Recent advances in dementia research in Japan: Alzheimer-type dementia


Kenji Kosaka MD, Department of Psychiatry, Yokohama City University School of Medicine, 3–9 Fukuura, Kanazawa-ku, Yokohama 236-0004, Japan.


In a previous article, recent reports by Japanese researchers on non-Alzheimer-type degenerative dementias were reviewed. In the present article, recent Japanese reports on Alzheimer-type dementia (ATD) are reviewed. Alzheimer-type dementia has received great attention and has been studied from various viewpoints in Japan as well as in Europe and the Americas. In Japan, although it was believed that vascular dementia was the most frequent dementia in the elderly, ATD has recently been shown to be the most predominant type of dementia. Such a great number of papers on ATD have been reported in Japan that mainly the clinical, neuropathological, biochemical and molecular biological research papers alone are reviewed here.


Alzheimer-type dementia (ATD) is a generic term: it encompasses Alzheimer’s disease (AD; presenile type) and senile dementia of the Alzheimer type (SDAT; senile type). Alzheimer-type dementia research has advanced markedly during the past 20 years. Since the first Japanese reports on ATD brains in the 1920s, numerous studies of ATD have been published in Japan. We review here the main recent ATD papers reported by Japanese researchers.


It had been thought that vascular dementia (VD) was the most frequent dementing illness in the elderly in Japan, followed by ATD. However, recent epidemiological investigations have revealed that ATD is the most frequent type of dementia in Japan.1,2 A great number of clinical studies of ATD have been published in Japan. Therefore, we review here only the main papers on biological markers for the clinical diagnosis of ATD.

Cerebrospinal fluid (CSF) is an important material for neurochemical investigations, because it may reflect the metabolism or conditions related to neurotransmitter systems in the central nervous system (CNS). Tohgi et al. conducted a series of CSF studies to understand the changes in neurotransmitter-related markers in ATD and its disease specificity.3 The CSF concentrations of acetylcholine, choline, serotonin and kynurenine were found to be significantly lower in ATD patients than in the normal controls. These findings were in good agreement with those in ATD brains. Yasuda et al. measured concentrations of the peptides derived from prepro-vasoactive intestinal polypeptide (VIP) and the peptides derived from prepro-somatostatin in the CSF of ATD patients.4 They found significantly reduced levels of total peptide histidine methionine-immunoreactivity (IR), peptide histidine valine-IR, total somatostatin-IR and somatostatin-28-IR, and unaltered levels of VIP-IR and somatostatin-14-IR in the ATD group. These results suggest that the processing of the prepro-peptides of VIP and somatostatin, which may have some significant roles in the ATD pathogenesis, may be altered in ATD brains.

Because several lines of evidence have suggested that an acute-phase response may be involved in the formation of amyloid deposits in ATD, there is an interesting hypothesis that an interleukin-1 (IL-1)/IL-6-mediated acute-phase response, which augments the amyloidgenesis, may occur in the ATD brain. To test this hypothesis, Yamada et al. examined the IL-6 levels in the CSF of ATD patients, and found that the levels were significantly decreased.5 In particular, the early onset ATD patients showed a much greater IL-6 reduction compared to the control group. Yamada et al. speculated that this reduction of IL-6 levels may be involved in the neurodegeneration of the ATD brain.5

The most important neuropathological features in ATD are the deposition of A/β protein (Aß) in senile plaques (SP), and the deposition of abnormally phosphorylated tau and ubiquitin in neurofibrillary tangles (NFT). Accordingly, the assay of the levels of Aß and tau in the CSF and/or blood is thought to be the most useful clinical diagnostic method. Aß is released from the amyloid precursor protein (APP) through cleavage by proteases, and is secreted into the extracellular fluid during normal cellular processing. Two major forms of Aß have different carboxy termini, Aß1–40 and Aß1–42(43). Kosaka et al. reported a significant increase of the percentage of plasma Aß1–42 in familial AD patients with APP717 mutation,6 but Tamaoka et al. found no difference in the plasma concentrations of Aß1–40 and Aß1–42(43) between ATD patients and normal controls.7 The diagnostic meaning of CSF-Aß thus remains unclear.

