Latest / EHA Unplugged / How Blood Cells Age: From Stem Cells to Clones (with Dr Michael Milsom)

Transcript

How Blood Cells Age: From Stem Cells to Clones (with Dr Michael Milsom) · EHA Unplugged

0:00
I don't think aging is one thing.
0:02
I think aging is multiple different things.
0:04
And therefore I think it's highly likely that there are
0:07
different mechanisms that are driving the evolution of this
0:10
phenomena. So actually my own pets favorite
0:13
theory for what drives aging is that I'm a big fan of this
0:17
entropy model of aging. And basically what this relates
0:20
to is this idea that a cell or a system goes from an ordered to a
0:25
disordered state. And we believe that one of the
0:28
major drivers of this is cell division.
0:31
Welcome to EHA on Plot, the official podcast channel of the
0:34
European Hematology Association EHA.
0:37
I'm your host Isabella Rivera. Today we're joined by Doctor
0:42
Michael Nielsen. Dr. Nielsen is a division head
0:45
at the German Cancer Research Center and group leader at the
0:50
Heidelberg Institute for Stem Cell Technology and Experimental
0:54
Medicine in Germany. Doctor Mason is one of the
0:56
organizers of EHS Research Conference in 2026, Recon 2026.
1:02
And as an appetizer today, we're going to have a discussion about
1:07
hematopoiesis and how it changes during aging.
1:10
Thank you for being with us today.
1:11
Thank. You for having me?
1:13
So to start off, what do you mean by the traditional HSC
1:16
centric view of hematopoietic aging?
1:20
Why do you think everything is centre in HSC?
1:23
Well, when we think about how the hematopoietic system changes
1:26
and evolves during the process of aging, it's a really
1:29
attractive concept that the hematopoietic stem cells, these
1:33
cells that we think last throughout your lifetime and
1:36
they're essential for the ongoing production of blood
1:40
cells and responding to stress when you get depleted of mature
1:43
cells that these alterations that would be permanent and
1:47
progressive should be enacted at the level of hematopoietic stem
1:51
cells. And there are various kind of
1:54
conditions where we know that the hematopoietic stem cells,
1:58
their functionality is altered often detrimentally during
2:01
aging. So it kind of makes sense
2:05
logically that these permanent changes would be as a result of
2:09
the stem cells getting functionally compromised during
2:12
the ageing process. What is the main evidence that
2:15
supports this? CSC centric.
2:18
So I think that there are many lines of evidence that show that
2:21
hematopoietic stem cells lose their functional potency during
2:25
the process of ageing both in humans and in experimental model
2:30
systems like the mouse. So for example, in humans, if
2:33
you think back to criteria where we look to employee in terms of
2:38
looking at donors for stem cell transplant, we're typically not
2:43
using older donors in that setting because there's a higher
2:46
proportion of a graft failure in that setting.
2:49
And that's thought to result from this loss of functional
2:51
potency of the hematopoietic stem cells.
2:54
You can see this really nicely in in the mouse setting as well
2:57
and where paradoxically the number of immunophenotypic stem
3:01
cells increases with time. However, the number of
3:04
functional stem cells declines with time.
3:08
There also were the lines of evidence, and I think we'll talk
3:10
a little bit later about clonal hematopoiesis, but I think this
3:14
is another kind of excellent line of evidence that something
3:17
happens at the stem cells level where there's a clonal
3:20
contraction in the stem cell pool that now impacts on blood
3:24
cell production and the evolution of a wide range of age
3:27
associated diseases. So I think there's lots of
3:30
evidence that the stem cells alter with aging, but where I
3:34
think there's a big gap in knowledge is exactly how that
3:37
might impact on a lot of the clinically relevant features of
3:41
aging, the predisposed towards disease.
3:44
So for example, immune dysfunction, is that really
3:47
driven by altered biology of the hematopoietic stem cells or
3:51
other cell types, Anemia of aging, is that really driven by
3:55
alterations in the stem cells progenitor cell compartment or
3:59
something further downstream? And I think that that's
4:02
something that's very difficult to address mechanistically, but
4:05
I think it's a really important question.
