Is essential tremor a degenerative disorder or an electric disorder? Degenerative disorder

Abstract

Essential tremor (ET) is a highly prevalent neurologic disease and is the most common of the many tremor disorders. ET is a progressive condition with marked clinical heterogeneity, associated with a spectrum of both motor and non-motor features. However, its disease mechanisms remain poorly understood. Much debate has centered on whether ET should be considered a degenerative disorder, with underlying pathological changes in brain causing progressive disease manifestations, or an electric disorder, with overactivity of intrinsically oscillatory motor networks that occur without underlying structural brain abnormalities. Converging data from clinical, neuroimaging and pathological studies in ET now provide considerable evidence for the neurodegenerative hypothesis. A major turning point in this debate is that rigorous tissue-based studies have recently identified a series of structural changes in the ET cerebellum. Most of these pathological changes are centered on the Purkinje cell and connected neuronal populations, which can result in partial loss of Purkinje cells and circuitry reorganizations that would disturb cerebellar function. There is significant overlap in clinical and pathological features of ET with other disorders of cerebellar degeneration, and an increased risk of developing other degenerative diseases in ET. The combined implication of these studies is that ET could be degenerative. The evidence in support of the degenerative hypothesis is presented.

1. Introduction

Essential tremor (ET) is a highly prevalent and progressive tremor disorder, with its most recognizable feature being a 4–12Hz kinetic tremor (i.e., tremor occurring during volitional movement) of the arms. The clinical phenotype of ET has expanded from a monosymptomatic entity focusing on its kinetic arm tremor to one with several other types of tremors, other motor features, and a range of non-motor features (Louis, 2021). Thus, there is considerable clinical heterogeneity in ET, leading to the concept that it may well be a family of diseases. While it is currently accepted that motor circuits involving the cerebello-thalamo-cortical loop and brainstem circuits including olivo-cerebellar networks are involved in ET, the underlying pathobiology for how tremor is generated in the brain in ET remains controversial.

There are currently two main postulates for tremor generation in ET: (1) it is a degenerative disorder, (2) it is an electric disorder. The degenerative hypothesis was previously discarded, as early studies on ET claimed there was no definable pathology (Rajput, Robinson, & Rajput, 2004). Indeed, during the 100-year period from 1903 to 2003, there were only 15 published postmortems, mainly comprising isolated case reports with no conclusive findings (Louis & Vonsattel, 2008). These studies lacked rigor, as none used quantitative or immunohistochemical approaches nor compared ET to control brains, and clinical definitions lacked precision as many of the cases had dystonia or chorea rather than ET. This thinking persisted in the literature, with many textbook chapters and review articles simply omitting any discussion of ET pathology well into the 2000s.

In the early 1970s an alternative model took favor that ET is an electric/electrophysiologic disorder, much like epilepsy, due to overactivity of synchronous firing from pacemaking neurons in the inferior olivary nucleus (ION). This electric hypothesis of ET was fueled by studies on an animal model of action tremor induced by injection of harmaline into rats (Louis & Lenka, 2017). Harmaline primarily targets the ION where it enhances gap-junction mediated synchronous firing of olivary neurons. Overactivity in the ION could then induce rhythmical burst firing in cerebellar Purkinje cells (PCs) and cerebellar nuclei through excitatory climbing fiber (CF) input, leading to tremor by propagation through cerebello-thalamo-cortical output channels. There remain many problematic issues with the harmaline model as an accurate representation of human ET, as (1) it is an acute model of action tremor, lacking the chronicity of ET, (2) it does not respond to all the same pharmacological agents as ET, and (3) harmaline injection in rats causes massive destruction of PCs in parasagittal bands, a morphologic feature not seen in human autopsies of ET brains (Louis & Lenka, 2017). Furthermore, there are neurons with pacemaking properties in several other brain regions, including PCs, cerebellar nuclei, thalamus and locus coeruleus, among others. More recent studies in ET have not shown any consistent neuroimaging (Wills, Jenkins, Thompson, Findley, & Brooks, 1994) or pathological abnormalities in the ION (Louis, Babij, Cortés, Vonsattel, & Faust, 2013), raising issues with this single ION oscillator model. Pathology case studies of ET patients that developed olivary hypertrophy (Elkouzi, Kattah, & Elble, 2016) or olivary degeneration (Louis et al., 2018) demonstrated no change in their tremor characteristics, arguing against the hypothesis that the ION alone plays a crucial role in genesis of tremor in ET. Considering these findings, the focus of the electrical model of ET shifted to encompass network properties. It is proposed there is an oscillatory network comprising the ION, cerebellum, thalamus and other brainstem regions, and where all network components may act as oscillators that dynamically entrain each other (Helmich, Toni, Deuschl, & Bloem, 2013; Raethjen & Deuschl, 2012). However, this model is largely a theoretic construct with little empirical support.

Since 2003, increasing support for the neurodegenerative hypothesis of ET has grown through consideration of the clinical features of ET and its similarities with other degenerative disorders, advanced neuroimaging techniques demonstrating structural changes consistent with neuronal loss, and an increasing number of pathological studies demonstrating degenerative changes centered in the cerebellum in ET. Indeed, the idea that ET could be neurodegenerative was proposed long ago. In the landmark paper by Critchley and Greenfield, the authors wrote: “Although anatomical proof is as yet lacking, there are at least a number of clinical points to make question whether “essential tremor” may not, at times any rate, represent an incomplete or a premature variant of one of the cerebellar atrophies” (Critchley & Greenfield, 1948). In this chapter, we review this growing body of evidence from these three perspectives, which support the conclusion that ET is a degenerative disorder. We present evidence that these findings in ET fulfill central tenets for the definition of a degenerative disease: a chronic and progressive disorder, often with an insidious onset, with an intrinsic selective vulnerability of specific cell populations leading to cell loss and loss of structure and function in the nervous system.

2. Clinical features of ET as a degenerative disorder

ET is increasingly recognized as a clinically heterogeneous condition, with differences in age of onset, anatomic distribution of tremor, rate of disease progression and other degenerative disease associations. Many of these clinical features suggest that ET is a degenerative disorder (Table 1).

Table 1.

Clinical features suggesting that ET is neurodegenerative.

Evidence	Comment
Insidious onset	Tremor may initially be dismissed as just “nervousness” due to slow onset
Prevalence increases with age, and exponentially with advanced age	Prevalence across all ages ~1.33%, increases to 5.79% in individuals ≥65 years, over 20% in oldest old (≥95 years)
Older age of onset associated with more rapid disease progression	ET associated with aging processes
Somatotopic spread of tremor to cranium (neck, voice, jaw) with disease progression	Suggests progressive destruction of different brain regions, as seen in AD and PD
Intention tremor, gait ataxia, oculomotor abnormalities	Clinical resemblance to other cerebellar atrophies, suggesting shared disease mechanisms
Isolated rest tremor in arm only with longstanding disease	Pathological involvement of extra-cerebellar circuits or abnormal interactions with basal ganglia
Associated non-motor features Cognitive deficits Psychiatric features (depression, anxiety, social phobia) Mild olfactory dysfunction, hearing impairment, sleep dysregulation	Suggests more widespread involvement of different functional brain regions Increased prevalence or incidence of MCI and dementia in ET Many neurodegenerative diseases are neuropsychiatric disorders. Depression in ET may be primary Increased in some ET patients. Not specific to neurodegeneration
Increased risk of AD	Increased conversion from MCI to AD-type dementia. Late onset-ET increases risk of developing AD by ~2-fold
Increased risk of PD	Four- to fivefold increased risk of developing PD. Suggests overlap in disease mechanisms
Hereditability factors lead to earlier onset disease	Similar to that seen in other neurodegenerative diseases

AD, Alzheimer’s disease; ET, essential tremor; MCI, mild cognitive impairment; PD, Parkinson’s disease.

