Subpopulations of dystrophic neuritis in Alzheimer's ... - ScienceDirect

8 downloads 14932 Views 8MB Size Report
visible or very weak with Bielschowsky silver stain. Of DNs continuous with neurofibrillary tangles (NFTs), tau-1/PHF-1 double-labeled DNs were continuous with ...
BRAIN RESEARCH ELSEVIER

Brain Research 637 (1994) 37-44

Research Report

Subpopulations of dystrophic neuritis in Alzheimer's brain with distinct immunocytochemical and argentophilic characteristics Joseph H. Su, Brian J. Cummings, Carl W. Cotman * Irvine Research Umt m Brain Aging, University of Cahforn~a, Irvme, Irvme, CA 92717-4550, USA (Accepted 28 September 1993)

Abstract

Using two monoclonal antibodies, tau-1 and PHF-1, and a sequential staining method combining double-labeling immunofluorescence and Bielschowsky silver staining, we have demonstrated the presence of two populations of dystrophic neurites (DNs) with distinct immunocytochemical and argentophilic characteristics. Tau-1 and PHF-1 immunoreactivity were co-localized in many DNs. However, approximately 20% of the DNs were immunoreactive for PHF-1 only. PHF-1 single-labeled DNs were not visible or very weak with Bielschowsky silver stain. Of DNs continuous with neurofibrillary tangles (NFTs), tau-1/PHF-1 double-labeled DNs were continuous with intracellular NFTs only, while PHF-1 single-labeled DNs were continuous with extracellular NFTs only. Furthermore, the population of DNs that cluster around extraceUular NFTs is different from those that cluster around or within senile plaques. The combined use of tau-1 and PHF-1 immunocytochemistry may provide a more accurate indication of the number of extracellular DNs and extracellular NFTs, which may aid in the diagnosis of severe and advanced AD cases. Key words: Dystrophic neurite; Tau; Paired helical filament; Argentophilia; Immunocytochemistry; Neuropathology; Alzheimer's disease

1. Introduction

The number of dystrophic neurites (DNs) and neurofibrillary tangles (NFTs) in brain tissue affected by Alzheimer's disease (AD) has recently been reported to correlate with the degree of clinical dementia [20]. Accordingly, understanding the progression of pathology and the possible interactions between DNs and NFTs is essential. It appears that DNs arise from abnormal dendrites [13,19] and axons [32] of degenerating neurons that have developed NFTs [24]. NFTs appear to undergo several stages of maturation [1,32], changing from intracellular to extracellular NFTs. While it is known that neuronal DNs originate from NFTs, it is not known if DNs undergo parallel changes and mature from intracellular to extracellular DNs. Intracellular NFTs can be distinguished from extracel-

* Corresponding author. Fax: (1) (714) 725-2071. 0006-8993/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 3 ) E 1 3 6 1 - 6

lular NFTs by differential labeling with antibodies [4,15,30,40]. Since intracellular and extracellular NFTs have distinct antigenic properties, there may also exist two immunocytochemically distinct populations of DNs that are differentially related to intracellular and extracellular NFTs. It is possible to approach this problem with the combined use of various antibodies to epitopes located differentially in the amino-terminal portion and the carboxyl-terminal portion of the cytoskeletal protein tau. Tau, a microtubule-associated phosphoprotein, is the major antigenic constituent of paired helical filaments (PHF) [16,21], which are known to contribute to the structure of NFTs and DNs. Both DNs and NFTs are considered to be expressions of a widespread alteration of the neuronal cytoskeleton [29] including changes in the tau protein. DNs and NFTs originate as intracellular cytoplasmic aggregates of insoluble P H F [14] which contains a modified form of the entire, or nearly entire, tau protein [10,17,21]. When neurons containing intracellular N F T die, they are thought to

\

~ ~,~"

~

~

~i

~,,~|i ~i:~i~!!~!~ ~~"

