Hematopoietic Stem/Progenitor Cell Fate Specification in Health and Disease

September 25, 2020

Information

Dr. Diane Krause
Professor of Laboratory Medicine, Pathology and Cell Biology, Yale University
Blood and Bone Webinar
March 23, 2020

ID5677

To CiteDCA Citation Guide

00:00OK, so welcome. I'm going to be
00:03talking about hematopoietic stem and
00:05progenitor cell fate specification
00:06in health predominantly and disease
00:08and I just loved megakaryocytes,
00:11so I have a picture of a gorgeous
00:13megakaryocyte here you can see all of
00:16the nuclei inside this megakaryocyte.
00:18This megakaryocyte became polyploid
00:19in vitro, so the nuclei,
00:21or separate and abstained the cytoskeleton.
00:24The actin and tubulin and you can
00:26just see how gorgeous these cells are.
00:30So I have to be able to go forward.
00:33There we go.
00:34So this is just the hematopoietic tree
00:36for anybody who is new to him out of
00:39police is these are hematopoietic stem cells.
00:42They divide and differentiate,
00:44become hematopoetic progenitor cells and
00:46then the classic tree shows that they
00:48split into common lymphoid progenitors,
00:50which will make the B&T cells and NK cells
00:52and Common myeloid progenitor cells,
00:55amongst which then you get me
00:57P Mega Rich Royd Projectors,
00:59which is what I'll be focusing on today.
01:02Since it's bright orange
01:03granulocyte monocyte for janitors,
01:04the maker annulus sites and monocytes
01:07and these mega rich right for janitors
01:09that make the megakaryocytes and
01:11are it's red cells and we produce
01:13about 2 * 10 to the 11th.
01:15New platelets and red blood cells daily,
01:18and if you calculate that down it's
01:20about 2 million platelets in 2
01:22million red cells at every second.
01:24So these MVP are very,
01:26very busy and we'd like to better
01:28understand how the MVP Self Renew
01:30and choose a fate to become.
01:32With Droid Versus Megakaryocytic.
01:35This image just shows that there
01:37are over 14 million platelet
01:38transfusions worldwide annually,
01:40which we would like to be able
01:42to decrease the need for donors.
01:45An if there were ways of making
01:47platelets in vitro that would be
01:49great and this is from an excellent
01:52video that might all talk will
01:54be available on YouTube later.
01:56Jonathan has a video of Platelet formation
01:58and this these images are from that.
02:01It's spectacular.
02:02I highly recommend it.
02:03OK,
02:03so why do we study the megakaryocyte?
02:06Erythroid progenitor or me P,
02:08which makes Mega Mega Carey site
02:11progenitors which make megakaryocytes
02:13and platelets and erythroid
02:14progenitors that make red blood cells.
02:17So we studied this for lots of
02:19reasons and you guys could probably
02:21come up with additional good ones.
02:23But first of all it just by
02:25learning how the MVP makes it date,
02:27it's fate.
02:28Decisions were learning really about
02:29cell fate decisions in general,
02:31basics of stem cell biology,
02:32what makes us Ella stem cell?
02:34What makes a cell at progenitor cell,
02:36and how does it make its fate?
02:39Decisions for regenerative medicine purposes,
02:40it's good to understand how this
02:42actually occurs in vivo,
02:43so that we can improve our
02:45approaches for making platelets
02:47and red blood cells in vitro.
02:48For patients,
02:49and also because understanding this
02:52entire lineages and how it's regulated
02:55will lead to the discovery of potential
02:58targets for disorders of the Mega.
03:00Carey's itinerary thread lineages.
03:02So how can one distinguish an
03:04Emmy PA megathread progenitor
03:06cell functionally in FINA?
03:07Typically this is a challenge that
03:10was initially taken up by Ched Sonata
03:12when he was a postdoc in my lab,
03:14and then followed up by
03:16Juliana Javie Ferruccio,
03:17who then finish the work and what
03:19they did is they figured out a way
03:21where you start with a single cell.
03:24You hope it's a mega rich Rd progenitor,
03:26and if you keep that cell in
03:29culture for about 2 weeks,
03:30it will form a colony of cells where
03:33the only cells in that colony.
