Dr. Sergiu Pașcă

Thank you.

It's great to be here.

Well, autism is a complex condition.

It's a spectrum, as you said.

In a way, you could say autism and neurodevelopmental disorders.

It's behaviorally defined.

There's no biomarker.

So in a way, it's a condition that is defined exclusively by observing behavior, which is actually the case for most psychiatric disorders.

But it's essentially diagnosed by the presence and absence of certain behaviors in a certain period of time or up to a certain age.

And, of course, what triggered, I think, a lot of discussions in recent years is because the number or the prevalence of autism has increased.

So now it's close to almost 3% of the general population, which, of course, it's a big number.

3%?

Almost 3%, yes.

Wow.

So that has increased even since I was in medical school.

When I was in medical school, actually, it was considered a rare disease.

The reason why I actually studied autism, because it was a very rare disease and we had very few resources, so we thought studying a rare disease would be easier.

But now we also know so much more about this condition.

So we do know, for instance, that there is a strong genetic component to it.

which for a while, obviously, we didn't.

In fact, in early days, the psychoanalytic perspective dominated, especially in the 50s and 60s.

So it was thought that it was resulting from having very cold parents, in particular, a cold mother.

Emotionally cold?

Yeah, emotionally cold.

It was the so-called refrigerator mother hypothesis of autism.

And then in the 70s, some of the first biological studies were done, primarily in twins, that show something quite remarkable, that if you have twins that are identical, genetically identical, and one has autism, then the probability that the other one has autism is very, very high.

Even with different mothers?

Sure, yes.

But generally, we think that there is a strong heritable component to autism.

So that was like in the late 70s.

And really just in the last 10, 15 years, we've learned actually that there are genes associated with autism and certainly with very specific forms of autism.

So that's what we would call generally profound autism today, the conditions that are severe.

that are causing an impairment.

They are very often associated with other conditions such as intellectual disability, so low IQ, epilepsy.

So because it is a spectrum, of course, it creates a lot of confusion.

And certainly there's no doubt that there are individuals that have autistic traits that are fully functional in the general population.

But the reality is also that there are

kids that have autism who are very impaired and will require actually lifelong care of sorts.

Another way of thinking about autism is that autism is not one disease.

And I think no psychiatrist or even biologist who's studying autism will ever consider that this is one single disease.

The way I look at it sometimes is, think about the fever

of the 19th century in medicine, right?

So you see this very often in movies, right?

They will say, oh, he has a fever, high fever.

He's going to die from high fever.

Well, that fever could have been a viral infection, a bacterial infection, could have been cancer, metastatic cancer, right?

Could have been an autoimmune disease.

The treatments are very different.

But in that time, that's all we knew.

We were observing that behavior, in which case raising of the temperature, but we didn't know the biology.

Today, we will use very different treatments for those conditions.

And some of them, of course, we don't even treat, right?

We just observe.

So I think...

In autism research, as it is the case for many psychiatric conditions, they are defined behaviorally, but there's a disconnect with the biology.

Very often, we don't have biological markers by definition.

And so that disconnect, I think, creates a lot of confusion.

The ratio varies also based on severity, but generally it's been 1 to 4.

So more males than females.

Well, there are a lot of discussions, of course, like what causes this difference.

And, you know, some discussions are just in terms of diagnosis that perhaps some of the girls are not getting diagnosed properly.

We do know that some of them are very good at what we call like masking the symptoms or sort of like, you know, learning the skills, social skills and sort of like covering for that diagnosis.

But what we do know for sure is...

is that there are differences in how the male and the female brain, especially around birth, can actually take up injury.

So think, for instance, about premature birth.

You know, one of the best predictors for a premature baby in terms of outcomes, it's actually to be a female.

Just in general, females' preemies will do much better for whatever reasons.

The way the nervous system is built, the resilience.

We know that the maturation stage is also different for the male and the female.

Think about acquisition of certain milestones that happen much faster in girls.

They generally tend to speak a few months earlier, to walk a few months earlier.

So just the nervous system.

It's maturing at a different pace and can take injury differently.

So it could be that that is certainly the cause.

But at the same time, and as we were talking, since autism is not one single disease, it is very hard to point out to one specific factor that is behind it.

It's not part of the diagnostic criteria.

Interesting.

But it is one of the features that has been observed.

Of course, it also has to do with just in general, like joint attention is one of the earlier.

So, you know,

If you just tell a child, like, oh, look here.

So if they have that attention, if they engage in that attention, it's one of the features that is associated with autism.

It's not certainly diagnostic.

It's not...

pathognomonic, so to speak, so it's not specific to the disease in any way.

But there's certainly many deficits, and some of them can actually be compensated later.

Interesting.

Yeah.

So those are mostly anecdotic reports of patients who would have a very high fever.

And then, for instance, they were nonverbal.

So many patients with autism or individuals with autism will have, you know, will be nonverbal.

They have very few words or if they, you know, they're not able to communicate.

And so there are a few reports of parents saying that when they spike the very high fever, they'll start talking in sentences like very briefly or like engage.

I mean, that is known.

You know, kids in general, when they have a high fever, they tend to be more talkative.

It activates somehow the nervous system.