J H Su et al ~Brain Research 637 (1994) 37-44

leave extracellular NFTs, also called ghost tangles, because of the insoluble nature of the modified tau. In the transition from intracellular NFTs to extracellular NFTs in vivo, the chemical conformation of tau changes and it appears that several tau epitopes are lost, particularly those located on the amino-terminal half of the protein [4,33,41]. This is consistent with the binding characteristics of the tau protein to PHF. Biochemical and immunological studies have demonstrated that PHF has a protease-sensitive fuzzy outer coat and a structurally more regular protease-resistant core [37]. The amino-terminal half of tau has been shown to contribute to the pronase-sensitive fuzzy coat, whereas the carboxy-terminal third appears to tightly bind with PHF, resulting in a pronase-resistant core [16,38]. Thus, it is possible that in the transition from intracellular to extracellular NFTs the amino-terminal tau epitopes are cleaved, leaving the carboxy-terminal tau epitopes intact. The antibodies (tau-1 and PHF-1) used in this study are directed against antigens associated with the aminoand carboxy-terminals of the tau molecule. The tau-1 antibody's epitope on the tau molecule is in the amino half (amino acid residues 131-149) [3,17], while PHF-1 antibody's epitope is in the carboxy-terminal half of tau [10]. Seitelberger et al. ('91) [28] and Tabaton et al. ('91) [33] have shown that intracellular NFTs are tau-1 immunoreactive while extracellular NFTs are tau-1 negative. Bondareff et al. ('90) [4] and Yamaguchi et al. ('91) [41] have demonstrated that both intracellular NFTs and extracellular NFTs immunostain with the antibodies tau-C4 and 7.51, the epitopes of which are in the carboxyl third of the tau protein. Thus, NFTs can be identified as intracdlular NFTs that bind antibodies to tau epitopes on both the amino and carboxyterminal halves or as extracellular NFTs that only bind antibodies to tau epitopes on the carboxy-terminal half [4,26,33,41]. Extracellular NFTs can also be distinguished from intracellular NFTs by their differential reactivity to Bielschowsky's silver stain and antibodies to glial fibril-

39

lary acidic protein (GFAP). GFAP immunopositive glial processes can be found specifically around or within extracellular NFTs [25,40]. In addition, it has been suggested that late stage or extracellular NFTs stain weaker with Bielschowsky's silver stain [42]. Bielschowsky silver stain is commonly used in the identification of pathology in AD tissue, thus, weak or absent silver staining of NFTs in their late-stage or mature extracellular form may contribute to an underestimation of the severity or progression of the disease. Accordingly, we have also analyzed whether or not intracellular and extracellular subtypes of NFT pathology can be differentiated with the use of routine Bielschowsky silver staining. In the present study, we have investigated the possibility that there are two immunocytochemically distinct populations of DNs related to intracellular and extracellular NFTs. To accomplish this, we examined the hippocampus and entorhinal cortex of AD cases and aged control cases using double-label immunofluorescence techniques followed with Bielschowsky's silver staining of the same sections. We demonstrate the presence of two populations of DNs with distinct immunofluorescent and argentophilic characteristics. Furthermore, these two populations of DNs were differentially associated with intracellular NFTs, extracellular NFTs and senile plaques.

2. Materials and Methods Six cases of clinically and neuropathologlcally defined AD (average age, 80; average post mortem delay, 8 h) and four age-matched control cases (average age, 73; post mortem delay, 7 h) were used in this study Fresh brain tissue was fixed in 4% paraformaldehyde in 0.1 M Sorensen buffer, pH 7.3 for 24 h and stored in 20% sucrose in 0.1 M phosphate buffered saline. Samples were taken from the hippocampal formation, cut on a Vlbratome (40-60 /zm) and collected free floating in Tris-HC1 buffer (0.1 M, pH 7.4). Two monoclonal antibodies were used to detect DNs; tau-1 and PHF-1. Tau-1 monoclonal antibody is commeroally available (tau-l: clone PC1C6, Boehringer Mannhein, IN, 1:20,000). This antibody stares insoluble and abnormally phosphorylated tau on Western blots [12]. PHF-1

Fig. 1. Photomicrographs showing two populations of DNs m the AD brain. A: the AD brain contains many PHF-1 positive DNs. They consist of straight or curly fibers with variable lengths and diameters. Many of them (open arrows) are not co-localized with tau-1 lmmunostaining (shown in B). B" same section as A showing a few tau-1 positive DNs. Note most of them (arrowheads) co-localize with PHF-1 positive DNs (shown in A). C" s a m e s e c t i o n a s A and B showing Blelschowsky's silver staining of PHF-1 smgle-labeled and tau-1/PHF-I double-labeled DNs. The silver staining of DNs doubled-labeled for tau-1/PHF-1 (arrowheads) is strong, whereas that of DNs single-labeled for PHF-1 (arrows) is absent. All scale bars: 10/.~m. Fig. 2. Two populations of DNs are dlfferenUally derived from intracellular and extracellular NFTs. A: photomicrograph showing three PHF-1 positive NFTs and the DNs continuous with them. Note that the upper NFT (open arrow) has a loose appearance and is tau-1 negative (shown m B) B: same section as A showing two tau-1 posltwe NFTs and the DNs continuous with them (arrowhead). They are co-localized with PHF-1 positive NFI's and the DNs continuous with them (shown in A) C" same section as A and B counterstained with Bisbenzimide showing that the two NFTs doubled-labeled for tau-1/PHF-1 have nuclei (arrowheads), whereas the nucleus of the NFT single-labeled with PHF-1 is not detectable (open arrow). D' same section as A, B, and C showing the Bielschowsky's silver staining of two populations of NFI's and DNs. Two NFTs, and the DNs continuous with them, double-labeled for tau-1/PHF-1 exhibit strong silver staining (arrowheads), whereas the NFT, and the DN continuous with it, single-labeled for PHF-1 (shown in A) is not argentophdic (arrow). All scale bars: 10 #m.