03:35Are megakaryocytes and red blood cells,
03:37so this is this was the goal and this is
03:40what Chad and you figured out how to do.
03:43So basically what they would
03:45do is fax sort candidate MPs,
03:47megathread progenitor cells and
03:48put them into semisolid media.
03:50Very, very seriously.
03:51So none of the two colonies would
03:53be bumping up against each other
03:56and then waited 13 to 15 days and
03:58at the end of that period they
04:00stained with antibodies.
04:01They staying with an antibody
04:03against Glycophorin A,
04:04which is a red blood cell antigen.
04:06And again CD41A,
04:07which is on Maggie Carey sites and
04:09you can see we have some colonies
04:12that are erythroid only and some
04:14colonies that are Mega Carey site
04:15only and then we have some colonies
04:18that have both colors in them and
04:20these are colonies that we can see.
04:22If you mke because they have both
04:25megakaryocytes and erythrocytes.
04:26So using this functional readout
04:28for a by potent progenitor chat and
04:31then shoe went through Facs sorting
04:33protocols to try to identify the best
04:36enrichment they could for primary
04:38human megathread progenitors and
04:40what they discovered is if you gate
04:43within the 34 positive Lynn negative 135.
04:45That's all three receptor negative CD,
04:4845 RA negative population.
04:49If you look for Maple,
04:51which is the receptor for thrombo
04:53poet and the vast majority of
04:56these cells are Maple positive.
04:58It's not limited to megakaryocytes,
05:00we know he mad about extending
05:02projector cells expressed nipple,
05:03but then there was also this nipple
05:06low population and if we look at that
05:09and compared to excuse me CD 38 which
05:11I'm sorry that should be labeled better.
05:14What we can find is that the
05:16erythroid progenitors tend to have
05:18less nipple and more CD 38 so
05:20if we enrich for this population
05:22they grow erythroid only colonies.
05:24If you then take this population which
05:27is really in the middle for 3038,
05:29it's not the most negative and
05:31it's not the most positive.
05:33And then you get on CD 41.
05:35You can find CD 41 positive Meg
05:38progenitors that make Mega Carey
05:39site only colonies and these MVP,
05:41which made colonies that have
05:44both Meg Anorith Royd cells.
05:46When they enriched for these populations,
05:48this is what the colonies look like,
05:51and you'll see a lot of these
05:53graphs in my talk,
05:55so I'm going to go through this slowly.
05:57What I'm showing you here on
05:59the Y axis is
06:00the number of colonies per 100 cells plated,
06:03and what I'm showing you in blue is the
06:06number of those colonies that had both
06:08megakaryocytes and erythrocytes in them,
06:10so a by potent colony and you can see
06:13about half of the colonies that grow.
06:16Have megakaryocytes and erythrocytes and E.
06:18Only colony has just erythrocytes and a Meg.
06:21Only colony has just megakaryocytes.
06:23We also enriched as I said for Meg
06:26Progenitors which are in the CD 41
06:29positive population and erythroid
06:31progenitors which are in the.
06:33Nipple low population that has more CD 38.
06:36So we were able to enrich for these cells.
06:39We always get some erythroid, only an MB.
06:42Only colonies in our cultures of MVP.
06:44So one of our questions is really,
06:46is this a uniform population where by
06:49chance some of the colonies will be
06:51erythroid only and some will be Meg only?
06:54Or we really contaminated here
06:55with some erythroid progenitors and
06:57some Meg progenitors.
06:58So what we next did is to better
07:00understand this Anne how fake
07:02decisions actually occur is performed.
07:04Single cell RNA seq on the
07:06different populations in this work.
07:08Was done by Yishun,
07:09or will Lou in my laboratory.
07:12So what he did is he fax sorted out the
07:15candidate Mega Rich Ride Projectors.
07:18The Meg Progenitors the erythroid
07:20progenitors and then the upstream
07:21common myeloid progenitors,
07:23and this was done with single cell seq
07:26that was run by Amazong in the Yale
07:29stem cell center or genomics core.
07:32The data were then analyzed with
07:33the assistance of Nathan Salamone's
07:35at University Cincinnati,
07:36and I want to take a minute to
07:38really look at this heat map,
07:40'cause it gave us a lot of information.