First, ET has an insidious onset (Critchley, 1949; Larsson & Sjogren, 1960). Age of symptom onset can be difficult to precisely recall, and individuals with ET may initially dismiss the tremor as just “nervousness.” In contrast, electric disorders such as seizures, often have an abrupt onset.

As in many degenerative diseases, the prevalence of ET increases markedly with age and, furthermore, increases in an exponential manner with advanced age. While prevalence estimates for ET differ across studies due to a variety of differences in methodology and population demographics, reproducible patterns of increasing prevalence with age are still obtained. In the largest meta-analysis to date, based on 42 studies across 6 continents and 23 countries reported between 1960 and 2020, the pooled prevalence of ET (all ages) was 1.33% (95% CI=0.88%–2.02%) (Louis & McCreary, 2021). In individuals age ≥65 years, prevalence increases to 5.79% (95% CI=4.14%–8.05%) and is even higher in oldest age groups (80s, 90s and older) ranging from 1.2% to 42.9% (mean=11.4%, median=9.3%). ET prevalence increases by 74% for every decade increase in age (P <0.00001) (Louis & McCreary, 2021). In another meta-analysis based on 29 studies reported between 2000 and 2019, the estimated global prevalence of ET was 0.32% (95% CI=0.12%–0.91%) (Song et al., 2021). The prevalence of ET increased dramatically with advancing age, where the prevalence estimate in people aged under 20 years was 0.04% (95% confidence interval=0.00%–0.29%) and that in elderly aged 80 years and above was 2.87% (95% confidence interval=1.07%–7.49%). Thus, aging is a clear risk factor for ET, and similar exponential increases in prevalence are found in patients with Parkinson’s disease (PD), Alzheimer’s disease (AD), and other degenerative diseases of the nervous system (Louis, 2019).

ET is a chronic disorder with a gradual yet progressive clinical course that proceeds over many years (Louis, Ford, & Barnes, 2000; Putzke, Whaley, Baba, Wszolek, & Uitti, 2006). A few longitudinal studies have attempted to quantify the clinical rate of tremor progression from baseline in ET patients, estimated as 12% per year (Putzke et al., 2006) or a 3.1%–5.3% average annual increase in tremor severity (Louis, Agnew, Gillman, Gerbin, & Viner, 2011). In a prospective study of ET patients, disease duration and age were independently associated with tremor severity in ET (Louis, Jurewicz, & Watner, 2003). These findings raise the possibilities that reported increases in tremor severity in ET may be related not only to inherent worsening of disease pathophysiology with increase duration but also with age and age-related processes such as neuronal attrition. In a large clinical cohort of 348 ET cases, older age of onset was associated with more rapid disease progression, indicating an intimate relationship between disease progression and aging (Louis, Faust, Vonsattel, Honig, Henchcliffe, et al., 2009a). With disease progression in ET, there is a somatotopic spread to other body parts, including head (neck), voice and jaws (Louis, 2021), which may reflect spreading of disease to different brain regions as seen in other degenerative disorders. Last, once the disease begins there are no reported cases of spontaneous remission of ET, which might occur if the pathophysiology of ET was based on a central oscillating pacemaker.

A wide spectrum of cerebellar symptoms may be observed with disease progression, supporting the notion that the cerebellum might be centrally involved in ET. These include intentional tremor, gait ataxia, dysarthria, oculomotor abnormalities, and deficits in hand-eye coordination (Benito-León & Labiano-Fontcuberta, 2016; Louis & Faust, 2020a, 2020b). For instance, as many as 50% of patients with ET have intention tremor, which may involve arms as well as head and legs, and this is usually more evident in advanced ET cases. Similarly, gait disturbances such as ataxia typically occur with longer disease duration in ET, and those patients with head tremors have the greatest gait impairment. These ET-related gait abnormalities may reflect onset of disease in the medial cerebellum that controls balance and locomotion, consistent with prominent atrophy of the cerebellar vermis in a magnetic resonance imaging (MRI) study (Quattrone et al., 2008) and greater cerebellar pathology in a postmortem study of ET vermis (Louis et al., 2011) among patients with cranial tremors. A variety of oculomotor abnormalities occur in ET, and they are particularly evident among those ET patients also exhibiting intention tremor (Helmchen et al., 2003). In sum, these other cerebellar abnormalities in ET resemble that seen in cerebellar atrophies, suggesting the possibility of shared disease mechanisms (Louis & Faust, 2020a, 2020b).

Rest tremor involving only the arms may occur in as many as 20%–30% of ET patients, associated with a kinetic tremor that is more severe, more disseminated and of longer duration (Cohen, Pullman, Jurewicz, Watner, & Louis, 2003). In a postmortem analysis of 9 ET cases with rest tremor, Lewy body pathology was not present in the basal ganglia (Louis et al., 2011). Rest tremor in PD, and more broadly in other disorders (e.g., ET, dystonia), is postulated to be induced by abnormal basal ganglia activity, generated by the thalamus, and modulated or reinforced by the cerebellum (Duval, Daneault, Hutchison, & Sadikot, 2016). Thus, both ET and PD have circuits involving cerebellar-thalamo-cortical networks associated with rest and kinetic tremor generation.

As with many neurodegenerative movement disorders (e.g., PD, Huntington’s disease), the clinical features in ET extend beyond the motor system. These non-motor features may affect several distinct domains, including cognitive, psychiatric (depression, anxiety, social phobia), sensory (olfactory dysfunction, hearing impairment) and sleep disturbances (Louis, 2016, 2021). Many studies have demonstrated cognitive deficits in ET, ranging from mild cognitive impairment to frank dementia, and a greater rate of cognitive decline occurs in older ET patients, all more than seen in age-matched controls (Benito-León, 2014; Radler et al., 2020). Some of this impairment might reflect degenerative pathology in the cerebral cortex (Farrell et al., 2019; Pan et al., 2014) although there are some biochemical distinctions in neocortical tau isoforms in ET versus other tauopathies with an increase in tau without N-terminal inserts (Kim et al., 2021). A cerebellar contribution is also possible, as a cerebellar cognitive affective syndrome is well-known to occur and includes hallmark deficits in executive function (Janicki, Cosentino, & Louis, 2013). Mild olfactory dysfunction occurs in a subset of ET patients, and does not correlate with disease duration or severity, indicating it may occur early in the disease process. Furthermore, emerging evidence suggests that some of these features (e.g., cognition, depression) could be primary and may even pre-date motor features in ET (Benito-León, 2014; Louis, 2016).

ET itself has been associated with other degenerative disorders. There is a longstanding relationship between ET and PD from a clinical, epidemiologic, and genetic perspective (Clark & Louis, 2018; Thenganatt & Jankovic, 2016), including a four- to fivefold increased risk of developing PD in ET patients (Benito-León, 2014). In older onset ET, the risk of developing AD is increased nearly twofold (Louis, 2019). Family history is also a strong risk factor for ET. The disease is familial in as many as 30%–70% of ET patients and this genetic predisposition lowers age of disease onset (Clark & Louis, 2018; Louis, 2019), as seen in other degenerative disorders with a hereditary basis. These various associations suggest that ET may share pathogenic mechanisms with these disorders.

While many of the clinical features discussed in this section are not specific to degenerative diseases when occurring in isolation, the constellation of findings, all present in the same disease, is more compelling (Table 1).

3. Neuroimaging findings suggest cerebellar degeneration in ET

Neuroimaging studies provide an in vivo modality to localize disease related changes in the brain. Several recent reviews have compiled the ever-increasing number of neuroimaging studies in ET that use a broad array of methods, including magnetic resonance (MR) volumetry, MR spectroscopy, diffusion tensor imaging (DTI), functional MR imaging (fMRI), other MR imaging, and positron emission tomography (PET) (Cerasa & Quattrone, 2016; Louis, Huang, Dyke, Long, & Dydak, 2014; Pietracupa, Bologna, Tommasin, Berardelli, & Pantano, 2021). Several fMRI studies and a considerable number of PET studies have identified neuronal activity changes in the cerebellar-thalamic network in ET, clearly demonstrating involvement of these brain regions for tremor generation. As fMRI and PET do not necessarily assess neuronal loss or degeneration, they will not be discussed further here. Although there is still significant variability in results across these various neuroimaging studies, overall, the combined structural data provide the greatest support for a cerebellar neuronal degeneration in ET (Table 2).

