It is such a pleasure to be here with you all this morning. I'm Li-Huei Tsai. I'm the director of the Institute for Learning and Memory.
And my laboratory focuses on age-related neurodegenerative diseases, including Alzheimer's, dementia. So we know that we are a aging nation. Especially in recent years, the population of seniors over age 65 has increased sharply, whereas the proportion of children under the age of 18 has decreased gradually. So by the year of 2034 for the first time the seniors are going to overtake the children. And in fact, by 2060, a good 25% of our population is going to be seniors over 65, OK? So with these demographic changes it comes with a sharp increase in age-related chronic diseases, most notably dementia.
So currently, there are about 55 million people worldwide living with dementia, and this number is expected to triple by 2050. And dementia poses a particular problem, among all the other chronic illnesses, such as cardiovascular disease, diabetes, or even cancer, because all these other diseases the death rate has decreased over the years, but the deaths from dementia has increased year over year. And we don't have any treatment currently that can prevent or even slow the progression of dementia. So with that, I want to tell you a little bit about what we do. So we take three different experimental approaches to tackle dementia, especially Alzheimer's disease.
First, we look at large-scale brain circuits and systems activity to really learn how intervention at a very broad level, such as manipulating the natural brain waves, or brain rhythms, can stimulate healthy activity among cells and tissues within Alzheimer diseased brain. And to learn how a brain differs between health and disease we collaborate with computer scientists to do large-scale, but very fine-grained, machine learning analysis of how the various brain cell types carry out their genetic instructions. So by doing big data analysis we gain insight into how things may be going wrong at the molecular and cellular level. So back in the laboratory we can test new hypotheses and try out new ideas for intervention. Lastly, we combined stem cell biology, genome editing, and tissue engineering to probe what are the genetic and environmental risk factors that affect the etiology of Alzheimer disease. And also, we want to learn why certain individuals are more resilient to age-related neurodegeneration.
So here I'm showing you an example of the brain's vasculature that we created in a Petri dish from patient-derived stem cells. So I'm going to tell you a little bit more about this later. But first, I want to say a few more words about our non-invasive approach to stimulate the brain and to treat diseases as Rafael told you earlier.
So you all heard of brain waves. So we have nerve cells called neurons in the brain, and when they act in concert they produce these brain waves, or rhythms, at different frequencies. And the relatively fast frequency band known as gamma is associated with higher order brain function, such as spatial navigation, sensory processing, attention, and so on.
And we and others have shown that in Alzheimer disease model mice and in some patients gamma is compromised. So we initially try to boost gamma power in the brain of Alzheimer disease mice and found that it reduced beta amyloid, a key pathological hallmark feature of Alzheimer's disease. So as Rafael told you, we eventually figured out to use a very non-invasive approach, basically using light and sound to increase gamma power in the brain.
And I really want to thank Emery Brown in the audience, because he really gave us this idea to be as non-invasive as possible. So this is a cartoon depicting how we use live flashes and some clicks presented at gamma frequency to stimulate the mice. And we can see a profound increase in gamma power rhythms in many regions of the mice. So when we expose Alzheimer's model mice to gamma light and sound stimulation we found very widespread neuroprotective effects. Let me just give you some examples.
For instance, we found that there is profound preservation of the brain volume. As you can see the size of the brain in the Alzheimer's mice is reduced, and also, the ventricles are very much enlarged. But after gamma stimulation, you can see the size is preserved and the ventricle size is very much reduced.
And we also saw a profound reduction of this pathology you heard about associated with Alzheimer's, the plaques which is shown in magenta here after gamma is reduced, and also, tau tangles pathology also is very much reduced. But in addition to that, we saw functional improvement. We saw the circuit connectivity of the brain is increased and learning and memory function is improved. So often when I presented this work a lot of people will say, this is not possible. This is too good to be true. So we know that we have to learn how this increased gamma power works in the brain to be more compelling.
So we found increased gamma power in the brain can change gene expression as well as biochemical properties of not just neurons, but also other supporting glial cells, such as microglia and oligodendrocytes. And this leads to increased synaptic transmission and network connectivity. But we see something beyond that.
So here I'm showing you another cartoon to depict that the circulation of blood flow into the brain is increased. But also, the size of the blood vessel in the brain is also markedly increased. So this means, I need more nutrients, such as oxygen and other nutrients can get into the brain to support neurons.