07:42That was some of it quite surprising.
07:45So what you can see here is these are
07:47single cells from top to bottom that are
07:49from the common myeloid progenitor gate.
07:52These are single cells from the
07:53Mega Rich Royd Projector Gate.
07:55These are cells from the megakaryocyte
07:57progenitor gate and the erythroid
07:58progenitor gate and what you can
08:00see is there are distinct gene
08:02expression patterns that are unique.
08:03To arrest rate for janitors.
08:05Jeans are from left to right here.
08:07So these are genes that are expressed
08:09uniquely in Eryth Roid Progenitors.
08:11These are genes that are expressed
08:13uniquely in Meg progenitors.
08:14Mazer jeans that are expressed
08:16predominantly in common myeloid progenitors.
08:18But there's no such group of jeans
08:20for the Mega Mega Carey Cider
08:22Ridge Rd for generators.
08:24They really seem to be a
08:26transitional population.
08:26That's between the CMP where some
08:28of the CMP jeans are still on,
08:31but they're all going to turn
08:33off when they be picked.
08:34A fate to be Mega Rich,
08:36Royd,
08:37and then you see some that are
08:39me P an MB only,
08:40and some that are me P Anorith Droid only,
08:43and what that taught us is that
08:45MVP really are a unique population.
08:48They are not contaminated
08:49with Meg Progenitors.
08:50'cause these jeans are
08:51not on yet and they're not
08:53contaminated with the rich Roy Projectors.
08:55They really are at their own
08:57unique population that has a
08:59little bit of expression of E.
09:01Only jeans, MK only jeans,
09:03and some leftover from this
09:04common myeloid progenitor.
09:05So true transitional population.
09:08And we wanted to then look at the
09:10specific genes that turn on and off and
09:13understand whether the fate decisions
09:14are made according to what had been
09:17classic dogma in the literature.
09:19So the classic dug in the literature
09:21is than any pee pics to be a
09:23mega Carey Cider Inner it's roid
09:25commitment by expression of KLF one,
09:27or flee once OK LF one is also
09:30known as a rich Rd krupa like
09:32factor or E KLF and flea one is a
09:35transcription factor that is known
09:37to be necessary for Meg progenitors.
09:39To differentiate into megakaryocytes so when
09:41we looked at the gene expression of these,
09:43we thought we would find is that
09:45these cells are uniquely flea.
09:47One positive these cells are uniquely
09:49KLF one positive and these cells
09:51pick one or the other, or neither.
09:53But that's not what we got.
09:54Oh, and it's already been shown.
09:56I'm sorry in the dogma that
09:58fully one turns off KLF one.
10:00Tail F1 turns awfully one in cell lines,
10:03but when we looked at the gene expression
10:05patterns of all the various gene patterns,
10:07we could have jeans that are on and
10:09then off and the other lineages,
10:11or often CMP in them on in
10:13the other lineages,
10:14we found that there were very few
10:17transcription factors that really
10:18told us the story to as to how
10:20these state decisions are made.
10:22So just looking at flea one and KLF
10:24one what you can see is now what
10:26I've done is from the heat map.
10:29We still have.
10:29These are individual CMP's individual MPs,
10:31individual Meg progenitors and
10:32individual rich Rd projectors.
10:34And these are comb plots of individual genes,
10:37so KLF one.
10:38As you can see,
10:39is expressed in some me P in
10:41some Meg progenitors.
10:42It's predominantly expressed
10:43in the erythroid progenitors,
10:45but it's really not showing you that it's
10:48clearly making a fate decision of any sort.
10:50Flea.
10:51One is also expressed in almost
10:53every Lenny Edge at well.
10:55Yes,
10:55it is expressed strongly in Meg Progenitors.
10:57It's also expressed throughout MVP,
10:59and if you look for Co expression
11:02of flea one and KLF one in MVP.
11:04Sometimes it is coexpressed in,
11:06sometimes it's not,
11:07so it wasn't quite so simple.
11:08We also looked at Gotta One and
11:10got it two which are known to
11:12be critical for a rich Rd in
11:14megakaryocytic maturation.