Table 2.

Neuroimaging studies in ET demonstrating structural cerebellar changes.

Method	Authors (year)	# Subjects	Main findings in ET patients versus controls
MRI volumetry	Quattrone et al. (2008)	50 ET, 32 controls	Marked atrophy of cerebellar vermis in ET patients with head tremor (n = 20)
	Cerasa et al. (2009)	46 ET, 28 controls	Reduced cerebellar volume in ET patients with head tremor (n = 19)
	Benito-León et al. (2009)	19 ET, 20 controls	Gray and white matter losses in bilateral cerebellum and other regions (parietal lobes; right frontal, insular, limbic lobes; left medulla)
	Bagepally et al. (2012)	20 ET, 17 controls	Gray matter losses in bilateral cerebellum, vermis, bilateral frontal and occipital lobes
	Bhalsing et al. (2014)	25 ET, 25 controls	Gray matter loss in anterior and posterior cerebellar lobes, right medial frontal gyrus in ET patients with cognitive impairment
	Choi et al. (2015)	45 ET, 45 PD, 45 controls	Reduced cerebellar volume in ET patients with head tremor (n = 19); no change in PD versus controls
	Shin et al. (2016)	39 ET, 36 controls	Reduced cerebellar volume in several vermal areas, most evident in ET patients with (n = 20) versus without (n = 19) cerebellar signs
	Dyke, Cameron, Hernandez, Dydak, and Louis (2017)	47 ET, 36 controls	Decreased cerebellar volume in the vermis and several cerebellar lobules in ET patients with cranial tremors
DTI	Shin et al. (2016)	10 ET, 8 controls	Reduced FA in bilateral cerebellum as well as pons, midbrain and cerebral cortex (orbitofrontal, lateral frontal, parietal, temporal)
	Nicoletti et al. (2010)	25 ET, 15 PD, 15 controls	Reduced FA in dentate nucleus and superior cerebellar peduncle correlating with ET disease duration; not seen in PD patients
	Klein et al. (2011)	14 ET, 20 controls	Reduced FA in right inferior cerebellar peduncle; Increased MD of inferior cerebellar peduncle bilaterally and left parietal white matter
	Jia, Jia-Lin, Qin, Qing, and Yan (2011)	15 ET, 15 controls	Increased diffusion values in red nucleus; only limited regions examined that did not include cerebellum
	Saini et al. (2012)	20 ET, 17 controls	Increased diffusivity in cerebellar hemispheric white matter, thalamus, brainstem and bilateral cerebral hemispheres
	Novellino et al. (2016)	67 ET, 39 controls	Increased diffusivity in cerebellum
	Pietracupa et al. (2019)	19 ET, 15 controls	White abnormalities in corticospinal tract, cerebellar peduncles, corpus callosum, several associative white matter bundles
MRS	Louis et al. (2002)	16 ET, 11 controls	Reduced mean cerebellar cortical NAA/Cr ratio in ET patients
	Pagan, Butman, Dambrosia, and Hallett (2003)	10 ET, 10 controls	Reduced NAA/Cr ratio in bilateral cerebellar hemispheres
Other	Novellino et al. (2013)	24 ET, 25 controls	Whole brain voxel-based analysis consistent with increased iron deposits in globus pallidus, substantia nigra, right dentate nucleus
PET	Boecker et al. (2010)	8 ET, 11 controls	Increased cerebellar dentate, ventrolateral thalamus and premotor cerebral cortex uptake of 11C-flumazenil
	Gironell et al. (2012)	10 ET, no controls	Significant correlation of 11C-flumazenil uptake and tremor rating scales

CR, creatinine; DTI, diffusion tensor imaging; ET, essential tremor; FA, fractional anisotropy; MRI, magnetic resonance imaging; MRS, magnetic resonance spectroscopy; NAA, N-actetylaspartic acid; PD, Parkinson’s disease; PET, positron emission tomography.

In a comprehensive review published in 2014 (Louis et al., 2014), four of six studies using voxel-based morphometry (i.e., MR volumetry) reported a reduction in cerebellar volume in ET (Bagepally et al., 2012; Benito-León et al., 2009;Cerasa et al., 2009; Quattrone et al., 2008), and this was particularly prominent in patients with head tremor in two studies (Cerasa et al., 2009; Quattrone et al., 2008). In another recent review, cerebellar gray matter volume reductions were the most common structural finding in ET, seen in 7 of 12 studies that examined the cerebellum (Pietracupa et al., 2021). Two additional MR volumetry studies confirmed greater cerebellar volume losses in ET patients with head tremors, either more globally present and distinguishing patients with ET from those with PD (Choi et al., 2015) or present within vermis and particular cerebellar lobules (Dyke et al., 2017). ET patients with clinical cerebellar signs (e.g., ataxia) had atrophy in several contiguous regions of the cerebellar vermis versus those lacking cerebellar signs (Shin et al., 2016). Last, ET patients with cognitive impairment had gray matter loss in anterior and posterior lobes of the cerebellum, as well as medial frontal gyrus, in comparison to controls (Bhalsing et al., 2014); this is of interest as posterior cerebellar lobes are implicated in high-level cognitive functions.

In addition to gray matter volume reductions, most studies using diffusion imaging have demonstrated microstructural changes in cerebellar white matter between ET cases and controls. Five of seven studies reported as of 2014 demonstrated differences between ET cases and controls (Table 2), and the two studies with null results had important methodologic limitations (Louis, Lee, Babij, et al., 2014). A subsequent study also showed diffusion changes within the cerebellum (Novellino et al., 2016). More widespread white matter changes have been detected in cerebellar peduncles as well as corticospinal tracts, corpus callosum and several associative white matter bundles (Pietracupa et al., 2019). In sum, these diffusion imaging studies demonstrate that some orientation dependent aspect of tissue microstructure is abnormal in ET. The bulk of studies show evidence of axonal changes within the cerebellum itself along with abnormalities in other brain regions, many of which are part of the cerebellum’s tremor outflow pathways.

Some diffusion imaging studies have highlighted a relationship between areas of cerebellar degeneration and specific disease features in ET. Increased diffusion changes in the dentate nucleus region completely distinguished familial ET patients from PD patients and controls, and patients with longer ET disease duration had more severe degeneration of white matter within/around the dentate (Nicoletti et al., 2010). In patients with late-onset compared to early-onset ET, more degenerative changes were seen in the superior cerebellar peduncle, which may relate to faster rate of tremor progression in late-onset ET (Louis, Faust, Vonsattel, Honig, Henchcliffe, et al., 2009a; Saini et al., 2012).

MR spectroscopy will detect in vivo concentrations of certain cellular metabolites such as N-acetyl aspartate (NAA), creatine (Cr) or choline (Cho). NAA is a marker of neuronal integrity, thus reduced NAA is suggestive of processes that cause neuronal damage, neuronal dysfunction, and neuronal loss. There are two currently available MRS studies that have examined the cerebellum and provide evidence of cerebellar involvement in ET based on reduced NAA (Louis et al., 2002; Pagan et al., 2003). The reduced NAA/cr ratio inversely correlated with tremor severity in ET (Louis et al., 2002).

Brain iron accumulation is a common phenomenon in neurodegenerative processes, as previously demonstrated in AD, PD and other degenerative disorders. In a study of 24 ET compared with 25 controls, a whole brain voxel-based analysis demonstrated significant differences consistent with increased iron deposition in bilateral globus pallidus, substantia nigra and right dentate nucleus (Novellino et al., 2013). This study provides indirect evidence supportive of degeneration. Iron accumulation has yet to be studied in the postmortem ET brain.