4:08
But I think up until now we focused a lot on the
4:11
hematopoietic stem cell compartment, so.
4:13
You argue that this view is too narrow to focus on HSCS is too
4:18
narrow. What are we missing?
4:20
If we focus only on HSCS, what else is there?
4:24
OK. So I think that before I jump
4:27
into why we should look outside HSCS, the hematopoietic stem
4:31
cell compartment, it's probably important to kind of take a step
4:34
back and think about broadly what is aging and what does that
4:38
look like in different individuals.
4:41
And often times the way that we study this, we'll kind of use
4:45
something like an experimental mouse cohort, young and aged,
4:50
and basically look at the things that change on average in those
4:53
populations in terms of the composition of the blood, the
4:56
bone marrow, the functionality. And clearly one of the the
5:01
problems with that is that aging isn't just a kind of a process
5:07
where you go from A to B directly.
5:09
I think that actually aging is going to multiple different
5:13
destinations and potentially also at different rates.
5:17
So if you think about humans, not everybody looks the same in
5:20
terms of how they elicit age associated phenotypes and not
5:24
everybody develops exactly the same kind of malignant or non
5:29
malignant age associated hematopoietic pathologies and
5:33
people develop those at different time points.
5:36
I think historically we've written that off as alterations
5:40
in the environment or genetic predisposition.
5:43
But I can tell you in our experimental mice that are
5:45
inbred or mice look very different in terms of their
5:49
hematopoietic composition, old age.
5:52
So I think that it's kind of problematic to think of aging as
5:56
a single thing, like we, for example, think about acute
6:01
myeloid leukemia. In reality, that's a spectrum of
6:05
different diseases with different drivers that have
6:08
different aggressiveness and I think aging is the same.
6:13
Now if we kind of pivot back in a very long winded way back to
6:17
the question about hematopoietic stem cells and what we're
6:20
missing, I think that if we just focus in on that population, we
6:25
probably miss a lot of important information.
6:27
So I think it's possible that a lot of the phenotypes that
6:31
develop during aging aren't necessarily as a result of the
6:35
hematopoietic stem cells being altered.
6:37
So if our focus is predominantly on the stem cells, we missed
6:40
that and we're looking in the wrong place for a mechanism
6:43
that's driving the phenotype. Now I think there are certain
6:46
things that are driven predominantly by hematopoietic
6:49
stem cell dysfunction. So I think in a range of
6:52
different hematologic malignancies, it's quite clear
6:56
that the hematopoietic stem cell was likely the cell of origin
6:59
that became dysfunctional and then later on evolved into a
7:03
full blown leukemia. So I'm kind of proposing a more
7:07
holistic view where we maybe take a step back, don't just
7:10
focus on one cell compartment. And actually, even if we're
7:14
interested in hematopoietic stem cells, try to really assess
7:18
whether in the more mature cell types or for example, in the
7:21
niche, whether this is the thing that's really driving the
7:24
disease phenotype. So the more mature immune cells
7:30
show ageing phenotypes and that are independent of age.
7:34
Yeah, I think that there are a number of clear examples of
7:38
that. So if you think about defects in
7:40
the adaptive immune system that evolve during aging, these are a
7:45
major cause of a range of age associated predispositions to
7:50
developing disease. So if you look at for example,
7:54
by people generally have a poor response to viral infection like
7:59
we saw during the COVID pandemic or how people respond really
8:02
poorly to influenza infection. These are generally elicited
8:07
through defects in in T cell production and potentially also
8:10
B cell production and functionality.
8:13
And I think it's highly likely that the defects there are not
8:16
related to hematopoietic stem cells.
8:20
I think that also if we think about some age associated
8:23
hematologic malignancies that relate more to the lymphoid
8:27
branch of hematopoiesis, they're the likely cell of origin
8:31
probably isn't a hematopoietic stem cell.
8:33
So these are kind of examples of things that have evolved
8:36
independent of the hematopoietic system.