11:15And what you can see is that gotta
11:17one comes on when the cells commit
11:19to the megathread progenitor
11:20lineages and then stay on.
11:21In both lineages got it too is
11:24expressed throughout as his NFE 2.
11:26So what really is going on and
11:28what's making these fate decisions?
11:29We looked at all of the different
11:32gene expression patterns and
11:33found some pretty interesting
11:34genes and genes that are
11:36somewhat me specific that we're
11:37pursuing further in my lab.
11:39But really, what the data ended
11:41up showing is that it's not one
11:43specific group of jeans I wanted
11:45to make sure to stop and say again,
11:48this is going to be.
11:50Publicly available on YouTube at
11:52this URL that I have down here and
11:54it's also available in the paper,
11:56you can actually put any gene
11:58of interest in an get the comb
12:00plots that you're interested in
12:02from the are single cell data.
12:04OK, So what ended up happening is
12:06when we analyze these data using gene
12:08ontogeny and other approaches is
12:10the cell cycle was amongst the most
12:12the genes of the cell cycle where
12:14the most differentially expressed.
12:15When you compared me, PETA,
12:17Meg progenitors and me Peter
12:19it's red projectors.
12:20You can see regulation of the
12:21cell cycle comes up here and then.
12:24All of these different differences
12:25between MVP and MKP and this actually
12:27kind of hit a nerve with us 'cause
12:30we already had a finding that
12:32suggested this might be the case.
12:34What we already knew.
12:35We had tried some candidate
12:37drugs to see if we could.
12:39If they affect the Mega Rich Rd
12:40fate decision and we had added for
12:42example all trans retinoic acid
12:44at low and high concentrations and
12:46seeing a dose dependent increase in
12:48megakaryocyte fate specification.
12:49Similarly with Rappo Mison with a dose
12:51dependent increase in Meg fate specification.
12:53And this was true for a lot of other
12:55drugs that we had tried that were
12:58known kinase inhibitors and what
12:59became clear to us is the thing
13:01that all of these inhibitors had.
13:03All of these drugs head is that they inhibit.
13:06Or slowed the cell cycle.
13:07So we decided to test that.
13:09So what we've done here is just
13:12treating with all trans retinoic acid,
13:14the 50 nanomolar,
13:15which is the same thing is down
13:17here you take any pee,
13:19stain them with CFC and then as
13:21the cells are in culture for
13:2372 hours that CFC gets diluted.
13:25So from as cells become lower in
13:28their fluorescence for CFC that means
13:30that they have proliferated more
13:31times and what you can see is the
13:34control population or blue population here.
13:36Has proliferated more times than
13:38the population that was treated
13:40with Aptra and which gave you a
13:41Meg lineages bias suggesting that
13:43this lower cell cycle might be
13:45associated with the Meg bias.
13:47Similarly with Rappa Mice and we see
13:49a slower cell cycle that's shown here
13:51in red is associated with the Meg bias.
13:54So what we did is we tested just
13:56inhibiting the cell cycle with
13:58the CD K46 Inhibitor and here you
14:00can see that the cells that were
14:02treated with the inhibitor or
14:04practically not dividing at all,
14:06whereas the control population is Dividing.
14:08And again we see a dose.
14:10Specific increase in the Meg
14:12progenitor cell fate decision.
14:14So the next thing we had to do
14:16is figure out a way to increase
14:19this cell cycle speed,
14:20and for this we ended up trying
14:23to knock down CD 21 and CD 57,
14:25but those things didn't work very well.
14:28They killed the cells so we
14:30tried something else and what we
14:32did is from the Vascular lab.
14:34We got two different constructs.
14:36One is C DK2 cycle independent,
14:38two cyclin dependent kinase
14:39to driving cycling.
14:40E phosphorylation and cycling depending
14:42kinase for which promotes cycling.
14:44D phosphorylation,
14:45and we coexpressed either CD K2 and
14:48Cyclin E or C DK foreign cyclin
14:50D in our MVP and what you can
14:53see up here is for the site CD.
14:55K for Cyclin D, which we call for.
14:58D actually promotes G one of the cell
15:01cycle and CK2 Cyclin E which we call
15:042 E promotes the G1 to S transition and down.
