Identification of sequence variants and analysis of the role of the glycogen synthase kinase 3 beta gene and promoter in late onset Alzheimer's disease

Alzheimer's disease (AD) is a disorder characterised by a progressive deterioration in memory and other cognitive functions. Glycogen synthase kinase 3 beta (GSK3 beta) phosphorylates the microtubule associated protein tau at sites that are aberrantly phosphorylated in AD. GSK3 beta binds to presenilin 1 and plays a role in wnt and insulin signalling cascades, both of which have been proposed to be linked to AD. Moreover GSK3 beta activity may be altered in AD brain. These observations suggest a central role for GSK3 beta in AD and led us to investigate GSK3 beta as a candidate gene for AD. We sought to identify sequence variations in the gene and its promoter, as these could have an effect on activity and expression leading to abnormal function. Sequencing over 3000 bp of the GSK3 beta putative promoter revealed there to be five sequence variations, two of which were common (>10%). However on further examination none of these, either alone or in synergy, had any association with late onset AD. Stratification of the data by APOE epsilon 4 status also produced no significant association. Sequencing of the GSK3 beta coding region revealed no variations. This would suggest that the aberrant phosphorylation of tau by GSK3 beta in AD is not due to sequence variations in the gene or its promoter.     

Characteristics of the binding of thioflavin S to tau paired helical filaments

Binding of histological benzothiazole dye thioflavin S (ThS) to protein aggregates has been related with the presence of amyloid beta sheet structure in those protein aggregates. Paired helical filaments (PHF) from Alzheimer's disease (AD) patients, (whose main component is the microtubule associated protein, tau) bind to thioflavins. By using a novel immunofluorescence method, the binding of ThS to isolated tau filaments was tested. Also, the characteristics of this binding of ThS to PHF or to the in vitro assembled tau filaments, have been analyzed. Our results suggests that ThS binds to PHF with a higher affinity than to the straight filaments (SF), also found in AD.     

Kinases and phosphatases and tau sites involved in Alzheimer neurofibrillary degeneration

Microtubule associated protein (MAP) tau is abnormally hyperphosphorylated in Alzheimer's disease (AD) and related tauopathies; in this form it is the major protein subunit of paired helical filaments (PHF)/neurofibrillary tangles. However, the nature of protein kinases and phosphatases and tau sites involved in this lesion has been elusive. We investigated self-assembly and microtubule assembly promoting activities of hyperphosphorylated tau isolated from Alzheimer disease brain cytosol, the AD abnormally hyperphosphorylated tau (AD P-tau) before and after dephosphorylation by phosphoseryl/phosphothreonyl protein phosphatase-2A (PP-2A), and then rephosphorylation by cyclic AMP-dependent protein kinase (PKA), calcium, calmodulin-dependent protein kinase II (CaMKII), glycogen synthase kinase-3beta (GSK-3beta) and cyclin-dependent protein kinase 5 (cdk5) in different kinase combinations. We found that (i) dephosphorylation of AD P-tau by PP-2A inhibits its polymerization into PHF/straight filaments (SF) and restores its binding and ability to promote assembly of tubulin into microtubules; (ii) rephosphorylation of PP-2A-dephosphorylated AD P-tau by sequential phosphorylation by PKA, CaMKII and GSK-3beta or cdk5, and as well as by cdk5 and GSK-3beta, promotes its self-assembly into tangles of PHF similar to those seen in Alzheimer brain, and (iii) phosphorylation of tau sites required for this pathology are Thr231 and Ser262, along with several sites flanking the microtubule binding repeat region. Phosphorylation of recombinant human brain tau(441) yielded similar results as the PP-2A dephosphorylated AD P-tau, except that mostly SF were formed. The conditions for the abnormal hyperphosphorylation of tau that promoted its self-assembly also induced the microtubule assembly inhibitory activity. These findings suggest that activation of PP-2A or inhibition of either both GSK-3beta and cdk5 or one of these two kinases plus PKA or CaMKII might be required to inhibit Alzheimer neurofibrillary degeneration.     

