Cell Cycle Defects and Chromosome Mis-segregation in AD

One mechanism by which Aß evidently causes neurodegeneration is by interfering with cell cycle. For example, we proposed that AD subjects would exhibit chromosome mis-segregation and the accumulation of aneuploid, particularly trisomy 21 cells and then used primary skin fibroblasts to demonstrate such trisomy 21 mosaicism and a trend toward chromosome 18 mosaicism in both familial and sporadic AD (SAD) patients [8]–[11]. AD lymphocytes also showed trisomy 21 mosaicism and premature centromere division, a mechanistic precursor to chromosome mis-segregation [9], [12]–[16]. Work from several laboratories confirmed and extended these results to buccal cells and brain neurons, with trisomy 21 constituting 10% of neurons in late stage AD brain [17]–[21].

We investigated the mechanism of this cell cycle defect and found that mutations in the presenilin 1 (PS1) and amyloid precursor protein (APP) genes that cause FAD directly induce chromosome mis-segregation and up to 20% aneuploidy in lymphocytes and neurons of transgenic mice and in transfected cells. Indeed the Alzheimer Aß peptide, the proteolytic product of the PS1 based γ-secretase enzyme on the APP protein and the key player in AD pathogenesis, also induces abnormal spindle structure and chromosome mis-segregation, including human trisomy 21 and mouse trisomy 16 in transgenic mice, transfected cells, and cell-free Xenopus egg extracts [22]–[24], [25]–[28]. We found that his profound defect in mitosis in AD results from Aß inhibition of certain mitotic kinesin motor proteins, including kinesin5/Eg5, that are essential for the structure and function of the mitotic spindle [24].

That chromosome mis-segregation and the consequent development of aneuploidy plays an essential role in AD pathogenesis is indicated by the finding that following the development of 30% aneuploid neurons during the early stages of AD, the specific loss of these aneuploid neurons in the transition from mild cognitive impairment to late stage AD can account for 90% of the neuronal cell loss observed at autopsy [21].

The finding of trisomy 21 mosaicism in AD is particularly interesting because of the universal presence of AD-like pathology and neurodegeneration in full trisomy 21 Down syndrome (DS) patients [8], [29]–[31], and the finding that early onset inherited AD can be caused by a single duplicated APP gene on one chromosome 21 [32], [33]. Evidently a 50% excess of APP and its product Aß are sufficient for the development of AD. Case studies of patients with trisomy 21 mosaicism and no intellectual impairments of the DS type who developed AD by age 40 demonstrates that a small percentage of trisomy 21 cells can, over a lifetime, lead to and/or contribute to the pathogenesis of AD [8], [25], [34]–[37].

Disruption of Cholesterol Homeostasis in Neurodegenerative Diseases

Because the generation and loss of aneuploid neurons appears to underlie the majority of AD neurodegeneration, it is essential to understand the mechanism by which such aneuploidy arises during the preclinical stage of the disease. Clearly, the overproduction of the Aß peptide may be partly responsible. However, defects in cholesterol homeostasis may also play a role in AD pathogenesis [38]–[43]. For example AD, and cardiovascular disease (CVD) share risk factors (e.g., high plasma cholesterol and a high cholesterol and/or saturated fat diet), and the severity of atherosclerotic lesions in AD brains correlates with the extent of AD pathology [38], [40]. Likewise, FAD transgenic mice fed a high cholesterol diet develop accelerated amyloid burden and steeper cognitive decline compared to animals on regular chow [39], [43]. Conversely, inhibition of endogenous cholesterol synthesis with statins has been associated with a reduced risk of AD and decreased amyloid deposits in humans and animals [42]. Finally, ApoE is the major cholesterol transporter in the brain, and inheritance of the ApoE ε4 allele is the strongest genetic risk factor for sporadic AD [44]. ApoE particularly ApoE4 directly associates with Aß and catalyzes its polymerization into neurotoxic assemblies [7].

