ALZHEIMER'S DISEASE AND BRAIN DEVELOPMENT: COMMON MOLECULAR PATHWAYS

Kelly Jordan-Sciutto and Robert Bowser

Department of Pathology, University of Pittsburgh School of Medicine. Pittsburgh, PA 15261

Received 11/9/97 Accepted 11/15/97

2. INTRODUCTION

Alzheimer's Disease (AD) has been the subject of intense investigation in recent years. Yet, in spite of many advances, the mechanisms leading to neurodegeneration remain elusive. As in other diseases of aging, it has proven to be a multifaceted process involving many gene products and cellular pathways only recently the subject of serious investigation. This review will examine the role of molecular mechanisms implicated in neural development that are re-expressed during neurodegeneration in AD.

AD is characterized by several neuropathologic features. These include beta-amyloid (A-beta) containing neuritic plaques, neurofibrillary tangles (NFTs), dystrophic neurites, and glial activation. Although the number of NFTs correlates best with disease progression and are used in staging AD progression (1), there is strong genetic evidence supporting a role for A-beta in disease progression (Reviewed in 2, 3). Yet, one must remember that aging is a prerequisite of AD, even in familial forms of AD. This suggests that AD is a disease that results from an accumulation of insults over a lifetime that can be accelerated by genetic defects in the amyloid precursor protein (APP), Presenilin 1, Presenilin 2 (PS1 and PS2), the presence of specific ApoE alleles, or additional currently unknown genes (2, 4, 5). Such insults have been reported to include accumulation of oxidative damage, mitochondrial defects, and activation of immunologic mechanisms (6-9).

In this review, we will discuss gene products that function during brain development and are re-expressed or re-activated in AD brain. Such proteins and the molecular pathways in which they function will hold important clues as to the regenerative and degenerative events that regulate neurodegeneration not only in AD but potentially in other neurologic diseases. A brief discussion of current models for AD will first be provided. Although neurotrophic factors, neurotransmitters and their respective receptors exhibit altered gene expression in AD, these will not be discussed in the present review. Instead, we will focus our discussion to transcription factors, cell cycle proteins and developmentally regulated genes and how these proteins may participate to control the formation and progression of AD.

2.1. The Amyloid Hypothesis

A-beta plaques consist of fibrils formed from a proteolytic product of the amyloid precursor protein (APP). APP has one transmembrane domain, a short cytoplasmic tail and a larger extracellular domain (For review see (10)). It is normally cleaved to form a 90 kD secreted protein that stimulates cell proliferation and mediate cell substratum adhesion or neurite outgrowth. In AD patients the APP protein is processed to a 40-42 amino acid fragment that includes part of the transmembrane and extracellular domains. It is this fragment that is found in A-beta containing plaques. Other studies also suggest A-beta plays an active role in AD neurodegeneration. Not only is A-beta localized to these pathologically affected regions of the AD brain, it has been demonstrated that aggregated A-beta is neurotoxic to cultured neurons (11-15) and upregulates a number of genes involved in apoptosis (Bax, Caspase 3, etc.) (16, 17). This is suggestive of a mechanism by which A-beta deposition may lead to AD.

The amyloid hypothesis is further supported by genetic studies that show a mutation on chromosome 21 in the APP gene which results in increased production the highly agregable A-beta1-42, develop premature AD (age of onset 40 - 65 years; For review see (2)). Finally, Down's syndrome patients which have trisomy of chromosome 21 also develop AD at an accelerated rate (40 years of age onset). Although there is much evidence to establish a role for A-beta in AD progression, current studies fail to explain how A-beta is produced, what is the connection with other existing pathology, and why it takes a lifetime to develop the disease.

