Acta Neuropathol (1999) 98 : 141–149
© Springer-Verlag 1999
R E G U L A R PA P E R
Panteleimon Giannakopoulos · Enikò Kövari · Armand Savioz · Fabienne De Bilbao · Michel Dubois-Dauphin · Patrick R. Hof · Constantin Bouras
Differential distribution of presenilin-1, Bax, and Bcl-XL in Alzheimer’s disease and frontotemporal dementia Received: 27 October 1998 / Revised: 4 January 1999 / Accepted: 5 January 1999
Abstract We have previously reported that presenilin-1 (PS-1)-immunoreactive neurons survive in late-onset sporadic Alzheimer’s disease (AD). To examine if this is also the case in other dementing conditions, and if it is associated with changes in the expression of the main apoptosisrelated proteins, a quantitative immunocytochemical study of presenilin-1, Bax, and Bcl-XL in the cerebral cortex of non-demented and AD patients, and patients with frontotemporal dementia (FTD) was performed. In nondemented cases, the frequency of neurons showing PS-1 immunoreactivity was 25–60%, Bax immunoreactivity 36–54%, and Bcl-XL immunoreactivity 26–63% depending on the cortical area. The frequency of NFT-free neurons which contained PS-1 or Bax was consistently increased in all of the areas in AD. In FTD cases, the percentage of PS-1-, but not Bax-immunoreactive neurons was increased only in areas displaying a substantial neuronal loss. Conversely, there was no difference in the den-
P. Giannakopoulos (쾷) · E. Kövari · A. Savioz · F. De Bilbao · M. Dubois-Dauphin · C. Bouras Department of Psychiatry, HUG Belle-Idée, University of Geneva School of Medicine, 1225 Geneva, Switzerland e-mail:
[email protected] Tel.: +41-22-3055001, Fax: +41-22-3055040 P. R. Hof · C. Bouras (쾷) Neurobiology of Aging Laboratories and Fishberg Research Center for Neurobiology, New York, NY 10029, USA e-mail:
[email protected] Tel.: +41-22-3055358, Fax: +41-22-3055350 P. R. Hof Departments of Geriatrics and Adult Development, and Ophthalmology (PRH), Mount Sinai School of Medicine, New York, NY 10029, USA Present addresses: P. Giannakopoulos, Clinic of Geriatric Psychiatry, HUG-Belle-Idée, CH-1225 Geneva, Switzerland C. Bouras, Division of Neuropsychiatry, HUG Belle-Idée, 2 chemin du Petit Bel-Air, CH-1225 Geneva, Switzerland
sities of Bcl-XL-containing neurons among the three diagnosis groups. These data suggest that surviving neurons in affected cortical areas in AD show a high expression of PS-1 and Bax, indicating that these proteins play a key role in the mechanisms of cell death in this disorder. In FTD, neurons containing PS-1 are preserved, further supporting a neuroprotective role for this protein in other neurodegenerative disorders. Key words Apoptosis · Dementia · Neurofibrillary tangles · Neuroprotection · Presenilin
Introduction Several studies have indicated that neuronal apoptosis is an important mechanism of cell death in Alzheimer’s disease (AD) [7, 29, 36, 42, 45–47]. Recently, an expanding family of genes encoding homologous proteins has been identified as the Bcl-2 family, and appears to determine whether or not a cell will undergo apoptosis [1, 13, 16, 24, 28, 30, 35, 52, 53]. Among them, the Bcl-2 and Bcl-XL gene products can block apoptotic cell death and promote survival of neurons [1, 13, 16, 24, 30, 53], whereas the Bax gene product (Bcl-2-associated X protein) may act as a promoter of cell death [35]. Bax heterodimerizes with Bcl-XL abrogating their anti-apoptotic function [52], and it has been postulated that the ratio of Bax to Bcl-XL plays a key role in determining the relative sensitivity of neurons to apoptotic cell death [1, 28, 35, 52]. In AD, a Bcl-2 down-regulation and a Bax up-regulation have been reported in neurons prone to degeneration [29, 46], while a loss of Bax immunoreactivity may render neurons more resistant to programmed cell death [29]. Subsequently, it has been proposed that an up-regulation of Bax may act by promoting both neurofibrillary tangle (NFT)-related and NFT-unrelated neuronal loss [47]. Although Bcl-XL-positive microglia has been frequently seen colocalized within senile plaques (SP) [9], the role of this protein in neuronal death in AD is still unclear. The two recently identified genes carrying mutations in the
142
majority of early-onset familial AD, presenilin 1 and 2 (PS-1, PS-2), are also thought to be involved in the regulation of cell death in AD [17, 19, 25, 26, 41]. We have previously reported that PS-1-immunoreactive neurons are prone to survive in late-onset sporadic AD and Pick’s disease, two dementing conditions characterized by prominent cytoskeletal pathology ([15], and P. Giannakopoulos, C. Bouras, E. Kövari, J. Shioi, L. Buée, P. R. Hof, and N. K. Robakis, Presenilin-1 expression in Pick’s disease, Acta Neuropathol. (P. Giannakopoulos et al., submitted)). However, it is not clear whether PS-1 overexpression occurs in dementing conditions without cytoskeletal pathology, and whether it is associated with changes in the expression of the main apoptosis-related proteins. To address these issues, we performed a quantitative immunocytochemical study of PS-1, Bax, and Bcl-XL proteins in the cerebral cortex of non-demented (ND) individuals, and patients with AD and frontotemporal dementia (FTD).
