Delineation of Early Changes in Cases with Progressive Supranuclear Palsy‐Like Pathology. Astrocytes in Striatum are Primary Targets of Tau Phosphorylation and GFAP Oxidation

Abstract

Progressive supranuclear palsy (PSP) is a complex tauopathy usually confirmed at post‐mortem in advanced stages of the disease. Early PSP‐like changes that may outline the course of the disease are not known. Since PSP is not rarely associated with argyrophilic grain disease (AGD) of varible intensity, the present study was focused on AGD cases with associated PSP‐like changes in an attempt to delineate early PSP‐like pathology in this category of cases. Three were typical clinical and pathological PSP. Another case presented with cognitive impairment, abnormal behavior and two falls in the last three months. One case suffered from mild cognitive impairment, and two had no evidence of neurological abnormality. Neuropathological study revealed, in addition to AGD, increased intensity and extent of lesion in three groups of regions, striatum, pallidus/subthalamus and selected nuclei of the brain stem, correlating with neurological impairment. Biochemical studies disclosed oxidative damage in the striatum and amygdala. Together the present observations suggest (i) early PSP‐like lesions in the striatum, followed by the globus pallidus/subthalamus and selected nuclei of the brain stem; (ii) early involvement of neurons and astrocytes, but late appearance of tufted astrocytres; and (iii) oxidative damage of glial acidic protein in the striatum.

INTRODUCTION

Progressive supranuclear palsy (PSP) is a rare neurodegenerative disease, with an age‐adjusted prevalence of 6.4 per 100 000 (43), clinically manifested by movement disorders and cognitive deficits. These include bradykinesia, postural instability with falls (usually backwards), parkinsonism, nuchal dystonia and gaze supranuclear palsy. Frontotemporal impairment and dementia can follow the appearance of movement disorders 20, 39).

Clinical diagnosis of PSP is based on both inclusion and exclusion criteria of other neurodegenerative diseases with similar clinical features as Parkinson's disease (PD), corticobasal degeneration, multiple system atrophy and Lewy body disease 28, 29). Progressive Supranuclear Palsy Rating Scale and Staging system permits an approach to the clinical staging and prognosis of the disease (19).

The main neuropathological findings in PSP are neuronal loss, astrocytic gliosis, and hyperphosphorylated‐tau accumulation in neurons, astrocytes and oligodendrocytes. Cortical and subcortical structures are affected at the terminal stages of the disease. The globus pallidus, striatum, subthalamic nucleus, nucleus basalis of Meynert, colliculi, tegmentum, periaqueductal gray matter, substantia nigra, red nucleus, reticular formation of the midbrain and pons, basis pontis, and dentate nucleus are involved in the majority of cases; the cerebral cortex, locus ceruleus, oculomotor complex and inferior olive are variably affected 21, 22, 29). Tau‐positive aggregates in PSP are neurofibrillary tangles (NFT) and pretangles in neurons, neuropil threads, coiled bodies in oligodendroglial cells, and heterogeneous inclusions in the cytoplasm of astrocytes that are subdivided into thorn‐shaped or bush astrocytes and protoplasmic tufted astrocytes 4, 26, 46, 47). Hyperphosphorylated tau in PSP is mainly composed of 4R‐tau isoforms that are resolved in two phospho‐tau bands of 68 kDa and 64 kDa in Western blots of sarkosyl‐insoluble fractions 11, 14).

In contrast with other degenerative diseases of the nervous system, as Alzheimer's disease (AD) or PD, practically nothing is known about early changes in PSP. This is because, in part, of the rarity of PSP in comparison with AD and PD. Yet PSP is often associated with argyrophilic grain disease (AGD) 31, 49), another tauopathy much more common than PSP. AGD was first described as a degenerative disease characterized by argyrophilic grains in the entorhinal cortex, hippocampus, amygdala and neighbouring temporal cortex in a subset of patients who had suffered from adult onset dementia 7, 8). However, AGD is frequently asymptomatic depending on the extension of the lesions 42, 52); the presence of AGD has been estimated at 5%–9% in adult autopsy series 9, 50). The main neuropathological findings in AGD are tau hyperphosphorylation and accumulation in grains, pretangle neurons, tangles, neuropil threads, coiled bodies, and cytoplasm of astrocytes. Gel electrophoresis of sarkosyl‐insoluble fractions has shown that AGD is characterized by a double band of 68 kDa and 64 kDa 16, 18, 49, 53, 54). The use of specific anti‐4R tau antibodies has further categorized AGD as a 4R‐tauopathy (49).

The neuropathological screening of AGD is relatively easy with the use of phospho‐tau immunohistochemistry in a single section of the anterior hippocampus, and this method is routinely carried out in many laboratories because it also permits the screening of AD. Following this procedure, we have found 45 AGD cases, which represent 4% of the total autopsies, in a consecutive autopsy series in a general adult hospital. Seven cases had additional lesions of mild, moderate or severe intensity that resembled those seen in PSP. Three of them had suffered from typical clinical symptoms of PSP for 4–6 years—the neuropathological changes were also typical of terminal PSP—and these cases were diagnosed as PSP with associated AGD. Another case presented with cognitive impairment, abnormal behavior and two falls in the last 3 months. One case had suffered from mild cognitive impairment with no motor abnormalities. The remaining two cases had no evidence of neurological deficits according to the clinical report and the neurological examination.

