Molecular mechanisms and therapeutic application of NSAIDs and derived compounds in Alzheimer's disease.

Alzheimer's disease (AD) is the most common form of neurodegenerative dementias worldwide. Amyloid-ß deposition, neurofibrillary tangle formation and Neuroinflammation are the major pathogenetic mechanisms that in concert lead to memory dysfunction and decline of cognition. To date, there is no curative treatment for AD. Epidemiological analysis support the notion that sustained intake of non-steroidal anti-inflammatory drugs (NSAIDs) reduce the risk and delay the onset of AD. In contrast, therapeutic studies testing NSAID efficacy in AD patients have not yielded positive results. This suggests that either the investigated drugs have not addressed the mechanism of action required for mediating beneficial effects or that NSAIDs are effective at stages way before clinical onset of symptoms. The NSAIDs concerned are pleiotrophic in nature and interact with more than one pathomechanism. Therefore evidence for more than one neuroprotective action of NSAIDs has been put forward and it seems likely that some of the drugs act at multiple levels through more than one molecular mechanism. Some, even may not only be beneficial, but negative actions may be overruled by protective effects. Within these mechanisms, modulation of γ-secretase activity, the activation of the peroxisome proliferator-activated receptor-γ, binding to prostaglandin receptors or interactions at the blood-brain barrier may account for the observed protection from AD. This article reviews the current knowledge and views on the above mechanisms and critically discusses current obstacles and the potential as future AD therapeutics.

A subset of NSAIDs lower amyloidogenic Abeta42 independently of cyclooxygenase activity.

Epidemiological studies have documented a reduced prevalence of Alzheimer's disease among users of nonsteroidal anti-inflammatory drugs (NSAIDs). It has been proposed that NSAIDs exert their beneficial effects in part by reducing neurotoxic inflammatory responses in the brain, although this mechanism has not been proved. Here we report that the NSAIDs ibuprofen, indomethacin and sulindac sulphide preferentially decrease the highly amyloidogenic Abeta42 peptide (the 42-residue isoform of the amyloid-beta peptide) produced from a variety of cultured cells by as much as 80%. This effect was not seen in all NSAIDs and seems not to be mediated by inhibition of cyclooxygenase (COX) activity, the principal pharmacological target of NSAIDs. Furthermore, short-term administration of ibuprofen to mice that produce mutant beta-amyloid precursor protein (APP) lowered their brain levels of Abeta42. In cultured cells, the decrease in Abeta42 secretion was accompanied by an increase in the Abeta(1-38) isoform, indicating that NSAIDs subtly alter gamma-secretase activity without significantly perturbing other APP processing pathways or Notch cleavage. Our findings suggest that NSAIDs directly affect amyloid pathology in the brain by reducing Abeta42 peptide levels independently of COX activity and that this Abeta42-lowering activity could be optimized to selectively target the pathogenic Abeta42 species.

Identification of amplified genes from SV40 large T antigen-induced rat PNET cell lines by subtractive cDNA analysis and radiation hybrid mapping.

Primitive neuroectodermal tumors (PNETs) such as human medulloblastomas are genetically heterogeneous and therefore poorly understood. In a rat model the SV40 large T antigen was used to induce neoplasms with characteristic features of PNETs. Tumor development requires a latency period of 8-11 months implicating secondary genetic alterations. To identify such secondary alterations we performed comparative analyses of two phenotypically identical PNET-derived cell lines. Indeed, these cell lines displayed distinct high-level amplification sites. Using a combination of subtractive cDNA analysis and radiation hybrid mapping we have now identified genes in the amplicon regions of the two cell lines. Interestingly, one of these genes encodes the rat homolog of a cytosolic branched chain aminotransferase (BCAT(C)) previously shown to be amplified in a mouse teratocarcinoma cell line. We propose that this simple cloning strategy may serve as a powerful tool for the isolation of genes implicated in known chromosomal aberrations in primary tumors and tumor cell lines.

Low frequency of SV40, JC and BK polyomavirus sequences in human medulloblastomas, meningiomas and ependymomas.

