Neuropathologic findings in essential tremor

A fundamental question about the biology of essential tremor (ET) is whether an underlying pathology can be identified.¹ There have been only 25 published postmortem brain examinations, with most reported as single cases.^1,2 Many were published 50 to 100 years ago and included little clinical information and questionable diagnoses. Moreover, pertinent structures such as the cerebellum and brainstem nuclei often were not examined in a comprehensive manner, and most studies did not perform immunohistochemistry. Most importantly, these studies did not include control brains for comparison. The Essential Tremor Centralized Brain Repository (ETCBR) at Columbia University was established to focus attention on the pathology of ET. Our hypothesis, based on clinical and functional imaging studies,³ was that pathologic findings in ET might involve the cerebellum. The ETCBR is the result of a collaborative effort between investigators at several medical centers and the International Essential Tremor Foundation; both archival and prospectively collected postmortem brains have been obtained.

Methods.

Archival (n = 6) and prospectively collected (n = 4) postmortem ET brains and 12 control brains were obtained. The tissue was processed according to a standardized protocol. Details on case ascertainment, tissue processing methods, microscopic methods (quantifying Lewy bodies, Purkinje cells, torpedoes, Bergmann astrogliosis) and data analyses are provided on the Neurology Web site (www.neurology.org).

Results.

ET cases were similar to controls (table). Case 5 had carried a diagnosis of ET for 65 years and developed Parkinson disease (PD) with dementia in the last 11 years of life. In Case 8, slowness and stooped posture were noted at age 81 years but not at reexamination at age 83 years; the patient died at age 85 years. None of the remaining eight ET cases had tremor at rest or other motor features of parkinsonism.

Six (60%) of 10 ET cases (Cases 5 through 10) had brainstem Lewy bodies on alpha-synuclein–stained sections vs 2 (16.7%) of 12 controls (odds ratio 7.5, 95% CI 1.04 to 54.1; p = 0.035; table E-1). The final examination of these six cases had taken place a median of 3 months before death (mean 6.7 months, range 1 to 24 months).

In two ET cases (Cases 5 and 6), there was moderate to severe neuronal loss in the substantia nigra pars compacta (SNC) (caudal > rostral involvement) and dorsal vagal nucleus (DVN), and there were similar numbers of Lewy bodies in the DVN, locus ceruleus (LC), and SNC. These two cases met pathologic criteria for PD (table E-1).

Cases 7 through 10 had an atypical and somewhat restricted pattern of Lewy body distribution in the brainstem. This distribution showed a much greater density of Lewy bodies in the LC than in the well-pigmented DVN or SNC (table E-1 and figure). None of the four had parkinsonism on examination (three were examined in person by their treating physician 1, 2, and 4 months before death, and one had a videotaped examination 24 months before death).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Lewy body pathology in essential tremor (A through F = Case 9 and G and H = Case 4). (A) No Lewy bodies on a luxol fast-blue and hematoxylin and eosin (LH&E)–stained section of the dorsal vagal nucleus (200×). (B) A single Lewy body (arrow) on an alpha-synuclein–stained section of the dorsal vagal nucleus (200×). (C) Multiple Lewy bodies (arrows) on an LH&E-stained section of the locus ceruleus (200×). Each of the neurons denoted by an arrow has 2 or more Lewy bodies, and 7 Lewy bodies are visible in this field. (D) Multiple Lewy bodies on an alpha-synuclein–stained section of the locus ceruleus (200×). Two arrows denote Lewy bodies, although more than 10 Lewy bodies are visualized in this field. (E) No Lewy bodies on an LH&E-stained section of a well-pigmented substantia nigra pars compacta (100×). (F) No Lewy bodies on an alpha-synuclein–stained section of the substantia nigra pars compacta (100×). (G) LH&E-stained section of cerebellar folium showing torpedoes (two in box) (100×). (H) Two torpedoes (in box in G) (400×).

The presence of Lewy bodies was not correlated with age or postmortem interval. When we adjusted for age and postmortem interval, the association between presence of Lewy bodies (dependent variable) and diagnosis (ET vs control) remained significant (p = 0.04).

