Abstract
Background: Because motor manifestations of ALS begin focally and progress contiguously, the anatomic distribution of underlying lower motor neuron and upper motor neuron degeneration should correlate to onset.
Objectives: To assess the rostral–caudal distribution of lower motor neuron loss in relation to the region of clinical onset.
Methods: We evaluated 19 ALS postmortem nervous systems from patients whose motor manifestations had begun in different body regions. In each, we looked at four neuraxis levels: hypoglossal nucleus and cervical, thoracic, and lumbar spinal cord. We used light microscopy and devised a technique of particle counting that indexed lower motor neuron loss.
Results: The average overall loss of lower motor neurons in ALS nervous systems was 55%, and the range of loss had a normal distribution that ranged between 8% and 90%. The distribution of lower motor neuron loss was graded within the nervous system relative to onset (p = 0.02 by analysis of variance). In 14 of the 19 nervous systems, the regional lower motor neuron loss within the nervous systems was graded radially away from the region of onset. In 1, radial degeneration seemed likely but did not meet significance. In 2, radial degeneration was apparent but loss was greatest in a region different than that identified as the region of onset. In the remaining 2, lower motor neuron loss was minimal and not graded (both from patients whose motor manifestations had been predominantly upper motor neuron).
Conclusion: Lower motor neuron degeneration in ALS is a focal process that advances contiguously, summates over time, and creates graded loss. Stage of degeneration in the nervous system is a function of anatomic location.
Motor manifestations of ALS begin focally and progress contiguously outward.1–3 Because they are caused by simultaneous degeneration of upper motor neurons (UMNs) and lower motor neurons (LMNs), throughout the course of the disease including at the time of death, there should be a relationship between clinical onset and distribution of motor neuron loss—for LMNs, a rostral–caudal gradation; for UMNs, a lateral–medial gradation. It is difficult if even possible to do this for UMNs because of difficulties with identification, localization, standardization, and quantification.4–8 But it is feasible to do this for LMNs because they are distinctly located in motor columns, readily identified, obviously correlated with body regions, and well characterized morphometrically. We therefore did so for LMN loss in a sample of nervous systems from patients whose disease had begun in different body regions. We found that LMN loss was usually radial—greatest at the region of onset and decreased outward. This supports the idea that ALS motor neuron degeneration is a focal process that advances contiguously and summates over time and suggests that stage of degeneration in the nervous system is relative to anatomic location.
METHODS
Selection and demographics.
We evaluated 19 ALS and 8 control nervous systems from our ALS tissue repository. Inclusion criteria for the ALS nervous systems were as follows: nervous systems came from patients who had met modified El Escorial criteria for definite ALS9; disease was sporadic; disease had no atypical clinical or neuropathologic features; disease onset was from representative regions of the body (bulbar, arm, trunk, and leg); and histology was available for all four neuraxis levels—medulla, cervical, thoracic, and lumbar regions. Demographics are shown in table 1. In the ALS group, nervous systems were from 12 men and 7 women; the mean age was 67 years (range 41 to 84 years). In the control group, nervous systems were from 7 men and 1 women; the mean age was 63 years (range 38 to 80 years). Postmortem intervals averaged 5.25 hours (range 2.0 to 12 hours). All nervous systems had been acquired with an Investigational Review Board and Health Insurance Portability and Accountability Act–compliant informed consent process.
Table 1 Demographics
Histologic sampling.
We used 6-μm-thick sections from formalin-fixed, paraffin-embedded tissues stained with cresyl violet acetate. We studied four neuraxis levels: hypoglossal nucleus at the midmedulla, midcervical spinal enlargement, midthoracic spinal region, and midlumbar spinal enlargement. We standardized rostral–caudal sampling of the hypoglossal nucleus by its relation to the medial longitudinal fasciculus and by the morphology of the fourth ventricle.10 We standardized rostral–caudal sampling of spinal cord regions by choosing middle segments—the cervical region has uniform counts between C4 and C8,11,12 the lumbar region has uniform neuron counts between L3 and S2,13–15 and we assume the midthoracic region has uniform counts. Because motor neuron presence varies from one histologic section to another,11,15–17 we evaluated a total of eight sections from each neuraxis level, choosing every 10th section for a sampling interval of 60 μm from a series of 84 consecutive sections spanning 504 μm, a sampling technique modified from Method 1 of Tomlinson et al.16 We separately imaged each side of the spinal anterior horns and hypoglossal nucleus under 50× with a Leica DM2500 microscope with a Spot Insight 4 digital camera and Spot Advanced version 4.5.7 software (Diagnostic Instruments, Inc., Sterling Heights, MI) and stored each image in Tagged-Image File Format.
Counting.
