EXPERtMENTAL. NEUROLOGY Is!& 172-179 (19%)

Early Association of Reactive Astrocytes with Senile Plaques in Alzheimer’s Disease

CHRISTIAN J. PIKE,’ BRIAN J. CUMMINGS, AND CARL W. COTMAN

Organized Research Unit in Bmin Aging and Dementia, Department of Psychobiology, University of California, Irvine, California 92717-4550

The fibrillar f3-amyloid protein (Al9 plaques of Alzhei- mer’s disease (AD) are associated with reactive astro- cytes and dystrophic neurites and have been suggested to contribute to neurodegenerative events in the dis- ease. We recently reported parallel in vitro and in situ iindings, suggesting that the adoption of a reactive phenotype and the colocalization of astrocytes with plaques in AD may be mediated in large part by aggre- gated A& Thus, Ai3-mediated effects on astrocytes may directly affect disease progression by modifying the degenerative plaque environment. Alternatively, plaque-associated reactive astrocytosis may primarily represent a glial response to the neural injury associ- ated with plaques and not significantly contribute to AD pathology. To investigate the validity of these two positions, we e zamined the differential colocalization of reactive astrocytes and dystrophic neurites with plaques. Hippocampal sections from AD brains- ranging in neuropathology from mild to sever-were triple-labeled with antibodies recognizing Ap protein, reactive astrocyt’es, and dystrophic neurites. We ob- served not only plaques containing both or neither cell type, but also plaques containing (1) reactive astro- cytes but not dystrophic neurites and (2) dystrophic neurites but not reactive astrocytes. The relative pro- portion of plaques colocalized with reactive astrocytes in the absence of dystrophic neurites is relatively high in mild AD but signiikantly decreases over the course of the disease, suggesting that plaque-assoc iated astro- cytosis may be an early and perhaps contributory event in AD pathology rather than merely a response to neuronal injury. These data underscore the poten- tially signi&ant contributions of reactive astrocytosis in modifykg the plaque environment in particular and disease progression in general. 0 1896 Academic Res* I~c

IN-IRODUCTlON

The senile plaques found in the neuropil of Alzhei- mer’s disease (AD) brain are frequently associated with

l To whom correspondence should be addressed. Fax: (714) 824. 2071.

reactive glia and dystrophic neurites. This classic obser- vation, coupled with recent molecular and genetic find- ings (34, 491, provides supportive evidence for the hypothesis that the insoluble A9 deposited in plaques significantly contributes to the initiation and or progres- sion of AD neuropathology. Thus, in order to under- stand and eventually interrupt the disease process, detailed knowledge of the cellular responses to plaques appears to be necessary. Specifically, determination of the factors underlying both the colocalization of astro- cytes and neuritic processes with plaques and the patho- logical responses of these plaque-associated cells will increase our understanding of neurodegenerative events in AD and thus should facilitate therapeutic efforts.

The findings of previous studies suggest that one factor contributing to cellular association with plaques is the developmental stage of the plaque. Plaques are generally categorized into three basic types (diffuse, primitive, neuriticl, which have been theorized to repre- sent stages along a temporal continuum of plaque development. More specifically, plaques are believed to begin at the diffuse stage, which are deposits of (&uny- loid protein (A81 that lack the ordered p-sheet conforma- tion (as detected by thioflavine and Congo red stains) characteristic of amyloid (9, 291. These A8 deposits are presumed to gradually adopt a fibrillar structure charac- terized by p-sheet conformation (primitive stage) and eventually form a compact A9 core enveloped with fibrillar, p-sheet A8 (neuritic stage) (44, 62). Both dystrophic neurites (28, 31, 45, 571 and reactive astro- cytes (19,20,30,46,56,62) have been observed in only a subset of plaques, presumably localized with primitive and neuritic plaques but not diffuse plaques. Thus, the conformational state of A8 (diffuse deposits versus p-sheet fibrils) within plaques may contribute to cellular association with plaques (121. In fact, assembled but not soluble A8 peptides are associated with reactive astrocy- tosis (8,26,411, neuritic dystrophy (22, 391, and neuro- toxicity (7, 32, 38, 42, 43, 511 in vitro. Further, in quantitative analyses of triple-labeled AD brain tissue, we have recently extended these in vitro findings, dem- onstrating that both reactive astrocytes (411 and dystro- phic neurites (14) are associated, almost exclusively with

0014-4666/95$6.00 Copy-right 8 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

172

COLOCALIZATION OF ASTROCYTES AND NEURITES WITH PLAQUES 173

A8 plaques exhibiting an assembled, thioflavine-positive structure.

One interpretation of existing data is that A9, in a conformation-dependent manner, is a primary stimulus underlying the colocalization of both astrocytes and neurites with senile plaques. Accordingly, as A8 as- sembles into a P-sheet fibrillar structure, plaques evolve in parallel from diffuse into primitive and neuritic types, and local astrocytes and neurites respond to the confor- mationally altered A8, becoming increasingly involved with the plaques (12, 40). However, factors other than A8 may contribute to cellular colocalization with plaques. For example, astrocytes may colocalize with plaques and adopt a reactive phenotype largely in response to signals from degenerating neuritic processes. Conversely, neu- ritic processes may be attracted to plaques in response to trophic signals (e.g., basic fibroblast growth factor) emanating from reactive, plaque-associated astrocytes (15). By determining the temporal colocalization of reactive astrocytes and dystrophic neurites with plaques, we should be able to gain insight into the validity of these competing hypotheses. To experimentally address these questions, we have quantitatively studied neuritic and astrocytic colocalization with plaques in triple- immunolabeled brain tissue from AD cases exhibiting neuropathology ranging from mild to severe.