The present study is focused on the neuropathology of the four cases with AGD and associated lesions that were evaluated as consistent with early PSP‐like changes. It is worth considering that in spite of certain similarity of lesions in these cases to those encountered in advanced PSP, we will never know whether these individuals would have suffered from clinical PSP if they had survived for a long time. Therefore, the present study is geared to increase understanding about possible steps of PSP‐like pathology in certain subjects rather than to establish a rigid and universal scheme of PSP stages of disease progression. In addition to the neuropathological study, gel electrophoresis and western blotting to glycoxidative and lipoxidative markers, as well as 2D gel electrophoresis and mass spectrometry, have been carried out to evaluate whether oxidative damage is an early event associated with PSP‐like pathology.

MATERIAL AND METHODS

Cases

A summary of the cases studied is shown in Table 1.

Table 1.

Summary of the clinical and neuropathological diagnosis in the present series. Abbreviations: MCI = mild cognitive impairment; PSP = progressive supranuclear palsy; AGD = argyrophilic grain disease; NFTs = neurofibrillary tangles; EC = entorhinal cortex; p.‐m. delay = post‐mortem delay (h).

	Age	Gender	p.‐m. delay	Neurological diagnosis	Neuropathological diagnosis
1	68	Man	12 h	Normal	AGD + tauopathy
2	79	Man	4 h	Normal	AGD + tauopathy
3	66	Woman	6 h	MCI	AGD + tauopathy
4	68	Woman	16 h	MCI + falls	AGD + PSP‐like
5	75	Man	18 h	PSP	PSP + AGD
6	54	Man	3 h	Normal	No lesions
7	35	Man	8 h	Normal	No lesions
8	82	Woman	11 h	Normal	A few diffuse plaques
9	75	Woman	6 h	Normal	No lesions
10	80	Woman	3.5 h	Normal	A few diffuse amyloid plaques; NFTs in EC
11	73	Woman	7 h	Normal	No lesions
12	58	Woman	4 h	Normal	No lesions

Case 1: The patient was a 68 years old man who was admitted in the hospital because of fever, respiratory insufficiency, pancytopenia, lung infiltrates and positive cultures to pseudomona. He was diagnosed of chronic lymphocytic leukemia, bilateral pneumonia, septic shock and multiorganic failure. He died 7 days later. No clinical evidence of neurological disorder was recorded in the clinical file. An interview to the relatives after the neuropathological diagnosis was made further excluded major neurological deficits, including abnormal behavior and cognitive impairment.

Case 2: The patient was a 79 years old man with chronic respiratory insufficiency currently visited in the hospital. No evidence of neurological symptoms and signs was recorded in the clinical files. He was admitted in the hospital because of sudden respiratory failure caused by a lobar pneumonia. He died 48 h later.

Case 3: The patient was a 66 years old woman with loss of recent memory and cognitive decline during the last 2 years. The last CT examination carried out 3 months before her admittance in the hospital showed mild global cerebral atrophy. At the same time, the neurological examination did not reveal motor or sensory disturbances, gait disorders or motor ocular anomalies. The patient was admitted because of sudden respiratory failure caused by pulmonary thromboembolism and she died 4 h later.

Case 4: The patient was a 68 years old woman with progressive cognitive decline, abnormal behavior with irritability and emotional instability for the last 4 years, clinically categorized as possible Alzheimer's disease. Two falls were recorded in the last 3 months. The patient also suffered from arterial hypertension and renal failure under clinical study. The neurological examination 1 month before her admittance to the hospital revealed slight tremor and a discrete disorder of the gait; the motor ocular movements were not affected. The patient came to the hospital because of urinary infection and sepsis.

Case 5: The patient was a 75 years old man with progressive cognitive beginning at the age of 70 and characterized by loss of memory, impaired orientation and progressive aphasia, personality changes with apathy, frequent changes of mood, irritability, and insomnia. He also suffered from the last 3 years of frequent falls, tremor, abnormal gait, facial hypomimia and motor ocular disturbances, including slowness of saccadic movements and difficulties to look upward. The patient was admitted in the hospital because of respiratory insufficiency caused by lobar pneumonia.

Cases 6–12: No neurological abnormalities were recorded in these patients. Causes of death were respiratory infections (3), cardiac infarction (1), disseminated cancer (2) and pulmonary thromboembolism (1).

Neuropathological methods

The time between death and tissue processing was between 2 and 14 h.