Materials and methods A total of 30 patients (18 women, 80.5 ± 2.1 years old; 12 men, 76.5 ± 2.0 years old), who died and were autopsied in the Hospitals of the University of Geneva School of Medicine, were included in the present study. Among them, 11 patients (6 women, 82.5 ± 1.8 years old; 5 men, 77.8 ± 1.4 years old) had preserved cognitive abilities and were considered as non-demented. Most of these patients were admitted to the hospital with symptoms of cardiac failure, acute pulmonary insufficiency or chronic vascular disease. Their mean Mini-Mental State Examination (MMSE) score at the final admission was 28.6 ± 1.2, and the score of the extended Clinical Dementia Rating (CDR) Scale applied retrospectively was 0.5 ± 0.3 [11, 22]. The clinical diagnosis was confirmed neuropathologically by the absence of significant histopathological changes in the ND group. Ten patients (7 women, 83.3 ± 2.6 years old; 3 men, 82.6 ± 2.3 years old) showed severe cognitive deterioration and were classified clinically as AD according to the DSMIV criteria. Their hospitalization was motivated by the presence of major behavioral disturbances such as psychomotor agitation, marked aggressiveness, delusions of persecution, and suicidal thoughts. Detailed neuropsychological evaluation performed at least twice during the 6 months prior to death revealed a significant decline of higher cortical functions characterized by severe memory impairment, temporal and spatial disorientation, language impoverishment, apraxia, and agnosia in all of the cases. These demented patients form the AD group of this study. The mean MMSE score of this group was 14.5 ± 3.8, and extended CDR score was 3.5 ± 0.5. The neuropathological investigation according to the Consortium to Establish a Registry for Alzheimer Disease (CERAD) criteria confirmed the clinical diagnosis of AD [31]. The remaining nine patients (6 women, 75.1 ± 1.8 years old; 3 men, 70.7 ± 2.4 years old) showed early frontal symptomatology, speech disorders and mild memory impairment and were classified clinically as FTD [14]. The neuropathological confirmation was based on the Lund and Manchester criteria [2]. All of the brains were obtained at autopsy (postmortem delay: 3–10 h), fixed in a 10% formalin solution for at least 6 weeks, and cut into 1-cm-thick coronal slices. Following macroscopic examination, tissue blocks were dissected from the hippocampal formation, the entorhinal cortex, the superior and middle frontal cortex, middle and inferior temporal cortex, and the inferior parietal cortex. In FTD cases, additional tissues blocks were taken from the amygdala, thalamus, striatum, substantia nigra, brain stem and cerebellum. For microscopic purposes, blocks were washed in a series of graded sucrose solutions (12%, 16%, 18%, and 30%) in cold phosphate-buffered saline (PBS), frozen, and cut at 12-µm-
thick sections. For routine neuropathological evaluation, tissues were stained with Nissl, hematoxylin-eosin, and Globus silver impregnation stains [51]. PS-1 expression was detected using a fully characterized polyclonal antibody raised against the N-terminal of the PS-1 sequence (TELPAPLSYFNRKC-NH2) [10], as previously described (see Fig. 1 a) [15]. This antibody recognizes both the 46-kDa holoprotein and the 27-kDa PS-1 N-terminal derivative. Bax and Bcl-XL expression was studied used two polyclonal antibodies raised against N-terminal sequences of human Bax (N20) and Bcl-XL (Santa Cruz Biotechnology, Santa Cruz, Calif.). In the present study, specificity of staining was confirmed by preadsorption with the corresponding synthetic peptide. The percentages of PS-1-, Bax-, and Bcl-XL-immunoreactive NFT-free neurons were obtained from Nissl-stained sections. Briefly, 12-µm-thick cryostat sections were rinsed in PBS followed by treatment for 10 min with potassium permanganate (0.25%) to mask lipofuscin fluorescence [15]. To see if potassium permanganate treatment masks most of the lipofuscin autofluorescence, copper sulfate treatment was used in adjacent sections in three randomly selected cases from each diagnosis groups as previously described [40]. Subsequently, the pattern of lipofuscin autofluorescence in these sections was compared to that of sections treated with potassium permanganate. After rinsing in PBS, slides were treated with 1% oxalic acid and 1% potassium metabisulfite in PBS for 2 min. After incubation overnight with the primary antibodies, sections were incubated with a peroxidase-conjugated anti-rabbit secondary antibody for 1 h followed by rinsing with PBS. The sections were treated with 3, 3′-diaminobenzidine as a chromogen and counterstained with Nissl stain [15, 29]. In addition to staining with PS-1, Bax and Bcl-XL, double labeling was performed in adjacent sections with a modified thioflavin S stain to visualize NFT and SP in AD cases [51]. To assess whether the antibodies used in the present study cross-reacts with lipofuscin, a series of ten randomly selected, non-treated sections from all ND, AD and FTD cases were examined to detect lipofuscin autofluorescence, and the pattern of lipofuscin distribution was compared to that of PS-1, Bax, and Bcl-XL in each case in immediately adjacent sections stained with the corresponding antibodies. In all of the brains, total and PS-1-, Bax-, and Bcl-XL-immunoreactive neuron densities were estimated in the anterior CA1–3 fields of the hippocampus and hilus of the dentate gyrus, subiculum, layers II and V of the entorhinal cortex, and layers II–III and V–VI of areas 9 (superior frontal cortex) and 20 (inferior temporal cortex). In addition, the total number of NFT-containing neurons and SP and those immunoreactive for PS-1, Bax, and Bcl-XL were determined in all of