There is considerable evidence that defects in the GABAergic system underly disease pathogenesis in the ET brain (Gironell, 2014). For instance, the majority of drugs used to treat ET act via GABAergic pathways. One ET case-control PET study examined ¹¹C-flumazenil binding to the GABA_A complex, demonstrating increased ligand uptake in cerebellar dentate, ventrolateral thalamus and premotor cerebral cortex in ET patients, raising the possibility of a receptor upregulation due to either GABA deficiency or abnormal/dysfunctional GABA_A receptors in ET (Boecker et al., 2010). Further evidence for GABAergic dysfunction in ET was provided by tissue-based studies of ET cerebellum demonstrating a localized GABAergic defect due to cerebellar PC loss and reduced GABA receptors in cerebellar dentate nucleus (Louis & Faust, 2020c; Paris-Robidas, Brochu, Sintes, et al., 2012), supporting the presence of cerebellar degeneration in ET. A subsequent study in ET patients only demonstrated correlation between ¹¹C-flumazenil uptake and tremor rating scales, again suggesting an abnormality at the level of GABA_A receptors in ET (Gironell et al., 2012).

In summary, neuroimaging studies in ET have provided convincing evidence that the cerebellum plays a key role in disease pathophysiology, with many of these studies supporting a degenerative process. More widespread changes are also seen in other brain regions, many of which are part of the cerebellum’s tremor outflow pathway. Some of the changes outside the cerebellum may correlate with cognitive impairments in ET, as detailed elsewhere (Pietracupa et al., 2021). Distinct challenges remain for these neuroimaging methods, particularly with respect to issues of both spatial and temporal resolution and to fully capture the clinical heterogeneity of ET.

4. Pathological findings in ET demonstrate cerebellar degeneration

A fundamental question about the degenerative nature of a neurological disorder is whether there are histopathological changes in the brain that provide evidence of neuronal loss. In this section we discuss the expanding literature demonstrating that there is indeed selective vulnerability of specific cell populations (e.g., cerebellar PCs) leading to neuronal loss and loss of structure and function in the nervous system.

Considerable accumulating evidence now suggests that ET is a disease of cerebellar degeneration. A chronology of major pathological findings identified in the ET cerebellum since the turn of the 21st century is listed in Table 3, with many studies that now include modern methodologies as well as comparisons with both neurologically normal controls and various degenerative disease controls. These studies summarize growing evidence that there are indeed structural abnormalities centered on the cerebellar PC along with altered PC synaptic contacts by neighboring basket neurons and olivary CFs in cerebellar cortex. Indeed, it has been proposed that ET may be a “Purkinjopathy” (Grimaldi & Manto, 2013), where dysfunction of PCs results in dis-inhibition of the cerebellar nuclei, which then impacts functioning of cerebello-thalamo-cortical pathways and cerebellar-brainstem loops. In this section, a “best guess” anatomical-physiologic model is presented, based on findings identified cross-sectionally at the time of death. This model places the pathological findings in ET cerebellum within a putative cascade of “early,” “middle” and “late” changes (Louis & Faust, 2020c), and highlights how these changes provide support for a degenerative disease process (Fig. 1).

Table 3.

Chronology of Neuropathology studies in ET during the 21st century.

Authors (year)	# Subjects	Main findings
Rajput et al. (2004)	20 ET, no controls	Mild PC loss in 2 of 20 ET cases. Qualitative methods only
Ross et al. (2004)	11 ET, 11 control	Some ET cases with more pathology in brainstem and cerebellum. Qualitative methods only
Louis et al. (2006)	10 ET, 12 control	Two disease patterns in ET: increased torpedoes and Bergmann glia (4/10), brainstem Lewy bodies often with atypical distribution (6/10)
Louis et al. (2007)	33 ET, 21 control	PC loss (25%, P < 0.01), ~7 × increase in torpedoes (P < 0.001). Qualitative dentate degeneration in 2 cases. Lewy bodies in 8 ET cases, which did not have significant cerebellar changes
Axelrad et al. (2008)	14 ET, 11 control	PC loss as high as 38%. PC loss correlated with age and number of torpedoes
Shill et al. (2008)	26 ET, 21 control	Increased cerebellar Bergmann gliosis and PC loss in 7/26 ET cases. Qualitative methods only
Louis, Faust, Vonsattel, Honig, Henchcliffe, et al. (2009a)	9 ET, no controls	Older age of ET onset associated with more degenerative pathology (PC loss, torpedoes) and more rapid disease progression
Louis, Faust, Vonsattel, Honig, Rajput, et al. (2009b)	40 ET, 25 control, 21 AD, 14 PD/DLBD	Increased torpedoes in ET versus controls, intermediate number of torpedoes in AD and PD
Louis, Yi, Erickson-Davis, Vonsattel, and Faust (2009)	4 ET, no controls	Torpedoes contain disorganized neurofilaments and organelles in unmyelinated PC axon segment
Erickson-Davis et al. (2010)	37 ET, 21 control	Structural changes in basket cell-PC contacts (“Hairy Baskets”); correlate with torpedo and PC loss pathology
Louis, Asabere, et al. (2011)	9 ET, no controls	PC loss and increased torpedoes, but no Lewy body pathology in basal ganglia, in ET with isolated rest tremor
Louis, Faust, et al. (2011)	24 ET, 10 controls	Increase torpedoes in ET vermis versus controls (2.7-fold), particularly in ET with cranial tremors
Kuo et al. (2011)	35 ET, 21 control, 11 PSP	Increased heterotopic PCs in ET cerebellar molecular layer, not seen in controls or PSP cases
Kuo, Faust, Vonsattel, Ma, and Louis (2011)	20 ET, 19 control	Similar parallel fiber counts and density in ET and controls
Rajput, Robinson, Rajput, and Rajput (2011)	7 ET, 2 control, 6 PD	No significant difference in PC counts among ET, PD-tremor and control groups. Underpowered study, type II statistical error
Yu et al. (2012)	20 ET, 19 control	Increased number of PC dendritic swellings in ET
Paris-Robidas et al. (2012)	10 ET, 16 control, 10 PD	Reduction in GABA receptors in the dentate nucleus in ET, not in PD
Louis, Babij, Cortés, et al. (2013)	14 ET, 15 control	No morphological differences in inferior olivary nucleus in ET
Kuo et al. (2013)	11 ET, 12 control	Increased Lingo-1 expression in ET cerebellum and elongated basket cell pinceau
Babij et al. (2013)	49 ET, 39 control	Abnormal PC axonal anatomy in ET with changes in shape and connectivity
Louis, Babij, Lee, Cortés, and Vonsattel (2013)	32 ET, 16 control	Reduced PC linear density in 100 micrometer thick sections, inversely correlated with torpedo counts
Louis, Lee, Cortés, Vonsattel, and Faust (2014)	25 ET, no controls	Clinical laterality or symmetry of tremor correlates with laterality of axonal pathological changes
Lin et al. (2014)	12 ET, 13 control	Abnormal climbing fiber-PC synaptic connections in ET
Louis, Lee, Babij, et al. (2014)	27 ET, 27 control	Reduced PC dendritic arborization and loss of dendritic spines in ET using Golgi silver staining
Lee et al. (2014)	16 ET, 13 control	Decreased EAAT2 expression in ET cerebellar cortex
Delay et al. (2014)	9 ET, 16 control, 10 PD	Increased LING0–1 expression in ET cerebellum, not seen in PD
Symanski et al. (2014)	56 ET, 62 control	ET is not associated with cerebellar PC loss. Potential problems with study design
Louis, Lin, Faust, Koeppen, and Kuo (2015)	37 ET, no controls	Climbing fiber synaptic changes correlate with clinical features in ET
Béliveau et al. (2015)	9 ET, 16 control, 10 PD	Insoluble Aβ42 accumulates in the cerebellar cortex of ET, but not in age-matched PD
Choe et al. (2016)	50 ET, 25 control	PC loss in ET, assessed by random sampling method and nearest neighbor analysis
Kuo et al. (2016)	39 ET, no controls	DBS treatment changes pattern of abnormal CF-PC connections in ET cases
Wang et al. (2016)	61 ET, 25 control	Increased EAAT2 expression in ET dentate nucleus, decreased expression ET cerebellar cortex
Kuo et al. (2017)	20 ET, 25 control, 30 disease controls	Degenerative movement disorders differ in CF-PC synaptic pathology, extent of PC loss and torpedo formation
Kuo et al. (2017)	60 ET, 30 control	Early onset and late onset ET cases share similar cerebellar postmortem features
Louis, Kuo, Tate, Kelly, and Faust (2017)	60 ET, 30 control	Familial and sporadic ET cases share similar cerebellar postmortem features
Louis et al. (2018)	43 ET, 22 control, 31 SCAs	Heterotopic PCs and PC loss in ET are along spectrum of changes seen in SCAs
Lee, Gan, Faust, Louis, and Kuo (2018)	60 ET, 30 control	Similar climbing fiber-PC synaptic changes across clinical subtypes of ET
Lee et al. (2019)	50 ET, 25 control, 27 SCA, 25 PD	“Empty baskets” provides an indirect measure of PC loss, correlates with PC loss and torpedoes
Louis et al. (2019)	50 ET, 25 control, 81 disease controls	Distinctive patterns of pathologic changes in ET compared to primary cerebellar degenerations and other disorders involving cerebellum
Pan et al. (2020)	15 ET, 15 control	75% reduction in cerebellar GluRδ2 protein levels, inverse correlation with metrics of CF rewiring
Hartstone et al. (2021)	25 ET, 25 control	No difference in dentate nucleus neuronal density or total neuronal number in ET versus control
Wu et al. (2021)	15 ET, 15 control	Increased climbing fiber-PC crossings in ET, correlates with tremor severity
Gionco et al. (2021)	50 ET, no controls	No pathological differences in cerebellar cortex between ET versus “ET-plus” cases

