WEBVTT

00:00:03.497 --> 00:00:10.497
This is the Convergent Science Network podcast. Leading researchers in the domain

00:00:10.497 --> 00:00:16.777
of neuroscience, brain theory and technology are interviewed by Paul Verschoor and Tony Prescott.

00:00:21.137 --> 00:00:25.297
This is Paul Verschoor for the Convergent Science Network podcast

00:00:25.517 --> 00:00:29.337
here at the barcelona cognition brain

00:00:29.337 --> 00:00:35.977
and technology summer school of 2018 and um i'm with me elena galea welcome

00:00:35.977 --> 00:00:41.677
to the podcast elena thank you very much happy to be here great um and your

00:00:41.677 --> 00:00:47.877
your talk was about the astrocytes and also was immediately um,

00:00:48.877 --> 00:00:53.157
instructed not to call them glia anymore because they're astrocytes.

00:00:53.517 --> 00:00:57.717
So maybe this is also a good starting point for a discussion, right?

00:00:57.777 --> 00:01:02.157
So why do we have to drop this notion glia and focus on astrocytes?

00:01:03.377 --> 00:01:09.577
Well, because it doesn't make any sense to classify cells between neurons and

00:01:09.577 --> 00:01:14.777
non-neuronal cells because the non-neuronal cells are so different among themselves.

00:01:14.777 --> 00:01:17.077
So this is like saying in the whole body.

00:01:18.214 --> 00:01:24.774
That the body is divided into heart and the rest of the organs, intestines and liver.

00:01:24.874 --> 00:01:27.674
It makes no sense because those organs are so different.

00:01:27.874 --> 00:01:32.634
They have different problems and different features. You have to focus on them individually.

00:01:33.254 --> 00:01:39.514
So the non-neuronal cells include cells that are so different in their functions

00:01:39.514 --> 00:01:44.054
and in molecular phenotype, in their contribution to pathology,

00:01:44.054 --> 00:01:46.374
contribution to higher brain functions,

00:01:46.554 --> 00:01:49.094
that just classifying them,

00:01:49.314 --> 00:01:51.914
just putting them together in a single concept,

00:01:52.314 --> 00:01:59.434
it just doesn't force us to look into the details. I think details are very, very important.

00:01:59.854 --> 00:02:05.754
Names are very, very important because they guide research and they guide the

00:02:05.754 --> 00:02:07.674
understanding of problems.

00:02:08.034 --> 00:02:12.754
So it makes sense one century ago to see neurons and the rest of the brain.

00:02:12.754 --> 00:02:14.914
Now it doesn't make sense any longer.

00:02:15.294 --> 00:02:20.374
Okay, so basically you're saying with many cell types that build the brain,

00:02:21.174 --> 00:02:22.274
some of those are spiking.

00:02:22.534 --> 00:02:25.234
They generate action potentials. Those are what we call neurons.

00:02:25.734 --> 00:02:30.654
But then the many other cells, which is a very variable set of cells that formerly

00:02:30.654 --> 00:02:37.554
were known as LEA, the non-neuronal cells, but it is not in any way descriptive of this variability.

00:02:38.394 --> 00:02:44.054
And then among this highly variable set of cells, We have a subtype called estrocytes, right?

00:02:44.134 --> 00:02:49.734
So what then sets the estrocytes apart from these other non-neuronal cells?

00:02:49.734 --> 00:02:52.434
Well, first, molecularly, they're different already.

00:02:52.734 --> 00:02:54.094
They have a different identity.

00:02:55.034 --> 00:03:01.114
And they carry functions that the others don't. They don't generate myelin,

00:03:01.314 --> 00:03:03.114
like oligodendrocytes.

00:03:03.194 --> 00:03:09.174
They don't deliver blood, like the vascular cells. cells, they don't protect

00:03:09.174 --> 00:03:14.374
neurons from some danger like microglia.

00:03:14.554 --> 00:03:17.314
So at least they don't do what other things do.

00:03:17.554 --> 00:03:23.074
And perhaps they do things like providing lactate to neurons,

00:03:23.434 --> 00:03:28.394
removing potassium from the extracellular medium.

00:03:28.714 --> 00:03:34.194
They transform glutamate into glutamine. So they have very, very specific functions.

00:03:36.255 --> 00:03:39.135
On the astrocytes too and none on the rest of the cells too.

00:03:39.475 --> 00:03:43.795
So this is what is important to focus on them.

00:03:44.075 --> 00:03:47.955
But they stand apart also on purely morphological grounds?

00:03:51.375 --> 00:03:57.755
Morphologically, yes, depending on what comparison you carry out.

00:03:58.255 --> 00:04:03.435
Because there are other cells in the brain called NG2 oligodendrocytes.

00:04:03.435 --> 00:04:07.075
They are like between astrocytes and microglia.

00:04:07.255 --> 00:04:10.155
So if you compare an astrocyte with a neuron, they're very different.

00:04:10.975 --> 00:04:16.635
So neurons have their dangerous spines and axons, while astrocytes are very

00:04:16.635 --> 00:04:22.815
complicated cells, bushy cells with very intricate distal processes.

00:04:23.735 --> 00:04:31.515
But NG2 have also primary, secondary processes and some dense distal processes.

00:04:31.655 --> 00:04:35.155
So they are intermediate. And in microglia, they have very low processes,

00:04:35.195 --> 00:04:36.875
and they don't have the distal processes.

00:04:37.135 --> 00:04:44.795
So it seems that there is a morphological gradient in the brain that differentiates,

00:04:46.255 --> 00:04:51.395
neurons from microglia, but not so much NG2 from astrocytes.

00:04:51.635 --> 00:04:53.995
Vascular cells are very different from the rest of the cells.

00:04:54.135 --> 00:04:57.455
All the goals are very similar to neurons in many ways.

00:04:57.455 --> 00:05:03.195
So, depending on the comparisons you tell us, that the morphological differences

00:05:03.195 --> 00:05:07.775
are larger or smaller, but yeah, morphologically, they're very,

00:05:07.835 --> 00:05:09.235
very different, absolutely.

00:05:10.215 --> 00:05:13.915
Molecularly and morphologically, very different. So, in the human brain,

00:05:13.955 --> 00:05:18.795
we have about 90 billion neurons, right, according to the most recent counts.

00:05:20.355 --> 00:05:24.195
Do we have then 90 billion non-neuronal cells?

00:05:27.001 --> 00:05:30.741
Marisa, I don't know, what is her name? Herculano. So those people,

00:05:30.861 --> 00:05:34.021
that's a fantastic word. There are papers.

00:05:34.181 --> 00:05:38.481
She has analyzed that with very objective technology.

00:05:38.861 --> 00:05:45.841
And so the answers are there. I don't think the answer is that equal amount of neurons. 50-50?

00:05:46.281 --> 00:05:51.881
I think it's 50-50. Depends on the animal and depends on the area of the brain.

00:05:52.421 --> 00:05:54.081
So I don't think it's 50-50.

00:05:54.721 --> 00:05:57.621
She has calculated those. Those numbers.

00:05:57.921 --> 00:06:01.181
But now, the often non-neuronal... Do you think this is relevant?

00:06:01.341 --> 00:06:03.821
Like numbers of... Well, just plenty of ratios.

00:06:04.241 --> 00:06:08.341
Ratios. It was like... Well, you gave some numbers actually in your talk, right?

00:06:08.381 --> 00:06:14.101
Where you said that... Where do I have... On average... In humans...

00:06:14.101 --> 00:06:21.281
On average, like two astrocytes per each 10 neurons. I think that's an average calculation.

00:06:21.961 --> 00:06:25.921
On average. In mouse, it was 1 to 20 neurons per astrocyte, right?

00:06:25.961 --> 00:06:31.821
You said. Yeah, or 1 to 20 or 2 to a… it's in that world park.

00:06:31.821 --> 00:06:35.901
It's in difference for humans as well, I would say. Yeah, probably. So it would be the same.

00:06:36.581 --> 00:06:41.581
But what I meant is that there was, for many years, the notion that astrocytes

00:06:41.581 --> 00:06:44.801
were 10 times more abundant than neurons. Exactly.

00:06:45.942 --> 00:06:51.282
First, this is not true. And I don't think the numbers are so relevant in the

00:06:51.282 --> 00:06:55.962
sense that the fact that you have more of something makes that something more

00:06:55.962 --> 00:06:58.502
relevant, perhaps not necessarily.

00:06:59.442 --> 00:07:02.222
I think it's the function that is more relevant.

00:07:02.302 --> 00:07:08.222
You may have something that is just present in very low amounts,

00:07:08.322 --> 00:07:11.662
but it's for some reason incredibly important for the function of the brain.

00:07:12.242 --> 00:07:15.322
So numbers are interesting, but I don't think that is so relevant.

00:07:15.322 --> 00:07:20.362
And definitely, astrocytes are not 10 times more abundant in the brain than...

00:07:20.362 --> 00:07:25.622
But as the non-neuronal cells, how abundant are the astrocytes compared to the

00:07:25.622 --> 00:07:28.942
other non-neuronal cells? I don't have the numbers. Herculano has them.

00:07:29.942 --> 00:07:38.142
But just in your imagination, astrocytes are a relatively small subpopulation

00:07:38.142 --> 00:07:40.662
of cells making up the brain?

00:07:42.182 --> 00:07:48.522
So they're not that dominant? if I know it's directed so and then is if you

00:07:48.522 --> 00:07:49.762
look at the non-neuronal cell types,

00:07:50.322 --> 00:07:55.382
is there anything special about their distribution in the nervous system if

00:07:55.382 --> 00:07:59.442
we go for more primitive parts of the brain like the brain stem and the spinal

00:07:59.442 --> 00:08:04.202
cord or two more frontal areas do you see any kind of differentiation across

00:08:04.202 --> 00:08:08.362
these non-neuronal cell types I don't think that this has been analyzed.

00:08:09.322 --> 00:08:17.262
We spoke about tiling, and I think tiling is a feature of astrocytes that neurons don't have.

00:08:18.022 --> 00:08:22.882
And perhaps there is more tiling the more frontal we go.

00:08:23.062 --> 00:08:26.782
So the more developed...

00:08:28.000 --> 00:08:33.720
The area, the more tiling. Regular, if you want. More regularly and perhaps tighter.

00:08:34.240 --> 00:08:39.240
So the tiling function, whatever it is, is more developed. That's perhaps the case.

00:08:39.360 --> 00:08:42.540
But those kind of analysis have been performed.

00:08:42.820 --> 00:08:47.460
And from an evolutionary perspective, if we go to, say, primitive vertebrates,

00:08:47.520 --> 00:08:51.020
let's say, lamprey or fish,

00:08:51.340 --> 00:08:59.600
or we go to invertebrates, Do we see any kind of patterning of the non-neuronal cell types?

00:09:00.000 --> 00:09:07.300
I'm not sure about that. I know the difference between mice and humans,

00:09:07.460 --> 00:09:13.260
that astrocytes are more complicated structurally in humans than in astrocytes,

00:09:13.300 --> 00:09:14.580
and there are more subtypes.

00:09:14.860 --> 00:09:18.720
And I don't know exactly, but I think in invertebrates, yeah,

00:09:18.780 --> 00:09:25.680
there are astrocytes. but I'm not sure about their colonization or whether they

00:09:25.680 --> 00:09:29.460
are But it also means people might not have really looked at that. No.

00:09:29.800 --> 00:09:36.400
Okay. So that's interesting, right? Because one thing that you also made clear

00:09:36.400 --> 00:09:40.920
in your talk is that the whole idea already of the morphology of astrocytes has changed a lot.

00:09:41.000 --> 00:09:44.600
Because initially because of just the immunocyte chemistry available.

00:09:45.040 --> 00:09:47.480
Because there are better techniques now to the astrocytes, yeah?

00:09:47.540 --> 00:09:50.960
Initially And actually, they look like very sort of punctuated,

00:09:50.960 --> 00:09:52.480
star-like... Star-like cells.

00:09:52.700 --> 00:09:55.000
It's mostly with long processes. Yeah.

00:09:55.380 --> 00:09:59.860
And they're not. They're like balls of a lot of hairs.

00:10:00.040 --> 00:10:02.860
Exactly. That's the way they are. Right. And that's probably indicative of their

00:10:02.860 --> 00:10:05.260
function. Really? In the way... Yeah, absolutely.

00:10:05.640 --> 00:10:09.660
So, and that's also some of your own work that you showed. Yeah.