2.2. The Presenilins, Support of the Amyloid Hypothesis?

Recent advances in understanding AD etiology have come from studying familial AD (FAD). Using affected families, two additional genetic defects have been identified on chromosomes 1 and 14. These genes, named presenilin 1 and 2 (PS1 and PS2), are highly homologous (67%) (For review see (2, 4, 5). The presenilins contain six to eight transmembrane domains and have been localized to the endoplasmic reticulum and golgi of neurons, suggesting a role in protein processing or trafficking (18, 19). Mutations in the presenilins may increase intracellular levels of calcium and increase oxidative stress, contributing to the neurodegenerative processes in AD (20). Both proteins are proteolytically cleaved between transmembrane domains 6 and 7 to form amino and carboxy terminal fragments which appear to exhibit different subcellular localization (21). For example, using antibodies generated against amino acids 331-360 (C-terminal PS1 fragment), it has been localized to amyloid containing plaques (22), but using an antibody to amino acids 263-280 (N-terminal PS1 fragment), PS1 has been localized to NFTs (23). However, the interpretation of these antibody results has become somewhat complicated by other disparate results. Using antibodies generated to different epitopes within either proteolytic fragment, different pathologic markers or even no lesions at all have been recognized as presenilin immunoreactive (24-26). Regardless of their subcellular localization, it is clear that PS1 and PS2 affect APP processing. Patients carrying mutations in PS1 or PS2 express higher levels of the amyloidgenic A-beta1-42 (27). Transgenic mice that express human mutant PS1 and humans carrying PS1 or PS2 mutations have increased levels of APP processed to A-beta1-42 (28). Similar results have been observed in cells transfected with mutant presenilins (29). These data suggest that PS1 and/or PS2 may regulate APP processing. This is further supported by recent reports of direct interaction between presenilins and APP demonstrated by co-immunoprecipitation and the yeast dihybrid screen (30-32). Overall, these data further support A-beta as a major factor in regulating FAD neurodegeneration. However, many questions remain, including the formation and significance of NFTs, and how disease is initiated in sporadic AD where the A-beta1-42 burden is significantly lower than in FAD.

2.3. Cytoskeletal Abnormalities

The other major neuropathologic feature of AD, NFTs, consist of paired helical filaments (PHF-tau) composed of polymerized tau protein (For review see (33-35). Tau plays an important role in microtubule organization, an activity that is regulated by phosphorylation on specific serine and threonine residues. PHF-tau is aberrantly phosphorylated at additional sites. Hyperphosphorylated tau is protease resistant and able to form PHF in vitro. Tau is detected in both NFTs and dystrophic neurites in AD brain. The quantitation of NFTs in specific brain regions is currently the best correlate to dementia. NFTs are present in early stages of disease within the entorhinal cortex and the CA1 region of the hippocampus, regions that show early and severe neuronal abnormalities. The presence of extracellular "ghost tangles" in later stages of AD suggests that NFT formation directly leads to neuronal cell loss (36). Thus, the priming event for AD may include activation of a kinase that inappropriately phosphorylates tau, or inactivation of a phosphatase. Although a variety of kinases have been found to phosphorylate tau in vitro (glycogen synthase kinase b, cyclin dependent kinase 5; reviewed in (37)), the promiscuity of kinases within in vitro assays precludes the confirmation of the kinase responsible for aberrant tau phosphorylation in vivo. Since the nature of tau phosphorylation is likely an early and necessary step, it suggests that some earlier event, such as activation of a signal transduction mechanism, exists to activate the tau kinase. Again, as with the amyloid hypothesis, we are left not only with intriguing results and hypotheses regarding AD formation but equally important unanswered questions.

2.4. Alternate Pathways

Elucidating the causes of AD must include reconciliation of the two existing models, both with ample scientific evidence supporting their importance. Recently, other models of AD progression have gained validation that may begin to link these two seemingly disparate events. Mounting evidence suggests that early events leading to AD include accumulation of oxidative stress and glial activation. The role of oxidative damage in AD is supported by the presence of glycated proteins, lipid peroxidation, and the expression in AD brain of proteins that reduce oxidative damage (7, 38-43). This model of increased oxidative stress and protein glycation is interesting because it accounts for the prerequisite aging prior to disease onset. Interestingly, the receptor for advanced glycation end products (RAGE) appears to bind A-beta (44, 45), further integrating this model into the current paradigms.

The role of inflammation in AD has recently received considerable support. Numerous reports indicate that patients who take non-steroidal anti-inflammatory drugs (NSAIDs) have reduced risk of developing AD (46-49). There is also evidence for microglial and astrocytic activation surrounding neuritic plaques, along with elevated levels of cytokine activity (50). However, these results do not prove that inflammatory mechanisms directly cause neurodegeneration, as opposed to being a secondary event that stimulates a response to the primary neuronal injury. Thus, additional studies are required to demonstrate that molecular cascades initiated by factors released by activated glia play a role in AD formation and progression.

One common theme that emerges from each model is the reliance on a change in the molecular constituents present in the cells. Whether it is a kinase, a receptor, oxidative stress, or inflammatory cascade, there is a common thread of alternative gene expression and activity in AD. Many of the genes alternatively expressed in AD function during brain development. Further investigation of alternative gene expression in AD will yield pathways, both novel and pre-existing, that underlie the mechanisms of AD and may unite the current hypotheses into a cohesive model of disease formation and progression.