METHODS

Immunohistochemistry

Fifty-micrometer sections of paraformaldehyde-fixed hippocampal formation were triple-immunolabeled with minor modifications of previously described methods (16). Briefly, free-floating sections were pretreated for 5 min with 70% formic acid (to enhance A3 immunolabel- ing) and for 20 min with 1% hydrogen peroxide (to quench endogenous peroxidase). Next, we applied the first label, our anti-A8 81-42 (17) diluted 1:lOOO. All primary antibodies were applied following 2% bovine serum albumin (BSA) blocking and incubated with tissue overnight (25°C) under constant orbital shaking (50 r-pm). A8 was visualized by sequential exposure to biotinylated anti-rabbit antibody, avidin-biotin com- plex, and red alkaline phosphatase substrate, according to manufacturer’s instructions (Vector ABC alkaline phosphatase kit; Burlingame, CA). Sections were post- fixed for 2 h in 37% formaldehyde to attenuate cross- reactivity in subsequent labeling (61). The second immu- nolabel was PHF-1 (1:1600), a monoclonal antibody generously provided by Dr. Sharon Greenberg that is directed against hyperphosphorylated tau/paired helical filament (PHF) and recognizes dystrophic neurites and neurofibrillary tangles (24). PHF-1 was visualized with diaminobenzidine (Vector) after incubation in anti- mouse biotinylated antibody and avidin-biotin complex (Vector Elite anti-mouse ABC peroxidase kit). Sections

were incubated for 20 min in 1% hydrogen peroxide to quench unreacted product from the diaminobenzidine reaction. The final label was a polyclonal antibody against glial fibrillary acidic protein (GFAP 18000; DAKO, Carpinteria, CA), an astrocytic antigen that is markedly increased in reactive states (21). GFAP was visualized with SG blue (Vector), following sequential exposure to anti-rabbit antibody and avidin-biotin com- plex (Vector Elite anti-rabbit ABC peroxidase kit). Triple- labeled sections were mounted on gelatin-coated slides, air-dried, dehydrated by serial rinses in graded ethanol solutions, and cover-slipped in Depex mounting me- dium.

Case Selection

Ten brains, obtained from the O.R.U. in Brain Aging Tissue Repository (U.C.I.), were utilized in this study. All cases were neuropathologically evaluated and exhib- ited senile plaque density (via Al3 immunoreactivity) consistent with the diagnosis of probable AD; 8 of 10 cases were also clinically assessed. Age of subjects ranged from 62 to 86 years, with a mean age of 75 f 2.7 years. Postmortem delay ranged from 0.5 to 6.5 h, with a mean delay of 4.1 ~fr 0.5 h. For the purposes of some data analyses, cases were binned according to neuropathologi- cal severity (mild, N = 3; moderate, N = 3; severe, N = 4) based upon extent of neuropil dystrophic neu- rites and threads, as determined by overall levels of PHF-1 immunoreactivity; previous studies have demon- strated that neuropil dystrophic neurites correlate ex- tremely well (r m 0.9) with clinical dementia in AD patients (33, B.J.C., C.J.P., and C.W.C., unpublished observations). In those cases with clinical evaluation (81101, available neuropsychological measures were con- sistent with neuropathological categorization.

Quantification of Immunoreactivity

Quantification of colocalization of dystrophic neurites and reactive astrocytes with plaques was conducted as previously described (411, with minor modifications. First, the proportion of plaques exhibiting cellular colo- calization was compared with that lacking colocahza- tion. In each triple-labeled section, 100 8142~immuno- reactive plaques were scored immunopositive (+I if associated with GFAP and/or PI-IF-1 immunoreactivity and negative if lacking both GFAP and PHF-1 immuno- reactivity. No more than four plaques were scored from any single field; fields were taken from all plaque- containing regions (hippocampus CA fields, dentate gyrus, hilus, entorhinal cortex). In subsequent analyses, plaques exhibiting cellular colocalization were analyzed to determine the relative prevalence of plaques contain- ing (a) only GFAP( + 1 cells, (b) only PHF-l( + 1 processes, (c) both GFAP(+) and PHF-l(+). One-hundred 81- 42(+) plaques were scored in each section using the

174 PIKE, CUMMINGS. AND COTMAN

following quota for different regions of hippocampal formation: 10 dentate gyrus, 10 hilus, 10 CA3, 20 CA2/CAl, 25 medial entorhinal cortex, and 25 lateral entorhinal cortex. In cases with insufficient numbers of plaques within a particular region, additional plaques were counted in the remaining regions in a representa- tive fashion. All determinations were made simulta- neously by two observers at 200x final magnification (400 x as necessary).