15:07Here we can see the data the two E.
15:11And four deconstructs gave
15:12us Anorith Royd Bias,
15:13which is what we were looking for.
15:15It's the opposite of the Meg bias that
15:17we get when we slowed the cell cycle.
15:20When you just had to Ian 40 to Meg
15:22progenitors in Eryth Roid Progenitors,
15:24you don't see any specific change
15:26in their fate specification.
15:27And up here,
15:28we're just showing that when you
15:30overexpress the 2E or 4D in the cells,
15:32you get a faster cell cycle,
15:34or the CFC is more diluted.
15:35So this really suggested to us
15:37that the faster cell cycle is
15:39associated with the River police is.
15:41Slower cell cycle with Mega Carey
15:43side of Louise is so can we assess
15:46cell cycle in vivo and now I'm bout
15:48to present to you some unpublished
15:50work using and now recently published
15:53fluorescent Reporter mouse.
15:54They shun ching glow at Yale University
15:56has developed and what she did is
15:59she made a mouse that basically
16:01tells you the cell cycle speed of
16:03any given cell that you look at.
16:05She did this by overexpressing a cell
16:08cycle timer protein that starts out
16:10blue and then gradually becomes red overtime.
16:13And this has been described previously.
16:15She fused it with Histone
16:16H2B so it was nuclear.
16:18And what you can see in the math is
16:20all in this paper that's available
16:22in bio archive soon to be out
16:25in a peer reviewed publication.
16:27Basically what happens is the blue
16:29because it is expressed for just
16:31a short time after the protein
16:33comes on is always steady,
16:35whereas the red gets brighter and
16:37brighter and brighter as as the cell.
16:40Proliferates So what you can see is
16:42if you look at the blue red ratio
16:44you get a sense of how quickly
16:46that cell has been proliferating.
16:48So when we did this and I think the
16:51main thing here is to look down here,
16:53I'll tell you what we did.
16:55What we did is we took urine cells from
16:58these Reporter Mice and we gated on them.
17:00You Ring me pee Meg.
17:02Progenitor eryth roid for genitor
17:04or fully differentiated cells that
17:06were Lynn positive and then for
17:08every cell we looked
17:09at the ratio of blue to red.
17:11If you have a lot of red,
17:13so a high red to blue ratio,
17:15that means you're proliferating much,
17:17much faster and what you can see is that the
17:19erythroid cells the erythroid progenitors
17:21are the fastest cells in the proliferation.
17:24The Meg progenitors were in the middle.
17:26And the MVP,
17:27the by potent progenitors upstream,
17:29where the slowest.
17:30So this was new information to us,
17:32although it was consistent with
17:33what we've seen in vitro for human
17:36cells is that MVP are probably
17:37the slowest proliferating cells,
17:39then the Meg progenitor cells,
17:40then the erythroid progenitor cells.
17:42So our model right now is that
17:44me PR is quite slow.
17:46They do self renew and then
17:47if you pick up the cell cycle,
17:50so it's a little faster than that,
17:52you're a Meg progenitor and a faster
17:54cell cycle, Anorith raid progenitor.
17:56And we haven't.
17:57Still answered the question
17:58as to how this happens.
18:00That's a really,
18:01really exciting,
18:02important question.
18:02We think there are lots of things
18:05going on in addition to maybe
18:07epigenetic things with DNA methylation.
18:09There are probably also changes
18:11in phosphorylation of critical
18:13transcription factors when you
18:15have a slower fast cell cycle.
18:17So I'm going to move on to now is
18:20whether there are clinical scenarios
18:22in patients where MP fate might
18:24play a critical role and this
18:26is work that was done again bij
18:28Juliana Javie Ferruccio and now in
18:30collaboration with Vanessa Scanlon,
18:32an instructor in my laboratory and
18:34what we're looking at is iron deficiency.
18:36So it's long been known anecdotally,
18:38the patients who become iron deficient
18:40have elevated platelet counts,
18:42and we decided we really wanted
18:44to look at that and see whether
18:46it might be an MVP fake decision.
18:49So what you're seeing here is just
18:51data from published accounts of
18:53patients with iron deficiency anemia,
18:55specifically in.
