Neuropathology of Speech Network Distinguishes Bulbar From Nonbulbar Amyotrophic Lateral Sclerosis

Abstract

Bulbar amyotrophic lateral sclerosis (ALS) is a debilitating neurodegenerative subtype affecting speech and swallowing motor functions as well as associated with the burden of cognitive deficits. The neuroanatomical underpinnings of bulbar ALS are not well understood. The aim of this study was to compare neuropathology of the speech network (SpN) between 3 cases of bulbar-onset ALS (bALS), 3 cases of spinal-onset ALS (sALS) with antemortem bulbar ALS (sALSwB) against 3 sALS without antemortem bulbar ALS (sALSnoB) and 3 controls. Regional distribution and severity of neuronal loss, TDP-43 (transactive response DNA-binding protein of 43 kDa), and tau proteinopathy were examined. All 3 bALS cases showed marked neuronal loss and severe proteinopathy across most SpN regions; sALSwB cases showed no neuronal loss but mild and variable TDP-43 pathology in focal regions; sALSnoB cases demonstrated an absence of pathology. Two bALS cases had coexisting tauopathy in SpN regions, which was not noted in any sALS cases. The findings suggested that bALS may have a distinct neuropathological signature characterized by marked neuronal loss and polypathology in the SpN. Milder TDP-43 pathology in the SpN for sALSwB cases suggested a link between severity of bulbar ALS and SpN damage. Findings support a clinicopathologic link between bulbar symptoms and pathology in the SpN.

INTRODUCTION

Amyotrophic lateral sclerosis (ALS) is a multisystem neurodegenerative disease encompassing both motor and extramotor systems (1). The disease is fatal and typically fast progressing with a median survival of 3–5 years from disease onset (2). The progressive loss of upper (UMN) and lower motor neurons (LMN) leads to muscle spasticity, hyperreflexia, weakness, and atrophy, leading to eventual paralysis. Extramotor manifestations include behavioral and cognitive-linguistic deficits of frontotemporal dysfunction; 10% present with overt frontotemporal dementia (i.e. ALS-FTD) (3), while up to 50% present with detectable cognitive-linguistic deficits on neuropsychological assessment (4, 5).

Bulbar ALS is a clinically recognized phenotype of ALS affecting speech and swallowing musculature (6). Thirty percent of patients present with bulbar-onset ALS (bALS). The remaining patients present with spinal-onset ALS (sALS), yet, nearly 80% of sALS patients develop bulbar ALS with disease progression (2). Motor speech difficulties, or dysarthria, are the most common bulbar symptoms affecting 93% of patients with bulbar ALS (7). Bulbar ALS is arguably one of the most devastating subtypes of ALS—onset of bulbar signs and symptoms are associated with a faster rate of functional decline resulting in a shorter survival time (i.e. <2 years) (2, 8, 9) and reduced quality of life (10–15). In addition to the devastating motor consequences of bulbar ALS, neuropsychological investigations have linked bulbar symptoms to greater extramotor cognitive-linguistic deficits (16–19), the presence of which may be an independent adverse prognostic factor in ALS (20). The extent of the link between bulbar motor and extramotor changes is not yet known.

Neuropathological examination plays an important role in understanding and diagnosing ALS. Postmortem examination confirms clinical diagnosis, identifies specific subtypes of ALS (21–23), and allows for documentation and staging of coexisting neurodegenerative phenomena. At gross examination, classic ALS shows atrophy of the anterior nerve roots and the precentral gyrus (i.e. primary motor cortex) (24, 25). Microscopically, affected regions show neuronal loss with concomitant gliosis and spongiosis, and the accumulation of intracellular misfolded protein aggregates primarily composed of transactive response DNA-binding protein of 43 kDa (TDP-43) found within the primary motor cortex, anterior horns, and motor cranial nerve nuclei in the surviving neurons (26, 27). Degeneration of descending corticospinal tracts and anterior nerve roots is also prominent (28). Coexisting frontotemporal lobar degeneration (FTLD) is seen in most ALS-FTD cases and in 25%–50% of “pure” ALS cases (29, 30) and is characterized by variable atrophy of frontal and temporal regions with TDP-43 dystrophic neurites (DNs) and neuronal inclusions (FTLD-TDP43) (27). As such, a clinical and neuropathological spectrum between ALS and FTD has been supported, with differences between phenotypes partially explained by pathological variations in the neuronal type (i.e. motor vs extramotor neurons), and the morphology and density of TDP-43 pathology (27, 29). Neuropathological differences in motor and extramotor regions between bulbar and spinal variants have not yet been empirically examined.

