse association** with ALS (protective).
- A systematic review identified 15 metals showing genotoxicity relevant to ALS pathogenesis (Kim et al., 2025), with particular focus on aluminum, arsenic, cadmium, chromium, cobalt, copper, iron, lead, manganese, mercury, nickel, selenium, uranium, vanadium, and zinc.

**Counter-evidence:**
- Many historical case-control studies found **no association** (e.g., Gresham et al., 1986).
- Exposure assessment relies on recall-based questionnaires or geographic linkage, introducing misclassification bias.
- Effect sizes are moderate; the attributable fraction remains small relative to overall sALS incidence.

**Subgroup implications:**
- If a toxic-metal–exposed subgroup exists, interventions targeting **copper homeostasis**, **metal chelation**, or **antioxidant pathways** could be subgroup-specific.
- **Copper hypothesis (Min et al., 2024):** Both copper toxicity (ROS, aggregate seeding) and copper deficiency (cuproenzyme loss) coexist in ALS, creating a vicious cycle. Restoration of copper balance via copper delivery agents + chaperones + serine is proposed.

**Trial tractability:** ⭐⭐ Low to moderate. Environmental exposure is difficult to measure retrospectively with accuracy. A prospective trial would require detailed exposure questionnaires plus biomarker confirmation (urinary/serum metal levels). The most realistic path is to enrich a trial with individuals showing high metal exposure *and* blood biomarkers of oxidative stress.

### 7.2 BMAA / Cyanobacterial Toxin Hypothesis

**Hypothesis:** Chronic exposure to β-N-methylamino-L-alanine (BMAA), a non-protein amino acid produced by cyanobacteria, drives a subset of sALS through protein misincorporation and TDP-43 aggregation.

**Evidence:**
- The **Western Pacific ALS/Parkinsonism-Dementia Complex (ALS/PDC)** epidemic on Guam, the Kii Peninsula (Japan), and Papua (Indonesia) is the strongest natural experiment. Cycad seed exposure (BMAA, cycasin/MAM) correlated with disease clusters that have now **largely disappeared** after dietary/exposure cessation (Spencer, 2020, 2022).
- BMAA reproduces ALS/PDC neuropathology (TDP-43 inclusions, cortical degeneration) in animal models when administered perinatally or chronically (Arnold et al., 2023).
- BMAA is produced by cyanobacteria worldwide (including in Australian freshwater blooms); human exposure occurs through drinking water and seafood bioconcentration.
- A two-hit mouse model (TDP-43 Q331K + chronic low-dose BMAA) developed a motor phenotype absent with either insult alone (Arnold et al., 2023).

**Counter-evidence:**
- Critics (Chernoff et al., 2017) argue the BMAA–ALS causal relationship is not established, citing analytical inconsistencies in BMAA detection and lack of robust epidemiological linkage outside Guam.
- The Guam cluster had unique demographic features not replicable in other sALS populations.

**Serine supplementation hypothesis:**
- BMAA is structurally related to L-serine and may compete for protein incorporation. A Phase I clinical trial of L-serine in ALS is complete; Phase II trials are underway based on this mechanistic hypothesis (Spencer et al., 2022).

**Trial tractability:** ⭐⭐ Moderate. If validated, a trial would target individuals with documented BMAA exposure (residence near cyanobacterial blooms, elevated urinary/serum BMAA) combined with low-grade cognitive/motor deficits. However, exposure assessment in most healthcare settings is currently impractical.

---

## 8. Immune and Autoimmune Subgroups

### 8.1 Autoantibody Subset

**Hypothesis:** A minority of sporadic ALS cases harbour pathogenic autoantibodies against neuronal ion channels, creating a functionally autoimmune motor neuropathy.

**Evidence:**
- **Voltage-gated calcium channel (VGCC) antibodies** are detected in ~6.9% of MND patients (Donaldson et al., 2016).
- ALS IgGs bind the α1 subunit of L-type VGCCs (Kimura et al., 1994; Offen et al., 1998), increase calcium currents in motoneuron hybrid cell lines (Mosier et al., 1995), inhibit dopamine release (Offen et al., 1998), and alter neuromuscular junction function (Gonzalez et al., 2011).
- **Passive transfer models:** Transfer of ALS IgG into mice induces calcium-dependent motoneuron injury (Smith et al., 1996).

**Counter-evidence:**
- Donaldson et al. (2016) found **no difference in progression rate or survival** between antibody-positive and antibody-negative MND patients.
- Antibody titers are often low and may represent epiphenomena.
- Low prevalence means the attributable fraction is small even if the association is causal.