And also, the brain waste, the toxic metabolites, can be cleared from the brain more efficiently. But most exciting, I want to share with you this very new result from my laboratory. This is the glymphatic system.
So the glymphatic system is actually discovered recently as a very powerful waste clearance system in the brain. And this involves glial cells, as well as the lymphatic vessels. So what happened is that we all have this cerebrospinal fluid in our brain, which normally resides in the ventricles and also in the subarachnoid space near the meninges.
So this is beneath the skull but above the brain tissue. So this CSF fluid they can enter the brain. So I'm going to show you this cartoon to depict this. So they travel along the blood vessels to enter the brain, and along the way they flush out this brain waste. It's almost like a car wash to flush out all of this bad metabolites from the brain, including beta amyloid.
And what we found is that increased gamma power significantly increased the glymphatic clearance of the brain waste. So with all of this beneficial effects now we have in mind, we eventually-- so here is just, so you see a piece of the data, just very pretty [INAUDIBLE] meninges. They're prepared by my graduate student. So that the yellow label here is the lymphatic vessels, OK? And they are pretty abundant in the meninges. You can see this zoomed in view of this glymphatic vessel.
And here I'm showing you a segment of the glymphatic vessel before the simulation and after the simulation in Alzheimer's diseased mice, and you can see the vessel size nearly doubled. And this also means that the glymphatic clearance doubled. So we eventually put together a team of researchers at MIT to bring this light and sound simulation approach to human subjects. And I just want to first show you our device.
So we have a 2 foot by 2 foot LED light box, which can emit gamma light flashes, as well as a high definition sound bar, which produces some clicks at gamma frequency. So a subject can sit about 5 to 6 feet away from the device and be stimulated. So we installed this device in our subject's home so they can turn on the device to stimulate themselves. And our phase II trial mild Alzheimer's disease patients is very small, but we still see significant results.
So we saw that this stimulation is very safe and produces very minimal side effects, but it's significantly improved network connectivity. And it preserved the volume of the hippocampus. The hippocampus in the brain is a part of the brain that is responsible for learning and memory, and also reduced the size of the ventricles. So basically this is the result that we saw in our Alzheimer diseased mouse models, and this is reproduced in human subjects. And we also saw improved face-name association memory. And two other studies now using similar approach to increase gamma power in the brain in humans produce corroborating results.
And an MIT spinoff company, Cognito Therapeutics, based on the promising phase II human trial results, they were granted an FDA breakthrough device designation. So they are starting their phase III trials in Alzheimer's disease. And here, at MIT, we are about to start a prevention trial.
So we identify human subjects that don't have memory impairment yet, but they have a family history of Alzheimer disease. So we wanted to give these people the stimulation to see whether this approach can delay the onset of memory impairment. And we are also doing a small scale Parkinson disease as well as Down syndrome trial. So with that, I would like to quickly tell you about our other approach, which is to build trackable models of the human brain to unravel the genetic and molecular complexity of neurodegenerative diseases.
So our approach is to combine human genetics, single cell transcriptomics, stem cell biology, and tissue engineering. So I collaborate very closely with many other colleagues in different schools. Manolis Kellis is a scientist at computer science and artificial intelligence who I work very closely together. And together we have collected millions of single-cell gene expression profiles from thousands of postmortem human brain samples with or without Alzheimer's disease.
So we are sitting on huge data sets that no one else has yet. And based on this data, we are able to stratify patient populations, as well as the single-cell transcriptomics, or gene expression profiles, based on genetic background of those people, or their medication history, or their lifestyle, or other factors, OK? So for instance, we can take stem cells from the subjects carrying a particular risk gene, and then we can differentiate the stem cells now into different brain cell types. And then study them in depth to find out what molecular pathways are perturbed in the cell types, because of the risk gene. And then we can come up with new preventative and therapeutic opportunities.
So I want to give you an example of this very well-established Alzheimer's gene. In fact, the most significant genetic risk factor for Alzheimer disease known as APOE, or apolipoprotein E. So in the human population there are two or three different flavors of the APOE gene. And this APOE4 version significantly increases the risk of developing Alzheimer's.