Based on the findings that serum and CSF levels of α1-antichimotrypsin (ACT) were significantly and specifically higher in ATD patients, Matsubara et al. concluded that these could be useful diagnostic markers for ATD.8 Yoshiiwa et al. showed that patients with sporadic SDAT had significantly higher frequencies of the A allele of the ACT gene, as well as of the ɛ4 allele of the apolipoprotein E (apoE) gene compared to controls.9 An elevated CSF tau level has been reported to be a useful marker for the diagnosis of ATD.10[11]–12 Isoe et al. also reported that the CSF tau level well reflects the severity of ATD and may be useful for clinical assessment in ATD.13 Urakami et al. studied the levels of superoxide dismutase (SOD), an antioxidase enzyme, in the serum, CSF and fibroblasts of ATD patients.14 Although the SOD levels in the serum and CSF did not differ significantly between the ATD and control groups, the skin fibroblast SOD and SOD-mRNA levels were significantly higher in the early-onset ATD group, while they were lower in the late-onset ATD group compared to the control group. The determination of the SOD level of skin fibroblasts may be useful for the diagnosis of ATD.

Based on the finding of an abnormal accumulation of phospholipase (PLC)-δ in ATD brains15 as described later, Matsushima et al. examined the protein level of PLC-δ and its enzyme activity in platelets obtained from ATD patients.16 Their western blot analysis disclosed that the protein level of PLC-δ was significantly higher in the cytosolic fraction prepared from the ATD patients, and that the PLC-δ activity in the cytosolic fraction from ATD platelets was significantly reduced. They concluded that these findings are consistent with those of ATD brains, and that aberrant phosphoinositide metabolism may be present in non-neuronal tissues as well as in the brains in ATD patients.

The pupil dilatation test by tropicamide, which had been reported to be useful for the diagnosis of ATD, was re-examined. Arai et al. described a limitation of this test as a diagnostic marker of ATD.17,18


The presence of numerous SP is the most important finding for the pathological diagnosis of ATD. A great number of reports on SP have been published in Japan. In the 1970s and 1980s Ishii and Haga identified immunoglobulins and also complements in amyloid fibrils of SP,19,20 and Kitamoto et al. established a new method using formic acid pretreatment for enhancing the immunoreactivity of amyloid.21 Using this method, Yamaguchi et al. and Ikeda et al. reported in detail the light and electron microscopic features of ‘diffuse plaques’.22[23][24][25]–26 Since then, SP have been classified into three types: diffuse plaques, primitive (immature) plaques and typical (mature) plaques. Diffuse plaques are now thought to be the early type of SP.

New silver impregnation staining methods such as the methenamine silver method and the methenamine Bodian method were devised for enhancing the staining of SP and NFT.27[28]–29 Using these methods, Kosaka et al. described in detail the distribution pattern of various types of SP in both ATD and non-demented brains.30 In addition, Iseki et al., Tsuchiya and Kosaka, and Yoshimura et al. described in detail the distribution and morphology of SP in subcortical areas including the amygdala, striatum, pallidum, brain stem and cerebellum of ATD brains.31[32]–33 Amyloid deposits in the cerebral white matter of ATD brains were reported by Uchihara et al. and Iwamoto et al.34,35 The latter researchers confirmed the close relationship between Aß deposits and blood vessels first pointed out by Miyakawa et al.36

Since it was disclosed that Aß constitutes the amyloid in ATD and is derived from APP, various antibodies against Aß and APP have been developed for neuropathological research. Iwatsubo et al. reported that diffuse plaques are Aß1–42(43)-positive but Aß1–40-negative, and that amyloid of typical SP and amyloid angiopathy consist mainly of Aß1–40.37 This means that Aß1–42(43) is the initially deposited Aβ species in the ATD brain, and that Aß1–40 does not appear until almost a decade later and then only in a subset of SP. Fukamoto et al. recently described amyloid Aß deposition in normal ageing as having the same characteristics as that in ATD.38 In addition, Takamatsu et al. found that subpial amyloid plaques in the cerebellum were invariably positive for Aß1–42(43), and that two-thirds of them were associated with Aß1–40.39 Itoh et al. reported that subpial Aß deposits and amyloid angiopathy showed a significant positive correlation in their severity.40 Akiyama et al. reported that granules in the glial cells of ATD patients are immuno-positive for the C-terminal but not for the N-terminal sequences of Aß-amyloid protein.41

