Finding Common Ground on the Site of Onset of Amyotrophic Lateral Sclerosis

Abstract

The fundamental origin of amyotrophic lateral sclerosis (ALS) has remained an enigma since its earliest description as a relentlessly progressive degeneration with prominent neuromuscular manifestations that are associated with upper and lower motor neuron dysfunction. Although this remains the hallmark of ALS, a significant proportion of patients will also demonstrate one or more features of frontotemporal dysfunction, including a frontotemporal dementia (FTD). Understanding whether these 2 seemingly disparate syndromes are simply reflective of the co-occurrence of 2 distinct pathologic processes or the clinical manifestations of a common pathophysiologic derangement involving the brain more widely has gripped contemporary ALS researchers. Supporting a commonality of causation, both ALS and FTD show an alteration in the metabolism of TAR DNA-binding protein 43, marked by a shift in nucleocytoplasmic localization alongside a broad range of neuronal cytoplasmic inclusions consisting of pathologic aggregates of RNA-binding proteins. Similarly, several disease-associated or disease-modifying genetic variants that are shared between the 2 disorders suggest shared underlying mechanisms. In both, a prominent glial response has been postulated to contribute to non-cell-autonomous spread. A more contemporary hypothesis, however, suggests that syndromes of cortical and subcortical dysfunction are driven by impairments in discrete neural networks. This postulates that such networks, including networks subserving motor or cognitive function, possess unique and selective vulnerabilities to either single molecular toxicities or combinations thereof. The co-occurrence of one or more network dysfunctions in ALS and FTD is thus a reflection not of unique neuroanatomic correlates but rather of shared molecular vulnerabilities. The basis of such shared vulnerabilities becomes the fulcrum around which the next advances in our understanding of ALS and its possible therapy will develop.

The last several decades have seen an explosion in our understanding of the pathophysiology of ALS, moving it well beyond the classical clinicopathologic correlates that date to the early works of Charcot, Joffroy, and others¹ into a far more complex conceptualization driven by a deeper knowledge of the interconnectedness of the CNS across anatomic, cellular, and molecular levels. This article reflects the substance of a point: counterpoint presentation by the 2 authors at the International UKMND ALS/MND Symposium in December 2021 during which we were asked to debate the origins of ALS as beginning either distally (i.e., lower motor neuron [LMN] in nature) or cortically. In considering this, we argued that the premise of 2 seemingly distinct theoretical sites of origin fails to take account of this complexity. By considering this, a concept of ALS arises that is more holistic and takes into account this interconnectedness. We conclude that the focus of ALS research should be redirected if a mechanism-based, effective, disease-modifying treatment is to be found.

ALS as a Syndrome of Motor Neuron Degeneration

Charcot's description of ALS remains the basis for current diagnostic criteria in which the focus remains on progressive neurogenic muscular weakness and wasting, often asymmetric, with fasciculations and usually associated with upper motor neuron (UMN) signs.^1-4 ALS is a brain degeneration that especially involves the UMN and LMN,⁵ although LMN degeneration is the primary driver of prognosis.^6,e1 Brain pathology in ALS is strikingly anterior, largely without involvement of the posterior brain.⁵ There is corticomotor and prefrontal cortical pathway pathology with atrophy of the precentral cortical motor strip. ALS thus involves the effector parts of the brain. For example, there are no recognized clinical abnormalities in visual function, parietal lobe function, or hearing, and there are no cerebellar signs. Primary sensory modalities are always spared, although sometimes there is subtle alteration of vibration sense. There is now also consensus that frontotemporal lobar degeneration, which can be manifest as a frontotemporal dementia (FTD) or more subtle frontotemporal dysfunction not fulfilling the full diagnostic FTD criteria, affects nearly 50% of people with classical ALS.⁷ In contrast, FTD commonly presents without neuromuscular manifestations, although these may develop later. Hence, the phenotype of the ALS syndrome is now considered to be much broader than a pure motor system degeneration.

The idea that ALS may have more than a single site of origin and then spread from that point is attractive but not new. Indeed, Charcot² himself distinguished between deuteropathic atrophy, involving the UMN and LMN, and protopathic atrophy, affecting the LMN alone. He wondered, in the former case, without supportive evidence, whether corticospinal degeneration might precede anterior horn cell degeneration. This conceptualization has continued to find some support clinically and especially through the observation of the split hand phenomenon of selective lateral (thenar) hand muscle wasting in ALS.⁸ It has been suggested that this is consistent with a primary motor cortical origin for ALS, but there are other possible noncortical causes, especially more prominent usage, or differential distal motor end-plate susceptibility to denervation among small hand muscles.⁹ In contrast, a century ago, Gowers¹⁰ was especially critical of the speculation that a UMN lesion could give rise to progressive LMN dysfunction because UMN lesions of various causes are not ordinarily followed by progressive denervation.