~L

J.H. Suet al. / Bram Research 637 (1994) 37-44 monoclonal antibody (1. 800) was kindly provided by Dr S Greenberg. This antibody recognizes an abnormally phosphorylated epitope on PHF-tau protein, but not normal adult human tau [9]. Additionally, antlbo&es &rected against residues 1-42 of the 13amyloid protein (1:1000) (B. Cummings; personal commumcation) and against glia fibrillary acidic protein (GFAP) (Dako, CA, 1. 1000) were used to identify senile plaques and astrocytes, respectively Tissue sections were immunostained using indirect fluorescence techniques. Sections were incubated with the primary antibody for 18-24 h at room temperature. Before incubation with tau-1 antibody, sections were pretreated with 400 /zg/ml of alkaline phosphatase (VII-L, Sigma, MO) in 0.1 M TrIs-HCL, pH 8.0 for 1.5-2 h at 37°C to enhance staining [11]. No change in staining pattern or intensity was observed with increased phosphatase concentration [11]. Prolonged incubation also didn't appear to affect staining intensity. Bound antibodies were detected using AMCA-conjugated IgG, FITC-conjugated IgG, TRITC-conjugated IgG (Vector Labs, CA), and CY3-1abeled IgG (red fluorescence, Jackson ImmunoResearch, PA). In double or triple immunofluorescent staining using primary antibodies from the same species, after localizing the first antigen, the section was immersed in 37% formaldehyde solution for 2-2.5 h to mmtm~ze possible crossreactwlty. Paraformaldehyde has been shown to selectwely destroy the antigen-binding sites of second layer anti-IgG antibo&es, while having httle or no effect on the color or intensity of the initial fluorochrome [34]. Sections incubated in parallel without primary anUbody failed to develop specific staining No significant second fluorochrome labehng was seen when the second primary antibody from the same species was omitted from the second staining cycle The same results were observed by repeat experiments mverting the order both of the primary antibodies and the secondary fluorochromes Some sections were counterstalned with Bisbenzimlde (blue fluorescence; 00001%, pH 7.4) [27] to identify the nucleus of NFTs Bielschowsky's staining, performed according to Yamamoto and Hirano modification [42] was also used after immunofluorescence on some ~mmuno-stalned sections to detect DNs, NFTs and senile plaques. In these cases, the immunofluorescence-labeled DNs were photographed prior to silver staining

3. Results 3.1. DNs were immunopositive for tau-1 and PHF-1 Both tau-1 a n d P H F - 1 a n t i b o d i e s consistently imm u n o l a b e l e d n u m e r o u s D N s in the n e u r o p i l of A D brain, a n d D N s l a b e l e d for tau-1 were rare w h e n i m m u n o s t a i n i n g sections w i t h o u t prior p h o s p h a t a s e t r e a m e n t . I m m u n o r e a c t i v e D N s were identified as straight or curly fibers with a r a n g e of diameters. Most of these n e u r i t e s a p p e a r e d to be i n d e p e n d e n t of o t h e r structures, w h e r e a s some were c o n t i n u o u s with N F T s