The speech neuroanatomy has been well established based on fMRI, electrical stimulation, and lesion studies in humans (31–34). One of the most influential, well-tested, and anatomically mapped computational models of speech production is the DIVA model by Guenther et al (33). The DIVA model maps a feedforward control system that plans and initiates speech movements and a feedback control scheme that guides the movements using somatosensory and auditory feedback onto cortical and subcortical neuroanatomy. The established neural substrates underlying this model include the ventral (oral) portion of the primary motor cortex (bulbar PMC); premotor and supplementary motor areas in the prefrontal cortex related to planning speech sound targets and initiating a chosen movement; ventral somatosensory cortices (i.e. bulbar PSC) receiving real-time somatosensory feedback used to correct motor commands; auditory cortices (i.e. primary acoustic cortex [PAC] in the transverse temporal gyrus) receiving real-time auditory feedback used to correct motor commands; inferior frontal gyrus (IFG) contributing to sequence planning and timing release of planned speech movements (34); posterior superior temporal gyrus (pSTG), an auditory association area integrating auditory feedback and target; and cerebellum and basal ganglia, contributing to precisely timed commands, together forming a large-scale speech network (SpN) (33, 34). The SpN encompasses both motor and extramotor regions, some of which are also involved in cognitive-linguistic processing. Specifically, the IFG and pSTG are known to be involved in mapping sound-to-meaning and in syntactic structuring and processing of sentences (35–38). The high occurrence of speech dysfunction caused by bulbar motor disease and the possible link between bulbar symptoms and cognitive-linguistic deficits suggest that an impairment of SpN in bulbar ALS is highly likely.

Neuropathological regional analyses of speech/language areas have been performed previously on cases with primary progressive aphasia (PPA). PPA is a subtype of FTD, which is considered part of a disease spectrum with clinical, neuropathological, and genetic overlap with ALS (1). The results of these analyses have demonstrated a distribution of neuronal loss and TDP-43 pathology that was clinically concordant with the aphasic phenotypes; namely, leftward asymmetry and greater concentration in language-related regions: the IFG, pSTG, and inferior parietal lobule in logopenic PPA, and IFG and middle frontal gyrus in the nonfluent variant of PPA (39–41). Yet, the extent and distribution of SpN pathology in ALS remain unknown.

Our recent systematic review compared the neuropathology of bALS to sALS and found that the relevant studies were very few (42). Existing studies reported neuronal loss and TDP-43 inclusions in the IFG and pSTG in selected bALS cases but not in any sALS cases (43, 44). Differences in proteinopathy between bALS and sALS were also reported—some bALS cases had basophilic inclusions and/or tau-immunoreactive neurofibrillary tangles (NFTs) in the frontal and temporal lobes of the cortex (43, 45–48). Together, the findings suggested that bALS might be distinct from sALS with neuroanatomic and compositional differences in the underlying pathology. However, the neuropathological studies to date were not designed to compare ALS subtypes, and as such the reported cases were not matched for demographic or disease-related variables. The presence of antemortem bulbar motor impairments, particularly for sALS cases, was not indicated in the majority of studies to allow comparisons between bulbar variants (i.e. bALS vs sALS patients that develop bulbar dysfunction with disease progression). Reports of region-based results were also rare. In addition, existing blocking protocols did not routinely include sampling of the SpN, likely contributing to our poor understanding of the pathogenesis of bulbar symptoms in ALS. There is a need to study hypothesis-specific regions in the brain that may be vulnerable in bulbar ALS.

The main aim of the present study was to further our understanding of the differences between bulbar and nonbulbar variants of ALS by examining postmortem neuropathological changes in SpN regions of the brain. Specific objectives of this case-control study were to compare the neuroanatomic regional distribution, severity, and composition of neuropathology in the SpN regions between bALS cases, sALS cases with antemortem bulbar motor disease (sALSwB), and “pure spinal” sALS cases without antemortem bulbar motor disease (sALSnoB). We hypothesized that the severity and neuroanatomic distribution of pathology in SpN regions would be greater in cases with bALS compared to sALSwB and sALSnoB cases.

MATERIALS AND METHODS

Study Cohort

This study was based on 9 individuals who had a clinical and autopsy-confirmed diagnosis of sporadic adult-onset ALS (49) at the ALS/MND Clinic at Sunnybrook Health Sciences Centre, Toronto, Canada between the years of 1998 and 2019. As of January 2019, the complete autopsy database comprised 86 cases with ALS. Autopsy cases predating 2009 followed a different autopsy sampling protocol and so were excluded to minimize detection bias, remaining 44 cases for inclusion. A single reviewer (author S.S.) who was blinded to the contents of the autopsy reports (e.g. neuropathological diagnosis, presence or absence of coexisting neurodegenerative phenomena, extent and distribution of neuropathology) performed retrospective chart reviews to pseudorandomly select 9 ALS cases for inclusion in the study. Specifically, the first 3 cases in each of the 3 ALS groups that met the a priori inclusion criteria and with accessible archived formalin-fixed cadaveric brain tissue were included in this study. The 3 ALS groups that were investigated were as follows: bALS (n = 3), sALS with reported antemortem bulbar dysfunction sometime during the disease course (i.e. sALSwB) (n = 3), and “pure spinal” sALS cases with reports of intact antemortem bulbar function through the disease course (i.e. sALSnoB) (n = 3). Operational definition of antemortem bulbar dysfunction is described in the section below. In addition, 3 sex- and age-matched control cases without a neurological cause of death were included in this study.