8:39
So actually we've been carrying out a number of studies in the
8:44
laboratory mouse where it's relatively straightforward to
8:48
measure multiple different hematopoietic metrics from the
8:51
same mouse. So for example, we would
8:54
transplant hematopoietic stem cells, but also in the same
8:57
individual, look at anemia, look at the level of peripheral blood
9:02
cells, look at the, the, the, the gene expression programs,
9:07
the different cell types. And the amazing thing to me is
9:10
that actually we see huge diversity there, but we see a
9:13
very poor correlation between stem cell dysfunction in that
9:18
individual as measured by what the stem cell does when we
9:21
transplant it into a recipient and the actual heap
9:26
hematopoietic phenotypes in the primary animal.
9:29
That doesn't seem to be a good correlation which would be
9:31
predicted in the literature. And in fact, after we've
9:35
transplanted the stem cells and these cells have regenerated
9:37
hematopoiesis in the recipient animal, we see a poor
9:41
correlation in blood cell parameters between the donor and
9:44
the recipient, which you wouldn't predict if the stem
9:48
cells were actually driving what was happening in the primary
9:50
mouse. Why is is there this difference?
9:53
What else is driving aging in this mice?
9:56
Yeah, So this is the $1,000,000 question.
9:57
What's the mechanistic driver? So again, I'd I'd revert back to
10:01
one of my original comments that I don't think aging is one
10:05
thing. I think aging is multiple
10:07
different things. And therefore I think it's
10:09
highly likely that there are different mechanisms that are
10:12
driving the evolution of this phenomena.
10:15
So we can see alterations in niche, niche composition.
10:20
This is also something that we've studied and found a kind
10:23
of lack of correlation between this and what's happening in
10:26
terms of the blood cell compartmentalization.
10:29
And, and actually at the moment, we're really struggling to find
10:33
anything that correlates particularly well with the
10:36
phenotypes that we're observing in the mice, even when we look
10:39
at things that have been previously published in the
10:41
literature. So actually my own pets favorite
10:45
theory for for what drives aging is that I'm a big fan of this
10:49
entropy model of aging. And basically what this relates
10:53
to is this idea that a cell or a system goes from an ordered to a
10:58
disordered state. And we believe that one of the
11:02
major drivers of this is cell division.
11:05
So for example, on a very basic level, we know that when you
11:10
replicate your genome, it's very challenging to replicate it
11:13
perfectly. I think the same thing holds
11:15
true, for example, when we try to replicate the epigenome
11:19
status of a daughter cell compared to the parent cell.
11:23
So you end up with a diversification and the more
11:26
cell divisions, the more you differ from the original cell.
11:30
So this we think is potentially A mechanism through which
11:33
hematopoietic stem cells achieve diversification.
11:38
And this diversification can also drive an evolutionary
11:41
process. So this kind of creates a system
11:44
of clonal competition, outgrowth of clones with, for example,
11:48
favorable growth characteristics and then potentially also
11:53
disordered functionality afterwards.
11:56
So this is attractive to me, but that's extremely difficult to
11:59
study because most of our computational analysis methods
12:03
tend to focus on what's coherently changing rather than
12:07
to what degree do we see diversification.
12:10
So your work has shown that inflammatory stress, even early
12:15
in life can accelerate HSE ageing.
12:17
What can you tell us about this? Yeah.
12:20
So we were interested in inflammation and aging for quite
12:23
a long period of time. So actually I used to work in
12:26
the the field of inherited bone marrow failure, specifically
12:29
looking at Franconi anemia. And it was very clear from the
12:32
literature that there was this link between inflammation and
12:36
bone marrow failure in that setting.
12:39
And so this kind of piqued my interest and I was working on
12:41
this in the setting of Franconi anemia.
12:43
And it really was kind of one of these Eureka moments when I
12:48
started working in Heidelberg and saw the work of Andreas
12:51
Trump, who showed that these pro inflammatory cytokines can push
12:55
stem cells from this long term quiescent or dormant state into
13:01
active cell cycle. So now this idea that cell
13:05
division is something that drives loss of functionality, it
13:08
kind of made sense to me that potentially the inflammation was
13:12
making the stem cells divide. And the more inflammation you
13:15
would be exposed to, the more those cells would diversify away
13:19
from the original cells that you'd got in there.