Several reports have suggested a role for polyomaviruses in the pathogenesis of human brain tumors. This potential involvement is not conclusively resolved. For the present study, a highly sensitive PCR-assay with fluorescence-labelled primers was developed to search for polyomavirus sequences in human brain tumor and control DNA samples. The assay was shown to detect approximately one viral large T-antigen (TAg) gene per 250 cells. We identified simian virus 40 (SV40)-like sequences in 2/116 medulloblastomas, in 1/131 meningiomas, in 1/25 ependymomas and in 1/2 subependymomas. A single case of ependymoma contained SV40 VP-1 late gene sequences. Moreover, one of the meningioma samples showed JC virus sequences. In contrast, 60 hepatoblastoma samples and 31 brain samples from schizophrenic patients were consistently negative. BK virus sequences were not detectable in any of our samples. Immunohistochemical analysis of two SV40 positive tumor biopsies failed to detect large TAg in the tumor cells. In the JC positive meningioma, immunoreactivity for the viral late gene product (VP-1) was not observed. Our data do not entirely rule out SV40 and JC virus as an initiative agent with a hit-and-run mechanism. However the low frequency of virus sequences and the absence of TAg protein expression argue against a major role of these viruses in the pathogenesis of human medulloblastomas, meningiomas and ependymomas.

Chromosomal imbalances and DNA amplifications in SV40 large T antigen-induced primitive neuroectodermal tumor cell lines of the rat.

Comparative genomic in situ hybridization analysis of four cell lines derived from SV40 large T antigen-induced primitive neuroectodermal tumors of the rat revealed non-recurrent chromosomal copy number changes and DNA amplifications at chromosomal bands 2q34, 4q43qter and 15q12qter in cell lines TZ102, TZ103 and TZ107, respectively. Semi-quantitative PCR and western blot analysis demonstrated amplification and over-expression of the rat N-ras proto-oncogene in TZ102. Furthermore, all cell lines displayed aneuploid cell populations and variable chromosome numbers as assessed by flow cytometry and cytogenetics. These findings suggest that DNA amplification as well as genomic instability may contribute to the pathogenesis of SV40 large T antigen-induced primitive neuroectodermal tumors of the rat.

Migratory potential of transplantable neural tumor cell lines.

Experimentally induced primitive neuroectodermal tumor (PNET) cell lines were transplanted into neonatal and adult rat brain and examined neuropathologically for their tumorigenic potential. Both cell lines showed a striking migratory behavior in both neonatal and adult brain. Migration of tumor cells was found in host brain parenchyma, along white matter tracts and associated with CSF pathways. These neural tumor cell lines provide a valuable tool for the development of strategies against strongly migrating neural tumors.

Analysis of presenilin 1 and presenilin 2 expression and processing by newly developed monoclonal antibodies.

Because distinct mutations in presenilin 1 and presenilin 2 are a major cause of early-onset familial Alzheimer's disease, we generated four monoclonal antibodies for the identification, localization, and investigation of presenilins in various cell lines and tissues from patients and controls. We show that these antibodies are specific for the N- and C-terminal domains of human presenilin 1 and presenilin 2. They recognize presenilin full-length proteins and their approximately 28-35 kDa N-terminal fragments and approximately 18-20 kDa C-terminal fragments. None of the antibodies showed cross-reaction in their specific detection ability. We demonstrated that presenilin 1 and presenilin 2 are proteolytically processed in human glioma cell lines, transfected and untransfected human neuroblastoma SH-SY5Y cells, COS-7 cells, rat cerebellar neuronal ST15 cells, mouse and human brain. Remarkably, we observed that presenilin 2 is alternatively cleaved during apoptosis, producing smaller C-terminal fragments. By analyzing the subcellular distribution of presenilins, we found reticular and fine vesicular staining throughout the cell bodies. In addition, staining of Golgi compartments and the perinuclear envelope was observed. Alzheimer's disease brain showed strong immunoreactivity of presenilin 1 in reactive astrocytes and senile plaques. This high expression of presenilin 1 may explain the increased production and accumulation of the amyloid-beta peptide in patients with sporadic Alzheimer's disease in the absence of familial presenilin mutation.

Increased oncogenicity of subclones of SV40 large T-induced neuroectodermal tumor cell lines after loss of large T expression and concomitant mutation in p53.

A model for medulloblastoma-like primitive neuroectodermal tumors was established in rat using retrovirally transduced SV40 large T antigen (LT) as an inducing agent (O. D. Wiestler et al., Brain Pathol., 2: 47-59, 1992). A cell line isolated from such a tumor and clonal derivatives thereof were biologically and molecularly characterized. In the parental tumor cell line, TZ870, which had been selected for G418 resistance, virtually all cells expressed LT and wild-type p53, which were complexed to each other. When plated in soft agar, these cells grew relatively slowly and formed disperse colonies. However, when grown without selection pressure, these cells reproducibly gave rise to LT-negative and G418-sensitive derivatives, LT-0 cells. Surprisingly, these latter cells exhibited a higher degree of malignancy both in vitro, growing readily to large colonies in soft agar, and in vivo, where they gave rise to a rapidly growing malignant tumor. Clonal selection from TZ870 cells revealed two types of clones: in one type, LT expression was stably maintained, even without selection pressure, whereas the other type lost the LT coding sequences. All LT-negative clones exhibited the same phenotype as the LT-0 cells. Reexpression of LT had no effect. However, LT no longer formed complexes with p53, and p53 was metabolically stable, suggesting that it had been mutated. Sequence analyses and diagnostic restriction digests of the p53 gene revealed that (a) both the parental LT-transformed cells and their derivatives contained only one complete p53 allele and (b) all LT-positive clones expressed wild-type p53, whereas all LT-negative clones expressed a mutant allele with a common mutation at Cys-174-->Tyr, indicating their clonal origin. We assume that the loss of LT coding sequences is the consequence of the p53 mutation, perhaps by inducing genomic instability, and that both the p53 mutation and additional genetic alterations that accompany the loss of LT coding sequences might contribute to enhanced malignancy.