The tissue archives used in this study followed approved procedures for the donation and storage of clinical information in accordance with Sunnybrook Research Ethics Board guidelines and the Declaration of Helsinki.

Clinical and Medical Autopsy Reports

All patients were clinically assessed by a single experienced neuromuscular neuropathologist (author L.Z.) at the ALS/MND clinic at Sunnybrook Health Sciences Centre, Toronto, Canada. Clinical information and demographics were extracted from the patient clinical chart and included age, sex, clinical diagnosis, hand dominance, presence of antemortem bulbar and/or cognitive dysfunction at onset and during the disease course, dates of disease and bulbar symptom onset, and site and laterality of disease onset. Antemortem bulbar dysfunction was operationally defined as any finding of bulbar signs and/or symptoms as follows: neurological examination indicating UMN or LMN signs in orofacial musculature, conducted by a speech-language pathologist; patient-reported changes on the speech, swallowing, or salivation domains on the ALS Functional Rating Scale—Revised (50); or patient reports of dysarthria, dysphagia or related functional changes (e.g. choking, drooling, soft voice, etc.) at any disease stage. Cognitive dysfunction was operationally defined as a score <26 on the last-recorded Montreal Cognitive Assessment (MoCA [51]) or a symptom report of behavioral and/or cognitive changes.

Following the selection of the 9 ALS cases, the neuropathologic autopsy reports for each case were reviewed (by author S.S.), and the final neurodegenerative diagnoses and stages were recorded.

Neuropathological Evaluation

Regions of Interest

A total of 8 blocks were evaluated per case: 6 blocks contained 8 SpN cerebral regions and 2 blocks contained 3 brainstem regions. Cortical SpN regions and the cross-sectional fibers of the pSTG deep white matter (WM) region, corresponding to the arcuate fasciculus/superior longitudinal fasciculus, were chosen for analysis based on previous literature indicating their role in speech control (52–55) and due to the frequent involvement of cerebral changes in ALS patients. Brainstem nuclei that innervate bulbar musculature (i.e. tongue, lips, and jaw) were also assessed in order to characterize the relative contribution of LMN loss to bulbar presentation. Table 1 lists the blocks, corresponding regions, and landmarking details used for tissue sampling. Figure 1 shows the anatomical mapping of the cortical SpN regions that were assessed.

TABLE 1.

A List of the Included Blocks With the Corresponding SpN and Brainstem Regions and Anatomic Sampling Locations

Block #	Region(s) of Interest	Anatomic Sampling Location
Cerebral cortex
1	(1) Bulbar primary motor cortex (PMC)	Pre- and postcentral gyri taken at the level of the occipital horn, 1 slice anterior to calcarine cortex, corresponding to the “oral” region on the motor and sensory homunculi
	(2) Bulbar primary somatosensory cortex (PSC)
2	(3) Supplementary motor area (SMA)	Superior (SFG) and middle frontal gyri (MFG) taken at the level of the hypothalamus
	(4) Premotor area
3	(5) Inferior frontal gyrus (IFG)	Inferior frontal gyrus taken at the level of the caudate head, corresponding to the pars triangularis and/or opercularis
4	(6) Posterior superior temporal gyrus (pSTG)	Posterior superior temporal gyrus taken at the level of the pulvinar
5	(7) pSTG deep white matter (WM)	White matter underlying the pSTG taken at the level of the pulvinar
6	(8) Primary acoustic cortex (PAC)	Transverse temporal gyrus (TT) taken at the level of the subthalamic nucleus, corresponding to Heschl’s gyrus
Brainstem
7	(9) Hypoglossal nucleus (CNXII)	Lower medulla
8	(10) Trigeminal motor nucleus (CNVmo)	Superior pons
	(11) Facial nucleus (CNVII)

FIGURE 1.

Macroimage of the left cortex illustrating the location of the 8 cerebral SpN regions that were sampled in this study. The hypoglossal, trigeminal motor, and facial nuclei in the brainstem were also sampled (not shown). PMC, primary motor cortex; PSC, primary somatosensory cortex; SFG, superior frontal gyrus; SMA, supplementary motor area; MFG, middle frontal gyrus, IFG, inferior frontal gyrus; PAC, primary acoustic cortex; pSTG, posterior superior temporal gyrus; WM, white matter.