The non-Aß component (NAC) of AD amyloid and its precursor (NACP) have received increasing attention since Ueda et al. discovered them in 1993.42 By immuno-electron microscopy, Ueda et al. revealed that NAC is localized on amyloid fibrils,42 and Iwai et al. demonstrated that NACP is present on the membrane of synaptic vesicles.43

The close relationship between apoE and ATD is well known. ApoE immunoreactivity in SP was shown for the first time by Namba et al.44 By using immunohistochemical methods on autopsied brains, Yamaguchi et al. found that SP were consistently labelled with apoE antiserum even in the early stage of SP formation.45 The extracellular NFT were also consistently labelled for apoE, whereas only a small minority of them were positive for Aß protein. These findings suggest that apoE accumulates in the early stage of SP formation and that this apoE accumulation precedes the Aß deposition in extracellular NFT and amyloid angiopathy. Nishiyama et al. used confocal laser scan microscopy to investigate apoE in SP in ATD brains with or without the ɛ4 genotype.46 Although there was no difference between the ɛ3/ɛ3 brains and ɛ4/ɛ4 brains in terms of the pattern of apoE deposition in SP and its correlation with Aß deposition, they found clear differences in the distribution and staining pattern of plaque-shaped deposits by apoE and Aß antiserum. The apoE deposits were generally larger than the Aß deposits in the same region, and several Aß deposits were included in diffuse plaque-like apoE deposits. Some apoE deposits did not show any Aß-immunoreactivity, and typical SP tended to be composed mainly of Aß-immunoreactivity. These findings suggest that the discrepancy between apoE and Aß depositions reflects the different stages of SP formation, and that apoE is deposited in SP before the beginning of the massive Aß deposition.

The appearance of numerous NFT is another important marker for the pathological diagnosis of ATD. A great number of papers on NFT have been reported in Japan. Using methenamine silver-stained sections, Iseki et al.47 examined the regional order of occurrence of NFT qualitatively and quantitatively, and concluded that in non-demented elderly cases, NFT appear in the temporal isocortex a considerable time after they are first formed in the transentorhinal pre-α, whereas in ATD cases this time-lag is much shorter. Ikeda et al. observed that degenerated neurite-bearing ghost tangles in which roughly dispersed 15-nm straight and occasionally twisted tubules were penetrated by proliferated astrocytic processes had lost their immunoreactivities to anti-NFT, -tau, and -ubiquitin antibodies.48 In addition, Ikeda et al. observed the coexistence of paired helical filaments and glial filaments in astrocytic processes within ghost tangles.49 Using a monoclonal antibody against paired helical filaments, Wakabayashi et al. found a large number of spherical NFT in the dentate granule cells of patients with ATD, which consisted of straight tubules.50 Umahara et al. demonstrated NFT and neuropil thread-like structures but no amyloid deposits in the indusium griseum of ATD patients.51 Ohara et al. examined the incidence of NFT in the subcortical nuclei of advanced-stage ATD cases; the highest incidence of NFT was found in the nucleus basal of Meynert, and in the severely affected group the raphe nucleus, anterior thalamic nucleus, locus ceruleus and claustrum also showed NFT.52 A scanning electron microscopic study by Itoh et al. revealed that the NFT consisted of two types of filamentous structures: straight and helical filaments.53

Neuropil threads were also referred to as curly fibers by Ihara.54 Using the Gallyas electron microscope method, Ikeda et al. observed neuropil threads not only in dendrites but also in axons.55 Iwatsubo et al. showed by double labelling for tau/ubiquitin that ~40% of the threads lacked ubiquitin immunoreactivity at one or both ends, where only tau immunoreactivity was present.56 This observation suggests the possible bi-directional growth of neuropil threads.