Aβ, amyloid beta; AD, Alzheimer’s disease; DBS, deep brain stimulation; EAAT2, excitatory amino acid transporter type 2; ET, essential tremor; GABA, gamma-amino-butyric acid; GluRδ2, glutamate receptor delta 2; PC, Purkinje cell; PD/DLBD, Parkinson’s disease/Diffuse Lewy Body disease; PSP, progressive supranuclear palsy; SCA, spinocerebellar ataxia.

Fig. 1.

Pathological changes in ET cerebellum. A putative cascade of early (A–F), middle (G–J) and late (K–N) changes is presented. Early changes: (A–D) Focal PC axonal swellings (“torpedoes”) identified in LH&E stain (A, arrow), calbindinD28k immunostain (B, arrow), phosphorylated-neurofilament immunostain (C, arrow), and electron microscopy (D). There is abnormal accumulation of phosphorylated-neurofilaments in the cell body of a PC harboring a torpedo (C). In addition to rounded swellings, PC axons may be diffusely thickened (B, arrowhead). Dendritic abnormalities include focal dendrite swellings (E, arrowhead, LH&E stain) and excessive pruning of PC dendrites (F, arrows, Golgi-Kopsh stain). Middle changes: Decreased PC linear density and Bergmann gliosis in LH&E stain (G, ovals). PC loss detected in calbindinD28K-GAD dual immunostain by increased empty baskets (H, arrows). Complex PC axonal changes revealed by calbindinD28k immunostain, including excessive branching (I, black arrows), recurrent collaterals (I, arrowheads) and terminal axonal sprouting (J, K, white arrows). Late changes: Basket cell axonal plexus hypertrophy (K, Bielschowsky stain). Heterotopic PCs (L, M, arrowheads) typically lack a surrounding basket cell axon plexus (L, Bielschowsky stain) and may have torpedoes along their axon (M, arrows, LH&E stain). Abnormal CF-PC synapses (N, arrows, VGlut2 immunostain) in the outer molecular layer (demarcated by dotted line). GAD, glutamic acid decarboxylase; LH&E, luxol fast blue-hematoxylin & eosin stain.

4.1. Putative “early” events: Purkinje cell axonal swellings and dendritic changes

Pathological events that indicate stress and underlying compromise in PCs are likely to occur early in the disease, prior to attempted compensatory responses or cell death in susceptible cells (Fig. 1A–F). This involves both axonal and dendritic compartments of PCs in ET.

4.1.1. Purkinje cell axonal torpedoes

One of the first pathological changes identified in cerebellum of postmortem ET cases was an increased number of PC torpedoes, which are round or ovoid swellings of the proximal portion of the PC axon (Louis et al., 2006, 2007; Louis, Vonsattel, Honig, Ross, et al., 2006) (Fig. 1A). Subsequent studies have demonstrated a 2- to 12-fold increase in number of torpedoes in ET versus age-matched controls, in both the motor region of cerebellar cortex (Babij et al., 2013; Kuo, Wang, et al., 2017; Louis et al., 2007, 2017; Louis, Kuo, Tate, et al., 2017) and cerebellar vermis (Louis, Faust, et al., 2011). Torpedoes are also increased in ET over that seen in other neurodegenerative disorders such as PD, AD, and dystonia (Louis et al., 2019; Louis, Faust, Vonsattel, Honig, Rajput, et al., 2009b). Ultrastructurally, torpedoes contain massive accumulations of disorganized neurofilaments admixed with mitochondria and smooth endoplasmic reticulum and occur along unmyelinated segments of the PC axon (Fig. 1D) (Louis, Yi, Erickson-Davis, Vonsattel, & Faust, 2009). A quantitative Western blot analysis in postmortem cerebellar cortex showed significantly higher levels of both phosphorylated and non-phosphorylated forms of neurofilament heavy chain subunit in ET compared to controls, suggesting these molecular differences could underly torpedo formation in ET (Louis et al., 2012). Abnormal phosphorylation of neuronal cytoskeletal proteins is one of the major pathological hallmarks of degenerative disorders such as AD, PD and amyotrophic lateral sclerosis (Petzold, 2005).

Torpedoes are not specific to ET, being seen in a wide range of conditions with cerebellar degeneration, such as spinocerebellar ataxia (SCA), multiple system atrophy, progressive supranuclear palsy and paraneoplastic cerebellar ataxia (Louis et al., 2019; Louis, Kuo, Vonsattel, & Faust, 2014). They do not prominently accumulate with normal advanced aging (Louis, Faust, Vonsattel, & Erickson-Davis, 2009). There is general agreement that torpedo formation is an injury associated alteration in PCs with resultant inhibition of both anterograde and retrograde axonal transport, consistent with the presence of phosphorylated neurofilaments in the cell body of PCs harboring a torpedo (Fig. 1B; Louis et al., 2012). This axonal injury might ultimately lead to “neuronal strangulation” (Louis & Faust, 2020c). While it remains unclear whether torpedoes represent a preterminal cellular response or a regenerative response, the robust inverse correlation between number of torpedoes and PC counts in ET and other cerebellar diseases suggests that torpedoes are part of a degenerative cascade (Axelrad et al., 2008; Choe et al., 2016; Louis, Kuo, Vonsattel, & Faust, 2014).

4.1.2. Purkinje cell dendritic changes

Dendritic abnormalities traditionally occur in degenerative diseases where they reflect derangements in cellular homeostatic processes that often precede neuronal death. An increased number of PC focal dendritic swellings in ET compared to control brains has been demonstrated in two studies (Louis et al., 2019; Yu et al., 2012) (Fig. 1E). The robust correlation between number of PC dendritic swellings and number of torpedoes in a brain (r = 0.60, P−0.005) (Yu et al., 2012) suggests a cascade of interconnected events in the degenerative process involving PCs.