There have been a lot of hypotheses about this.

Some of them having to do with how the neurodegenerative system is activating during fever.

Others saying that there are some of the cytokines, the immune molecules that are present during fever that are somehow getting into the brain, activating the nervous system.

And others as simple as, oh, ion channels, right?

Ion channels will open up

more when the temperature rises.

So something about the circuits functioning differently during that.

But it's mostly anecdotic at this point.

And it's certainly, again, probably not present in all individuals with autism, also because autism is, again, not one single disease.

So we would not expect it to be present in all.

I think, again, almost everything has been associated or thought to be causal.

But generally, demonstrating this is very, very difficult.

So we cannot deny that perhaps improving the microbiome will improve the quality of life of some of these individuals.

But whether it's really causal, there's no clear evidence for it.

Think about it.

Just to give you another example, think about sleep.

Many patients will report, especially the ones that are profoundly impaired, will have severe sleep disturbances.

I mean, 70%, 80% of them, they can have nights where they sleep very little, then do that for like a week.

So just imagine even just improving the quality of sleep for those patients can do miracle.

I mean, all of us, right?

If we don't sleep for three, four days, our social skills, you know, we become socially impaired.

So I think, of course, correcting a lot of these issues.

So for instance, many patients are picky eaters.

You know, they don't like certain textures.

So they will never eat, for instance, veggies.

So that creates in the early days, for instance, we thought that there are dietary disturbances that really at the core.

Of course, it remains to be seen whether just simply correcting those is going to be just improving or certainly improving.

reversing some of these forms.

But again, most of the evidence points out towards a very strong genetic component behind it.

And in fact, we now have hundreds of genes that we know when they are mutated, they are strongly associated with specific forms of autism.

Yeah, absolutely.

And those are really elegant experiments that he's done.

Many of the genes, you know, they fit in different categories.

Like you would have genes that would produce proteins that sit at synapses, which was sort of like to be expected.

Some of them are, you know, ion channels.

They're proteins that would let ions inside or outside of a neuron.

There are many of these conditions, so-called channelopathies.

Then there are the ones that are like synaptic related, so synaptopathies.

There are a lot of chromatin genes, so like proteins that pack the DNA in cells.

Those are chromatinopathies.

So there are really, again, many, many categories of genes.

And then what is also interesting is that many of these genes are also expressed in the periphery.

So I think the experiments that you are mentioning are really elegant.

because it showed that indeed that can perturb the development of the nervous system, even if they're affecting just the periphery.

Of course, now in patients, there are present also in the central nervous system, so it's always difficult to distinguish.

But just missing some of these critical periods or perturbing some of these critical periods of development can have certainly devastating effects later on.

No, there's no single treatment for autism, again, in the context of this not being one single disease.

What we can say today is that if, you know, family walks into the clinic with the diagnosis of autism or perhaps like they receive it into the clinic, there's still like a 20% probability that they will leave the clinic with the genetic diagnosis, meaning that it will be pointed out to them that this gene is mutated in your child.

And it may be sometimes a mutation that was present in one of the parents and got transmitted, or maybe it was present in both and somehow the child got two copies that were modified now, or many of the genes were actually mutated de novo, meaning that the mutation was not present in either parents, but something went wrong during development, perhaps early in the sperm cell, in the egg cell, or perhaps in early stages of development, and a new mutation was acquired.

But that is also, we acquire a lot of mutations, all of us, we have a lot of new mutations, right?

About like 80 new mutations, 30 of them are protein truncating.

So certainly the challenge very often is to, even when you see a gene that is mutated, to know whether that gene is truly causing the disease.

So very often the way we know is that we find many patients that have a similar presentation clinically.

Let's say maybe they'll have syndactyly.

So they're webbing of the finger and they have autism and let's say epilepsy.

And they all have a mutation in one single channel.

let's say in a calcium channel.

So that would be Timothy syndrome, a genetic form of autism, where the mutation is very clear.

Actually, there's one single letter in the genome that is changed and causes a relatively similar presentation in all of these patients.

So about 20% of the patients will get a genetic diagnosis.

Now, sadly, that doesn't do that much today because we don't really have specific therapies for those forms.

I think the hope is that perhaps we will have individual treatments, whether they're going to be genetic or otherwise.

So being part of that community is generally useful.

And then the rest of the patients will essentially fit into this larger category of idiopathic, meaning that we don't really know the precise cause.

Yeah.

Well, certainly the increase is still puzzling, right?

So I think on one hand, there's no doubt that the changes in diagnostic criteria, which have happened over time.

I mean, we had to just refine what autism really is.

That changed, you know, to some extent, the prevalence.

We've also seen, you know, a diagnostic migration, so to speak.

So some children, for instance, you know, 30 years ago would have been diagnosed with intellectual disability.

And today they fit the criteria for autism.

About a third of individuals with autism also have intellectual disability.

So there is also a great overlap between the conditions.

So there's been a move sometimes between the diagnosis over time, of course,

There are all kinds of discussions about availability of services and to what extent that is also contributing.

But we don't truly understand all the reasons behind this increase.

There's no doubt.

We can't explain.

We know that it's highly heritable based on genetic studies.