To meet the realities of the counting task, we devised a method of counting neuron “particles” to index relative neuron presence rather than determining absolute neuron counts18—stereology technologies do not apply to the motor neurons19 and automated particle counting software was inefficient, experience reported by others.20 Our method had three observers (J.R., P.L., and Y.F.) independently count motor neuron particles in each image file according to the following criteria learned in training sessions to standardize counting: neurons were located in the anterior portion of the anterior horns of the spinal cord segments or in the hypoglossal,10 had relatively deep Nissl staining, were generally multiconcave, and were larger than 25 μm in diameter in cervical and lumbar regions and larger than 15 to 20 μm in thoracic and hypoglossal regions. Both sides were counted. We were careful to avoid counting neurons in the perihypoglossal nuclei, the intermediolateral cell column, and Clarke's column. Because there is no significant shrinkage of motor neurons in ALS vs controls,20–22 we did not correct for split cell error.23,24
Blinding and controls.
Observers were blind to sample identification, although in many cases, ALS nervous systems were identified by the obvious motor neuron loss; in such cases, the site of clinical onset was not known and not obvious. Control nervous systems were age-matched with the ALS nervous systems, although in our controls, contrary to other reports,14,25,26 age-related changes were not pronounced.
Statistics and calculations.
We compiled particle counts in a spreadsheet and then transferred them to the statistical package Stata (version 9.2) (StataCorp LP, College Station, TX) for statistical analysis by multiway analysis of variance (ANOVA). To normalize variation across observers and across neuraxis levels, we transformed counts to percentages of control values. To do this, we expressed each observer's mean count per level for each ALS nervous system as a percentage of that observer's mean count for that level in controls, thus indexing observers to themselves. We then averaged these percentages across observers for subsequent calculations and ANOVA. Neuron loss was expressed as percent and calculated as 100 counts. Overall neuron counts and neuron loss were calculated as an even-weighted average of the four neuraxis levels and did not adjust for their different lengths or neuron densities because we were interested in relative changes along the rostral–caudal span.
RESULTS
The particle counts for each nervous system are shown in table 2 and are displayed graphically arranged by age in figure E-1 on the Neurology Web site at www.neurology.org. The calculated overall loss of LMNs in the ALS nervous systems averaged 55%, was normally distributed, and ranged between 8% and 90% (figure 1). In general, nervous systems from younger aged patients had greater degrees of loss (figure E-1). The distribution of LMN loss was usually graded within the nervous system (p = 0.02 by ANOVA; figure 2). In 14 of the 19 ALS nervous systems, loss was greatest at the neuraxis level that corresponded to the region of onset—4 of 6 nervous systems from patients with bulbar onset, 7 of 7 nervous systems from patients with arm onset, 3 of 3 nervous systems from patients with truncal onset, and 0 of 3 nervous systems from patients with leg onset. In 1 nervous system (No. 29), radial degeneration was apparent but did not meet statistical significance. In 2 nervous systems (Nos. 6 and 41), radial degeneration was apparent but loss was greatest in a region different from that identified as the region of onset, suggesting possible misappropriation. In 2 nervous systems (Nos. 34 and 36), LMN loss was minimal and radial degeneration was not seen—both were from patients who had had predominant UMN signs clinically. In all nervous systems, LMN loss averaged 68% in the region of clinical onset and 44% in remoter regions. In nervous systems from onset at the anatomic extremes, bulbar and leg onset patients, LMN loss averaged 54% in the region of onset and 22% in the remotest region. The composite data for each region of onset are presented in table 3 and also show this gradation except in the nervous systems from leg onset patients, where 1 of the 3 seemed misappropriated and 1 other had predominantly UMN involvement.
Table 2 LMN particle counts
Figure 1 Frequency histogram of distribution of overall LMN loss in each ALS nervous system
The total overall LMN loss calculated for each nervous system from the four neuraxis regions (hypoglossal nucleus, cervical anterior horn, thoracic anterior horn, and lumbar anterior horn) is highly variable and normally distributed. LMN = lower motor neuron.
Figure 2 Rostral–caudal distribution of motor neuron degeneration by region of onset
Focal onset and contiguous advancement of motor neuron degeneration frequently creates radial degeneration, the greatest neuronal loss in the region of onset, and progressively less loss in remoter regions. All images are consistently arranged anterior up and right side to the right. The densely staining cells standing out from the background are motor neurons. From the top, patients are 7, 14, 16, 21, and 29. (Cresyl violet acetate; 100×; scale bar is 100 μm.)