However, the concept that the neurodegenerative process of ALS can be defined solely on the basis of clinical markers of motor system degeneration has been challenged by the neuropathologic observation that pathways such as gracile columns and spinocerebellar tracts in the cervical cord, not primarily motor, may also be involved. With increasingly sensitive markers of corticospinal tract degeneration (i.e., inflammatory markers such as CD68 immunoreactivity), approximately 50% of cases presenting with clinically isolated LMN involvement (progressive muscular atrophy) will show evidence of corticospinal tract degeneration.¹¹

A major turning point in our understanding of ALS and FTD occurred with the discovery that TAR DNA-binding protein 43 (TDP-43) is the primary constituent of intraneuronal inclusions in both ALS and approximately 50% of FTD cases (FTD-TDP).^12,13 TDP-43, normally predominantly nuclear, undergoes a significant nucleocytoplasmic shift with nuclear depletion and an associated formation of neuronal cytoplasmic inclusions in Betz cells of the motor cortex, medullary motor neurons, and spinal motor neurons in >95% of ALS cases.¹⁴ There is an associated loss of a broad range of TDP-43 functions that include impairments in RNA transport, transcription, and splicing.¹⁵ Of particular relevance to this discussion of the site of origin of ALS and the relationship to FTD has been the discovery of the association of the UNC13A gene with ALS, ALS with frontotemporal dysfunction, and FTD-TDP.^16,17 The protein encoded by UNC13A is involved in the priming of presynaptic vesicles; disruption of this function disrupts the exocytosis of a broad range of neurotransmitters and the function of neural networks (further discussed later).^e2 Depletion of UNC13A in ALS and FTD-TDP is driven by the presence of a cryptic exon in UNC13A, in which a nonconserved intronic sequence is included in the mature UNC13A RNA, leading to a loss of the transcript and protein. Nuclear depletion of TDP-43, as seen in both ALS and FTD-TDP, results in cryptic exon inclusion in UNC13A transcripts through a failure of TDP-43–mediated cryptic exon splicing and the resultant loss of UNC13A.^18,19 As will be discussed, this linkage between a loss of nuclear TDP-43 and the resultant inclusion of a cryptic exon in UNC13A provides a powerful basis for moving from a purely anatomic basis of the origins of ALS to one in which molecular pathology is considered critical.

Lower Motor Neuron Involvement as the Major Clinical Feature of ALS

Axonal neurofilamentous inclusions of >20 μm in diameter are prominent in spinal motor neurons in ALS and are thought to contribute to altered axoplasmic transport.^20,21,e3 Neuropathologic studies have generally confirmed Greenfield's assertion that ALS is associated with central and central-distal dying back axonopathy,²² implying degeneration in both first- and second-order motor neurons,¹² although more prominently in second-order motor neurons in the spinal cord. Focal segmentally distributed neurogenic change, detected clinically or by EMG, precedes more generalized muscular atrophy. Motor neurons innervating fast contracting, fatigable motor units are most susceptible. Fasciculations, signifying neuronal or axonal hyperexcitability, are a characteristic early feature and are usually more widely distributed than neurogenic change. Distal upper limb muscles are more affected than lower limb muscles, and tongue involvement is frequent. The earliest detectable abnormality in EMG studies is increased neuromuscular jitter, followed by impulse blocking, motor unit denervation, and partial reinnervation.²³ Disease spread between adjacent motor neurons remains poorly understood, although the concept that the disease process of ALS can spread between neuronal populations and indeed between neuronal and non-neuronal cell populations has been largely confirmed by clinical, neurophysiologic, and neuropathologic studies for the majority of cases.^e4-8

Motor cortex (UMN) hyperexcitability in ALS was suggested by studies in SOD1 familial ALS using cortical magnetic stimulation, recording from small hand muscles, in which cortical hyperexcitability (or reduced inhibition) was the earliest detectable abnormality, even before typical clinical features were evident, perhaps even preceding EMG evidence of LMN degeneration.²⁴ However, the methods used for detecting UMN and LMN dysfunction in these studies differ both in their technology and sensitivity. Therefore, whether this cortical reduced inhibition truly represents causation or, more likely, is simply an early feature of the physicobiological anomaly characteristic of ALS is uncertain.