So if you have one copy, you probably have three times more likely incidence of developing Alzheimer's and two copies account for 10 to 15 higher risk. And the APOE3 version is a non-risk allele. So what we did was to take the skin biopsy from human subject carry APOE3 and a human subject carry APOE4. We then reprogram the skin cells into stem cells. And then we can use genome editing technology to actually change the nucleic acid, so to make APOE3 version into APOE4 version and vice versa, correct APOE4 version into APOE3 version, OK? And now we have this really beautiful so-called isogenetic pair of human stem cells. So we can compare their properties of this pair of cells.
They basically have identical genetic background with the only exception of the APOE4 locals. So then we can induce the stem cells into different cell types of the brain, such as neurons, astrocytes, microglia, and then we can see how APOE4 impacts these different cell types. So in this study, we show APOE4 really impacts all these different cell types to develop Alzheimer's-like pathology, it increases the amyloid production in neurons, increases lipids and cholesterol. So triglyceride and cholesterol accumulation in astrocytes and APOE4 makes these microglial cells more prone to inflammation and less supportive of neurons.
So not just these individual cell types, we can also make blood vessels, as I showed you earlier from stem cells. And then we can make the blood vessels from APOE3 or APOE4 stem cells. And we also saw differences, OK? So we found that this APOE4 blood vessels are much more prone to accumulate amyloid along these vessels. So this is a very well-established pathology known as cerebral amyloid angiopathy, this amyloid accumulating along the blood vessels. So then with these different systems we can then go in and find out what pathways are perturbed, and then identify small molecule drugs that can ameliorate these pathological features.
So I'm very happy to tell you that just by looking at some of the FDA approved drugs, we found that this drugs can have repurpose effects. For instance, FK506 and cyclosporine A, they are used to treat organ transplant patients to suppress their immune system, but we found that these drugs actually are very, very effective in reducing amyloid accumulation along the blood vessels. And another example I think is particularly cool is this essential nutrient choline. So choline is present in our food, such as it's in salmon, and eggs, and nuts. It turns out we found that high levels of calling in astrocytes can reduce triglyceride and cholesterol accumulation in these cells.
And recently we found that when we treat APOE4 mice with choline, high levels of choline, it can reduce amyloid and reduce inflammation. So this really gave us confidence to continue our work to be more ambitious to try to reconstitute a whole brain in a dish. So we are now differentiate the stem cells into all seven different brain cell types, and then we cultured them together in a three-dimensional manner.
We call this tissue [? my ?] [? brain, ?] and it can sub-sample into intricate cell networks, and then we can use this [? my ?] [? brain ?] system to not just studying risk gene and lifestyle environmental effects, but we can do drug screens and we can even do clinical trial in the dish. So you may ask, how does this work? So basically, here, everybody, I can take your skin cell, and then I can reprogram your skin cells into stem cells. And then now with the stem cells I can create a mini brain of you, your own [? my ?] [? brain ?] in a tissue culture dish, OK? And then with this personalized [? my ?] [? brain ?] I can now do, say, test candidate drugs. So most likely not everyone going to respond to a particular drug equally. So some of you will respond to a drug very, very nicely, but some of you may not respond to a drug at all.
But before you take it, we will already know that using your [? my ?] [? brain ?] system. So with this matrix, you can see, we can look at the extent of response from these drugs, and then we can even see whether a combination of different drugs may be better for some of you than some others. So we can optimize the drug treatment protocol for each individual.
So this is called personalized medicine, and we can also track your response over time. So our eventual goal is to define not just drug efficacy, but also pharmacokinetics, pharmacodynamics properties of the drug and achieve optimal therapeutic window. So I would like to end my talk by telling you that we are now putting together a so-called [? de-risking ?] center. So the center will be dedicated to better advanced therapies to clinical and commercial utility.
So we want to do drug development, but we also want to do preclinical research, especially using [? my ?] [? brain. ?] I want to tell you that is [? my ?] [? brain ?] concept is so exciting that the NIH actually organized a workshop next month to discuss the utility of [? my ?] [? brain, ?] not just for drug discovery, but for a really bona fide preclinical testing utility to use side-by-side with the traditional mouse models or even one day to replace the mouse model. And we also want to do early phase clinical testing in humans. And I want to thank MIT administration for very kindly offer us a 12,000 square foot new space for us to get started with [? de-risking ?] center. And we really want to make the world better for everybody.
Thank you very much for your time. [APPLAUSE]
2022-06-23