NetLogo User Community Models

(back to the NetLogo User Community Models)

Download If clicking does not initiate a download, try right clicking or control clicking and choosing "Save" or "Download". Try It in NetLogo Web

## WHAT IS IT?

This is an agent-based model simulating a 2-dimensional slice of the hippocampus, intending to analyze the relationship between microglia (resident immune cells) and the factors that contribute to the progression of Alzheimer’s Disease, as guided by available research.

This version of the model includes three agent breeds, representing mobile microglial cells and stationary neuronal cells, either healthy or damaged. The microglia are initialized with a generalized goal of phagocytosing (clearing out) all damaged neurons in any given iteration of the simulation. They accomplish this task through pseudo-random movement in the patch space until they encounter a chemokine (a kind of signaling protein) being secreted from a damaged neuron, which signals the microglia to follow the source of the chemokine until reaching the damaged neuron, and phagocytosing it with some variable probability.

This model utilizes solely agent-dependent patch characteristics. In this version of the model, only a damaged neuron has the ability to affect the patches that surround it in the model world. Patches do not spawn with any patch characteristics innately, unless there are damaged neurons present which, by design, will follow a procedure that provides nearby patches with user-created patch characteristics. The specifics of this process will be elaborated on in greater detail later on.

This model sheds light on the impacts of microglial activity as they relate to damaged neurons. Rather than highlighting and examining damaged neurons' role in the proliferation and progression of Alzheimer’s Disease, which is well researched, this model seeks to explore which spatial and temporal microglial factors can influence creation of response to damaged neurons.

## HOW IT WORKS

The model is intended to have many variables that can be chosen by the user. Initialization involves setting the number of microglia agents, healthy (pink) neuron agents and damaged (red) neuron agents. The user also has the option to adjust the temperature, inflammation radius, the probability of successful phagocytosis, the efficiency of the microglia sensing, the distance of the neuron links, and the chance of a neuron becoming damaged. The turtles representing neurons are stationary. They can be healthy or damaged. In the damaged state they secrete inflammation. In both states, neuron turtles can connect to one another via links. These links communicate the current state of the neurons with those they connected to, infecting with a chosen probability. The microglia turtles move randomly about the world. Their task is to seek out damaged neurons and clear them from the world. The simulation ends once the microglia have cleared all the damaged neurons.

### Inflammation

The turtles representing damaged neurons secrete inflammation, a patch characteristic dependent on a user-created procedure that mimics the diffusal of chemicals secreted by cells in the body (specifically IFN-γ, a cytokine that signals microglia to activate). Inflammation signals nearby microglia turtles to attempt phagocytosis. The radius of this inflammation can be adjusted by the user with the INFLAM-RADIUS slider. The inflammation spreads from the damaged neurons until it reaches the chosen radius.

### Phagocytosis

The microglia turtles’ main function is to seek the world for damaged neurons and phagocytose them. Phagocytosis is the process of ridding the hippocampus of the damaged neurons. The probability of successful phagocytosis can be modified by the user with the EAT-PROBABILITY slider.

### Sensing

The microglia turtles have a sensing radius that allows them to sense the damaged neurons that they are hunting for. The effectiveness of these turtles accurately sensing a damaged neuron can be altered by the user using the SENSING-EFFICIENCY slider.

### Neurons

The number of healthy and damaged neuron turtles in each run is determined by the user using the INIT-HNEURON and INIT-DNEURON sliders, respectively. If the neuron is damaged, they diffuse inflammation (described above). The neurons are stationary, and their main purpose is to be consumed by the microglia. The neurons can connect to one another via links.

### Links

The turtles representing neurons (both healthy and damaged) can connect to one another via links. These links communicate the status of the neuron turtles to one another. If there is a damaged neuron connected to a healthy neuron, the damaged neuron can “infect” the healthy neuron by a set probability. The chance that this happens is once per healthy neuron per tick.

## HOW TO USE IT

In the interface tab, the user is presented with a variety of sliders to adjust the simulation. Here is a brief description of each of these sliders.

* INIT-MICROGLIA: sets the number of microglia turtles that the simulation starts with.

* INIT-HNEURON and INIT-DNEURON: sets the number of healthy and damaged neurons, respectively.

* TEMPERATURE: adjusts the temperature of the microglia turtles, which affects their speed of movement.

* INFLAM-RADIUS: changes the radius of the inflammation secreted by the damaged neurons.

* EAT-PROBABILITY: adjusts the probability that a microglia turtle will successfully perform phagocytosis.

