Abstract

All Down’s syndrome individuals develop Alzheimer’s disease (AD) neuropathology by the age of 40 years. To unite the two diseases under one hypothesis, we have suggested that classical AD, both of the genetic and late-onset sporadic forms, might be promoted by small numbers of trisomy 21 cells developing during the life of the affected individual. Recent evidence from several laboratories will be presented, which strongly supports the trisomy 21 hypothesis that defects in mitosis, and particularly in chromosome segregation, may be a part of the AD process. Specifically, genetic mutations that cause familial AD disrupt the cell cycle and lead to chromosome aneuploidy, including trisomy 21, in transgenic mice and transfected cells; cells from both familial and sporadic AD patients exhibit chromosome aneuploidy, including trisomy 21. The possibility that many cases of AD are mosaic for trisomy 21 suggests novel approaches to diagnosis and therapy.

Introduction

Alzheimer’s disease (AD) arises when neurons in regions of the brain involved in memory and cognition are damaged and ultimately killed. A key step in this process is the oligomerization or polymerization of the amyloid (A)β peptide that is derived from proteolytic processing of the Alzheimer amyloid precursor protein (APP). The Aβ filaments then aggregate in the brain as the characteristic neuropathological lesions known as amyloid plaques, which are thought to be the end product of the pathogenic pathway to neuronal cell death in AD.[1-5]

An important clue to the mechanism of AD was the discovery that Down’s syndrome (DS) patients who live beyond the age of 40 years develop brain neuropathology indistinguishable from that observed in classical AD.[6-9] DS is caused by the presence of three copies of chromosome 21, instead of the usual two, in every cell of the body from the moment of conception. The implication of this finding is that trisomy for chromosome 21 not only causes DS, but is also sufficient to cause AD later in life.[10]

These and other curious links between AD and DS led us to hypothesize that if complete ‘trisomy 21’ could lead to early AD in DS, perhaps the slow development of some trisomy 21 cells over a lifetime could cause or at least help promote the AD that affects elderly individuals.[11] This model explained many seemingly unrelated facts about AD and could also account for both the inherited and more common, nonfamilial form of the disorder, depending upon whether the defect in chromosome segregation that led to trisomy 21 mosaicism was the result of a genetic mutation or some environmental insult.

The trisomy 21 model for AD makes at least three important and testable predictions:[11]

Alzheimer patients should have accumulated a small number of trisomy 21 cells in their somatic tissues. Such trisomy 21 mosaicism would be produced over time by unequal chromosome segregation during mitosis;

There should be alterations in the mitotic spindle apparatus or in mitosis-related proteins in AD cells that could lead to chromosome missegregation and trisomy 21 mosaicism;

Mutations causing AD should occur in genes coding for proteins that are directly or indirectly involved in mitosis and chromosome segregation.
Each of these predictions has been tested by experimentation in our laboratory as well as others. Together, the evidence indicates that chromosome instability/missegregation and trisomy 21 mosaicism occurs in AD and likely contributes to the disease, with important implications for Alzheimer diagnosis and therapy.

Trisomy 21 Mosaicism in AD

An early hint that chromosome missegregation and trisomy 21 mosaicism might be related to AD was provided by epidemiological studies showing that women in some Alzheimer families in which the disease is inherited as an autosomal dominant mutation, gave birth to a significantly higher than normal number of DS children.[12,13] Other studies failed to confirm this finding, but the samples were too small for the lack of increased DS in Alzheimer families to be statistically significant.[14-16]

A strong association between DS and sporadic AD was also found in a retrospective study, which showed that young mothers of DS children have a fivefold greater risk of developing AD later in life, when compared with either older DS mothers, or the general population.[17,18] Schupf and her colleagues interpreted the effect as a novel form of accelerated aging. We interpreted this result instead, as indicating that the young DS mothers were most likely mosaic for trisomy 21 (as if they had a predisposition to chromosome missegregation that was reflected in their trisomy 21 offspring and their own increased risk for AD). Finally, trisomy 21 mosaicism has been found in a few women who came to medical attention because they developed Alzheimer-like dementia by 40 years of age.[19-20] These results demonstrate that trisomy 21 mosaicism can be present in some individuals with Alzheimer dementia. However, standard chromosome analysis of peripheral blood lymphocytes from sporadic AD patients has failed to show trisomy 21 mosaicism.

