Late Onset Dementia with Argyrophrc Grains and Subcortical Tangles or Atypical Progressive Supranuclear palsy? Eliezer Masliah, MD, Lawrence A. Hansen, MD, Sylvia Quijada, BS, Richard DeTeresa, BS, Michael Alford, RA, Joachim Kauss, MD, and Robert Terry, M D

Most clinically demented elderly patients are found at autopsy to have Alzheimer’s disease, multi-infarct dementia, Parkinson’s disease, Picks disease, or Creutzfeldt-Jakob disease. We studied 5 patients clinically characterized by late onset dementia whose brains showed no pathological evidence of Alzheimer’s disease, or any other specific neuropathological diagnosis. We found argyrophilic grains, coiled bodies, abundant Alz-50-positive and thioflavine S-negative neurofibrillary tangles, and neuropil threads in the hippocampus, entorhinal cortex, locus ceruleus, substantia nigra, subthalamic nucleus, and inferior olives. Ultrastructurally, the grains, threads, and tangles were composed of straight tubulofilamentous structures, 25 nm in diameter, similar to those found in patients with progressive supranuclear palsy but different from the paired helical filaments of patients with Alzheimer’s disease. These findings suggest that the late onset dementia with argyrophilic grains syndrome is also characterized by the presence of tangles and threads with the topographical distribution of progressive supranuclear palsy. Masliah E, Hansen LA, Quijada S, DeTeresa R, Alford M, b u s s J, Terry R. Late onset dementia with argyrophilic grains and subcortical tangles or atypical

progressive supranuclear palsy. Ann Neurol 1991;29:389-396

Most clinically demented elderly patients have Alzheimer’s disease (AD), and only a minority are diagnosed as having vascular dementia, Parkinson’s disease (PD) with dementia, Pick’s disease, Creutzfeldt-Jakob disease (CJD), or dementia lacking a distinctive histopathology (DLDH). Still, after excluding all these pathological entities, about 1% of patients with dementia do not fit into any current histopathological category. One possible explanation is that these patients correspond to an early stage of A D in which the preamyloid deposits are not readily identifiable with conventional staining methods and are detectable only with special techniques ( e g , Bielschowsky or anti-amyloid immunohistochemistry) [ 1, 2). We, therefore, studied a group of 5 such patients, at autopsy, with non-AD dementia histopathologically, expecting to find diffuse amyloid deposits. Instead, we encountered argyrophilic grains, coiled bodies, and tangles mostly in the entorhinal cortex. To determine whether these patients fit into the category described by Braak and Braak [3, 41 as dementia with argyrophilic grains, or if another underlying pathological process was present, we studied subcortical areas, which are affected by tangles in various neurodegenerative diseases. We also examined, in

From the Department of Neurosciences M-024, University of Cali-

forriia-San Diego School of Medicine, La Jolla, CA. Received Jun 27, 1990,and in revised form Sep 7 and Sep 20,1990. Accepted for publication Sep 21. 1990.

these patients, neocortical neuronal cell counts, choline acetyltransferase levels (ChAT), immunoreactivities of the lesions with anti-paired helical filament (anti-PHF) antibody, Alz-50, and anti-beta amyloid protein, and the ultrastructural composition of the threads and tangles. Materials and Methods Fifteen patients, at autopsy, were studied, that is, 5 patients with dementia without AD, 5 with dementia with AD, and 5 age-matched control subjects without dementia. Left hemibrains were fixed in 4% formaldehyde and processed as previously described { 5 ] . Paraffin sections of midfrontal (MF), inferior parietal (IP) and superior temporal (ST) neocortex, hippocarnpusientorhinalcortex, basal ganglia, nucleus basalis of Meynert (NbM), cerebellum, mesencephalon, and pons from all patients were stained with hematoxylin and eosin (HE), thioflavine S, and Bielschowsky [l}and Gallyas methods {6, 18, 191. Additional sections of the MF, hippocampus, NbM, thalamus, cerebellar dentate, and inferior olive were immunolabeled with anti-PHF (from Dr D. Selkoe) 171, Alz50 (from D r P. Davies) 181, and anti-amyloid protein (from Dr C. Masters) [2) as previously described [ 5 , 161. Sections of the hippocampus immunostained with Ah-50 were counterstained with thioflavine S. Twenty-micron-thickcresyl vio-

Address correspondence to Dr Masliah, University of California, San Diego, School of Medicine, Department of Neurosciences M-024, La Jolla. CA 92093-0624.