So we know the heritability is very high, one of the highest for psychiatric disorders that we know of.

But of course, we don't have the genes for every single form.

So it is likely that some of them are very rare.

So essentially, just think of it as like, they're individually rare form, but collectively common.

So it will take a while until we sort of like map all of them.

And then of course, there are environmental factors that we do know historically can contribute to this.

So there are

various exposures to environmental factors, like in early days, thalidomide, for instance, was one of them that we know increases the risk for autism.

Yeah.

So there's certainly, you know, it's quite complex because, first of all, the definition of the condition is quite difficult, right?

And I think that is in general like the challenge with psychiatric disorders, right?

And perhaps one of the reasons we've made such slow progress in understanding these conditions because, of course, the power of modern medicine is in molecular biology, right?

We kind of deploy this remarkable force of an understanding.

And in order to do that, you need two things.

You need, first of all, to have a very clear definition of what that disease is, generally, biologically.

Think about myocardial infarction, very clearly defined in terms of what it actually means.

You immediately have biomarkers.

The patient walks in.

You take blood.

You can immediately tell, yes, in 20 minutes, you can tell that they have a myocardial infarction based on a biomarker.

And then the other one, which is certainly very important, which to a large extent is the source of all the work that we've done, is the unbearable inaccessibility of the human brain, so to speak.

And to a large extent, the human brain is inaccessible for most of its development.

And so if you look actually across branches of medicine, you can see that there is a very strong correlation between how accessible an organ is

And how many cures or therapies we actually have.

Think even just in cancer.

Think about in cancer, which used to be, of course, an incurable disease a century ago.

Think about leukemias in children.

They were like 90% lethal in the 50s and the 60s.

Today, they are maybe 10% lethal.

And that is because a lot from these patients, right, it's very easy to collect.

We've been bringing it to the lab, studying it, what goes wrong, and then deploying molecular biology to develop therapeutics.

With the brain, sadly, you know, there's no way of doing it.

And so largely to, you know, what we've been trying to do is like find a way of shortcutting that process.

But I do believe that the major challenges that we're facing in understanding brain disorders, whether they're neurological or psychiatric, are on one hand problematic.

you know, the inaccessibility of the organ of interest, the brain.

And on the other hand, our challenges are very often defining some of these conditions with biological markers because they're much more complex.

And there's not.

Epidemiologically, there's no evidence.

Right.

And then, of course, the other issue is also that these conditions are disorders of the human brain.

So if you think about it, even talking about schizophrenia, right, hallucinations or phenomena that are very difficult to study.

And, of course, we don't know this.

We know that schizophrenia is present in almost every population that we know of, even isolated population at 1%.

And again, it's a little bit easier because it's done in adults.

I think in children, it's much more difficult.

And in fact, many of the genes that were early on identified for autism were identified in these populations, in the Amish populations, for instance.

There is a very classic example of a gene that is associated with severe epilepsy and autism that was identified there for the first time.

It's present in other places as well.

So, yeah, I think, of course, the complexity of the problem is that you also want to make sure that you don't just associate something.

You also want to reverse it in a way, right?

So you would want to do the other experiment where you change it and it goes away.

But you can never do that in the human brain.

We can't just turn things on and off to see whether they're truly causal.

And then, of course, human brain development also takes an incredibly long period of time.

Right.

If anything, it seems that the human nervous system has done everything possible to slow down that process.

I mean, we myelinate all the way to the third decade.

Neurons are born and migrating through the nervous system into early postnatal years.

There's evidence that myelination, especially in the frontal areas of the brain, are continuing up to the third decade.

Yeah, no, the prevalence for autism has been actually reported to be higher in other countries even before this.

Some of the early reports many years ago showed that in Korea, for instance, the prevalence was very high.

Now that the studies are done, also like in Scandinavian countries, it shows that it's probably around the same, you know, kind of like rate, 1 in 30 to 1 in 40, so somewhere between.

And perhaps another argument, which is very important to bring, is that we find the same mutations.

I mean, the same mutations, if we're talking, let's say, a mutation in a specific calcium channel, you'll find it in a patient in Denmark, as well as one in Africa or in, let's say, Australia.

So I think some of these genetic mutations are sort of like the same.

So gene therapy is a rather actually broad term, and it covers many ways in which you can correct generally a gene or a genetic defect that we think it's causal.

So on one extreme, for instance, you can envision a gene is broken, has a mutation.

So what you want to do is you want to put it back.

So those were some of the early efforts where you would put it in a virus and deliver it to the patient.

An adult.

In an adult or in a child, depending on like the condition.

With the idea is that the gene is not there or like there's not enough of it.

So I'm just going to deliver more.

That's one extreme.

Many of the studies were done for blood disorders, of course, because it was easier, so you would inject them.

Of course, the other possibility is sometimes you don't want to put the gene, you want to put the protein already made.

And that is the case for many conditions where an enzyme, so a protein that, you know, does some interesting chemical reactions that are essential to a cell is missing.

So sometimes you just make that enzyme and then you deliver that.

It's not always working, but in some cases actually works really well.

Now, the other thing that you can do is you can try to correct that defect directly.