These suggestive but inconclusive and sometimes negative results led us to use an alternative technique to assess chromosome aneuploidy in AD. Specifically, we used FISH to determine the extent of aneuploidy, particularly chromosome 21 trisomy, in fibroblasts from Alzheimer and normal individuals.[21,22] FISH allowed us to better assess poorly dividing cells and avoid lymphocytes that are under strong selective pressure to remain normal. We found that fibroblasts from AD patients exhibited more than twice the frequency of trisomy 21 as did cells from age-matched, normal individuals (Figure 1). The greater frequency of trisomy 21 cells in AD patients compared with controls was significant (p = 0.007) and was not related to the age of the affected individuals. A small parallel study of chromosome 18 showed similar aneuploidy, indicating that the chromosome missegregation likely affected all chromosomes.
    
When this chromosome study was initiated, we found that chromosome missegregation occurred in all AD individuals, including those with sporadic AD and those who are now known to carry a familial (F)AD-causing mutation in either presenilin (PS)-1 or -2[23-28] Chromosome missegregation in sporadic AD patients has been elegantly confirmed in blood lymphocytes by Migliore et al.,[29-30] who avoided any in vivo selection bias by observing aneuploidy develop in real time. Trisomy 21 mosaicism, together with other chromosome abnormalities, has also been reported in the brains of sporadic AD patients in the laboratories of Herrup and Arendt.[31,32] The findings by the laboratories of Schupf that young mothers of DS children are at a fivefold increased risk of developing AD and also results by Migliore that they are mosaic for trisomy 21, are further confirmation that trisomy 21 mosaicism is commonly associated with AD. Indeed there is a hint of a dose-response effect, in that AD pathology is observable in DS brain by the age of 20 years, but not until middle age in FAD patients or later in the sporadic AD group to which belonged the trisomy 21 mosaic patients who were found to have had an increased frequency of DS conceptions. Furthermore, the recent discovery of families that develop autosomal dominant inherited AD only because the APP gene on one chromosome 21 is duplicated, indicate that the cause of Alzheimer’s in both DS and trisomy 21 mosaicism is likely to be the extra APP protein and Aβ peptide produced by the extra APP gene.[33,34]

Defects in Mitosis in AD

During the course of the studies described above, several lines of investigation provided independent evidence that defects in mitosis and/or in mitosis-specific proteins may be present in AD patients - supporting another prediction of the trisomy 21 mosaicism model.[35] Such defects could be expected to lead to chromosome missegregation, and thus could result in the trisomy 21 mosaicism and other aneuploidy observed in AD cells and individuals, and may also lead to other features of the disease, such as apoptosis and changes in APP processing.

The most direct evidence for mitotic defects in AD has been provided by experiments whereby the mitotic spindles in dividing cells from AD patients were observed to be abnormal. For example, treatment of cells from AD and control patients with the microtubule-disrupting agent colchicine, caused many of the Alzheimer’s cells to exhibit separated chromatids in metaphase spreads; the individual chromatids lay parallel to each other, separated by a clear gap, rather than being correctly connected at their centromeres.[29,36] This metaphase chromosome pattern is similar to the spontaneous premature centromere division that makes cells and patients prone to chromosome missegregation.[37]

Further evidence for the potential involvement of cell cycle defects in AD comes from the finding that both the APP and the microtubule-associated protein tau found in Alzheimer paired helical filaments become increasingly phosphorylated during mitosis [Geller and Potter, Unpublished observations].[38-41] Furthermore, phosphotau and other mitosis-specific phospho-proteins and enzymes are present in neurons of AD patients but not in neurons from normal subjects.[42-49] One possible reason why these post-mitotic cells acquire mitosis-specific proteins is that they have begun an aberrant mitosis. In turn, this could have led to a gain or loss of chromosomes, to changes in mitosis-specific gene expression, or to apoptosis in response to an untenable physiological state - any or all of which could stimulate the development of Alzheimer’s pathology.[28,31,49] Alternatively, the apparent tendency of Alzheimer brain neurons to attempt to re-enter the cell cycle may be a cell cycle defect that is independent from the chromosome missegregation occurring in neuronal precursors and glial cells.