Numbers of torpedoes (p = 0.009) and Bergmann glia (p = 0.046) were increased in ET cases compared with controls (especially Cases 1 through 4; see table, table E-1, and figure). Number of torpedoes was weakly associated with duration of the postmortem interval (r = 0.42, p = 0.05) but not with age, sex or race; number of Bergmann glia was not associated with any of these confounders. In a linear regression analysis that adjusted for postmortem interval and age, the association between number of torpedoes (dependent variable) and ET remained significant (β = 7.74, p = 0.004).

Additional (secondary) regions of interest were examined in ET cases and controls. Neither cell loss nor gliosis was present in the cerebellum or these secondary regions (red nucleus, thalamus, inferior olivary nucleus, and motor cortex).

ET cases with Lewy bodies were clinically similar to ET cases without Lewy bodies in terms of age, duration, and family history of ET (see table), although small numbers precluded meaningful comparisons.

Discussion.

ET cases clustered into two preliminary groupings: those with brainstem Lewy bodies and those with mild cerebellar changes. Other investigators⁴ recently noted a higher proportion of Lewy body pathology in the brainstem and the cerebellum in patients with bilateral action tremor compared with controls. The current results suggest that ET could be pathologically heterogeneous.

Strengths of this study included the collaborative, multicenter effort; prospective collection of four ET brains; detailed and standardized assessment of cerebellar pathology; careful approach to quantification of Lewy bodies in all relevant brainstem nuclei; and use of a control group.

Four ET cases had mild cerebellar changes, consistent with studies suggesting that the cerebellum is involved in ET.³ Bergmann cells are a nonspecific pathologic response to injury. Torpedoes, which consist of massive accumulations of disoriented neurofilaments,⁵ occur in degenerating and possibly regenerating Purkinje cells. They have also been described in other disease processes involving destruction of cerebellar tissue (e.g., cerebellar ataxias). Our failure to detect a reduction in Purkinje cell numbers in ET patients may reflect our semiquantitative approach; quantitative (stereologic) methods, which are preferable, are planned.

Lewy bodies were present in a substantial proportion of ET cases. LC Lewy bodies occur rarely (0 to 11.5%) in normal elderly controls.⁶ The rarity of LC Lewy bodies in our control brains further suggests that the Lewy bodies in our cases were not the result of normal aging. The spectrum of Lewy body disease includes entities other than PD (diffuse Lewy body disease and Alzheimer disease with Lewy bodies).⁷ That a significant proportion of ET cases could have Lewy body disease lends support to the literature suggesting a link between ET and Lewy body disease. The mechanism whereby Lewy bodies mainly in the LC could result in action tremor is not clear. The LC is the principal source of norepinephrine in the CNS.⁸ An LC lesion could result in a diminution of stimulatory output from the LC to the inhibitory Purkinje cells (i.e., resulting in a net reduction in inhibitory output from the Purkinje cells).

Patients with ET may over time develop PD. However, the particular topographic pattern of Lewy bodies seen in ET Cases 7 through 10 (a much greater density of Lewy bodies in the LC than in the well-pigmented DVN or SNC) has rarely been reported in presumed preclinical PD or in clinically evident PD. For example, postmortem examinations were performed in 110 individuals (69 individuals with incidental Lewy bodies [presumed preclinical PD] and 41 individuals with clinical diagnoses of PD), and a staging scheme was proposed for preclinical and clinical PD (Stage 1 = slight involvement of the DVN; Stage 2 = moderate involvement of the DVN and slight involvement of the LC; Stage 3 = moderate to severe involvement of the DVN, slight to moderate involvement of the LC, and slight to moderate involvement of the SNC).⁹ The topographic pattern of Lewy bodies observed in Cases 7 through 10 was not described in any of the 110 individuals.⁹ Similarly, in another study¹⁰ of 254 individuals (196 incidental Lewy body cases and 58 with parkinsonism or cognitive changes), the typical pattern of Lewy body involvement (DVN > LC > SNC) described by the previous investigators⁹ occurred in 253 (99.6%) of 254 cases.

Our finding of Lewy bodies in ET raises several questions. What proportion of ET cases has Lewy bodies? Does the action tremor in these cases represent the arrested development of Lewy body pathology in the LC? What proportion of ET cases will develop more widespread Lewy body pathology and motor signs of PD? Research at the ETCBR will continue to address these questions.