Proctor and Gray. (2010). GSK3 and p53 - is there a link in Alzheimer's disease?
July 2013, model of the month by Martina Fröhlich
Original model: BIOMD0000000286
Alzheimer's disease (AD) is the most common form of dementia . It was first reported by Alois Alzheimer in 1906 and is most often diagnosed in people over 65 years (exception: early-onset AD). Most common symptoms are difficulty in remembering recent events, confusion, agitation, trouble with language, and at later stages loss of bodily functions and death. At present, there exists no possibility to cure the disease.
AD is characterized by the loss of neurons and synapses and the atrophy of the affected regions in the cerebral cortex, temporal and parietal lobe, and the hippocampus , regions that are important for language, processing of auditory and sensory inputs and memory.
Figure 2 Schematic representation of the model. A: reactions involved in p53 and Mdm2 turnover. B: rections involved in tau turnover and aggregation. C: reactions involved in the DNA damage response, Aβ production and aggregation. The image was reproduced from . The tables listing the individual reactions can be found in .
Under unstressed conditions, their model predicts no DNA damage, low p53 levels as well as low levels of Aβ plaques and tau tangles during the whole simulation (Figure 3A).
An irradiation event of 1 min applied 1 hour after start of the simulation resulted in DNA damage, oscillations of p53 and Mdm2 protein levels and a slight increase in Aβ and phosphorylated tau. However, under those conditions, no aggregates were formed and the DNA damage decreased over time (Figure 3B).
The same in silico experiment performed with reduced DNA repair capacity, resulted in an accumulation of reactive oxygen species (ROS). Aβ and tau aggregates slowly accumulated and in some simulations DNA damage starts to increase again after a prior decrease (Figure 3C).
To simulate what happens in ageing cells, the ageing process was artificially speeded up by increasing the aggregation rate. This resulted in about 76% of cells accumulating DNA damage. A high level of variability was observed between individual simulations, which might imply that the disease onset might be due to stochastic effects. Furthermore, their model predicts that tau tangles started to form prior to Aβ plaques (see Figure 3D).
Figure 1 The GSK3/p53 hypothesis for AD. For description read the main text. The image was reproduced from .
Two abnormalities in the brains of AD patients serve as diagnostic hallmarks of the disease , namely extracellular amyloid beta (Aβ) plaques and intracellular tau tangles. The amyloid precursor protein (APP) is cut by β- and γ-secretases, thereby releasing the Aβ protein, which forms plaques outside the cells and progress through the brain. Within the cells the microtubule stabilizing protein tau is hyper-phosphorylated. This leads to its detachment from microtubules and thereby to the destabilisation of the cytoskeleton and the formation of intracellular tau tangles.
It is still unclear, what the main cause of AD would be? Originally, it was believed to be a deficiency in the production of acetylcholine (cholinergic hypothesis), but this might not be the cause but rather a result of the widespread tissue damage. It might be possible that cytotoxic effects of Aβ aggregates (amyloid hypothesis) or tau tangles (tau hypothesis) are the cause of the disease.
Hooper et al.  proposed the glycogen synthase kinase 3 (GSK3) hypothesis for AD. Over activity of GSK3 leads to tau hyper-phosphorylation, increased production of Aβ and the expression of typical features of AD, such as inflammatory responses, reduction in acetylcholine synthesis and memory impairment.The tumour suppressor protein p53 interferes with AD related GSK3 signalling in various ways :
- p53 is reported to be increased in sporadic AD
- Aβ binds to p53 promotor and enhances transcription
- p53 affects GSK3β activity
- induces tau phosphorylation (maybe via GSK3β)
- co-immunoprecipitates with GSK3β after DNA damage
- binding of p53 activates GSK3β
- GSK3β activates transcriptional activity of p53
- GSK3 negatively regulates p53 levels through phosphorylation of Mdm2
Based on these observations, Proctor and Gray  proposed the GSK3/p53 hypothesis for AD: Under low levels of p53, p53 is bound to Mdm2, ubiquitinated and degraded. However, when the p53 levels increase due to cellular stress, p53 forms complexes with GSK3β. This leads to an increased activity of GSK3β, hyper-phosphorylation of tau and increased production of Aβ. That results in a positive feedback, by the production of even more p53 due to its enhanced transcription and diminished degradation (see Figure 1).
The model presented here by Proctor and Gray  is a modification and extension of two of their previous models ([5-BIOMD0000000188],[6-BIOMD0000000105]). It includes the turnover of p53 and Mdm2 and the interaction of p53 with GSK3β (Figure 2A), tau turnover and aggregation (Figure 2B) as well as Aβ production and aggregation, and DNA damage response (Figure 2C).
Figure 3 Simulation results. A: Example of an individual simulation of a cell at unstressed conditions, B: cell at stressed conditions (irradiation event for 1 min applied 1 hour after start of the simulation), C: cell with reduced DNA repair capacity, D: increased aggregation rate. (green: p53, red: Mdm2, grey: GSK3β_p53, purple: damaged DNA, blue: tau tangles, cyan Aβ plaques). This figure is a combination of few plots taken from various figures of .
- Squire et al. Fundamental Neuroscience. (2008).
- Wenk GL. Neuropathologic changes in Alzheimer's disease. J Clin Psychiatry. (2003), 64 Suppl 9:7-10.
- Hopper et al. The GSK3 hypothesis of Alzheimer's disease. Neurochem. (2008), 104(6):1433-9.
- Proctor and Gray. GSK3 and p53 - is there a link in Alzheimer's disease?. Neurodegener. (2010), 5:7.
- Proctor and Gray. Explaining oscillations and variability in the p53-Mdm2 system. BMC Syst Biol. (2008), 2:75.
- Proctor et al. An in silico model of the ubiquitin-proteasome system that incorporates normal homeostasis and age-related decline. BMC Syst Biol. (2007), 1:17.