Identification of Microtubule-Associated Protein Tau Isoforms in Alzheimer's Paired Helical Filaments

AD66 proteins derived from sodium dodecylsulfate (SDS) insoluble paired helical filaments (PHF) were isolated from Alzheimer's brain using a purification procedure developed previously in this laboratory, and characterized by immunologic and chemical cleavage methods. AD66 proteins were immunoreactive with antibodies that recognize the amino terminal, tubulin-binding, and carboxy terminal domains of microtubule-associated protein tau indicating the presence of the entire tau sequence in AD66 proteins. These proteins were reactive with antibody 423 that binds to PHF but not human adult tau. Immunologic and chemical cleavage studies indicated that only two of the six tau isoforms were present in these proteins. AD66 proteins were comprised of tau proteins containing only three tubulin binding domains with either a 29 amino acid insert or no amino terminal insert. For comparative purposes, SDS soluble PHF-tau (A68 proteins) was purified from Alzheimer's brains and normal adult tau purified from control brains. Antibody Alz-50 was immunoreactive with PHF-tau or normal tau regardless of alkaline phosphatase treatment while immunoreactivity was only observed with dephosphorylated AD66 proteins. A second phosphorylated epitope on AD66 proteins but not PHF-tau or normal tau proteins was demonstrated with antibody PHF9. These data suggest that AD66 proteins represent a more phosphorylated form of tau than PHF-tau or normal tau proteins. Two-dimensional gel electrophoresis demonstrated that AD66 proteins have higher apparent molecular weights and lower pI values than normal tau, differences possibly due to the greater phosphorylation observed in these proteins.     

In vitro polymerization of oxidized tau into filaments

Paired helical filaments (PHF) are abnormal neuronal polymers characteristic of Alzheimer's disease (AD). Although tau appears to be a major constituent of PHF, the mechanism for the polymerization of tau or its integration into PHF remains unknown. Here, we show that the oxidation of bovine tau in vitro induces an apparent dimerization of this protein and polymerization into filaments. These observations suggest that the oxidation of tau in vivo may contribute to the development of PHF in individuals with AD.     

Genetic variation in the tau kinases pathway may modify the risk and age at onset of Alzheimer's disease

Tau abnormal hyperphosphorylation and the formation of neurofibrillary tangles in the Alzheimer's disease (AD) brain is the result of upregulation of tau kinases. In a group of 729 Spanish late-onset AD patients and 670 healthy controls, we examined variations into a set of 20 candidate genes of kinases involved in tau phosphorylation at AD-related sites (PRKACB; CAMK2A; MARK1, 2, 3 and 4; CSNK1D; CDC2; RPS6KB1 and 2; p38α and β; IB1; JNK1, 2 and 3; MEK1 and 2; ERK1 and 2), to address hypotheses of genetic variation that might influence both AD risk and age at disease onset. There was an increased frequency of RPS6KB2 (intron 2, rs917570) minor allele in patients (50%) versus controls (39%) (OR = 1.52; 95% CI 1.30-1.77; p = 1.24 × 10-5 Bonferroni corrected), and the presence of this minor allele was significantly (p = 4.2 × 10-5) associated with a 3-years later onset of AD (mean age 74.1 years) when compared to age at onset of non-minor allele carriers (mean age 71.1 years). In APOE non-ε4 allele carriers, the combined effect of AD-associated risk alleles from the genes of CDC2, RPS6KB1 and 2, p38α, JNK (1, 2 and 3), MEK2, and ERK2 was significantly (p = 0.002) associated with a late-onset (>76 years) of AD. The CDC2 AGC haplotype derived from SNPs in introns 3 (rs2448347), 5 (rs2456772), and 7 (rs1871447) showed a protective effect against AD in APOE non-ε4 allele carriers (permutation p = 1.0 × 10-4) with a frequency of 9% in cases and 15% in controls. Common genetic variation in the tau kinases pathway does underlie individual differences not only in susceptibility to AD but also in disease phenotype (age at disease onset).     