41

or densely clustered a r o u n d or within both extracellular N F T s a n d senile plaques. T h e s e observations are c o n s i s t e n t with those of previous studies [2,5,13,18, 35,41]. I n this report, we refer to D N s in the n e u r o p i l that are n o t associated with o t h e r structures as indep e n d e n t DNs. D N s in the n e u r o p i l that are located a r o u n d a n d within N F T s a n d senile p l a q u e s are referred to as b e i n g clustered a r o u n d or within these structures. D N s that are a t t a c h e d a n d c o n t i n u o u s with N F T s are r e f e r r e d to as c o n t i n u o u s DNs. D o u b l e i m m u n o s t a i n i n g for tau-1 a n d P H F - 1 showed that tau-1 a n d P H F - 1 i m m u n o r e a c t i v i t y were co-localized in m a n y DNs. However, approximately 20% of the D N s were i m m u n o r e a c t i v e for P H F - 1 only (Fig. 1A,B). D N s exhibiting positive i m m u n o s t a i n i n g for tau-1 but not P H F - 1 were rare. Both t a u - 1 / P H F - 1 doublel a b e l e d D N s a n d P H F - 1 single-labeled D N s consisted of straight or curly fibers with variable lengths a n d diameters, thus there were n o obvious morphological differences b e t w e e n tau- 1/ P H F - 1 d o u b l e - l a b e l e d D N s a n d P H F - 1 single-labeled DNs. Bielschowsky's staining of previously i m m u n o s t a i n e d sections showed that the a r g e n t o p h i l i a of D N s d o u b l e - l a b e l e d with t a u - 1 / P H F - 1 was strong, w h e r e a s that of D N s single-labeled with P H F - 1 was usually weak or a b s e n t (Fig. 1C).

3.2. Two populations of DNs are contmuous with two populations o f NFTs E x a m i n a t i o n of D N s c o n t i n u o u s with N F T s revealed that D N s single-labeled for P H F - 1 were only c o n t i n u ous with NVFs which were also single-labeled for P H F 1. Such N F T s usually had a loose a p p e a r a n c e (Fig. 2A,B). T h e same section c o u n t e r s t a i n e d with Bisbenzimide prior to Bielschowsky's staining, revealed that the nuclei of such loose N F T s were usually not detectable (Fig. 2C) a n d the a r g e n t o p h i l i a of these N F T s were weak or a b s e n t (Fig. 2D) indicating that these were extracellular NFTs. Triple i m m u n o f l u o r e s c e n c e for tau-1, P H F - 1 a n d G F A P showed that the N F T s single-labeled for P H F - 1 were also closely associated with G F A P - p o s i t i v e astrocytes (not shown). I n contrast, D N s d o u b l e - l a b e l e d for t a u - 1 / P H F - 1 were only c o n t i n u o u s with tau- 1/ P H F - 1 d o u b l e - l a b e l e d N F T s (Fig. 2A,B). T h e nuclei of these N F T s were

Fig. 3 Photomicrographs showing the populations of DNs clustered around and within a senile plaque. A PHF-1 positive DNs clustered around and within a senile plaque. A few DNs (open arrow) are not co-localized with tau-1 lmmunostaining (shown in B). B: same section as A showing tau-1 positive DNs Many of them co-localize with PHF-1 (curved arrows; shown in A). Note that one tau-1 positive DN is not co-localized with PHF-1 (arrowhead). C: same section as In A and B counterstained with Bielsehowsky's silver stain. Note that DNs double-labeled for tau-1/PHF-1 (curved arrows; shown In A and B) and DN single-labeled for tau-1 (arrowhead; shown in B) exhibit strong silver staining, whereas DNs single-labeled for PHF-1 (arrow, shown in A) are not argentophilic All scale bars: 10/~m Fig. 4. Photomicrographs showing DNs clustered around an extracellular NFT A. PHF-1 positive DNs (arrowheads) clustered around an NFT B: same section as A showing tau-1 positive DNs (arrowheads) clustered around the NFT Note tau-1 positive DNs co-localize with PHF-1 (shown in A) and the NFT is positive only for PHF-1 indicating that it is an extracellular NFT. All scale bars' 10 p~m.

42

J H Suet a l / Brain Re~earch 637 (1994) 37-44

detectable with Bisbenzimide and these NFTs were strongly argentophilic (Fig. 2C,D), indicating that these were intraceilular NFTs. 3.3. DNs clustered around or wtthm senile plaques and extracellular NFTs

Both DNs single-labeled for PHF-1 and doublelabeled for t a u - l / P H F - 1 were clustered around and within senile plaques (Fig. 3A,B). Like independent DNs in the neuropll, the number of double-labeled DNs was greater than that of single-labeled DNs and the argentophilia of DNs single-labeled with PHF-1 was weak or absent (Fig. 3C). In addition, DNs which were single-labeled for tau-1 could be detected near or within senile plaques (Fig. 3A,B). In contrast to the variety of DNs clustered around and within senile plaques, only DNs double-labeled for t a u - 1 / P H F - 1 were found to be clustered around extracellular NFTs (Fig. 4A,B). 3.4. DNs m the control bram

Within the hippocampal formation of normal aged brain, DNs double-labeled for t a u - 1 / P H F - 1 in the neuropll were much less in number as compared to DNs in AD cases. DNs single-labeled for either tau-1 or PHF-1 were rare. T a u - 1 / P H F - 1 double-labeled DNs clustered around extracellular NFTs were not detected in control brains. Very few t a u - 1 / P H F - 1 doublelabeled DNs and even fewer PHF-1 single-labeled DNs were observed clustered around or within senile plaques in control brains. DNs that were continuous with NFTs were observed in control tissue, but with a much lower rate of incidence. The pattern of immunoreactivity was identical to that observed in Alzheimer's tissue in all respects.