Tissue Sampling

Our study cases were subjected to a 2-step tissue sampling methodology. All cases had previously undergone standard neuropathologic diagnostic protocols at the time of autopsy. This included the right hemisphere, hemicerebellum, hemibrainstem, and 3 levels of spinal cord being snap frozen. The remaining left cerebral hemisphere, hemibrainstem, hemicerebellum, and levels of spinal cord were fixed in formalin for at least 2 weeks. After formalin fixation, the left cerebral hemisphere was sectioned into approximately 1-cm-thick coronal sections; the hemibrainstem and spinal cord tissue were sectioned into approximately 3-mm-thick axial sections, and the cerebellum was sectioned sagittally. A standard ALS blocking protocol was applied to the left brain at the time of autopsy which included 23 blocks submitted from the sensorimotor cortex, frontal and temporal cortices, hippocampus, deep gray nuclei, 3 levels of brainstem, cerebellum, and cervical, thoracic and lumbar spinal cord. These tissue blocks were then processed, embedded in paraffin, cut into 7-μm-thick sections, mounted on glass slides, and stained with hematoxylin and eosin/Luxol fast blue (H&E/LFB) and immunohistochemistry for p62, TDP-43, β-amyloid, α-synuclein, and tau (AT8). These slides were examined by an experienced neuropathologist and enabled a neuropathological diagnosis of ALS with or without coexisting FTLD to be made. The presence and staging of coexisting Alzheimer-type neuropathologic change and/or Lewy bodies were also assessed. A diagnostic neuropathologic autopsy report was then issued.

For the purposes of this study, paraffin-embedded blocks created at the time of autopsy were retrieved, cut into additional 7-μm-thick sections, and stained to assess pathology in brainstem nuclei. SpN regions were not blocked at the time of general autopsy; thus, a second step of tissue sampling was required. To this end, the previously dissected and archived formalin-fixed left brain tissue was retrieved and supratentorial regions of interest were identified and blocked using predefined anatomical landmarks as denoted in Table 1. Five of the supratentorial blocks contained 2 adjacent gyri; in order to help with orientation under the microscope, 1 of the 2 gyri was nicked to be noted during tissue sampling. PMC regions were anatomically confirmed through microscopic identification of motor neurons within the tissue section.

In addition, the original stained slides from each case were retrieved. The original tau (AT8), β-amyloid, and α-synuclein stained slides were re-reviewed by an experienced neuropathologist (author J.K.), and the presence of Alzheimer-type neuropathologic change and synucleinopathy were re-assessed and staged.

Stains and Histological Features

The stains assessed for the purposes of this study were H&E/LFB and immunohistochemical stains for p62, TDP-43, and/or tau (AT8) (Supplementary Data Table).

Neuronal loss, pallor of myelin staining within the WM, gliosis, and spongiosis were assessed using H&E/LFB staining. Immunohistochemical stains were used to visualize the burden of phosphorylated tau and TDP-43 immunopathology. Features of neurodegeneration and TDP-43 pathology were chosen for analysis because they represent the classic neuropathologic findings in ALS and FTLD-TDP43 (25, 29). Burden of tau was assessed based on previous findings of tauopathy in bALS cases (42, 56, 57).

All slides were assessed at a multiheader microscope by 2 raters: an experienced neuropathologist (author J.K.) who was blinded to the case ID and clinical diagnosis, and a trainee (author S.S.). The order of cases to be evaluated was chosen at random. Slides were scored semiquantitatively on a 5-point Likert scale with 0 indicating a complete absence of pathology and 3+ indicating severe or marked pathology. The 2 raters achieved consensus on all assessments. Figure 2 shows examples of ratings for TDP-43 scores in the cortex.

FIGURE 2.

Photomicrographs demonstrating semiquantitative ratings of: mild, 1+ (A), moderate, 2+ (B), and severe, 3+ (C) density of TDP-43 pathology in the cortex (shown for MND12: posterior superior temporal gyrus, superior frontal gyrus, and inferior frontal gyrus regions, respectively).

In order to establish intrarater reliability on the ratings, a random subset of cases (n = 4) were evaluated a second time after 6 months. Scores were compared with the initial ratings for each histological feature in each region and showed excellent reliability with an intrarater variance of <3.5%, calculated as the percentage of regions with an altered score. For those with an altered score, the first rating was replaced with the second rating. Furthermore, to ensure consistency in ratings for regions with complicated neuroanatomy (e.g. the trigeminal motor nucleus [CNVmo], arcuate fasciculus fibers without the deep pSTG WM, and transverse temporal gyrus), slides for each region were grouped together across cases and reevaluated a second time. Selected slides were digitally scanned using a Leica Biosystems Aperio AT Turbo model scanner, and digital photomicrographs were taken using Aperio software.

RESULTS

Clinical Descriptors and Diagnostic Neuropathologic Autopsy Summaries

Table 2 summarizes the demographics, clinical information, and neuropathologic autopsy findings for each examined ALS case. All cases were right-handed, indicative of left hemisphere being language dominant. None of the subjects was treated with invasive ventilation, but all subjects had used noninvasive ventilation (i.e. BiPAP) during the course of the illness. Disease durations were calculated from date of reported symptom onset.

TABLE 2.