Other changes in the PC dendritic compartment in ET also support a degenerative process. In a detailed quantitative study of PC dendrite arbors in 27 ET and 27 age-matched controls using the Golgi-Kopsh method, there were significant reductions in spine density (19% reduced, P = 0.03) and several measures of dendritic complexity (21%–47% reduced, P = 0.001–0.01) in ET (Louis, Lee, Babij, et al., 2014) (Fig. 1F). The dendritic arbor and spine density variables were all highly intercorrelated, indicating a system-wide set of regressive changes. In addition, tremor of greater severity was correlated with more pruning of the PC dendritic arbor. Regressive dendritic changes are not specific to ET, being seen in patients with hereditary ataxias (Shintaku & Kaneda, 2009) as well as chronic alcoholics (Ferrer, Fabregues, Pineda, Gracia, & Ribalta, 1984), but may be considered a marker of neurons in extremis. Structural perturbations of the PC dendrite, such as swellings and changes in complexity, would likely alter the interface with the multiple synaptic inputs to PCs. These are likely to have physiological implications.

4.2. Putative “middle” events: Complex PC axonal changes and selective PC death

Although symptomatic loss of function in degenerative diseases is frequently attributed to the loss of neurons, events before this neuronal loss might well result in disease manifestations. Following early cellular changes, middle events represent ongoing responses in PCs that lead to neuronal remodeling, and failed attempts leading to eventual PC death (Fig. 1G–I).

4.2.1. Remodeling of PC axons

The cellular response of PCs to injury is not typical of that seen in most CNS neurons, as they survive for prolonged periods after axonal injury despite the inability to regenerate their severed axon (Rossi, Gianola, & Corvetti, 2006). An initial formation of torpedoes and hypertrophy of recurrent collateral branches (which form connections with neighboring PCs) is followed by a vigorous inclination towards axonal sprouting along the intracortical segment of their axon. These PC axonal reorganizations might enable PCs to access trophic factors by establishing additional connections with other PCs or granule layer neurons in their attempt to preserve connections and prevent cell death (Carulli, Buffo, & Strata, 2004). However, the extent to which these plasticity-type changes in PC axons remain neuroprotective versus inducing further degeneration is not known.

Morphologic analysis of the PC axonal compartment was performed in 49 ET and 39 control brains using calbindinD28k immunohistochemistry in 100-μm thick cerebellar cortical tissue sections (Babij et al., 2013), enabling detailed visualization of the trajectory of PC axons as they traverse the granule cell layer. In addition to axonal torpedoes, other axonal alterations were all present to an increased degree in ET cases versus controls, including changes in axonal shape (diffusely thickened axonal profiles, 17% increase, P = 0.006) and axonal connectivity (axonal recurrent collaterals, 2.3× increase; axonal branching, 2.2× increase; terminal axonal sprouting, 2.5× increase; all P< 0.001) (Fig. 1C, G, H). These changes in axonal shape and connectivity were significantly intercorrelated and were three- to fivefold more frequently seen on axons of PCs with torpedoes, indicating a set of biologically related processes. These findings indicate there is a broader rewiring of PCs in ET, with potential for formation of aberrant synapses and abnormal circuits that might precede or coincide with neuronal loss.

In a subsequent clinicopathological study of ET patients, the extent of these complex PC axonal changes was quantified in relation to laterality or symmetry of tremor on neurological examination (Louis, Lee, Cortés, et al., 2014). Thus, if these pathological changes were tremor-producing, they should be more prominent in the cerebellar hemisphere ipsilateral to the arm with more severe action tremor. There was a high concordance between clinical and pathological features in 18 of 25 (72%) ET cases (P = 0.007). A positive correlation between the degree of clinical asymmetry and pathological asymmetry (r = 0.52, P = 0.01) was even more robust in the subset of ET patients that had the greatest clinical asymmetry (r = 0.78, P = 0.39). This study provides important construct validation that the pathological changes observed in the cerebellum in ET are of patho-mechanistic importance.

4.2.2. Purkinje cell loss

Many pathology series, but not all, have identified a significant decrease in PCs in ET compared to controls. The first postmortem study to examine PC loss in the largest number of ET specimens available at that time compared changes in 33 ET and 21 age-matched controls, demonstrating 25% lower mean PC counts per 100× microscopic field in all patients with ET (P <0.01) and 31% lower in a subset of 25 ET cases that did not have Lewy bodies (P <0.01) (Louis et al., 2007).

Among the current pathological findings in ET, the issue of PC loss remains controversial. Some of the earlier studies still lacked ET-control comparisons (Table 3) (Rajput et al., 2004) or did not use quantitative methods (Ross et al., 2004; Shill et al., 2008). Two subsequent studies did not reproduce the finding of PC loss in ET (Rajput et al., 2011; Symanski et al., 2014). As previously described in detail (Louis & Faust, 2020c), these studies were hampered by several methodologic limitations. A follow-up study on a new cohort of brains from 50 ET versus 25 age-matched controls used a more sophisticated random sampling scheme and a nearest neighbor analysis and replicated the finding of significant PC loss in individuals with ET (Choe et al., 2016). In this study, 94% of individuals with ET had a PC count below the central tendency (the mean for the norm) for the controls, and the mean distance between PC bodies was greater in ET, further consistent with PC loss.

A novel assay was subsequently developed to further examine the question of PC loss in ET by quantifying “empty baskets” (Lee et al., 2019). Basket cells are interneurons in the cerebellar molecular layer whose axons form a plexus around the PC soma. With PC loss, the basket plexus appears empty (Fig. 1J). In a study of 50 ET and 25 controls, the median percentage of empty baskets in ET patients was 1.5 times higher than controls (48.8% versus 33.5%, P <0.001). Even higher percentage of empty baskets was found in 22 SCA cases, including SCA1 (59.7%, P = 0.011), SCA2 (77.5%, P <0.003), and SCA6 (87.0%, P <0.001). The presence of empty baskets, particularly in SCA cases where PC loss is often marked, indicates that the basket cell plexus persists even after the loss of PCs. Thus, quantification of empty baskets provides an indirect and alternative method to quantify PC loss. Furthermore, the strong correlation between the percentage of empty baskets, PC loss and torpedo counts suggests that they are all part of the same pathophysiological cascade.

In sum, these studies suggest that some PCs become overwhelmed in ET, resulting in death of these cells. The reason for the selective nature of PC loss is not known.

4.3. Putative “late” events: Further remodeling of basket cell plexus and PCs and reorganization of climbing fiber-PC synapses

Several pathological changes have been observed in ET that might be a response to PC death, including remodeling of neuronal populations that normally synapse with PCs (basket cells, CFs) or involving the PCs themselves (Fig. 1K–N).

4.3.1. Basket cell axonal remodeling

Basket cells are GABAergic inhibitory interneurons whose axonal processes form a pericellular basket around the PC body and in a “pinceau” structure around the PC axon initial segment (AIS). The density of the basket cell plexus around the PC soma on Bielschowsky-stained sections of neocerebellum was examined in 37 ET, 48 disease control brains (patients with AD, PD, progressive supranuclear palsy or dystonia) and 21 normal controls using a semiquantitative hypertrophy rating scale (Erickson-Davis et al., 2010). A dense and tangled appearance of the perisomal basket cell plexus (termed “hairy baskets”) was seen in 25% of ET brains as compared with 4% of other brains (P = 0.001) (Fig. 1K). This was the first observation, subsequently confirmed in additional studies (Kuo, Wang, et al., 2017; Louis, Kuo, Tate, et al., 2017; Louis, Kuo, Wang, et al., 2017), indicating that structural changes in ET are not restricted to the PC but also involve their functional network. It is postulated that in the context of PC loss in ET, relative preservation of basket cells can lead to basket cell axon sprouting onto remaining PCs, resulting in increased process density and hypertrophic appearance of the basket cell plexus.