It has recently been speculated that apotosis plays a role in ATD. In a study of immunohistochemical staining for Bcl-2 protein, which has been found to protect neurones from apotosis, Satou et al. reported that immunoreactivity for this protein within neurones in ATD brains increases with disease severity.57 Nishimura et al. described Fas-positive astrocytes in ATD brains.58 In addition, evidence is accumulating that inflammatory components such as cytokines and complement proteins play a significant role in exacerbating neuronal damage in ATD.59 Terai et al. observed prominent staining for nuclear factor-κB in neurones and their processes, NFT and dystrophic neurites of the hippocampal formation and entorhinal cortex.60 An increased expression of IL-1 receptor antagonist in neurones and SP was reported by Yasuhara et al.61 Akiyama et al. provided evidence that microglias are cells with properties of myeloid lineage and that in ATD lesions, these cells are activated in a manner comparable to that observed in chronic inflammatory states.62

Inoue et al. described eosinophilic bodies, which were thought to be compatible with one type of axonal spheroids, as being scattered in the cerebral cortex of ATD cases.63

Some researchers tried to differentiate ATD brains from non-demented control brains by the quantitative investigation of SP and/or NFT.30,64 Mizutani et al.65,66 recently proposed new neuropathological diagnostic criteria for ATD: (i) exclusion of the diseases causing dementia; (ii) the numbers of SP and/or NFT above the upper limit of the physiological range; and (iii) marked laminar cortical degeneration in the second and third cortical layers. Those authors also described a neocortical type and a limbic type of ATD. Moreover, Mizutani and Kasahara suggested that the degeneration of the entorhinal cortex and the perforant pathway plays a considerable role in the development of the hippocampal atrophy.67

Subcortical lesions have also been studied in ATD. Morphometric examinations by Uchihara et al. revealed selective neuronal loss and an unexpectedly large number of NFT in the substantia nigra of ATD brains.68,69 In the advanced stage of ATD, the cerebral white matter is usually affected. Yamada et al. found that many reactive astrocytes in the white matter as well as the gray matter had cholesteryl ester transfer protein-like immunoreactivity.70 These astrocytes were thought to play a role in ATD pathology such as in tissue repair.

The occurrence of some patients with atypical ATD has been reported. Kato et al. proposed an occipital subtype of SDAT characterized by marked psychotic symptoms and dilatation of the posterior horns of the lateral ventricles.71 A patient with a ‘vascular variant’ of Alzheimer’s disease characterized by severe plaque-like β protein angiopathy was reported by Yamada et al.72


Since 1976, when a significant reduction of choline acetyl-transferase (CAT) activity was reported in ATD brains, many studies regarding the cholinergic hypothesis have been conducted worldwide. In Japan, Nakamura’s group and Arai et al. have led this field and suggested that ATD is a multineurotransmitter system disorder.73,74 Relevant studies have resulted in the development of new drugs for ATD, namely tacrine and E2020. However, the efficacy of these drugs is not sufficient, and further investigations related to the cholinergic deficits and other neurotransmitter abnormalities are now ongoing. Tanaka et al. used chronic administration of an acetylcholinesterase (AchE) inhibitor, ENA-713, to treat senescent rats.75 Although the acetylcholine level, CAT activity and the maximum number (Bmax) of muscarinic M1 receptor binding sites were significantly decreased in many regions of the treatment group brains, these changes could be prevented in the treated rats, suggesting that the drug may have protective, neurotrophic and therapeutic effects on ageing-induced cholinergic dysfunction, and may be useful for the treatment of age-related dementia. Nanri et al. also studied the effectiveness of cholinergic drugs, but their focus was on nicotinic rather than muscarinic transmission.76 The effects of subchronically administered GTS-21 [3-(2,4-dimethoxybenzylidene)-anabaseine dihydrochloride], a selective nicotinic agonist, were studied in rats with a lesion of the nucleus basalis magnocellularis. Neuronal cell loss was observed in the lesioned rats at layers II–III of the parietal cortex. However, when GTS-21 was orally administered once daily for 20 weeks after 2 weeks of the bilateral lesion state, the cell loss at these layers could be attenuated. Therefore, nicotinic agonists may have a protective action against the neuronal cell death and a beneficial effect on neurodegenerative disorders.

