Genetics molecular biology and animal modeling of Alzheimers disease

Super Memory Formula

Dementia Holistic Treatment

Get Instant Access

P.H. St George-Hyslop1,2, P.E. Fraser1, D. Westaway1, J. McLaurin1, E. Rogaeva1, G. Schmitt-Ulms, A. Tandon1, H.T.J Mount1, J. Robertson1, and C. Bergeron1

Our group at the Center for Research in Neurodegenerative Disease, at the University of Toronto, first became interested in several aspects of the biology of Alzheimer's disease (AD) during the early 1980s. The relative homogeneity of the clinical and neuropathological features of (AD) had led to the prevailing assumption that AD was likely to be a single homogeneous disorder. At that time, standard biochemical methods were being applied to dissect the protein composition of both the amyloid plaque and the neurofibrillary tangle (NFT). While these biochemical studies were on the edge of providing important clues to the biochemical pathogenesis of AD, mechanistic insights into the disease remained elusive. One notable exception was the observation that AD clustered in some families and was often inherited as an autosomal dominant trait. This observation alone, however, was insufficient to provide much traction. Indeed, some early attempts to define the chromosomal location of the disease genes using classical blood group antigens as genetic markers had been largely unfruitful. We were aware that individuals with Trisomy 21 had an increased incidence of an AD-like disease after the age of 40 years. Unfortunately, at that time, there were no useful polymorphic markers for genes on chromosome 21. However, we knew that phosphofructokinase-liver type (PFK-L), an enzyme encoded on chromosome 21 and involved in energy metabolism is decreased in AD brain tissue. In 1984, in collaboration with Donald Crapper-MacLachlan, we measured PFK-L activities in the brain of patients with AD but failed to uncover any difference in enzymatic activity. While this finding made variants in PFK-L unlikely to be causal for AD, it did not exclude the possibility that other genes on chromosome 21 might be associated with familial AD.

Fortuitously, in 1983, studies by Ray White, Mark Skolnick and David Botstein led to the discovery that restriction fragment length polymorphisms (RFLPs), arising from nucleotide sequence variations in genomic DNA, could be used as tools to map the chromosomal locations of disease traits. James Gusella, who had been a classmate of P.H. St George-Hyslop and had worked as a postdoctoral fellow of David Botstein, was proposing to use RFLPs to map the gene for Huntington's disease (HD). This approach led St George-Hyslop, in 1985, to initiate, in Gusella's lab, the genetic analysis of families with autosomal dominant familial AD, beginning with RFLP markers on chromosome 21. Gusella was already independently generating these markers to make a genetic map of chromosome 21. The details of the various interactions that followed over the next several years, and which resulted in the collection of several very large families in collaboration with Luigi Amaducci, Amalia Bruni, Jean-Francois Foncin, Peter Frommelt, Linda Nee, Lorenzo Pinessi, Ron Polinsky, Daniel Pollen, Innocenzo

1 Centre for Research in Neurodegenerative Diseases, Departments of Medicine, Medical Biophysics, Laboratory Medicine and Pathobiology, Tanz Neuroscience Bldg, University of Toronto, Toronto, Ontario, Canada M5S 3H2

2 Division of Neurology and Toronto Western Hospital Research Institute, Toronto Western Hospital, Toronto, Ontario, Canada

Rainero, Sandro Sorbi, and many others, are presented in the book "Hannah's Heirs" by Daniel Pollen (1993).

The initial genetic linkage studies conducted in 1985-1987 using RFLP markers from chromosome 21 enhanced the suspicion that there was a genetic susceptibility locus for AD on the long arm of chromosome 21 (St George-Hyslop et al. 1987a), in close proximity to the location of the amyloid precursor protein (APP) gene (Tanzi 1987a) (but see also below). Even more importantly, these studies immediately allowed a direct test of the hypothesis that AD is a single homogeneous disorder (St George-Hyslop et al. 1990). Specifically, the collaborative analysis of a large cohort of pedigrees assembled in conjunction with Jonathan Haines, John Hardy, Alison Goate, and Christine van Broeckhoven led to the clear demonstration that a subset of pedigrees was linked to a region of chromosome 21, at or near the APP gene. However, these genetic linkage analyses also unequivocally demonstrated that a larger subset of pedigrees was not associated with chromosome 21. The resulting conclusion of etiologic heterogeneity in AD, which is now so "mainstream" that it is taken for granted, has had a profound effect on all clinical and basic research studies in AD. Every clinical trial and every basic research study now inspects their results based upon some concept of heterogeneity (e.g., upon age-at-onset, the presence of absence of Apolipoprotein E e4, etc.).