00:10:10.743 --> 00:10:19.083
To get to the tiling, so it looks like we have these bushy astrocytes that are

00:10:19.083 --> 00:10:24.583
forming sort of a tessellated pattern throughout the brain,

00:10:24.783 --> 00:10:28.903
which is structured in a three-dimensional way.

00:10:30.083 --> 00:10:33.963
So do you also think this tiling is highly structured in three dimensions?

00:10:34.443 --> 00:10:37.823
Oh, absolutely. Like bricks? Absolutely. Like we shouldn't hit the boxes?

00:10:38.003 --> 00:10:40.063
Absolutely. It's like a boronite tessellation.

00:10:40.663 --> 00:10:44.283
Absolutely, it is. I know. We have published a paper, actually,

00:10:44.303 --> 00:10:51.923
where the tiling of astrocytes, it's really well modeled when you do borneotisolation of the brain.

00:10:52.003 --> 00:10:58.483
So you take the brain and you just do this kind of tisolation from,

00:10:58.483 --> 00:11:05.103
you just have seeds there and then you stabilize the distances between the seeds

00:11:05.103 --> 00:11:06.763
and you draw walls there.

00:11:07.063 --> 00:11:09.103
And that way you divide the...

00:11:10.063 --> 00:11:14.163
The brain, like a given volume in different voronoi volumes,

00:11:14.603 --> 00:11:21.083
and it's identical to the desolation, to the natural desolation of astrocytes.

00:11:21.923 --> 00:11:25.943
Indicating that, I think this is indicative of the way they grow.

00:11:26.163 --> 00:11:31.383
So probably when they grow, either from radionuclidea or from clonal expansion

00:11:31.383 --> 00:11:37.043
from local astrocytes, they sort of grow until they find another astrocyte.

00:11:37.043 --> 00:11:42.523
And that's actually how voronoi's deflation occurs and develops.

00:11:43.023 --> 00:11:46.103
But that's for the neocortex, mainly, no?

00:11:46.223 --> 00:11:50.503
Yeah, that's probably the neocortex, and probably the rest of the brain is like

00:11:50.503 --> 00:11:54.923
that, but with more mistakes, with lower quality, perhaps.

00:11:55.683 --> 00:12:02.103
So my idea that's in the neocortex, the signaling that determines the tannin,

00:12:02.103 --> 00:12:06.163
I mean, the synodyms that I think are repulsive pathways,

00:12:06.503 --> 00:12:14.923
like hip frames or semaphore frames, probably is more developed and works better.

00:12:15.183 --> 00:12:20.523
But in the rest of the brain, that occurs, but perhaps in a more primitive fashion,

00:12:20.763 --> 00:12:26.043
more, taking a picture, in a more primitive fashion.

00:12:26.703 --> 00:12:32.483
So, probably the same elements all over the brain, but with different quality.

00:12:33.303 --> 00:12:40.703
Okay. So, the ester sites, they have this really dense projective field of processes,

00:12:41.363 --> 00:12:43.903
which is, you said, about 50 micron across?

00:12:45.043 --> 00:12:51.683
The whole size, yeah. So, the ball is the diameter, yes, up to 50 microns.

00:12:51.883 --> 00:12:57.323
Okay. The whole ball from 35 to 50. 50 is a good number.

00:12:58.363 --> 00:13:03.483
So would that be compatible to something like a cortical column?

00:13:05.983 --> 00:13:06.843
Um...

00:13:09.933 --> 00:13:15.093
I think columns are, aren't they larger? I'm not sure about this.

00:13:15.293 --> 00:13:19.433
And they're wider, I think they're wider than that. But do you see it aligned

00:13:19.433 --> 00:13:22.693
with, let's say, the structures that neurons would be like?

00:13:23.013 --> 00:13:26.833
There is one paper where they do fate analysis,

00:13:27.113 --> 00:13:33.333
so they label the astrocytes, and they see in the adults where they are with

00:13:33.333 --> 00:13:39.513
radial glia, and they do see columns, but they are not all over the place.

00:13:39.933 --> 00:13:45.413
So, it's not that columns have astrocytes associated to that column, all the columns.

00:13:46.453 --> 00:13:51.833
The notion is that some of the columns have astrocytes, indicating that the

00:13:51.833 --> 00:13:53.873
origin of astrocytes is mixed.

00:13:54.253 --> 00:13:58.933
Some of them come from the radial clea that gives rise to the column,

00:13:59.013 --> 00:14:00.973
and some of them come from elsewhere.

00:14:02.113 --> 00:14:07.413
And that creates a complex pattern. And the astrocytes would only be in the

00:14:07.413 --> 00:14:10.853
gray matter of the nervous system. No, there are also white matter astrocytes.

00:14:10.853 --> 00:14:13.113
Absolutely. They're different as well as they are.

00:14:13.713 --> 00:14:18.093
And they have a different morphology and different molecular phenotype too.

00:14:18.753 --> 00:14:22.773
But there's essentially just a continuous matrix, if you want,

00:14:22.913 --> 00:14:26.453
of astrocytes. So if I take a brain.

00:14:27.992 --> 00:14:31.532
If I could visualize all the exercise in that brain, it would be just the whole

00:14:31.532 --> 00:14:35.912
volume would be covered with sort of very regularly spaced exercise.

00:14:37.152 --> 00:14:42.872
Yeah, absolutely. With the different densities, different morphologies, depending on the area.

00:14:43.032 --> 00:14:47.112
But it would be a continuous. Yeah, that's a nice experiment to do with a clear

00:14:47.112 --> 00:14:48.412
brain technique. Absolutely.

00:14:48.752 --> 00:14:51.292
Before you were thinking about that. Yeah, that's a good experiment.

00:14:51.372 --> 00:14:52.292
But that would be the prediction.

00:14:52.432 --> 00:14:56.392
To stain estros exclusively. And yeah, you will, we will see,

00:14:56.392 --> 00:15:01.272
uh, different tiling of astrocytes all over the brain. Yeah, absolutely.

00:15:01.492 --> 00:15:05.692
That would be nice. So a nice picture to have. So would that structure then,

00:15:05.772 --> 00:15:09.712
so what's the density then, if you take just a bit of brain,

00:15:09.912 --> 00:15:12.212
a cubic millimeter, right?

00:15:13.132 --> 00:15:19.092
In your mind, then these, these mushy astrocytes would really fill up that That

00:15:19.092 --> 00:15:27.592
whole volume and then the neurons and the blood vessels are sort of squeezed inside that matrix?

00:15:29.652 --> 00:15:36.572
Everything is squeezed there. And probably the arrangement of all the cell types

00:15:36.572 --> 00:15:39.172
is highly coordinated and regulated.

00:15:39.672 --> 00:15:42.852
And vessels probably play a big role too.

00:15:43.012 --> 00:15:48.252
So everything, the vessels grow and everything adapts to the rest of the space.

00:15:48.252 --> 00:15:50.932
And creates their own surface there.

00:15:52.072 --> 00:15:57.912
But then, couldn't you speculate that these astrocytes are there to give some

00:15:57.912 --> 00:16:02.872
sort of uniform mechanical support to the rest of the system, right? That could be it.

00:16:04.572 --> 00:16:11.912
So I think that when astrocytes were described as a glue, initially named glia.

00:16:14.408 --> 00:16:21.088
Glue is not a bad term for astrocytes, actually, in the sense that they're all over the place.

00:16:21.708 --> 00:16:27.208
The difference is that glue implies it's a passive support, and this is possible.

00:16:27.268 --> 00:16:32.868
Perhaps there is some, as you say, structural support for the rest of the structures.

00:16:33.108 --> 00:16:38.428
But I do think myself that that structure is also actively exchanging information

00:16:38.428 --> 00:16:43.108
with the rest of the cells. So, it's not just a passive thing there that you

00:16:43.108 --> 00:16:46.268
may have break or real glue. It's an active glue.

00:16:46.988 --> 00:16:52.768
So, whether it has also supportive elements and guiding elements to its structures, it is possible.

00:16:53.888 --> 00:17:00.128
Because would you know of any other cell type, non-neuronal or neuronal,

00:17:00.268 --> 00:17:06.688
that would have this feature of such a uniform distribution locally and globally, right?

00:17:08.248 --> 00:17:12.528
Microglia too. Microglia is all over the place, exactly.

00:17:13.848 --> 00:17:16.368
Also like a three-dimensional matrix. Yeah, yeah, yeah, yeah.

00:17:16.508 --> 00:17:19.228
And NG2 cells, which I find fascinating.

00:17:19.948 --> 00:17:23.748
And we don't know much about them, but they are all over the place, are highly abundant.

00:17:24.388 --> 00:17:30.728
And they're also, they have a tile. Microglia and NG2, they are also tiled.

00:17:31.248 --> 00:17:32.348
They're also territorial.

00:17:33.808 --> 00:17:36.328
It's just neurons that are not territorial, but yes, they are.

00:17:36.328 --> 00:17:39.828
And then if you could state them, you actually see them beautifully arranged,

00:17:39.988 --> 00:17:43.268
regularly arranged all over the brain. Right. Absolutely.

00:17:43.628 --> 00:17:48.828
So in some sense, you want to have a very strongly symmetric structure to distribute

00:17:48.828 --> 00:17:52.808
all mechanical forces so that you can allow neurons to be, let's say,

00:17:52.808 --> 00:17:55.888
asymmetric in how they process things. That's an interesting idea.

00:17:56.828 --> 00:18:00.748
Uh-huh. Okay. You think like an engineer.

00:18:02.188 --> 00:18:10.628
I don't know. Do I? that's a continuing thought but why not um well but so i think myself about.

00:18:11.842 --> 00:18:15.902
For me, the timeline, it's about information processing.

00:18:16.902 --> 00:18:20.542
Well, that's indeed the next question, right? For me, that's the idea.

00:18:20.862 --> 00:18:27.242
So they segregate information processing in a set of neurons.

00:18:27.322 --> 00:18:32.502
Either they synchronize that activity or they provide energy for a given set of neurons.

00:18:32.802 --> 00:18:37.322
So for me, that's the notion. It is related to information processing.

00:18:37.582 --> 00:18:44.122
Like in a circuit, it's a mini circuit. And I have no idea what they do there,

00:18:44.302 --> 00:18:51.722
but tallying is more associated to synchronization of activities in a given module in the brain.

00:18:52.722 --> 00:18:56.462
Okay, so let's look now at that function, right?

00:18:56.542 --> 00:19:04.562
So here we have the disaster sites, equally tessellated.

00:19:04.562 --> 00:19:07.322
It um we don't know what

00:19:07.322 --> 00:19:11.542
they do and it's actually a significant chunk

00:19:11.542 --> 00:19:14.542
of the of the cell volume of the brain that's

00:19:14.542 --> 00:19:21.062
correct so you spoke of the dark matter of of the brain and now we can start

00:19:21.062 --> 00:19:24.382
to speculate about what they do because actually we don't know much about that

00:19:24.382 --> 00:19:29.342
no so why has it taken us so long to even start to worry about what they might

00:19:29.342 --> 00:19:33.102
do functionally why no we we i that's a great question.

00:19:33.342 --> 00:19:41.542
I realize, and actually this is sort of a recent understanding about things, research.

00:19:42.642 --> 00:19:48.482
Develops very, very slowly, incredibly slowly, which is for the people that

00:19:48.482 --> 00:19:52.442
are not very patient, that includes me, that's very striking.

00:19:52.562 --> 00:20:00.242
So it takes us years to find something relevant, events, and sometimes the techniques

00:20:00.242 --> 00:20:02.982
are a limitation or an advance.

00:20:03.262 --> 00:20:06.042
So in the case of astrocytes, we spoke about this this morning,

00:20:06.522 --> 00:20:10.502
the reason why we know so much about calcium astrocytes is because we have calcium

00:20:10.502 --> 00:20:13.162
imaging, and that was fantastic.

00:20:13.402 --> 00:20:18.962
And since we don't have imaging or another readout for another phenomenon that

00:20:18.962 --> 00:20:22.082
may be more relevant, then we cannot talk about that phenomenon.

00:20:22.082 --> 00:20:26.282
So sometimes techniques determine concepts.

00:20:26.442 --> 00:20:32.962
We have to be very aware of that. So as to why it has taken so long, I think a big problem is,

00:20:34.342 --> 00:20:43.242
in addition to science being slow in nature, that the neuronal people haven't

00:20:43.242 --> 00:20:46.742
been interested in non-neuronal cells. Even...