these areas in AD cases. Neuron densities per mm3 were estimated using the optical dissector, an unbiased stereological counting method allowing that all regions within the structure of interest have an equal chance of being analyzed (i.e., there is no bias in sampling), and that counts do not depend on variables such as the size and shape of neurons and degree of cellular atrophy [44]. The technique relies on a three-dimensional counting box located entirely within the tissue section, and objects are quantified by focusing in the section depth (i.e., in the z axis). The fact that the three-dimensional counting box is located within the thickness of the section and the existence of exclusion (forbidden) planes guarantee that any neuron may be counted only once [44]. Total neuron and NFT numbers were not obtained in the present study due to the fact that only the anterior portion of the hippocampus and only topographically equivalent samples of areas 9 and 20 were available for analysis. These regions were outlined at a low magnification on the computer display and the area included in these outlines was calculated on each section. The average section thickness was subsequently estimated at high magnification (× 63) by focusing on the top and bottom faces of the sections and measuring the distance between them in at least five randomly distributed locations within the outlines. The software then placed within each laminar boundary a set of optical disector frames in a systematic-random fashion and neurons were counted in a series of disector stacks, the thickness of which was kept constant throughout the study. To achieve this, the computer software was moving the stage in the z axis in regular steps of 1 µm. The volume of the
143
a
b
c
d
Fig. 1 a–d Thick perinuclear PS-1 staining in a pyramidal neuron (a) and glial cell (b, arrow) of the CA1 field in a non-demented 72-year-old patient. c, d Double labeling visualization of Bcl-XLimmunoreactive NFT-bearing neurons (c) and SP (d) in the entorhinal cortex in a 80-year-old patient with AD. Visualization of neurons containing PS-1 was made using an antibody against the PS-1 in Nissl counterstained sections (a, b). Double labeling with an antibody against the Bcl-XL and modified thioflavin S stain was performed to visualize Bcl-XL-immunoreactive NFT-bearing neurons (c) and SP (d). Arrows indicate PS-1 staining in a glial cell (b), and Bcl-XL staining in NFT-bearing neurons (c), and SP (d) (PS-1 presenilin-1, NFT neurofibrillary tangle, SP senile plaque, AD Alzheimer’s disease). Bars a, b 10 µm, c, d 20 µm regional samples was variable rendering comparison of total cell numbers impossible from case to case. For this reason, neuron and NFT densities per mm3 were assessed in a 1 in 10 series of sections, 500 µm apart, using a Zeiss 63x PlanNeofluar objective (numerical aperture 1.4). The numbers of SP were determined in each area and their mean densities per mm2 were calculated. All analyses were performed by two independent investigators (E.K. and C.B.) with a reliability of 0.95, using a computer-assisted image analysis system consisting of a Zeiss Axioplan microscope, a high sensitivity LH-4036 camera (LHESA Electronic), a COMPAQ Deskpro 386/20 microcomputer and a SAMBA 2005 software system developed by TITN Inc. (ALCATEL, Grenoble, France). The percentages of PS-1-, Bax-, and Bcl-XL-immunoreactive neurons in ND and FTD cases, as well as the percentages of PS-1-, Bax-, and Bcl-XL-immunoreactive NFT-free and NFT-containing neurons and SP in AD cases were evaluated within each cortical
layer for each selected area. Statistical differences in the frequency of PS-1-, Bax-, and Bcl-XL-containing neurons between ND cases and each of the two other diagnosis groups and between NFT-free and NFT-containing neurons in AD cases were assessed by MannWhitney U test.
Results All of the three proteins studied showed the same granular and punctate immunostaining pattern. They were intensely concentrated in the soma, as well as in the proximal segments of basal and apical dendrites in both pyramidal and granular neurons (Fig. 1 a). Scarce glial cells were also PS-1, Bax or Bcl-XL immunoreactive, in particular around SP. Their number was very low in ND brains but increased significantly in AD brains (Fig. 1 b). In addition, a weak immunoreactivity for these proteins was observed by double labeling in NFT-containing neurons and SP (Fig. 1 c, d). In particular, Bax immunoreactivity was confined to dystrophic neurites, whereas PS-1 and Bcl-XL staining was observed in both amyloid fibrils and dystrophic neurites (Fig. 1 d). Both potassium permanganate and copper sulfate efficiently masked nearly all lipofuscin autofluorescence. Moreover, no colocalization
144 Table 1 Total neuron and PS-1-immunoreactive neuron densities in ND, AD, and FTD cases. Results represent neuron number ± SEM/mm3 in each area. Note that a statistically significant neuronal loss is observed in all cortical areas in AD cases compared to ND cases. Note also the preserved neuron densities in the hilus of the dentate gyrus and entorhinal cortex in FTD cases, and the relative preservation of PS-1-immunoreactive neurons in most cortical areas in both AD and FTD cases. Cortical layers are indicated by Roman numerals (PS-1 presenilin-1, ND non-demented, AD Alzheimer’s disease, FTD frontotemporal dementia) Area/layer
ND
AD
FTD
Nissl-stained neurons CA1 32615 ± 2252 CA2–3 64303 ± 1976 Hilus 29077 ± 2206 Subiculum 36243 ± 3737 Entorhinal II 46999 ± 4510 Entorhinal V 50576 ± 3618 20II-III 50421 ± 2937 V-VI 54346 ± 6858 9II-III 65579 ± 2833 V-VI 72551 ± 3848
14172 ± 3018a 33523 ± 3343a 14332 ± 1231a 28105 ± 2416c 23240 ± 2755a 39233 ± 2710c 26919 ± 3186a 32513 ± 2580b 39635 ± 4762a 55940 ± 2532c