Reverse relationship between beta-amyloid precursor protein and beta-amyloid peptide plaques in Down's syndrome versus sporadic/familial Alzheimer's disease.

Strong genetic evidence has been accumulated in favor of a central role of beta-amyloid precursor protein (APP) and beta-amyloid peptide (betaA4) in the pathogenesis of Alzheimer's disease (AD). We employed four newly developed APP and betaA4 antibodies and performed a comparative neuropathological study of patients with Down's syndrome (DS), early-onset familial AD and sporadic AD to investigate the distribution of APP and betaA4 plaque densities in the cerebral cortex of these disorders. Quantitative analysis of APP versus betaA4 plaques revealed that brains with early-onset familial AD and sporadic AD showed significantly more betaA4 plaques than brains with DS (P < 0.05). In contrast, APP plaques were more abundant in DS cerebral cortex (P < 0.02). These observations suggest that the development of pathological changes in DS brains does not parallel that observed in AD, which might be attributable to different causes in the pathogenesis of betaA4 formation. A comparison of these disorders may be useful to further complement our knowledge of the mechanisms leading to plaque development.

Prominent expression of presenilin-1 in senile plaques and reactive astrocytes in Alzheimer's disease brain.

Mutations in the presenilin 1 (PS-1) gene account for most cases of autosomal dominant early-onset familial Alzheimer's disease (AD). In order to elucidate the cellular expression profile of PS-1 we used a novel N-terminal monoclonal antibody against human PS-1. Immunohistochemical staining was observed strongly in senile plaques, and reactive astrocytes of gray and white matter. Neuronal immunoreactivity, however, was found to be only moderate. RT-PCR analysis of PS-1 mRNA revealed expression throughout human development as well as in human glioma cell lines. Altered PS-1 function may contribute to plaque formation in AD.

Characterization of neural cell lines derived from SV40 large T-induced primitive neuroectodermal tumors.

We recently reported intriguing properties of neural tumors generated by retrovirus-mediated transfer of the SV40 large T antigen into fetal rat brain transplants. Histopathologically, these neoplasms displayed characteristic features of primitive neuroectodermal tumors (PNET) and exhibited a striking potential for migration into the host brain. In the present study, four cell lines were derived from these PNETs and characterized. Two lines with an immature phenotype expressed the embryonal form of the neural cell adhesion molecule and nestin. They showed spheroid formation and delicate cell processes. The remaining cell lines had a flat, epitheloid appearance and were immunoreactive for synaptophysin, neurofilament proteins and glial fibrillary acidic protein. These cells constitute valuable tools to study the cellular origin(s) and molecular basis of PNETs, differentiation of neural progenitors and tumor cell migration in the brain.

Amyloid precursor-like protein 1 accumulates in neuritic plaques in Alzheimer's disease.

The Alzheimer's disease (AD) beta-amyloid precursor protein (APP) and the amyloid precursor-like protein 1 (APLP1) and 2 (APLP2) are members of a super-family of proteins that appear functionally related. Although APLPs are highly homologous to APP in the N- and C-terminal domains, they lack the beta A4/amyloid peptide, i.e., the main constituent of neuritic plaques in AD. To assess a potential role of APLP1 in AD, we have determined its immunohistochemical distribution in human hippocampal formation, a structure which is strongly affected in AD, and compared it with APP immunoreactivity. There was a considerable overlap of APP and APLP1 regional expression patterns. Significant APLP1 immunoreactivity was observed in neuritic plaques. Large pyramidal neurons of the subiculum showed an accumulation of APLP1 protein in their dendritic compartment. Some astrocytes elicited perinuclear APLP1 staining, but this was observed in both AD and control brains. These findings raise the possibility that APLP1 may contribute to the pathogenesis of AD-associated neurodegeneration.