Amyloid-β toxicity and tau hyperphosphorylation are linked via RCAN1 in Alzheimer's disease

Amyloid-β peptide (Aβ) toxicity and tau hyperphosphorylation are hallmarks of Alzheimer's disease (AD). How their molecular relationships may affect the etiology, progression, and severity of the disease, however, has not been elucidated. We now report that incubation of fetal rat cortical neurons with Aβ upregulates expression of the Regulator of Calcineurin gene RCAN1, and this is mediated by Aβ-induced oxidative stress. Calcineurin (PPP3CA) is a serine-threonine phosphatase that dephosphorylates tau. RCAN1 proteins inhibit this phosphatase activity of calcineurin. Increased expression of RCAN1 also causes upregulation of glycogen synthase kinase-3β (GSK3β), a tau kinase. Thus, increased RCAN1 expression might be expected to decrease phospho-tau dephosphorylation (via calcineurin inhibition) and increase tau phosphorylation (via increased GSK3β activity). Indeed, we find that incubation of primary cortical neurons with Aβ results in increased phosphorylation of tau, unless RCAN1 gene expression is silenced, or antioxidants are added. Thus we propose a mechanism to link Aβ toxicity and tau hyperphosphorylation in AD: In our hypothesis, Aβ causes mitochondrial oxidative stress and increases production of reactive oxygen species, which result in an upregulation of RCAN1 gene expression. RCAN1 proteins then both inhibit calcineurin and induce expression of GSK3β. Both mechanisms shift tau to a hyperphosphorylated state. We also find that lymphocytes from persons whose ApoE genotype is ε4/ε4 (with high risk of developing AD) show higher levels of RCAN1 and phospho-tau than those carrying the ApoE ε3/ε3 or ε3/ε4 genotypes. Thus upregulation of RCAN1 may be a valuable biomarker for AD risk.     

M1 muscarinic agonists target major hallmarks of Alzheimer's disease--an update

The M1 muscarinic receptor (M1 mAChR), preserved in Alzheimer's disease (AD), is a pivotal target that links major hallmarks of AD, e.g. cholinergic deficiency, cognitive dysfunctions, beta-amyloid (Abeta) and tau pathologies. Some muscarinic agonists, while effective in AD, had limited clinical value due to adverse effects and lack of M1 selectivity. The M1 selective muscarinic agonists AF102B [Cevimeline], AF150(S) and AF267B - i) elevated alphaAPPs, decreased Abeta levels and tau hyperphosphorylation, and blocked Abeta-induced neurotoxicity, in vitro, via M1 mAChR-modulation of kinases (e.g. PKC, MAPK and GSK3beta); ii) restored cognitive deficits, cholinergic markers, and decreased tau hyperphosphorylation in relevant models with a wide safety margin. AF267B decreased brain Abeta levels in hypercholesterolemic rabbits and decreased CSF Abeta42 in rabbits and removed vascular Abeta42 deposition from cortex in cholinotoxin-treated rabbits. In 3x transgenic-AD mice that recapitulate the major pathologies and cognitive deficits of AD, chronic AF267B treatment rescued cognitive deficits and decreased Abeta42 and tau pathologies in the cortex and hippocampus (not amygdala), via M1 mAChR-activation of ADAM17/TACE and decreased BACE1 steady state levels and inhibition of GSK3beta, extending findings from above. CONCLUSIONS: A comprehensive therapy should target all AD hallmarks, regardless of the culprit(s) responsible for the disease. In this context, AF267B is the 1(st) reported low MW CNS-penetrable mono-therapy that meets this challenge. Clinical trials will determine if AF267B may become an important therapy in AD.     