Some protocols of acetylcholine replacement therapy have been proposed. In Japan, tacrine is not available, and E2020 has been clinically evaluated. Since oestrogen administration was reported to increase CAT activity, Ohkura et al. intensively examined the efficacy of oestrogen replacement therapy in ATD patients.77 They found that a 6-week oestrogen treatment regimen improved not only the patients’ performance on psychometric assessments, but also their regional cerebral blood flow and electroencephalogram activity. As a second step, a long-term oestrogen regimen was administered; improved cognitive function, dementia and daily activities were the result in women with mild-to-moderate ATD. Mimori et al. reported an interesting finding regarding the property of AchE in ATD brains.78 In contrast to the AchE obtained from normal human brains, AchE present in the SP-rich fraction isolated from ATD brains or AchE solubilized from SP was not so markedly inhibited by many AchE inhibitors, including physostigmine, E2020, amiridine, tacrine and nicegorine. This finding is of importance, because this hydrophobic property of anomalous AchE may serve as a ‘seed’ for amyloid fibrils in SP. Irie et al. have developed radioactive analogs, N[14C]methyl-4-piperidyl acetate and propionate, as radiotracers for measuring brain AchE activity by positron emission tomography (PET) study.79

In addition, their group reported that the tracers had successfully sufficient sensitivities to detect AchE activity changes in rat cortex with cholinergic deficits, indicating the usefulness of the tracers for the early diagnosis of ATD using PET.80 This method can be also used to test the suitability of cholinergic drugs for treating demented patients. In addition to studies of the cholinergic hypothesis, ATD brains have been analysed chemically in terms of the pathogenesis of ATD. Shimohara et al. investigated changes in the PLC-δ protein level and its specific activity in ATD brains.15 Phosphoinositide-specific PLC is a key enzyme in signal transduction, and the immunoreactivity of its iso-enzyme, PLC-δ, was detected in NFT in ATD brains. Western blot analyses disclosed that the concentration of PLC-δ protein was significantly higher in the ATD cortex, while its activity, that is, the hydrolysis of phosphatidylinositol, was not significantly different between the ATD and control brains. These findings indicate that the specific activity of PLC-δ is decreased in ATD brains, suggesting that this inactivation might be related to the neural dysfunction or neurofibrillary changes in ATD brains.

Kumashiro et al. and Nagata et al. measured the concentration of free D-serine in ATD brains.81,82 Recent studies have raised the possibility that D-serine is an endogenous positive modulator of N-methyl-D-aspartate (NMDA) receptors in brains, and disturbed neural transmission via the NMDA receptors has been implicated in ATD as well as in schizophrenia. In light of these findings, D-serine was expected to be involved in the pathophysiological processes of ATD. However, no alteration in the concentration of D-serine was detected in ATD brains, suggesting that the metabolism of D-serine was relatively constant despite the cerebral dysfunction in ATD brains.


Recent molecular biological and molecular genetic studies have successfully revealed several genetic foci in familial AD (FAD). Familial AD is now thought to be a polygene disease. Although there are many different mutations and modulating foci, the most important and compelling topics in this area of ATD research are the depositions of Aß and tau protein.

Aß protein and APP

It should be noted that several Japanese research groups have been leaders in the early stage in these fields. In particular, the studies on Aß protein by Kitaguchi et al. and those on NFT by Ihara’s group and Nishimura’s group contributed to breakthroughs in ATD research.83[84]–85

Because the physiological functions of APP are as yet unknown and some effects of APP on neural development have been speculated, Hayashi et al. investigated the effect of APP on the proliferation of neural stem cells.86 Two secretory forms of APP (sAPP770 and sAPP695) were purified from conditioned media of COS-7 cells transfected with genetically modified APP cDNA. Both secretory APP promoted the growth of neural stem cells, but the effect of sAPP770 was greater than that of sAPP695. These findings suggest that APP possessing the protease inhibitor domain regulates the growth of neuronal precursor cells during the development of the nervous system.

Recent studies on the mutations of APP using transfected cells showed that the APP670/671 mutations located immediately adjacent to the amino-terminal of Aß protein increase the Aß secretion by five to eight times, whereas the APP717 mutation located close to the carboxyl-terminus of the Aß protein increases the percentage of Aß1–42.