The left hemisphere was immediately cut on coronal sections, 1‐cm thick, frozen on dry ice and stored at −80°C until use. For morphological examination, the brains were fixed by immersion in 10% buffered formalin for 2 or 3 weeks. The neuropathological study was carried out on dewaxed 4‐µm thick paraffin sections of the frontal (area 8), primary motor, primary sensory, parietal, temporal superior, temporal inferior, anterior gyrus cinguli, anterior insular, and primary and associative visual cortices, entorhinal cortex and hippocampus, caudate, putamen and globus pallidus, medial and posterior thalamus, subthalamus, Meynert nucleus, amygdala, midbrain (two levels), pons and bulb, and cerebellar cortex and dentate nucleus. The sections were stained with haematoxylin and eosin, Klüver Barrera, and, for immunohistochemistry to glial fibrillary acidic protein (Dako, Barcelona, Spain dilution 1:250), CD68 (Dako, dilution 1:100), βA‐amyloid (Boehringer, Ingelheim, Germany, dilution 1:50), tau AT8 (Innogenetics, Gent, Belgium, dilution 1:500), tau 4R and tau 3R (Upstate, Gent, Belgium, dilution 1:200 and 1:2 000, respectively), phosphorylation‐specific tau Thr181, Ser202, Ser214, Ser262, Ser396 and Ser422 (all of them Calbiochem, LaJoya, CA, USA, dilution 1:100, except Thr181 1:250), and αB‐crystallin (Abcam, Cambridge, UK, dilution 1:100), α‐synuclein (Chemicon, Millipore, MA, USA, dilution 1:500) and ubiquitin (Dako, dilution 1:200). AD stages were established according to the amyloid deposition burden and neurofibrillary pathology following the nomenclature of Braak and Braak (10). Stages of amyloid deposition refer to initial deposits in the basal neocortex (stage A), deposits extended to the association areas of the neocortex (stage B), and heavy deposition throughout the entire cortex (stage C). Stages of neurofibrillary pathology correspond to transentorhinal (I–II), limbic (III–IV) and neocortical (V and VI). AGD stages were established following the nomenclature of Saito et al (42), slightly modified (18). Stage I is characterized by mild involvement of the anterior entorhinal cortex, cortical and basolateral nuclei of the amygdala, and hypothalamic lateral tuberal nucleus. Stage II is defined by moderate involvement of the entorhinal cortex, anterior CA1, transentorhinal cortex, cortical and basolateral nuclei of the amygdala, presubiculum, hypothalamic lateral tuberal nucleus and dentate gyrus. Stage III involves the entorhinal cortex, CA1, perirhinal cortex, presubiculum, amygdala, dentate gyrus, hypothalamic lateral tuberal nucleus. In addition, there is mild involvement of CA2 and CA3, subiculum, other nuclei of the hypothalamus (ie, mammillary bodies), anterior temporal cortex, insular cortex, anterior gyrus cinguli, orbitofrontal cortex, nucleus accumbens and septal nuclei.

Sarkosyl‐insoluble fraction extraction

Frozen samples of about 2 g of the amygdala and striatum were gently homogenized in a glass tissue grinder in 10 vol (w/v) with cold suspension buffer (10 mM Trishydroxymethylaminomethane (TRIS‐HCl), pH 7.4, 0.8 M NaCl, 1 mM Ethylene‐glycol tetraacetic acid (EGTA), 10% sucrose). The homogenates were first centrifuged at 20 000 g, and the supernatant (S1) was retained. The pellet was rehomogenized in 5 vol of homogenization buffer and recentrifuged. The two supernatants (S1 + S2) were then mixed and incubated with 0.1% N‐lauroylsarcosynate (sarkosyl) for 1 h at room temperature while being shaken. Samples were then centrifuged at 100 000 g in a Ti70 Beckman rotor. Sarkosyl‐insoluble pellets (P3) were resuspended (0.2 mL/g, starting material) in 50 mM TRIS‐HCl (pH 7.4). Protein concentrations were determined with the Bradford method using bovine serum albumin (BSA) as a standard.

Brain homogenates

Brain samples (0.1 g) of the striatum and amygdala of cases 2, 3 and 5, and controls 6, 7 and 8 were homogenized in 1 mL of lysis buffer (40 mM Tris, pH 7.5, 7 M urea, 2 M thiourea, 4% 3‐[(3‐Cholamidopropyl)dimethylammonia]‐1‐propanesulphonate (CHAPS) (BioRad, Barcelona, Spain) and complete protease inhibitor cocktail (Roche Molecular Systems, Barcelona, Spain), and centrifuged at 15 000 rpm for 10 minutes at 4°C. The pellets were discarded and protein concentrations of the supernatants were determined by Bradford method with bovine serum albumin (BSA) (Sigma, Barcelona, Spain) as a standard.

