Neuropathology of primary lateral sclerosis

Published descriptions of the neuropathology of clinically defined primary lateral sclerosis (PLS) are reviewed in order to clarify the pathogenesis and the relationship between PLS and classical amyotrophic lateral sclerosis (ALS). Degeneration of the primary motor cortex and corticospinal tracts with preservation of lower motor neurons (LMN) has been reported in most cases. Studies that employed immunohistochemistry found ubiquitin and/or TDP-43-positive neuronal inclusions in the motor cortex and often in the extramotor neocortex. Ubiquitin/TDP-43-immunoreactive inclusions in LMN have been reported in just over half of cases; however, these have never been numerous. The finding of TDP-43 pathology in most cases indicates that PLS and ALS are closely related conditions; however, the fact that cases of PLS consistently show minimal involvement of LMN suggests that PLS represents a distinct entity, rather than an early stage of ALS.

There have been surprisingly few reports describing the postmortem neuropathology of patients with clinical primary lateral sclerosis (PLS), with many of these being published prior to the development of modern diagnostic techniques. Interpreting the pathological findings and how they might shed light on the relationship between PLS and other neurodegenerative conditions requires an understanding of the neuropathological features of classical amyotrophic lateral sclerosis (ALS), the overlap between ALS and frontotemporal dementia (FTD), and how this knowledge has evolved over the past decades.

The pathology of classical ALS is characterized by chronic degeneration of both upper and lower motor neurons (UMN and LMN, respectively), the associated corticospinal tract (CST) and motor nerves, ultimately leading to neurogenic atrophy of skeletal muscle. Histopathological changes in the primary motor cortex include loss of the large primary motor neurons (Betz cells), associated with nonspecific secondary changes of reactive gliosis and microglial activation. UMN axonal degeneration is often best appreciated by reduced myelin staining in the subcortical white matter and CST. There is always an appreciable loss of LMN from the ventral gray matter of the spinal cord and the affected brainstem motor nuclei, with some of the remaining neurons appearing shrunken, chromatolytic, vacuolated or undergoing neuronophagia. In addition, the remaining LMN usually contain various types of neuronal cytoplasmic inclusion bodies (NCI), some of which are visible with routine histochemical stains, while others require immunohistochemistry (IHC) for detection. Bunina bodies are small, brightly eosinophilic, granular inclusions that have been recognized for many decades as a characteristic pathological feature of ALS. With the growing use of IHC in the late 1980s, the presence of ubiquitin immunoreactive (-ir) NCI in LMN became recognized as an even more sensitive and specific pathological diagnostic feature of ALS ([1],[2]). In 2006, the identity of the ubiquitinated pathological protein that forms these NCI in most cases of ALS was discovered to be the transactive response DNA binding protein Mw 43 (TDP-43); a DNA/RNA binding protein, involved in many aspects of RNA metabolism ([3]). TDP-43-ir NCI in LMN and in the primary motor cortex are now considered to be the defining pathology feature of sporadic ALS and most cases of familial ALS (fALS); the important exceptions being cases of fALS caused by mutations in SOD1 and FUS, in which the cellular inclusions are composed of the respective mutant proteins and not TDP-43 ([4]).

Over the past few decades, there has been a growing recognition that ALS and FTD are closely related conditions with overlapping pathogenesis. At a clinical level, it is now appreciated that a significant proportion of patients who present with ALS go on to develop FTD or more subtle evidence of frontal lobe dysfunction and that patients with FTD often have evidence of pyramidal system dysfunction ([5]). In early 1990s a novel type of pathology was first recognized in ALS patients with dementia; NCI and dystrophic neurites in the extramotor neocortex and dentate granule cells of HC that, at the time, could only be detected with IHC for ubiquitin ([6],[7]). A few years later, various groups began reporting similar ubiquitin-ir cortical pathology in some patients with clinically pure FTD without motor features and this pathology became known as frontotemporal lobar degeneration with ubiquitinated inclusions (FTLD-U) ([8]). It was subsequently appreciated that most cases of FTD that did not have tau-ir pathology had FTLD-U, and that FTLD-U was the most common pathological substrate for FTD ([9]). The 2006 discovery that TDP-43 is the pathological protein in most cases of FTLD-U, as well as ALS, was the strongest evidence to date for a shared pathogenesis ([3]). This overlap has become even more apparent in recent years as several genes have been identified in which mutations may cause either ALS or FTD with TDP-43-ir pathology (including C9orf72, TBK1 and VCP) ([[10]]).

UMN features may be seen in a wide range of neurodegenerative disorders, but are usually a minor component of a more complex neurological syndrome. However, PLS has been reported to be the predominant phenotype in rare cases of common causes of dementia (such as Alzheimer's disease and Lewy body disease) ([13]) and in some conditions that more often present as either atypical parkinsonism or FTD (including progressive supranuclear palsy, neuronal intermediate filament inclusion disease and globular glial tauopathy) ([[14]]). PLS has also been described as a rare clinical presentation of a variety of systemic diseases, including several disorders of the immune system and metabolism ([[17]]). For the remainder of this review, we will focus on cases that fulfill clinical criteria for PLS ([20]) (with or without dementia), that do not meet the pathological criteria for some other established disorder.

It is difficult to glean much from PLS autopsy reports prior to the use of IHC (Table 1). Most describe the expected pathological correlations with loss of Betz cells and chronic degenerative changes in the primary motor cortex and CST, with preservation of LMN ([[20]]). The two cases reported by Younger et al. ([24]) did not show obvious changes in the motor cortex but did have CST degeneration that extended into the upper brainstem and internal capsule. The report by Imai et al. ([25]) did not include a description of the cerebral cortex but did indicate CST degeneration starting at the level of the cerebral peduncles. This case was also said to show a "slight decrease" in the number of spinal cord LMN. Although some of these cases had emotional lability interpreted as pseudobulbar affect, none was reported to have dementia or a family history of neurodegenerative disease.