Studies using the transgenic mouse method are now common and have provided much useful information regarding the pathology of Aß deposition and pathogenesis of ATD. Shoji et al. successfully generated transgenic mice carrying many types of Aß gene with or without mutations.87 First, they generated transgenic founder mice overexpressing a 99-amino acid carboxy-terminal fragment of the human APP, based upon a chicken β-actine promotor combined with a cytomegalovirus enhancer to obtain a possible overproduction of Aß protein. Concerning the Aß deposition, although the brains of the transgenic mice did not show consistent Aß depositions immunohistochemically, the pancreas showed a massive Aß deposition, accompanied by cell degeneration.87

Moreover, some of these founder mice exhibited decreased behavioural activity with neurodegeneration and gliosis in the hippocampus and decreased levels of synaptophysin, a synaptic marker.88 These findings suggest that these mouse models are suitable for investigations of APP processing in vivo, and that the overexpression of the carboxyl-terminal fragment of APP has the potential to elicit the Aß deposition in the pancreas, which might induce cell degeneration and also neurodegeneration in vivo without an appreciable production of Aß fibrils. Further important clues can be expected to be obtained in future studies.

Apolipoprotein E

Apolipoprotein E initially received attention because Namba et al. had found ApoE-LI in SP and NFT in ATD brains.44 Several researchers reported that the incidence of ApoE ɛ4 allele was significantly associated with sporadic late-onset ATD.89 The apoE ɛ4 allele is now considered to be the most common risk factor for ATD. Regarding the specificity of ApoE4 association with ATD, Isoe et al. examined the phenotypes of ApoE in the plasma, as well as the ApoE mRNA level in skin fibroblasts in ATD and VD patients.90 In both groups, the frequency of ApoE4 was significantly higher and the skin fibroblast ApoE4 mRNA level was significantly lower than in the control group. These findings suggest that ApoE4 may be a risk factor not only for ATD but also for VD. Isoe et al. also speculated that the reduced level of ApoE mRNA may disturb its neurotrophic function in the CNS, and may contribute to the development of dementia in both ATD and VD.90 However, Itoh et al. investigated whether the ApoE genotype influences the severity of cerebral amyloid angiopathy in elderly individuals with or without ATD, and their data suggested that the ɛ4 allele is not a risk factor for cerebral amyloid angiopathy in elderly people.91

The relationship between ApoE and Aß protein may be one of the most important clues to disclose the pathogenetic mechanism of ATD. The localization of ApoE in brains has been intensively examined, and the recent Japanese studies have been noted here. The next question is which cells, neurones or glial cells produce ApoE in the brain. Although ApoE was postulated to be synthesized by astrocytes and taken up by microglias and neurones, reactive microglias have been observed close to amyloid cores of SP, while astrocytes have been seen forming outer shells around SP. Nakai et al. addressed the question of the source of ApoE by using a rat primary culture system and a reverse transcriptase-polymerase chain reaction (RT-PCR) analysis; they found the expression of ApoE mRNA in cultured rat microglias.92 An RT-in situ-PCR also showed positive staining for the PCR product of ApoE mRNA. Therefore, microglias could be one of the main sources of ApoE in the brain.

Presenilin 1 and 2

Recent molecular genetic studies revealed that the majority of early onset FAD cases are caused by mutations in the presenilin 1 (PS1) and 2 (PS2) genes, located on chromosomes 14 and 1, respectively, and are highly homologous (67%). Tabira’s group has been analysing Japanese FAD families and sporadic ATD cases to identify reported or novel missense mutations in PS1, PS2 and APP genes.93 Concerning PS1 mutations, the group found a FAD patient with a C to A transversion, responsible for a missense mutation of 280 Glu→Ala (Glu280Ala), and another FAD patient with a G to C transversion, responsible for a missense mutation of 384 Gly→Ala (Gly384Ala). These mutations have been found in other countries, and seem to have importance in different ethnic groups. It is also of interest that one sporadic ATD patient had an A to G transition, responsible for a missense mutation of 163 His→Arg, which has been reported as a missense PS1 mutation found in FAD patients. This is the first report of a sporadic ATD patient having PS1 mutation.

