Ubiquitin-positive tau-negative intraneuronal inclusions in dementia with motor neuron disease

The 50th Anniversary of Japanese Society of Neuropathology

Abstract

We first reported ubiquitin-positive tau-negative intraneuronal inclusions in the hippocampal granular cell layer and entorhinal cortices in patients with amyotrophic lateral sclerosis (ALS). We then found that those inclusions occur frequently in patients with presenile dementia and motor neuron disease. The ultrastructure of the inclusions consists mainly of granules with a few filaments. In 2006, TDP-43 was identified as a major component of the inclusions specific for frontotemporal lobar degeneration and ALS. Here, we review the current knowledge regarding ubiquitin-positive tau-negative intraneuronal inclusions.

BACKGROUND

In 1964, Yuasa¹ described a patient with both neurological features typical of amyotrophic lateral sclerosis (ALS) and behavioral and psychiatric symptoms of frontotemporal dementia. However, autopsy findings were not reported. In 1985, Mitsuyama² reviewed the clinicopathological findings of 26 patients with presenile dementia and motor neuron disease (MND) in Japan. Pathologically, there were nonspecific mild degenerative changes throughout the CNS, and he suggested the possibility of a new disease. Thereafter, we used (mainly in Japan) the term “Yuasa–Mitsuyama-type” dementia with MND to describe these patients.³ MND and ALS were used almost synonymously.

At that time, we studied the pathological findings of senile changes in the autopsied brains from 21 patients with sporadic ALS, aged 42–81 years. Paraffin-embedded sections were examined with the Bielschowsky method and by imunohistochemical staining with antibodies directed against β-protein, tau and ubiquitin. We suggested that aged ALS patients accelerate senile plaque formation.⁴ During these studies, we chanced to find ubiquitin-positive tau-negative intracytoplasmic inclusions in the hippocampal granular cells of some patients with sporadic ALS. These inclusions had not been previously reported, and similar inclusions are not found in routinely autopsied brains. Therefore, we studied their morphology and their specificity to ALS.

FIRST REPORT OF THE INCLUSIONS⁵

We studied the brains of 27 patients with clinically and pathologically confirmed sporadic ALS (aged 42–84 years), including one patient with dementia and ALS. Fifty non-ALS patients were also studied. Formalin-fixed and paraffin-embedded sections were stained with HE, KB, Nissl, phosphotungstic acid hematoxylin, Marolly, periodic acid–Schiff, Congo red, Bodian, and the modified Bielschowsky methods. Immunohistochemical studies were also performed using various antibodies, including those directed against ubiquitin, neurofilament, tau, paired helical filament (PHF), β-tubulin, β-protein, α-actin, GFAP and desmin. In seven of the 27 ALS patients, ubiquitin-positive intracytoplasmic inclusions were observed in the neurons of the hippocampal granular cell layers (Fig. 1). The inclusions formed a crescent or circular pattern around the nucleus and were seen in approximately 1–10% of the remaining granular cells. The inclusions were not seen with routine HE staining, nor did they show anilinophilia, argentophilia or congophilia. These seven ALS patients also showed similar inclusions in the small neurons of the second and third layers of the lateral part of the entorhinal cortices. The incidence of the inclusions was almost the same in the granular cell layer and the entorhinal cortex. In one patient who suffered from dementia with ALS, many ubiquitin-positive inclusions were seen in both the hippocampal granular cells and the frontal and temporal cortices. No similar inclusions were seen in the 50 control brains. We first differentiated the inclusions from other known intracytoplasmic inclusions, such as Alzheimer neurofibrillary tangles (NFT) and Pick bodies. They did not stain for tau or PHF, and no argentophilia was observed, which excluded the possibility of NFT and Pick bodies.

Figure 1. Many ubiquitin-positive intraneuronal inclusions in the granular cell layer, ×200.

Because of poor fixation and the relatively small amounts of filamentous material available, it was difficult to demonstrate clearly the fine structure of the ubiquitin-positive inclusions with a conventional electron-microscopic examination. Therefore, we performed an immunoelectron-microscopic examination, using a pre-embedding method with anti-ubiquitin antiserum. Immunoperoxidase products were seen in the cytoplasm of the hippocampal granular cells and in the small neurons of the entorhinal and frontal cortices of the ALS patient with dementia, and loosely arranged lineal filaments and granular material were also observed.