00:20:48.257 --> 00:20:52.757
If we look at the number of references, I have a student, my student,

00:20:52.797 --> 00:20:54.597
Abel, I just defended his thesis.

00:20:55.117 --> 00:21:00.677
He had a picture about the number of citations, so the number of articles related

00:21:00.677 --> 00:21:02.717
to astrocytes, the number of articles related to glia.

00:21:03.217 --> 00:21:09.997
They're per year. So we have like 20,000 articles for neurons versus,

00:21:10.097 --> 00:21:13.077
I don't remember, 2,000 articles for astrocytes.

00:21:13.677 --> 00:21:20.997
So then that explains a lot. So the number of people working on neurons is larger

00:21:20.997 --> 00:21:22.917
than the number of people working on astrocytes.

00:21:22.997 --> 00:21:26.037
That determines also how fast discovery goes.

00:21:26.737 --> 00:21:30.737
And the second thing is that neuronal people are not interested in astrocytes.

00:21:31.937 --> 00:21:39.217
That's it. And then many of the neurons, the advance in neurons is earlier because

00:21:39.217 --> 00:21:41.837
the axon potentials were discovered in the 50s.

00:21:42.177 --> 00:21:50.657
Advancement in astrocytes started to happen in the 90s when we discovered Gaussian

00:21:50.657 --> 00:21:51.757
imaging for astrocytes.

00:21:51.877 --> 00:21:58.057
So we have been studying neurons for 50 years already.

00:21:58.277 --> 00:22:03.177
By the time we start to realize astrocytes do something else, other than being a glue.

00:22:03.417 --> 00:22:06.277
So that can explain a lot of things. The techniques are different.

00:22:06.497 --> 00:22:11.657
But I think a big problem is that neuronal people are not interested in non-neuronal stuff.

00:22:11.717 --> 00:22:14.957
But conversely, you can also say that astrocyte people don't care too much about

00:22:14.957 --> 00:22:15.937
the neurons. Yeah, exactly.

00:22:16.137 --> 00:22:20.577
But the impact of that in the neuronal field is minimal.

00:22:20.797 --> 00:22:26.077
I think the impact of all these people working on neurons of not caring about

00:22:26.077 --> 00:22:30.697
astrocytes is much larger. Yeah, but I think that's a very common impression

00:22:30.697 --> 00:22:35.257
people have, that whatever they do is not considered relevant by their colleagues.

00:22:35.497 --> 00:22:38.597
Right, it's true. I think it's just a sign of general fragmentation.

00:22:38.977 --> 00:22:44.857
But in our case, we have numbers. So systems neuroscience concerns mostly neurons.

00:22:45.357 --> 00:22:51.597
So in the big meetings that are specialized in systems neuroscience,

00:22:51.837 --> 00:22:56.917
we calculated that only 1% of presentations are dedicated to non-neuronal cells.

00:22:57.417 --> 00:23:03.057
You know, maybe the problem here is that right now there is not a clear proposition

00:23:03.057 --> 00:23:08.597
what astrocytes could contribute functionally to neural systems, right?

00:23:08.657 --> 00:23:12.777
And I think this is the big challenge. In some sense, if you say 2,000 papers

00:23:12.777 --> 00:23:16.637
a year on astrocytes, then you can also wonder, okay, what the hell are these

00:23:16.637 --> 00:23:20.877
papers about, given that their overall understanding is still so limited. Right.

00:23:21.782 --> 00:23:25.062
So the question is how astrocytes contribute to higher brain function.

00:23:25.262 --> 00:23:27.402
That should be our challenge, right? That's our challenge.

00:23:27.762 --> 00:23:34.142
Beyond clearing the test, which is very important, or beyond providing fuel

00:23:34.142 --> 00:23:38.262
to neurons, which is very important too, but it did do something else that is

00:23:38.262 --> 00:23:39.622
relevant to higher brain functions.

00:23:39.822 --> 00:23:42.602
And if we didn't have astrocytes, we wouldn't have those functions.

00:23:43.102 --> 00:23:49.082
So let's take a look. What do we know about the interaction between astrocytes

00:23:49.082 --> 00:23:54.722
and neurons? What are the outstanding features that astrocytes have that might

00:23:54.722 --> 00:23:56.822
be relevant if you are a neuron?

00:24:00.122 --> 00:24:06.062
Let's go for the data. So the data shows that you manipulate astrocytes and

00:24:06.062 --> 00:24:07.922
you change the response of neurons.

00:24:08.162 --> 00:24:12.722
So before any further interpretation of why astrocytes are necessary for neurons.

00:24:12.722 --> 00:24:20.622
So, in local circuits or in networks, doing things to astrocytes,

00:24:20.662 --> 00:24:22.502
manipulating their responses,

00:24:22.942 --> 00:24:28.882
changes the circuit without, changes the network without.

00:24:29.082 --> 00:24:33.502
That's a fact. That's the evidence, the experimental evidence.

00:24:34.122 --> 00:24:38.802
And I could say, look, if I just occlude the blood vessel, also something happens

00:24:38.802 --> 00:24:40.782
to neurons, but it's not necessarily informative.

00:24:41.022 --> 00:24:48.382
No, but one thing is that it's different if neurons suffer because they lack

00:24:48.382 --> 00:24:51.842
basic support, they lack oxygen.

00:24:52.202 --> 00:24:59.122
And another thing is that the neurons don't have action potentials or they have

00:24:59.122 --> 00:25:06.922
reduced action potentials because astrocytes are releasing different amounts of neurotransmitters.

00:25:06.922 --> 00:25:11.642
So, the signaling that is involved there is very different.

00:25:13.062 --> 00:25:16.842
Yeah, but for instance, what I was looking for, as was discussed,

00:25:17.202 --> 00:25:24.282
there's direct exchange of key ions that neurons need to be active,

00:25:24.422 --> 00:25:29.842
to change their membrane potentials, to generate action.

00:25:29.842 --> 00:25:32.582
But like there's sodium exchange you have potassium exchange.

00:25:34.962 --> 00:25:43.642
There's a calcium exchange between neurons and estrocytes right so apparently these ions are,

00:25:44.422 --> 00:25:49.482
exchanging so what are estrocytes doing with it what is our estrocytes doing

00:25:49.482 --> 00:25:53.122
with potassium are they buffering it are they they're buffering using it are

00:25:53.122 --> 00:25:58.682
they they're buffering just they bring it out of the estrosolar space to the

00:25:58.682 --> 00:26:02.862
blood that's the the standard knowledge that you have a lot of potassium channels.

00:26:04.030 --> 00:26:09.910
And in diseases, it's well known that it's a good therapeutic target in the

00:26:09.910 --> 00:26:16.510
sense that they get dysregulated, meaning that probably if they don't buffer potassium correctly,

00:26:16.770 --> 00:26:19.890
that contributes to disease and to neuronal activity.

00:26:20.270 --> 00:26:22.610
So the interpretation goes again to a more homostatic direction,

00:26:22.850 --> 00:26:27.770
like, okay, it's sort of just removing the potassium from extracellular space.

00:26:28.050 --> 00:26:32.310
Right. But another interpretation, it's like basic computation. mutation.

00:26:33.470 --> 00:26:41.210
So the fact that we spoke about it, filtering, so if there is a circuit,

00:26:41.490 --> 00:26:46.330
imagine that it's a circuit and neurons are part of the circuit and astros are

00:26:46.330 --> 00:26:48.050
part of the circuit as another element.

00:26:48.410 --> 00:26:51.910
So they may be doing things to that circuit. That's what I'm trying to figure out.

00:26:51.930 --> 00:26:59.950
Like they're maybe gating, changing gain, changing extras on the...

00:26:59.950 --> 00:27:02.930
Could they release potassium again in the extracellular space?

00:27:04.270 --> 00:27:07.330
They could release potassium. They could release glutamate. That goes back.

00:27:07.390 --> 00:27:12.970
It's like a third neuron that goes back to the neuron and changes the way the

00:27:12.970 --> 00:27:14.530
neuron responds to the next stimuli.

00:27:14.630 --> 00:27:20.690
So it means that the astrocyte essentially can control whether a neuron will

00:27:20.690 --> 00:27:23.390
respond at all to the next stimulus. It would modulate.

00:27:23.630 --> 00:27:26.890
I think it's better to modulate. In an extreme case, it could actually prevent

00:27:26.890 --> 00:27:28.090
it. Yeah, exactly. Exactly.

00:27:28.710 --> 00:27:35.950
Or anything could change. In the studies in the local circuits,

00:27:36.270 --> 00:27:43.610
it is known that depending on the input, so the inhibitory neurons, for instance,

00:27:43.930 --> 00:27:47.330
depending on the kind of response they have,

00:27:47.630 --> 00:27:53.910
they produce different responses and astrocytes that in turn relay different

00:27:53.910 --> 00:27:59.250
outposts to, etc. That's well known.

00:27:59.490 --> 00:28:04.430
And that's pretty delicate. So it's not just homostasis.

00:28:04.550 --> 00:28:14.550
It is pretty balanced reading of the neuronal activity and routing it to the

00:28:14.550 --> 00:28:16.690
other neuron in the circuit.

00:28:17.130 --> 00:28:20.210
So for me, this is beyond homostasis.

00:28:21.150 --> 00:28:28.210
This is the astros are acting like another neuron. and they can may change actually

00:28:28.210 --> 00:28:34.730
the output the global output of the receding circuit can be changed by astrocytes

00:28:34.730 --> 00:28:39.790
in a precise manner beyond like.

00:28:41.750 --> 00:28:47.710
Vascular cells delivering oxygen to neurons they can change that for you that

00:28:47.710 --> 00:28:50.610
doesn't mean that astrocytes may contribute to,

00:28:51.643 --> 00:28:56.443
Higher brain functions, being just there, there are metabolic roles.

00:28:56.883 --> 00:29:00.803
I think it would be great if exercise would help out.

00:29:02.223 --> 00:29:06.403
So I have no objections against that. So it's more about looking at what they

00:29:06.403 --> 00:29:10.803
can really do and what spatial temporal scale that could play out. That's such a question.

00:29:10.923 --> 00:29:15.283
If we talk about the potassium dynamics, this could play out at,

00:29:15.363 --> 00:29:19.663
let's say, millisecond level, the buffering and release of potassium.

00:29:19.663 --> 00:29:22.683
Or do you see this as a slow process?

00:29:23.003 --> 00:29:26.563
I think it's more of a slow process, but I have to look into it. And glutamate?

00:29:27.363 --> 00:29:31.623
Glutamate, the glutamate uptake, I think it goes in the millisecond scale.

00:29:32.743 --> 00:29:35.603
Glutamate release, it's in the second scale.

00:29:36.083 --> 00:29:42.063
So when the astrocytes respond to neurons by releasing glutamate, that takes seconds.

00:29:42.503 --> 00:29:47.143
But in hundreds of milliseconds and then in tens of seconds.

00:29:47.143 --> 00:29:50.523
And this releases the nonspecific in extracellular space?

00:29:50.843 --> 00:29:56.663
Well, it's sort of paracrine. Let's say paracrine, which is not a bad term because

00:29:56.663 --> 00:30:01.423
there are neurons like the neurodegenerative neurons that work in a paracrine

00:30:01.423 --> 00:30:02.943
manner in volume transmission.

00:30:04.143 --> 00:30:08.763
So sometimes relevant things happen globally, not just locally.

00:30:09.023 --> 00:30:12.603
And probably local control is different from global control.

00:30:12.703 --> 00:30:19.323
They both have different roles. So you're saying there's no evidence that astrocytes

00:30:19.323 --> 00:30:22.423
would have the specificity to have a very highly,

00:30:22.423 --> 00:30:27.843
highly temporally precise and locally specific or global….

00:30:44.803 --> 00:30:49.443
So the point is to look at what kind of global roles, synchronization,

00:30:50.653 --> 00:30:56.173
different neurons at the same time, are relevant for astrocytes to control.

00:30:57.293 --> 00:31:00.613
Now we talk about different neurotransmitters and presumptuosum.

00:31:01.193 --> 00:31:04.813
Are these only evantotropic receptors that astrocytes have?

00:31:05.013 --> 00:31:07.273
Or do you also see metabotropic receptors?

00:31:08.633 --> 00:31:12.093
Oh, they have metabotropic receptors for glutamate, for instance.

00:31:12.413 --> 00:31:15.353
They are typically present in astrocytes.

00:31:16.353 --> 00:31:20.073
Doesn't it make you suspicious that there is actually a real functional role for this?