4. Discussion

In this study, we were able to differentiate between two populations of DNs based on their PHF-1 and tau-1 immunostaining. In general, DNs that were continuous with intracellular NFTs were immunopositive for both tau-1 and PHF-1 while DNs that were continuous with extracellular NFTs were immunopositive for PHF-1 only. This implies that as pathology progresses, the amino-terminal portion of tau is cleaved or degraded (i.e., the epitopes recognized by tau-1) while the carboxy-terminal portion of tau remains (i.e., the epitopes recognized by PHF-1). Furthermore, we demonstrated that the population of DNs clustered around extracellular NFTs is different from that of DNs clustered around and within senile plaques. We also demonstrated that intracellular NFTs and t a u - 1 / PHF-1 double-labeled DNs stain strongly with the

modified Bielschowsky Sliver stain while the silver staining of extracellular NFTs and PHF-1 single-labeled DNs was weak or absent. NFTs are generally considered to be either lntracellular or extracellular NFTs. When neurons first become abnormal in neurodegenerative diseases, intracellular NFTs develop as cytoskeletal changes take place. The neurons eventually die and the intracellular NFTs are transformed into extracellular NFTs as soluble cellular debris are cleared away. The following observations from the present study suggest that two populations of DNs are differentially associated with intracellular and extracellular NFTs: (1) T a u - 1 / P H F - 1 doubled-labeled NFTs were compact, strongly argentophilic, and their nuclei were detectable with Bisbenzimide (a fluorescent dye which stains only nuclei) [27], indicating that these are intracellular NFTs. (2) PHF-1 single-labeled NFTs were usually loose in appearance, weak- or non-argentophllic, and reactive with G F A P antibodies. In addition, no nuclei of PHF-1 single-labeled NFTs were detectable with Bisbenzimide. This indicates that these NFTs were extracellular NFTs. (3) PHF-1 single-labeled DNs were only continuous with PHF-1 single-labeled extracellular NFTs. T a u - 1 / PHF-1 double-labeled DNs were only continuous with t a u - 1 / P H F - 1 double-labeled intracellular NFTs. One explanation for such a difference in immunoreactivity may be that as in NFTs, the tau protein is modified in DNs. Tau-l lmmunoreactivity is lost as lntracetlular NFTs are transformed to extracellular NFTs, possibly because the tau protein is cleaved in this process. It follows that DNs undergo a parallel maturation process, losing tau-1 immunoreactivity as they mature. Based on this and the observations outlined above, we suggest that DNs can be subdivided into intracellular DNs which are double-labeled for t a u - 1 / P H F - 1 antibodies and extracellular DNs which are single-labeled for PHF-1 antibody. Furthermore, we suggest that these two populations of DNs are associated with intracellular and extracellular NFTs, respectively. Our results are consistent with previous reports that suggest that DNs cluster around and within senile plaques and extracellular NFTs. It has been suggested that extracellular NFTs and senile plaques contain biologically active substances that attract or promote the growth of neurites around them as well as toxic substances that trigger or promote the formation of NFTs. Several biologically active substances have been reported to be co-localized with extracellular NFTs and senile plaques, including /3-amyloid protein, basic fibroblast growth factor (bFGF), heparan sulfate glycosaminoglycan (HSGAG), and catabolic proteinases [6-8,22,31] b F G F and H S G A G have been reported to