Non-Cell-Autonomous Factors

An astrocytic dysfunction hypothesis is attractive and links to the insight that neuronal degeneration in ALS is non-cell-autonomous.²⁵ Although ALS is typically classified as a primary motor neuronal degeneration, motor neurons exist in close functional relationship with glial cells, and astrocytes are important in motor neuron survival.²⁶ Indeed, the ability for astrocytes to induce motor neuron injury through non-cell-autonomous mechanisms has been shown for C9orf72 and G93A mutant SOD1.^27,28 Although abnormal astrocytic glutamate transporter EAAT2 RNA processing giving rise to impaired astrocytic glutamate uptake is thought, in part, to underlie toxic CSF glutamate levels in ALS,²⁹ it remains to be determined whether abnormal glutamate metabolism is a primary or initiating causative factor.³⁰ The astrocytic abnormality suggests that the what comes first question is less relevant. Indeed, astroglial characteristics differ in various brain regions, perhaps allowing regional susceptibility. The tempo of the disease, as distinct from its onset, varies from case to case. Although ALS progresses inexorably, it does not progress in a linear fashion from its onset implying variable local susceptibility and resistance to the disease process.

ALS as a Neurodegenerative Disorder Driven by Common Thresholds of Molecular Vulnerability

Many avenues of evidence support the concept that ALS is a neurodegenerative disorder or syndrome for which the classical clinical expression is defined by the limited repertoire by which motor neuron dysfunction can be expressed.³¹ That the phenotypic expression of ALS is restricted in this manner, however, is refuted by the spectrum of ALS-associated frontotemporal dysfunctions.⁷ These include an FTD typical of the Neary, Hodges, or Rascovsky criteria (ALS-FTD), impairments of either cognitive or behavioral function (ALSci and ALSbi, respectively) or both (ALS-cbi), or other dementias (i.e., vascular cognitive impairment and Alzheimer disease).^e9-e11 Although around 50% of patients with ALS will be spared nonmotor neocortical involvement, with upward of 15% either initially presenting with FTD or developing FTD in the course of their illness, the remainder may develop manifestations of nonexecutive cognitive dysfunction, executive dysfunction, behavioral impairment, or, less commonly, Alzheimer or mixed dementia.⁷ It is less clear whether syndromes of frontotemporal dysfunction in ALS will inexorably progress to ALS-FTD, or indeed, whether the co-occurrence of cognitive and motor involvement in ALS represent different trajectories of cortical nonmotor and motor network degenerations.³²

If the co-occurrence of progressive motor neuron degeneration typical of ALS and nonmotor neocortical degeneration are not simply chance occurrences, the question becomes how such a co-occurrence is initiated. Is it a matter of shared thresholds of neuronal degeneration driven by common molecular mechanisms, a reflection of shared susceptibilities of neural networks, or some more broadly based pathophysiology in which neuronal dysfunction and ultimately loss is simply the byproduct of a set of neuronal entrapment in a disease process? Or some combination thereof?

A clue to this puzzle of the origin of ALS lies in the molecular underpinnings of neuronal dysfunction in the range of neurodegenerative disease states. One of the most fundamental advances in our understanding of ALS, and indeed neurodegeneration as a whole, has emerged from our increasing understanding of the regulation of gene expression through RNA-mediated biogenesis and transcriptional regulation.³³ For those of us who trained some time ago, the simple linear concept of DNA transcription giving rise to messenger RNA (mRNA), which is in turn faithfully translated into an array of proteins that then undergo extensive posttranslational modifications, is an outmoded concept. Although it is true that alterations in gene transcription will be reflected in alterations in steady-state mRNA levels (for instance as measured by RNAseq), RNA itself is a highly dynamic intermediary whose genesis, compartmentalization, translational regulation, and stability are regulated through an array of mechanisms that respond rapidly to alterations in cellular homeostasis. The interconnectedness of alterations in steady-state mRNA levels, and thus to some degree of gene expression, has led to the concept that neurodegenerative disorders can be understood in terms of alterations in molecular frameworks of gene expression that extend beyond transcription mechanisms alone. For example, in a computational study of published data sets of alterations in gene expression across ALS-FTD, Alzheimer disease, and Lewy body dementia, 88 genes were observed to be upregulated at the intersection of each of these disorders, whereas a further 45 genes were found to be downregulated.³⁴ These could be further grouped at the level of functional pathways into alterations in innate immunity, the cytoskeleton, transcriptional regulation and RNA processing, and the mitochondrial electron transport chain. Many of these processes encompass both neuronal and non-neuronal origins, and while the listing is far from exhaustive, it lends support to the concept that there is a commonality of alterations in gene expression across neurodegenerative disease states and that no single molecular derangement can account for all aspects.