00:31:20.073 --> 00:31:25.053
Because that would mean that you really actively, in a very specific way,

00:31:25.113 --> 00:31:31.373
are regulating processes in the astrocytes depending on the presence of glutamate. Oh, absolutely.

00:31:31.873 --> 00:31:35.193
So this would be also a little bit beyond just buffering, no?

00:31:35.393 --> 00:31:41.693
No, absolutely not. Yes, even transport of glutamate signals to some functions,

00:31:41.893 --> 00:31:43.513
even just the mere transport.

00:31:44.033 --> 00:31:48.033
So what kind of reactions do you see in astrocytes to glutamate?

00:31:48.973 --> 00:31:55.933
You may have changes, for instance, in the physical changes in the way the astrocytes wrap neuron.

00:31:57.133 --> 00:32:03.033
So the same way that spines change their morphology in order to get closer to

00:32:03.033 --> 00:32:08.933
our spine to retract, that has been described also in astrocytes in those distal

00:32:08.933 --> 00:32:11.093
processes in relationship with neuron.

00:32:11.313 --> 00:32:15.513
And this is regulated by glutamate and calcium. So you would argue that the

00:32:15.513 --> 00:32:22.373
glutamate also regulates issue on the formation of these synaptic structures

00:32:22.373 --> 00:32:27.773
that would then create post-synaptic neural processes plus gastrocytes.

00:32:27.793 --> 00:32:32.853
Yes, and an area of gastrocytes that I'm very interested in,

00:32:32.933 --> 00:32:36.073
and it hasn't been explored, are long-term changes.

00:32:36.393 --> 00:32:38.813
So we know that gastrocytes...

00:32:40.102 --> 00:32:46.182
Modulate memory, because when we manipulate the astrocytes with chemogenetics or transgenic tools,

00:32:46.562 --> 00:32:50.122
there are different stages that has been done with chemogenetics,

00:32:50.162 --> 00:32:55.462
and different stages in the memory paradigm from acquisition,

00:32:56.162 --> 00:33:03.402
consolidation, replay, so that exactly where they're acting is getting characterized.

00:33:04.422 --> 00:33:10.042
But we don't know exactly what are the long-term changes in astrocytes,

00:33:10.082 --> 00:33:16.142
the same way that neurons have long-term potentials, LTP, which causes changes

00:33:16.142 --> 00:33:18.482
in the molecular makeup of the neurons.

00:33:18.622 --> 00:33:21.542
That hasn't been characterized in astrocytes.

00:33:21.762 --> 00:33:25.322
And I do think this is an area that I'm interested, very interested,

00:33:25.522 --> 00:33:29.722
because I think that can be very fruitful to understand. To understand.

00:33:29.722 --> 00:33:32.662
Because for instance, there is this whole problem at the neuron side.

00:33:32.882 --> 00:33:36.522
If I build a synapse, how do I stabilize my synapse?

00:33:36.702 --> 00:33:41.662
And there is the idea that this also has to do with intercellular processes

00:33:41.662 --> 00:33:45.442
that depend on the molecules, like chemokinase 2, as an example,

00:33:45.542 --> 00:33:47.342
and John Lisman hypothesis.

00:33:48.382 --> 00:33:53.022
But that would suggest that you need similar kind of memory processes at the

00:33:53.022 --> 00:33:56.642
level of the astrocyte if they form part of that synapse.

00:33:56.782 --> 00:34:02.502
Yes. So, is there anything known about this kind of intercellular signaling

00:34:02.502 --> 00:34:03.902
pathways in estrocytes?

00:34:04.322 --> 00:34:09.222
No, not at all. Okay. That's a totally infant area of study.

00:34:09.282 --> 00:34:12.722
So, you're saying one of these 2,000 papers a year are all about, right? Yes.

00:34:12.922 --> 00:34:18.262
But then the other possible control that estrocytes have is capillaries.

00:34:18.542 --> 00:34:23.422
Absolutely. They can directly control blood flow and exchange,

00:34:23.822 --> 00:34:25.482
right? They do. The blood-brain barrier.

00:34:25.782 --> 00:34:33.062
They do. So how do you think that can be or is used and on what timescale is it used by astrocytes?

00:34:34.746 --> 00:34:39.326
I think, well, that's what the data shows in the time of milliseconds.

00:34:41.506 --> 00:34:47.666
Absolutely, that happens. So when a neuron in somatosensory stimulation,

00:34:48.046 --> 00:34:55.386
so a neuron in mice, the whiskers touch, the neuron detects that it has action

00:34:55.386 --> 00:34:57.966
potentials and calcium inactivation,

00:34:58.666 --> 00:35:01.886
then seconds later, milliseconds later, hundreds of milliseconds,

00:35:01.886 --> 00:35:07.106
seconds, astrocytes see it, and they transduce that to the vessel. That's a fact.

00:35:08.926 --> 00:35:13.986
And it's interesting when we talk about this this morning that they could do this earlier,

00:35:15.026 --> 00:35:20.666
before the task happens because there is a prediction that is going to happen

00:35:20.666 --> 00:35:26.406
and so it prepares the territory for a better response later on. But yeah, absolutely.

00:35:26.686 --> 00:35:31.226
The relationship of gastrocytes and vessels, for me, is very important. It's very, very clear.

00:35:31.886 --> 00:35:35.526
And I think astrocytes multiplex, so they can talk. And structurally,

00:35:35.606 --> 00:35:42.886
they are different because the distal processes are one thing,

00:35:42.886 --> 00:35:52.246
and the amphid is another morphological structure that is different and wraps the vessels.

00:35:52.406 --> 00:35:54.426
And it doesn't have the bulbous structure.

00:35:55.028 --> 00:35:59.988
Very dense network, this lattice in the distal process.

00:36:00.168 --> 00:36:03.028
It's just like a, yeah, it's like a sheet.

00:36:03.528 --> 00:36:09.548
Totally, it wraps the message. So the astrocytes have the multiplex.

00:36:09.888 --> 00:36:16.188
They're able to multitask, and one task is regulating the connection between

00:36:16.188 --> 00:36:19.328
vessels and the rest of the brain. Absolutely.

00:36:20.248 --> 00:36:27.688
What's the delay? So when we have neural activity, with what delay do the astrocytes

00:36:27.688 --> 00:36:30.428
start to then change blood vessel response?

00:36:30.988 --> 00:36:33.308
And what species is it in? Hundreds of milliseconds.

00:36:35.328 --> 00:36:38.288
And that's a long lasting? How long does this?

00:36:38.788 --> 00:36:42.048
I would say seconds. I'm not sure about that, but seconds. Okay,

00:36:42.188 --> 00:36:47.168
and then I would assume that this is not, if you have one spike,

00:36:47.388 --> 00:36:53.848
I don't think the astrocyte will start to change the dilation of the capillary.

00:36:54.248 --> 00:36:56.008
Oh, that you need to make a threshold?

00:36:56.868 --> 00:37:02.468
I probably, I don't know about it, but probably. So would you say then that

00:37:02.468 --> 00:37:08.388
the astrocyte looking at some average lower response and giving this average,

00:37:08.508 --> 00:37:09.908
it would control the capillary?

00:37:10.488 --> 00:37:14.748
That's a great question. I don't think it's none. Okay. Whether or not it responds

00:37:14.748 --> 00:37:17.908
to, because usually the experiments are more one-to-one.

00:37:19.268 --> 00:37:24.168
Bad balancing that has been analyzed, whether you need the minimum of neurons

00:37:24.168 --> 00:37:26.068
to be stimulated in order to produce.

00:37:26.548 --> 00:37:30.988
Now, one reason I'm asking is, of course, a lot of the work on magnetic resonance

00:37:30.988 --> 00:37:37.628
imaging and how you can use that to assess a function in the brain is by looking

00:37:37.628 --> 00:37:39.728
at this bold signal that's derived from blood flow.

00:37:39.908 --> 00:37:43.868
Right. And then there's a belief that this is related to neural activity.

00:37:45.099 --> 00:37:47.979
But now we see that this is actually mediated by astrocytes.

00:37:48.159 --> 00:37:49.579
Well, it could be mediated by astrocytes.

00:37:49.699 --> 00:37:55.339
The astrocytes consume oxygen, that's the point, but not all the time. So it's not that thing.

00:37:55.639 --> 00:38:00.719
Well, BOL signaling is actually very complicated because what it really shows

00:38:00.719 --> 00:38:03.719
is that it's hemodynamic response.

00:38:04.159 --> 00:38:08.919
So it is an increase in blood flow, but actually oxygen consumption doesn't increase.

00:38:09.259 --> 00:38:14.619
It's the paradox of BOL that was described several years ago.

00:38:15.099 --> 00:38:19.559
So meaning that oxygen consumption doesn't change very much.

00:38:19.699 --> 00:38:21.959
What changes is that there is more blood flow.

00:38:22.159 --> 00:38:27.779
And as a result, the ratio of hemoglobin that is oxidized versus the hemoglobin

00:38:27.779 --> 00:38:30.139
that is not oxidized changes. This is what you're looking.

00:38:30.339 --> 00:38:34.199
You're looking at increasing flow. You're looking at increasing oxygen consumption.

00:38:34.719 --> 00:38:40.679
Both is really difficult to interpret. But in general, it has been used as a

00:38:40.679 --> 00:38:42.639
surrogate for neuronal activation.

00:38:43.199 --> 00:38:47.639
And this is what I think is wrong. Not totally wrong, it's just neurons can

00:38:47.639 --> 00:38:50.299
contribute to the signal too. Not all the time.

00:38:51.039 --> 00:38:54.599
So my point is that we need to get into the details.

00:38:55.059 --> 00:39:00.939
I think the paradigm is very relevant because the notion that the task-dependent

00:39:00.939 --> 00:39:04.839
activity is the model of brain activity is wrong.

00:39:05.299 --> 00:39:07.139
So we need to look into different ways.

00:39:08.514 --> 00:39:14.714
Models, including that one of high glutamateric activation, but also brain,

00:39:14.874 --> 00:39:16.354
increasing brain activity.

00:39:16.734 --> 00:39:21.294
I'm sure that the both signaling increasing brain activity is very different.

00:39:21.494 --> 00:39:28.234
Gustavo Leco works on that, from task-dependent activity, changes during the

00:39:28.234 --> 00:39:30.794
day, changes in different nuclei.

00:39:30.914 --> 00:39:36.134
So we need to get out of average measurements and very specific models that

00:39:36.134 --> 00:39:42.054
we think are the models of all the reactivity because the devil's in the details here.

00:39:42.274 --> 00:39:48.574
And I'm sure that in very specific circuits and in very specific moments,

00:39:48.734 --> 00:39:50.674
astrocytes consume a lot of oxygen.

00:39:51.014 --> 00:39:55.614
So there are some papers and literature where people look at this relationship

00:39:55.614 --> 00:40:01.614
between reactivity, astrocyte response, and blood flow that are mainly done

00:40:01.614 --> 00:40:04.094
with optogenetics. And they're being questioned.

00:40:04.254 --> 00:40:07.974
Why don't they actually tell us Because whether they're formative or not,

00:40:08.094 --> 00:40:11.874
I almost understand people questioning because for many neuroscientists,

00:40:11.934 --> 00:40:16.274
they have a lot at stake to just believe in the dogma that, you know,

00:40:16.274 --> 00:40:17.874
the bold signal reflects neural activity.

00:40:18.154 --> 00:40:23.854
Right. But what is the truth of the matter in your view?

00:40:24.114 --> 00:40:29.074
Is it like the neural contribution will be like a tiny fraction of the bold

00:40:29.074 --> 00:40:33.614
signal, let's say 10%? or it will strongly vary on conditions.

00:40:34.754 --> 00:40:37.414
Because it's high variability, it's uninterpretable.

00:40:38.754 --> 00:40:43.054
So where do you... I think that the right measurements need to be done.

00:40:43.234 --> 00:40:45.874
I wouldn't be able to calculate.

00:40:47.454 --> 00:40:53.674
And this is like the traditional biophysics and the brain physiology.

00:40:53.754 --> 00:40:56.134
So we need to really calculate that.

00:40:56.514 --> 00:41:03.394
But you could also use... But I wouldn't dare to say that astrocytes use more

00:41:03.394 --> 00:41:05.554
oxygen than neurons or vice versa.

00:41:05.794 --> 00:41:12.494
So it is clear that action potentials are very demanding of energy.

00:41:12.754 --> 00:41:13.774
That's absolutely very clear.