RJZSU LTS

Triple-Immunolabeling Reveals Four Plaque Types

Upon initial examination of each immunolabeled AD tissue section, we observed classic pathological indica- tions of the disease, including PHF-1 immunopositive (+ 1 neurofibrillary tangles and dystrophic neurites, Ap(+) senile plaques, and GFAP(+) reactive astrocytes. Detailed examination of sections revealed heteroge-

neous populations of plaques in terms of cellular colocal- ization. Specifically, we found the two general plaque populations: diffuse types lacking cellular colocalization and classic primitive/neuritic types exhibiting both PHF- l(+) fibers and GFAP(+l reactive astrocytes (Figs. 1A and 1B). We also observed two additional, previously unreported plaque types: plaques exhibiting dystrophic neurites but not reactive astrocytes [GFAP(-)/PHF- l( + )] and plaques colocalized with reactive astrocytes but lacking dystrophic neurites [GFAP( + )/PHF-l( - )] (Figs. 1C and 1D).

Prevalence and Regional Distribution of Plaques Colocalized with GFXP and/or PHF-1

Among those plaques exhibiting cellular colocaliza- tion [GFAP(+) and/or PHF-l(+)l, the most prevalent plaque type is GFAP(+)/PHF-l(+ 1 (Fig. 2). Across all cases, the GFAP(+)/PHF-l(+) subtype accounts for 60.5 + 8.8% of plaques associated with astrocytes

FIG. 1. Triple-immunolabeling with antibodies against AP, GFAP, and PHF-1 reveals four distinct plaque subtypes. (A) GFAP-)/PHF- lf-) plaques are consistent with the diffuse subtype classification. Note the presence of PHF-l(+) fibers (arrow) and faintly GFAP(+), nonreactive astrocytes (arrowhead) nearby but not specifically colocalized with these plaques. (Bl GFAP(+)/PHF-l(+) plaques represent the classic neuritic plaque subtype. (0 GFAPC-l/PHF-l(f) plaques are relatively scarce except in the CA1 region of hippocampus. Arrowheads denote nearby, nonreactive astrocytes. (D) GFAP(+)/PHF-lf-1 plaques are particularly abundant in mild AD cases. Note the hypertrophic morphology of these astrocytes and the infiltration of the plaque by their processes. Scale bar, 50 pm.

COLOCALIZATION OF ASTROCYTES AND NEURITES WITH PLAQUES 175

II I GFAP(+)/PHF-I(+)~ iI m GFAP(+)/PHF-I(-) 1

overall

1 J-VPHF-I(+)

T T

FIG. 2. Distribution of cell-associated plaque subtypes across different regions of hippocampal formationientorhinal cortex. In general, distribution of plaque subtypes is relatively uniform except in CA1/2, where a significantly greater proportion of GFAPt-l/PHF- l(+) plaques are observed. Asterisk denotes P < 0.05, relative to overall values; between group comparisons performed by analysis of variance followed by Scheffe F test.

and/or neuronal processes. In comparison, the GFAP(+)/PHF-l(-) type comprises 29.2 2 9.2% and the GFAP( -)/PHF-l(+) type 10.3 & 2.8%. This distribu- tion of plaque types is relatively well-preserved across most regions in hippocampal formation (Fig. 2). The most striking deviation from this pattern is observed in CAl, where a significantly greater level of GFAP(-)I PHF-l(+) plaques are found in comparison to other examined regions; this increase was apparent in all cases containing CA1 plaques. A notable yet nonsignificant trend consistently observed was a relative increase in GFAP(+)/PHF-1(-J plaques (41.2 & 11.2%) in medial entorhinal cortex; a similar but more variable trend was also observed in dentate gyrus.

Differential Distribution of Plaque Types According to Disease Severity

If the progression of AD pathology is paralleled by increased cellular colocalization with plaques, then the relative proportions of the four observed plaque types may be expected to vary in accordance with disease severity. Consequently, we compared the proportions of each plaque type across AD cases with mild, moderate, and severe neuropathology, as assessed by levels of neuropil dystrophic neurites. In agreement with the above theory, we observed the highest proportion of GFAP(-)/PHF-1(-J or diffuse plaques in mild cases, lower relative levels in moderate cases, and still lower relative levels in severe cases (Fig. 3). Note that the reported values for mild AD are extremely conservative since cases showing very early AD neuropathology- which lack appreciable cellular colocalization with plaques-were excluded from this study.

In a pattern complementary to the decline of

i DG hilus CA3 CA112 EC-med EC-la

Brain Region

0

Mild Moderate Severe

AD Pathology

FIG. 3. Relative proportion of diffuse [GFAPt-)/PHF-1(-l] plaques across varying levels of AD pathology. Diffuse plaques consti- tute a decreasing proportion of the total number of plaques with increasing disease severity. Single asterisk denotes P < 0.05, relative to severe group; double asterisk denotes P < 0.05, relative to mild group. Between group comparisons performed by analysis of variance of raw data followed by Scheffe F test.