A dysregulation of cholesterol homeostasis has also been implicated in other chronic neurodegenerative diseases, including Niemann-Pick (NPC), Huntington, and Parkinson’s disease [39], [43]. Specifically, a direct association between disrupted intracellular cholesterol trafficking and consequent neuronal loss, gliosis, and formation of neurofibrillary tangles (NFT) in individuals carrying a mutation in the NPC1 or NPC2 gene has been established. Loss-of function mutations in these genes cause sequestration of unesterified cholesterol in late endosomes/lysosomes in a number of cells and tissues, but are most deleterious for brain development and health, triggering progressive neuropathology, ataxia, dementia and premature death in early adolescence [43]. Recent studies utilizing human samples and cellular and animal models of NPC revealed interesting parallels between AD and NPC neuropathogenesis, which include endosomal/lysosomal dysfunction, glial-mediated inflammation, NFTs and Aß accumulation especially in ApoE ε4 carriers, cholesterol dyshomeostasis, and cell cycle reactivation [27], [28], [39], [43]. However, unlike AD [8]–[21], the presence of chromosomal abnormalities such as aneuploidy in peripheral and brain cells has not been reported in NPC.

These pathological correlations and our finding that Aß induces chromosome mis-segregation prompted us to investigate whether cholesterol might also induce aneuploidy in AD and possibly NPC and atherosclerosis. The latter hypothesis was reinforced by reports that each atherosclerotic plaque harbors a monoclonal proliferation of aneuploid smooth muscle cells [45]–[52].

To test these hypotheses, we asked whether cholesterol and lipoproteins induce chromosomal mis-segregation in vivo and in vitro and then investigated the mechanism of their observed novel aneugenic effect.

Cholesterol Dyshomeostasis Induces Aneuploidy in vivo

Niemann-Pick (NPC1) mutant neurons, glia, and fibroblasts accumulate elevated levels of trisomy 21.

To investigate the role of disrupted peripheral and central cholesterol homeostasis in chromosome segregation in a longer-term human neurodegenerative disease model, we turned to primary cells derived from patients diagnosed with Niemann-Pick Type C (NPC) disease (Table 1). Mutations and deletions in the NPC1 and NPC2 genes cause impaired trafficking of unesterified cholesterol (both low density lipoprotein receptor (LDLR)-internalized and endogenously synthesized) and other lipids (e.g., glycosphingolipids), which accumulate in late endosomes and lysosomes and fail to travel to the plasma membrane and endoplasmic reticulum (ER), resulting in severe cognitive deficits [56], [57]. Loss/defect in the NPC1 gene has also been shown to promote atherosclerosis in animal models [58]. Interestingly, as was first reported in AD, mitosis-specific proteins and/or phospho-epitopes have been found to be upregulated in NPC1 brains [27], [28], [53], [54], [59]–[65].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the Niemann-Pick Type C (NPC) and Control Samples.

https://doi.org/10.1371/journal.pone.0060718.t001

To test whether cholesterol mal-distribution affects chromosome stability in NPC disease, we assessed the level of aneuploidy in primary NPC fibroblasts and brain cells hybridized with FISH probes for chromosomes 12, 21, 14 and/or 18. We observed a 4-fold increase in trisomy 21 in NPC1 fibroblasts (Figure 2A,B) compared to normal fibroblasts and an increase of aneuploidy 12 that did not reach statistical significance (2.1% versus 0.4%, p = 0.12). Quantitative FISH analysis of Neu-N-positive and Neu-N-negative brain cells showed 4 to 6-fold increases in trisomy 21 in both neurons and glia with the NPC1 mutation, compared to the cells from control brains (Fig. 2C,D). However, there was no increase in aneuploidy of chromosomes 12, 14 or 18 in NPC brain cells compared to controls (0.55% vs. 0.24%, p = 0.11; 1.61% vs. 1.46%, p = 0.41; and 0.75% vs. 0.78%, p = 0.45, respectively).

These results are the first demonstration of a genomic instability and trisomy 21 mosaicism in Niemann-Pick disease and indicate that a cholesterol accumulation/localization defect causes chromosome mis-segregation in humans, as a high cholesterol diet did in mice.