A second finding involved the basket cell plexus pinceau in ET. Genomic association studies in ET identified a sequence variation in the gene encoding leucine rich repeat and immunoglobulin-like domain-containing nogo receptor-interacting protein 1 (Lingo-1) (Clark & Louis, 2018). Lingo-1 immunohistochemistry labeled the basket cell pinceau, and ET cases had abnormally long pinceau processes targeting more distal segments of the PC axon over that seen in control brains (Kuo et al., 2013). Elongated Lingo-1-positive pinceau strongly correlated with number of PC axonal torpedoes and the basket cell hypertrophy rating, features more commonly seen in ET brains. This basket cell pinceau remodeling in ET could represent a disturbance in the AIS and neuronal polarity, as accumulations of nonphosphorylated neurofilaments in the PC AIS (Louis et al., 2012), and phosphorylated neurofilaments in PC soma (Erickson-Davis et al., 2010) have also been observed in ET cases (Fig. 1B).

Two studies have demonstrated increased cerebellar expression of Lingo-1 in ET versus control brains (Delay et al., 2014; Kuo et al., 2013). As Lingo-1 is a negative regulator of axonal regeneration and neurite out-growth, its upregulation might be a response to the increased axonal sprouting and branching of PC axons observed in ET, and possibly in other neuronal populations yet to be explored. Lingo-1 upregulation is seen in several disorders, such as PD, multiple sclerosis and spinal cord injury (see Delay et al., 2014), and would be consistent with a degenerative process in ET cerebellum.

4.3.2. Heterotopic Purkinje cells

PCs are normally located as a monolayer between the molecular layer and granule cell layers of the cerebellar cortex. Heterotopic PCs, whose cell body is mislocalized in the molecular layer (Fig. 1L–M), were quantified in the ET cerebellum in two studies, comparing ET brains to both neurologically normal controls as well those from individuals with progressive supranuclear palsy (PSP) and SCAs (Kuo, Erickson-Davis, et al., 2011; Louis, Kuo, et al., 2018). The mean or median number of heterotopic PCs was twofold higher in ET than in normal controls (P <0.05), and heterotopic PCs in ET were seen singly rather than in clusters, most consistent with a degenerative process (Yang et al., 2000). Heterotopic PCs were not increased in PSP brains but were even more marked in several of the SCAs. In this regard, ET could represent an intermediate or less advanced state of spinocerebellar ataxia (Louis et al., 2019; Louis & Faust, 2020b, 2020c). As the number of heterotopic PCs and number of PCs are inversely related in ET cerebellum (Louis, Diaz, et al., 2018; Louis, Kuo, et al., 2018), these heterotopic PCs may reflect an untethering associated with anatomical rearrangement in the molecular layer, perhaps due to the PC dendritic pruning and PC loss observed in ET. Combinations of pathological changes are also seen in heterotopic PCs, with a loss of the surrounding inhibitory basket cell axonal plexus and axonal torpedoes along their axon (Fig. 1L–M), further suggesting these PCs are functionally abnormal.

4.3.3. Reorganization of climbing fibers

CFs from olivary neurons provide one of the major excitatory inputs to PCs. These synaptic connections are essential for normal cerebellar-mediated motor control and are dynamically regulated by activity and disease states (Watanabe, 2008). Normally, CFs form synapses mainly on the thick, proximal PC dendrites in the inner portion of the molecular layer, whereas parallel fibers from granule cells form synapse in the outer molecular layer and on the thin, distal PC spiny branchlets. In a study of ET cerebellum versus age-matched controls, three main abnormalities in CF distribution were detected in ET, including decreased CF-PC synaptic density (33% reduction, P< 0.001), an abnormal extension of CFs to the outer portion of the molecular layer (1.9-fold increase, P <0.001) and more CF-PC synapses on the thin PC distal branchlets (2.2-fold increase, P <0.003) (Lin et al., 2014). The abnormal extension of CFs into domains normally occupied by parallel fiber synapses indicates a redistribution of synaptic wiring in the molecular layer in ET. Correlations between CF extension into the outer molecular layer, number of torpedoes and extent of PC loss in ET suggest that this change is a downstream response to PC loss (Lin et al., 2014). There was also an unexpected strong negative correlation between CF synaptic density on PC distal branchlets and total tremor scores in ET (Lin et al., 2014; Louis et al., 2015). Thus, the pattern of CF-PC innervation may change with disease severity in ET, where PC degeneration and/or persistent long-standing tremor may cause preferential pruning of distal PC branches and CFs, obscuring and/or reversing the pattern that is seen in mild ET.

Another important feature of CFs and PCs is a 1:1 pattern of innervation, where one CF innervates one PC (Kano, Watanabe, Uesaka, & Watanabe, 2018). Using a novel immunohistochemical staining technique that labeled a subset of cerebellar CFs, CFs in ET were found to cross a greater number of PC somas (19% increase, P = 0.023) and PC dendrites (39% increased, P = 0.014) than in control cases, and these increases in CF-PC crossings positively correlated with tremor severity (Wu et al., 2021). While the pathophysiologic implication of these changes is uncertain, it might indicate a rewiring, with increased synaptic contacts between CFs and PCs and/or potentially multiple PC innervations by CFs.

The redistribution of CF-PC synapses shows some specificity to ET, as this pathological change is enriched in ET cases compared to other cerebellar degenerations and movement disorders, including SCAs, multiple system atrophy, PD, and dystonia (Kuo, Lin, et al., 2017; Louis et al., 2019). They are consistently observed in ET with varying clinical features (Lee et al., 2018), suggesting this synaptic pathology, among others, might be a core feature of ET.

Similar redistribution of CF synapses is also seen in a mouse model with defective glutamate receptor δ2 (GluRδ2), a key protein controlling the territorial distribution of CF and parallel fiber synapses on PC dendrites (Pan et al., 2020). In an analysis of 15 ET and 15 controls, patients with ET had approximately 75% reduction in mean GluRδ2 protein expression level, and GluRδ2 levels inversely correlated with the percentage of CFs extending into parallel fiber synaptic territory (Pan et al., 2020). Interestingly, GluRδ2-deficient mice develop an ET-like tremor, and CF-PC synaptic pruning deficits drive excessive cerebellar oscillatory activity that is necessary for tremor generation (Pan et al., 2020). This study was also the first to use cerebellar electroencephalography to confirm the presence of excessive cerebellar oscillations in ET patients compared to controls, and that the extent of these oscillations correlate with clinical tremor scores. Thus, the combination of CF structural alterations identified in ET across these multiple pathological studies might result in entrainment of PC activity to a more synchronous and rhythmic mode, leading to tremor. It is also possible that the abnormal CF-PC synaptic connections in ET could represent a compensatory change, resulting from longstanding rhythmic firing of CFs onto PCs.

4.3.4. Changes in astroglia

Pathological studies have identified an increased presence of Bergmann glia in ET, a stereotypic reaction in cerebellum associated with PC loss (Louis et al., 2007; Shill et al., 2008). Genetic polymorphisms in Solute carrier family 1 (glial high affinity glutamate transporter), member 2 (SLC1A2) have been linked with ET (Clark & Louis, 2018). SLC1A2 encodes excitatory amino acid transporter type 2 (EAAT2), which clears glutamate from the synaptic cleft. Using both immunohistochemistry and Western blot analyses, EAAT2 expression was decreased by 40%–65% in cerebellar cortex of ET compared to controls (Lee et al., 2014). This finding raises the possibility that ET cerebellar cortex might be more vulnerable to excitotoxic damage than those of controls, resulting from its numerous glutamatergic synapses.