Indeed, there is increasing evidence to suggest that the intersection of distinct pathophysiologic processes may underlie many of the neurodegenerative disorders and act in a synergistic manner in which the net effect is significantly greater than predicted from each insult alone. This is of particular importance when considering the co-occurrence of motor and nonmotor manifestations in ALS. Disease survival is significantly shorter in individuals in whom FTD and ALS develop than in ALS in the absence of FTD.³⁵ Recent studies using transgenic rats expressing ALS-associated mutant TDP-43^M337V, in which a motor neuronal degeneration develops consistent with the key neuropathologic features of ALS, demonstrated synergism of pathology induction when also inoculated with a rAAV9 adenoviral construct expressing an ALS-associated pathologic tau construct.^36,37,e12-15 Of particular interest, although tau pathology was not observed in spinal motor neurons, the expected TDP-43 pathology of spinal motor neuron degeneration in the transgenic rat model was also increased. Such experimental evidence provides support for the concept of synergism of pathologies that do not require coexpression within the same cell populations but can also act at a distance, in essence the non-cell-autonomous conceptualization. Given the observation that the restricted expression of TDP-43^M337V in astrocytes can also cause non-cell-autonomous motor neuron death in transgenic rats,³⁸ it is quite plausible that the robust glial activation observed in this model plays a role in the synergistic effects.

With the realization that the majority of cases of ALS are associated with alterations in RNA metabolism has come the understanding that such alterations are also protean, even at the level of a single cell.³⁹ We have focused on the role of alterations in RNA-binding protein function in mediating RNA instability in ALS as a proposed molecular class effect and have shown coaggregation of RNA-binding proteins into neuronal cytoplasmic inclusions in single spinal motor neurons in ALS.⁴⁰ This lends support to the premise that the designations of subtypes of ALS as being reflective of specific opathies (i.e., TDP-43-opathies and FUS-opathies) might best be restricted to individual genetic variants of ALS in which specific mutations within such genes are clearly causative, whereas the majority of cases of ALS are reflective of more widespread molecular class effects. In this light, the pathologic sequestration of RNA-binding proteins fundamentally impairs RNA-mediated oversight of gene expression. For example, disruption of the physiologic nucleocytoplasmic distribution of TDP-43 as is characteristic of ALS is associated with the impairment of a feedback loop that regulates the nuclear processing of key microRNAs that in turn regulate TDP-43 expression, fundamentally leading to alterations in the stoichiometry of neuronal intermediate filament protein metabolism.⁴¹

If the molecular underpinnings of neuronal degeneration in ALS can be so highly variable and potentially driven by synergism of toxicities, is this sufficient to account for the distinctive syndromic nature of the disorder? Insights into this process have been provided through the neural network hypothesis of neurodegeneration, which predicts that syndromic atrophy patterns should be based on healthy functional networks.^42,43 In a comparative analysis of patients affected by Alzheimer disease, behavioral variant FTD, semantic dementia, progressive nonfluent aphasia, cortical basal syndrome, and a cohort of age-matched control patients, Seeley and colleagues⁴² described 5 distinct, human integrated neural networks. These networks created a framework onto which each of the neurodegenerative disorders could be mapped, in essence, by mapping the human brain connectome, composed of a series of interconnected neural nodes that determine physiologic functions. Such connectomes can be defined in several ways, including the presence of synchronous baseline activity, a unified cortical trophic fate, and selective vulnerability to neurodegenerative illness. They can also be defined on the basis of both neural and synaptic functions or can be macroscopic or structural in nature in which case they define anatomic connections linking a set of neural elements.

In applying the neural network degeneration hypothesis, disorders such as Alzheimer disease can be seen to predominantly involve the default mode network; behavioral variant FTD involves the salience network; semantic dementia the temporal pole, subgeneral, cingulate, ventral striatum, and amygdala network; progressive nonfluent aphasia is driven by the frontal operculum, supplementary motor cortex, and inferior parietal involvement; and cortical basal syndrome can be seen to be related to the dorsal sensorimotor-associated network. In this manner, what appear to be diffuse unrelated degenerative processes can be related more clearly to network degeneration and thus to widespread neurodegenerative networks.