Lipoprotein/cholesterol Induces Aneuploidy in vitro

Lipids and Cholesterol-induced Aneuploidy and Spindle Defects are Mediated through Membrane Fluidity and Calcium Homeostasis

Aneugenic Effect of Lipoproteins and Cholesterol: Implications for AD and NPC

The present data complement and extend previous work showing that up 30% of AD human and mouse neurons and other cells are aneuploid due to either environmental or complex genetic influences. Although cholesterol and other aneugens, such as mutant PS or APP genes or the Aß peptide, affect all chromosomes in cultured cells, the lack of selection against trisomy 21 neurons probably accounts for their preponderance in AD and NPC-1 brain. The finding that LDL/cholesterol induces chromosome mis-segregation in the absence of APP indicates that the aneugenic pathways initiated by Aß and LDL are, at least in their early steps, separate and independent. While our analyses of actively dividing peripheral cells and precursor brain cells were limited to a few human and only one validated mouse chromosome probe, a comprehensive investigation of other peripheral and brain cells (i.e., neurons and glia) is warranted to fully understand the rate and type of chromosome instability in relation to changed cholesterol homeostasis.

Although there is some indication that aneuploid neurons develop normally in human brains, possibly increasing neuronal diversity [17], [95], the large number of aneuploid neurons in AD and NPC must arise from mitosis, either from disease-related neurogenesis, from reactivation of the cell cycle in normally-post-mitotic neurons, or both. A number of studies reported neuronal aneuploidy and tetraploidy of several chromosomes in the human brains diagnosed with AD [17], [18], [20], [21]. Here we did not observe a cholesterol-dependent increase in tetraploid cells (i.e., cells that had duplicated every chromosome, but failed to divide, or had not yet divided), but mitosis specific epitopes have been observed in a small number of neurons in AD and NPC brain or in mouse models of AD, and have been interpreted as indicative of such reactivation of the cell cycle and the consequent development of aneuploid neurons [25], [27], [28], [53]–[64]. However, the frequency of such tetraploid (4 n) neurons in AD is certainly less than 10% of the frequency of aneuploidy cells (with DNA content between 2 n and 4 n) [21], and some researchers report that the number of tetraploid cells is actually the same in brains from AD and control individuals and are exclusively non-neuronal [96]. Nevertheless, if neurons are induced to re-enter the cell cycle in response to neurodegenerative disease, the resulting mitosis and cell division might well be abnormal and result in the observed increase in aneuploidy, rather than tetraploid neurons.

While we failed to observe a change in ploidy in the brains of mice fed a high cholesterol diet, possibly due to the 1) length of dietary exposure, 2) reduced neurogenesis, 3) young age of the mice, and 4) low or nonexistent exchange of cholesterol and its metabolites across the intact blood-brain barrier (BBB), a peripheral induction of trisomy 16 in mitotically active tissues may be an initiating event that precedes the induction of brain aneuploidy once the extracerebral damage reaches a critical point and the BBB is compromised.

In conclusion, AD and NPC1 exhibit cell cycle abnormalities, to which excess lipoproteins, particularly LDL, can contribute and lead to the chromosome mis-segregation, aneuploidy and neurodegeneration observed in these diseases. Furthermore, the fact that aneuploid neurons also characterize other cognitive brain diseases, such as Frontotemporal Dementia [97] (also Granic et al., unpublished), Down syndrome and Ataxia-Telangiectasia [20], suggests that chromosome mis-segregation and the development of aneuploidy neurons may be a general feature of neurodegeneration.

Chromosome Instability in Atherosclerosis: Implications for AD

Several studies have reported the hyperproliferation of aneuploid smooth muscle cells within atherosclerotic plaques [45]–[52], which our data suggest may be induced by elevated LDL-cholesterol. Specifically, during the initiation of atherosclerosis, a single smooth muscle cell that, as a consequence of aneuploidy, suffers altered gene regulation may lose checkpoint control or become more responsive to autocrine or paracrine growth factors (for example by becoming trisomy for chromosome 7, which encodes the PDGF receptor) and then hyper-proliferating to generate the observed thickening of the vascular wall. The fact that aneuploid cells are prone to apoptosis fits this model well, for apoptosis of smooth muscle cells is thought to be an essential contributor to the formation of atherosclerotic plaques [98]. Cholesterol induced aneuploidy would also be expected to initiate atherosclerosis in the brain, which can lead to the reduced blood flow and the development of ‘vascular dementia’ often seen AD patients [40]–[43], [99], [100].