Western blot

For monodimensional gel electrophoresis, 20 µg of each sample from total homogenates, 250 ug from sarkosyl‐insoluble fraction of case 2 and 150 ug from the same fraction of case 3 and case 5 were loaded for 10% SDS‐polyacrylamide gel electrophoresis (SDS‐PAGE) electrophoresis and then transferred to nitrocellulose membranes (400 mA for 90 minutes). Mini‐protean system (BioRad) was used for brain homogenates, and maxi‐protean system (16 × 20 cm) was used for sarkosyl‐insoluble fractions. Immediately afterwards, the membranes were incubated with 5% skimmed milk in tris buffered saline Tween 20 (TBS‐T) buffer (100 mM Tris‐buffered saline, 140 mM NaCl and 0.1% Tween 20, pH 7.4) for 30 minutes at room temperature, and then incubated with the primary antibody in TBS‐T containing 3% BSA (Sigma, Madrid, Spain) at 4°C overnight. The mouse monoclonal anti‐carboxy‐ethyl‐lysine (CEL) and anti‐carboxy‐methyl‐lysine (CML) (TransGenic, Kumamoto, Japan), the mouse‐monoclonal anti‐advanced glycation end products (AGE) (TransGenic), and the rabbit polyclonal anti‐malondyaldehyde‐lysine (MDA‐L) (Biomedical, Houston, TX, USA) antibodies were used diluted 1:1000. The rabbit polyclonal anti‐glial fibrillary acidic protein (GFAP) (Dako) was used at a dilution of 1:4000. The monoclonal antibody to β‐actin (Sigma) was used at a dilution of 1:30 000 as a control of protein loading. In sarkosyl‐insoluble fractions, rabbit polyclonal antiphospho‐tau(Ser422) (Calbiochem) was used at a dilution of 1:1000. Subsequently, the membranes were incubated for 45 minutes at room temperature with the corresponding secondary antibody labeled with horseradish peroxidase (Dako) at a dilution of 1:1000, and washed with TBS‐T for 30 minutes. Protein bands were visualized with the chemiluminescence ECL method (Amersham, Barcelona, Spain).

Control and diseased cases were processed in parallel.

2D gel electrophoresis

A 150‐µg protein was mixed with 2% Byolites (v/v), 2 mM tributylphosphine solution and bromophenol blue in a final volume of 150 µL. In the first‐dimension electrophoresis, 150 µL of sample solution was applied to an immobilized 7‐cm pH 3–10 nonlinear gradient ReadyStrip immobilized pH gradient (IPG) strip (Bio‐Rad) at both the basic and acidic ends of the strip. The strips were actively rehydrated for 12 h at 50 V, and the proteins were focused at 300 V for 1 h, after which time the voltage was gradually increased to 3500 V within 6 h. Focusing was continued at 3500 V for 12 h and at 5000 V for 24 h. For the second dimension separation, IPG strips were equilibrated for 10 minutes in 50 mM Tris‐HCl (pH 6.8) containing 6 M urea, 1% (wt/v) SDS, 30% (v/v) glycerol and 2% dithiotreitol, and then reequilibrated for 10 minutes in the same buffer containing 2.5% iodacetamide. The strips were placed on 10% polyacrylamide gels and electrophoresed at 0.02 A. For gel staining, Comassie Biosave staining (Biorad) was used as described by the manufacturer.

Two 2D gels for every case were run in parallel, one for Comassie staining and the other transferred to a nitrocellulose membrane (200 mA for 1 h 30 minutes). Diseased cases were processed in parallel with control cases. After incubation with 5% skimmed milk in TBS‐T buffer for 30 minutes at room temperature, nitrocellulose membranes were blotted with mouse monoclonal anti‐AGE (TransGenic). Membranes were stripped by two incubations of 20 minutes at 64°C with stripping buffer (0,1 mM B‐mecap, 2% SDS, 62.5 mm Tris HCL pH 6.8) and incubated with 5% skimmed milk in TBS‐T buffer for 30 minutes at room temperature. Membranes were then incubated with rabbit polyclonal anti‐GFAP (Dako) used at a dilution of 1:4000. Subsequently, the membranes were processed as previously indicated for monodimensional gels. Several combinations of disease cases with at least three control cases were performed in order to prove reproducibility.

In‐gel digestion

Proteins were in‐gel digested with trypsin (sequencing grade modified, Promega, Barcelona, Spain) in the automatic Investigator^TM ProGest robot of Genomic Solutions, Michigan, USA. Briefly, excised gel spots were washed sequentially with ammonium bicarbonate buffer and acetonitrile. Proteins were reduced and alkylated with 10 mM DTT solution for 30 minutes and 100 mM solution of iodine acetamide for 15 minutes, respectively. After sequential washing with buffer and acetonitrile, proteins were digested overnight at 37°C with trypsin 0.27 nM. Tryptic peptides were extracted from the gel matrix with 10% formic acid and acetonitrile. The extracts were pooled and dried in a vacuum centrifuge.

Acquisition of Mass spectrometry (MS) and MS/MS spectra

Proteins manually excised from the 2D gels were digested and analyzed by CapLCnano‐ESI‐MS‐MS mass spectrometry. The tryptic digested peptide samples were analyzed using on‐line liquid chromatography (CapLC, Micromass‐Waters, Manchester, UK) coupled with tandem mass spectrometry (Q‐TOF Global, Micromass‐Waters, Manchester, UK). Samples were resuspended in 12‐µL 10% formic acid solution, and 4 µL was injected for chromatographic separation into a reverse‐phase capillary C18 column [75‐µm internal diameter and 15 cm in length (PepMap^TM column, LC Packings, Amsterdam, The Netherlands)]. The eluted peptides were ionized via coated nano‐ES needles (PicoTip^TM, New Objective, Woburn, MA, USA). A capillary voltage of 1800–2200 V was applied together with a cone voltage of 80 V. The collision in the collision‐induced dissociation was 25 to 35 eV, and argon was employed as the collision gas. Data were generated in PKL file format and submitted for database searching in Mascot server (Matrix Science, Boston, MA, USA) using the NCBI database with the following parameters: trypsin enzyme, one missed cleavage, carbamidomethyl (C) as fixed modification and oxidized (M) as variable modification, and mass tolerance of 150–250 ppm.