GFAP(- l/PHF-1(-l plaques, we observed an increase in the relative levels of GFAP(+)/PHF-l(+) plaques with increasing levels of pathology (Fig. 4). As stated above, across all cases approximately 60% of plaques exhibiting cellular colocalization were both GFAP( + 1 and PHF-l( + 1. However, comparison with disease sever- ity revealed that the proportion of GFAP(+ )/PHF-l(+ 1 plaques in cases with mild pathology was only (30.7 2 17.3%) and a significantly higher 80.5 2 6.0% in severe

100

0

FIG. 4.

I GFAI’(+)/PHt-l(c) * m GFAP(+)/PHF-I(-)

J overall Mild Moderate severe

AD Pathology

Distribution of cell-associated plaque subtypes across varying levels of AD pathology. Levels of GFAP(+)/PHF-l(+) are relatively low in mild AD but significantly increase to become the predominant subtype in severe AD. Conversely, GFAP(+l/PHF-1(-l plaques are the most common subtype in mild AD pathology, but significantly decrease in prevalence to become the minority subtype in severe AD pathology. GFAPt-l/PHF-l(+) plaques were observed to remain at a constant proportion throughout all levels of the disease. Single asterisk denotes P < 0.05, relative to mild GFAP(+)/PHF-l(+) group; double asterisk denotes P < 0.05, relative to mild GFAP(+)/ PHF-l( -) group. Between group comparisons performed by analysis of variance of raw data followed by Scheffe F test.

176 PIKE, CUMMINGS, AND COTh4AN

pathology cases. This proportional increase in GFAP( + ) I PHF-l(+) plaques with increasing AD severity was paralleled by a significant proportional decrease in GFAP(+)/PHF-1(-l plaques; the relative levels of GFAP(-)/PHF-l(+) plaques did not significantly vary with disease severity within the examined cases (Fig. 4).

DISCUSSION

This paper presents quantitative immunohistochemi- cal evidence that reactive astrocytes colocalize with senile plaques during the relatively early stages of AD and that this association appears to occur prior to and thus independent of the presence of PHF-l(+) dystro- phic neurites. Predictably, we observed A9 plaques that lack celldar association and others that are associated with both reactive astrocytes and dystrophic neurites. Further, we demonstrate that two additional plaque types are found: plaques associated with reactive astro- cytes but not dystrophic neurites (as defined by PHF-1 immunoreactivity) and plaques associated with dystro- phic neurites but lacking reactive neurites. The exis- tence of these latter plaque subtypes suggests that the colocalization of both reactive astrocytes and dystrophic neurites can occur independently of the presence of the other cell type.

The significance of these findings is that astrocytes appear to colocalize with plaques and adopt a reactive morphology not solely in response to cues from degener- ating neurites, but perhaps also under the influence of plaque component(s). If astrocytes associate with plaques and adopt a reactive phenotype in response to neuritic degeneration, then reactive astrocytes should be ob-

. served nearly always in the presence of dystrophic neurites but never in their absence. In contrast to this prediction, we observe appreciable numbers of both GFAP(+)/PHF-U-1 and GFAP(-)/PHF-l(+) plaques. This absence of a strict correlation between the plaque colocalizations of GFAP(+) and PHF-l(+) processes suggests that astrocytic colocalization with plaques may be driven in part by signals from one or more plaque- associated components, rather than solely from neuritic degeneration.

A potential shortcoming of our interpretation of the reported data concerns the sensitivity in detecting dys- trophic neurites. To visualize dystrophic neurites, we utilized the PHF-1 antibody (241, which recognizes hyperphosphorylated epitopes of the cytoskeletal pro- tein tau. Other markers of dystrophic neurites, for example immunoreactivities for A3 precursor protein (16, 27, 50) and specific neurotransmitters (3-5, 60), label subpopulations of dystrophic neurites. Yet, in previous studies, we have found the PHF-1 antibody to be the most sensitive and broadest marker for the detection of dystrophic neurites in confirmed AD tissue (55). However, it is likely that other markers will be

developed that may detect either earlier changes in tau (48, 54) or more generalized indications of dystrophy. Should this occur, it would be of interest to examine if astrocytic colocalization with plaques also precedes the appearance of these markers.

Since PHF-l(+) dystrophic neurites do not appear to be the primary stimulus of plaque-associated reactive astrocytosis, identification of the relevant stimuli must be extended to other plaque stimuli. One such stimulus that is a major component of all plaques and may contribute to the observed reactive astrocytosis is fibril- lar, p-sheet AP. In support of this possibility, Canning et al. and Hoke et al. (8, 26) have reported in vitro and in uivo evidence consistent with the theory that A9 contrib- utes to reactive gliosis. In addition, we recently reported that cultured astrocytes treated with fibrillar, P-sheet A3 adopt a reactive phenotype and envelop A9 aggre- gates in a manner morphologically similar to the associa- tion of reactive astrocytes with plaques in AD (41). Similarly, fibrillar AP is associated with dystrophic neurites, both in vitro (22, 39) and in AD (6, 141, suggesting that the colocalization of PHF-l( +) neurites with plaques occurs at least partially in response to AP.