The discovery of non-allelic heterogeneity also had a very profound effect on the search for AD genes. Until that time, the obligatory assumption was that the disease was homogeneous, and therefore all genetic results had to be considered using a single locus model. This assumption had the confounding effect that positive linkage information from a subset of pedigrees linked to a given locus (e.g., to the APP on chromosome 21) would be obscured by negative genetic results arising from pedigrees with a genetic defect at another chromosomal location. However, upon the discovery of heterogeneity, it became appropriate to subgroup pedigrees according to some other a priori feature (e.g., age at onset) or to analyze single pedigrees individually. Concomitantly, in 1990, Christine van Broeckhoven and Blas Frangione reported that hereditary cerebral hemorrhage with amyloidosis of the Dutch type (HCHWA-D) was due to a missense mutation in the Ap domain of APP. These two sentinel observations led, in 1991, to a re-investigation of the APP gene for mutations just in the pedigrees showing linkage to chromosome 21. This research eventually culminated in the discovery of missense mutations in the APP gene in a handful of families with early onset familial AD (initially by Goate andMullan, and by us and others; Karlinskyet al. 1992). Subsequently, three other lines of evidence have suggested that, in addition to these missense mutants, AD might also be associated with other non-coding nucleotide sequence variants in the APP gene. First, strongly positive linkage results for chromosome 21 markers near APP gene had been generated for several of our original large families in 1987. These families turned out to have no missense mutations in the APP gene but to have missense mutations in the PS1 gene as the predominant cause of AD. While this result was initially ascribed to an artifact, it now seems likely that co-segregation of single nucleotide polymorphisms (SNPs) in non-coding regulatory sequences in the APP gene of these families may also have contributed to some of the risk for AD and/or to some of the variation in the AD phenotype in these families. Second, positive genetic association results were generated for SNPs near APP in several sporadic AD:control cohorts. These results are also interpreted to reflect weak disease-associated/disease-

Genetics, molecular biology, and animal modeling of Alzheimer's disease 233

modifying variants in non-coding elements of APP. Finally, a small number of FAD pedigrees have duplication of the entire APP gene, resulting in its misexpression.

The final important effect of the discovery of non-allelic heterogeneity was that it facilitated attempts to clone the other AD genes. This led to the discovery, in 1993, of the association between AD and Apolipoprotein E by Allen Roses, together with members of this group (Saunders et al. 1993), and the discovery in 1995 of mutations in presenilin 1 (Sherrington et al. 1995) and presenilin 2 (Rogaev et al. 1995) by this group.

In addition to the purely genetic experiments, our group also played a key role in showing that the presenilin mutations caused an alteration in APP processing with the increased production of Ap42 (Citron et al. 1997). This discovery, when coupled with similar observations about the effects of mutations in the APP gene, lent great weight to the focus upon Ap as a primary player in the pathogenesis of AD that had arisen from the seminal observation in 1984, by George Glenner, that the Ap fragment of APP accumulated in the brains of patients with Down's syndrome and AD. However, it was the discovery, by our group and by others, that pathogenic mutations in APP, PS1 and PS2 all altered APP processing, and all resulted in Ap accumulation in the brain (Citron et al. 1997), that conclusively proved that the accumulation of Ap was causal in the pathogenesis of AD.

Our group has also played a key role in characterizing the biology of the presenilin complex and its role in generating Ap. In addition to cloning the initial two members of this family, we defined their membrane topology, we showed that the presenilins existed in biologically active high molecular weight complexes, and we isolated three additional components of these complexes. We showed that the loop domain of PS1 interacted with p-catenin (Levesque et al. 1999). We isolated nicastrin (Yu et al. 2000) and showed that it was likely to be the substrate-binding molecule of the presenilin complexes (Chen et al. 2001). We further showed that the presenilin complexes were necessary for y-secretase activities (Donoviel et al. 1999) and that mutations in the presenilin complexes caused alterations in y-secretase activity by increasing the production of Ap42 (Citron et al. 1997). More recently we have identified TMP21 as a novel modulator of y-secretase activity (Chen et al. 2006). Significantly, TMP21 suppresses y-secretase-mediated production of Ap but permits physiological £-secretase cleavage, raising hope that it may be mimicked therapeutically.