Summary of Clinical and Diagnostic Neuropathological Data Obtained From the Medical Charts and Autopsy Reports for Each ALS Case

Subgroup	bALS			sALSwB			sALSnoB
Case ID	MND12	MND10	MND11	MND05	MND03	MND19	MND18	MND17	MND08
Sex	M	F	M	M	F	M	M	M	M
Site of onset	Bulbar	Bulbar	Bulbar	UE (bi)	LE (R)	LE (bi)	UE (bi)	UE (NR)	UE (L)
Age at death	73	67	52	73	66	62	63	66	71
Disease duration (months)	7	18	23	30	56	113	18	39	87
Bulbar disease duration (months)	7	18	23	6	27	24	NA	NA	NA
Clinical signs at onset
Bulbar signs	++	++	++	−	−	−	−	−	−
UE signs		+	−	+ (LMN-p)	−	−	++	++ (++head drop)	++ (++head drop)
LE signs	−	−	−	−	++ (R)	++ (LMN-p)	++	+	−
Cognitive dysfunction	−	−	−	−	−	−	−	−	−
Other	−							Flail-arm variant	Flail-arm variant
Clinical signs with progression
Bulbar signs	+++	+++	+++	++	++	+	−	−	−
UE signs	+		++	++	−	+	+++	++	+++
LE signs	+	++	++	+	+++	+++	+++	+	++ (L)
Cognitive dysfunction	−	++	−	−	−	−	−	−	−
Other			(pseudobulbar affect)
Diagnostic neuropathologic findings
ALS	Y	Y	Y	Y	Y	Y	Y	Y	Y
FTLD-TDP43	Y	Y	−	Y	−	−	−	−	−
Braak NFT stage (I–VI)	II	−	VI	−	I	I	I	−	−
CERAD plaque stage (A–C)	B	−	−	−	−	−	B	−	−
Lewy bodies (distribution)	−	−	−	−	−	−	Y (brainstem)	−	−

bALS, bulbar-onset ALS; sALSwB, spinal-onset ALS with antemortem bulbar disease; sALSnoB, spinal-onset ALS without antemortem bulbar disease; R, right side; L, left side; bi, bilateral; NA, not applicable; NR, not reported; M, Male; F, Female; UE, upper extremities; LE, lower extremities; LMN-p, lower motor neuron predominant; Y, yes; Bulbar signs: +, UMN and/or LMN signs in the tongue (fasciculations, needle EMG abnormalities); ++, functional oromotor changes (dysarthria and dysphagia); +++, anarthria; cognitive dysfunction: Montreal Cognitive Assessment (MoCA) scores (51) in addition to symptom reports.

The control cases consisted of 2 males and 1 female with an average age of 63 years (SD = 7). Causes of death were epiglottitis for 1 control subject and myocardial infarction for the other 2.

Neuropathological Findings

Patterns of Neuronal Loss, Gliosis, and Superficial Spongiosis Across Case Groups

Figure 3 shows the anatomical distribution and severity ratings for neuronal loss (in shades of green), gliosis (blue), and superficial spongiosis (pink/purple) for each region and case as assessed on H&E/LFB staining. An absence of pathology is shown as gray cells; missing data are shown as blank cells. Nearly all SpN cerebral regions were severely and widely affected in the 3 bALS cases: Both motor and extramotor SpN regions showed neuronal loss and secondary changes exclusively in the bALS cases. By contrast, the 2 sALS groups had mostly unremarkable SpN cortical regions on routine staining. However, WM pallor underlying the pSTG was noted in at least 1 case per variant. Figure 4 compares 0 pallor (Fig. 4A, C) with 2+ pallor (Fig. 4B, D) in the pSTG deep WM. For brainstem regions—particularly for CNXII—neuronal loss and secondary reactive changes were observed across nearly all cases, including those with purely spinal presentation. None of the 3 control cases had neuronal loss or degenerative phenomena in any of the examined regions.

FIGURE 3.

Severity ratings for neuronal loss (shades of green), gliosis (blue), and superficial cortical spongiosis (pink/purple) for each cerebral SpN and brainstem lower motor neuron region across cases, as seen on the hematoxylin and eosin/Luxol fast blue routine stain. Gray cells indicate absence of pathology. Blank spaces indicate missing or nonapplicable data. bALS, bulbar-onset ALS; sALSwB, spinal-onset ALS with antemortem bulbar dysfunction; sALSnoB, spinal-onset ALS without antemortem bulbar dysfunction; NL, neuronal loss; PMC, primary motor cortex; PSC, primary somatosensory cortex; SFG, superior frontal gyrus; SMA, supplementary motor area; MFG, middle frontal gyrus, IFG, inferior frontal gyrus; pSTG, posterior superior temporal gyrus; WM, white matter; PAC, primary acoustic cortex; CNVmo, trigeminal motor nucleus; CNVII, facial nucleus; CNXII, hypoglossal nucleus.

FIGURE 4.