Plaque component(s) in addition to A9 may also contribute to cellular colocalization with plaques. For example, neuritic fibers may associate with plaques partially in response to the presence of the neurite- stimulating factor (59) basic fibroblast growth factor (bFGF). In support of this possibility, reactive, plaque- associated astrocytes exhibit increased bFGF immunore- activity (23, 41, 531, senile plaques in AD exhibit bFGF immunoreactivity (23,53), and all such bFGF-immuno- reactive plaques are associated with dystrophic neurites (15). This possibility is consistent with our observations that particular regions of hippocampal formation (i.e., CA1 and medial entorhinal cortex) exhibit differential distributions of the cell-associated plaque subtypes. Since different brain regions are likely characterized by variations in their particular extracellular microenviron- ments, plaques may be expected to exhibit correspond- ing variations in their plaque-associated components, which in turn may modify cellular responses to plaques.

The presented data also support the theory that senile plaques proceed through a developmental continuum, ranging from A(3 deposition in a diffuse form which lacks both p-sheet conformation and cellular association, to an intermediate or primitive stage that exhibits limited evidence of A3 p-sheet conformation and cellular colocal- ization, and eventually to a classic neuritic form which includes p-sheet, fibrillar AP, and profuse neuritic and astrocytic association (44, 45, 62). In agreement with this hypothesis, we observe decreased proportions of diffuse plaques with increasing neuropathological sever- ity of the disease. Consistent with cell culture experi- ments (32, 38, 41, 43, 51), we suggest that as the A9 within difluse plaques assembles into a fibrillar form

COLOCALIZATION OF ASTROCYTES AND NEURITES WITH PLAQUES 177

with a P-sheet conformation, it becomes a stimulus which induces reactive changes in astrocytes and degen- erative changes in neurons. Thus, as AD progresses, an increasing proportion of plaques may become colocalized with cells and cellular processes. As suggested by our mildest cases, astrocytes may initially associate with primitive plaques followed by PHF-l(+) neurites; this sequence is supported by our data showing decreasing proportions of GFAP( + 1 /PHF- l( - ) plaques and increas- ing relative levels of GFAP(+)/PHF-l(+) plaques with increasing disease severity. Thus, plaque-associated re- active astrocytosis may be a relatively early event in AD which contributes to disease neuropathology rather than merely responding to it.

These data suggest that astrocytes, in addition to neurons, are directly affected by p-sheet, fibrillar A3 deposits and that these interactions can occur indepen- dently of each other. The independent and apparent early association of reactive astrocytes with plaques underscores their potential involvement in modifying the plaque environment. For example, the upregulation of bioactive substances in reactive astrocytes may ini- tiate a series of molecular cascades affecting both local neurons and microglia (11, 13). Numerous such bioac- tive substances exhibit increased expression during reactive astrocytosis (21) and are implicated in AD neuropathology, including apolipoprotein E (35, 47), ai-antichymotrypsin (1, 37), Al3 precursor protein (10, 16, 361, proteoglycans (18, 521, bFGF (23,531, interleu- kin-l (251, and transforming growth factor p (58). Perhaps in parallel, recent culture studies demonstrate that Al3 induces in astrocytes a reactive phenotype (8, 41) and can increase their expression of bFGF (2, 41), interleukin-1 (2), and chondroitin sulfate (8). In sum- mary, these data suggest that plaques constitute a degenerative environment which, over the course of the disease, appears to mediate pathological changes of progressive intensity in both neurons and astrocytes. Thus, one key in elucidating neurodegenerative events in AD is to determine further not only A@mediated effects on neurons, but also how Al3 and other plaque components affect astrocytes, and in turn, how these astrocytic responses influence disease progression.

ACKNOWLEDGMENTS

The authors thank Dr. Sharon Greenberg for providing PHF-1 antibody and Mr. Arman Afagh for technical assistance. This study was supported by a program project grant from the NL4 (AGOO538).

REFERENCES

1. ABRAHAM, C. R., D. J. SELKOE, AND H. POTTER. 1988. Immuno- chemical identification of the serine protease inhibitor oi- antichymotrypsin in the brain amyloid deposits of Alzheimer’s disease. Cell 52: 487-501.

2. ARAUJO, D. M., AND C. W. COTMAN. 1992. 6Amyioid stimulates

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

glial cells in vitro to produce growth factors that accumulate in senile plaques in Alzheimer’s disease. Bmin Res. MB: 141-145.

ARMSTRONG, D. M., W. C. BENZING, J. EVANS, R. D. TERRY, D. SHIELDS, AND L. A. HANSEN. 1989. Substance P and somatostatin coexist within neuritic plaques: implications for the pathogenesis of AIsheimer’s disease. Neuroscience 31: 663-671.

ARMSTRONG, D. M., S. LEROY, D. SHIELDS, AND R. D. TERRY. 1985. Somatostatin-Iike immunoreactivity within neuritic plaques. Brain Res. 338: 71-79.