Copyright 0 1991 by the American Neurological Association

389

Table I . Comparison of Brain Weight, h7eocorticalMorphometry. and Neurocbemisty Cell Counts" Group

Agec

Brain Weightd

MF

IP

NAD AD AMC

82 k 11 80 7 87 t 3

1,122 ? 143 1,126 lr 152 1,039 ? 154

394 ? 180 276 2 120 350 k 90

381 177 277

*

-t

k ?

ChATb

194 125 91

ST

MF

IP

ST

440

211 i 85 93 lr 64' 223 lr 20

241 k 149 77 73' 266 f 60

552 39 467

? 121 181 2 93'307 k 107

? f

?

491 14* 151

"Large neurons (> 90 k2)per mm' (mean -+ SD). bCholine acetyltransferase (nmole producedihri 100 mg protein, mean 2 SD). 'Mean (yr) SD. dMean (gm) 5 SD. ' p < 0.02 for AMC. 'p < 0.001.

*

MF AD

= =

midfrontal; IP = inferior parietal; ST = superior temporal; NAD = non-AD dementia with argyrophilic grains and tangles (n Alzheimer's disease dementia (n = 5 ) ; AMC = age-matched controls ( n = 5).

=

5);

Table 2. Microscopic Findings with Silver Sta2n.r and Immunohistochemistty in Diffrrent Areus of the Brain (?f the PatientJ- with Non-AD Dementiu

Neocortex Hippocampus i Entorhinalcortex & Nucleus basalis Subthalamic nucleus Inferior olive Dentate nucleus of the cerebellum

*

-

-

-

-

-

-

-

-

-

?

+ - + + +++ + - + t +++ +++ - +++ +++ +++ + + t +++ +++ - + + - + + - + ++ - + + + + + b NA NA NA ++ ++ ++b N A NA NA ++ ++ +

-

=

Alzheimer's disease; PHF

-

-

=

NA NA - =

absent:

2 =

paired helical filaments; pl

NA few;

=

Results All 5 patients with non-AD dementia were clinically characterized by dementia typical of the AD type lasting approximately 6 years. Additionally, 1 of the 5 patients was diagnosed with parkinsonism refractory to medication and a clinical picture suggestive of PSP, Vol 29 No 4

+

=

April 1991

~

-

mild;

plaques; t

let acetate-stained paraffin sections of neocortex (MF, ST, IP) were used for cell counts with a Cambridge Instruments 970 Quantimet image analyzer (Buffalo, NY) 19- 11). Additional formalin-fixed blocks of hippocampus from patients with non-AD dementia, and from 2 patients with typical AD and 2 with progressive suprmuclear palsy (PSP) were sectioned with an Oxford vibrotome (Technical Products, MI) and immunolabeled with AIz-50, and postfixed in 2% glutaraldehyde followed by 1 osmium tetroxide. Selected areas of the hippocampus were embedded and mounted in epoxy resin, nanoplast (R.Bachhuber, Ulm, Germany), or both, and sectioned with a Reichert OM U3 (Austria) ultramicrotome for ultrastructural examination with a JEOL lO0CX (Tokyo,Japan) electron microscope.

390 Annals of Neurology

-

~

-

Semiquantitative assessment of the lesions: "Fibrillar material. bCell loss and gliosis. AD

-

=

-

-

++

=

moderate;

tangles; g

=

-

-

-

~

-

-

+++

grains; NA

=

=

-

+ - a

-~

a

-

-

_ - _ - _ - _

~

abundant.

not available.

characterized by vertical gaze palsy, truncal ataxia, and absent gag reflex. The macroscopic examination of the brains of the patients with non-AD dementia showed mild atrophy of the frontoparietal region. The basal ganglia and thalamus appeared normal, and the locus ceruleus and substantia nigra displayed a mild decrease in pigmentation. Table 1 summarizes ages, brain weights, cell counts, and ChAT levels of the patients with non-AD dementia with argyrophilic grains and the patients with AD dementia, and the age-matched control subjects. Table 2 summarizes the microscopic observations based o n results of silver staining and immunohistochemistry in the patients with non-AD dementia. Argyrophilic grains were Ah-50 positive and were located in the entorhinal cortex (Fig 1A) and, to a lesser extent, in the NbM and hypothalamus. These areas also contained neurofibrillary tangles and threads. Many threads and tangles were also found in CA2 and CAI of the hippocampus (Fig 1C, D, E). Few argyro-