Photomicrographs of white matter myelin pallor scores of 0 in MND18 (A, scale bar: 3 mm; C, scale bar: 200 μm) and 2+ in MND11 (B, scale bar: 3 mm, D, scale bar: 200 μm) for the white matter underlying posterior superior temporal gyrus, hematoxylin and eosin/Luxol fast blue. The areas with cross-sectional axons corresponding to the arcuate fasciculus were assessed (black box). Black arrows indicate the tail of the caudate nucleus in each slide.

Patterns of TDP-43 and Tau Immunopathology Across Case Groups

Figure 5 shows the anatomic distribution and severity of TDP-43 (in shades of red/orange/yellow) and tau (in shades of blue) proteinopathy for each examined case and region as seen on the TDP-43 and AT8 immunohistochemical stains. All observed TDP-43 and tau-immunopositive inclusions were also immunoreactive for p62, which did not highlight any additional neurodegenerative type inclusions.

FIGURE 5.

Severity ratings of TDP-43 (shades of red/orange/yellow) and tau (shades of blue) proteinopathy for each cerebral SpN and brainstem lower motor neuron region across cases, as seen on the TDP-43 and AT8 immunohistochemical stains. bALS, bulbar-onset ALS; sALSwB, spinal-onset ALS with antemortem bulbar dysfunction; sALSnoB, spinal-onset ALS without antemortem bulbar dysfunction; TDP-43 -NCI, TDP-43-immunoreactive neuronal cytoplasmic inclusions that have “skein-like” or “circumferential” morphology; TDP-43 -DN, TDP-43-immunoreactive dystrophic neurites; NFT, neurofibrillary tangles; PMC, primary motor cortex; PSC, primary somatosensory cortex; SFG, superior frontal gyrus; SMA, supplementary motor area; MFG, middle frontal gyrus, IFG, inferior frontal gyrus; pSTG, posterior superior temporal gyrus; WM, white matter; PAC, primary acoustic cortex; CNVmo, trigeminal motor nucleus; CNVII, facial nucleus; CNXII, hypoglossal nucleus; g, inclusions also found in glial cytoplasm; •, NCIs had a “dot-like” morphology; +, NCIs had a “filamentous” morphology; △, NCIs had a granular morphology; ∼, tau-positive DNs were also seen in superficial (II and III) and deep (>IV) cortical layers; O, TDP-43 inclusions were seen in superficial and deep cortical layers; w, inclusions also seen in subcortical white matter.

TDP-43 pathology

Two of the 3 bALS cases showed severe (up to 3+) and widespread spatial distribution of TDP-43 pathology across SpN cerebral regions. The third bALS case had isolated TDP-43 inclusions in the bulbar PMC only. By contrast, TDP-43 burden was less (<2+) and focal within selected SpN regions in 2 out of 3 sALSwB cases, and absent from the SpN in sALSnoB cases (except for MND08 with 1+ inclusions in the bulbar sensorimotor cortex). Up to 3+ TDP-43 pathology was seen in the brainstem regions, particularly in the hypoglossal nucleus (CNXII), across all cases/variants.

The 2 predominant morphologies of TDP-43 inclusions seen in the cortical and brainstem neurons were “circumferential,” shown in Figure 6A, and “skein-like” neuronal cytoplasmic inclusions, shown in Figure 6B (58). A third common inclusion type was TDP-43-positive DNs, shown in Figure 6D (outlined arrows). One bALS case (MND12) had round, dot-like neuronal cytoplasmic inclusions, shown in Figure 6C, D (filled arrows), as well as granular TDP-43-positive cytoplasmic aggregates, in many of the examined SpN regions. The same case (MND12) also had an isolated TDP-43- and P62-positive glial cytoplasmic inclusions in the pSTG deep WM, shown in Figure 6E.

FIGURE 6.

Photomicrographs of TDP-43 and tau-immunopositive inclusions. “Circumferential” TDP-43 neuronal cytoplasmic inclusions (NCIs) in the inferior frontal gyrus (IFG) of MND10 (A, TDP-43). “Skein-like” TDP-43-positive NCIs in the CNVmo of MND07 (B, TDP-43). “Dot-like” TDP-43- and P62-immunoreactive NCIs in the bulbar PMC of MND12 (C, TDP-43; D, P62, filled arrows). TDP-43 dystrophic neurites (DNs) in the bulbar PMC of MND12 (D, outlined arrow, P62). Isolated TDP-43 and P62-positive glial cytoplasmic inclusion in the pSTG deep white matter of MND12 (E, P62). Tau-positive neurofibrillary tangles (filled arrows) and DNs (outlined arrows) in the IFG of MND11 (F, AT8).

