Monthly Updates
New: April 27, 2001
We have incorporated a new algorithm to determine
neuron health.
This is still work in progress. (Depending on
what web browser you are using, you may need to clear
your cache to get the new simulations to load.)
Summary of Updates: Oct 15, 2000
Changes to User Interface
We have eliminated the individual buttons for each cytokine
or chemical, and instead implemented a scroll-bar for selecting
the display type. This will allow a much larger number of options
to be incorporated into the display types.
Astrocytes activation states
The state of an astrocyte
(see summary) is now shown explicitly in the display
by the color of the cell. Cells become progressively darker as they
change from inactive to receptive, to motile. In the most
active state, the astrocytes secrete substances that block diffusion
and transport of chemicals.
Display astrocytic blocking effects
A new option for the background display is the level of astrocytic
effects including the local depression of diffusion and transport
of chemicals and cytokines.
Adjusted astrocyte rules.
The motion and behaviour of the astrocytes were adjusted
to be more realistic. Here are the
updated rules.
Implemented genetic typing
We are implementing the ability to select genetic type.
We are currently
testing this feature using APOE and its effect on amyloid fiber
deposition. Using the work of
Yamauchi et al 1999 we have incorporated the distinction between
the APOE 2, 3, and 4 types as a scaling of the rate that
soluble amyloid is converted to fibrillar amyloid.
E4 leads to the largest rate of fiber deposition, and E2, E3,
are at the ratios of 7/9, 7/12 of this maximal rate. This
set of ratios is based on the ratios of binding affinities of
APOE types to A-Beta.
To implement this feature, the program reads a parameter
file over the internet. This may lead to other
network applications.
Introduced two more case studies
We explain the progression
and variability of the
current simulation.
Summary of Updates: Sept 15, 2000
I. Consultation with Experts
Consultation with P McGeer in August pointed to important
areas that merit investigation and inclusion in the
simulation. These include: the effects of complement, of
apo-lipoprotein E, and of alpha-1-antichymotrypin on
the susceptibility to Alzheimer's Disease.
II. Research and Updates to Knowledge Base
Extensive updates to the knowledge base were
carried out in the above three new directions, i.e. on
complement,
apo-lipoprotein E, and
alpha-1-antichymotrypin.
Numerous papers, both quantitative and descriptive were
scanned, and their citations and abstracts incorporated
into the current knowledge base.
In some cases, citations are now linked directly to the
online journal articles. (Some restrictions apply
to remote use, or to users without subscriptions to these
e-journals.)
III. Summaries of the Literature
A comprehensive set of summaries
was prepared for the above three
topics. This will gradually be enlarged to include other
pertinent literature that has already been used in
implementing the past aspects of the simulation.
As further detailed quantitative
information is obtained, the simulation will
be enhanced to include these intermediates.
IV. Progress on Parameters of Neurotoxicity
We have initiated preliminary phases of identifying the
possible effects of cytokines on neurotoxicity by direct
experimental procedures. Discussions are under way with
Jill McEachern (graduate student) and Prof Chris Shaw
(UBC and Vancouver General Hospital) on in vitro and
in vivo toxicity assays. Currently, Jill is reviewing
the scientific literature to summarize what is known, what were
the experimental conditions, and how these new experiments
should be carried out.
V. Astrocyte Movement and Effect
A thorough revision was carried out for the role
of astrocytes, their motion ("gathering at plaque sites"),
and their effects. A full description of the current
rules is provided.
The speed of motion of astrocytes was obtained
from a newly-found
paper in the literature.
VI. Technical
Currently under consideration are ways to improve the
speed of the simulation and streamline the computations.
Summary of Updates: Aug 15, 2000
Two new simulation features have been added to enhance the
capabilities of the computation and visualization.
I. Batch Simulations
The user now has an option to run several simulations in parallel,
while varying any one parameter. (Select
Batch Runs.)
The results are compared and
plotted at the end of the runs. Plots of neuron health, and
of the final displays (both the amyloid level and the state
of neuron health in the region) are the final outputs.