_

Fig 1 . Alz-50 immunohi.ttochemi.ctry of the hippocampus and entorhinal cortex in patients with non-Alzheimer’s disease (non-AD) dementia. (A)In the entorhinal cortex, abundant Alz-50-positive neurons (Nj, threads, and gpains (arrowheads) werefaand. (Original magnification, x 150.) (B) In the CA2 area of the hippocampus,groups of Alz-50-positive dilated neurites were faund (arrows)awanged in a plaque-likefahion. (Original magnification, x 321 .) (C) Alz-50-positive atypical tangles IT) were seen in C A I . (Original magnification, x 321.I (D} The CAI area of the hippocampus displayed neurofibrillavy tangles (7) of vaqiing morphology. (Original magnification, x 300.) (E) In the presubiculum and subiculum, abundant threads (arrowheads) and curb fbers (arrowheads)were found. (Original magnifcation, x 125.) IF) Pyramidal neurons in the hippocampus of patients with nun-AD dementia showed Alz-50-positive, perinuclear reaction (arrowhead).(Original magnification, X 460.)

philic grains were present in the entorhinal cortex of the patients with AD, and none were seen in the control subjects. Thioflavine S did not stain the grains, threads, or most of the tangles. The Alz-50-positive grains (see Fig 1Aj were also anti-PHF positive. Alz-50 immunolabeled a wide variety of abnormal structures in the hippocampus, entorhinal cortex, and subcortical nuclei of the patients with non-AD dementia. Some neurons displayed a diffuse A1z-50 reaction in the perikaryon as well as in neurites (see Fig lA), but some pyramidal cells presented a clear tangle immunoreactivity (see Fig lC, Dj. The ad-

jacent neuropil showed, in addition to abundant Alz50-positive threads (see Fig l D , Ej and grains (see Fig lA), the presence of plaque-like structures (Fig 1B). These lesions had Alz-50-positive, dilated neurites and did not contain amyloid, as evidenced by the negative thioflavine S stain. A unique finding, in all 5 patients with non-AD dementia, was the presence of a strong perinuclear Alz-50-positive reaction in some pyramidal neurons in CA1 (Fig 1F). This perinuclear staining was found in a small proportion of the pyramidal neurons of the CA1 area. Anti-beta amyloid detected few fibrillar deposits in the entorhinal cortex of the patients with non-AD dementia. In contrast, the patients with AD showed extensive diffuse immature and mature plaques. The age-matched control subjects presented rare diffuse amyloid deposits. The subthalamic nucleus and inferior olive showed neuronal loss and atypical tangle formation (Fig 2B, C) in some neurons. These tangles were thioflavine S negative but Bielschowsky, Alz-50, and anti-PHF positive. A typical patient with PSP (used as a positive control) showed similar tangles in the inferior olive but in greater numbers (Fig 2A). The globus pallidus seemed normal in the HE sections, although the Bielschowsky stain showed few argyrophilic neurons. The dentate nucleus of the cerebellum showed

Masliah er al: Late Onset Dementia

391

Fig 2. (A, B, Cj Bielschowsky silver impregnations of inferior olives. (A)In a patient with classicprogressive supranuclear palsy. most of the neurons showed neurofibrillavy tangles (T). (B, C ) The patients with non-Alzheimer’s disease dementia presented neuronal loss and silver-positive atypical neurofibrillay tangles in some of the olivavy neurons (arrows). (Original magnification, X 285.) fD, E, F ) Alz-50 immunoreactivity of the brainstem. (0)Some pigmented neurons (FN) of the substantia nigra displayed Alz-SO-positive, neurojibrillary tangle formation and pigment loss. (Original magnification, X 440.) (E) Alz-50-positive, neurofibrillavy tangles (T)in the locus ceruleus. (Original magnification, x 440.) i F j Additionally, Alz-50-positively labeled grains (arrowheads)in the neuropil ofthe locus ceruleus. (Original magnification, x 250.1

gliosis and neuronal loss but no tangle formation in the patients with non-AD dementia. All 5 patients with non-AD dementia presented different degrees of neuronal loss and gliosis in the substantia nigra and locus cemleus. Three of these patients had Alz-50-positive tangles in the substantia nigra (Fig 2D), oculomotor nucleus, and locus ceruleus (Fig 2E). The other 2 patients presented Alz-50-positive cells but no tangles. Argyrophilic grains and Alz-50-posirive, fusiform structures were found focally in the neuropil of the substantia nigra and locus ceruleus (Fig 2F). In the patients with non-AD dementia, microscopic examination of the H E sections of the neocortex, locus ceruleus, and substantia nigra did not show Lewy bod392 Annals of Neurology Vol 29 No 4 April 1991