J H Suet al. /Bram Research 637 (1994) 37-44

promote neurite outgrowth [8,31] and it has been suggested that/3-amyloid and catabolic proteinases induce neuritic degeneration and hydrolysis of the tau protein [7,23]. Thus, it appears that neurites are first attracted to grow and form a cluster around extracellular NFTs and senile plaques, then become dystrophic and develop into intracellular DNs. Our results show that the population of DNs that cluster around extracellular NFTs is different that those clustered around senile plaques. Both intracellular and extracellular DNs cluster around and within senile plaques, while only intracellular DNs cluster around extracellular NFTs. It appears that the intracellular DNs that cluster around and within senile plaques undergo further degeneration and transform into extracellular DNs. There are several possibilities why only intracellular DNs cluster around extracellular NFTs. One hypothesis is that extracellular tangles induce tau and PHF in neuronal processes surounding the tangle, but that process or neuron does not degenerate to form extracellular residue. It is also possible that as the neuron from which these clustering DNs originate degenerates, the processes are retracted (including those clustered around the extracellular NFT) before the neuron dies and is transformed to an extracellular NFT. Another possibility is that as the neuron from which the clustered DNs originate degenerates, the neurites lose both tau-1 and PHF-1 epitopes on the tau protein and remained unlabeled in the present study. If tau is not tightly bound to PHF in newly formed neurites, both the carboxy and amino terminal portions of tau may be vulnerable to proteases and as a result the DNs would not be labeled by either tau-1 or PHF-1. The difference between the populations of DNs that cluster around extracellular NFTs and senile plaques may be a result of differences in the biologically active substances in senile plaques versus the extracellular environment around NFTs that induce the degeneration and degradation of the DNs. In support of this hypothesis, recent results suggest that the /3-amyloid protein present in extracellular NFTs is in a different conformation than that present in senile plaques [22]. Thus, it may be that the different conformations of /3-amyloid protein in combination with different proteinases contained in extracellular NFTs versus senile plaques has different affects on the neurites clustered around them. In effect, /3-amyloid may have a role in the direction or the extent of the degeneration of the neurites. Our results showing that the Bielschowsky silver staining of extracellular DNs and extracellular NFTs is weak or absent is significant in light of the fact that Bielschowsky's silver stain is one of the widely-used techniques for the detection and diagnosis of pathological lesions in AD. There is a significant correlation

43

between the number of DNs and NFTs and the degree of clinical dementia, suggesting that DNs and NFTs play a significant role in the expression of dementia in AD [20,36]. It has been shown that compared to other routine staining, Bielschowsky's silver stain is one of the most sensitive methods for detecting the presence of NFTs and senile plaques [39,42]. However, our results indicate that mature or late-stage extracellular DNs and extracellular NFTs may not be detected with Bielschowsky's silver stain. It should be noted that when NFTs become extracellular, they lose not only the amino terminal but also a portion of the carboxy terminal of tau protein [33]. Thus, it is possible that an additional small subset of extracellular NFTs is not detected with PHF-1 antibody. Therefore, we suggest that more than 20% of NFTs and DNs remain undetected when using routine Bielschowsky's silver staining. This underdetection of pathology should be considered when solely using Bielschowsky's silver stain to detect neuropathological changes for research interest or to evaluate the severity of disease. The combined use of tau-1 and PHF-1 immunocytochemistry may provide a more accurate indication of the number of extracellular DNs and extracellular NFTs, which may then be used as a criteria for the accurate diagnosis of severe and advanced AD cases. Acknowledgments. The authors would hke to thank Dr. J S Kahle for comments on this manuscript and Dr. S. Greenberg for providing the antibody PHF-1 This work was supported by Grants AG0538 and AG07918.

5. References [1] Bancher, C, Brunner, C., Lassmann, H., Budka, H., Jelhnger, K., Wlche, G., Seiteberger, F , Grundke-Iqbal, I., Iqbal, K and Wlsnlewskl, H.M., Accumulation of abnormally phosphorylated tau precedes the formation of neurofibrlllary tangles Jn Alzhelmer's disease, Brain Res, 477 (1989) 90-99. [2] Barcikowska, M., Wismewskl, H.M., Bancher, C and GrundkeIqbal, I , About the presence of paired helical filaments in dystrophic neurites partlopating m the plaque formation, Acta Neuropathol (Berl.), 78 (1989) 225-231. [3] Binder, L I, Frankfurterer, A. and Rebhun, L.I, The &stributlon of tau m the mammalian central nervous system, J Cell Btol. 101 (1985) 1371-1378 [4l Bondareff, W., Wlschlk, C.M., Novak, M., Amos, W.B., Klug, A. and Roth, M., Molecular analysis of neurofibrdlary degenerahon in Alzhelmer's disease. An Immunohistochemical study, Am J Pathol, 137 (1990)711-723. [5] Braak, H , Braak, E , Grundke-Iqbal, I. and Iqbal, K, Occurrence of neuropil threads m the semle human brain and m Alzhelmer's disease: a thtrd location of paired helical filaments outside of neurofibrdlary tangles and neuritic plaques, Neurosct Lett, 65 (1986) 351-355. [6] Cataldo, A.M. and NLxon, R.A., Enzymattcally actwe lysosomal proteases are associated with amyloid deposits m AIzhelmer's brain, Proc Natl Acad. Scl USA, 87 (1990)3861-3865. [7] Cataldo, A M., Thayer, E.D., Bird, E.D, Wheelock, T R. and NLxon, R A , Lysosomal protemase are prominently locahzed