Tau pathology

The 1 bALS case with minimal TDP-43 pathology in the SpN (MND11) had up to 2+ burden of tau-positive NFTs (Fig. 6F, filled arrows), glial cytoplasmic inclusions and DNs (Fig. 6F, outlined arrows) across multiple SpN regions. This case had no neurological symptoms that were suggestive of an FTLD-tau unifying diagnosis. Another bALS case (MND12) that had dot-like and granular TDP-43 inclusions also had an isolated tau inclusion in the pSTG. None of the sALSwB nor sALSnoB cases had tauopathy in any examined regions. None of the 3 control cases had proteinopathy in any of the examined regions.

DISCUSSION

Summary of Findings

This study characterized the neuropathology of the SpN regions in bulbar and nonbulbar clinical variants of ALS and found differences in the distribution and types of proteinopathy between these variants. All 3 bALS cases were characterized by severe neuronal loss and secondary reactive changes and a high density of proteinopathy across majority of SpN cerebral regions. In contrast, none of the 6 sALS cases showed any neuronal loss in SpN regions. However, mild and variable TDP-43 pathology was seen in focal SpN regions for the sALS cases with sALSwB. Further, coexisting tauopathy was noted for 2 bALS cases and none of the sALS cases. The stark neuronal loss and polypathology within the SpN in bALS cases suggested a unique neuropathological profile for bulbar-onset disease. The observed gradation of TDP-43 pathology from bALS, sALSwB to sALSnoB suggested a link between bulbar motor severity and extent of TDP-43 damage across SpN. The findings contribute to our understanding of the neuropathological underpinnings of bulbar ALS and may have significant clinical and research implications.

PMC and Brainstem Neuropathology of ALS Phenotypes

As expected, all 3 bALS cases presented with severe neuronal loss and proteinopathy in the bulbar PMC. This region was relatively spared in the sALS phenotypes, including in sALS cases with antemortem bulbar disease. A single sALSwB case (MND03) with the longest bulbar disease duration as compared to the rest of the group showed 1+ TDP-43 pathology in this region. The 2 remaining sALSwB cases were clinically described as LMN predominant at onset, presumably explaining the lack of UMN loss seen in their bulbar PMCs. Limb subregions on the PMC may show neuropathological changes in these cases, but they were not included in our tissue sampling. ALS cases with confirmed clinical UMN signs but no evidence of neuronal loss in the PMC have also been reported previously (24).

Bulbar PMC was spared in all “pure spinal” sALS cases (i.e. sALSnoB) except for 1 case (MND08) with low density of TDP-43 neuronal cytoplasmic inclusions and DNs that were atypically localized to the deep cortical layers (>IV) and subcortical WM. This case did not appear to differ from the other 2 sALSnoB cases in clinical characteristics; however, descriptions of antemortem bulbar dysfunction were limited to the clinical neurological assessment of bulbar dysfunction and symptom report; more sensitive measures of bulbar disease may have indicated the presence of antemortem bulbar abnormalities (59–61).

Brainstem regions showed comparable neuronal loss and TDP-43 inclusions across all examined cases and phenotypes, including in cases without antemortem bulbar symptoms. This finding is consistent with previous ALS literature indicating a universal vulnerability of brainstem motor nuclei in ALS, specifically of the hypoglossal nucleus innervating tongue musculature and controlling respiration (6, 62, 63).

SpN Neuropathology of ALS Phenotypes

Neuronal Loss and Tau Copathology in SpN Distinguished Bulbar-Onset From Spinal-Onset ALS

The 3 bulbar-onset cases differed from the 6 spinal-onset cases showing moderate-to-severe neuronal loss across all SpN cerebral regions. The stark differences in neuronal density within the SpN between these groups suggested a unique neuropathological signature for bALS, characterized by greater frontotemporal degeneration, which is consistent with previous reports (42, 64). The observed differences in neurodegeneration could not be explained by longer ALS or bulbar disease symptom durations (65, 66). One possible explanation is a longer prodromal phase of bALS, enabling greater cortical atrophy before onset of symptoms. This notion is supported by some large population-based reports showing a more “aggressive” presentation in bALS characterized by a faster rate of overall decline and shorter survival postdiagnosis (2, 8, 9). These observations were not explained by more prominent risk factors posed by greater bulbar disease, as both onset types showed an equal rate of bulbar decline (2, 67).

The SpN areas that showed neuronal loss exclusively in bALS included the secondary motor areas, SMA and premotor cortex; primary sensory areas, the bulbar PSC and PAC; and the IFG and pSTG (33). Previous neuropathological and in vivo structural imaging studies frequently reported degeneration in the prefrontal cortex, pSTG, and IFG in ALS, particularly in bALS (68–71). Structural abnormalities of the PSC and PAC have not been reported, although functional and metabolic changes have been previously shown (72–74). The findings from our study widened the current understanding of the extramotor disease effects in ALS to include neurodegeneration of primary sensory processing areas.