 Potential Mitotic Function of the Presenilin Proteins

The fact that many Alzheimer individuals whose fibroblasts we analyzed for trisomy 21 mosaicism belonged to large families that were subsequently shown to carry FAD mutations in two related genes, PS-1 and -2, provided the first indication that FAD genes are likely to be involved in mitosis and chromosome missegregation - an important prediction of the chromosome instability model of AD.

The presence of multiple transmembrane domains and potential target sites for the mitosis-specific cdc-2 kinase in the presenilins, together with the chromosome missegregation results with PS-mutant human cells, led us to suggest that the PS-1 and -2 proteins might reside in the nuclear membrane and might be involved in chromosome organization and segregation during the cell cycle.[28] We then tested this prediction by determining the normal intracellular location of the PS proteins.[50] We found that both PS-1 and -2 in dividing cells localized in part to the centromeres, the nuclear membrane and the kinetochores, and during interphase they are associated with the nuclear membrane. Several other Alzheimer researchers have confirmed that a major location for the PS in the cell is the nuclear envelope, centrosomes and kinetochores.[51-53]

Functional results are also consistent with the PS being involved in mitosis. The FAD PS mutations have been found to inhibit the cell cycle [Tanzi, Personal Communication; also see below],[54] to increase sensitivity to apoptosis,[55-57] and to prevent the full-length proteins from being translocated to the nuclear envelope.[52] Confirmation that the PS proteins are involved in chromosome segregation and AD has recently been provided by the findings that a polymorphism in the PS-1 gene may be associated with both an increased risk of developing AD,[58-61] and also with an increased risk of having a DS child owing to a meiosis II defect.[62]

It is not clear how the PS proteins influence chromosome segregation. One possibility consistent with the data is that the FAD mutations affect the ability of the PS proteins to link the chromosomes to the nuclear membrane and to release them at the appropriate time during mitosis, thus leading to chromosome missegregation and other consequent abnormalities seen in cells carrying mutant presenilin genes, such as inappropriate apoptosis.

Chromosome Aneuploidy Induced in Transgenic Mice by Presenilin Mutation

Because mutant PS can cause chromosome aneuploidy such as trisomy 21 in humans, we asked whether the same held true in transgenic mice.[63] Whole brains from PS-1 transgenic and nontransgenic mice were processed to yield primary neuronal cultures. The isolated neurons were then hybridized with a mouse chromosome 16 BAC probe. Most neurons were disomic, that is exhibiting two signals, but up to 4% of neurons from PS-1-FAD (M146L) transgenic mice showed trisomy 16 (Figure 2) Neurons from the PS-1 wild-type (WT) transgenic mice or nontransgenic mice showed almost no trisomy 16 aneuploidy. Similar results were found in mice in which the PS-1 gene had been knocked in and thus was expressed in the correct temporal and anatomical pattern (not shown).

Chromosome Aneuploidy Induced in PS-1 Transfected Cells

To determine whether the aneuploidy observed in FAD-PS-1 transgenic and knockin mice was caused directly by PS-1 gene expression, we transfected PS-1 genes into the hTERT-HME1 cell line, a primary human mammary epithelial cell line that expresses the telomerase reverse transcriptase from a permanently-transfected hTERT plasmid (Clontech, CA, USA) and is thus both immortal and karyotypically stable.