00:41:14.114 --> 00:41:20.454
The needs, energetic needs of astrocytes are not known, for instance.

00:41:20.814 --> 00:41:26.694
We don't know. No, there are pathways and there are phenomena in astrocytes

00:41:26.694 --> 00:41:31.134
that require ATP, bermanic, oxytocin, calcium removal.

00:41:31.434 --> 00:41:34.254
So there are many pumps in the astrocytes.

00:41:35.054 --> 00:41:40.614
Perhaps we can say that they don't do as much ATP as neurons, but I wouldn't say.

00:41:41.544 --> 00:41:44.484
Dare to estimate it unless I do the calculations.

00:41:45.004 --> 00:41:50.304
So the energy requirements of the astrocytes are not known at all.

00:41:50.644 --> 00:41:54.524
Is it measurable for you? I think it could be measurable. Yeah,

00:41:54.524 --> 00:41:58.224
I think we just, it's like they have been measured in neurons. So why not?

00:41:58.564 --> 00:42:03.904
And the thing that we know now is that at least they are able to produce ATP

00:42:03.904 --> 00:42:08.064
through fatty acid oxidation that we think they do, which is very,

00:42:08.084 --> 00:42:10.404
it's very efficient because there's a lot of ATP.

00:42:10.404 --> 00:42:16.224
But we think that they don't produce, they don't have fatty oxidation all the time.

00:42:16.504 --> 00:42:21.284
We think they're versatile. The beauty of astrocytes, if I can introduce this

00:42:21.284 --> 00:42:26.464
word, beauty, in a scientific talk, is that they're versatile.

00:42:26.624 --> 00:42:29.824
So they use whatever they have, and they have glycolysis.

00:42:29.824 --> 00:42:35.884
And probably because the ATP produced by glycolysis is necessary for certain

00:42:35.884 --> 00:42:42.104
pathways that are locally by the membrane and related to the release of glutamate

00:42:42.104 --> 00:42:44.444
or to uptake of glutamate.

00:42:45.364 --> 00:42:51.824
But they do generate enough ATP to provide the other oxidation at some points in the life.

00:42:52.004 --> 00:42:57.544
I'm not sure when because we haven't carried out experiments in vivo, but I'm sure they do.

00:42:57.544 --> 00:43:03.524
And in certain moments, but I don't, I wouldn't dare to say whether that accounts

00:43:03.524 --> 00:43:07.584
for what percentage of ranoxetamines.

00:43:07.804 --> 00:43:12.524
Astrocytes can operate in an energetically efficient regime,

00:43:12.704 --> 00:43:15.824
but they can also be energetically inefficient.

00:43:17.625 --> 00:43:21.765
It can be both. You say at some point in their life. I think that they can use

00:43:21.765 --> 00:43:25.145
only glycolysis. And what happens there is they don't die.

00:43:25.465 --> 00:43:29.985
But one thing is not to die, and I think it's to be doing well. To be useful.

00:43:30.145 --> 00:43:35.285
To be performing right. One thing is survival, and I think it's the right performance.

00:43:35.745 --> 00:43:39.005
No, because again, we'd argue against this homeostasis interpretation,

00:43:39.305 --> 00:43:41.365
because if it was homeostasis,

00:43:41.885 --> 00:43:45.825
you would expect that substrate to be very energetically efficient,

00:43:46.085 --> 00:43:50.445
because otherwise, how can you be usefully contributing to maintaining homeostasis

00:43:50.445 --> 00:43:55.225
of your neurons if you burn a lot of energy yourself and produce waste products and so on?

00:43:55.425 --> 00:43:58.865
That's actually correct. Yeah, I think we need to understand the energy metabolism

00:43:58.865 --> 00:44:03.925
and astrocytes, not just when they use, what pathway, in what moment.

00:44:04.645 --> 00:44:09.465
Also, I think this notion that ATP is not stored perhaps is not true,

00:44:09.565 --> 00:44:13.585
so there is something that is very labile, you produce ATP and then either you

00:44:13.585 --> 00:44:15.245
don't use it or it's gone.

00:44:15.585 --> 00:44:23.285
So perhaps the astrocytes store ATP too. So that allows them to resort to glycolysis

00:44:23.285 --> 00:44:28.765
when neurons need a lot of oxygen, then within the oxygen goes to neurons, not to astrocytes.

00:44:28.885 --> 00:44:31.525
But the astrocytes are still there doing something.

00:44:31.685 --> 00:44:34.125
So perhaps they can use ATP from other sources.

00:44:34.765 --> 00:44:38.025
So there are questions there that need to be answered.

00:44:38.025 --> 00:44:46.885
And I do believe that energy is very, very important and that many of the anatomical

00:44:46.885 --> 00:44:54.225
designs in nature are determined by energy usage and energy usage optimization.

00:44:55.405 --> 00:45:01.065
So that could be probably a very fruitful area of research to try to understand

00:45:01.065 --> 00:45:06.845
how astrocytes, the interplay of energy metabolism between astrocytes and neurons,

00:45:06.845 --> 00:45:09.065
and how much oxygen each one uses.

00:45:09.725 --> 00:45:14.105
So is there any evidence that astrocytes can really, let's say,

00:45:14.125 --> 00:45:18.305
turn circuits on and off, and then it also happens under a realistic condition?

00:45:18.505 --> 00:45:24.505
I would, more experimentally, I have to find the papers. I would say yes.

00:45:24.765 --> 00:45:29.345
Probably if you ask this question to Alfonso Arago, all the leaders in the field,

00:45:29.445 --> 00:45:31.245
yeah, I think they can actually turn.

00:45:31.525 --> 00:45:37.045
So that means energy management. Well, yeah, because of the energy,

00:45:37.245 --> 00:45:43.705
but also because they stop releasing a given biotransmitter than neurons do.

00:45:44.385 --> 00:45:47.645
Yes, I think they can, yeah. Mm-hmm. Yes. Right, okay.

00:45:48.225 --> 00:45:52.425
Not just because of the energy, but because of the signaling. Right. Yeah.

00:45:53.607 --> 00:45:58.287
But the other thing that I always find very exciting about estrocytes,

00:45:58.467 --> 00:46:02.667
also because I'm ignorant, are the gap junctions, right?

00:46:02.767 --> 00:46:11.047
So they have gap junctions that allows potassium to disperse across membranes between estrocytes.

00:46:11.047 --> 00:46:19.007
And that would, in theory, set up possibilities for a more rapid signal transduction

00:46:19.007 --> 00:46:23.127
than what you would get if you have action potentials going over axons.

00:46:23.227 --> 00:46:24.027
Would you agree with that?

00:46:24.387 --> 00:46:27.207
I'm not sure about that because the unit is not.

00:46:27.207 --> 00:46:35.427
So the potassium that is diffused out of the astrocytes would have reached enough

00:46:35.427 --> 00:46:41.427
concentration to change the neuronal activity in other neurons that are farther away?

00:46:41.587 --> 00:46:44.567
No, let's just worry about the network of astrocytes and they're coupled with

00:46:44.567 --> 00:46:45.787
these gap junctions, right?

00:46:46.227 --> 00:46:51.367
Right, but in order to have an effect on neurons, you need to have really reached

00:46:51.367 --> 00:46:58.047
certain levels of potassium. So I tend to think about diffusion as not being

00:46:58.047 --> 00:47:00.827
very controlled, it's just passive, they are removed.

00:47:01.447 --> 00:47:04.567
It depends on the gradients, right? Right, exactly.

00:47:04.847 --> 00:47:11.707
That has to be modeled. So the answer could be… Yeah, but look,

00:47:11.787 --> 00:47:18.487
if neurons are reactive, they are transiently buffering a lot of potassium, right? Right.

00:47:18.507 --> 00:47:22.627
So that could possibly set up these kinds of gradients.

00:47:23.927 --> 00:47:28.927
No, because then you would have, so then the potassium in the exercise network

00:47:28.927 --> 00:47:33.607
would be rapidly attracted to these areas where there is activity.

00:47:34.287 --> 00:47:40.787
But a gradient, for me, is a process that is regulated in a fine manner,

00:47:40.867 --> 00:47:46.627
and this is incompatible for me with diffusion. Isn't that incompatible?

00:47:48.407 --> 00:47:54.927
So I tend to think that diffusion is just, yeah. But that could be modeled, I think.

00:47:55.007 --> 00:48:03.687
That idea could be put into a model and tested different levels of potassium

00:48:03.687 --> 00:48:08.247
production and potassium removal by estrogen to see that creates gradients.

00:48:08.367 --> 00:48:11.727
Yeah. And those gradients change neural activity. Yes.

00:48:12.027 --> 00:48:17.647
Maybe they, what you mentioned about the priming of neural activity.

00:48:17.647 --> 00:48:23.167
Maybe this is created by gradients of potassium around a given circuit that

00:48:23.167 --> 00:48:28.947
makes the neurons respond more or less, or beyond a threshold or not reaching the threshold.

00:48:29.247 --> 00:48:32.007
So maybe that's, but that can be modeled.

00:48:33.470 --> 00:48:36.950
Yeah, because what you would have, let's say you have neurons and become very

00:48:36.950 --> 00:48:41.710
active, so they start to buffer a lot of potassium rapidly.

00:48:42.390 --> 00:48:48.610
This might then deprive their neighbors from also becoming active because there's no potassium around.

00:48:49.070 --> 00:48:57.350
And then the astrocytes allow potassium to flow in from other parts of the network that are, let's say.

00:48:57.350 --> 00:49:00.870
So in some sense, it also will give you a competitive system,

00:49:01.110 --> 00:49:07.410
because if I have different islands of activity, via the exercise,

00:49:07.630 --> 00:49:10.750
I can actually compete over the potassium. I see. Right?

00:49:11.370 --> 00:49:16.750
And then this would play out very quickly if the gradients are big enough.

00:49:17.110 --> 00:49:20.810
So would you buy this kind of fantasy scenario, or do you think this is really

00:49:20.810 --> 00:49:25.250
completely… I would like to see the models.

00:49:25.730 --> 00:49:29.830
Okay. I would like to see the models that support that kind of scenario.

00:49:30.510 --> 00:49:35.210
Okay. Because the same thing might also be then going on with your sodium.

00:49:35.990 --> 00:49:39.090
Exactly. But the sodium doesn't diffuse over the gap junctions,

00:49:39.110 --> 00:49:40.750
right? That stays local in the astrocyte.

00:49:40.850 --> 00:49:44.550
Right. It's taken out of the astrocytes. It's true that it's taken out of the astrocytes, yes.

00:49:44.810 --> 00:49:48.430
So it's only the potassium that could do this. Yes. So the other thing I was

00:49:48.430 --> 00:49:52.190
thinking about that is exciting about these potassium networks

00:49:52.430 --> 00:50:02.070
or communication channels across the astrocyte network is it might give you

00:50:02.070 --> 00:50:05.910
a channel to start to prime different areas of the brain, right?

00:50:05.990 --> 00:50:09.450
So you mentioned this also earlier where you say, well, we expect a vision.

00:50:09.490 --> 00:50:13.350
And there are experiments that also have shown already that you have anticipatory

00:50:13.350 --> 00:50:17.790
changes in the volt signal that do not correlate to neural activity.

00:50:18.750 --> 00:50:21.670
And this would suggest that this is driven by the astrocytes.

00:50:22.210 --> 00:50:25.090
Then the question, how can the astrocytes know, right?

00:50:25.130 --> 00:50:28.490
So there must be some communication channel that says, look,

00:50:28.590 --> 00:50:31.590
we expect a visual stimulus, so increased blood flow.

00:50:31.790 --> 00:50:37.010
Well, but this is also the prediction brain, that the brain has already the

00:50:37.010 --> 00:50:41.530
information about the outer world and is renewing that information all the time.

00:50:41.970 --> 00:50:46.390
And whenever new information comes in, they just reset.

00:50:47.370 --> 00:50:52.870
And perhaps astrocytes are already in that circuit, meaning that they are actively

00:50:52.870 --> 00:50:55.670
processing information all the time.

00:50:55.790 --> 00:51:03.090
So premonition is part of the system that is replaying the information all the time.

00:51:04.500 --> 00:51:07.800
Don't you agree with that? Well, I was more thinking about that you just have,

00:51:07.840 --> 00:51:09.960
let's say, a conditional event that you're conditioned.

00:51:10.400 --> 00:51:13.700
When you hear a sound, there will be a visual stimulus five seconds later.