Furthermore, the fact that inheritance of the ApoE ε4 allele is the strongest known risk factor for developing sporadic AD other than age [101] (i.e., ApoE being the major cholesterol carrier protein in the brain [102]), also strengthens the potential significance of lipoprotein/cholesterol-induced aneuploidy for the pathogenesis of AD. LDL and ApoE4 also promote atherosclerosis [103], which is characterized by hyperproliferation of aneuploid cells, as discussed above [45]–[52], extending the connection between lipid-induced aneuploidy and both CVD and AD. Finally, the fact that Aß causes chromosome mis-segregation and aneuploidy [23] and that atherosclerosis arises in the AD brain and in mouse models of AD [40], [104] suggests that Aß itself may trigger the mis-segregation event that initiates the atherogenic process in AD.

Mechanisms by which Aneuploidy may Contribute to Neurodegeneration in AD

There are several ways in which cholesterol (or Aß)-induced aneuploidy may contribute to cognitive decline in AD and related disorders. The most straightforward mechanism would be that aneuploid cells, with their altered genetic complements and thus altered gene expression are inherently prone to apoptosis/degeneration, as has been shown in many experimental systems [21], [105]–[107]. Indeed, as discussed above, the specific loss of (especially of non-trisomy 21) aneuploid neurons accounts for 90% of neurodegeneration in AD [9]–[23]. When the aneuploid neuronal progenitors and neurons harbor trisomy 21 or 16 and accumulate in the brains of AD patients and FAD transgenic mice respectively, overexpression of APP and other chromosome 21/16 genes will occur. That such imbalance may further contribute to disease onset and progression is evident in individuals harboring trisomy 21 mosaicism [25], [34]–[37] or an APP gene duplication [32], [33]. Similarly, the reduced neurogenesis and reduced neuronal number in Down syndrome [108], may result from the constitutive over expression of APP and production of Aß and the consequent induction of the cell cycle defect of chromosome mis-segregation.

Another possibility supported by the data, is that, like Aß, cholesterol disrupts chromosome segregation through inhibition of normal MT function. In addition to inducing defective neurogenesis as we demonstrate above, MT dysfunction could also affect neuronal plasticity. For example, we have found that Aß causes chromosome mis-segregation directly by binding to and inhibiting certain MT-dependent kinesin family motor proteins that are essential for mitotic spindle function [24]. These enzymes are also present in neurons, where they direct plasma neurotransmitter and neutrophin receptor localization and function at the plasma membrane [Ari et al. unpublished data]. As altered cholesterol homeostasis also disrupts microtubule structure and function, it could similarly affect receptor localization.

Finally, the counteracting effect of ethanol on lipid-induced chromosomal instability in vitro supports the hypothesis that the chromosome mis-segregation induced by lipoproteins is attributable to the increased membrane rigidity induced by cholesterol. We hypothesize that membrane stiffness upon cholesterol exposure may be restored by the fluidizing action of ethanol, making the cell less prone to mitotic errors. The fluidizing effect of ethanol on cell membranes is well established [83], [88], [89] and was proposed here as a possible mechanism for the reduction in aneuploidy in the cells co-incubated with cholesterol and alcohol, although no measurements of membrane fluidity were employed. Significantly, Aß peptide also increases membrane rigidity [109]. In the light of these data and the finding of aneuploid cells in AD brain and atherosclerotic plaques discussed above, it is interesting that moderate ethanol consumption has been linked to reduced risk of both CVD and AD [85]–[87], [110]. Prospective studies of microtubule stabilizers and/or ethanol or other modifiers of membrane fluidity may open new approaches to the prevention of neurodegenerative and cardiovascular disease.

Conclusion

In this study we have presented in vitro and in vivo evidence that cholesterol dyshomeostasis, a common risk factor for NPC-1 and two age-related diseases (i.e., AD and atherosclerosis) may exert its pathogenic effect in part through disruption of the cell cycle. Specifically lipoprotein/cholesterol-induced chromosome instability/mis-segregation may be an underlying cytogenetic trait and part of the pathogenic pathway of multiple disorders, which opens new avenues for research and therapy.