13987 ± 2133a 39867 ± 3211b 23566 ± 2287 24554 ± 2417b 38065 ± 3081 46766 ± 4026 14464 ± 1970a 24427 ± 1972a 35944 ± 1942a 61285 ± 3366c
PS-1-immunoreactive neurons CA1 12837 ± 918 CA2–3 32557 ± 3111 Hilus 16439 ± 1274 Subiculum 16008 ± 2015 Entorhinal II 20841 ± 1438 Entorhinal V 18676 ± 1319 20II-III 17555 ± 1556 V-VI 16588 ± 1639 9II-III 19797 ± 827 V-VI 18173 ± 2919
9140 ± 1809 26688 ± 2343 10156 ± 1234 16383 ± 744 19776 ± 1844 20446 ± 1183 13749 ± 1225 18172 ± 2707 20988 ± 1353 20381 ± 808
9368 ± 1152 31122 ± 3100 13543 ± 1778 16607 ± 1659 18844 ± 1697 21370 ± 2074 8067 ± 856c 14511 ± 1085 16947 ± 726 21133 ± 1064
Statistical analysis was performed by Mann-Whitney U test: aP < 0.001; bP < 0.01; cP < 0.05, compared to ND cases
was found between lipofuscin autofluorescence and PS-1, Bax and Bcl-XL staining [15]. Estimates of Nissl-stained neuronal densities demonstrated that AD cases had consistently lower neuron densities in all of the areas studied. In FTD cases, a profound neuronal loss was found in areas 20 and 9 and, to a lesser degree, in CA1–3 fields and subiculum, whereas the hilus of the dentate gyrus and entorhinal cortex showed preserved neuron densities (Table 1). Despite the neuronal loss, the number of neurons showing PS-1 and Bax immunoreactivity in the areas studied in all AD cases was comparable to that found in ND cases (Tables 1, 2). In contrast, there was a substantial decrease in the densities of Bcl-XL-immunoreactive neurons in most of the areas studied in AD compared to ND cases (Table 2). Except in layers II and III of area 20, PS-1-containing neuron densities did not significantly differ between the ND and FTD brains. Areas 9 and 20 displayed a significant decrease in Bax- and Bcl-XL-containing neuron densities in FTD compared to ND cases (Table 1). Since counting of immunoreactive neurons in AD included both NFT-free and NFT-bearing neurons, double labeling was performed to differentiate these two neu-
Table 2 Bax- and Bcl-XL-immunoreactive neuron densities in ND, AD, and FTD cases. Results represent neuron number ± SEM/mm3 in each area. Note the relative preservation of Bax- but not Bcl-XL-immunoreactive neurons in most cortical areas. Cortical layers are indicated by Roman numerals Area/layer
ND
AD
FTD
Bax-immunoreactive neurons CA1 16085 ± 2179 CA2–3 29972 ± 3765 Hilus 14853 ± 2486 Subiculum 15853 ± 1771 Entorhinal II 20764 ± 1740 Entorhinal V 20145 ± 2103 20II-III 21151 ± 1577 V-VI 25393 ± 2670 9II-III 25249 ± 2857 V-VI 25327 ± 2949
11572 ± 2440 23893 ± 3100 11556 ± 2564 15031 ± 1527 19744 ± 1748 22974 ± 968 18118 ± 2228 20429 ± 1781 23764 ± 2377 26935 ± 764
10062 ± 1414 23001 + 3778 16111 ± 1449l 12783 ± 1629 17074 ± 1956 22648 ± 2693 7522 ± 736a 12590 ± 1366b 15983 ± 961b 25476 ± 1953
Bcl-XL-immunoreactive neurons CA1 16047 ± 1638 CA2–3 32867 ± 2813 Hilus 17646 ± 1171 Subiculum 14693 ± 2159 Entorhinal II 20107 ± 2310 Entorhinal V 18444 ± 1218 20II-III 19759 ± 1595 V-VI 19411 ± 2707 9II-III 21112 ± 1567 V-VI 18831 ± 2819
7078 ± 1861b 18134 ± 2879b 8000 ± 774b 10618 ± 853 12241 ± 1120c 13033 ± 1570 12226 ± 1045b 12218 ± 1409c 15394 ± 940b 15047 ± 1161
7309 ± 1048a 23549 ± 2666 13775 ± 1711 9539 ± 1102c 18743 ± 941 14199 ± 1233 6016 ± 824a 9883 ± 1199a 13540 ± 901a 18063 ± 1698
Statistical analysis was performed by Mann-Whitney U test: aP < 0.001; bP < 0.01; cP < 0.05, compared to ND cases
ronal subpopulations, and the percentage of PS-1-, Baxand Bcl-XL-containing neurons was calculated separately (Table 3). In the ND group, the frequency of neurons showing PS-1 immunoreactivity was 25–60%, Bax immunoreactivity 36–54%, and Bcl-XL immunoreactivity 26–63% depending on the cortical area. The CA2–3 fields and hilus of the dentate gyrus displayed the highest percentages of immunoreactive neurons for all three proteins, whereas in area 9 fewer than 40% of all neurons contained PS-1, Bax or Bcl-XL (Table 3). The frequency of NFT-free neurons which contained PS-1 or Bax was increased in all of the areas studied in AD (Table 3; Fig. 2 a, b, d, e). Conversely, there was no difference in the densities of Bcl-XLcontaining neurons between the two diagnosis groups (Table 3; Fig. 2 g, h). In FTD cases, the percentage of PS1-containing neurons was significantly increased only in areas which displayed a substantial neuronal loss (see Tables 1, 3). For example, the percentages of neurons showing PS-1 immunoreactivity in each of the CA1 field, and layers II–III and V–VI of area 20 were 70%, 60%, and 55%, respectively. In the corresponding areas of control brains, the percentages of PS-1-containing neurons were 40%, 35%, and 32%, respectively (Table 3; Fig. 2 a, c). No difference was found in Bax- and Bcl-XL-immunoreactive neuron densities between ND and FTD cases (Fig. 2 d, f, g, i). Only a few NFT-containing neurons displayed PS-1, Bax or Bcl-XL labeling in the areas studied. For example,
145 Table 3 Frequency of PS-1, Bax and Bcl-XL in ND, AD, and FTD cases. Results represent the percentages of PS-1-, Bax- and Bcl-XL-immunoreactive neurons in each area. In AD cases, these percentages were calculated in NFT-free neurons. Cortical layers are indicated by Roman numerals Area/layer
ND
AD
FTD
PS-1 CA1 CA2–3 Hilus Subiculum Entorhinal II Entorhinal V 20II-III V-VI 9II-III V-VI
39.7 ± 2.7 50.6 ± 4.5 60.3 ± 2.9 44.1 ± 2.8 45.1 ± 2.7 37.4 ± 2.9 34.8 ± 2.0 31.8 ± 2.4 30.4 ± 1.6 25.0 ± 2.3
65.5 ± 1.7a 64.7 ± 3.8b 67.1 ± 4.2c 59.5 ± 4.5b 71.1 ± 3.2a 52.4 ± 2.0c 62.6 ± 1.5a 55.0 ± 4.8a 54.8 ± 4.4a 36.8 ± 2.5c
70.4 ± 3.3a 73.0 ± 2.8a 55.6 ± 3.2 61.6 ± 2.9b 52.9 ± 2.4 44.1 ± 6.1 59.7 ± 4.6a 55.1 ± 2.8a 48.5 ± 3.3b 34.5 ± 2.6c
Bax CA1 CA2–3 Hilus Subiculum Entorhinal II Entorhinal V 20II-III V-VI 9II-III V-VI