More than 30 missense mutations in PS1 and PS2 genes have been identified to date in FAD pedigrees. However, information about the metabolism and biological functions of PS1 and PS2 is still incomplete.

To investigate the function of PS1, Suzuki et al. examined regional and cellular distributions of PS1 gene expression in normal brains and other tissues, using in situ hybridization histochemistry, RT-PCR and northern blot analysis.94 Their study provided the first report of the determination of the gene expression of PS1 in human tissues at the cellular level; PS1 mRNA was found to be expressed in various human organs including the heart, spleen, kidney, liver, testis and brain. It was of interest that the PS1 mRNA transcript was increased in the testis, since the PS1 putative product has considerable homology with the Caenorhabditis elegans sperm integral membrane protein SPE-4, which appears to be involved in the formation and stabilization of the fibrous body–membrane organelle complex during spermatogenesis. In the brain, PS1 mRNA was found to be expressed predominantly in neuronal cells, and only at a low level in glial cells.

Takashima’s group has also contributed to the progress of PS1 research.95[96]–97 They examined the subcellular localization by using a monoclonal anti-PS1 antiserum.95 PS1 was found to be localized on the cellular membrane in COS-7 cells overexpressing PS1. The finding of PS1-LI on cell–cell contact regions in the plasma membrane suggested that PS1 may have a role on the cell membrane as a cell adhesion molecule. Takashima et al. then used rat pheochromocytoma PC12 cells and a rat glioma cell line (C6) transiently transfected with PS1 constructs containing cDNA for wild-type or one of several Alzheimer-associated mutations (Met146Val and Ala246Glu,96 and Gly384Ala, Leu392Val and Cys410Tyr97) to see whether such mutations affect the PS1 processing step in which 47-kDa full-size PS1 generated in PC12 cells with transfection of the wild-type is usually processed into a 28 kDa N-terminal fragment and a 19-kDa C-terminal fragment. Interestingly, Met146Val, Ala246Glu and Cys410Tyr inhibited such a proteolytic processing of PS1, while the other mutations failed to generate the fragments. These data suggest that FAD-related missense mutations may act via a mechanism of impaired PS1 proteolytic processing, but not all of the mutations previously found in FAD have such effects, and can be divided into two groups in terms of their effect on such PS1 processing. Waragai et al. addressed the question of whether PS1 directly binds to APP by analysing the results of two-hydrid interaction assays between these proteins in yeast cells, using bait plasmids for normal and mutant PS1 and prey plasmids for APP fragments.98 Their findings indicate that PS1 binds to multiple portions of APP, and that these interactions are based mainly on the structure of PS1, and not on that of APP, suggesting that PS1 may be directly involved in the Aß metabolism. However, it should also be noted that the mutant PS1 had the same features in these interactions as did the wild-type PS1. Therefore, the occurrence of a direct interaction between PS1 and APP is not enough to explain the abnormal Aß metabolism which would be induced by mutant PS1; and the elucidation of steps after the direct interaction between PS1 and APP may explain the pathological features found in ATD brains. Similar examinations of PS2 have been conducted around the world, and many interesting findings and speculations have been reported. In Japan, for example, Tomita et al. expressed cDNA for wild-type PS2 and PS2 with a Volga German (Asn141Ile) mutation in cultured cells and examined the metabolism of the transfected proteins; PS2 was identified as a 50–55-kDa protein and usually cleaved to the fragments.99 However, Tomita et al. used COS1 cells doubly transfected with cDNA for the PS2 mutation and APP, and Neuro2a neuroblastoma cells transfected with the PS2 mutation, and they found that these cells secreted 1.5- to 10-fold more Aß protein ending at residues 42(43) compared with those expressing the wild-type PS2. These findings suggest that a PS2 mutation may alter the Aß/APP metabolism to foster the Aß42(43) production which readily deposits in amyloid plaques.

A longstanding and controversial question is which type of degeneration, SP/Aß deposits or NFT, is more important in terms of the pathogenesis of ATD. The presenilin studies seem to support the Aß cascade theory, in which the primary and pathogenetically important change is the deposition of Aß protein in ATD brains.