The active form of glycogen synthase kinase-3β is associated with granulovacuolar degeneration in neurons in Alzheimer's disease

Glycogen synthase kinase-3beta (GSK-3beta) is a physiological kinase for tau and is a candidate protein kinase involved in the hyperphosphorylation of tau present in paired helical filament (PHF)-tau of neurofibrillary tangles (NFT) in Alzheimer's disease (AD). GSK-3beta is also a key element of several signaling cascades (including cell death cascades). We have investigated the immunocytochemical localization of GSK-3 immunoreactivity in AD. Neurons exhibiting strongly GSK-3-immunoreactive granules were observed in AD, with a much higher frequency than in control subjects. This immunoreactivity was found to co-localize with the granulovacuolar degeneration (GVD) and to be associated with the granules of the granulovacuolar bodies. The GVD granules showed a strong GSK-3alpha and GSK-3beta immunoreactivity, and this immunoreactivity was abolished by preabsorption with recombinant GSK-3. In addition, the GVD immunoreactivity was observed with an antibody against the tyrosine-phosphorylated and active form of GSK-3. Some granules of the granulovacuolar degeneration were also intensely labeled with an antibody specific for tau isoforms containing insert 1 (exon 2) and with antibodies specific for tau phosphorylated on Ser262 and for tau phosphorylated on Thr212/Ser214, two phosphorylation sites generated in vitro by GSK-3alpha and beta. GSK-3beta was expressed in neurons containing NFT but only a small proportion of intracellular NFT were observed to be GSK-3beta immunoreactive. Immunoblotting analysis of fractions enriched in PHF-tau did not reveal any GSK-3beta immunoreactivity in these fractions, indicating that GSK-3beta was only loosely associated to NFT. These results suggest that neurons developing GVD sequester an active, potentially deleterious, form of GSK-3 in this compartment and that increased GSK-3 immunoreactivity in a subset of neurons quantitatively differentiates normal aging from AD.     

Alzheimer neurofibrillary degeneration

Neurofibrillary degeneration has primary and pivotal involvement in the pathogenesis of Alzheimer disease (AD) and other tauopathies. The inhibition of this lesion offers a promising therapeutic approach. The microtubule- associated protein (MAP) tau is abnormally hyperphosphorylated in the brain of patients with AD, and in this form it is the major protein subunit of paired helical filaments/neurofibrillary tangles (PHF/NFT). The abnormal tau that is polymerized into PHF/NFT is apparently inert and has no effect on microtubule assembly in vitro. The cytosolic abnormally hyperphosphorylated tau from AD brain, the AD P-tau, does not promote in vitro microtubule assembly but, instead, sequesters normal tau, MAP1, and MAP2 and inhibits microtubule assembly. The AD P-tau readily self-assembles in vitro into tangles of PHF/straight filaments, and this self-assembly requires the abnormal hyperphosphorylation of this protein. Although, to date, an up-regulation of the activity of a tau kinase has not been established, the activity of phosphoseryl/ phosphothreonyl protein phosphatase (PP)-2A, which regulates the phosphorylation of tau, is compromised in AD brain. Thus, modulation of the activities of pp-2A and one or more tau kinases and inhibition of the sequestration of normal MAPs by AD P-tau offer promising therapeutic opportunities to inhibit neurofibrillary degeneration and the diseases characterized by this lesion. Development of high-throughput screening assays for potential drugs aimed at these therapeutic targets is currently under way.     