ies. Of the 5 patients with A D and the 2 patients with PSP analyzed, 2 had rare Alz-50-positive grains in the entorhinal cortex. Ultrastructurally, grains (Fig 3A), threads (Fig 3B), and tangles consisted of tubular or filamentous straight structures, 25 nm in diameter (Fig 3C, D), often closely associated with membranes of mitochondria (see Fig 3B, C , D), nuclei, or presynaptic terminals. These structures differed from the PHF (9-11 nm in diameter) of patients with A D (Fig 3E) but were similar to the so-called straight fibers or filaments of patients with

Fig 3. Ultrastructural appearance of the cytoskeleton abnomlities in hippocampus of patient with non-Alzheimer’s disease (non-AD)dementia. (A)Lvw-power view of an argyrophilicgrain (arrows) (original mgnifiration, X 5,000) and (B) o f a threadlike structwe (arrows) composed of abundant long tubules. (Original magni$cation, X 5,000.1 (C) In a pyramidzl neuron, tangles mde of the straight tubules und,filaments were close4 ussociated with membranes. (Original rnagniJcation, x I I .000.) (D) High-power vieto of the straight tubules revealed thick, 25nm structures (arrows)close4 associated with u membrane. (Original magntfication, X 26,000.j (E) Comparative electron micrograph of paired helical filaments in patient with typical AD. (Original magnification, X 26,000.)

Masliah et al: Late Onset Dementia

393

394 Annals of Neurology Vol 29 No 4 April 1991

PSP (17 nm in diameter in our patients), although wider. The epoxy resin-embedded or nanoplast-embedded, Alz-50-immunolabeled sections showed that the immunoreacting material in the tangles (Fig 4A, B), grains (Fig 4C), and threads corresponds to the tubulofilamentous structures (Fig 4D, E). In cross-sections, these tubules had thick, Alz-50-positive walls and narrow lumens (see Fig 4D, E). The Alz-50-immunolabeled filaments of typical tangles of patients with PSP presented similar tubdofilamentous structures (see Fig 4D, E). In some neurons of patients with non-AD dementia, the Ah-50-positive material presented a granular appearance associated with the nuclear membrane.

Discussion In the present study, 5 patients with late onset dementia laclung A D changes at autopsy had abundant argyrophilic grains and many silver-positive but thioflavine S-negative neurofibrillary tangles and neuropil threads in the hippocampus, entorhinal cortex, locus ceruleus, substantia nigra, subthalamic nucleus, and inferior olive. Plaques and tangles were absent from the neocortex. Additionally, neocortical morphometry and neurochemical studies of these patients with non-AD dementia showed normal values compared with control subjects. Identically processed specimens from a group of patients with A D showed characteristic thioflavine S-positive plaques and tangles, neuronal loss, and diminished ChAT levels consistent with previous reports of patients with AD [11-15]. The findings in the patients with non-AD dementia raise the following several diagnostic possibilities: o n e is the entity recently described by Braak and Braak [ 3 , 41 as adult-onset dementia with argyrophilic grains. In favor of this diagnosis is the presence of argyrophilic grains and coiled bodies in different areas of the brain in the absence of amyloid plaques o r any other histopathological changes of AD, PD, or CJD. The morphometric and neurochemical findings in the neocortex also point to the disorder described by Braak and

Fig 4. Ah-50 immunoelectron microscopy. (At Low-power view of an Alz-50-immunolabeled atypical tungle in the hippocampus. u~ (Original magnification, x 3,300.) (B) A higher power ~ i e of the Ah-50-positive tubules in the tangle. (Original magnification, X 10,OOO.j (C) Low-pozuer appearance of so-called argyrophilic grains. (Original mugniJication, X 5,500.) (0)The tabules are composed of a thick wall positively immunolabeled and a narrow lumen-. (Original mgn$cation, X 26,000.) (Ej The nanoplaJt section shows a higher resolution imdge of the tubules. (Origindl m&gn&cution, x 26>000.)(F) Comparative appearance of the Alz-SO-immunolabeled tubule.r in a threadlike (original mgnification, X 10,000) and (G) a tangle in u putient with typical progressive supranuclear palsy. (Original magnification, X 26,000.)