00:51:15.000 --> 00:51:18.980
So I hear the sound. And now in an anticipatory response to this association,

00:51:19.360 --> 00:51:23.340
I start to prime my visual area through the exercise to say,

00:51:23.440 --> 00:51:25.500
okay, we're going to get a visual stimulus, wake up.

00:51:26.120 --> 00:51:33.100
So you can sort of put areas in idle mode, like low energy consumptions,

00:51:33.100 --> 00:51:36.680
and then reactivate it from there by just three exercise network.

00:51:37.120 --> 00:51:41.900
It could be, yeah. Is there any data that would support that idea? No.

00:51:42.240 --> 00:51:47.580
Except this anticipatory bolt response innovation. Exactly. Was that Critchfield?

00:51:48.100 --> 00:51:55.140
I don't know. I don't remember. Yeah. Okay. So it could be one of the reasons for tylene.

00:51:55.880 --> 00:52:02.640
So that creates different islands of energy usage or potassium gradients along the brain.

00:52:02.760 --> 00:52:08.960
If I'm an estrocyte network, could I control the diffusion of the potassium?

00:52:09.160 --> 00:52:11.600
Can I open and close my gap junctions?

00:52:12.720 --> 00:52:15.680
Are they actively controlled or are they just passive? Nothing, they're passive.

00:52:17.180 --> 00:52:27.080
Yes. So now we made a little index of different possible functions of the estrocytes.

00:52:27.960 --> 00:52:29.900
And you also mentioned this glymphatic system.

00:52:31.200 --> 00:52:34.920
And that, of course, pushes us a little bit more in the direction of homeostasis.

00:52:35.680 --> 00:52:41.060
I've seen that's homeostasis, clearly. So why was this glymphatic system relevant

00:52:41.060 --> 00:52:46.600
in this discussion? Why do you think this is an important insight in terms of

00:52:46.600 --> 00:52:48.760
what astrocytes could be doing?

00:52:49.000 --> 00:52:56.220
Well, because it's, I think, a very important issue as to how the brain removes waste.

00:52:57.911 --> 00:53:05.071
And this is highly, highly relevant for normal function. It is highly relevant for diseases.

00:53:06.331 --> 00:53:10.171
So I think it is worth mentioning. And it's a very recent discovery,

00:53:10.291 --> 00:53:11.171
so it's worth mentioning.

00:53:11.951 --> 00:53:16.891
And moreover, it is probably regulated by the brain itself.

00:53:17.691 --> 00:53:22.211
So it's not really clear at this point, but the locus coeruleus and others,

00:53:23.131 --> 00:53:30.831
some of these nuclei that are globally broadcast, broadcasting throughout the

00:53:30.831 --> 00:53:33.271
brain, probably regulate that too.

00:53:34.271 --> 00:53:37.791
And there are studies showing that this can regulate also.

00:53:38.051 --> 00:53:43.131
You spoke about gradients of potassium determining circuit activation.

00:53:43.951 --> 00:53:54.291
There are studies showing that these fluxes inside the brain control,

00:53:54.871 --> 00:54:00.891
the glucose availability, the availability of nutrients.

00:54:01.271 --> 00:54:04.631
So they're not trivial. They're not just things that are there.

00:54:04.711 --> 00:54:06.371
They're very relevant to brain

00:54:06.371 --> 00:54:10.291
physiology. So the astrocytes are sort of regulating blood-brain barrier.

00:54:10.491 --> 00:54:14.771
This sets up a convective flow of what?

00:54:15.651 --> 00:54:16.651
Of liquid.

00:54:18.351 --> 00:54:23.571
It's liquid that moves from the artery side to the venous side.

00:54:23.931 --> 00:54:25.931
It's just inter-solar space.

00:54:26.471 --> 00:54:31.791
What's the liquid? It's a solutes, salts probably. It's convection.

00:54:32.871 --> 00:54:36.231
It's pure convection. And it's probably water.

00:54:37.451 --> 00:54:41.011
So it's really like you're washing out the brain. You're washing out the brain.

00:54:41.031 --> 00:54:42.371
It's going to flow towards.

00:54:42.971 --> 00:54:47.831
It flows to the venous side that eventually goes to the lymphatic system. Uh-huh.

00:54:49.242 --> 00:54:52.762
It's actually beautiful physiology.

00:54:53.142 --> 00:54:58.282
I'm very fond of myself in addition to computation, to brain physiology studies.

00:54:59.822 --> 00:55:04.642
And these studies are actually very elegant because it's actually very simple.

00:55:04.922 --> 00:55:09.222
Technically, they use fluorescent tracers and they just put them on one side

00:55:09.222 --> 00:55:11.502
and see where they can find them.

00:55:11.602 --> 00:55:15.942
And then they find their way out of the brain and they can track their pathway. way.

00:55:16.062 --> 00:55:24.282
So they absolutely there is water, the brain is clean every day with a lot of

00:55:24.282 --> 00:55:30.642
things get out and that is mediated by the astrocytes in the sense that they are a wall.

00:55:30.822 --> 00:55:36.622
At least they're a wall that those things have to cross and that wall probably

00:55:36.622 --> 00:55:42.422
is also selectively regulating the passage and when that wall is not working

00:55:42.422 --> 00:55:44.882
well then that passage doesn't occur.

00:55:44.882 --> 00:55:49.622
And that changes also during day and night. I find it fascinating.

00:55:50.122 --> 00:55:57.502
The finding that this clearance works only at night is fantastic,

00:55:57.762 --> 00:56:03.062
meaning that the intracellular volume of the brain changes during the day and

00:56:03.062 --> 00:56:04.142
during the night. Right.

00:56:04.282 --> 00:56:09.902
So, it's not sure if the estrocytes shrink or something, and that changes the

00:56:09.902 --> 00:56:15.042
convection towards the venous side of the brain.

00:56:15.202 --> 00:56:17.562
Well, distances are covered in that way, you think.

00:56:18.102 --> 00:56:21.082
I'm not sure. So, what's the density of this venous density?

00:56:21.162 --> 00:56:25.322
I'm not sure, but probably that answer can be found in the literature,

00:56:25.622 --> 00:56:27.262
yeah. But it's fantastic.

00:56:27.722 --> 00:56:31.702
But it also means that there must be different chambers along the artery,

00:56:31.862 --> 00:56:36.262
no? Because you must accumulate this fluid that's being released.

00:56:37.494 --> 00:56:41.434
It must be extracted from the blood, stored, and then released.

00:56:41.654 --> 00:56:47.254
Well, it comes from the CSF. The CSF is in contact with that liquid.

00:56:47.614 --> 00:56:51.494
But that's all physiology is well known. It comes from the CSF.

00:56:51.514 --> 00:56:57.294
The CSF is circulating all the time and it's in contact with the intracellular space.

00:56:58.014 --> 00:57:03.194
Probably they have different contents, but there is a flow. There is a flow

00:57:03.194 --> 00:57:10.634
from the putter in between the arteries and the astrocytes that goes in a different

00:57:10.634 --> 00:57:12.834
direction from the blood. This is interesting also.

00:57:13.074 --> 00:57:16.394
So the blood goes in one direction and that flow goes in the opposite direction.

00:57:16.774 --> 00:57:21.134
And eventually, because of these convection forces, crosses,

00:57:21.494 --> 00:57:23.674
the brain came out to the venous side.

00:57:23.834 --> 00:57:26.534
But the astrocytes must be controlling both walls.

00:57:26.954 --> 00:57:31.794
Yeah, they do, absolutely, yeah. So they must be controlling the artery wall. Absolutely.

00:57:32.054 --> 00:57:37.834
And then this CSF chamber. Yes, absolutely. Okay.

00:57:39.074 --> 00:57:42.734
So do you know when in sleep to sleep? In which sleep stage?

00:57:43.714 --> 00:57:50.094
I'm not sure. I'm sure they have looked at it, but I don't have the data.

00:57:50.354 --> 00:57:56.394
But as it occurs, the paper has very beautiful pictures and it's very traumatic.

00:57:56.394 --> 00:58:02.974
The way that those fluorescent tracers go through the brain at night,

00:58:03.134 --> 00:58:07.234
but they don't go through the brain during the day in mice.

00:58:07.294 --> 00:58:13.794
It's really, really traumatic, the change. But so now at the more behavioral level,

00:58:13.934 --> 00:58:20.614
you highlighted the number of correlations we are aware of with respect to exercise

00:58:20.614 --> 00:58:25.254
activity and behavior, like in sensor processing,

00:58:25.674 --> 00:58:29.534
state switching, and fear response regulation. Yes.

00:58:30.374 --> 00:58:38.874
So, which of these results for you are most informative about the possible functional

00:58:38.874 --> 00:58:40.794
role of astrocytes? Our brain states.

00:58:41.034 --> 00:58:44.494
I think network activity regulation.

00:58:44.834 --> 00:58:51.474
Because of this global notion, I think the findings in the local circuits are

00:58:51.474 --> 00:58:54.094
just a part of a larger role.

00:58:54.354 --> 00:58:59.734
So they find them because they look. But I think the astrocytes are more important

00:58:59.734 --> 00:59:02.294
in the regulation of networks.

00:59:02.394 --> 00:59:08.554
So they are more relevant to these brain state transitions. That's my thing.

00:59:08.934 --> 00:59:15.594
If I would study something, I would study the role of astrocytes in neuromodulation.

00:59:15.834 --> 00:59:20.594
So global phenomena, neuromodulation, brain state, slow oscillations,

00:59:20.614 --> 00:59:24.354
that would be, for me, the… So in the brain state example,

00:59:24.934 --> 00:59:36.494
you use optogenetics to stimulate… Astrocytes, Kira post-cancer, the paper.

00:59:36.494 --> 00:59:37.494
Yeah, right, exactly.

00:59:37.814 --> 00:59:43.714
And Raphael, Justin, and Kira. Right, exactly. But you're stimulating in the… which area are we?

00:59:43.754 --> 00:59:46.514
Is it the somatosensory cortex, or was it the barrel cortex?

00:59:47.074 --> 00:59:51.254
Yeah, I think it's the somatosensory cortex. Yeah, it's the cortex.

00:59:51.674 --> 00:59:53.694
So we're measuring LFPs.

00:59:56.490 --> 01:00:00.750
What we then see is that after a five-second stimulation.

01:00:01.690 --> 01:00:10.090
You essentially see a peak at about 100 seconds later of really low frequency,

01:00:10.210 --> 01:00:15.270
like below, let's say here, half to two hertz or a million delta range, right?

01:00:15.430 --> 01:00:21.310
Right. Which is a frequency range that under the control condition does not appear, right?

01:00:21.370 --> 01:00:24.310
So this here. Right, there's a change in the power.

01:00:25.070 --> 01:00:30.490
So… at the low frequency, that is increasing, stimulating. Like there's actually

01:00:30.490 --> 01:00:31.930
no power in a control condition.

01:00:32.290 --> 01:00:38.310
Exactly. And it jumps up when you stimulate the astrocytes with optogenetics. Exactly.

01:00:38.550 --> 01:00:41.770
But now, isn't that a surprise? That's no surprise.

01:00:42.030 --> 01:00:46.270
Because if you drive the astrocytes, you stimulate the astrocytes for that long,

01:00:46.490 --> 01:00:49.770
you might lead to, let's say, pathological change in the local.

01:00:50.170 --> 01:00:53.070
But not quite pathological. I wouldn't say pathological.

01:00:53.650 --> 01:01:01.770
It's not known. Well, low delta would correlate with really deep sleep or deep anesthesia, right?

01:01:01.870 --> 01:01:07.170
So those would mean these neurons are deprived of just any ion to do anything.

01:01:07.390 --> 01:01:09.430
That's such a good thing. So essentially shutting down.

01:01:09.850 --> 01:01:15.290
So this is like you have just shut down the neurons by depleting them of any resource.

01:01:15.730 --> 01:01:17.490
What would be the right control to use?

01:01:18.410 --> 01:01:23.110
Well if you want to see state change you want to look at oscillatory frequencies

01:01:23.110 --> 01:01:27.270
that you know have functional relevance right so you want to look at.

01:01:28.270 --> 01:01:32.870
Frequencies that's either in gamma or in theta or alpha but certainly above

01:01:32.870 --> 01:01:37.790
delta because delta would reflect deep sleep so my question is can we speak

01:01:37.790 --> 01:01:44.030
of a state that is functionally relevant so you have to look at a power of a

01:01:44.030 --> 01:01:48.010
frequency range that you know has functional relevance Right.

01:01:48.630 --> 01:01:54.370
This finding has been reproduced, actually, it is an early study.