Parallel cultures of hTERT cells were transiently transfected with WT PS-1, mutant PS-1 (M146L and M146V) and control empty vector (pcDNA3), and FISH was used to count the number of copies of chromosome 21 and 12. Overexpression of both WT PS-1 and FAD mutant PS-1 induced chromosome missegregation and the development of both trisomy 12  and trisomy 21 cells, with the mutant PS-1 showing a stronger aneugenic trend (Figure 3)

Potential Involvement of Other Genes and the Environment in Chromosome Missegregation in AD

Thus far, it is clear that trisomy 21 mosaicism is associated with many sporadic cases of AD, and there is strong circumstantial evidence favoring an induced trisomy 21 mosaicism as one explanation for how the FAD mutations in the PS genes exert their AD-causing effect. Thus, a key prediction of the trisomy 21 mosaicism model for AD has been, at least partly, fulfilled. Is it possible that other forms of AD also involve trisomy 21 mosaicism? Here the evidence is intriguing. For example, FAD mutations in the APP gene result both in apoptosis and in increased production of Aβ1-42 over Aβ1-40, which is also characteristic of trisomy 21 and of the PS mutations.[64-66] The APP protein not only becomes hyperphosphorylated during mitosis,[38] but APP phosphorylation also modulates APP processing.[67-71] Furthermore, endogenous APP localizes to the centrosome and the nuclear membrane in dividing cells, just as we have found for the PS proteins.[41,72-74] Also, in our study of aneuploidy in AD cell lines, the two lines harboring an APP mutation also exhibited increased trisomy 21.[22] Perhaps the APP mutations act not only directly, but also indirectly, on APP processing, by altering the cell cycle and its pattern of phosphorylation of proteins including APP itself. Finally, knocking out a gene that is highly homologous to APP, termed APLP2, in mice is lethal since the early embryonic cells undergo massive chromosome missegregation.[75-76]

The other major genetic risk factor for AD is the inheritance of the ε4 allele of apoE. Although the best evidence indicates that the apoE4 protein promotes AD by promoting the polymerization of the Aβ peptide into neurotoxic amyloid filaments,[77,78] it is possible that apoE4 has additional functions related to mitosis and chromosome segregation. Such a conclusion was reached, for example, by Avramopoulos and colleagues, who demonstrated that young mothers of DS children are significantly more likely to carry an apo ε4/ε4 genotype.[78] The authors suggested that the demonstrated inability of the apoE4 protein (compared with apoE3) to bind to the microtubule-associated protein tau[79] disrupts the balance of microtubule polymerization/depolymerization, and thus affects the mechanics of chromosome segregation. This explanation is interesting in light of the facts that a similar group of DS mothers was shown by Schupf et al. to have an increased risk of developing AD[17] and that the highest level of trisomy 21 mosaicism in our sample of AD patients was found in an apo ε4/ε4 individual. Further biochemical and chromosome segregation studies are underway to test directly whether apoE4 affects chromosome segregation.

Although improper chromosome segregation can result from a genetic mutation, it can also be caused by many environmental agents. Of the many exogenous factors that influence chromosome segregation, microtubule-disrupting agents such as colchicine, and low doses of radiation are perhaps the best studied.[80] Thus, it is possible that the large proportion of AD cases that arise in a sporadic manner not directly attributable to the inheritance of a genetic mutation can also be understood in the light of the chromosome 21 trisomy model.