That means you need to operate at the DNA level.

So somehow you need to get into every single cell that is affected and correct that.

And that's where CRISPR comes into play, where presumably you could at one point deliver the guides, so the tiny pieces of nucleic acid that tell you where to go on the DNA.

And then an enzyme, they will do the cutting and then the putting back or various other versions of this that you would correct.

Of course, there are challenges with that.

Yeah, well, presumably you could inject into the brain as well, right?

There are ways in which you can inject through either surgery or through an injection in the spinal canal, like intrathecally.

So that's certainly one way in which you can do it.

It is very challenging though, because of course the brain has a lot of cell types.

And, you know, you very often, the way you deliver this, like through a virus or through other modalities, you know, there's only so much

of that virus that you can actually put inside the nervous system.

And the efficiency is not yet like very high.

So another way is to go like one level down.

So that gene will produce an RNA that will produce a protein.

So perhaps we don't have to correct the DNA everywhere, but perhaps we can correct something that happens downstream.

And that's so like being the strategy that we've been using primarily.

Just mostly because at this point, and probably in the future it will be possible, who knows, like in 10 years or maybe even earlier, we'll be able to deliver very effectively some of the genetic therapies using CRISPR.

And they're very often modified to make sure that they don't cause disease.

Of course, another limitation of that is that if the gene is really large,

It simply won't fit in a virus.

So, for instance, that would be the case if you think about a calcium channel.

Calcium channel is a gigantic gene.

It would be very difficult to fit inside a virus.

And then, of course, the other thing is like with these viruses very often, especially with the adenoviruses or AAVs, is that you have one shot.

meaning that you have to inject once and hopefully it would work because next time, you know, you may have an immune reaction, right?

You'll have, you'll produce antibodies and so you won't be able to deliver again.

So there, again, there are all kinds of challenges that, you know, people are working really hard to solve.

And I have no doubt that in the next decade we'll see, you know, therapies or, you know, perhaps even cures for some of these conditions.

Of course, and I think you were bringing this up, one of the challenges is like when we do this.

because especially for disorders of the brain, neurodevelopmental disorders, so autism and other neurodevelopmental disorders, the question is always how early it is too late.

You know, how much damage has it done?

Has it been done?

And how much can I actually correct?

And that's one of the things that, you know, we're only now starting to really explore as we're thinking about some of the first clinical trials in the space.

And then for gene therapies, again, in the context of what you're mentioning is some of this, again, they're irreversible.

So once you put the gene in and it goes into a cell, let's say through a lentivirus that will integrate, you can't take it out anymore.

That would be very difficult.

It will get inactivated over time.

So that's why we have to be extra careful with some of these therapies and make sure that we don't do more harm, which I guess it's always what we try.

Absolutely.

Let's start first with stem cells and what they are, because I think it's also important to define them.

So stem cells are cells that have two properties.

First of all, they in principle can become other cells.

And if they are of the most potent type, they will be totally potent, so they can make everything.

If they're pluripotent, they can make almost everything.

And then, of course, there are lower levels of potency for the cells.

So we all carry stem cells in us, right?

Not in the brain, or fewer in the brain for sure, but in the liver and in other organs, like in the gut, as we renew the gut,

You know, every few weeks, that is done primarily through the stem cells.

But those are restricted.

They can make everything.

They can make mostly that specialized cell type for which they have been so, like, primed.

Now, the earliest, earliest of stem cells, like those pluripotent, they are very important.

Those are present at early stages of development of the embryo.

And, of course, that happens post-conception.

So the challenge has been that you have to remove them from a fertilized egg.

And if life starts at conception, then, of course, you're interfering.

So I think a lot of the ethical debates have started because of that.

But you know, in early days, even if you were to do that, you wouldn't be able to keep those cells.

It turns out that the cells are very difficult to maintain.

And this brings us actually to the second property of the cells, which is that in principle, they can be maintained forever.

If you provide the right conditions, they will divide and stay the same forever.

Those are the two properties.

So, you know, you can keep them forever.

You can freeze them down, put them in a, you know, liquid nitrogen, bring them out anytime, and they'll start exactly where they left.

And then with the right guidance, they can become other cell types.

So only around, you know, 1998 was that when we could actually maintain some of the cells in a dish.

So somebody figured out a soup recipe.

of chemicals that you can add and these cells will survive.

Because up to that point, it was not possible.

So that triggered, of course, the promise of this field that now would be able to

take those cells and derive various organs, perhaps transplant them, replace organs.

Of course, that ended up being much more complicated.

And of course, there were all these ethical debates related to the source of those cells and what does it actually mean to use these embryonic stem cells.

And yet we've learned a lot about those cells in early days.

What are the properties of those cells?

And then almost 20 years ago,

Shinya Yamanaka was a scientist in Japan at the UCSF, came up with an absolutely brilliant idea.

You know, we were always thought that the development, the development of the human or of any, it's a one-way street.

Once you go down development, you never come back.

So once you start making, you know, a stem cell that is more restricted, and then at the end you make, let's say, a liver cell, you can never go back and become that pluripotent stem cell again.

And that generally is thought to be useful to protect us from like cancer or like any others where we don't have, you know, parts of our hands like differentiating into something else.