Probability‐based Mowse score was used to determine the level of confidence in the identification of specific isoforms from the mass spectra. This probability equals 10 (‐Mowse score/10). Mowse scores higher than 50 were considered to be of high confidence of identification.

RESULTS

Neuropathology

Representative lesions are shown in Figure 1. Characteristic lesions in individual cases 1, 2, 3 and 5 are shown in Figure 1S–4S (supplementary data).

Figure 1.

Phospho‐tau pathology in cases with AGD and PSP‐like pathology. A. caudate, case 2. B. amygdala, case 3. C. globus pallidus, case 5. D. CA1, case 3. E. subthalamus, case 5. F. superior colliculus, case 5. G. ventral pons, case 3. H. midbrain, case 3. I. locus coeruleus, case 3. J. gyrus cinguli, case 3. Paraffin section slightly counterstained with haematoxylin. Abbreviations: AGD = argyrophilic grain disease; PSP = progressive supranuclear palsy.

AGD pathology was characterized by the presence of grains and pretangles in the entorhinal cortex, CA1 region of the hippocampus and amygdala in every case, although the intensity of lesions was variable depending on the presence of associated lesions. The amygdala was slightly damaged in cases 1 and 2, but it was severely affected in cases 3, 4 and 5. Grains and pretangles, together with deposits in dentate gyrus neurons, were also seen in the temporal cortex and subiculum in cases 3, 4 and 5. Astrocytic inclusions and coiled bodies were noticed in every case, although with variable intensity.

On the basis of neuropathological data, staging of AGD was established as follows: case 1: early stage 2, case 2: stage 2, and cases 3, 4 and 5: stage 3.

Neurofibrillary tangles in the same regions were categorized as stage I in cases 1, 2 and 4; and stage II in case 5. A few diffuse plaques were seen in the temporal and orbital cortices in case 3 (stage A of Braak and Braak).

PSP‐like pathology was examined in three different groups of regions: caudate/putamen. globus pallidus/sbthalamic nucleus, and brain stem nuclei (substantia nigra, locus ceruleus, colliculi, periaqueductal gray matter and ventral pons).

Lesions in the caudate/putamen were characterized by astrocytic gliosis, as revealed with GFAP‐immunostaining, and gel electrophoresis and GFAP immunoblotting (Figure 2), phospho‐tau‐immunoreactive astrocytes, and phospho‐tau‐immunoreactive neurons. These lesions were present in every case, although with variable intensity. The intensity of lesions was mild in case 1 and 2, moderate in cases 3 and 4, and severe in case 5.

Figure 2.

Monodimensional gel electrophoresis and western blotting to GFAP in the amygdala and striatum in cases 2, 3 and 5, and age‐matched controls. Several bands of high molecular weight and strong density (arrows) are seen in diseased cases when compared with controls. Abbreviation: GFAP = glial fibrillary acidic protein.

Lesions in the globus pallidus and subthalamic nuclei were scanty in cases 1, 2 and 3, moderate in case 4 and severe in case 5. Neuronal pathology rather than astrocytic pathology predominated in these regions.

Lesions in the brain stem were characterized by neurofibrillary tangles and phospho‐tau‐immunoreactive inclusions in astrocytes. Discrete lesions were restricted to the ventral pons and ceruleus in cases 1 and 2. The intensity of lesions increased in these nuclei and the distribution of lesions extended to the substantia nigra, colliculus, red nucleus and peri‐aqueductal gray matter in case 3 and particularly in case 4. Severe phospho‐tau deposition in neurons and astrocytes was found in case 5.

Interestingly, in spite of the substantial numbers of phospho‐tau‐immunoreactive astrocytes, tufted astrocytes were absent in case 1, exceptional or very rare in cases 2 and 3, occasional in case 4 and frequent in case 5.

Coiled bodies in these regions were absent in cases 1 and 2, they were mild in case 3, moderate in case 4 and severe in case 5. These lesions were stained with anti‐4R‐tau antibodies. 3R‐tau immunoreactivity was restricted to small numbers of tangles in the entorhinal cortex, thus suggesting combined stage I–II AD (data not shown).

Mild involvement of the Meynert nucleus occurred in cases 1 and 2, moderate in cases 3 and 4, and severe in case 5. The frontal cortex was involved in cases 3, 4 and 5, the parietal cortex was additionally involved in case 4, and the occipital cortex in case 5.

A summary of neuropathological observations is shown in Table 2, whereas individual changes in cases 1–5 are shown in the corresponding Tables 1S–5S (supplementary data).

Table 2.

Summary of main neuropathological findings related with tau pathology. Involvement of three combined regions, presence of tufted astrocytes and AGD changes are expressed semiquantitative ly. +: mild; ++: moderate; +++: severe; +/− indicates lack of involvement of one of the mentioned regions. Abbreviation: AGD = argyrophilic grain disease.