 How Trisomy 21 Aneuploidy May Lead to AD

Several potential mechanisms could explain how the generation of trisomy 21 or other aneuploid cells could lead to AD.[10,35,62] For example, aneuploid cells might be prone to apoptosis. Indeed, cortical neurons from DS fetuses undergo spontaneous apoptosis in vitro.[81] Such apoptosis would lead to neurodegeneration directly, although the number of aneuploid cells is sufficiently small that their loss might alone not be very detrimental and therefore a non-cell-autonomous effect seems more likely. Furthermore, the level of apoptosis in the AD brain is debated. Alternatively, apoptosis could indirectly affect APP processing and the production of the Aε peptide. Support for this latter hypothesis is provided by the findings that DS fetal brains and adult sera contain a higher ratio of the pathogenic Aε1-42 compared with Aε1-40,[82] and that inducing apoptosis in normal human neurons by serum starvation or other treatments increases their secretion of the Aε peptide.[83,84] Furthermore, trisomy 21 microglia overexpress IL-1 and begin an inflammatory cascade that we have shown will lead to increased amyloid formation.[77,85] Finally, if aneuploidy in AD arises from a defect in microtubule function, then other effects should occur in parallel, such as poor protein trafficking and targeting in neurons.[86-88] Also, as we predicted some time ago based on its localization in mitotic structures and ability to induce aneuploidy, PS turns out to be a tumor suppressor gene.[50,89,90] Thus, trisomy 21 aneuploidy may be both an effect of microtubule dysfunction as well as a cause (through overexpression of the Aε peptide, which induces tau phosphorylation and lower ability to stabilize microtubules), thus generating a positive feedback loop that would further promote AD progression.

Implications of Trisomy 21 Mosaicism in AD for Future Diagnosis & Therapy

The mechanistic implication of the findings discussed above is that an early step in the pathogenic pathway to AD may involve a defect in chromosome segregation that leads to trisomy 21 mosaicism. This conclusion can be exploited in our search for more effective diagnoses and treatments for AD. For example, similar hypersensitivity of Alzheimer and DS individuals to cholinergic agonists or antagonists may serve as an ADdiagnostic test.[91] Another very straightforward potential diagnostic test based on our finding of trisomy 21 cells among Alzheimer patients’ fibroblasts, would be to use chromosome analysis to directly assess the level of trisomy 21 mosaicism in an individual in, for instance, skin needle biopsies or buccal cells from the mouth.

The implication of the data presented in this review that trisomy 21 mosaicism may be one of the first steps in the Alzheimer pathogenic pathway also suggests new approaches to AD therapy.[35] Drugs that would rectify the mitotic defect could be researched. Alternatively, it may be possible to strengthen the fidelity of the chromosome segregation process prophylactically. Several approaches can be envisioned. It may be possible to identify and eliminate environmental toxins that cause chromosome missegregation, or to develop agents that counteract the effect of the toxins. A more difficult, but potentially equally effective approach to therapy, might be to target trisomy 21 cells for removal from the body by exploiting unusual features of their gene expression or cell biology.

In summary, results from several laboratories show that individuals of all ages accumulate aneuploid cells with the majority of individual AD patients having two- to threefold more trisomy 21 cells than that of age-matched controls. The mechanism(s) by which these abnormal cells arise and whether/how they contribute to the pathogenesis of the disease are areas under active investigation. Further exploration of these novel findings has the potential to contribute to the development of diagnoses and therapies for AD.

Future Perspective

It is well established that cancer is partly caused by chromosome missegregation and the results presented and reviewed here indicate that AD may also be caused or at least promoted by this cell cycle defect. We and others are now attempting to determine the mechanisms linking the cell cycle to AD and are developing therapeutic methods to prevent the problem. On a more general note, in considering the developing literature and our own research in several areas, I am forced to conclude that chromosome mis-segregation, the resulting aneuploid cells, and their aberrant physiology may prove to underlie many diseases of aging, potentially including aging itself.

 Funding Information    
The research was supported by the Alzheimer’s Association grant IIRG-2 96-038, the Eric Pfeiffer Chair for Research in Alzheimer’s Disease at the Suncoast Gerontology Center at USF, the Johnnie B Byrd Sr Alzheimer’s Center and Research Institute and private donors. Additional funding provided by NIA grant AG09665.

Disclaimer
No writing assistance was utilized in the production of this manuscript.

Reprint Address
Johnnie B Byrd Sr Alzheimer’s Center & Research Institute, Eric Pfeiffer Chair for Research in Alzheimer’s Disease, Department of Molecular Medicine, University of South Florida College of Medicine, FL, USA. Tel.: +1 813 866 1600; Fax: +1 813 866 1601; hpotter@byrdinstitute.org

Future Neurol.  2008;3(1):29-37.     ©2008 Future Medicine Ltd.