And he thought that maybe you could do that, not in a natural way, in an artificial way.

And that, of course, would be very useful.

So what he did is he went and he looked at the genes that are expressed in pluripotent stem cells at very, very high levels.

So very, very high levels.

And almost as gene therapy, because we were talking about gene therapy, he took like the top couple of dozens of these genes and then started adding them.

inside skin cells.

So he took skin cells, initially from mice and then from human, and then started adding them one by one, two by two, three by three, four by four, five by five, six by six, to see whether any of those cells, once they have this combination of genes that are expressed in pluripotent stem cells, would somehow get confused

and think that they're actually a pluripotent stem cell, and then go back in time and actually become a pluripotent stem cells.

And he showed indeed that a combination of four is enough.

Of course, you can have six.

And that ended up being what we today call the Yamanaka factor.

In a way, it was almost like alchemy, right?

Where you sort of like, you know, transform something into something else, right?

You make out of this metal, you make gold.

It was pretty much like that.

It was kind of like the essence of alchemy.

And it turns out that that discovery was so profound because suddenly you could take a skin cell from anybody and put those genetic factors in

turn those cells into pluripotent stem cells that we'd later on learn they're almost identical to those embryonic stem cells, and now have those cells from any of us and use them for various purposes, perhaps for, let's say, making blood cells in the future, or perhaps to model something outside of the body.

And I was finishing my clinical training around that time.

And I remember even seeing that paper

And of course, in my naivete at that time, I thought, wow, this is it.

This is going to be, you know, the entry point for studying human neuroscience.

I was doing experiments at that time, studying actually the cortex and recording from animals, electrical activity of those neurons.

And always like thought, it's like this disconnect between what I was seeing in the clinic, which were these patients with severe profound autism.

And then recordings from the brain and thinking, we're never going to be able to do that.

How are we going to understand this complex disorder of the brain if we cannot even listen to the activity of those cells live?

And then suddenly, like seeing that discovery, you know, again, naive at that time, thought, well, that could be perhaps the way in which we could make neurons from any patient.

And so very soon after I came to Stanford, which I guess where we met, with sort of like this idea in mind that we will be able to make neurons from these patients

and rebuild maybe some of the cells or some of the circuits of the brain outside of the body without doing any harm because we're not doing a biopsy of the brain or anything invasive, just essentially creating a replica of some of those cells outside of the body and then finally study them at will in a dish and do all kinds of experiments where you remove things and add things and perhaps one day even develop therapeutics.

And here we are.

16 years later, since that process really started, took a long time.

But now for the first time, we've gotten such a good understanding of some of these conditions, and one of them in particular, that actually a therapeutic is inside and we're preparing for the first clinical trial that is really arising exclusively through studies done with this human stem cell models without actually using any animal models, just essentially creating, recreating cells and circuits outside of the brain of those patients.

So those cells that are collected from the umbilical cord are stem cells, but they're already quite restricted in what they can make.

So their applications are also restricted, mostly to blood disorders.

So I think it's important to keep in mind that they're not so like a universal solution to anything that would ever involve pluripotent stem cells in the future or stem cell therapies in the future.

So again, I think it's important to know that while they have certain applications and there have been quite clear cases where the availability of those cells were useful in a blood disorder in that child later on, they're certainly not...

You know, they have these universal uses as maybe sometimes they're being advertised.

Well, that's one of the problems.

Very often we don't even know what is being injected.

I think that is like a very important aspect.

We don't know what is in.

Sometimes are the cells from the patient that are being collected.

Sometimes this umbilical cell, sometimes we don't even know what cells are being injected.

Like it could be cells from somebody else.

Yeah, they're incredibly risky procedures.

Of course, they've never really been observed.

There have been very few of any clinical trials trying to really address it in a very systematic way.

And very often that's also the case, you know, that's also because they're not really justified.

So in the context of autism, this is very often like done, you know, and it's done not just in South America.

Sometimes there are places in Europe where you can get an injection of some stem cells for autism.

Some cells that are collected from the patient.

It depends a little bit on where it's done and how it's actually done.

But again, even from a biological point of view, what are those stem cells collecting?

presumably doing, let's say, in autism.

You know, we don't think that there is a cell type that is missing in the brain, so it's not like those cells can go.

And I think, as I was mentioning before, most of the cells are already restricted in their potential.

They can no longer make any cell types.

So, you know, the idea that you take these pluripotent stem cells and you just inject them, let's say, in the knee, and it will, like, miraculously grow, you know, cartilage, it's very often not really the case because those cells are not even capable of making cartilage.

So I think there's, you know, very often, you know, a lack of understanding of what these therapies really are.

And then, of course, there is sadly a lack of understanding of what is actually being injected.

So, you know, for autism, this is unfortunately happening much more often than you would think.

So I very often get like parents or families that are asking me desperately, you know, with like exhausted old resources.

We don't know what else to do.

We've tried behavioral therapy.

We've tried these therapies.

Nothing works.

And everybody's recommended that we should just go now to South America and do this injection.

Should we do it or not?

And of course, my answer is always like no, because again, there's no reason that that would work.