13:23
So in that context, we start to carry out studies.
13:26
We had some very elegant mouse models that have been developed
13:29
by Andreas Trump and and her and Marika S, as a former postdoc in
13:33
his laboratory, is now also a group leader and works in the
13:36
next office to me. And these mouse models allowed
13:38
you in vivo to pull apart the stem cells that didn't divide so
13:42
much with the passage of time. You can even see dormant cells
13:45
that barely divide during the lifetime of a mouse.
13:48
And actually, interestingly, these cells retain a young
13:51
phenotype. And then the stem cells that
13:54
have divided more, and now you can expose to inflammation.
13:58
And the interesting thing is that even though all these stem
14:01
cells are exposed to inflammation, it's the ones that
14:04
divide that lose the functional potency.
14:07
And the ones that can still stay dormant, they still retain their
14:12
functional potency. So importantly, this kind of
14:17
brings into question this concept of self renewal, this
14:20
idea that a stem cell can make a perfect stem cell door to cell.
14:25
And I think one of the take home messages from our work is stem
14:29
cells can divide to make more stem cells, but very, very
14:33
rarely do those stem cells have the same functional potency as
14:37
the parent. So you end up in a situation
14:40
where each round of inflammation draws some of the stem cell pool
14:44
into cycle, it loses functional potency and it never gets
14:49
regained by those cells. So it's irreversible.
14:54
And what that means is that now later on week, a month, a year
14:59
later, if you have another, another stimulus that draws more
15:02
of the stem cell pool into cycle and they lose their functional
15:06
potency, you can now have an additive or cumulative depletion
15:11
of the stem cell pool. And this to me is a really
15:13
attractive mechanism to a kind of explain clonal succession
15:18
with aging, but also clonal contraction during aging.
15:22
But this is happening in the HSC cells, so it's been driven by
15:26
the environment of an inflammation.
15:29
How much of this increased vulnerability comes from the?
15:33
Yeah. So what I was talking about
15:35
there was really with the focus on the hematopoietic stem cell
15:38
compartment. So trying to enumerate the
15:41
number of hematopoietic stem cells actually in these mice
15:46
that we've exposed to multiple rounds of inflammation where
15:50
we've depleted the stem cell pool.
15:52
Actually, hematopoiesis looks relatively normal in that
15:55
setting. And I think that's also what you
15:57
would predict based on what we know of how things work in
16:00
humans. Yeah, so we do.
16:01
We are exposed to multiple rounds of infection, sometimes
16:05
pretty severe infection, and we always kind of bounce back to
16:09
these levels where we're able to maintain homeostatic production
16:13
of blood cells. And interestingly, there are
16:16
studies of very old humans, so centenarians who've had the a
16:20
massive contraction of their stem cell pool.
16:23
Maybe they've got one or two stem cells that are supporting
16:25
hematopoiesis, but they're still able to maintain normal levels
16:29
of peripheral blood cells. And I think that's kind of an
16:32
important concept because although the stem cell pool
16:35
contracts, we still have a system that that has evolved in
16:39
order to be able to sustain blood production during life.
16:42
So these mice generally their blood cell systems look pretty
16:46
normal. I think that what's interesting
16:49
is when you start to dial in things like these mutations that
16:52
you see in in clonal hematopoiesis.
16:55
So for example, the group of Grand Chalon and Catherine King
16:58
have done very elegant studies looking at DNMT 3, a loss of
17:01
function and then driving, adding on top of that bacterial
17:05
infection. And you can see this expansion
17:07
of the mutant clones in that particular setting.
17:10
And in the mice, I don't think that there's a strong phenotype
17:15
illustrated at the level that the mature blood cells.
17:17
Yeah, and maybe that relates to, you know, missing ingredients in
17:22
terms of our mouse models that are important in humans.