Significance and mechanism of Alzheimer neurofibrillary degeneration and therapeutic targets to inhibit this lesion

Abnormally hyperphosphorylated tau which is the major protein subunit of paired helical filaments (PHF)/neurofibrillary tangles is the pivotal lesion in Alzheimer disease (AD) and related tauopathies. The cosegregation of tau mutations with disease in inherited cases of frontotemporal dementia has confirmed that abnormalities in this protein can be a primary cause of neurodegeneration. Unlike normal tau that promotes assembly and maintains the structure of microtubules, the abnormally hyperphosphorylated protein sequesters normal tau, MAP1 and MAP2 and consequently disassembles microtubules. The abnormal hyperphosphorylation also promotes the self assembly of tau into tangles of PHF. The hyperphosphorylation of tau in AD is probably due to a protein phosphorylation/dephosphorylation imbalance produced by a decrease in the activity of protein phosphatase (PP)-2A and increase in the activities of tau kinases which are directly or indirectly regulated by PP-2A. Two of the most promising pharmacologic therapeutic approaches to AD are (1) the development of drugs that can inhibit the sequestration of normal MAPs by the abnormally hyperphosphorylated tau, and (2) the development of drugs that can reverse the abnormal hyperphosphorylation of tau by correcting the protein phosphorylation/dephosphorylation imbalance.     

Glycogen synthase kinase 3 alteration in alzheimer disease is related to neurofibrillary tangle formation

Highly phosphorylated τ protein is the main component of paired helical filaments (PHF), which comprise the neurofibrillary tangles (NFT) in some neurons of patients with Alzheimer disease (AD). Glycogen synthase kinase 3 (GSK3) phosphorylates τ in vitro at several sites also found to be phosphorylated in PHF-τ, τ is phosphorylated at these sites in both AD and normal control (NC) brains, although the extent of phosphorylation is far greater in τ from AD. If GSK3 levels are increased in AD, then τ phosphorylation and perhaps PHF formation may occur. To quantify GSK3, blots of AD and NC brain supernatant and particulate fractions were probed with antibodies to GSK3. In particulate fractions of AD compared to NC, GSK3α immunoreactivity did not increase, but in fact, decreased 40%, and GSK3β immunoreactivity decreased 30%. GSK3α and GSK3β levels correlated well with each other. GSK3 levels correlated negatively with numbers of NFT. The authors dedicate this article to the memory of Tsunao Saitoh, who died May 7, 1996.     

Defective insulin signaling pathway and increased glycogen synthase kinase-3 activity in the brain of diabetic mice: parallels with Alzheimer's disease and correction by insulin

We have evaluated the effect of peripheral insulin deficiency on brain insulin pathway activity in a mouse model of type 1 diabetes, the parallels with Alzheimer's disease (AD), and the effect of treatment with insulin. Nine weeks of insulin-deficient diabetes significantly impaired the learning capacity of mice, significantly reduced insulin-degrading enzyme protein expression, and significantly reduced phosphorylation of the insulin-receptor and AKT. Phosphorylation of glycogen synthase kinase-3 (GSK3) was also significantly decreased, indicating increased GSK3 activity. This evidence of reduced insulin signaling was associated with a concomitant increase in tau phosphorylation and amyloid beta protein levels. Changes in phosphorylation levels of insulin receptor, GSK3, and tau were not observed in the brain of db/db mice, a model of type 2 diabetes, after a similar duration (8 weeks) of diabetes. Treatment with insulin from onset of diabetes partially restored the phosphorylation of insulin receptor and of GSK3, partially reduced the level of phosphorylated tau in the brain, and partially improved learning ability in insulin-deficient diabetic mice. Our data indicate that mice with systemic insulin deficiency display evidence of reduced insulin signaling pathway activity in the brain that is associated with biochemical and behavioral features of AD and that it can be corrected by insulin treatment.     

Sodium tungstate decreases the phosphorylation of tau through GSK3 inactivation

Tungstate treatment increases the phosphorylation of glycogen synthase kinase-3beta (GSK3beta) at serine 9, which triggers its inactivation both in cultured neural cells and in vivo. GSK3 phosphorylation is dependent on the activation of extracellular signal-regulated kinases 1/2 (ERK1/2) induced by tungstate. As a consequence of GSK3 inactivation, the phosphorylation of several GSK3-dependent sites of the microtubule-associated protein tau decreases. Tungstate reduces tau phosphorylation only in primed sequences, namely, those prephosphorylated by other kinases before GSK3beta modification, which are serines 198, 199, or 202 and threonine 231. The phosphorylation at these sites is involved in reduction of the interaction of tau with microtubules that occurs in Alzheimer's disease.     