To tie together these concepts of neuron-specific patterns of molecular dysfunction and disruptions of independent neural networks subserving discrete functions, Warren and colleagues have suggested the existence of molecular nexopathies as “specific, coherent conjunctions of pathogenic protein and intrinsic network characteristics that define network signatures of neurodegenerative pathologies.”^44,45 This hypothesis implies that the degenerative process preferentially targets network elements that are selectively vulnerable to the pathologic molecular species and infers that compensatory mechanisms are insufficient. Indeed, the presence of such molecular networks could then be theorized as underlying spread or the susceptibility to spread. An example of such might be the evidence that motor neurons share a specific pool of miRNAs (termed motomiRs), which, when altered in their expression, give rise to specific variants of motor neuron dysfunction.^e16 Furthermore, such a proposal predicts that with progression of the molecular insult culminating in neuronal dysfunction and cell death, the phenotypic expression will be a function of the interconnectedness of the affected neural networks. Expressed differently, the clinical manifestations of ALS are driven not by the specific metabolic insult per se, but rather by the susceptibility of specific neuronal populations to the insult and the interconnectedness of the pool of dysfunctional neurons. Progression can be viewed as extension of pathologic dysfunction within an individual connectome, while the acquisition of new neurodegenerative features reflects the crossing of a threshold of individual neuronal populations.

Finding Common Ground

In our model, ALS is a neurodegenerative syndrome based on the progressive failure of functionally related neural networks. Clinical (phenotypic) heterogeneity is driven by the rate of progression and relative involvement of the neural networks, likely driven by shared, propagating molecular dysfunctions. The genesis of each syndromic variant may vary to some degree but what remains at the core, however, is the progressive degeneration of motor neurons.

If one accepts this, the issue of the anatomic site of disease origin becomes less relevant. It is instead the nature of the underlying molecular or proteomic insult that drives the degeneration of discrete neural networks, including those subserving motor function. Even here, our understanding of the underlying genetics of ALS is telling us quite clearly that there are shared thresholds of neuronal vulnerability by which an array of molecular insults can manifest similar network dysfunctions. Tempo and severity may vary but not the underlying susceptibility. Understanding how thresholds of molecular vulnerability are determined, including issues of shared phylogeny of seemingly disparate neuronal populations, or of how rates of progression or severity may be driven by factors exogenous to the underlying molecular pathophysiology, remains a critical frontier for ALS researchers.

This conceptualization does not, however, imply that such postulated processes occur in isolation. Equally critical is the need to understand by what mechanisms environmental factors affect these processes.⁴⁶ Much has been written about the various metabolic derangements that are associated with ALS, each of which can affect very specific aspects of neuronal function. We have focused on a narrow range, including alterations in RNA metabolism and the potential role for glutamatergic damage, to support our argument for a more broadly based network dysfunction. The range of pathogenic mechanisms underlying the degeneration of ALS includes defects in nucleocytoplasmic transport, impaired proteostasis, impaired DNA repair, and mitochondrial dysfunction, in addition to axonal and vesicular transport defects.^47,48 Both neuroinflammation and non-cell-autonomous effects are important in such a discussion.^49,50,e17 Understanding the interplay between each of these processes, and how they may contribute as determinants of the rate of progression, is a fundamental ongoing research question.

In conclusion, our remit at the invitation to this debate was to resolve the issue of where the anatomic site of origin of ALS lies. We have described how the classical anatomic view of origins can no longer apply as our understanding of the molecular basis of ALS evolves. Rather, we have begun to understand that the nervous system does not necessarily function in discrete classical anatomic pathways, but rather through pathways of interconnectedness (neural networks), that when perturbed lead to specific dysfunctional syndromes. The challenge now is to understand how this knowledge can be harnessed by a new generation of researchers to develop new therapeutic interventions.

Glossary

ALS

amyotrophic lateral sclerosis

FTD

frontotemporal dementia

LMN

lower motor neuron

TDP-43

TAR DNA-binding protein 43

UMN

upper motor neuron

Appendix. Authors

Study Funding

The authors report no targeted funding.

Disclosure

The authors report no disclosures relevant to the manuscript. Go to Neurology.org/N for full disclosures.