II. Zoom Feature
The normal (single run) simulation window has a new zoom feature
which allows the user to get a magnified view of any selected
area in the simulation window.
III. Bibliography
The scientific literature base has been expanded. Several new
topics are being studied for the next round of expansion.
These include APO-E, alpha 1 antichymotripsin, and complement.
Information about the antagonistic and synergistic effects
of these intermediates is being assembled.
Currently in progress
A more accurate representation of the participation of astrocytes
is under progress. Evidence that these cells gather at plaques, and that
their secretions influence the development of the plaques indicates
that more careful details are needed here.
Future Plans
Currently being considered is the ability for
the user to add new
molecular intermediates to the simulation. The
program would prompt the user for the name, properties,
and parameters associated with the given chemical,
the cells that secrete it, and approximate uptake
rates. This will allow customization of the software
for individual explorations.
Summary of Updates: July 1, 2000
I. Documentation
The first phase in extensive documentation of the source code was
undertaken this month. This was the largest
single area of progress. Further documentation is planned.
II. Refinements of the Simulation
The appearance of the simulation was refined in the following
ways:
- (a) The simulation now has three check boxes that
toggle between the view with images for cells and neurons
and those with the old format (disks and white background).
- (b) The neuronal health display has been improved. By
increasing the number of blocks of tissue that are displayed,
we show the spreading wave of neuronal stress in a more accurate
way. This eliminates the large black rectangular regions that appeared
in the previous displays of stressed neurons.
- (c) The graph of neuronal health has been improved with
better labeling of the axes.
The speed and efficiency of the simulation is undergoing
evolution:
- Currently under progress is redesign of the algorithm
to compute chemical diffusion - from the previous explicit
method to an implicit method that will save on computation
time and speed up the simulation.
- Other techniques for
speeding up the simulation are under consideration based
on the recent Java 1 conference and discussions with
local Java expert.
III. Validation experiments
A series of tests was carried out with the new settings and
parameter values and is described under the heading
Behaviour and Reproducibility.
IV. New functionalities under design
- Initial stages in the planning of automatic recomputation
over a range of parameters have been carried out. The
design of this system will gradually evolve.
Summary of Updates: June 1, 2000
I. Progress on Parameter Estimation
An extensive update of numerous parameters was carried
out this month. We identified and summarized literature about the
cell-surface receptors for the cytokines IL-1B, IL-6, and
TNF, and found cited values for the rates of binding and
unbinding of ligands to these receptors.
A table
listing these results is available. We used values
for the number of receptors per cell and the number of cells per
unit volume to convert these values to realistic rates of uptake of
cytokines. A web page listing the literature values was
created, and linked to extensive documentation about the
scientific literature on this subject.
We also refined our estimates of amyloid fibrillogenesis by using
a number of citations about the effects of ionic composition
similar to that of the brain on the aggregation of soluble amyloid.
Updates for the supporting
web page are under progress.
II. Refinements of the Simulation
The simulation underwent major revision to correct
for the way that cytokines and amyloid are removed by cells.
This resulted in fewer parameters, more well-defined parameters,
and changes in the assumptions (previously ad hoc) to
reflect mass action kinetics of receptors and ligands.
In a future revision, we will be able to incorporate receptor
down-regulation and internalization, though this is not yet
in place.
A significant revision was carried out to the visual aspect
of the simulation: images of actual neuronal tissue can now be
incorporated as a possible background, and as the glial
cells are represented by more realistic cell icons, in place
of ovals in a new version of the program.
This change was in response to a suggestion by
Prof Sue Griffin to improve the visual attributes. It
required some new programming tricks (double buffering)
to eliminate problems with flashing and choppy display.
In future refinements, we may show the changes in morphology that
accompany activation of cells, or a variety of neuronal tissue
background.
We revised the way that cells are represented in the region.
The size of our current volume is 16 x 105 cubic
microns. Thus, a lower estimate for
number of microglia in this region would be 500
and an upper estimate would be 1800. As it is unreasonable to
display and recompute positions of so many individual cells,
we now use each moving "unit" to represent the effect of
multiple cells while preserving the proportions
40 microglia : 40 neurons : 100 astrocytes.