01:01:57.444 --> 01:02:02.744
People have shown that if we remove the capacity of astrocytes to respond to

01:02:02.744 --> 01:02:06.544
the nucleus basalis, the nucleus basalis are minor, they also regulate brain state.

01:02:07.024 --> 01:02:09.784
And this has been measured with local field potentials.

01:02:10.424 --> 01:02:14.604
But that's very different. But the result is very similar.

01:02:14.664 --> 01:02:19.824
In this case, they are not stimulating us with optogenetics because optogenetics

01:02:19.824 --> 01:02:25.124
didn't exist. I think we're talking about a study from 2000-something.

01:02:25.184 --> 01:02:26.584
I don't remember the year.

01:02:26.844 --> 01:02:31.804
But the result is similar in the sense that the astrocytes don't have calcium responses.

01:02:32.364 --> 01:02:38.604
The nucleus basalis of minor doesn't regulate the brain states.

01:02:39.584 --> 01:02:43.004
The nucleus basalis of minor would be the cholinergic projection.

01:02:43.004 --> 01:02:49.884
Right, but meaning, talking about, so in that case, we cannot talk about astrocytes

01:02:49.884 --> 01:02:56.624
stimulation depleting neurons from things they need, from shedding, because it's actually,

01:02:57.084 --> 01:02:59.864
they aren't actually, it's just the opposite experiments.

01:03:00.324 --> 01:03:03.324
Astrocytes are not being stimulated. They just don't respond.

01:03:03.604 --> 01:03:07.024
Therefore, they cannot be depleting neurons from anything.

01:03:07.264 --> 01:03:10.684
And it's the same result in the sense that they regulate brain states.

01:03:10.684 --> 01:03:16.864
But now, in this optogenetic experiment, what does it mean to activate an astrocyte?

01:03:17.344 --> 01:03:20.904
Because they're genetically manipulated to respond to light,

01:03:21.084 --> 01:03:24.164
right? That's a key question.

01:03:24.704 --> 01:03:27.364
How are they coupled? To which receptors are they coupled? No, it's not known.

01:03:27.924 --> 01:03:34.964
That's a key question. That's the finding. They express Whereas our carotidopsin,

01:03:35.164 --> 01:03:39.384
there, they stimulate with light, and they see the calcium increases.

01:03:39.924 --> 01:03:43.884
But how it happens… The calcium inside the astrocyte increases?

01:03:44.184 --> 01:03:48.184
The calcium transients inside the astrocyte increase, and the only thing that

01:03:48.184 --> 01:03:55.004
they know is that their carotidopsin hyperpolarizes cells and the neurons that

01:03:55.004 --> 01:03:57.664
triggers the neurons are inhibited.

01:03:57.664 --> 01:04:03.004
But in astrocytes, there are also voltage-dependent channels,

01:04:03.204 --> 01:04:04.884
but not as active as in neurons.

01:04:05.084 --> 01:04:11.024
So probably the astrocytes are detecting the hydropolarization somehow,

01:04:11.184 --> 01:04:14.064
and that triggers somehow an

01:04:14.064 --> 01:04:19.464
increase in calcium concentration in the saddle. But it's not known how.

01:04:19.864 --> 01:04:23.584
First thing. Second thing, it happens mostly in the processes.

01:04:23.584 --> 01:04:27.904
So, it's a very important thing.

01:04:29.105 --> 01:04:34.585
Whether the calcium increases are happening in processes,

01:04:34.825 --> 01:04:39.405
in the distal processes, or in the primary process, or in the soma,

01:04:39.485 --> 01:04:45.105
because the mechanisms that regulate calcium release are very different in different compartments.

01:04:46.805 --> 01:04:53.305
And interestingly enough, with our pyrrhodopsin, they happen mostly in the processes,

01:04:53.585 --> 01:04:59.285
indicating that those voltage-dependent channels are perhaps richer there, but it's not known.

01:04:59.525 --> 01:05:04.605
And I think this is a limitation. It's very interesting because it's actually

01:05:04.605 --> 01:05:10.445
interesting to be able to stimulate as-processed-reverence-in-assault type in

01:05:10.445 --> 01:05:15.545
a time-specific manner and not using transgenic mice that are,

01:05:16.365 --> 01:05:22.165
it's a really elegant tool but the fact that we don't know exactly how it works

01:05:22.165 --> 01:05:27.505
I think is a limitation because we don't know the physiological relevance of the response.

01:05:27.825 --> 01:05:30.525
It is a limitation. This is a bit problematic then, right?

01:05:30.585 --> 01:05:35.905
Because we do we can speak of a state change but we don't know whether we have

01:05:35.905 --> 01:05:38.365
induced some pathological state of the tissue.

01:05:38.545 --> 01:05:41.965
I don't know. It's also sort of not sufficiently controlled.

01:05:42.345 --> 01:05:45.785
It's probably you need to refine the technique, understand mechanisms better,

01:05:45.985 --> 01:05:49.325
and find the same things in other approaches.

01:05:49.825 --> 01:05:54.065
So we're still in the dark then about where in the dark is the glass, right?

01:05:55.145 --> 01:06:00.565
But now in your case you are also interested in finding well,

01:06:00.645 --> 01:06:02.345
there's another problem we see here.

01:06:02.425 --> 01:06:06.105
Everything is so much tied to calcium imaging, right? It's too much.

01:06:07.665 --> 01:06:12.725
Probably yes and this of course is also biasing very much how we think about

01:06:12.725 --> 01:06:14.305
because we think oh these guys are really slow,

01:06:14.925 --> 01:06:20.765
right but this might be due to the way we observe them so what do you expect

01:06:20.765 --> 01:06:23.905
if we would have the magic imaging techniques,

01:06:24.665 --> 01:06:27.865
or measurement techniques to some kind of.

01:06:28.625 --> 01:06:33.625
Electrophysiologically like techniques do you expect them to be as fast as neurons

01:06:33.625 --> 01:06:37.745
in terms of time constants at work or do you really Would your prediction be

01:06:37.745 --> 01:06:41.925
that they are a system that are slower than neurons? I would say they're slower.

01:06:42.285 --> 01:06:46.425
Perhaps they are faster than seconds.

01:06:46.545 --> 01:06:52.045
So maybe we can move now the time scale from hundreds of milliseconds to tens of milliseconds.

01:06:52.725 --> 01:06:55.385
But still, that's slower than neurons. Yes.

01:06:56.094 --> 01:07:03.374
Right. I would say they're slower. And my point is that we have to find the

01:07:03.374 --> 01:07:07.294
moments where the brain needs that time scale.

01:07:07.534 --> 01:07:11.914
So we don't need to make astrocytes faster than they are. That's a mistake.

01:07:12.394 --> 01:07:16.494
So the field has been trying for a very long time to make astrocytes look like

01:07:16.494 --> 01:07:17.574
neurons. So that's a mistake.

01:07:17.914 --> 01:07:22.194
We just have to find what they are. And once we characterize them without forcing

01:07:22.194 --> 01:07:29.594
the interpretations, try to find out where that time scale is relevant for.

01:07:29.854 --> 01:07:34.754
Is it relevant to any phenomenon in the brain? I think this is the way to go.

01:07:34.934 --> 01:07:41.154
So these These very ultra-fast representations of neurons that last a few milliseconds,

01:07:41.834 --> 01:07:46.734
that astrocytes have nothing to do with it, obviously, because that's beyond their time scope.

01:07:46.814 --> 01:07:52.714
But other phenomena in the brain that happen there, where the astrocytes can

01:07:52.714 --> 01:07:56.314
be useful, that's, I think, the way to go. Okay.

01:07:56.894 --> 01:08:02.074
So now, could you imagine astrocytes as a substrate of memory?

01:08:02.074 --> 01:08:08.154
Like, his memory is to carry information over time, and you do that by stable structural change.

01:08:09.694 --> 01:08:17.294
Do you think exercise has the properties, the basic, also, biomechanical and

01:08:17.294 --> 01:08:22.714
the biochemical tools to form memories and store information over time?

01:08:22.874 --> 01:08:27.234
I think it's a biophysical question whether they have something that can have

01:08:27.234 --> 01:08:32.114
many different positions to store information. So, that's the question,

01:08:32.174 --> 01:08:37.134
and the question is, can different calcium transients do that?

01:08:37.834 --> 01:08:41.554
We don't know. We have to ask that question and analyze it.

01:08:42.174 --> 01:08:45.994
Because the question is, of course, then, if they provide context in which neurons

01:08:45.994 --> 01:08:50.594
operate, this context might have to also be associative. Right.

01:08:52.434 --> 01:08:58.014
The question is whether this context is just supportive, metabolically supportive,

01:08:58.414 --> 01:09:04.034
or more or structurally supported, or it's really an intelligent support and

01:09:04.034 --> 01:09:07.014
it's doing something for neurons. Right.

01:09:09.670 --> 01:09:16.190
There are the questions of what are the variables that neurons encode is not

01:09:16.190 --> 01:09:23.570
totally clarified because they know that some variables like odor or sounds or colors.

01:09:23.690 --> 01:09:29.530
I know that the neurons encode those variables and in their presentations that

01:09:29.530 --> 01:09:33.810
are happening in the time scale of milliseconds. seconds, but there are other

01:09:33.810 --> 01:09:36.630
variables than ever, for instance.

01:09:36.850 --> 01:09:40.870
There are many other variables that the brain uses, and I don't think we have

01:09:40.870 --> 01:09:45.410
discovered them, and perhaps some of those variables are bolder,

01:09:45.410 --> 01:09:49.970
and that is the role I'm afraid to precise.

01:09:51.210 --> 01:09:56.170
But they're very important too, like associations, contests,

01:09:56.170 --> 01:09:59.430
providing averages, perhaps they do that.

01:09:59.430 --> 01:10:04.390
They carry out the statistics of the brain in this model of prediction coding,

01:10:04.730 --> 01:10:09.130
where the brain is so tied in doing the statistics about how their predictions

01:10:09.130 --> 01:10:12.650
are confirmed by experience. But that's a lot of error calculation.

01:10:13.850 --> 01:10:21.010
And some of the papers that describe that, some of the parameters around the error calculation...

01:10:22.327 --> 01:10:25.967
Placing it in a timescale of seconds, actually. Some of them are really fast,

01:10:26.187 --> 01:10:30.667
and some of them are slower, the ones that require more computation.

01:10:31.567 --> 01:10:35.087
And that perhaps is this kind of buffering. Well, probably.

01:10:35.467 --> 01:10:38.347
It's perhaps what the kind of things that astrocytes do.

01:10:38.547 --> 01:10:44.147
So they're not hunting the code for concepts or for colors or for others.

01:10:44.187 --> 01:10:50.767
I think those neurons do that, and that is related to the lack of specialization

01:10:50.767 --> 01:10:52.887
of astrocytes. So neurons are highly specialized.

01:10:53.227 --> 01:10:58.627
And for me, that means that they encode for many different things, plays and music.

01:10:58.967 --> 01:11:02.447
I don't think astrocytes do that. I think astrocytes do general things,

01:11:02.747 --> 01:11:10.127
general computations, providing general context, net time scale of seconds.

01:11:10.487 --> 01:11:13.887
So the question is what variables are encoding.

01:11:14.367 --> 01:11:18.587
That's the question. But I think there are different variables that are broader,

01:11:18.587 --> 01:11:21.807
there, but those variables are probably important, too.

01:11:23.447 --> 01:11:26.347
One way to get a handle on that is, of course, you could look at the different

01:11:26.347 --> 01:11:29.147
pathologies where astrocytes are involved.

01:11:29.887 --> 01:11:33.987
So what are the outstanding pathologies? They're involved in all pathologies.

01:11:34.447 --> 01:11:38.607
Any brain pathology has,

01:11:40.827 --> 01:11:46.847
sick astrocytes. The question there, so they're morphologically changed and molecularly changed.

01:11:46.887 --> 01:11:50.447
The question is, what is the contribution of that astrocyte to the pathology.

01:11:50.587 --> 01:11:54.547
How many diseases are there that specifically attack the astrocytes?

01:11:54.547 --> 01:11:57.007
There are Alexander's disease.

01:11:57.147 --> 01:12:01.567
There are some diseases where GFAP has mutations and that causes changes.

01:12:02.067 --> 01:12:05.947
So what are the specific phenotypes you get? They are retarded.

01:12:06.667 --> 01:12:11.667
The children that have a problem with the blood-brain barrier.