BENZING, W. C., M. D. IKONOMOVIC, D. R. BRADY, E. J. MUFSON, AND D. M. ARMSTRONG. 1993. Evidence that transmitter- containing dystrophic neurites precede paired helical filament and AIz-50 formation within senile plaques in the amygxhda of nondemented elderly and patients with Alzheimer’s disease. J. Comp. Neural. 334: 176-191. BENZING, W. C., E. J. MUFSON, AND D. M. ARMSTRONG. 1993. Alzheimer’s disease-like dystrophic neurites characteristicaUy associated with senile plaques are not found within other neuro- degenerative diseases unless amyloid P-protein deposition is present. Brain Res. 608: 10-18. BUSCIGLIO, J., A. LORENZO, AND B. A. YANKNER. 1992. Methodologi- cai variables in the assessment of beta amyloid neurotoxicity. Neurobiol. Aging. 13: 609-612. CANNING, D. R., R. J. MCKEON, D. A. DEWITT, G. PERRY, J. R. WUJEK, R. C. A. FREDERICKSON, AND J. SILVER. 1993.8Amyloid of Alzheimer’s disease induces reactive ghosis that inhibits axonal outgrowth. Exp. Neural. 124: 289-298.

CAST~~O, E. M., AND B. FRANGIONE. 1988. Biology of disease. Human amyloidoses, AIzheimer’s disease and related disorders. Lab. Invest. 58: 122-132.

COLE, G. M., E. MASLIAH, E. R. SHELTON, H. W. CHAN, R. D. TERRY, AND T. SAITOH. 1991. Accumulation of amyloid precursor fragment in AIzheimer plaques. Neurobiol. Aging 12: 85-91. COTMAN, C. W., B. J. CIJMMINGS, AND C. J. PIKE. 1993. Molecular cascades in adaptive versus pathological plasticity. Pages 217- 240 in A. Giorio, Ed., Neuroregeneration. Raven Press, Ltd., New York. COTMAN, C. W., ANII C. J. PIKE. 1994.6Amyloid and its contribu- tions to neurodegeneration in AIsheimer’s disease. Pages 305- 315 in R. D. Terry, R. Katzman, and K Bick, Eds, Alzheimer’s Disease. Raven Press, New York. COTMAN, C. W., C. J. PIKE, AND B. J. CUMMINGS. 1993. Adaptive versus pathological plasticity: Possible contributions to age- related dementia. Pages 35-45 in F. J. Se& Ed., Advances in Neurology. Raven Press, New York. CUMMINGS, B. J., C. J. PIKE, AND C. W. COTMAN. 1994. PHF-1 immunopositive dystrophic neurites are preferentially associated with thioflavine positive B-pleated plaques and not diffuse amy- loid plaques. Neurobiol. Aging 15: S152.

CIJMMINGS, B. J., J. H. SU, AND C. W. COTMAN. 1993. Neuritic involvement within bFGF immunopositive plaques of AIshei- mer’s disease. Exp. Neural. 124: 315-325. CUMMINGS, B. J., J. H. Su, J. W. GEDDES, W. E. VAN NOSTRAND, S. L. WAGNER, D. D. CUNNINGHAM, AND C. W. C&MAN. 1992. Aggregation of the amyloid precursor protein within degenerat- ing neurons and dystrophic neurites. Neuroscience 48: 763-777.

CUMMINGS, B. J., J. S. Su, C. W. COTMAN, R. WHITE, AND M. J. RUSSELL. 1993. 8-Amyloid accumulation in aged canine brain: a model of early plaque formation in AIzheimer’s disease. Neuro- biol. Aging 14: 547560. DEWITT, D. A., J. SILVER, D. R. CANNING, AND G. PERRY. 1993. Chondroitin sulfate proteoglycans are associated with the lesions of AIsheimer’s disease. Exp. Neurol. 121: 149-152. DICKSON, D. W., J. FARLO, P. DAVIES, H. CRYSTAL, P. FIJLD, ANLI

178 PIKE, CUMMINGS, AND COTMAN

20.

21.

22.

23.

24.

25.

26.

‘27.

28.

29.

30.

. 31.

32.

33.

34.

35.

36.

37.

S. C. YEN. 1988. A double-labeling immunohistochemical study of senile plaques. Am. J. Pathol. 13% 86-101. Dumw, P. E., M. RAPPORT, AND L. GRAF. 1990. GIiaI fibriilary acidic protein and Aizheimer-type senile dementia. Neurology 30~778-782.

38.

EDDLESMN, M., AND L. MUCKE. 1993. Molecular profile of reac- tive astrocytes-implications for their role in neurologic disease. Neuroscience 54: 15-36. FRASER, P. E., L. L~FXQUE, AND MCLACHIAN. 1994. AIzheimer A8 amyloid forms an inhibitory neuronal substrate. J. Neuro- chem.62: 1227-1230.

39.

40.

G~MEZ-PINILLA, F., B. J. t%MMINGS, AND C. W. COTMAN. 1990. Induction of basic fibroblast growth factor in AIzheimer’s disease pathology. NeuroRepoti 1: 211-214. GREENBERG, S. G., AND P. DAVIES. 1990. A preparation of AIzheimer paired helical filaments that displays distinct tau proteins by polyacryhunide gel electrophoresis. Proc. Natl. Acad. Sci.USA81: 5827-5831.

41.

GRIFFIN, W. S. T., L. C. STANLEY, C. LING, L. WHITE, V. MACLEOD, L. J. PERROT, C. L. WHITE III, AND C. ARAOZ. 1989. Brain interleukin-1 and S-100 immunoreactivity are elevated in Down’s syndrome and AIzheimer’s disease. Proc. Natl. Acad. Sci. USA 86: 7611-7615.