Two of the parameters that we have added, "microglia
concentration" and "astrocyte concentration" are the factors
used to designate how many cells are represented by one
moving "unit". We are thus able to more accurately depict the
effect of the glial cell density in the region.
III. Content and Format Changes
There were numerous cosmetic changes as requested
during the month, including links on most pages to navigate back to
overview topics.
IV. Validation Experiments
A number of experiments have been carried out with the
current simulation, as described in detail on
summary web pages.
These are among the earliest scientific
results of the simulation, and use some of the
features that we have incorporated. One example is the ability to
use the mouse to "click away" a stimulus source
during the course of a run. We used this feature to
explore the consequences of removing the initial stimulus
at various times.
These experiments show the effect
of removal of the early stimulus, and the resulting
delay in development of partial neuronal degeneration.
More validation experiments are
planned.
V. Other Changes
Further updates of the literature base and extensive links
from papers to tables, to abstracts were incorporated.
VI. Future Plans
- Documentation of the code will take most of June.
- The ability to run multiple simulations, while
varying one parameter will be considered, and initial
stages in this programming task will be planned.
- Improving the neuronal stress and death estimates will
be the next parameter estimation priority, for the month of July.
- Investigating how to scale the simulation to longer time scales
will be considered in the upcoming months.
Summary of Updates: May 1, 2000
I. Progress on Parameter Estimation
Discussions were held with Prof S Griffin to narrow down
parameter values for amyloid critical concentration at
which fibers form, and for the neurotoxicity of cytokines.
Prof Griffin was unable to help us identify these values.
We are continuing to search the literature.
II. Refinements of the Simulation
Astrocyte chemotaxis and motion was removed following
advice by neuroscientists. The astrocytes are now stationary.
III. Content and Format Changes
An extensive list of references and abstracts were put onto the
web, and arranged by topic. This was one of the major
labour intensive developments
of the past month.
Important aspects of these citations
were highlighted. Where possible, links from parameters to these
references were indicated. This base of references will be
expanded and further links will be made in due time.
IV. Other Changes
The simulation now includes a new feature: the ability to
remove a source of stress-inducing signal at any point
during the run. This feature will enable us to determine
whether removal of this stimulus early enough could prevent
some of the effects leading up to neuronal death.
V. Future Plans
(a) We plan to introduce the possibility of partial neuronal recovery
and examine the effect of various interventions on that
recovery.
(b) Following advice from Prof Griffin,
we plan to make the simulation more visually appealing
by using realistic pictures of cells (in place of ovals), and
showing changes in their morphology as they are activated.
(c) We are making the growth of plaques more realistic, and
incorporating a direct link between amyloid and neuronal health.
(d) We are considering navigational aids and clearer
organization of the web pages.
(e) We plan to gradually update and improve the parameter
descriptions, the simulation description, and other
aspects of the web pages.
Summary of Updates: April 1, 2000
New Look for Demo and Simulation
The simulation has been given a new look. All front-end buttons are now
placed to the right of the active simulation window to avoid the
need for scrolling up and down while making selections.
This should reduce the distractions due to browser display difficulties.
TNF Added
The cytokine, TNF-alpha, has been added to the simulation. Users
can now see the effects of TNF, which is currently secreted by astrocytes
and taken up by neurons. The current implementation
allows the user to specify whether TNF is helpful or harmful to
neurons (the default is for TNF to protect neuron health).
The parameter TNF effect determines whether TNF is neurotoxic
(negative values) or neuroprotective (positive values). The effect
of TNF interacts with the effect of IL-6, as described in the
simulation details.
Manipulate Graph Data
A new button (labeled "Show data") has been added to the graph window for
overall neuron health. Selecting this option
enables the user to get the coordinates that make up points on the graph.
The data can be copied to a file which can then be manipulated
by any software package such as Excel to summarize the results
of multiple runs.