**Trial implications:**
- If a true autoimmune subgroup exists, it may respond to **immunoglobulin depletion, immunosuppression, or plasma exchange**—therapies that have been largely abandoned in broad ALS trials after negative results.

**Trial tractability:** ⭐⭐ Low. The prevalence is too low for a scalable trial without a dedicated screening program. However, a retrospective analysis of prior immune-modulatory trial failures stratified by VGCC antibody status might reveal a hidden responder subgroup.

---

## 9. ALS-FTD Spectrum Subgroup

### 9.1 Cognitive/Behavioral Subtypes

**Hypothesis:** Frontotemporal dysfunction (executive impairment, behavioral changes) defines a distinct pathological subtype within sALS, with implications for prognosis and trial enrichment.

**Evidence:**
- ALS and FTD share TDP-43 proteinopathy and C9orf72 HRE as unifying molecular pathologies.
- ALS-FTD patients survive significantly less than pure ALS (Wolf et al., 2014).
- Cognitive screening (ECAS) is now standard in many ALS clinics and identifies a subset with executive dysfunction that may reflect more widespread neurodegeneration.
- The Edinburgh Cognitive and Behavioural ALS Screen differentiates language, executive, and behavioral domains, enabling enrichment for or exclusion of the FTD spectrum.

**Trial implications:**
- ALS-FTD and non-FTD ALS may require different primary endpoints and possibly different interventions.
- The C9orf72 carrier state is enriched in ALS-FTD compared to pure ALS.

**Trial tractability:** ⭐⭐⭐ Moderate. Cognitive screening is feasible at trial entry but adds complexity. The C9orf72-specific subset is already being targeted separately.

---

## 10. Summary: Most Tractable Subgroups for Focused Trials

| Subgroup | Basis | Validation Strength | Trial Feasibility | Suggested Approach |
|----------|-------|--------------------|---------------------|-------------------|
| **Molecular subtype 1 (Synaptic)** | Blood/motor cortex transcriptomics | **Strong** (cross-continental AUC ~0.88) | **High** | Synaptic stabilizers, glutamate modulators (e.g., perampanel, memantine) in enriched cohort |
| **Molecular subtype 2 (Oxidative/Apoptotic)** | Blood/motor cortex transcriptomics | **Strong** | **High** | Antioxidants (edaravone already approved; subgroup-specific dosing), mitochondrial enhancers |
| **Molecular subtype 3 (Neuroinflammatory)** | Blood/motor cortex transcriptomics | **Strong** | **High** | Anti-microglial agents (e.g., masitinib), CSF1R inhibitors, IL-6 pathway blockers |
| **Low baseline NfL** | Serum NfL (Simoa) | **Very strong** (multiple trials) | **Very high** | Enrich trials for individuals with low NfL where drug effect may be detectable (e.g., neuroprotective agents, anti-inflammatory therapies) |
| **C9orf72-negative, TDP-43-positive** | Genetics + UNC13A cryptic exon | **Strong** (genetic + plasma biomarker) | **High** | Target TDP-43 loss-of-function (e.g., STMN2 splice correction, TDP-43 stabilizers) |
| **C9orf72-positive** | Repeat-primed PCR | **Very strong** (genetic gold standard) | **Very high** | C9orf72-targeted ASOs (already in trials: WVE-004, BIIB078) |
| **Hyperexcitable MEP:CMAP** | Transcranial magnetic stimulation | **Moderate-Strong** (N=743 cohort) | **Moderate** | Enrich with cortical hyperexcitability for GABA/glutamate modulator trials |
| **Flail arm / Flail leg / PLS** | Clinical phenotype | **Strong** (decades of evidence) | **High** | Disease-modifying trials in slowly progressing populations where longer observation windows reduce sample size needs |
| **Environmental (metal-exposed)** | Exposure history + biomarkers | **Moderate** (case-control data, recall bias) | **Low-Moderate** | Subgroup analysis within trials or dedicated chelation/antioxidant studies |
| **Autoimmune (VGCC-positive)** | Antibody testing | **Weak-Moderate** | **Low** | Retrospective analysis of immune trial data; small-scale proof-of-concept |

---

## 11. Recommended Priority for Future Trial Design

### Immediate (1–2 years)
1. **Prospective blood-transcriptomic or plasma-proteomic molecular subtyping** at trial enrollment to assign patients to mechanism-matched arms. The Marriott three-cluster framework or the new 19-protein presymptomatic panel (Ran et al., 2025; Nature Medicine) are the leading candidates.
2. **NfL stratification** in all neuroprotective trials to enrich for low-NfL populations with preserved neuronal substrate and to avoid futility in late-stage axonal burnout.
3. **C9orf72 genetic testing** combined with UNC13A cryptic exon testing to confirm TDP-43 pathology status in non-C9 sporadic cases.