Some parents come back and of course they report an improvement.

And which is generally temporary to the extent that we know, of course, it's never really been studied in a very systematic way.

Partly, of course, there is a very strong placebo effect, which you can, you know, especially in parents, like by proxy, when you have a child who's like very sick, those placebo effects are very, very strong.

These parents really want those kids to improve.

So they will see things that are improving.

Plus, those are still developing kids, right?

So week by week, they may acquire new milestones.

And then the other thing, which of course could be part of this, is that there is an inflammatory effect very often.

And so that's almost like the fever in a way, right?

Like would increase perhaps some of the cytokines, will create a fever, perhaps that is associated.

We don't really know.

But certainly there are dangers associated with procedures like this that lack the rationale, first of all.

And then, of course, then they lack any regulatory framework.

Absolutely.

And especially when it is in the brain where there's not enough space.

So we know that anything that grows in the cranial cavity will actually push down vital centers.

So there are certainly risks associated with that.

So around like 15, 16 years ago, we were able for the first time to get some of the cells that are now known as induced pluripotent stem cells.

These are the Yamanaka.

Yes.

Or iPS cells.

So induced because they've been induced to become pluripotent in an artificial way.

But again, they stay like that.

So you can share them with anybody else like afterwards.

So we got some of those first cells in those early days.

And now the question was, how do we make neurons?

And what you do is you really kind of like leverage the...

everything that is known in developmental biology.

So we already know that there are certain molecules that are very important for making neurons.

So all you do is you put those cells in a dish, in a plastic dish, in a petri dish, and then you start, almost like when you cook, you start adding various molecules on top, and you see what happens.

And we knew that it's actually quite easy to make neurons.

That was already known.

There have been a lot of experiments done the decade before that showed that even if you just remove some of the factors that maintain those cells pluripotent, those pluripotent stem cells will start now to differentiate, and they like to become neural cells.

By default.

Almost by default.

So it's actually not that difficult to make neurons.

So in those early days, you know, you would take those cells, play them nicely, those pluripotent stem cells in a dish, and then remove some of these factors.

And then within a few days, you will see that they'll change shape.

And within a few weeks, some of them will really look like neurons.

And when you look at them, you can even look at proteins that only neurons will have.

You can actually get an electrode inside a cell and listen to the electrical activity.

So it was very exciting, as maybe you remember in those days.

I mean, you know, this bursting curiosity is always sort of like the ATP of the life in the lab, so to speak.

I mean, you just kind of want to wake up and want to go see what happened to those cells.

And it was clear in those days that, you know, we would be able to make those cells, but would we actually see any abnormalities in those cells?

I think it was like the question.

You know, how would you know if you derive cells from a patient with autism, how would you know that you found anything abnormal?

I think that was like the question.

You know, we didn't even know what would be abnormal in the brain.

And so that's when we decided actually to focus on something that would be relatively predictable.

And that was this mutation in a calcium channel, which was discovered just a few years before in very few patients that had essentially one single letter in their entire genome changed.

in a gene that makes a protein known as a calcium channel, sits in excitable cells, meaning cardiac cells and brain cells.

And every time a cell receives electrical input, this protein opens up and lets calcium go inside the cell.

And that's very important because it couples electrical activity of the network with chemical activity inside the cells.

And what we knew about that mutation at that point, and that's pretty much all we knew in those early days, is that it probably...

allows the channel to stay open slightly longer, just a little bit longer.

So more calcium would go inside the cells.

Of course, there would be no way to know because you can't get a neuron or a cardiac cell from those patients to actually test it.

So what we did is essentially we made, we recruited some of these patients, we flew them to Stanford.

Then we got a tiny skin biopsy, made this iPS cells.

This takes months.

This takes already like four or five months.

And then we took those cells in a dish, started to deriving neurons.

And after about five, six, seven weeks, then we put them under a microscope and we started looking at the calcium.

You can measure calcium inside cells through a microscope and just literally look at it.

And I'll never forget that day, you know, when we did that experiment, was looking down the microscope and we essentially stimulated the neurons and you could just see how control cells will go.

Calcium goes inside the cells and then it goes out.

And then in patients that had Timothy syndrome, so in Timothy syndrome-derived neurons, you could see how the calcium will go, and then it will stay longer.

It takes longer to go out.

So it's like the first defect that we saw in patient-derived neurons that were actually not coming from a biopsy.

They were not coming.

So that was incredibly exciting, as you can imagine.

But it was still relatively simplistic, just a few neurons at the bottom of a dish.

Of course, for me, what was particularly frustrating was that we couldn't go very far in development.

So think about the cerebral cortex, the outer layer of the brain that presumably makes us human.

It has multiple layers, a large diversity of neurons.

You know, it takes 27 weeks.

to make all those cells in the cortex, 27 weeks to make all those neurons.

And we're not even talking about glial cells, the supporting cells that are coming much later for several years afterwards.

But just making those cells takes about 27 weeks.

And it turns out, something that we discovered through experiments done in a dish, is that the timing of the development of those cells, it's actually recapitulated in a dish as well.

So if you keep the cells in a dish, they'll actually essentially develop at the same pace.

They're not like much faster.