Pharmacological targets to inhibit Alzheimer neurofibrillary degeneration

Neurofibrillary degeneration appears to be required for the clinical expression of Alzheimer disease (AD) and related tauopathies. Given the polyetiology of these diseases and the pivotal involvement of neurofibrillary degeneration in their pathogenesis, inhibition of this lesion offers a promising therapeutic target. Studies from our laboratories have shown that there is a protein phosphorylation/dephosphorylation imbalance and that the microtubule associated protein tau is abnormally hyperphosphorylated in the brain of patients with AD and in this form it is the major protein subunit of paired helical filaments/neurofibrillary tangles (PHF/NFT). The abnormal tau which is polymerized into PHF/NFT neither promotes or inhibits in vitro microtubule assembly. In contrast the cytosolic abnormally hyperphosphorylated tau from AD brain, the AD P-tau neither associates with tubulin nor promotes in vitro microtubule assembly but instead it sequesters normal tau, MAP1 and MAP2 and inhibits microtubule assembly. The AD P-tau readily self-assembles in vitro into tangles of PHF/straight filaments under physiological conditions of protein concentration, pH, ionic strength and reducing conditions and this self assembly requires the abnormal hyperphosphorylation of this protein. The activity of phosphoseryl/phosphothreonyl protein phosphatase (PP)-2A which regulates the phosphorylation of tau, is compromised in AD brain. Thus, modulation of the activities of protein phosphatase-2A and tau kinases and inhibition of the sequestration of normal MAPs by AD P-tau offer promising therapeutic opportunities to inhibit neurofibrillary degeneration and the diseases characterized by this lesion.     

17β-estradiol attenuates glycogen synthase kinase-3β activation and tau hyperphosphorylation in Akt-independent manner

Decline of estrogen is associated with high incidence of Alzheimer's disease (AD) characterized pathologically with tau hyperphosphorylation, and glycogen synthase kinase-3beta (GSK-3beta) is a major tau kinase. However, the role of estrogen on GSK3beta-induced tau hyperphosphorylation is elusive. Here, we treated N2a cells with wortmannin (Wort) and GF-109203X (GFX) or gene transfection to activate GSK-3beta and to induce tau hyperphosphorylation and then the effects of 17beta-estradiol (betaE2) on tau phosphorylation and GSK-3beta activity were studied. We found that betaE2 could attenuate tau hyperphosphorylation at multiple AD-related sites, including Ser396/404, Thr231, Thr205, and Ser199/202, induced by Wort/GFX or transient overexpression of GSK-3beta. Simultaneously, it increased the level of Ser9-phosphorylated (inactive) GSK-3beta. To study whether the protective effect of betaE2 on GSK-3beta and tau phosphorylation involves protein kinase B (Akt), an upstream effector of GSK-3, we transiently expressed the dominant negative Akt (dnAkt) in the cells. We found that betaE2 could attenuate Wort/GFX-induced GSK-3beta activation and tau hyperphosphorylation with Akt-independent manner. It suggests that betaE2 may arrest AD-like tau hyperphosphorylation by directly targeting GSK-3beta.     