42.

43.

H~~KE, A, D. R. CANNING, C. J. MALEMUD, AND J. SILVER. 1994. Regional differences in reactive ghosis induced by substrate- bound $-amyloid. Exp. Neural. 130: 56-66. JOACHIM, C., D. GAMES, J. MORRIS, P. WARD, D. FRENKEL, AND D. SELKOE. 1991. Antibodies to non-beta regions of the beta-amyloid precursor protein detect a subset of senile plaques. Am. J. Pathol. 138~373-384.

44.

45. JOACHIM, C. L., J. H. MORRIS, AND D. J. SELKOE. 1989. Diffuse senile plaques occur commonly in the cerebellum in Alzheimer’s disease. Am. J. Pathol. 136: 309-319. KISILEVSKY, R. 1987. From arthritis to Alzheimer’s disease: current concepts on the pathogenesis of amyloidosis. Can. J. Physiol. Pharmacol. 65: 1805-1815.

46.

MANDYBUR, T. I., AND C. C. CHUIRAZZI. 1990. Astrocytes and the plaques of Aizheimer’s disease. Neurology 40: 635-639. EnANN, D. M. A., A. M. T. BROWN, D. PRINJA, D. JONES, ANTI C. A. DAVIES. 1990. A morphological analysis of senile plaques in the brains of non-demented persons of different ages using silver, immunocytochemical and lectin hi&chemical staining tech- niques. J. Neuropathol. Appl. Neurobiol. 16: 17-25.

47.

48.

IIA?TSON, M. P., K. J. TOMASELLI, AND R. E. RYDEL. 1993. Calcium-destabilizing and neurodegenerative effects of aggre- gated 8-amyloid peptide are attenuated by basic FGF. Brain Res. 621:35-49. MCKEE, A C., K. S. KOSIK, AND N. W. KOWALL. 1991. Neuritic pathology and dementia in Aizheimer’s disease. Ann. Neurol. 30: 156-165. MULLAN, M., ANLI F. CRAWFORD. 1993. Genetic and molecular advances in Alzheimer’s disease. Trends Neurosci. IS: 398-403. NAMEA, Y., M. TOMONGA, H. KAWASAKI, E. OTOMO, AND K. IKEDA. 1991. ApoIipoprotein E immunoreactivity in cerebral amyloid deposits and neurofibriIIaxy tangles in AIzheimer’s disease and kuru plaque amyloid in Creutzfeldt-Jakob disease. Brain Res. S41:163-166.

49.

50.

51.

PALMERT, M. R., M. R. PODLISNY, D. S. WITKER, T. OLTERSDORF, L. H. YOUNIUN, D. J. SELKOE, AM) S. G. YOUNKIN. 1988. Antisera to an amino-terminal peptide detect the amyloid precursor protein of Aizheimer’s disease and recognize senile plaques. Biochem. Biophys. Res. Commun. 158: 432-437. PASTERNACK, J. M., C. R. ABRAHAM, B. J. VAN DYKE, H. POTTER, AND S. G. YOUNKIN. 1989. Astrocytes in Alzheimer’s disease gray

52.

53.