And it's very difficult to keep neurons in a dish for 27 weeks to get all the neurons.

Essentially they peel off, you know, every time you start to move them to another plate, then at one point they just die.

And so then we thought, how about like never letting them to sit down on a surface?

How about just essentially aggregating them as balls of cells and then letting those float?

And in those early days, there was this amazing scientist from Japan, Yoshiki Sasai, who started doing really beautiful experiments where he was already moving some of these studies that he was doing of development in 3D culture.

So he showed you can make an optic cup, a part of the eye.

And so it was clear it was in the air, this revolution of actually moving cells from 2D flat cultures to 3D self-organizing.

And that actually unleashed amazing new properties of the cells.

So essentially all we did in those days is I ordered from Germany this plate

that were counterintuitively coded so the cells never stick, right?

I mean, every time we keep cells in a dish, you want them to stick.

That's the major problem.

So they were actually coded so the cells will never stick.

And then there were like these balls of cells, they were floating there.

And of course, I remember talking in the lab and everybody was like, oh, they're not going to survive.

It's going to be a couple of weeks and they're going to...

And then a week passed and two weeks passed, and then they kept growing and growing.

And, of course, the enthusiasm of every day to see, are they still alive?

And then we discovered that we can keep them for months.

And these three-dimensional cultures are now known as organoids, which is perhaps not the most fortunate name because it suggests that it's organ-like.

And of course, they're not an entire organ.

So they're not a representation of the entire brain.

But that's sort of like the term that we refer these days to anything that is sort of like three-dimensional and organizing in some way.

And so we started keeping these cultures.

And then at one point, actually, we discovered that we can pretty much keep them indefinitely.

My lab maintained the longest cultures that have ever been reported, like literally going for years, for two, three years in a dish.

And at one point in those early days when I was running out of funds in the lab and I came one day in lab meeting, I'm really, you know, determined to for us to actually like cut costs.

So I've told everybody, go into your incubators because we're spending so much money in feeding the cells and everybody throws out 20 percent of your cultures.

And then people started saying, so should I throw the ones that are like 500 days old?

And somebody was like, the ones that are 800 days old?

And I said, what?

You guys are keeping them for such a long time?

Yeah, they just keep growing.

They're in the incubator.

So then we actually did the first study.

And then we had a series of three studies done over the years of like trying to ask, how far do they go in development?

So if you have a clump of human neurons that you've made from pluripotent stem cells and you keep feeding them in a dish, how far do they go in development?

Do they...

Move much faster?

Do they move much slower?

Are they stuck at one point in development?

And it turns out that they actually keep track of development beautifully to such an extent that, for instance, we discover when they reach nine months of keeping them in a dish, so about the time of birth, they literally switch to a postnatal signature.

On their own.

In a dish?

In a dish.

So, you know, there's this classic example

In developmental neurobiology, there's this protein that usually changes around the time of birth.

It's an NMDA receptor.

So maybe some people know about NMDA receptors.

Binding glutamate, they're very important.

But they change a lot during development.

They're made out of different units, and the units change.

And it was very well known that during early development, so prenatal, before birth, you primarily have two B subunits.

And then after birth, they're primarily 2A.

So if you look in brain development, you just see how essentially 2B goes up and then it goes down and 2A goes up.

And when you look, they meet around birth.

So very often people thought that it's birth itself that triggers that switch.

That canonical, it's called a canonical switch because we all thought that it was like so classic.

And then you take an organoid that you maintain in the dish for 600 days.

And of course, we're not inducing birth.

We're not changing media.

We're not doing anything special.

And no hormones from mom.

No hormones changes.

Like, you know, we keep exactly the same media, which is certainly a very simplistic, you know, kind of like soup of chemicals, but we don't change it.

And then you just look at this two subunits and you see how like 2B goes down and 2A goes up and they pretty much meet that nine months of keeping them in a dish.

So that tells us that there's some sort of intrinsic clock.

Once you start a development,

the cells measure really, really well the time of development.

That does not mean that all aspects of development are going to now be recapitulated in a dish, but it tells us that there is this incredible ability of cells, especially in the nervous system, because of course those cells will keep for the rest of our lives.

No, we're never going to renew neurons.

It's going to be different for liver cells or gut cell.

But for neurons, probably in particular, they'll need to keep track of time really, really well.

So that was like the first discovery that we've sort of like made, which is still stunning today.

We still don't know the mechanism.

We're still working really hard on figuring out exactly how the cells are keeping track of time.

Because as you can imagine, if we understand what that molecular machinery is, we used to call it a clock.

We now call it a timer.

We think it's more of a timer than an actual clock.

But understanding what the molecular biology of that is will allow us actually to play with that clock.

So if you want to make neurons that are 70 years old neuron from a patient with Parkinson, I don't have to wait 70 years in a dish.

Could I make it in a few weeks?

Or perhaps could I take an aging neuron and somehow rejuvenate it by playing with that timer?

But just to make it clear, we still don't know.

We have some clues about what it may be, but I think it's still early days.

And I think that was one of the first things that this culture's

allowed us to do.

Just watch development, human brain development, outside of the human body in a dish and actually witness that some fundamental aspects of brain development are actually recapitulated even outside of the uterus and, of course, of the brain.