INCLUSIONS IN DEMENTIA PATIENTS WITH ALS⁶

We found no clinical or pathological differences between the seven inclusion-positive ALS patients and the 20 inclusion-negative ALS patients. However, we noticed ubiquitin-positive inclusions in many small neurons in the second layer of the frontal cortex of one patient with a history of dementia. Therefore, we studied the brains and spinal cords of 10 patients with clinically and pathologically confirmed presenile dementia and MND. All 10 patients had ubiquitin-positive tau-negative intracytoplasmic inclusions in the neurons of the hippocampal granular cell layers and in 1–14% of the remaining granular cells. No inclusions were seen in the pyramidal neurons of the hippocampus. The inclusions were frequently observed in the small neurons of the second and third layers of the entorhinal and frontal cortices. These results suggest that ubiquitin-related cytoskeletal abnormalities are common in cerebral non-motor small neurons in these patients.

In the following year, Wightman et al.⁷ confirmed our findings. In 1994, the Lund and Manchester Groups proposed clinical and neuropathological criteria for frontotemporal dementia, dividing it into three subgroups: the frontal lobe degeneration type, the Pick type and the MND type.⁸ The inclusions were described as a neuropathological marker of the MND type, in which “hippocampal dentate gyrus neurons show inclusions that are ubiquitin-positive but not silver or tau reactive”.

In 1998, Neary et al.⁹ proposed a consensus on the clinical diagnostic criteria for frontotemporal lobar degeneration (FTLD). However, FTLD is a heterogeneous entity, and the pathological diagnosis of FTLD includes tau-positive FTLD and tau-negative FTLD.¹⁰ Two variants of tau-negative FTLD are FTLD with and without MND. FTLD with ubiquitin-positive tau-negative neuronal inclusions was grouped as FTLD-U.

FINE STRUCTURE OF THE INCLUSIONS

In 1996, we examined the inclusions using paired routine electron-microscopic ultrathin sections and adjacent semithin sections.¹¹ After the removal of the epon, the semithin sections were stained with anti-ubiquitin antiserum. In the ubiquitin-stained semithin sections, the inclusions formed a crescent or circular pattern around the nucleus (Fig. 2). The adjacent ultrathin sections were examined by electron microscopy, and there was no limiting membrane around the area (Fig. 3). The area seemed to consist of ordinary cytoplasmic organelles, including lipofuscin, mitochondria, cytoplasmic reticulum, and many ribosome-like granules. There were a few filamentous structures. When the findings from immunoelectron-microscopic and semithin sections were compared, the ubiquitin-positive structures seemed to correspond to ribosome-like granules and filaments. The granules were less electron-dense and more irregular, with amorphous outlines, than the ribosomes in the non-ubiquitinated cytoplasm. These findings suggest the development of ribosome-associated and ubiquitin-related abnormalities in the neurons of the extra-motor cortices of these patients.

Figure 2. Ubiquitin-immunostained semithin section of a hippocampal cortical neuron. Lipofuscin granules were negative for ubiquitin, ×400. N, nucleus.

Figure 3. Routine electron-microscopic ultrathin section adjacent to that in Figure 2. The ubiquitin-positive area seems to consist of ordinary cytoplasmic organelles, with a few filaments. Ribosome-like granules are less electron-dense and more irregular, with amorphous outlines, than the ribosomes in the non-ubiquitinated cytoplasmic area, ×3.3 (original). N, nucleus; Lp, lipofucin.

NATURE OF THE INCLUSIONS

The inclusions were positive for ubiquitin-binding protein p62^12,13 and vacuole-creating protein.¹⁴ However, their main components were unknown. In 2006, Neumann et al.¹⁵ and Arai et al.¹⁶ found that the ubiquitin-positive tau-negative inclusions are composed of the 43-kDa TAR DNA-binding protein (TDP-43). Diseases that include TDP-43-positive inclusions have recently been classified as TDP-43 proteinopathy.¹⁷

ACKNOWLEDGMENTS

This work was supported by Grants-in-Aid from the Ministry of Health, Labour and Welfare of Japan, and also from the Ministry of Education, Culture, Sports, Science and Technology of Japan.