44

J H Sit et al /Brain Re,search 637 (1994) 37-44

within senile plaques of Alzhelmer's disease evidence for a neuronal origin, Brain Res, 513 (1990) 181-192 [8] G6mez-Plmlla, F . Cummings, B J and Cotman. C W, Induction of basic fibroblast growth factor in Alzhelmer's disease pathology, NeuroReport, 1 (1990) 211-214 [9] Greenberg, S.G and Davies, P, A preparation of Alzhelmer paired helical filaments that displays distract tau proteins by polyacrylamlde gel electrophoresls, Proc Natl Acad Sct USA, 87 (1990) 5827-5831 [10] Greenberg, S G . Davies, P, Scheln, J D and Binder, L I, Hydrofluoric acid-treated tau PHF proteins display the same biochemical properties as normal tau, J Blol Chem. 267 (1992) 564-569. [11] Grundke-Iqbal. I. lqbal. K, Tung. Y - C , Qulnlan, M, Wlsnlewskl, H M and Binder, L I, Abnormal phosphorylatlon of the mlcrotubule-assoclated protein ~" (tau) m Alzhelmer cytoskeletal pathology. Proc Natl Acad Scz USA. 83 (1986) 4913-4917 [12] Hanger. D P , Brlon, J.-P, Gallo, J - M , Cairns. N.L, Luther. P J and Anderton, B H , Tau in Alzhelmer's disease and Down's syndrome Is insoluble and abnormally phosphorylated. Btochem J, 275 (1991) 99-104 [13] lhara. Y, Massive somatodendrmc sprouting of cortical neurons In Alzhelmer's disease, Bram Res, 459 (1988) 138-144 [14] Ihara, Y. Abraham, C and Selkoe, D J , Antibodies to paired hehcal filaments in Alzhelmer's disease do not recognize normal brain protein, Nature, 304 (1983) 727-730 [15] Joachim, C.L, Morns, J H , Koslk, KS and Selkoe, D J , Tau antlsera recogmzes neurofibrlllary tangles in a range of neurogeneratwe disorders, Ann Neurol, 22 (1987) 514-520 [16] Kondo, J , Honda, T. Morl, H , Hamada, Y, Mmra, R , Ogawara, M and lhara, Y, The carboxyl third of tau is tightly bound to paired helical filaments, Neuron, 1 (1988) 827-834 [17] Koslk, K S., Orecchlo, L D , Binder, L, Trojanowskl, J Q , Lee, V.M-Y and Lee, G , Epltopes that span the tau molecule are shared with paired helical fdaments, Neuron, 1 (1988) 817-925 [18] Kowall, N W and Koslk, K S, Axonal disruption and aberrant locahzatlon ot tau protein characterize the neuropfl pathology of Alzhelmer's disease. Ann Neurol, 22 (1987) 639-643 [19] Mckee, A N, Kowall, N W and Koslk, K S, Mlcrotubular reorganization and dendritic growth response in Alzhelmer's disease, Anti Neurol, 26 (1989) 652-659 [20] McKee, A C. Koslk, K S and Kowall, N W , Neuritic pathology and dementia in Alzhelmer's disease. Ann Neurol. 30 (1991) 156-65 [21] Nuklna, N. and Ihara, Y., One of the antlgemc determinants of paired hehcal filaments is related to tau protein, J Btochem, 99 (1986) 1541-1544 [22] Perry, G , Cras, P, Sledlak, S.L., Tabaton, M and Kawal, M,/3 Protein lmmunoreactwlty is found in majority of neurofibnllary tangles of Alzhelmer's disease, Am J Pathol, 140 (1992) 283290 [23] Pike, C J., Cummings, B J. and Cotman, C.W, /3-Amylold induces neuritic dystrophy m vitro, similarities with Alzhelmer pathology, NeuroReport, 3 (1992) 769-772 [24] Probst, A , Anderton, B.H, Brlon, J P and Ulrich, J , Sende plaque neuntes fall to demonstrate antl-pmred helical filament and antz-mlcrotubule-assoclated protem-tau ~mmunoreactlve proteins in the absence of neurofibrfllary tangles in the neocortex, Acta Neuropathol, 77 (1989) 430-436 [25] Probst, A , Ulrich, J and Heltz, P , Senile dementm of Alzhelmer type. astroghal reaction to extracellular neurofibrfllary tangles in the hlppocampus an lmmunocytochemlcal and electron-microscopic study, Acta Neuropathol, 57 (1982) 75-79