Table 1 Pathological findings reported in cases of clinical PLS (20).

1 aAbstract only. bSame case as previously reported by Tan et al. (29), but with the addition of immunohistochemical findings.

2 AD: Alzheimer disease; AGD: argyrophilic grain disease; BG: basal ganglia; CN XII: hypoclossal nucleus; CST: corticospinal tract; CVD: cerebrovascular disease; Deg: degeneration; Dem/FTD: dementia not otherwise specified or frontotemporal dementia; Dur: duration; Fam Hx: family history; LMN: lower motor neurons; na: no information available; Poss: possible; Ref: reference; Ubi2q: ubiquitin.

Five subsequent case reports included IHC for ubiquitin only ([[26]]). These also described degeneration of the motor cortex and CST with only one reporting loss of LMN that was restricted to the hypoglossal nucleus ([26]). In three of these cases, IHC was performed on the cerebrum and demonstrated abundant ubiquitin-ir NCI and neurites in the motor cortex ([26],[27],[29]). These same cases also had ubiquitin-ir pathology in the prefrontal and temporal neocortex and the hippocampus, consistent with FTLD-U, and one was thought to have become demented a year prior to death ([29]). IHC was performed on sections of spinal cord in all five cases and ubiquitin-ir NCI were found in LMN in three ([26],[29],[30]). Notably, in all three of these cases the inclusions were reported to be infrequent (described as "few", "some", or "mild"). A similar case report by Konagaya et al. ([31]), that is often cited as an example of PLS neuropathology, also showed widespread cortical ubiquitin-ir pathology and "occasional" inclusions in LMN; however, the disease duration was slightly less than the three year minimum required for a PLS clinical diagnosis.

Somewhat surprisingly, in the decade since the identification of TDP-43 pathology, there have been only a few published autopsy reports of patients with clinically defined PLS. Lu et al. ([32]) described a patient with nine years of PLS and late-onset dementia, with autopsy findings of motor cortex and CST degeneration but preservation of LMN with no Bunina bodies or TDP-43-ir inclusions; however, this case was published as a conference abstract only. The report by Kosaka et al. ([33]) included one new case and a reexamination of the case originally reported by Tan et al. ([29]). Both cases had possible dementia and both showed similar pathological features, with prominent frontotemporal atrophy, abundant TDP-43-ir pathology throughout the cerebral neocortex and hippocampus (e.g., FTLD-TDP), but only a "few" inclusions in LMN, which were preserved in number. Finally, Hirsch-Reinshagen et al. ([34]) recently described a family with a TBK1 mutation, in which two brothers both presented with a combination of PLS and primary progressive aphasia. In both, TDP-43-ir pathology was present throughout the neocortex, limbic cortex and many subcortical regions; however, LMN were well preserved and only a single TDP-43-ir NCI was found in the lumbar spinal cord in one of the two cases (Figure 1).

PHOTO (COLOR): Figure 1 Neuropathological findings in a case of PLS (from ([34])). Primary motor cortex showed chronic degeneration with TDP-43 immunoreactive (-ir) neuronal cytoplasmic inclusions (arrows, A). Corticospinal tracts at the level of the medullary pyramids (*) showing severe reduction of myelin staining (B). Normal numbers of lower motor neurons (LMN) in brainstem and spinal cord (C). Single LMN containing TDP-43-ir cytoplasmic inclusion (D). TDP-43 immunohistochemistry, A and D; Luxol fast blue combined with hematoxylin and eosin (HE), B; HE, C. Scale bar: 55 µm, A; 1600 µm, B; 175 µm, C; 30 µm D.

There have also been several studies that have included cases classified pathologically as PLS, based on the combination of CST degeneration and preservation of the number of LMN, but with insufficient clinical information to determine if any of the patients would have fulfilled clinical PLS criteria ([[35]]). Interpreting the results is further complicated by the selection criteria (e.g., clinical FTD ([36],[37]), presence of FTLD-TDP ([36])) which would have significantly biased the results. The most relevant finding of these studies is confirmation that cases with FTLD-TDP pathology often have CST degeneration, which may or may not be symptomatic, and that these cases may also have TDP-43-ir NCI in LMN without neuronal loss or clinical manifestations.

The number of published PLS pathology case reports is few and may not be fully representative due to various acquisition and reporting biases, emphasizing the need for additional postmortem studies of clinically defined cohorts. None-the-less, the information currently available suggests that most cases of clinical PLS have cortical TDP-43-ir pathology as the substrate for the UMN signs. The finding that approximately half also have some TDP-43-ir NCI in LMN suggests that these cases are closely related to classical ALS. This relationship is further supported by the facts that both ALS and PLS may be associated with FTD and both may be caused by mutations in the same genes ([34],[38],[39]). However, this does not necessarily mean that PLS and ALS represent a disease continuum, with PLS always being a forme fruste of ALS. Reports of ALS patients dying prematurely of other causes ([40]) and autopsy findings in patients who are at genetic risk for ALS ([41]) indicate that LMN pathology is already advanced by the time LMN symptoms present. Therefore, if PLS was simply an early form of ALS, these cases should show a more equal distribution of LMN involvement, with some prodromal cases having near-ALS levels. However, all of the reports of PLS describe either absent or very mild involvement of LMN. This suggests that, PLS and ALS more likely represent distinct points on the MND spectrum, raising the intriguing possibility that cases of PLS may have some genetic or environmental factor that selectively protects against LMN degeneration.

The author has no conflicts of interest to declare.