18:57Um Irida, which I'll tell you bout,
19:00which is a temporal 6 Mutation population,
19:03but they're highly iron deficient and.
19:06Refractory to when you add iron.
19:08So iron refractory iron deficiency
19:10anemia and what you can see is
19:13in individual patients.
19:14These are 11 of individual patients.
19:16As the hemoglobin goes down,
19:18the platelet count goes up.
19:20So if we could look at the
19:22MVP in these patients,
19:24that would be really cool.
19:25'cause we could determine whether
19:27or not there are megakaryocytes
19:28biased in the iron deficient state,
19:30given that we couldn't get bone
19:32marrow from these patients,
19:33we did the next best thing,
19:35which is to look at Immuring
19:37model of this disease.
19:38So the mooring model of the diseases
19:40that Empress 6 knockout mouse an
19:42we were very fortunate that current
19:44finberg is here at Yale and she
19:46really helped define the temper
19:486 knockout in patients and she
19:49also had the mice.
19:51So Long story short,
19:53temper 6 normally keeps hepcidin levels low.
19:56Low hepcidin levels allow you to absorb iron.
20:00If you have high hepcidin
20:02then you don't absorb any
20:04iron, so Long story short,
20:05our temporal 6 knockout mice do not absorb
20:08iron because their hepcidin levels are high.
20:11And So what we did is we got the temporal
20:146 nicean we looked to see whether they
20:16have microcytic anemia and thrombocytosis,
20:18and they do so with what
20:20I'm showing you here.
20:21Here's the temper 6 knockout.
20:23Here's the wild type.
20:24The hemoglobin is low, the hematocrit is low.
20:26The MCD mean corpuscular volume is low.
20:28So that's why their microcytic their small
20:30red blood cells and the platelet counts
20:32are significantly elevated in these mice.
20:34So the next thing we could do is
20:36look at these mice at their MPs.
20:39So what we did is we fax sorted
20:41out murin megathread progenitors.
20:43And we grew colonies and what you
20:45can see is that the temper 6 knockout
20:48me peas have a megakaryocyte bias,
20:50which is exactly what we predicted.
20:52So this was super exciting to us
20:54and the next thing we did is well,
20:57do they proliferate slower?
20:59And the answer is yes.
21:00So again,
21:01it's consistent with our previous
21:03data that slower cell cycle goes is
21:06consistent with a higher megakaryocyte bias.
21:09We wanted to also study this in human
21:11cells an what we did initially didn't
21:14workout and what we did initially is.
21:16We just tried growing the cells
21:18in the presence of iron chelators
21:21hoping to get a low iron environment
21:23for the cells that would allow
21:25us to see the human cells.
21:27Reiterate these Mooring Dataware
21:29with human me pee when you grew them
21:31in low iron you'd get a Meg bias.
21:34However that didn't workout because
21:35whenever we lowered the iron in
21:38vitro the cells didn't grow colonies
21:40so we ended up instead using.
21:41A more biochemical approach,
21:43which was to knock down TF are two,
21:46so TF are let me just go through back that
21:49there are two different cvars transfusion.
21:53I mean transferrin receptors, there's
21:55transparent Receptor One which is CD 71,
21:57and it's an all proliferating cells and very,
22:01very highly expressed in Eryth roid cells.
22:03And there's transparent receptor 2,
22:05which is entirely different.
22:07Transferrin receptor two
22:08acts more as an iron sensor,
22:10but not as an iron transporter.
22:13And what it does is it binds
22:16to Jolo transparent.
22:17That's transparent,
22:18that has iron bound to it,
22:20and when it's bound to that.
22:24Transparent it activates Erk 1
22:25two and P38 map kinase signaling,
22:28so it's kind of a baseline.
22:30The cells are always expressing
22:31TF R2 they have iron,
22:33they're saying to the MVP.
22:35We got plenty of iron.
22:36Do what you need to do if you
22:39actually have no iron present
22:41so that you have transparent,
22:43but it's not bound to iron then that
22:45Fr two comes off the cell membrane.
22:48It gets internalised and you lose
22:50your Erk 12 signal so it's really an
22:53iron sensor in the presence of it.
22:55It's saying to the cell.