Table 3 Composite counts
DISCUSSION
The neuropathologic hallmark of ALS is degeneration and loss of motor neurons.27 While the stages of degeneration leading to loss are chromatolysis, somatodendritic attrition, and apoptosis,28,29 in fact they are not conspicuous and there is generally remarkably little to be seen of a spectrum of morphologic change. Most neuropathologic evaluations have been of LMNs—LMNs are stacked in columns that are arranged rostral to caudal in motor nuclei in the tegmentum of the brainstem for bulbar muscles10,30 and in the anterior horns of the spinal cord for limb muscles.11,13–15,30–33 In spinal anterior horns, three distinct populations of neurons exist: large alpha motor neurons, which have 35- to 100-μm diameters and are located in the anterior portions of the horn; intermediate-sized gamma motor neurons, which have 25- to 35-μm diameters and are distributed among the alpha motor neurons; and small neurons such as interneurons, which have diameters less than 25 μm and are located dorsally and medially in the horn.14,15,34,35
Several morphometric studies have quantified LMN degeneration in ALS at specific regions of the spinal cord—C6,17 C8,36 L3,20 L4,22 and L5.21 Nearly all of the morphometric studies of LMNs have demonstrated significant decreases in the numbers but not diameter sizes of the neurons and little or no dispersion or shift from large to smaller sizes, showing that neuronal atrophy does not occur or is inconspicuous.20–22 Only one study reported a decrease of neuronal area,5 and the explanation for its disparity with the other studies may be that degeneration affects neuron shape.37 All three populations of neurons in the anterior horns, small neurons, and alpha and gamma motor neurons are affected in ALS.20,22,38,39 Neuronal loss may be patchy, focal, and asymmetric and does not seem to affect particular motor cells or columns of cells more than others.36 Not explicitly addressed in any of these morphometric studies but apparent in the data are variability of the degree of neuronal loss—we calculate from the published data that in these studies the average loss of large motor neurons ranged from 51% to 94% and that within each study individual nervous system loss of large sized motor neurons ranged from 23% to 100%.17,20–22,36
Because of the focally progressive nature of motor deficits during the clinical course of the disease,3 underlying motor neuron degeneration should be a focally progressive process and lead to a relatively orderly summation and distribution of loss in the neuraxis. This has not been specifically analyzed, and we therefore pursued this in the LMN compartment. We found that indeed this was the case: LMN loss correlated well with the region of onset and degeneration was radially graded—greatest in the region of onset and decreased away. The true extent of the radial degeneration likely is greater than indicated by our counts. Our counting method was designed to compare between regions of the same nervous system and also to compare between regions of different nervous systems, and to do this in the presence of the marked variability of normal and pathologic anatomy, the true gradation was likely underrated.
Because degeneration advanced outward from the region of onset, summating LMNs loss in the fixed and finite neuronal population sometimes so reduced the numbers that radial gradation became unapparent—in these nervous systems, radial change was often observed in more remote regions. When LMN degeneration began in or near regions controlling respiratory function, death occurred early in the course of degeneration, and there was relatively little overall LMN loss. This is a unique feature of ALS—death occurs during the course of motor neuron degeneration rather than at the end of the degenerative process,40–42 which continues if respiratory function is supported,43 and contrasts with other neurodegenerative diseases, where death almost always occurs after underlying neuropathologic degeneration is essentially completed.
In two nervous systems, LMN loss was minimal and not obviously radially graded. These were from patients who had had predominantly UMN involvement during the clinical phase of the disease and therefore neuropathologic burden at the UMN level. This highlights that both UMN and LMN degeneration are simultaneous and independent at the two levels of the motor system5,6 and that either one alone is sufficient for progressive deterioration. UMNs are spread in laminar sheets that are arranged somatotopically lateral to medial in layer V of the motor cortex.4 They are a heterogeneous population of medium to large pyramidal-shaped neurons, and their most prominent components are the Betz cells, distinguished chiefly by their giant size.4,44–46 Morphometry of the UMN is more difficult than for LMN, but a few studies have been performed,5–8 two of which were not able to demonstrate neuron loss at the motor cortex at all, presumably because of technical aspects of the new stereology techniques.7,8 Morphometry demonstrating a possible lateral-to-medial gradation of neuron loss along the somatotopically organized motor gyri analogous to what we demonstrated for the LMN is difficult if even possible: the oblique orientation of the motor cortex defies traditional coronal sectioning; identification of the motor cortex may be difficult or uncertain in ALS because its chief distinguishing feature, the Betz cell, is or may be reduced or absent; somatotopic localization of body regions is not distinct by histology; the medial to lateral arrangement of sizes and densities of Betz cells along the gyrus is complex; and differentiation of UMNs from other pyramidal neurons in layer V is technically problematic.4,6–8,18
That motor neuron degeneration is a focal and contiguously advancing process that summates over time both within and between the LMN and UMN compartments in the three dimensions of the motor system explains the well-ordered but highly varied LMN loss that occurs both within and between different nervous systems, averaging 55% with a normal distribution between minimal and severe. This creates stages of motor neuron degeneration that in turn could be exploited by new genomic technologies that key directly on relevant neuronal and glial compartments to explore molecular mechanisms of degeneration in ways greater now than previously possible.47–52
Footnotes
-
See also page 1571
Supplemental data at www.neurology.org
Supported by grants from the National Institute of Neurological Diseases and Stroke (NS051738), Juniper Foundation, Moyer Foundation, and Benaroya Foundation.
Disclosure: The authors report no conflicts of interest.
Presented in part at the 58th Annual Meeting of the American Academy of Neurology, San Diego, CA, April 1–8, 2006.
Received December 21, 2005. Accepted in final form January 30, 2007.
REFERENCES
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.