Biochemical and ultrastructural study of neurofibrillary tangles in amyotrophic lateral sclerosis/parkinsonism-dementia complex in the Kii peninsula of Japan

Amyotrophic lateral sclerosis/parkinsonism-dementia complex of the Kii peninsula (Kii ALS/PDC) is a neurodegenerative disorder endemic to natives in the southern coast area of the Kii peninsula of Japan. The disorder closely resembles Guamanian ALS/PDC clinically and neuropathologically. The characteristic neuropathological finding is abundant neurofibrillary tangles (NFTs) without amyloid deposition. To elucidate the biochemical properties of hyperphosphorylated tau protein, the major component of the NFTs, we examined Kii ALS/PDC brains by immunoblotting and immunohistochemical analysis using well-characterized anti-tau antibodies specific to phosphorylation-dependent or -independent epitopes. Hyperphosphorylated tau in Kii ALS/PDC had phosphorylated epitopes common to tau of paired helical filaments (PHFs) in Alzheimer disease (AD): immunoblot showed triplet bands composed of 6 tau isoforms. Ultrastructurally, NFTs revealed a twisted filamentous shape similar to PHF of AD. The biochemical properties of its phosphorylated tau protein and the ultrastructural characteristics of the NFTs of Kii ALS/PDC are very similar, if not identical, to PHF tau in AD, although they are different taupopathies.     

Phosphorylation of human tau protein by microtubule-associated kinases: GSK3beta and cdk5 are key participants

Microtubules (MTs), primarily composed of alpha and beta tubulin polymers, must often work in concert with microtubule-associated proteins (MAPs) in order to modulate their functional demands. In a mature brain neuron, one of the key MAPs that resides primarily in the axonal compartment is the tau protein. Tau, in the adult human brain, is a set of six protein isoforms, whose binding affinity to MTs can be modulated by phosphorylation. In addition to the role that phosphorylation of tau plays in the "normal" physiology of neurons, hyperphosphorylated tau is the primary component of the fibrillary pathology in Alzheimer's disease (AD). Although many protein kinases are known to phosphorylate tau in vitro, the in vivo players contributing to the hyperphosphorylation of tau remain elusive. The experiments in this study attempt to define which protein kinases and protein phosphatases reside in the associated network of microtubules, thereby being strategically positioned to influence the phosphorylation of tau. Microtubule fractions are utilized to determine which of the microtubule-associated kinases most readily impacts the phosphorylation of tau at "AD-like" sites. Results from this study indicate that PKA, CK1, GSK3beta, and cdk5 associate with microtubules. Among the MT-associated kinases, GSK3beta and cdk5 most readily contribute to the ATP-induced "AD-like" phosphorylation of tau.     

Plaque biogenesis in brain aging and Alzheimer’s disease. II. Progressive transformation and developmental sequence of dystrophic neurites

Plaque-associated dystrophic neurites are a common pathological feature in the brains of patients with Alzheimer’s disease (AD). In the present study, we investigated the relative abundance and progressive transformation of the amyloid precursor protein (APP), neurofilament (NF) and paired helical filament (PHF) tau-positive dystrophic neurites, within plaques in non-demented controls versus plaque-associated dystrophic neurites in mild or severe AD using double and triple immunolabeling. We also determined the argentophilia of the various sub-populations of dystrophic neurites. In aged non-demented brain, approximately half of the APP-positive plaques contained NF-immunopositive dystrophic neurites; rarely were PHF/tau-positive dystrophic neurites detectable. In contrast, in the AD brain, three-fourths of the APP-positive plaques contained NF-positive dystrophic neurites and half contained PHF/tau neurites. We also observed focal patches of hyper-phosphorylated NF and/or PHF/tau within APP-immunopositive dystrophic neurites, which appeared similar to retrograde degeneration, whereas we never observed focal accumulations of APP within NF- or PHF/tau-positive fibers. We hypothesize that plaque-associated dystrophic neurites within plaques develop in a particular sequence: APP-positive dystrophic neurites appear first and are non-argentophilic. This is followed by the appearance of NF-positive dystrophic neurites, where a subset of NF-positive dystrophic neurites are lightly argentophilic. Over time, PHF/tau-positive dystrophic neurites develop and are strongly argentophilic. These data suggest that dystrophic neurites can develop retrogradely from focal plaque damage to induce somatic and dendritic degeneration and potentially contribute to neurofibrillary tangle formation. Received: 22 September 1997 / Revised, accepted: 15 April 1998