01:12:12.067 --> 01:12:17.687
Yes. But are there any diseases that specifically attack the astrocytes later in life?

01:12:18.547 --> 01:12:22.047
It's not known. It's not known because.

01:12:26.792 --> 01:12:32.092
Astrocytes also get older, and probably they are dysfunctional,

01:12:32.092 --> 01:12:36.332
but we don't know, and they are affected by disease.

01:12:36.492 --> 01:12:41.392
The question is whether this is an epiphenomenon, whether it's just the main

01:12:41.392 --> 01:12:44.992
cause of the disease, or it's just a lateral thing.

01:12:45.372 --> 01:12:49.252
That's the question, but absolutely, they don't work that way.

01:12:49.312 --> 01:12:53.292
But you are interested in being therapeutics. Yes, because we have theta,

01:12:53.532 --> 01:12:57.232
and we are very driven by theta.

01:12:57.412 --> 01:13:02.412
Then when we manipulate astrocytes, we manage to recover brain function very effectively.

01:13:02.992 --> 01:13:06.792
And that's in Alzheimer's disease. In traumatic brain injury and Alzheimer's

01:13:06.792 --> 01:13:11.192
disease, we are still trying to understand the molecular makeup of the astrocyte

01:13:11.192 --> 01:13:16.252
change. because the problem is that we cannot suffer.

01:13:16.492 --> 01:13:22.072
We don't have astrocytes isolated from human brains for obvious reasons.

01:13:22.712 --> 01:13:29.552
So we are trying to understand what happens to astrocytes in Alzheimer's disease.

01:13:30.112 --> 01:13:34.472
Happens means all these homeostatic functions, metabolic functions, are they changed?

01:13:34.632 --> 01:13:40.012
And if they are, how does it contribute to disease progression or not?

01:13:40.012 --> 01:13:44.292
Doesn't make any difference whether we treat them more or not.

01:13:44.492 --> 01:13:50.052
So my understanding is that we should go for combination therapies and focusing

01:13:50.052 --> 01:13:51.572
on the neurons is a big mistake.

01:13:52.782 --> 01:13:58.602
Because going back to the vascular system, even if you have the finest treatment

01:13:58.602 --> 01:14:05.722
for neurons in order to help them make these amazing spines, if the blood system,

01:14:06.042 --> 01:14:09.302
the vascular system is not working, the treatment is not going to work.

01:14:09.642 --> 01:14:17.002
So it's a very obvious thing, but the field is not realizing that some very

01:14:17.002 --> 01:14:19.722
obvious things are not being done.

01:14:19.722 --> 01:14:24.902
And for me, recovering the function of astrocytes, it's as important as recovering

01:14:24.902 --> 01:14:29.142
the function of blood vessels, you know, to cure any disease.

01:14:29.602 --> 01:14:34.042
And in addition, perhaps then we can target the neurons more specifically.

01:14:34.342 --> 01:14:41.222
But I think recovering astrocytes will have a major impact on disease recovery

01:14:41.222 --> 01:14:44.262
because of all the many things they do. Right.

01:14:44.902 --> 01:14:50.502
So that's a fantastic outlook, right? So in your career, you have,

01:14:50.542 --> 01:14:52.922
in some sense, you're in this minority field, right?

01:14:53.002 --> 01:14:59.642
That is not necessarily in the limelight of neuroscience, but that also,

01:14:59.702 --> 01:15:02.362
as you show now, there's a long application, as you can all promise.

01:15:02.542 --> 01:15:07.742
And in some sense, it is a scientific field per se, in the sense that there's

01:15:07.742 --> 01:15:09.702
still so much to discover, right?

01:15:10.022 --> 01:15:14.462
It is. And we're really just at the beginning of understanding this whole system.

01:15:14.462 --> 01:15:18.362
So you're, you're now pursuing this for quite a while.

01:15:18.762 --> 01:15:23.742
And so if others would like to follow you, understanding the brain from that

01:15:23.742 --> 01:15:27.642
perspective, what is Elena's law that we shoot at here too, what is.

01:15:28.296 --> 01:15:32.416
Just consider astrocytes and neurons

01:15:32.416 --> 01:15:36.056
and microdegenerative endocytes and vascular cells at the same time.

01:15:36.936 --> 01:15:38.916
So my point is like, yeah, absolutely.

01:15:39.796 --> 01:15:45.016
So getting astrocytes with neurons, absolutely.

01:15:45.236 --> 01:15:50.336
Stop standing cells in isolated systems. Inclusive.

01:15:51.096 --> 01:15:57.336
Always try to be as complex as possible. So we need to have simple models because

01:15:57.336 --> 01:15:59.516
otherwise we cannot really study them.

01:15:59.636 --> 01:16:04.336
But they need to be complicated enough in order to be meaningful for anything.

01:16:04.656 --> 01:16:10.136
So that's the point. So with astrocytes, we need to upgrade astrocytes and study circuits.

01:16:10.396 --> 01:16:18.336
We have to find the minimum circuit that is relevant for astrocyte functions

01:16:18.336 --> 01:16:22.336
and then analyze how they interact with neurons.

01:16:22.336 --> 01:16:28.596
So that's the key idea is stop you're interested in astrocytes stop studying

01:16:28.596 --> 01:16:32.316
astrocytes only get you know collaborate.

01:16:33.096 --> 01:16:38.056
With your own people and get outside and study astrocytes in addition to other

01:16:38.056 --> 01:16:42.196
subtypes stop the xenophobia oh absolutely yeah that's,

01:16:42.716 --> 01:16:48.336
well yeah that's to me otherwise we won't cure diseases absolutely unless we

01:16:48.336 --> 01:16:52.796
have an internal view of the brain that is based upon on circuits,

01:16:52.996 --> 01:16:55.276
for me, that's very important idea.

01:16:55.396 --> 01:16:57.576
Well, that's an interesting point, right, that you make here with,

01:16:57.616 --> 01:17:01.556
you see a lot of also big pharmaceutical companies have actually withdrawn from

01:17:01.556 --> 01:17:04.056
the brain because everything they're through, and in fact- Well,

01:17:04.056 --> 01:17:08.656
the failure, the failure has been- You would say that's because of this neurocentric perspective.

01:17:09.096 --> 01:17:13.976
Well, it's, well, it's neurocentric perspective is the not giving the drug in

01:17:13.976 --> 01:17:17.236
the right moment, depending on the disease, in the case of Alzheimer's disease,

01:17:17.456 --> 01:17:19.356
it's just all the treatments have been very late.

01:17:20.168 --> 01:17:22.708
When the brain is not destroying us, it's not doing that much.

01:17:23.128 --> 01:17:27.548
But yeah, I think we need to understand that the brain are circuits and we need

01:17:27.548 --> 01:17:32.768
to understand that and target and restore circuit function.

01:17:33.088 --> 01:17:37.528
To me, this is the key. It's not just restoring one neuron function or one astrocyte,

01:17:37.628 --> 01:17:42.468
restoring the function circuits and that's going to be more effective on the

01:17:42.468 --> 01:17:47.368
recovery of a given disease or in the prevention of a given disease because

01:17:47.368 --> 01:17:50.848
these circuits are impaired They are very early in disease.

01:17:51.128 --> 01:17:55.348
They're actually an excellent diagnostic marker for early alterations.

01:17:55.628 --> 01:18:01.748
And to me, that's the focus, the therapeutic focus and the focus for the basic

01:18:01.748 --> 01:18:06.888
research from the astrocyte point of view, from the neuron and from microarray

01:18:06.888 --> 01:18:09.168
point of view. And NG2, NG2 cells are fantastic.

01:18:09.748 --> 01:18:13.788
We don't really know. And they do have spikes, actually. So this classification

01:18:13.788 --> 01:18:17.228
of neurons and the spike and cells that don't spike,

01:18:17.388 --> 01:18:25.008
NG2 have spikes because they are like in the middle and they're fantastic too. So we.

01:18:25.668 --> 01:18:31.008
And you mentioned now at the end of our discussion. Right. But because I'm, I'm interested. So the.

01:18:31.508 --> 01:18:33.828
Are they as voluminous as ester sites? How?

01:18:34.408 --> 01:18:38.388
Are they as voluminous as ester sites? They're very big. They're too very big.

01:18:38.428 --> 01:18:40.988
And they share some developmental origin.

01:18:41.228 --> 01:18:44.728
And they talk to the ester sites as well. It's not known. Okay. It's not known.

01:18:45.748 --> 01:18:49.008
But I think we have to have comparison. Why are they not differentiated from

01:18:49.008 --> 01:18:50.148
neurons if they also spike?

01:18:50.868 --> 01:18:55.408
Why were they... Because they don't... They are not as sophisticated as neurons.

01:18:55.668 --> 01:18:59.588
They don't have actions, for instance. They don't transmit information.

01:18:59.788 --> 01:19:02.308
They don't have, like, long distance.

01:19:03.128 --> 01:19:06.808
Okay, but... They don't have spice, I think. Perhaps they do,

01:19:06.868 --> 01:19:08.068
because... Does it form synapses?

01:19:09.828 --> 01:19:17.528
I think there's a paper talking about the NG2 neuronal synapse. I'm not sure about that.

01:19:17.748 --> 01:19:22.508
Okay, that's your next talk. But yeah, the point for me is that we focus on

01:19:22.508 --> 01:19:25.668
our self-type because we need to focus on something.

01:19:25.788 --> 01:19:31.008
But don't lose the perspective that you're looking at a small thing and that

01:19:31.008 --> 01:19:34.048
if perhaps you're missing the connections with other elements,

01:19:34.208 --> 01:19:37.388
you're missing the whole point. That's my point.

01:19:38.948 --> 01:19:44.108
That's a point also well taken. And it's like embrace sort of this inclusive

01:19:44.108 --> 01:19:47.308
perspective. It's not only about all endurance, only exercise,

01:19:47.368 --> 01:19:53.348
also look at this whole variable collection of cells that form circuits that we call the brain.

01:19:53.508 --> 01:19:58.868
Also, moreover, I talked in the talk, the.

01:20:00.885 --> 01:20:05.905
Like, neurons are highly diverse. So even saying, maybe neuron is about work, too.

01:20:06.605 --> 01:20:11.525
Perhaps I'm going too far. But neuron is about work, because if the mouse,

01:20:11.605 --> 01:20:15.645
they discover in the somatosensory cortex 30 types of different neurons.

01:20:16.425 --> 01:20:20.985
Just there. So imagine in the whole brain, I have thousands of different types

01:20:20.985 --> 01:20:23.265
of neurons. So neuron is about turn, too.

01:20:24.005 --> 01:20:27.625
And if we don't understand that they're very different in different circuits,

01:20:27.905 --> 01:20:29.065
they may share some roles.

01:20:29.065 --> 01:20:34.465
But then now the difference may be really very relevant in the way they interact

01:20:34.465 --> 01:20:39.285
with other cells types and the way they encode and the way they store information

01:20:39.285 --> 01:20:41.005
that perhaps that's a problem.

01:20:41.165 --> 01:20:43.305
So let's stop saying neurons.

01:20:45.265 --> 01:20:49.905
So four years from now, I'll come visit you here in Barcelona in your lab.

01:20:50.045 --> 01:20:54.685
And we're going to check whether the prediction you made today was falsified or confirmed.

01:20:55.085 --> 01:20:59.765
What prediction? That's the question. What's the prediction that is most critical

01:20:59.765 --> 01:21:03.245
to your research that you would like to see analyzed?

01:21:03.325 --> 01:21:11.385
That astrocytes encode specific variables in information processing.

01:21:12.625 --> 01:21:16.445
That's my, that they encode. That's the prediction.

01:21:17.205 --> 01:21:22.425
The answer may be that they don't. There may be some variables that are relevant

01:21:22.425 --> 01:21:27.305
to higher brain functions are encoded by estrocytes.

01:21:27.565 --> 01:21:31.465
Okay. Very good. That's my... Galena, thank you very much for this conversation.

01:21:31.985 --> 01:21:32.825
Thank you. Thank you very much.

01:21:35.225 --> 01:21:40.905
The CSN podcast was produced by the Convergent Science Network of Biometrics

01:21:40.905 --> 01:21:47.385
and Biohybrid Systems, a project funded by the European Sevens Research Framework Program.

01:21:48.865 --> 01:21:54.145
For more interviews, recorded lectures, or upcoming conferences in the field

01:21:54.145 --> 01:22:00.385
of biometrics and biohybrid systems, go to csnnetwork.eu.

01:22:00.720 --> 01:22:08.560
Music.