So that was the first.

And then, of course...

I guess I'm a developmental neurobiologist by training and I've done a lot of circuit work in early days.

Of course, an obsession of mine was that especially for conditions as complex as autism and schizophrenia, we need to recapitulate some of the circuit properties of the brain.

So we now know that probably both for schizophrenia and for autism, it is very unlikely based on the evidence that we have so far that there are cells really missing from the brain.

We thought for a while that maybe some cells are missing or maybe other cells are in excess.

But now the studies that have been done, especially with single cell profiling of brains of patients that have already died, showed us that the composition of the brain, of the cortex in particular, it's very, very similar.

So it's unlikely that the cells are missing, but likely the way they're connected with each other is that makes a difference.

And of course, in the beginning, we were just making this clump of cells.

They're all for the cortex, but they're not connected to anything else.

So then came the idea of assembloids because most of the cells in the brain connect with cells across the nervous system.

And in fact, even more interestingly, cells do not reside in the place in which they're born in the nervous system.

We have the largest cell diversity of any other organ, almost 2000 cell types.

By the end of the first trimester, there are about 600 cell types in the human brain.

You know, think about the liver, right?

Maybe a couple of dozens.

The brain has to make, you know, hundreds of times more.

So how do you do that?

The only way is to actually make the cell types in different parts of the brain, provide local cues there,

And then once the cells have been specified, let them move and find their final position.

So the first assembly that we've actually made were of a very stereotypical canonical movement of cells in the nervous system, which has to do again with the cortex.

So the cortex, again, the outer layer of the brain has both excitatory and inhibitory neurons.

It turns out that most inhibitory neurons are not born in the cortex, but they're born deep in the brain.

So essentially, all we did is we made two brain regions, the ones that has excitatory neurons and the one that has inhibitor neurons.

And the plan was to put them together, hoping that at one point, the cells will know what to do.

And in fact, that was one of the first projects in my lab, planning that.

And I remember I gave to one of the students this very difficult task of figuring out how we're going to fuse these two cultures.

And they're about three millimeters in size, so you can see them by eye.

And I thought it was going to be very difficult to put them together.

So the student worked for months trying to figure out, like, biological glues, you know, kind of like using various electrodes and impaling them and everything else until somebody else came one day and said, like, it's very simple.

You just put them at the bottom of a tiny Eppendorf tube, which is the tiniest, like, of tubes that you get.

You put them there overnight, and next day they're completely fused.

But they're not just fused because now if you look inside, within a few days, the cells that are supposed to move start to actually point out towards the cortex.

They literally smell the chemicals from the cortex and they start to move in this very stereotypical way towards the cortex.

And so that was the first disembloid made around 2015.

And I still remember it was Ben, actually.

Ben was so excited.

Ben Barris was so excited about like seeing the cells.

He wanted to look at these movies every day.

And then he said, I still have this email from him where he was very preoccupied that he kept saying like this new preparation is not an organoid.

It's not a steroid.

It's something else.

You have to find another name.

He loved naming things.

He loved naming things.

So he kept insisting that I should find the name.

So I made this long list I still have, like, in my notebook.

Like, I had a long list of about 20.

And I would, like, keep sending Ben one.

And, you know, like, Ben was always awake, like, 24 hours.

Yeah, he didn't sleep much.

He never slept.

So I remember after sending many emails going back and forth, and he was just like, no, bad name, bad name.

I don't like it.

And then at one point, I thought, well, OID because it's like, and then Assemble because we're Assemble the circuits.

So I thought Assembloid.

And I sent this and it says, perfect.

I love it.

So you named Assembloids.

I named Assembloids and Ben sort of like blessed it, like one night at like 3 a.m.

And so that was the first Assembloid.

And the first Assembloid was for cells migrating.

But then the question was, cells have to find each other and form circuits.

And so within a couple of years, we started making Assembloids that will have exons.

So the long projections of neurons finding other partners.

And you know how, forgot who said this, must have been Rodolfo Linas or, you know, who said that the brain is sort of, you know, the next evolutionary step towards movement.

You know, so like the nervous system,

There's been this theory that has evolved as a way of like moving around.

Well, we should check it.

Like who actually said it.

But for us, that became like the next objective.

Like, can we actually build a circuit?

that will have a very clear output.

So we would know that we've actually built that circuit.

So what we did is essentially we thought about like the simplest circuit for movement, which is like the corticospinal tract.

So that means that a neuron in deep layers of the cortex sends long axons all the way to the spinal cord, finds a motor neuron, makes a connection, then the motor neuron leaves the spinal cord, goes to the muscle.

And essentially, you only have these two neurons, right, that are connecting with each other, with the muscle, two connections, one between the two of them, and one with the muscle.

So the simplest of circuits that you can have.

It's very simple.

And of course, like in other species, it's a little bit more complicated.

It turns out that in mice, there's an additional neuron there.

So there are some changes that happened over evolution.

But for us and in primates, it's as simple as this.

So what we did was we essentially made an organoid that resembles the cortex and has some of those neurons.

And then we made an organoid that resembles the spinal cord.