Cases/region affected	1	2	3	4	5
Caudate/putamen	+	+	+	+	++
Globus pallidus/subthalamus	+/−	+/−	++/+	+	++
Substantia nigra/coeruleus/colliculi/ventral pons	+/−	+/−	+	++/+	+++/++
Tufted astrocytes	no	exceptional	very rare	occasional	very common
AGD stage	2	2	3	3	3

No lesions were seen in cases 6–12. More explicitly, tau, α‐synuclein and ubiquitin inclusions, and β‐amyloid plaques were absent, excepting a few diffuse plaques in the orbital and temporal cortex in two cases and a few neurofibrillary tangles in the entorhinal cortex in one.

Tau banding pattern in sarkosyl‐insoluble fraction

Western blots of sarkosyl‐insoluble fractions incubated with antiphospho‐tau(Ser422) antibody revealed a double band of 68 kDa and 64 kDa in the striatum and amygdala in case 3 and in case 5 (established PSP). In addition to these two bands, other bands of about 50 kDa and lower were detected in the striatum and amygdala (Figure 3). Similar bands representing full‐length 4R tau and truncated forms of tau have also been described in PSP and AGD 18, 38). The pattern in case 2 was lightly different because the two upper bands appeared to be composed of doublets. The reasons are not clearly understood, but the signal in this case was low when compared with case 3 and canonical PSP. The protein needed was higher in case 2 (250 µg) than in case 3 (50 µg).

Figure 3.

Western blots of sarkosyl‐insoluble fractions in the amygdala and striatum in case 3 stained with the antiphospho‐tau(Ser422) antibody. Bands of 68 kDa and 64 kDa together with bands of about 50 kDa and lower are present in the amygdala and striatum. These bands are similar to hose currently found in PSP and AGD cases. Abbreviations: PSP = progressive supranuclear palsy; AGD = argyrophilic grain disease.

Oxidative stress markers (MDA‐L, AGE, CML and CEL)

Gel electrophoresis was carried out in total homogenates of the amygdala and striatum in three controls (cases 6, 7 and 8), and diseased cases 2, 3 and 5. Membranes were immunoblotted with anti‐MDAL, AGE, CEL and CML antibodies. Control and diseased cases were run in parallel.

Several bands were obtained in the amygdala in control and diseased cases with the different antibodies. Yet a band of about 50 kDa was detected with the four antibodies in cases 3 and 5. This band was not detected, or it was very faint in the three controls and in case 2. However, anti‐AGE antibodies recognized a band of about 70 kDa only in case 2 (Figure 4, upper panel).

Figure 4.

Monodimensional gel electrophoresis and western blotting to MDA‐L, AGE, CEL and CML in the amygdala and striatum in three controls, and in cases 2, 3 and 5. Increased density of one band of about 50 kDa is found in the amygdala in cases 3 and 5, and in the striatum in cases 2, 3 and 5 with in membranes processed with anti‐AGE antibodies. Lower intensity of the band in the striatum in case 5 when compared with case 2 and 3 can be related with the longer post‐mortem delay in this case. Abbreviations: MDA‐L = malondyaldehyde‐lysine; AGE =  advanced glycation end products; CEL = carboxy‐ethyl‐lysine; CML =  carboxy‐methyl‐lysine.

Several bands were also present in the striatum in control and diseased cases. Yet an AGE band of about 50 kDa was increased only in diseased brains, but not in controls (Figure 4, lower panel). A band of higher molecular weight (about 70 kDa) was also present in case 2. The intensity of MDA‐L, CEL and CML bands was similar in disease cases and controls, thus indicating a preferential damage related with advanced glycation end products.

2D gel electrophoresis, western blotting and mass spectrometry

Amygdala

2D gels stained with Comassie revealed three spots close to 50 kDa in case 3 when compared with controls. Parallel membranes blotted for AGE disclosed AGE immunoreactivity in these spots. The three spots were excised from the gels and identified by mass spectrometry as glial fibrillary acidic protein. Mowse scores higher than 50 (96, 204 and 273) were considered of high confidence of identification (Table 3).

Table 3.

Excised spots from 2D gels in the amygdala in case 3 and in the striatum in cases 2 and 3. Mowse scores higher than 50 are considered of high confidence of identification. Abbreviation: GFAP = glial fibrillary acidic protein.

Protein	Molecular weight	pI	Mowse score	Number of peptides matched	ID number
Amygdala case 3
GFAP	49.7 kDa	5.42	204	12	gi|38566198
GFAP	47.6 kDa	5.4	96	5	gi|119571954
GFAP	49.7 kDa	5.42	273	16	gi|38566198
Striatum case 2
GFAP	49.7 kDa	5.42	286	22	gi|38566198
GFAP	49.7 kDa	5.42	226	21	gi|38566198
GFAP	49.7 kDa	5.42	222	17	gi|38566198
Striatum case 3
GFAP	49.7 kDa	5.42	377	22	gi|38566198
GFAP	49.7 kDa	5.42	279	21	gi|38566198

Validation was carried out in 2D gels of one control and case 3 run in parallel and processed for Western blot with anti‐GFAP antibody. Increased expression of GFAP was found in case 3 as three spots of high density and about 50 kDa of molecular weight.