Cell Line

Primary Cells

Mice, Feeding and Diets

Housing of and procedures involving all animals, including the time-pregnant C57BL/6 mice were approved by the USF Institutional Animal Care and Use Committee.

Reagents and Kits

In Vitro Incubation with Lipoproteins/Cholesterol

Two days prior to treatment/incubation, freshly passaged hTERT cells (1–3×10⁵ cells/2 ml) were plated in 100 mm tissue culture dishes in MEBM to assure 60%–70% confluency on the day of the treatment. The cells were then incubated for 48 hr at the concentration of 20 µg/ml of lipoproteins or 2 and 4 µg/ml of WsCh.

Seven parallel experiments were used to assess the effect of lipoproteins on chromosome segregation (i.e., karyotype analysis and chromosome specific aneuploidy) and six separate experiments to test the aneugenic effect of cholesterol. The same cell passages were used for the set of parallel experiments involving Ca++ chelating reagents, BAPTA or EGTA performed as in [23].

Similarly, two days before lipid treatment, freshly passaged HASMC, human fibroblasts with and without a functional LDL receptor or hTERT-HME1 cells were seeded onto pre-coated (2 µg/ml of poly-L-lysine, Sigma) single glass chamber slides (8.6 cm² growth surface/well) (BD Bioscience) at the density of 100,000 and 200,000 cells per slide respectively. After 24 or 48 hr of lipid/cholesterol incubation, the cells were rinsed twice with 1X PBS without Ca and Mg (Cellgro), immediately fixed with either cold 3∶1 anhydrous methanol and acetic acid fixative or methanol, and incubated at −20°C for 30 min.

Mouse non-transgenic and APPKO spleen cells were similarly incubated with lipoproteins for 48 hr, harvested and fixed for FISH analysis as described [22], [23].

In Vitro Ethanol and Lipoprotein/Cholesterol Incubation

To assure an exact and consistent ethanol (EtOH) concentration in treated cells, we used a modified closed chamber system protocol by Adickes et al. [114]. Prior to ethanol and lipoproteins co-treatments, several pilot experiments were conducted to examine the viability of hTERT-HME1 cells in the closed chamber, and to establish cytotoxicity of 25 mM and 50 mM of EtOH for 6, 12 and 24 hr on cell morphology, proliferation and survival (Trypan Blue dye exclusion). Based on these results, triplicate experiments were conducted with 25 mM EtOH for 24 hr followed the next day by lipoproteins (20 µg/ml) and WsCh (4 µg/ml) and EtOH co-incubation for 48 hr. A small aliquot of EtOH containing media from two samples was collected at the baseline, 24 and 48 hr to measure alcohol concentration using colorimetric assay (Ethanol Assay Kit, BioVision). Lastly, the cells were harvested and fixed for FISH analysis.

Metaphase Chromosome Analysis

After colcemid treatment, hTERT-HME1 cells were harvested and the metaphase chromosomes counted as described [22], [23]. Genus 2.81® software was used for chromosome analysis (Applied Imaging). At least 40–45 metaphases were analyzed per each sample.

Fluorescence in situ Hybridization (FISH)

Prior to hybridization, all slides were aged at room temperature followed by hybridization procedures recommended by Abbott, Vysis and by protocols and mouse and human probes described above and previously [22], [23]. Hybridizations were done according to manufacturer’s instructions (Abbott, Vysis) for the selected probes in a HyBrite hybridization chamber (Vysis) followed by DAPI II (Abbott, Vysis) or Vectashield with DAPI (Vector Laboratories) counterstain.

FISH Followed by Immunocytochemistry of NPC1 Brain Cells

To assess neuronal versus glial cell aneuploidy in NPC1 and control brains, we used a combination of FISH and immunocytochemistry [23]. Briefly, immediately after FISH, brain cells were incubated in 1XPBS for 10 min, blocked in 10% goat serum/0.1% Triton X-100 PBS solution for 1 hr followed by overnight incubation in conjugated Ms X Neuronal Nuclei AlexaFluor 488 (Millipore) 1∶100 antibody prepared in 1% BSA/0.1% Triton X-100 1XPBS. After final washes in 1XPBS, the cells were stained with DAPI II counterstain.