EAAT2 expression was also found to be 1.3- to 1.5-fold higher in the dentate nucleus of ET patients compared to controls (Wang et al., 2016). The mechanism behind this differential expression pattern of EAAT2 in the ET cerebellar cortex versus dentate nucleus remains to be investigated. The balance between inhibitory inputs from multiple PCs and excitatory inhibitory inputs from mossy fiber and CF collaterals on dentate neurons is critical for modulating their spike frequency and thus important for cerebellar function and proper motor control (Person & Raman, 2012). It is possible that lower levels of inhibition in the dentate nucleus, due to PC loss in ET, could trigger a compensatory increase in EAAT2 to maintain excitation-inhibition balance in cerebellar nuclei.

4.3.5. Changes in the dentate nucleus

PCs send dense inhibitory GABAergic inputs to cerebellar nuclei that modulate dentate neuron spike frequency. In a study of 10 ET and 16 control brains, reductions in GABA_A (35%) and GABA_B (22%–31%) receptors was detected in ET (P <0.05) (Paris-Robidas et al., 2012). Concentrations of GABA_B receptors in dentate inversely correlated with duration of tremor symptoms in ET, suggesting this receptor loss follows disease progression, consistent with a degenerative process. A diminished GABA receptor efficacy in ET might lead to overactivity of cerebellar nuclei neurons and spread up through the cerebellar-thalamo-cortical circuit, contributing to the generation of tremors.

Current evidence does not support the presence of extensive neuronal loss in the dentate nucleus in ET. While neuronal loss was qualitatively noted in 2 of 33 (6.1%) ET patients in one case series (Louis et al., 2007), a larger quantitative study in 25 ET compared to 25 age-matched controls did not find differences in dentate nucleus neuronal density (Hartstone et al., 2021). A stereological analysis in a subset of ET and control brains in this study also did not show differences in total number of dentate neurons. These findings do not exclude that there are subtle anatomical changes, other than neuronal cell death, or neurochemical changes in the dentate that might contribute to tremor generation in ET.

5. Cerebellar pathology and clinical heterogeneity in ET

Considering the considerable clinical heterogeneity in ET, case definition has been an area of increasing focus. The term “ET-plus” was introduced in a recent consensus statement to distinguish and separate ET cases with signs of dystonia, cerebellar dysfunction, and rest tremor, among others, from those without such signs (Bhatia et al., 2018). This diagnostic concept has raised considerable controversy and its validity is not yet established. Indeed, “ET-plus” has not been distinguished from ET based on differences in genetics or prognosis. To address whether ET cases differ from “ET-plus” cases in underlying pathological changes in the postmortem brain, 50 ET cases (24 ET and 26 ET-plus) were examined using a set of 14 quantitative metrics of cerebellar pathology reflecting changes across the PC body, dendrites, and axon, and basket cell-PC and CF-PC synaptic alterations (Gionco et al., 2021). ET and ET-plus were similar with respect to 13 of 14 cerebellar pathologic metrics, and after correcting for multiple comparisons, there were no differences. Furthermore, none of the cerebellar pathologic metrics significantly differed when ET cases were grouped by rest tremor or intention tremor. This study demonstrated there were no pathological differences in motor cerebellar cortex between ET versus ET-plus cases, and thus do not currently support the notion that ET and ET-plus represent distinct clinical-pathological entities.

6. ET as a degenerative disorder: Insight into diseasemechanisms

The tissue-based research studies in ET and supportive findings from clinical features and neuroimaging studies now provide a patho-mechanistic model of ET—the cerebellar degenerative model. The confluence of abnormalities in the cerebellum suggest it may be a critical hub for generation of tremor. Although much work remains to be done to fully understand the underlying basis for this disease, two potential disease mechanisms are currently highlighted: (1) a diminution of GABAergic control and (2) reorganization of cerebellar cortex.

PCs provide the entire output from cerebellar cortex and provide dense GABAergic inhibitory inputs to cerebellar nuclei. Pathological changes observed in PCs, such as torpedoes, dendritic swellings, pruning of the PC dendritic arbor, and eventual PC loss might represent a primary and degenerative change that would decrease GABAergic inputs to cerebellar nuclei. In addition, the increased PC recurrent collateral formation as well as basket cell axonal plexus hypertrophy would act synergistically to further reduce PC inhibitory output onto remaining PCs (Louis & Faust, 2020c). These events may be additive and destructive to result in a hyper-excitable state in cerebellar nuclei that drives abnormal burst firing, as both the rate and timing of firing by cerebellar nuclear neurons under normal conditions is precisely regulated, at least in part, by the coordinated firing of PCs (Person & Raman, 2012).

Cerebellar control of movement requires precise and coordinated firing of its neuronal populations. A key feature of this control is the coordination, or synchronization, of PC oscillations, which converge in cerebellar nuclei to temporally organize cerebellar output (de Solages, Szapiro, Brunel, et al., 2008). Several pathological changes observed in ET cerebellum would lead to rewiring of cerebellar circuits and may facilitate generation of coordinated PC oscillations. Synaptic pruning deficits in CFs like that seen in human ET have been linked to tremor generation in a mouse model with ET-like tremor (Pan et al., 2020), and gap-junction coupling of ION neurons promotes synchronous PC firing. PC recurrent collateral synapses facilitate coordinated PC oscillations (Witter, Rudolph, Pressler, Lahlaf, & Regehr, 2016). Molecular layer interneurons shape PC firing patterns (Brown, Arancillo, Lin, et al., 2019), and their gap-junction mediated coupling within the molecular layer and ephaptic coupling at PC axonal initial segments by basket cells both enhance synchronous firing in cerebellar cortex (Han et al., 2018; Hoehne, McFadden, & DiGregorio, 2020). While many of these plasticity-type changes in PC inputs may initially represent compensatory responses, they are likely insufficient and possibly maladaptive or destructive, leading to progression of clinical signs in ET.

Our current understanding of ET pathology is largely based on extensive characterization within the motor region of cerebellar cortex, which is the region of greatest interest given the primary motoric nature of ET. Whether the morphological disease pattern seen in this region applies to other anatomic and functional compartments of cerebellum remains to be determined. The current disease model has revealed the greatest number of changes in the cerebellum, although it is theoretically possible that more subtle changes could extend more widely within the tremor circuit. Indeed, neuroimaging studies have identified a wide array of changes outside the cerebellum in ET, but their role in motor and non-motor features in this disease remain to be defined. Additional studies examining synaptic, axonal and other changes across regions in the physiologic tremor network is a challenge for future studies.

7. Conclusions

ET is a progressive, aging-associated condition with selective neuronal loss (affecting PCs) and other underlying structural changes in the cerebellum that disrupt normal function, all of which are features of a degenerative disorder. Both clinically and pathologically, there are numerous similarities and intersection points between ET and other disorders of cerebellar degeneration. Furthermore, considerable clinical associations suggest that ET may share disease mechanisms with other degenerative disorders, including PD and AD. The constellation of clinical non-motor findings in ET raises the possibility there is dysfunction in other brain regions, as seen in other degenerative disorders. A wide array of neuroimaging studies has identified functional and structural abnormalities in the cerebellum, although it remains to be established to what degree these contribute to clinical symptoms of ET. In the end, there are likely considerable overlaps between the degenerative and electric hypotheses for ET, including a current focus on central involvement of the cerebellum for tremor generation. There may be future merging of aspects of these hypotheses as we begin to explore whether cellular reactions occur elsewhere within the tremor network. A major challenge is to apply uniform and stringent inclusion criteria in ET disease studies and to organize the considerable clinical, pathological, and molecular heterogeneity within ET.

Abbreviations

AD

Alzheimer’s disease

CF

climbing fiber

DTI

diffusion tensor imagingfiber

EAAT2

excitatory amino acid transporter type 2

ET

essential tremor

fMRI

functional magnetic resonance imaging

ION

inferior olivary nucleus

MR

magnetic resonance

MRS

magnetic resonance spectroscopy

PC

Purkinje cell

PD

Parkinson’s disease

PET

positron emission tomography

PSP

progressive supranuclear palsy

SCA

spinocerebellar ataxia