[26] Schmldt, M L, Lee. V -Y and Trojanowskl. J Q , Comparative epltope analysis of neuronal cytoskeletal protein m Alzhelmer's disease senile plaque neurltes and neuropll threads, Lab Invest, 64 (1991) 352-357 [27] Schmued, L.C. Swanson. L W and Sawchenko, P E . Some fluorescent counterstalns for neuroanatomlcal studies. J Histochem Cytochem. 30 (1982) 123-128 [28] Seltelberger, F, Lassmann. H and Bancher, C. Cytoskeleton pathology m Alzhelmer's disease and related disorders, J Neural Transm. Suppl 33 (1991) 27-33 [29] Selkoe, D J , The molecular pathology of Alzhelmer's disease, Neuron. 6 (1991) 487-498 [30] Sledlak, S L. Cras, P, Kawal, M, Rtchey, P and Perry, G . Basic fibroblast growth factor binding is a marker for extracellular neuroflbrlllary tangles in Alzhelmer disease, J Hmochem Cytochem. 39 (1991) 899-904 [31] Su, J H , Cummings, BJ and Cotman, C W , Localization of heparan sulfate glycosammogylcan and proteoglycan core protern in aged brain and Alzhelmer's disease, Neurosctence, 51 (1992) 801-813 [32] Su, J H , Cummings, B J. and Cotman, C W, Identification and distribution of axonal dystrophic neurltes in Alzhelmer's disease, Bram Re~s, 625 (1993) 228-237 [33] Tabaton, M, Cammarata, S. Mancardl, G, Manetto, V . Autlho-Gambettl, L and Perry, G , Utrastructural locahzatlon of /3-amylold, ~-, and ublqultm epltopes m extracellular neurofibrll[ary tangles. Proc Natl Acad Sct USA, 88 (1991) 2098-2102 [34] Wang, B L and Larson. L.I, Sxmultaneous demonstranon of multiple antigens by indirect lmmunofluorescence or immunogold staining, Htstochemtstry, 83 (1985) 47-56 [35] Weghe, J V . Cras, P, Kawal, M, Sledlak, S L., Tabaton, M, Greenberg, B and Perry. G . Dystrophic neuntes infiltrate extracellular neurofibrlllary tangles in Alzhelmer disease. Bram Res, 560 (1991) 303-305, [36] Wdcock, G K and ESlrl, M M , Plaques, tangles and dementia, J Neurol Sct, 56 (1982) 343-356 [37] WlSchlk. C M. Novak, M, Edwards, P C , Klug, A , Tichelaar, W and Crowther, R A , Structural characterization of the core of the paired helical filament of Alzhelmer disease, Proc Natl Acad Scl USA. 85 (1988) 4884-4888 [38] Wlschlk, C M., Novak, M, Thogersen, H C. Edwards, P C . Runswlck, M J , Jakes, R . Walker, J E . Mllsteln, C, Roth, M and Klug, A , Isolation of a fragment of tau derwed from the core of the paired helical filament of Alzhelmer disease, Proc Natl Acad Sct USA. 85 (1988) 4506-4510. [39] WismewskL H M, Wen, G Y and Kam, K S , Comparison of four staining methods on the detection of neuntes plaques. Acta Neuropathol, 78 (1989) 22-27 [40] Yamaguchl. H , Morlmatsu, M, Hlrm, S and Takahashl, K, Alzhelmer's neurofibrlllary tangles are penetrated by astroghal processes and appear eosmophdlc in their final stages, Acta Neuropathol (Bed), 72 (1987) 214-217 [41] YamaguchL H , Nakazato. Y, Kawarabayashl, T, K, I. Ihara, Y. Morlmatsu, M and Hlral, S, Extracellular neurofibrdlary tangles asscclated with degenerating neurltes and neuropd threads in Alzhelmer-type dementia, Acta Neuropathol, 81 (1991) 603-609 [42] Yamamoto, T and Hlrano, A , A comparative study of modified Blelschowsky, Bodlan and thloflawn S stares on Alzhelmer's neurofibrlllaqry tangles, Neuropathol Appl Neurobiol, 12 (1986) 3-9