No differences were seen between controls and case 2 (data not shown).

Striatum

2D gels stained with Comassie revealed five spots (three from case 2, two from case 3) close to 50 kDa in cases 2 and 3 when compared with controls. The five spots were excised and identified by mass spectrometry as glial fibrillary acidic protein (GFAP) (Table 3). All the spots excised were consistent with GFAP. No other proteins or negative results were obtained in the present study. Mowse scores varied from 222 to 377, and they were considered to be of high confidence of identification (above 50).

Parallel membranes of cases 2 and 3, and corresponding controls showed the five spots immunoreactive for AGE and for GFAP (Figure 5).

Figure 5.

2D gel electrophoresis and western blotting to AGE shows the presence of differential spots in case 2 when compared with the corresponding control run in parallel (arrows). Parallel membranes blotted for GFAP show that AGE‐immunoreactive spots are also GFAP positive. Abbreviations: AGE = advanced glycation end products; GFAP = glial fibrillary acidic protein.

DISCUSSION

The present study was carried out in an attempt to delineate neuropathological modifications consistent with early changes in PSP. As this is a small number of cases, and they have been selected from a series of AGD with associate PSP‐like pathology, we can not exclude certain bias in their selection and therefore that results of the present series may be not universal for PSP. This remark is even more relevant considering the heterogeneity in terms of clinical manifestations and neuropathological findings in established PSP 6, 12, 24, 36, 44, 55). As stated at the beginning, the present observations do not attempt to establish a rigid and universal scheme of PSP progression, but rather to increase our understanding about early changes related with PSP‐like pathology.

Categorization of AGD pathology was carried out following established criteria 18, 42). Cases 1 and 2 were considered as stage 2, and cases 3, 4 and 5 as stage 3. Previous studies have shown that early stages in AGD are usually asymptomatic, and this is also observed in cases 1 and 2. AGD stage 3 is often associated with a variety of clinical symptoms, including cognitive decline, dementia, behavioral abnormalities, personality changes, and emotional and mood imbalance 8, 23, 41, 48, 49, 50, 51). Cognitive and behavioral abnormalities were also present in cases 3 and 4, and these changes were ascribed to AGD. Cognitive impairment in case 5 was considered as a combined result of AGD and PSP.

For instrumental purposes, three main groups of regions were considered in the study of PSP‐like pathology: (i) caudate/putamen; (ii) globus pallidus/subthalamus; and (iii) substantia nigra/ceruleus/colliculi/ventral pons.

The most remarkable lesions in cases 1 and 2 were astrocytic gliosis and hyperphospho‐tau accumulation in caudate and putamen. Changes increased in intensity in case 3, where involvement of the globus pallidus and subthalamus, together with mild involvement of brain stem structures, was also observed. The intensity of lesions, ranging from mild to moderate, augmented in all three groups of regions in case 4. All these regions were severely affected in case 5 (prototypical PSP).

No motor deficits were present in cases 1, 2 and 3. The neurological examination carried out at the time of admission was not relevant in cases 1 and 2. Cognitive impairment in case 3 was considered consistent with AGD. Mild motor deficits were present in case 4. Two falls in the last 3 months was the only additional complain in this patient. However, the neurological examination disclosed, in addition to memory loss and cognitive impairment, slow saccadic eye movements and slight tremor in the upper extremities, as well as certain loss of stability. Motor symptoms and signs we considered the consequence of PSP‐like pathology. Finally, neurological deficits and neuropathological findings in case 5 were typical of PSP.

Together, these observations point to early involvement in the caudate/putamen in cases with PSP‐like pathology, followed by globus pallidus/subthalamus and selected nuclei of the brain stem. It may be postulated that clinical symptoms appear as the intensity of lesions increase in these regions from a determinate threshold in a particular individual. Yet it is noteworthy that mild or moderate lesions in the caudate/putamen, even associated with mild tau pathology in the globus pallidus/subthalamus and selected nuclei of the brain stem, are not accompanied by significant clinical deficits, at least those discriminated in the current clinical practice.

Interestingly, midbrain hypometabolism has been identified as an early diagnostic sign in PSP (32), and dopaminergic dysfunction has been characterized in subclinical familial PSP (45). More pertinent in the present context is the observation of early striatal abnormalities in nonaffected individuals in a large kindred with autosomal dominant PSP linked to 1q31.1, as revealed by 18F‐dopa and 18‐fluordeoxyglucose positron emission tomography 15, 37, 40). Together, neuropathological and imaging data point to the caudate/putamen as the initial vulnerable region in PSP, followed by brain stem and globus pallidus/subthalamus.

