Prolonging the Rapid Antidepressant Effect of Ketamine

Major Depressive Disorder (MDD) is a debilitating psychological disorder characterized by symptoms such as low mood and fatigue. Although there are a variety of treatments for MDD, many of them are either ineffective or have a delayed effect. Ketamine, a drug, is commonly used as an anesthetic, but it can also be used as an antidepressant. What makes ketamine different than other treatments for MDD is that it is usually more effect, and the antidepressant effect works in a few hours instead of weeks or months. Many downsides, such as a transient antidepressant effect, make it inaccessible to many patients. The purpose of this project is to investigate a way to extend the antidepressant effect of ketamine by looking at the relationship between it and cortisol concentration, the chemical associated with stress. The hypothesis is that a higher stress level will extend the antidepressant effect since people can gain resilience after overcoming difficult experiences. A computer simulation was made to estimate how neuron firing rates, which is how depression is measured in this project, vary before and after applying ketamine under five different cortisol levels. The simulation results will be compared to the experimental results to see how well the simulation predicts neuron firing rates. Learning more about ketamine as an antidepressant is vital to making it accessible to more patients and to develop better treatments for MDD.

Transcript:

Hi, I’m Andrew and my project is called prolonging the rapid antidepressant effect of ketamine

So lets jump right into it

Depression is a psychological disorder characterized by symptoms such as difficulty concentrating and low self-esteem

Although many medications for depression exist, many of them are either ineffective or have a delayed effect

Ketamine, a drug, is commonly used as an anesthetic, but it can also be used as an antidepressant

What makes it better than other medications for depression is that it is more effective and works in a matter of hours instead of weeks or months

The downsides are that the antidepressant effect typically lasts only a few days, it has harsh side effects, and it has the potential to be abused which makes it inaccessible to many patients

Much research has been done on understanding how ketamine produces its rapid antidepressant effect, but not much has been done on extending it

The purpose of this project is to establish the relationship between the antidepressant effect of ketamine and cortisol concentration, the chemical responsible for stress

The hypothesis is that a higher stress level could lead to a longer antidepressant effect since people can gain resilience after overcoming difficult experiences

In this project, depression will be represented by a decrease in the number of hippocampus neurons spiking, or turning on, since patients with depression typically have low neuronal activity in the hippocampus, the part of the brain responsible for learning and memory

After applying ketamine, we would expect to see an increase in the number of neurons spiking since antidepressants can increase neuronal activity in the hippocampus for a period of time until the firing rate returns to baseline

This project consists of three parts, but because of many problems some parts were not able to be completed in time

(thinking) Just another thing I can’t do right

But I will talk about the model and simulation I developed to predict how different cortisol levels affect the duration of the increased neural spiking caused by ketamine

So for this project there is a grid with two types of neurons: a “normal” neuron that will spike at a normal rate with a resting potential of – 70 millivolts and a “depressed” neuron which will spike less often with a resting potential of – 90 millivolts. If the resting membrane potential reaches –55 millivolts, the neuron will spike. Otherwise, it will not.

Alongside this, ketamine will also be on a grid where it is administered in the center

As time passes, it will diffuse outward from the center until it’s completely gone

Ketamine will help the depressed neurons fire more while the normal neurons will be unaffected

Hi again. I thought we’d take a field trip to show how we can measure neuron firing rates in the lab

We can use one of these micro electrode arrays

We can plate hippocampus neurons on these, stimulate them with this machine, and the micro electrodes will be able to detect neurons spiking

Now, in order to simulate depression, a chemical called carbenoxolone will be applied to the experimental groups since it is known to cause neurons to fire less like that in depression

After that, ketamine, alongside one of five different concentrations of cortisol, will be applied to each dish in the experimental group

We’ll then record the neuronal activity for an additional two days

That’s the first part of this project

The second part has to do with this

This colorful image was produced using immunofluorescence, a lab technique used to image neurons

It will be performed on a group of neurons to see how ketamine affects dendrite length and number which are these branches

These are the two parts of the project that were not able to be completed in time due to complications, but they will be completed soon

(thinking) This project was a failure and this presentation isn’t going well either

Now that we have covered all of that, let’s get back to the simulation

In the actual experiment, the only variable that will change is cortisol concentration. In the simulation, alongside cortisol concentration, grid size, the drug dose and diffusion rate, and the level of depression, represented by what percentage of neurons on the neuron grid are normal or depressed can be changed

Also, since neurons spiking is somewhat random, there is an element of randomness

Let’s do the simulation with a 15 by 15 grid where 80% of neurons are depressed and 20% are normal with a 75 milligram dose diffusing at a rate of 6 micrometer squared per second starting with the lowest cortisol concentration.

We’ll see how the firing rate looks like before the drug is applied

It looks likes there are about 20 spikes at each time point

Now, let’s apply the drug

We see a sharp increase in the neuron firing rate which is expected since ketamine works fast with a max of around 60 spikes.

After that, we see a general decline back to baseline

One of the advantages of this simulation is that we can change the parameters to see how the firing rate changes

If we increase the dose to 150 milligrams, the effect lasts longer since there’s more ketamine available

If we increase the diffusion rate to 12 micrometer squared per second, the effect is shorter since ketamine is removed faster

If we increase cortisol concentration to the highest level, the effect is also longer which is the hypothesis for this project

We can also change the number of doses

If we give a patient a 75 milligram dose every 3 days, the simulation graph looks like this

It looks like the effect lasts longer with each additional dose.

This result agrees with the literature that with each additional dose of ketamine, the antidepressant effect is usually more effective

Now going back to one dose, we can use the simulation to develop a model equation depending on the cortisol concentration.

After collecting the experimental data, we want to compare how well the model fits it. We only get four data points from each microelectrode array and we need more to compare them so we can use interpolation to generate more data points like this

Now we can overlay the model and data graphs for that specific cortisol concentration and calculate S the standard error of the regression value which will tell us how well the model fits the data

For this sample data and model, we get an S value of 8 which means that on average the model and data differ by about 8 spikes

And now, the last part of this simulation has to do with efficacy and toxicity of ketamine and we’ll be looking at two variables: dose and diffusion rate

Starting with dose, too small of a dose will have no effect and a large dose will be more effective

However, like any drug, there is also toxicity we must consider which follows the same general pattern

Similar curves can be produced for the diffusion rate. A lower diffusion rate will be more effective and toxic, and a high diffusion rate will have little effect and toxicity

We want to find the dose and diffusion rate that maximizes efficacy and minimizes toxicity

Thus, we want to find the maximum of this net benefit equation shown on this graph. However, since many optimization algorithms are designed to find a minimum, we’ll have to flip this graph and find the minimum

From the graph, we can see the minimum occurs at a dose of 89.4 milligrams and a diffusion rate of 7.7 micrometer squared per second and if the graph is flipped back, we would see that this dose and diffusion rate maximizes net benefit

And that’s all for my project

But before I go, I want to address one final thing

Throughout this presentation, you have heard some of my depressive thoughts that I’ve had in the past

I’d be lying if I said that I still didn’t have them especially during this project

Is that a bad thing?

I don’t think so

It just reminds me of what I’m working towards and just in case I forget, I always carry this id badge with me

It represents three things: the past filled with many enduring hardships, the present filled with its ups and downs, and the future filled with unlimited opportunity

So, what’s next? The same thing that’s always next. I’m going to keep learning more and working hard so that I can work towards developing better treatments for depression

(thinking) Yeah, this project was a success and the presentation went well. I’m proud of it