51.0 ± 7.8 47.8 ± 4.7 53.9 ± 7.4 45.1 ± 5.3 45.7 ± 3.5 42.0 ± 6.7 43.3 ± 4.9 47.3 ± 5.6 39.3 ± 5.0 35.8 ± 3.0
83.2 ± 3.9b 69.1 ± 4.5b 82.2 ± 8.6b 58.9 ± 7.6c 80.8 ± 4.0a 60.2 ± 6.3c 68.0 ± 6.8c 63.4 ± 5.7c 60.9 ± 3.8b 48.8 ± 1.8c
61.6 ± 4.5 53.8 ± 4.5 64.2 ± 4.8 51.1 ± 4.0 52.7 ± 4.6 50.5 ± 7.0 52.2 ± 6.0 50.4 ± 2.8 42.6 ± 2.9 40.4 ± 2.3
Bcl-XL CA1 CA2–3 Hilus Subiculum Entorhinal II Entorhinal V 20II-III V-VI 9II-III V-VI
48.9 ± 2.1 50.9 ± 3.6 63.4 ± 5.9 40.1 ± 3.2 43.1 ± 4.2 36.8 ± 4.6 39.2 ± 2.3 36.9 ± 3.5 32.1 ± 1.6 25.9 ± 3.6
50.5 ± 7.5 52.6 ± 3.3 57.1 ± 5.6 38.9 ± 3.8 50.4 ± 3.4 31.3 ± 3.7 46.5 ± 3.6 38.6 ± 4.1 40.5 ± 3.8 27.3 ± 2.8
51.9 ± 3.0 53.7 ± 3.9 59.8 ± 3.5 39.2 ± 3.8 44.6 ± 4.4 31.9 ± 3.9 46.3 ± 6.4 36.8 ± 2.9 38.8 ± 3.6 29.2 ± 3.0
Statistical analysis was performed by Mann-Whitney U test: a P < 0.001; b P < 0.01; c P < 0.05 compared to ND cases. See text for details
the percentages of PS-1, Bax, and Bcl-XL-containing NFT-containing neurons were 15–20%, 11–18%, and 5–9% in hippocampal subdivisions and 11–20%, 15–19%, and 6–10% in neocortical areas, respectively. These percentages were significantly lower than those of NFT-free neurons in all of the cortical areas studied (P < 0.001). The densities of immunolabeled SP varied substantially within the cerebral cortex. For instance, 23% of SP were PS-1-immunoreactive, 20% Bax-immunoreactive, and 11% Bcl-XL-immunoreactive in the CA1 field, whereas these percentages were 50%, 37%, and 20% in the hilus of the dentate gyrus, 41%, 24%, and 18% in layers V–VI of area 20, and 35%, 30%, and 12% in layers II–III of areas 9 and 20.
Discussion The present data indicate that surviving neurons in affected cortical areas in AD are predominantly PS-1 and Bax immunoreactive, suggesting that these proteins participate in the regulation of cell death in this disorder. In FTD, neurons containing PS-1 are well preserved, supporting further a possible neuroprotective role for this protein in various neurodegenerative disorders [15]. Moreover, our results show that Bcl-XL, an anti-apoptotic protein, is not activated in AD and FTD. The regional distribution of PS-1, Bax, and Bcl-XL is comparable to that reported in previous immunocytochemical studies of ND and AD cases, in that these proteins show a granular punctate pattern of staining in the soma and the proximal segments of basal and apical dendrites of pyramidal neurons [3, 5, 15, 23, 29, 32, 33, 46, 49]. However, the quantitative analysis revealed marked differences in the distribution of these proteins within the cerebral cortex between AD and FTD cases. Although the increase in the percentage of PS-1- and Bax-immunoreactive neurons reported here may partly reflect an overexpression of these proteins which may take place in surviving neuronal subpopulations, it should be noted that this increase was statistically significant only in areas which exhibited a marked neuronal loss in both disorders [15]. Using a polyclonal antibody directed against a N-terminal epitope of PS-1, Hendriks and collaborators reported no difference in the pattern of PS-1 immunoreactivity in the hippocampus between sporadic AD and control cases, although no quantitative data were included in this study [22]. This confirms our results showing no difference in PS-1-immunoreactive neuron densities between ND and AD cases despite the presence of substantial neuronal loss in AD cases. With respect to PS-1, the present results extend previous studies that reported a relative preservation of morphologically intact neurons containing PS-1 in both AD [15] and Pick disease (P. Giannakopoulos, submitted), in that they imply that this is also the case in FTD. Consistent with this possibility, our preliminary data using double labeling with the antibody against PS-1 and Tdtmediated dUTP-biotin nick-end labeling (TUNEL) method in all three diagnosis groups indicates that there is no evidence of DNA damage in morphologically intact neurons which express PS-1 (data not shown). Several lines of evidence suggest that PS-1 is involved in the regulation of apoptotic cell death [17, 19, 25, 26, 41]. For instance, PS-1 knockout mice show substantial neuronal loss in the cerebral cortex [41], and PS-1 mutations sensitize neurons to apoptotic death induced by trophic factor withdrawal and Aβ peptide [19, 25]. In addition, an alternative cleavage of PS-1 by a caspase-3 family protease occurs during apoptosis and may contribute to the neuronal loss in AD [17, 26]. Recent studies have demonstrated that an induction of PS-1 gene takes place following transient ischemia [48], whereas the inhibition of PS-1 expression results in apoptosis and tumor suppression in immunodeficient mice [38]. In conjunction with these observations, the present find-
146
Fig. 2 a–i Representative examples of PS-1- (a–c), Bax- (d–f), and Bcl-XL- (g–i) immunoreactive neuron densities in a non-demented 75-year-old-patient (a, d, g), a 77-year-old patient with AD (b, e, h), and a 72-year-old patient with FTD (c, f, i). Note the substantial increase in the densities of PS-1-immunoreactive neurons in the AD (b) and FTD (c) cases compared to the non-demented case. Note the presence of higher Bax-immunoreactive neuron densities in the AD patient (e). There was no difference in the densities of Bcl-XLimmunoreactive neurons between the three cases. Bar 20 µm
ings indicate that PS-1 expression may protect against apoptotic neuronal loss not only in dementing conditions with prominent cytoskeletal pathology, such as AD and Pick’s disease, but also in dementia lacking distinctive histopathology, such as FTD. The increase in Bax-immunoreactive neuron densities in AD cases observed here agrees with a recent report showing an up-regulation of Bax in neurons lacking evidence of neurofibrillary changes [47], and suggests that Bax plays an important role in promoting NFT-independent death of these neurons. Moreover, MacGibbon et al. [29] have found a significant decrease in Bax staining in the dentate granule cells of the hippocampus, and have proposed that the loss of Bax may be related to the rela-