Braak [3, 41. In contrast, the unexpected finding of silver, anti-PHF-positive and Ah-5 0-positive, and thioflavine S-negative tangles and threads in subcortical nuclei and the hippocampus favor a diagnosis of PSP. In fact, 1 of our patients had clinical signs of PSP in addition to dementia. Alternative remote diagnostic possibilities to be considered in these patients are Guam parkinsonism-dementia syndrome and postencephalitic parkinsonism. These differential diagnoses, however, can be ruled out by clinical history and histopathology. The question then occurs whether the patients described by Braak and Braak 131,as well as our 5 patients, represent a combined form of late onset dementia with argyrophilic grains and PSP changes, or an atypical form of PSP. Our findings suggest that some histopathological lesions resembling PSP may present late in life with dementia rather than with gaze palsy and rigidity. Therefore, thioflavine S-negative but silver-positive tangles (presumably not composed of PHFs) should be sought in subcortical nuclei in patients with dementia without A D or any other specific neuropathological changes. Even though the topographical distribution of the tangles suggests PSP, however, there are discordant pathological and clinical features. By light microscopy, the tangles in the patients with non-AD dementia do not have the usual globose appearance typically seen in patients with PSP. Our patients would have to be considered clinically atypical for PSP as well because the presenting symptoms indicated dementia rather than rigidity and gaze palsy. IJltrastructurally, the neuronal neurofibrillary changes, as well as the threads and grains in the patients with non-AD dementia, consisted of thick, tubulofilamentous structures (25 nrn) closely associated with membranes. According to Braak and Braak 141, the coiled bodies and argyrophilic grains of the late onset dementia with argyrophilic grains syndrome are composed of 3-nm filaments and, according to Itagaki and colleagues {131, they are composed of 10- to 13-nm filaments. In patients with PSP, different authors have described a wide variety of straight tubules or filaments as ultrastructural correlates of the tangles. The diameters of these structures range from 15 to 22 nm 120-251. Such ultrastructural characteristics explain the negative thioflavine S staining. The differences in diameters of the tubulofilamentous structures in the patients with late onset dementia reported by others and us could be explained on the basis that all these studies were done in human, suboptimally fixed, postmortem tissue. In this regard, other studies have shown that pathological cytoskeletal elements such as PHF may vary ultrastructurally with fixation and preparatory procedures [26]. The patient with late onset dementia with argyrophilic grains reported by Itagaki { 131 was described as Masliah et al: Late Onset Dementia

395

having Alz-50-positive neurons in cortical and subcortical areas including the locus ceruleus and substantia nigra. In our patients, Ah-50 stained a wide variety of abnormal structures including diffusely positive neurons, tangles, threads, and neurons with Alz-50-positive perinuclear staining only. This last finding could reflect the presence of abnormally phosphorylated tau in the nuclear membranes of pyramidal neurons. In this regard, Ellisman and co-workers [17) have described a close association between the nuclear membrane and PHF in patients with AD. It seems that dementia with argyrophilic grains occurs at least sometimes with associated subcortical tangles and neuropil threads. When associated with the latter lesions, it could be considered an atypical form of PSP. If so considered, however, the clinicopathological spectrum of PSP would have to be expanded to include patients presenting with late onset dementia. A less nosologically traumatic compromise would be to enlarge the category of dementia with argyrophilic grains to include patients both with and without subcortical neurofibrillary tangle pathology in a PSP-like distribution. This work was supported by National Institutes of Health Grants AG08201, AG05131, and NS-07078, the Pew Charitable Trust, and the Alzheimer’s Disease and Related Disorders Association. We thank Dr H. Powell for providing one of the patients for the present study, and Drs D. Selkoe, P. Davies. and C. Masters for kindly providing the antibodies against PHF, A68, and amyloid beta protein, respectively. Thanks also to Mrs Margaret Mdory and Mrs Tanya Albright for their expert technical assistance.

References 1. Yamamoto T. Hirano A. A comparative study of modified Bielschowsky, Bodian and Thioflavin S stains on Alzheimer’s neurofibrillary tangles. Neuropathol Appl Neurobiol 1986;12:3-9 2. Masters CL, Multhaup G, Simms G, et al. Neuronal origin of a cerebral amyloid: Neurofibrillary tangles of Alzheimer’s disease contain the same protein as amyloid plaque cores and blood vessels. EMBO J 1985;4:2757-2763 3. Braak H, Braak E. Argyrophilic grains: characteristic pathology of cerebral cortex in cases of adults onset dementia without Alzheimer changes. Neurosci Lett 1987;76:124-127 4. Braak H, Braak E. Cortical and subcortical argyrophilic grains characterize a disease associated with adult onset dementia. Neuropathol Appl Neurobiol 1989;15:13-26 5. Masliah E, Cole G, Shimohama S, et al. Differential involvement of protein kinase C isozimes in Alzheimer’s disease. Neurosci 1990;1 0 2 113-2 124 5. Gallyas F. Silver staining of Alzheimer’s neurofibrillary changes by means of physical development. Acta Morphol Hung 1971; 19:1-8 7 . Ihara Y , Abraham C, Selkoe DJ. Antibodies to paired helical filaments in Alzheimer’s disease do not recognize normal brain proteins. Nature 1983;304:727-730