Immunocytochemical Quantification of Double Strand Breaks

Double strand breaks, an indication of DNA damage were identified in cholesterol-treated hTERT-HME1 cells by immunostaining for 53 BP1-containing foci. First, cell pellets were fixed in 70% cold ethanol, dropped onto Frosted slides (Fisher Scientific) and air-dried. Second, immediately after fixation in 4% paraformaldehyde for 10 min, and three 1XPBS washes, the cells were permeabilized in 0.2% Triton X-100 1XPBS solution for 10 min, blocked in 5% BSA 1XPBS/0.1% Tween-20 for 30 min, and incubated in 1% BSA 1XPBS/0.1% Tween-20 solution containing anti-53 BP1 rabbit polyclonal antibody (Abcam), 1∶400 for 1 hr followed by AlexaFluor 488 goat-anti-rabbit antibody (Invitrogen), 1∶300 incubation for the same time period. After several 1XPBS washes, the cells were stained with DAPI II (Vector).

Immunocytochemistry for Spindle Analysis

Upon fixation in methanol, cholesterol treated hTERT-HME1 cells were co-immunostained for mitotic spindles and centrosomes with anti-α-tubulin (Sigma) and anti-γ-tubulin antibodies (Sigma) 1∶1000, followed by AlexaFluor 488 goat-anti-mouse and AlexaFluor 594 goat-anti-rabbit secondary antibodies (Invitrogen, Molecular Probes), 1∶2000 incubation, respectively and Vectashield mounting medium with DAPI staining (Vector) as described [23].

Oil-Red-O Staining

Cell suspension of snap-frozen livers (2 from mice fed a regular and 2 from mice fed a high cholesterol diet) were made in 1XPBS, dropped onto pre-wetted slides, and stained with the Oil-Red-O staining kit following manufacturer’s procedures for frozen tissues with slight modifications (American MasterTech). Hematoxylin stained single cells and cell aggregates (blue) were evaluated under phase microscopy (Zeiss Imager M1) for the presence of lipid droplets (red).

Image Acquisition and Analysis

FISH hybridization signals were analyzed according to Abbott/Vysis guidelines as described in [22], [23] using a Nikon Eclipse E1000 fluorescence microscope with a 4912 CCIR high performance COHU CCD Camera and Genus 2.81® software (Applied Imaging) or a Zeiss Imager M1 microscope with a CV-M4+CL high resolution camera and Axiovision4.6 software (Zeiss) or a Zeiss Axiovert 135 with AxioCam HRc (Zeiss). The frequencies of nuclei containing 0, 1, 2 or 3 53 BP1 signals were recorded in treated and control cells, and mitotic spindle phenotypes were analyzed as described [22]. On average, between 700 and 1,000 interphases per each sample/treatment were scored for aneuploidy, and 400 NeuN positive and NeuN negative NPC1 brain cells were analyzed for DNA hybridization signals. More than 450–600 cells for each of three samples of both WsCh-treated and control were assessed for 53 BP1 foci, and more than 45–80 spindles from each of four samples of both treated and control cells were analyzed for spindle abnormalities.

Hematoxylin and Oil-Red-O stained mouse hepatocytes and cell aggregates from mice fed a regular and high cholesterol diet were compared and contrasted for amount of red staining under phase microscopy (Zeiss).

Statistical Analysis

Paired Student’s T-test was used to compare the aneuploidy induced by various lipoproteins with or without Ca++ chelating reagents and ethanol in different cell lines across multiple experiments and in spleen and brain cells of mice fed different diets, and to compare the percentages of nuclei containing 1 to 3 DSBs foci and the levels of abnormal spindle events between treated and untreated cells. Independent T-test was used to compare the aneuploidy between NPC1 fibroblasts and NPC1 brain cells and appropriate controls. Three to 7 treatments with lipids/cholesterol for each cell line, and triplicate experiments with chelators and ethanol were conducted and scored for aneuploidy. Four to 5 primary NPC1 cells sources and controls and at least 6–12 mice were analyzed for each graph.