Web Pages Updated
All the web pages have been expanded and updated to reflect these
changes (e.g. the tutorial).
Summary of Updates: March 1, 2000
The Demo and Simulation Pages
The extensive text on the bottom of these pages has been replaced by hypertext links to separate pages with details about the simulation and demo and information about how to run the simulation and the demo.
Tutorial and Instructions
Detailed instructions for running the simulation and a full description of the available options is now in place as a tutorial.
Parameter Values
Pages of Parameter values have been separated and tagged so that the links on the demo page lead directly to the tables and references supporting the values. New tables are in place, and others are being prepared. (A table for chemotaxis coefficients of cells has been added.)
Status of Parameter Values
A summary page has been prepared listing all the current parameter values used in the simulation and indicating the status of each of the parameters, i.e. whether based on biological values, calculated, or still not calibrated.
New Information about Neurotoxicity
An extensively referenced web-page with details about neurotoxicity has been added. This page represents new information found in the literature that will gradually be incorporated into the simulation.
Summary of Simulation Updates: Feb. 1, 2000
Units on Display
Lengths are measured in microns, time is measured in minutes, and concentrations are measured in nM.
Grid
- The simulation consists of a 40 x 40 grid.
- Each grid cell is 10 x 10 microns.
- We assume a uniform depth of 10 microns.
- The total volume represented in the simulation is 400 microns x 400 microns x 10 microns.
Time steps
- The display is updated every 0.5 minutes. (Some browsers take advantage of fast updating, and can compute more than one display step while showing a single display. In this case, it will look like the updates occur over longer time intervals, and consequently the simulation really moves.)
- There are 40 time steps for every display step. This is to compute the high diffusion rates accurately.
Changes to Amyloid-Beta
Soluble form
- The source has been changed to a maximum of 200 nM. (i.e. updated source concentration.)
- Updated the diffusivity so that it is half the actual value. This is in order to keep the diffusion process numerically stable while still seeing something happen in the simulation. (i.e. updated diffusivity.)
Fiber Formation
- Updated initial fiber occupancy.
New information on fiber growth has lead us to change the fiber formulation. It has been shown that fibers need a critical concentration of soluble Amyloid-beta to grow. Thus, the parameter critical sol-AB for fibers has been introduced to the simulation. Studies also reveal that fibers do not shrink, so fiber degradation has been removed from the simulation. Finally, the current formulation only allows for fibers to grow where they exist; they cannot grow spatially.
Changes to Microglia
Movement
- The speed that a microglia moves has been estimated. (i.e. updated motility time delay.)
- The percentage of cells that move in response to chemotaxis was found. (i.e. updated chemotactic sensitivity.)
The disruption of movement due to fibers has been reformulated. Studies show that microglia get stuck with an increase of the fiber concentration at a probability that fits a Michaelis-Menten relation. Thus, the parameter half hazardous fiber level is introduced to indicate at what fiber concentration the microglia have a 50% chance of getting stuck. However, an increase in the concentration of soluble amyloid-beta tends to lower the probability of getting stuck. We model this with an exponential decay, where the new parameter half soluble slipping level indicates the concentration of soluble amyloid-beta present for a stuck cell to have a 50% chance of moving.
Amyloid-Beta Absorption
The absorption of amyloid-beta by microglia also had to be reformulated to fit the data. The rate of absorption follows a Michaelis-Menten relation based on the concentration of soluble amyloid-beta present. Thus, we introduce two new parameters for these kinetics. Half the maximum sol-AB absorption rate occurs when the amyloid-beta concentration is at the half max amyloid binding conc.
IL-1B secretion
- Updated IL-1B secretion rate.
Changes to Astrocytes
- Updated motility time delay.
- Updated IL-6 secretion rate.
Changes to IL-1B and IL-6
- Found secretion rates (noted above).
- Scaled diffusivities relative to the diffusivity of amyloid-beta. (i.e. updated IL-1B diffusivity and IL-6 diffusivity.)
Changes to Neurons
No significant data as of yet. Simply scaled current parameter values to be close to others in the simulation.