* SENSING-EFFICIENCY: changes the probability of a microglia turtle to successfully sense the inflammation secreted by the damaged neuron turtles.

* NEURON-DISTANCE: affects the radius that a neuron turtle can connect to another neuron turtle through synapses (links).

* DAMAGE-CHANCE: adjusts probability that a healthy neuron with a link to a damaged neuron will become damaged.

The SETUP button resets the graphics and plots and places the turtles in random locations. The INFECT button causes a random healthy neuron turtle to become a damaged neuron turtle. This button can be used throughout the simulation. The GO button starts the simulation and the plotting function.

As the simulation runs the microglia turtles survey the world and clear it of any damaged neurons present. Once the microglia clear all the damaged neurons, the simulation ends.

The plot shows the population of each turtle with respect to time. The green line represents microglia turtles, and the red and pink lines represent the damaged and healthy neurons, respectively.

## THINGS TO NOTICE

When the simulation starts, notice:

* Healthy neurons that are linked to damaged ones might become damaged.

* Inflammation spreads from the damaged neurons.

* Microglia move toward damaged neurons if they encounter a patch with inflammation.

* Microglia change from green to yellow when surveying or attempting to eat.

## THINGS TO TRY

The user can run a number of experiments with many different variables to adjust.

The model includes an INFECT button. Try running a simulation and infect healthy neurons as the simulation runs. Observe how long it takes for the microglia turtles to clear the world of the damaged neuron turtles. How could this relate to real world diseases? If the infection is rampant, the microglia struggle to clear the damage. Try the same parameter values, but reduce the number of microglia. How does a lower number of microglia affect the recovery of a disease?

The model includes the slider INFLAM-RADIUS. Try adjusting the radius of the inflammation. If the radius is large, does this affect the efficiency of the microglia in clearing the damaged neurons? Or, is the efficiency of microglia clearing the damaged neurons entirely dependent on the SENSING-EFFICIENCY slider? Is there a threshold of when SENSING-EFFICIENCY is so high that microglia become overwhelmed by the inflammatory signals and stop properly functioning, or so low that they fail to have an appropriate response altogether?

Try experimenting with different values on each slider. For example, to have a good chance at seeing every facet of the model working together, try the starting values:

INIT-MICROGLIA: 3

INIT-HNEURON: 10

INIT-DNEURON: 10

TEMPERATURE: 37 C

INFLAM-RADIUS: 5 patches

EAT-PROBABILITY: 70%

SENSING-EFFICIENCY: 50%

NEURON-DISTANCE: 5 patches

DAMAGE-CHANCE: 0.50%

## EXTENDING THE MODEL

Like all computer simulations of biological systems, this model has simplified many aspects of the Central Nervous System. The model, therefore, provides numerous opportunities for expansion:

* Of chief priority in future models is to establish an effective phenotype breed switch for the microglia turtles. Microglia continually sense and respond to their environment, and in responding to certain signals, enter activated states described as the expression of a specific phenotype. These activated phenotype switching states would be an ideal biological phenomena to model and explore in later models.

* The model uses inflammation to depict neurons secreting cytokines, particularly Interferon-gamma (IFN-γ). A future model could implement other factors related to the progression of Alzheimer’s Disease, such as Aβ plaques.

* This model makes the simplifying assumption that IFN-γ is the only cytokine present in the hippocampal slice being simulated. In reality, there are many other cytokines in a human hippocampus, and future models could explore implementation of many different cytokines all interacting at once

* Microglia also emit their own cytokines that allow them to communicate with each other; future models could add a feature where microglia secret their own cytokines similar to the inflammation secreted by the damaged neurons.

## CREDITS AND REFERENCES

This project was supported by NSF grant 2245839 from the Mathematical Biology program. We are deeply grateful for this support.

We would also like to thank the users on Stack Overflow who replied to [this](https://stackoverflow.com/questions/77467956/how-to-create-a-diffuse-function-without-the-diffuse-primitive) thread.

## HOW TO CITE

If you mention this model or the NetLogo software in a publication, we ask that you include the citations below.

For the model itself:

Case, A., Mezzulo, E., Penland, A., Ty, C., & Larripa, K. (2024). NetLogo Microglia Model. Cal Poly Humboldt, Arcata, CA.

Please cite the NetLogo software as:

Wilensky, U. (1999). NetLogo. http://ccl.northwestern.edu/netlogo/. Center for Connected Learning and Computer-Based Modeling, Northwestern University, Evanston, IL.