tive resistance of these cells in AD. Altogether, these observations suggest that Bax protein may be a marker of cellular vulnerability to apoptosis in AD. By contrast, Bax immunoreactivity does not increase in FTD. This is consistent with a report showing that glutamate-induced death of cerebellar granular cells in culture does not involve Bax expression, and suggests that alternative mechanisms of cell death which do not require the activation of this protein occur in this disorder [8]. In contrast to PS-1 and Bax, the densities of Bcl-XLimmunoreactive neurons do not differ between the three diagnosis groups included in the present study. Consistent with two recent immunocytochemical studies [33, 49], this finding confirms the specificity of the increase in PS1 and Bax immunoreactivity reported here, and suggests that Bcl-XL is not recruited to protect neurons against apoptotic cell death in AD and FTD. However, it should be kept in mind that Western blot analysis has shown an increase in the level of Bcl-XL, but not Bax, protein in the temporal cortex in AD compared to ND cases [27]. Moreover, an increase in the expression of Bcl-2 has been reported in AD brains, yet sparse and equivocal Bcl-2 staining is observed in neurons in this disorder [34, 39]. Be-
147
sides methodological considerations such as antibody used and tissue preparation, the disparity between the immunocytochemical and immunoblot data may be explained by the high expression of Bcl-XL and Bcl-2 proteins in glial cells reported in the present and previous studies [27, 34, 39, 46]. Although the exact significance of this phenomenon remains unclear, it has been proposed that the expression of Bax and Bcl-XL in glial cells may interfere with neuronal viability by promoting or preventing neuronal death [27, 29, 34]. These observations imply that the balance between the activation of PS-1, Bax, and Bcl-XL in neurons and glial cells may be a key factor in the regulation of cell death in the dementing processes. With respect to AD lesions, our results show that NFT are rarely positive for the three proteins studied [4]. In particular, our observations agree with those of Uchihara et al. [50] and Hendriks et al. [22] in sporadic AD cases, who reported a marked decrease of PS-1 labeling in NFT. Although it has been proposed that a down-regulation of PS-1 and an up-regulation of Bax precedes and possibly triggers the development of NFT [15, 29, 46], this remains a controversial matter [33]. In agreement with previous studies [3, 15, 33, 46], the present data suggest that a secondary down-regulation of these proteins follows the formation of intraneuronal NFT. This possibility is supported further by the decrease in Bcl-2 immunoreactivity in neurons displaying abnormally phosphorylated tau and NFT reported previously [34, 39, 47]. The staining of SP with all three antibodies could reflect a possible relationship between amyloid deposition and the activation of these proteins in AD [4, 18]. For instance, β-amyloid induces apoptosis in cell cultures as well as in hippocampal slices, and it has been postulated that deposition of βamyloid in SP may induce transcriptional changes leading to apoptosis in AD [12, 20, 36]. However, the substantial variation in the percentages of PS-1-, Bax-, and Bcl-XLimmunoreactive SP within the cerebral cortex independently of the severity of dementia do not corroborate this hypothesis. Alternatively, the presence of these proteins in SP may represent a non-specific phenomenon. In agreement with this hypothesis, the antibody against Bax used in the present study (N-20), which recognizes unique bands at the appropriate level, does not stain amyloid fibrils [49]. A large number of proteins have been detected in association with SP, and it is possible that some of them are adsorbed non-specifically on to the amyloid fibers when released as a result of nerve cell death and lysis [37]. In conclusion, the present study reveals two distinct patterns of PS-1- and apoptosis-related gene expression in dementing conditions. First, PS-1 immunoreactivity appears to increase in surviving neurons in AD, Pick’s disease and FTD, suggesting that this protein is activated to protect neurons at risk in a disease-independent manner. Second, Bax activation takes place only in AD, and it is likely that the balance between PS-1, Bax, and possibly other proteins plays a key role in the survival of neurons at risk in this condition. However, it should be kept in mind that apoptosis is not the only mechanism involved in
cell death in AD [6]. The DNA fragmentation which occurs in the hippocampus of AD brains is not necessarily associated with an increased rate of cells displaying the morphological characteristics of apoptosis, and it is possible that necrosis also play an equally important role in AD-related cell destruction [43]. Future studies including both estimates of neurons showing DNA damage and expression of apoptosis-related proteins are warranted to understand better the molecular mechanisms of neuronal loss in AD and other dementing conditions. Acknowledgements We thank M. Surini and P. Y. Vallon for expert technical assistance. Supported by grant 31-45960.95 (CB and PG), and 4038-44006 (MDD and PG) from the Swiss National Science Foundation, Bern, Switzerland and grant AG05138 (PRH) from the National Institutes of Health, Bethesda, USA.