Coexisting tauopathy in frontotemporal SpN regions also differentiated between onset types: 2 of the 3 bALS cases had variable tau NFTs and DNs, while none of the 6 sALS cases showed tauopathy in any of the examined SpN regions. A greater prevalence of NFTs/pretangles in bALS versus sALS has been previously described, although reportedly seen only in the hippocampal/parahippocampal regions (see review in Ref. [42]). Coexisting tauopathy of similar morphologies in frontal areas has been previously highlighted as a component of some ALS cases with mild cognitive impairments (i.e. ALSci subtype) (57), although the presence of bulbar disease in these cases was not reported. The clinical consequence of tauopathy was unclear in our cases as they did not differ considerably in age, motor symptomology, or cognitive status compared to the other bALS case and sALS cases. The finding suggested a clinicopathologic link between bALS and polypathology.

Gradation of TDP-43 Pathology in SpN Between Bulbar Variants Suggests a Link Between Bulbar Motor Disease and Extent of SpN Damage

Although neuronal loss in SpN regions was largely restricted to the bALS cases, TDP-43 pathology was seen in both bulbar variants—bALS and sALSwB. The distribution and severity of TDP-43 pathology differed between variants—a gradation of pathology was observed from bALS cases with widespread and 3+ inclusions to sALSwB cases with a low density of inclusions. Cell-culture studies have demonstrated that TDP-43 aggregation precedes neuronal loss in the neurodegenerative process (75), which suggested that the observed TDP-43 pathology in sALSwB brains corresponded to the early stages of SpN degeneration. The bulbar presentation of these cases was milder than the bALS cases—dysarthria versus anarthria, respectively—suggesting a relation between bulbar motor severity and extent of SpN involvement. This link was further supported by the absence of SpN pathology in cases without bulbar disease (sALSnoB).

Recent neuroimaging and neuropathology studies have demonstrated a network-based pattern of neurodegeneration in ALS and other proteinopathies (76–79). The distribution of pathology appeared to be dependent on the structural connectivity between affected areas, which may partially explain the observed distributional patterns of our sALSwB cases (76, 80, 81). Both cases had TDP-43 pathology in the premotor, SMA, and bulbar PSC—the areas with direct axonal connections to the bulbar PMC (82–84). By contrast, the PAC, IFG, and pSTG, which were spared in most sALSwB cases, do not have direct connections to the PMC, but instead are indirectly connected through the premotor cortex (85). The TDP-43 findings from our study indirectly supported a network-based model of degeneration within the SpN, in which the axonal connections may serve as a conduit for disease propagation across areas with synaptic connectivity (76, 78, 86).

The neuroanatomical areas involved in speech production and language comprehension are not entirely separate, and instead include shared cortical regions that subserve both functions (31). Thus, the relation between SpN damage and bulbar ALS that is proposed in our findings may mechanistically explain the reported link between cognitive-linguistic deficits and bulbar ALS (17, 18, 87). Verbal fluency (18, 88–91), spelling (90, 92), and sentence comprehension deficits (90, 93) have been commonly reported in ALS, which have been mapped to the IFG, pSTG, and PAC (35, 37, 94–96). Neuronal loss and TDP-43 pathology in IFG and/or pSTG have been noted in cases with logopenic PPA and the nonfluent variant of PPA, who present with overt language deficits (39–41). Our findings propose a codevelopment of motor speech (dysarthria) and language dysfunction (aphasia) in ALS which may be mediated by the extent of SpN damage, which in turn informs the need for timely cognitive and language testing at the onset of bulbar symptoms (97, 98).

Study Limitations and Future Directions

The primary limitation of this work has been in the relatively small number of cases examined. Additionally, the tissue sampling was limited to the cerebral SpN regions; serial sampling of the whole brain should be performed to quantify pathology in subcortical SpN structures and to ascertain the extent of pathology beyond the SpN. Microscopic pathology was rated using a semiquantitative ordinal rating scale. Although excellent inter- and intrarater reliability were established, the method relied on subjective ratings of severity and the work should be replicated using digitized methods of quantification. Tissue sampling was conducted from the remnants of previously dissected archived cadaveric brain tissue resulting in missing data points for certain regions and a subtle variance in the level of sampling for the various anatomical regions. Lastly, the nature and severity of bulbar symptoms should be objectively and consistently quantified across all cases in the future works in order to provide more accurate phenotyping of all cases.

Conclusions

Uniquely, this study examined neuropathological differences in SpN regions in clinically defined bulbar and nonbulbar variants. The findings showed notable differences in neuronal loss and polypathology between bulbar-onset and spinal-onset cases. Further, patterns of TDP-43 pathology suggested a link between severity of bulbar motor symptoms and SpN damage across onset types, which may mechanistically explain the previously reported link between bulbar motor disease and cognitive-linguistic dysfunction.

This study was supported by the National Institutes of Health (R01DC013547), the ALS Society of Canada Bernice Ramsay Discovery Grant, and the Kappa Kappa Gamma Foundation of Canada Graduate Scholarship for Women.

The authors have no duality or conflicts of interest to declare.

Supplementary Data can be found at academic.oup.com/jnen.