matter express ai-antichymotrypsin mRNA. Am. J. Pathol. 135: 827-834. PIKE, C. J., D. BURDICK, A. J. WALENCEWICZ, C. G. GLABE, AND C. W. COTMAN. 1993. Neurodegeneration induced by 8-amyloid peptides in oitro: The role of peptide assembly state. J. Neurosci. 13:167&1687. PIKE, C. J., B. J. CUMMINGS, AND C. W. COTMAN. 1992.8Amyloid induces neuritic dystrophy in uitro: similarities with Alzheimer pathology. NeuroReport 3: 769-772. PIKE, C. J., B. J. CUMMINGS, AND C. W. COTMAN. 1995. Contribu- tions of 8-amyloid to reactive astrocytosis in Aizheimer’s disease. Pages 619-627 in K. IqbaI, J. Mortimer, B. Winblad and H. Wisniewski, Eds., Research Advances in Alsheimer’s Disease and Related Disorders. Wiley, Sussex. PIKE, C. J., B. J. CUMMINGS, R. MONZAVI, AND C. W. COTMAN. 1994. 8-Amyloid-induced changes in cultured astrocytes parallel reactive astrocytosis associated with senile plaques in Aizhei- mer’s disease. Neuroscience 63: 517-531. PIKE, C. J., A. J. WALENCEWICZ, C. G. GJAFIE, AND C. W. COTMAN. 1991. In vitro aging of 8-amyloid protein causes peptide aggrega- tion and neurotoxicity. Brain Res. 563: 311-314. PIKE, C. J., A. J. WALENCEH~CZ-WASSER, J. KOSMOSKI, D. H. CRIBBS, C. G. GLABE, AND C. W. COTMAN. 1995. Structure-activity analyses of 8-amyloid peptides: Contributions of the 825-35 region to aggregation and neurotoxicity. J. Neurochem. 64: 253-265. PROBST, A., H. BRUNNSCHWEILER, C. LAUTENSCHLAGER, AND J. ULRICH. 1987. A special type of senile plaque, possibly an initial stage. Acta Neuropathol. 74: 133-141. ROZEMULLER, J. M., P. EIKELENBOOM, F. C. STAM, K. BEYREUTHER, AND C. L. NIASTERS. 1989. A4 protein in AIzheimer’s disease: Primary and secondary cellular events in extracellular amyloid deposition. J. Neuropathol. Exp. Neurol. 48: 674-691. SCHECHTER, R., S.-H. C. YEN, AND R. D. TERRY. 1981. Fibrous astrocytes in senile dementia of the AIzheimer type. J. Neuro- pathol. Exp. Neurol. 40: 95-101. SCHMECHEL, D. E., A. M. SAUNDERS, W. J. STRITTMA~ER, B. J. GRAIN, C. M. HULETTE, AND S. H. Joo. 1993. Increased amyloid beta-peptide deposition in cerebral cortex as a consequence of apolipoprotein E genotype in late-onset Aizheimer’s disease. Proc. Natl. Acad. Sci. USA 90: 9649-9653. SCHMIDT, M. L., A. G. DIDARIO, V. M. LEE, AND J. Q. TRO- JANOWSKL 1994. An extensive network of PHF tau-rich dystro- phic neurites permeates neocortex and nearly a5 neuritic and diffuse amyloid plaques in AIzheimer disease. FEBS L&t. 344: 69-73. SELKOE, D. J. 1993. Physiological production of the 8-amyloid protein and the mechanism of Aizheimer’s disease. Trends Neurosci. 16~403409. SHOJI, M., S. HIRAI, H. YAMAGUCHI, Y. HARIGAYA, AND T. KAWAR- ABAYASHI. 1990. Amyloid beta-protein precursor accumulates in dystropbic neurites of senile plaques in AIzheimer-type demen- tia. Brain Res. 512: 164-168. SIMMONS, L. K., P. C. MAY, K. J. TOMASELLI, R. E. RYDEL, K. S. FUSON, E. F. BRIGHAM, S. WRIGHT, I. LIEBERBURG, G. W. BECKER, D. N. BREMS, AND W. Y. LI. 1994. Secondary structure of amyloid 8 peptide correlates with neurotoxic activity in vitro. Mol. Pharmacol. 45: 373-379. SNOW, A. D., AND T. N. WIGHT. 1989. Proteoglycans in the pathogenesis of Alzheimer’s disease and other amyloidoses. Neurobiol. Aging 10: 481-497. STOPA, E. G., A.-M. GONZALEZ, R. CHORSKY, R. J. CORONA, J. ALVAREZ, E. D. BIRD, AND A. BAIRD. 1990. Basic fibroblast growth factor in Ahheimer’s disease. Biochem. Biophys. Res. Commun. 17k690-696.

COLOCALIZATION OF ASTROCYTES AND NEURITES WITH PLAQUES 179

54. Su, J. H., B. J. CUMMINGS, AND C. W. COTMAN. 1994. Early phosphorylation of tau in Alzheimer’s disease occurs at Ser-202 and is preferentially located within neurites. NeuroReporr. 5: 2358-2362.

55. Su, J. H., B. J. CUMMINGS, AND C. W. COTMAN. 1994. Subpopula- tions of dystrophic neurites in Alzheimer’s brain with distinct immunocytochemical and argentophilic characteristics. Brain Res. 637: 37-44.

56. SUENAGA, T., A. HIRANO, J. F. LLENA, H. K~IEZAK-REDING, S.-H. YEN, AND D. W. DICKSON. 1990. Modified Bielschowsky and immunocytochemical studies on cerebellar plaques in Alzhei- mer’s disease. J. Neuropathol. Exp. Neural. 49: 31-40.

57. TAGLIAVINI, F., G. GIACCONE, B. FRANCIONE, AND 0. BUGIANI. 1988. Preamyloid deposits in the cerebral cortex of patients with Alzheimer’s disease and nondemented individuals. Neurosci. I&. 93: 191-196.

58. VAN DER WAL, E. A., F. GOMEZ-PINELLA, AND C. W. COTMAN. 1993.

59.

60.

61.

62.

Transforming growth factor-61 is in plaques in Alzheimer and Down pathologies. NeuroRepori 4: 69-72.

WALICKE, P., W. M. COWAN, N. UENO, A. BAIRD, AND R. GUILLE- MIN. 1986. Fibroblast growth factor promotes survival of dissoci- ated hippocampal neurons and enhances neurite extension. Proc. Natl. Acad. Sci. USA 83: 3012-3016. WALKER, L. C., C. A. KITT, R. G. STRUBLE, D. E. SCHMECHEL, W. H. OERTEL, L. C. CORK, AND D. L. PRICE. 1985. Glutamic acid decarboxylase-like immunoreactive neurites in senile plaques. Neurosci. L.&f. 59: 165-169. WANG, B., AND L. LARsoN. 1985. Simuhaneous demonstration of multiple antigens by indirect immunofluorescence or immuno- gold staining. Histochemistry 83: 47-56.

WISNIEWSKI, H. M., AND R. D. TERRY. 1973. Reexamination of the pathogenesis of the senile plaque. Pages l-26 in H. M. Zimmer- man, Ed., Progress in Neuropathology. Grune and Stratton, New York.