### Medium-term (2–5 years)
4. **Phenotype-guided adaptive trials:** Use OPM classification and MEP:CMAP ratios to build biomarker-driven adaptive platforms where intervention is matched to cortical excitability state.
5. **Environmental exposure enrichment:** Develop validated biomarker panels for BMAA, cyanotoxins, and heavy metals. Enrich prevention trials in high-risk exposed populations.
6. **Single-cell transcriptomics in CSF/blood** to refine the macrophage/microglial activation state within the "inflammatory" molecular subtype, enabling more precise anti-inflammatory targeting.

---

## 12. Key References

1. Marriott H, et al. "Unsupervised machine learning identifies distinct ALS molecular subtypes in post-mortem motor cortex and blood expression data." *Acta Neuropathologica Communications* 11, 183 (2023). doi:10.1186/s40478-023-01686-8
2. Pasternack N, Paulsen O, Nath A. "Machine learning predicts distinct biotypes of amyotrophic lateral sclerosis." *European Journal of Human Genetics* 33, 1290–1299 (2025). doi:10.1038/s41431-025-01920-y
3. Dragoni F, et al. "Whole transcriptome analysis of unmutated sporadic ALS patients' peripheral blood reveals phenotype-specific gene expression signature." *Neurobiology of Disease* 210, 106823 (2025). doi:10.1016/j.nbd.2025.106823
4. Pupillo E, et al. "Long-term survival of participants in a phase II randomized trial of RNS60 in amyotrophic lateral sclerosis." *Brain, Behavior, and Immunity* (2024). doi:10.1016/j.bbi.2024.08.044
5. Anjum F, et al. "Emerging biomarkers in amyotrophic lateral sclerosis: from pathogenesis to clinical applications." *Frontiers in Molecular Biosciences* 12, 1608853 (2025). doi:10.3389/fmolb.2025.1608853
6. Ranieri F, et al. "Cortical excitability as a prognostic and phenotypic stratification biomarker in amyotrophic lateral sclerosis." *Annals of Neurology* 98, 801–813 (2025). doi:10.1002/ana.27305
7. Wu F, et al. "Exposure to ambient air toxicants and the risk of amyotrophic lateral sclerosis (ALS): A matched case control study." *Environmental Research* 243, 117719 (2024). doi:10.1016/j.envres.2023.117719
8. Kim WW, et al. "Metal-Induced Genotoxic Events: Possible Distinction Between Sporadic and Familial ALS." *Toxics* 13, 493 (2025). doi:10.3390/toxics13060493
9. Min J-H, et al. "Copper toxicity and deficiency: the vicious cycle at the core of protein aggregation in ALS." *Frontiers in Molecular Neuroscience* 17, 1408159 (2024). doi:10.3389/fnmol.2024.1408159
10. Spencer PS. "Parkinsonism and motor neuron disorders: Lessons from Western Pacific ALS/PDC." *Journal of the Neurological Sciences* 434, 120021 (2022). doi:10.1016/j.jns.2021.120021
11. Månberg A, et al. "Altered perivascular fibroblast activity precedes ALS disease onset." *Nature Medicine* 27, 1051–1063 (2021). doi:10.1038/s41591-021-01295-9
12. Troakes C, et al. "An MND/ALS phenotype associated with C9orf72 repeat expansion: abundant p62-positive, TDP-43-negative inclusions." *Neuropathology* 32, 241–250 (2012). doi:10.1111/j.1440-1789.2011.01286.x
13. Wei Q-Q, et al. "Clinical and prognostic features of ALS/MND in different phenotypes." *Brain Research Bulletin* 143, 213–218 (2018). doi:10.1016/j.brainresbull.2018.09.005
14. Meyer T, et al. "Motor phenotypes of amyotrophic lateral sclerosis - a three-determinant anatomical classification." *Neurological Research and Practice* 7, 21 (2025). doi:10.1186/s42466-025-00389-w
15. Donaldson R, et al. "Clinical significance of cation channel antibodies in motor neuron disease." *Muscle & Nerve* 54, 228–231 (2016). doi:10.1002/mus.25046
16. Grima N, et al. "Multi-region brain transcriptomic analysis of amyotrophic lateral sclerosis reveals widespread RNA alterations." *Molecular Neurodegeneration* 20, 8 (2025). doi:10.1186/s13024-025-00820-5

---

*Document generated for scientific review. Classification of heterogeneous subgroups in sporadic ALS is an active, evolving field. All trial tractability assessments are provisional and subject to validation in prospective cohorts.*