In addition to these studies on the pathogenesis of AD, our group has also provided insights into the biophysics and assembly of Ap into neurotoxic oligomeric assemblies. Our high-resolution structures of an Ap amyloid fibril using magnetically aligned preparations of a central Ap domain provided clues as to the mechanism of amyloid assembly and identified potential targets for controlling aggregation (Serpell et al. 2000a). We unequivocally demonstrated that the structural similarity that defines amyloid fibers exists principally at the level of p-sheet folding of the polypeptides within the protofilament, whereas the different types of amyloid vary in the supramolecular assembly of their protofilaments (Serpell et al. 2000b). We also identified cofactors, such as glycosaminoglycans, chemical chaperones and membrane lipids, that modulate this process. We were one of the first groups to dissect the interactions of Ap with various lipid membranes (McLaurin et al., 1996,1998) and the first to demonstrate in situ that Ap disrupts membrane stability (Yip and McLaurin 2001). These observations suggest that the fibrillogenic properties of Ap peptide are in part a consequence of membrane composition, peptide sequence, and mode of assembly within the membrane. Furthermore, our elucidation of the mechanisms by which chemical chaperones control Aß assembly has led to an interest in naturally occurring small molecules, such as Hsp, as amyloid modulators (Yang et al., 1999).

Finally, we used our knowledge of the genetics of AD to generate robust animal models of AD, such as the TgCRND8 mouse (Janus 2000). This model, based upon a double mutant APP transgene, develops profound amyloid-based neuropathology, synaptic loss, significant microglial and astrocytic inflammation, defects in spatial learning and memory, and increased mortality. This mouse model, which has been shared with many investigators, has been invaluable in the preclinical evaluation of several potential therapeutics.

Since total brain Aß levels were not perturbed by clinically effective immunization, we inferred the existence in the brain of TgCRND8 mice of sub-varieties of Aß with differential biological effects (Janus 2000). This deduction, and the suspicion that Aß oligomer assemblies might be important in the pathogenesis of AD, led us to characterize the functional epitope recognized by Aß42-directed antibodies (amino acids 4-10) that are therapeutically effective as vaccines (Janus 2000). We showed that these vaccine- induced, anti-Aß antibodies specifically recognize Aß oligomers and fibrils in tissue sections but not monomeric and diffuse Aß deposits, an observation subsequently deployed by others as the TAPIR assay (McLaurin et al. 2002). Finally, we demonstrated that these antibodies have an anti-aggregant effect, disrupting neurotoxic Aß oligomer and fibril formation (McLaurin et al. 2002). The recognition of the significance of Aß oligomers arising from this and other studies was the basis for our discovery of a small molecule inhibitor of Aß aggregation. One such compound, scyllo-cyclohexanehexol, effectively inhibits several Aß-induced, AD-like phenotypes (including cognitive and memory impairment, cerebral amyloid deposition, gliosis, synaptic loss, and accelerated mortality) in the TgCRND8 model of AD (McLaurin et al. 2006). This compound, which has high oral and CNS bioavailability, will shortly enter therapeutic trials in AD.

In summary, between 1985 and 2006, our group made a series of fundamental discoveries that have been instrumental in forming the present mechanistic understanding of AD and in providing the basis for several different therapeutic targets currently being investigated clinically (e.g., y-secretase inhibitors, Aß anti-aggregant compounds, Aß vaccines). However, there is still great uncertainty about how Aß accumulation leads to the generation of tau-positive NFTs and ultimately to neuronal death. There is still uncertainty as to the identity of the remaining AD-causing genes (the four known genes account for only about one half of the genetic risk), and there is also little robust knowledge about environmental factors that increase risk for AD. Finally, the hypothesis that AD is initiated by the accumulation of toxic oligomeric species of Aß remains just that - a hypothesis. Although widely held, this hypothesis will only be fully validated by the observation of prevention or cessation of disease activity by anti-Aß therapies in human clinical trials. Until that has been achieved, the research to address these gaps must remain fully active.

Christine Van Broeckhoven
Christine Van Broeckhoven

Was this article helpful?

0 0
Unraveling Alzheimers Disease

Unraveling Alzheimers Disease

I leave absolutely nothing out! Everything that I learned about Alzheimer’s I share with you. This is the most comprehensive report on Alzheimer’s you will ever read. No stone is left unturned in this comprehensive report.

Get My Free Ebook

Post a comment