17:25
So most of the time our mice are kept in specified pathogen free
17:29
conditions where they're not readily changed, challenged by
17:33
pathogens. Maybe if we depleted the bloods,
17:35
the stem cell pool sufficiently, then added on top some kind of
17:40
infection, perhaps those mice would have a shorter lifespan or
17:44
not be able to respond so much to those challenges.
17:47
But yeah, in, in, in that setting, again, we've been very
17:50
much focused on the stem cell pool rather than any age
17:54
associated phenotypes in terms of mature blood cells.
17:57
You mentioned clonal hematopoiesis.
18:00
How does inflammation influence the clonal expansion?
18:05
Yeah. Again, I think that that's a
18:06
really interesting and important question.
18:10
I, I think that we're only now starting to delve into that to
18:14
some extent. So I mentioned before the work
18:16
of people like Grant Challon and Catherine King.
18:19
I think there's also been very exciting data recently from the
18:22
likes of her Parish Bias and John Dick looking at human
18:26
samples, where indeed it does look like exposures to
18:30
inflammatory stimuli and maybe infectious stimuli provide a
18:35
selection advantage for for example DNMT 3A mutant clones.
18:41
So I think it isn't 100% clear why that is the case, but I
18:46
think it data does seem to indicate that with each exposure
18:50
to inflammation and infection, you program the epigenome
18:56
somewhat along the lines of trained immunity where if you
19:01
execute a specific transcriptional program over and
19:04
over again, there's a residual memory of this in the stem
19:08
cells. And I guess the idea is that if
19:11
you now go in again with the same kind of stimulus like an
19:14
infection that the cells will be able to more rapidly respond or
19:19
be able to respond with a a greater magnitude effect that
19:23
same stimulus or something similar.
19:25
Or perhaps there's a little bit of a cost towards this because
19:28
hematopoietic stem cells have to be good all rounders that are
19:31
able to produce multiple different cell types and carry
19:33
different functions. So maybe if you program them in
19:37
a particular direction, then maybe they don't perform so well
19:42
as a stem cell. Now if you look at the DNNT 3A
19:45
setting, I think what's clear is that now these cells have less
19:51
of this priming effect when you give the inflammatory stemics.
19:54
So they have less of a epigenetic memory of that
19:58
exposure. And it's tempting to speculate
20:01
that that might actually facilitate these cells still
20:05
maintaining their stem cell attributes.
20:08
And maybe being able to be resistant had this inflammatory
20:12
challenge. So I think it's a little bit
20:14
different the the idea that, for example, a stem cell would get a
20:20
pro proliferative mutation and now this mutation would give a
20:23
positive selection advantage and allow it to expand more rapidly
20:28
than the rest of the stem cell pool.
20:30
But rather this idea that there's a constant ongoing
20:33
attrition of the stem cell compartment and maybe things
20:37
like the DNMT 3A mutations, T mutations might make the cells
20:42
more resistant to that. So they're fitter in that
20:46
context. So do you think any of these
20:49
features can be targeted for delaying or mitigating?
20:53
Aging, yeah, potentially. I mean, I think that I'm not a
20:58
big fan of this idea of rejuvenation.
21:04
I mean, it would be nice, yeah. But I think it's kind of hard to
21:07
think about exactly how that would work.
21:10
So even if you look in these extreme settings where people
21:13
formally demonstrate this. So I'm talking about things like
21:16
if you put reprogramming factors into cells that you've taken
21:20
from an elderly Organism, or actually if you generate A
21:23
transgenic mouse where you can transiently induce the
21:26
reprogramming factors. And now you partially reset the
21:31
epigenome and the mice by and large look younger and actually
21:36
have improved health metrics. So first of all, it it's
21:39
obviously very difficult to think about exactly how you're
21:42
going to facilitate that in the setting of a human.
21:45
But even when we start to think about fundamental mechanisms
21:49
that are taking place there, I don't think you're completely
21:52
resetting the clock on the changes that occurred during
21:55
aging. So obvious example, the somatic
21:58
mutations that have been acquired, they're still going to
22:01
be there. So I refer back now to the
22:05
comment that I made towards the beginning of the podcast where I
22:08
was talking about aging probably being multiple different things
22:12
and multiple different destinations.