22:56There's plenty of iron available.
22:58Do what you need to do in the absence is
23:00saying wait no eryth roid differentiation.
23:02I'll show you the data
23:04'cause we're low on iron.
23:05So what we do is we knocked down TF are two.
23:08This is just showing you the M RNA
23:10levels are decreased and true enough
23:11when you knock down TF are to the
23:14cells have slower proliferation.
23:15So that told us we might end up seeing
23:17a Meg Bias which is what happened.
23:20So what you're looking at here
23:22is the colony distribution and
23:23non transduced primary human MVP.
23:25Me P transduced with the scrambled
23:27SH RNA and transduced with an SH RNA
23:30against TF are one against TFR 2.
23:32Two different SH RNA's against TF
23:34R2 and in both cases we're seeing
23:36a megakaryocyte bias so this was
23:38consistent with the low iron environment.
23:40Both the mouse and human leading to
23:42a Meg bias Anna slower proliferation
23:44but what's the mechanism of this?
23:47So in order to assess the mechanism
23:49what we did is we went back to
23:52the Meyssan we fax sorted out the
23:54Mega Rich ride for janitors.
23:56And looked at weather.
23:57Looked at the gene expression patterns
24:00between wild type and temper 6
24:02knockout MVP and this work was done
24:05in collaboration with Tomata Baldy
24:07who is does hit our bioinformatics and
24:10what he showed is that there are many
24:13genes that are upregulated when you
24:15knock out temper 6 in these empiezan.
24:17Those tend to be targets of veg F,
24:20whereas there's downregulation of Erk
24:23target genes again consistent with the
24:25idea that you're losing Erk signaling
24:27because the TF R2 would be internalised.
24:30So what we did is we looked in
24:32those MVP's to see if their fasfa
24:35lurk levels were actually decreased
24:37and they were consistent with the
24:39decreased Erk target genes being
24:41phosphorylated and being expressed.
24:42We actually have decreased
24:44or toss work levels.
24:47Similarly, so that was in the mooring system.
24:49Similarly in the human system,
24:51when you knock down the TF R2,
24:53you have decreased phospho, Erk levels.
24:56So the model that we have for what's
24:58going on in this fake decision is that
25:02under normal iron conditions you have
25:04the cells making their decisions to go
25:07down the megakaryocytic versus areth
25:09ride progenitor cell lineages and in
25:11the presence in the case of iron deficiency,
25:14you get anemia at least in part
25:17because you have decreased TF R2.
25:20Which leads to decreased phospho Erk levels,
25:23decreased proliferation and a
25:24megakaryocyte bias which leads you
25:27to have elevated platelet count
25:29and a decrease erythroid count and
25:31the microcytic microcytosis is the
25:33downstream effect of lacking iron
25:35as the erythroid cells are maturing.
25:40So there are a lot of
25:42unanswered questions here and.
25:44Some of them are listed here.
25:46To what extent do me P self renew?
25:49I showed you that me PR this
25:52unique transitional state can
25:54me pee themselves proliferate.
25:56Does cell cycle speed actually
25:58change as cells are about to
25:59undergo their fate specification?
26:01Could we see cells starting to slow
26:03their cell cycle speed in that predicts
26:05there about to pick the megakaryocytic
26:07and speed up their their fate,
26:09their speed up their cell cycle,
26:11and start to pick the erythritol image?
26:14Are there cell characteristics that
26:16predict fate decisions like the
26:19cell size or the cell motility?
26:22What about if we changed the cytokines?
26:24Does that affect fate decisions?
26:26How did cells in the micro
26:28environment affect decisions?
26:29These are all unanswered questions that
26:32we're addressing in our laboratory now in.
26:34This brings me to the beautiful
26:36dynamic system that Vanessa Scanlon
26:38in my laboratory has developed.
26:41So what Vanessa is doing is she is live
26:44image Ng me P as they form colonies.
26:47So she fax or it's out the MVP and
26:49she puts them into semisolid medium,
26:52the same semisolid medium that we
26:54use for a colony forming assays.
26:56It's a collagen based medium.
26:58It's basically mega cult from
27:00stem cell technologies,
27:01and we've added Orris Republican,
27:03and she puts these into an
27:05Olympus Viva view system.