Accumulation of phospho‐tau in neurons and astrocytes occurred in cases consistent with early stages of PSP‐like pathology, whereas coiled bodies in oligodendrocytes were rarely seen. Moreover, abnormal astrocytes rarely had the morphology of tufted astrocytes. Early astrocytic lesions were rather characterized by fine tau‐immunoreactive networks and tau‐immunoreactive astrocytes. Tufted astrocytes were absent in case 1, exceptional or very rare in cases 2 and 3, and occasional in case 4. On the one hand, these findings suggest that tufted astrocytes and coiled bodies appear later in the course of the disease. On the other, these observations also indicate that astrocytes are early targets of abnormal tau phosphorylation in the caudate and putamen in cases with early PSP‐like pathology. These observations are in line with previous studies suggesting that tau accumulation (and pathology) in astrocytes is a degenerative rather than a reactive process in PSP (47).

Previous studies have shown increased lipid peroxidation, as revealed by increased tissue levels of MDAL and HNE in the subthalamic nucleus, midbrain and superior frontal cortex in PSP brains 1, 2, 30, 34). Moreover, the activity of antioxidant systems such as superoxide dismutase and reduced glutathione is increased in multiple brain regions of PSP with reactive gliosis (13). Finally, stress‐related proteins, as stress‐activated protein kinase (SAPK/JNK) and p38SAPK, are activated and expressed in neurons and glial cells in PSP 3, 17). In line with these findings, evidence of oxidative stress damage was also observed in the amygdala (selected as a vulnerable region in AGD) and striatum (selected as a susceptible region in cases with PSP‐like pathology). Single gel electrophoresis and western blotting showed increased lipoxidative and glycoxidative damage in diseased cases when compared with controls. Changes were observed in the amygdala and striatum thus showing that oxidative damage affects regions vulnerable to AGD and PSP‐vulnerable regions as well.

2D gel electrophoresis, western blotting, in gel digestion and mass spectrometry revealed GFAP as a major target of oxidative damage in the striatum in conventional PSP and in the two cases with PSP‐like pathology. GFAP was also oxidized in the amygdala in AGD associated with conventional PSP and in one case (case 3) with PSP‐like pathology. GFAP has been found modified by oxidation in AD 27, 35) and Pick's disease (33). GFAP oxidation has also been reported in conditions not associated with tau pathology as in aceruloplasminemia (25) and diabetic retina (5). Whether GFAP oxidation is the result of phospho‐tau deposition or an unrelated event in astrocytes associated with PSP‐like pathology is not solved.

Oxidative damage to GFAP is associated with increased expression of GFAP, as revealed with western blotting and with increased astrogliosis as shown by immunohistochemistry. The present findings further support the concept of early involvement of astrocytes (increased numbers, increased GFAP levels and increased astroglial tau phosphorylation), and of GFAP as a target of oxidative damage in cases with PSP‐like pathology.

Supporting information

Figure 1S. Case 1: GFAP (A) and AT8 (B–D) immunoreactivity in the caudate (A,B), globus pallidus (C) and subthalamic nucleus (D). Moderate gliosis is observed in the caudate together with phosphor‐tau accumulation in neurons and glial cells. Only rare neurons are immunoreactive with the AT8 antibody in the globus pallidus and subthalamus. Paraffin section slightly counterstained with haematoxylin. A, bar = 50 μm; B–D, bar in D = 50 μm.

Figure 2S. Case 2: AT8 immunoreactivity in the caudate (A), amygdala (B), CA1 region of the hippocampus (C) and dentate gyrus (D). Tau immunoreactivity is observed in astrocytes and in neurons in every region. In addition, grains are clearly visible in CA1. Paraffin section slightly counterstained with haematoxylin. Bar = 25 μm.

Figure 3S. Case 3: AT8 immunoreactivity in the putamen (A), caudate (B), amygdala (C), CA1 region of the hippocampus (D), dentate gyrus (E), nucleus basalis of Meynert (F), ventral pons (G), midbrain (H), ceruleus (I) and gyrus cinguli (J). Tau immunoreactivity is observed in astrocytes and in neurons in every region. Note grains in the CA1 region and amygdale. Rare tufted astrocytes are seen in the caudate (A) and midbrain (H). Paraffin section slightly counterstained with haematoxylin. Bar = 25 μm.

Figure 4S. Case 5: AT8 immunoreactivity in the putamen (A), amygdala (B), globus pallidus (C), dorsomedial nucleus of the thalamus (D), subthalamus (E), superior colliculus (F), ventral pons (G), frontal cortex (H) and CA1 region of the hippocampus. Paraffin section slightly counterstained with haematoxylin. A–G, I, bar in I = 25 μm; C, E, H, bar in H = 50 μm.

Table 1S. Summary of neuropathological findings in case 1. AGD classification: early stage 2. AD classification: ADI; no β‐amyloid deposition

Table 2S. Summary of neuropathological findings in case 2. AGD classification: stage 2. AD classification: ADI; no β‐amyloid deposition

Table 3S. Summary of neuropathological findings in case 3. AGD classification: stage 3. AD classification: ADIIA

Table 4S. Summary of neuropathological findings in case 4. AGD classification: stage 3. AD classification: ADI; no β‐amyloid deposition

Table 5S. Summary of neuropathological findings in case 5. AGD classification: stage 3. No β‐amyloid deposition

This material is available as part of the online article from: http://www.blackwellsynergy.com

Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Supporting info item

Supporting info item