References 1. Boise LH, Gonzàlez-Garcìa M, Postema CE, Ding L, Lindsten T, Turka LA, Mao X, Nùnez G, Thompson CB (1993) bcl-x, a bcl-2 related gene that functions as a dominant regulator of apoptotic cell death. Cell 74 : 597–608 2. Brun A, Passant U (1996) Frontal lobe degeneration of nonAlzheimer type. Structural characteristics, diagnostic criteria and relation to other frontotemporal dementia. Acta Neurol Scand Suppl 168 : 28–30 3. Busciglio J, Hartmann H, Lorenzo A, Wong C, Baumann K, Sommer B, Staufenbiel M, Yankner BA (1997) Neuronal localization of presenilin-1 and association with amyloid plaques and neurofibrillary tangles in Alzheimer’s disease. J Neurosci 17 : 5101–5107 4. Chui DH, Shirotani K, Tanahashi H, Akiyama H, Ozawa K, Kunishita T, Takahashi K, Makifuchi T, Tabira T (1998) Both N-terminal and C-terminal fragments of presenilin 1 colocalize with neurofibrillary tangles in neurons and dystrophic neurites of senile plaques in Alzheimer’s disease. J Neurosci Res 53 : 99–106 5. Cook DG, Sung JC, Golde TE, Felsenstein KM, Wojczyk BS, Tanzi RE, Trojanowski JQ, Lee VM-Y, Doms RW (1996) Expression and analysis of presenilin 1 in a human neuronal system: localization in cell bodies and dendrites. Proc Natl Acad Sci USA 93 : 9223–9228 6. Cotman CW (1998) Apoptosis decision cascades and neuronal degeneration in Alzheimer’s disease. Neurobiol Aging 19 : S29–32 7. Cotman CW, Anderson AJ (1995) A potential role for apoptosis in neurodegeneration and Alzheimer’s disease. Mol Neurobiol 10 : 19–45 8. Dessi F, Charriaut-Marlangue C, Khrestchatisky M, Ben-Ari Y (1993) Glutamate-induced neuronal death is not a programmed cell death in cerebellar culture. J Neurochem 60 : 1953–1955 9. Drache B, Diehl GE, Beyreuther K, Perlmutter LS, König G (1998) Bcl-xl-specific antibody labels acted microglia associated with Alzheimer’s disease and other pathological states. J Neurosci Res 47 : 98–108 10. Duff K, Eckman C, Zehr C, Yu X, Prada C-M, Pérez-Tur C, Hutton M, Buée L, Harigaya Y, Yager D, Morgan D, Gordon MN, Holcomb L, Refolo L, Zenk B, Hardy J, Younkin S (1996) Increased amyloid-beta42(43) in brains of mice expressing mutant presenilin 1. Nature 383 : 710–713 11. Folstein MF, Folstein SE, McHugh PR (1975) Mini-Mental State: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12 : 189–198 12. Forloni G, Chiesa R, Smiroldo S, Verga L, Salmona M, Tagliavini F, Angeretti N (1993) Apoptosis mediated neurotoxicity induced by chronic application of β-amyloid fragment 25–35. NeuroReport 4 : 532–536
148 13. Garcia I, Martinou I, Tsujimoto Y, Martinou J-C (1992) Prevention of programmed cell death of sympathetic neurons by the bcl-2 proto-oncogene. Science 258 : 302–304 14. Giannakopoulos P, Hof PR, Bouras C (1995) Dementia lacking distinctive histology: clinicopathological evaluation of 32 cases. Acta Neuropathol 89 : 346–355 15. Giannakopoulos P, Bouras C, Kövari E, Shioi J, Tezapsidis N, Hof PR, Robakis NK (1997) Presenilin-1-immunoreactive neurons are preserved in late-onset Alzheimer disease. Am J Pathol 150 : 429–436 16. Gonzàlez-Garcìa M, Garcìa I, Ding L, O’shea S, Boise LH, Thompson CB, Nùnez G (1995) Bcl-x is expressed in embryonic and postnatal neural tissues and functions to prevent neuronal cell death. Proc Natl Acad Sci USA 92 : 4304–4308 17. Grunberg J, Walter J, Loetscher H, Deuschle U, Jacobsen H, Haass C (1998) Alzheimer’s disease associated presenilin-1 holoprotein and its 18–20 kDa C-terminal fragment are death substrates for proteases of the caspase family. Biochemistry 37 : 2263–2270 18. Gschwind M, Huber G (1995) Apoptotic cell death induced by β-amyloid (1–42) peptide is cell type dependent. J Neurochem 65 : 292–300 19. Guo Q, Sopher BL, Furukawa K, Pham DG, Robinson N, Martin GM, Mattson MP (1997) Alzheimer’s presenilin mutation sensitizes neural cells to apoptosis induced by trophic factor withdrawal and amyloid beta-peptide: involvement of calcium and oxyradicals. J Neurosci 17 : 4212–4222 20. Harrigan MR, Kunkel DD, Nguyen LB, Malouf AT (1995) Beta amyloid is neurotoxic in hippocampal slice cultures. Neurobiol Aging 16 : 779–789 21. Hendricks L, De Jonghe C, Lubke U, Woodrow S, Vanderhoeven I, Boons J, Cras P, Martin JJ, Van Broeckhoven C (1998) Immunoreactivity of presenilin-1 and tau in Alzheimer’s disease brain. Exp Neurol 149 : 341–348 22. Heyman A, Wilkinson WE, Hurwitz BL, Helms MJ, Haynes CS, Utley CM, Gwyther LP (1987) Early-onset Alzheimer’s disease: clinical predictors of institutionalization and death. Neurology 37 : 980–984 23. Huynh DP, Vinters HV, Ho DH, Ho VV, Pulst SM (1997) Neuronal expression and intracellular localization of presenilins in normal and Alzheimer disease brains. J Neuropathol Exp Neurol 56 : 1009–1017 24. Kane DJ, Sarafian TA, Anton R, Hahn H, Gralla EB, Valentine JS, Ord T, Bredesen DE (1993) Bcl-2 inhibition of neural death: decreased generation of reactive oxygen species. Science 262 : 1274–1277 25. Keller JN, Guo Q, Holtsberg FW, Bruce-Keller AJ, Mattson MP (1998) Increased sensitivity to mitochondrial toxin-induced apoptosis in neural cells expressing mutant presenilin-1 is linked to perturbed calcium homeostasis and enhanced oxyradical production. J Neurosci 18 : 4439–4450 26. Kim TW, Pettingell WH, Jung YK, Kovacs DM, Tanzi RE (1997) Alternative cleavage of Alzheimer-associated presenilins during apoptosis by a caspase-3 family protease. Science 277 : 373–376 27. Kitamura Y, Shimohama S, Kamoshima W, Ota T, Matsuoka Y, Nomura Y, Smith MA, Perry G, Whitehouse PJ, Taniguchi T (1998) Alteration of proteins regulating apoptosis, Bcl-2, Bcl-x, Bax, Bak, Bad, ICH-1 and CPP32 in Alzheimer’s disease. Brain Res 780 : 260–269 28. Korsmeyer SJ, Shutter JR, Veis DJ, Merry DE, Oltvai ZN (1993) Bcl-2/Bax: a rheostat that regulates an anti-oxidant pathway and cell death. Semin Cancer Biol 4 : 327–332 29. MacGibbon GA, Lawlor PA, Sirimanne ES, Walton MR, Connor B, Young D, Williams C, Gluckman P, Faull RL, Hughes P, Dragunow M (1997) Bax expression in mammalian neurons undergoing apoptosis, and in Alzheimer’s disease hippocampus. Brain Res 750 : 223–234