22:15
Now if you ask me, do I think you can correct some of these
22:18
clinically relevant phenotypes, the answer to that will be yes.
22:23
And I think that it's not going to be some kind of panacea for
22:27
all aging phenotypes, but maybe in a more targeted fashion we
22:31
could think about addressing things like immune dysfunction,
22:34
so compromised adaptive immune system.
22:37
There are some of these hand hematopoietic rejuvenation
22:41
strategies that have been put out there.
22:43
So things like dietary restriction would be one of
22:45
them. Mtor inhibition, this idea of
22:48
depleting myeloid biased aged stem cells.
22:51
These are all things that have been put forward.
22:54
I think it'll be interesting to see how robustly, robustly those
22:57
kind of findings can be recapitulated across broad
23:00
different animal models. I think that they're clearly
23:04
benefits to things like dietary restriction, but rejuvenation?
23:08
I'm not so sure that that's a reality, but I will be very
23:12
happy to be proved wrong. Is there anything else that you
23:16
would like to add for the audience?
23:17
So I think that one of the things that's important if you
23:20
want to do research in this area is, is to really kind of take
23:26
take notice of the literature that's out there, but really
23:29
kind of take a step back and think about broadly, if we're
23:32
looking at aging, if we if we kind of think about what we
23:37
know, what's obvious about kind of human aging now that people
23:41
develop very kind of vastly different phenotypes at
23:45
different rates. That's kind of an important
23:48
thing, I think. And to purely study aging in the
23:52
context of old versus young. But this idea that we'll be able
23:57
to identify mechanisms that define what happens to every
24:01
individual, that seems kind of unlikely.
24:04
Maybe we need to carry out the entire analysis and design our
24:07
experiments in in slightly different ways to try and
24:11
somehow understand this diversification process.
24:15
Because I think if we can try to understand the basic mechanism
24:20
that underlies this collapse of the system, then we can actually
24:25
do meaningful research to think about whether it is possible to
24:29
delay that or reverse some of the features, or whether that
24:33
doesn't make any sense at all. And we need to think about
24:36
things like cell replacement therapies.
24:38
Is the opportunity to keep the system functioning for as long
24:42
as we need it to? So if I understand well, there's
24:46
this diversity that we knew already.
24:48
We know that, you know, during a lifetime very different things
24:52
happen, one person compared to another and that causes aging.
24:56
But you're saying that there's also this huge diversity in the
24:59
internal regular processes in the cells.
25:04
So the diversity comes from everywhere in the aging process,
25:08
not only from outside of the body, but also from our
25:11
biological processes. They're different from one
25:13
person to the other. So it's really hard to tackle.
25:17
Yeah, process. Although I mean, one of the
25:19
things that we often kind of debate within the group is we're
25:22
spending a lot of time and effort and trying to
25:25
characterize this stochasticity in terms of the diversification
25:31
of the function of the hematopoietic system.
25:33
And there I think the, the really exciting thing is not
25:36
just kind of characterize the level of an entire system, but
25:39
be able to dial into individual cells and clonal systems that
25:44
occur within the hematopoietic compartment to try to understand
25:48
what's happening there. And so, yes, that there is this
25:53
kind of really complex pattern of diversification, but the
25:57
debate that we have a lot in the laboratory is whether some kind
26:01
of extrinsic pressure can actually drive you towards a
26:05
more coherent system or whether it will actually, just as you
26:10
were saying, make the diversification even broader.
26:13
So I think that this is kind of next set of experiments that
26:16
we'd like to approach. But yeah, the problem is
26:21
studying aging. It takes a long time.
26:23
Well, thank you very much for this fascinating conversation.
26:27
Thank you for being with us today.
26:29
Thank you, I've enjoyed it. Thank you to the audience for
26:33
listening. If you enjoyed this episode,
26:35
please like it and subscribe to our channel.
26:38
Stay tuned for more episodes of EHM Blog.