396 Annals of Neurology Vol 29 No 4 April 1991

8. Wolozin BL, Pruchnicki A, Dickson DW. Davies PA. A neuronal antigen in the brains of Ahheimer patients. Science 1986;232:648-650 9. Terry RD, Hansen LA. Some morphometric aspects of Alzheimer’s disease and of normal aging. In: Terry RD, ed. Aging and the brain. New York: Raven Press, 1988:109-114 10. Terry RD, DeTeresa R, Hansen LA. Neocortical cell counts in normal human adult aging. Ann Neurol 1987;21:530-539 11. Terry R, Deteresa R, Schechter R,Horoupian DS. Some morphometric aspects of the brain in senile dementia of the Alzheimer type. Ann Neurol 1981;10:184-192 12. Fonnum F. Radiochemical micro assays for the determination of choline acetyltransferase and acetylcholinesterase activities. Biochem J 1969;115:465-472 13. Itagaki S, McGeer PL, Akiyama H, et al. A case of adult-onset dementia with argyrophilic grains. Ann Neurol 1989;26: 685-689 14. Bancher C, Lassmann H , Budka HI et al. Neurofibrillary tangles in Alzheimer’s disease and progressive supranuclear palsy: antigenic similarities and differences. Acta Neuropathol 1987;74: 39-46 15. Hansen LA, DeTeresa R, Davies P, Terry RD. Neocortical morphometry, lesion counts, and choline acetyltransferase levels in the age spectrum of Alzheimer’s disease. Neurology 1988; 38348-54 16. Love S, Saitoh T, Quijada S, et al. Alz-50, ubiquitin and tau immunoreactivity of neurofibrillary tangles, Pick bodies and Lewy bodies. J Neuropathol Exp Neurol 1988;47:393-405 17. Ellisman M, Rangmathan R, Deerinck T, et al. Neuronal fibrillar cytoskeleton and endomembrane system organization in Alzheimer’s disease. In: Perry G, ed. Alterations in the neuronal cytoskeleton in Alzheimer’s disease. New York Plenum, 1987: 61-73 18. Braak H, Braak E. Neuropil threads occur in dendrites of tangle-bearing nerve cells. Neuropathol Appl Neurobiol 1988;14: 39-44 19. Braak H, Braak E, Grundke-Iqbal I, Iqbal K. Occurrence of neuropil threads in the senile human brain and in Alzheimer’s disease: a third location of paired helical filaments outside of neurofibrillary tangles and neuritic plaques. Neurosci Lett 1986;6>:351-355 20. Probst A, Langui D. Lautenschlager C, et al. Progressive supranuclear palsy: extensive neuropil threads in addition to neurofihrillary tangles. Acta Neuropathol 1988;77:61-68 2 1. Montpetit V, Clapin DF, Guberman A. Substructure of 20-nm filaments of progressive supranuclear palsy. Acta Neuropathol 1985;68:31 1-3 18 22. Tornonaga M. Ultrastructure of neurofibrillary tangles in progressive supranuclear palsy. Acta Neuropathol 1977;37: 177-181 23. Bugiani S, Mancardi GL, BNsa A, Ederli A. The fine structure of subcortical neurofibrillary tangles in supranuclear palsy. Acta Neuropathol 1979;45 : 147-152 24. Powell HC, London GW, Lampert PW. Neurofibrillary tangles in progressive supranuclear palsy. Electron microscopy observations. J Neuropathol Exp Neurol 197433198-106 25. Takahashi H, Oyanagi K, Takeda S, et al. Occurrence of 15nm wide straight tubules in neocortical neurons in progressive supranuclear palsy. Acta Neuropathol 1989;79:233-239 26. Papasoaomenos S. Tau protein immunoreactivity in dementia of the Alzheimer type: 11. Electron microscopy and pathogenetic implications. Lab Invest 1989;60:37 5-389