30. Martinou JC, Dubois-Dauphin M, Staple JK, Rodriguez I, Frankowski H, Missotten M, Albertini P, Talabot D, Catsicas S, Pietra C, Huarte J (1994) Overexpression of BCL-2 in transgenic mice protects neurons from naturally occuring cell death and experimental ischemia. Neuron 13 : 1017–1030 31. Mirra SS, Heyman A, McKeel D, Sumi SM, Crain BJ, Brownlee LM, Vogel FS, Hughes JP, Belle G van, Berg L (1991) The Consortium to Establish a Registry for Alzheimer’s disease (CERAD). Part II. Standardization of the neuropathologic assessment of Alzheimer disease. Neurology 41 : 479–486 32. Moussaoui S, Czech C, Pradier L, Blanchard V, Bonici B, Gohin M, Imperato A, Revah F (1996) Immunohistochemical analysis of presenilin-1 expression in the mouse brain. FEBS Lett 383 : 219–222 33. Nagy ZS, Esiri MM (1997) Apoptosis-related protein expression in the hippocampus in Alzheimer’s disease. Neurobiol Aging 18 : 565–571 34. O’Barr S, Schultz J, Rogers J (1996) Expression of the protooncogene bcl-2 in Alzheimer’s disease brain. Neurobiol Aging 17 : 131–136 35. Oltvai ZN, Milliman CL, Korsmeyer SJ (1993) Bcl-2 heterodimerizes in vivo with a conserved homolog, Bax, that accelerates programmed cell death. Cell 74 : 609–619 36. Paradis E, Douillard H, Koutroumanis M, Goodyer C, LeBlanc A (1996) Amyloid beta peptide of Alzheimer’s disease downregulates Bcl-2 and upregulates bax expression in human neurons. J Neurosci 16 : 7533–7539 37. Robakis NK (1994) β-amyloid and amyloid precursor protein: chemistry, molecular biology, and neuropathology. In: Terry RD, Katzman R, Bick KL (eds) Alzheimer’s disease. Raven Press, New York, pp 317–326 38. Roperch JP, Alvaro V, Prieur S, Tuynder M, Nemani M, Lethrosne F, Piouffre L, Gendron MC, Israeli D, Dausset J, Oren M, Amson R, Telerman A (1998) Inhibition of presenilin 1 expression is promoted by p53 and p21WAF-1 and results in apoptosis and tumor suppression. Nat Med 4 : 835–838 39. Satou T, Cummings CW, Cotman CW (1995) Immunoreactivity for Bcl-2 protein within neurons in the Alzheimer’s disease brain increases with disease severity. Brain Res 697 : 35–43 40. Schnell SA, Wessendorf MW (1998) Reduction of lipofuscin autofluorescence in monkey tissue sections. Soc Neurosci Abstr 24 : 1062 41. Shen J, Bronson RT, Chen DF, Xia W, Selkoe DJ, Tonegawa S (1997) Skeletal and CNS defects in presenilin-1 deficient mice. Cell 89 : 629–639 42. Smale G, Nichols NR, Brady DR, Finch CE, Horton WE Jr (1995) Evidence for apoptotic cell death in Alzheimer’s disease. Exp Neurol 133 : 225–230 43. Stadelmann C, Bruck W, Bancher C, Jellinger K, Lassmann H (1998) Alzheimer disease: DNA fragmentation indicates increased neuronal vulnerability, but not apoptosis. J Neuropathol Exp Neurol 57 : 456–464 44. Sterio DC (1984) The unbiased estimation of number and sizes of arbitrary particles using the disector. J Microsc 134 : 127– 136 45. Su JH, Anderson AJ, Cummings BJ, Cotman CW (1994) Immunohistochemical evidence for apoptosis in Alzheimer’s disease. NeuroReport 5 : 2529–2533 46. Su JH, Satou T, Anderson AJ, Cotman CW (1996) Upregulation of Bcl-2 is associated with neuronal DNA damage in Alzheimer’s disease. NeuroReport 7 : 437–440 47. Su JH, Deng G, Cotman CW (1997) Bax protein expression is increased in Alzheimer’s brain: correlations with DNA damage, Bcl-2 expression, and brain pathology. J Neuropathol Exp Neurol 56 : 86–93 48. Tanimukai H, Imaizumi K, Kudo T, Katayama T, Tsuda M, Takagi T, Tohyama M, Takeda M (1998) Alzheimer-associated presenilin-1 gene is induced in gerbil hippocampus after transient ischemia. Mol Brain Res 54 : 212–218 49. Tortosa A, Lòpez E, Ferrer I (1998) Bcl-2 and Bax protein expression in Alzheimer’s disease. Acta Neuropathol 95 : 407– 412
149 50. Uchihara T, El Hachimi HK, Duyckaerts C, Foncin JF, Fraser PE, Levesque L, St George-Hyslop PH, Hauw JJ (1996) Widespread immunoreactivity of presenilin in neurons of normal and Alzheimer’s disease brains: double labeling immunohistochemical study. Acta Neuropathol 92 : 325–330 51. Vallet PG, Guntern R, Hof PR, Golaz J, Delacourte A, Robakis NK, Bouras C (1992) A comparative study of histological and immunohistochemical methods for neurofibrillary tangles and senile plaques in Alzheimer’s disease. Acta Neuropathol 83 : 170–178
52. Yin XM, Oltvai ZN, Korsmeyer SJ (1994) BH1 and BH2 domains of Bcl-2 are required for inhibition of apoptosis and heterodimerization with Bax. Nature 369 : 321–323 53. Zhong LT, Sarafian T, Kane DJ, Charles AC, Mah SP, Edwards RH, Bredesen DE (1993) bcl-2 inhibits death of central neural cells induced by multiple agents. Proc Natl Acad Sci USA 90 : 4533–4537