27:07And what this system is is.
27:09It's basically an incubator.
27:10And inside the incubator are these
27:13spots where you can put the dishes.
27:15You can put eight dishes in its hooked up
27:17to a computer an underneath the incubator.
27:20There is a fluorescence camera so
27:22you tell the fluorescence camera
27:23where to take pictures and how
27:25often to take pictures overtime
27:27and you since this is actually an
27:29incubator you can keep them in the
27:31incubator for up to two weeks,
27:33three weeks and really watch
27:35the individual colonies form.
27:36I should mention also that when.
27:39Vanessa was putting this system together.
27:41What she realizes she had to have
27:43some way of flattening the colonies so
27:45that we could watch the cells overtime
27:47and they didn't form a big stick.
27:49Big, thick 3 dimensional colony.
27:51So when she puts the cells in the
27:53first thing she does is she puts him
27:55in about 15 microliters and then
27:57put the cover slip on top of that.
27:59It doesn't affect the distribution
28:01of colony subtypes and it allows
28:03us to see the individual cells.
28:04So what we're seeing here is
28:06a colony with megakaryocytes.
28:07That's the green and erythrocytes in red.
28:10Here's a megakaryocyte only colony,
28:11and here's an Areth Royd Colony.
28:13I left out.
28:14The fact I'm sorry that towards
28:15the end of the culture period,
28:17she adds antibodies against 235 a.
28:19That's like a four and A and CD
28:2141 so that we can identify the
28:23different cell types.
28:24If you add these antibodies
28:26too early in the culture,
28:27then the cells will die.
28:29So there's some phototoxicity
28:30that we really needed to address,
28:32but we're still.
28:32We're still optimizing that,
28:34so we can add the antibodies sooner.
28:37So what I'm going to show you
28:39here is one of the initial images.
28:42Stacked images or movies that
28:43Vanessa was able to get when
28:45she sorted primary human MVP,
28:47put them in culture and
28:49watch them form a colony of
28:51megakaryocytes and erythroid cells.
28:57I'm still here, I just want you to watch.
29:01When you start to see the green color,
29:04that's when she's added
29:05the antibody against E 41,
29:06so those are the megakaryocytes an,
29:08then the pink cells are the ones
29:09that are standing with two,
29:1135 A and those are the erythroid cells.
29:14So the wonderful thing about
29:15having these time lapse images is
29:18that then you can play it back
29:19and forth and actually figure out
29:21which cell divided to Weikum which
29:23other cell type and make a tree.
29:25So here is it.
29:26Tree or lineages tree from a single
29:28by potent megathread for Genitor
29:30and what we've done with these
29:32trees is any cell that was green
29:34at the end or Meg committed at the
29:37end became a green cell at the end.
29:39Any cell at the end? That's red.
29:42That was a rich Rd committed if.
29:44There's a cell that has downstream of it,
29:46some green cells and some red cells.
29:48We call data by Potence Elan, it's blue.
29:52So one of the first things you can
29:54see here is the blue cells can self
29:56renew that me P can self renew so
29:58that had not previously been show.
30:00Phone and now we have that in every
30:02video we have with by potent colonies,
30:05we can see that the MVP themselves
30:07self renew in vitro.
30:08What you can also see is that
30:10there are different patterns.
30:11Sometimes we have a cell that commits to
30:13the erythroid lineages quite early on,
30:15and sometimes it takes much,
30:17much longer to commit to
30:18the erythroid lineage.
30:19Similarly with the megakaryocyte Lenny edge.
30:22We also have colonies that
30:23are eryth roid only.
30:25That came from our sorted me pee
30:26and colonies that are mega only.
30:28They come from our original MVP and
30:30what you can see is initially the.
30:33Cell cycle is relatively slow,
30:35but when it's a erythroid only colony
30:37they get very very fast overtime,
30:39whereas the megakaryocytes or
30:41have a slower colony formation.
30:45So this gave us the opportunity to
30:47address to start to address some of
30:50the many many questions that we have,
30:53and the analysis is on going.
30:55I'll just give you a glimpse as to some
30:58one of the stories that has become more
31:00clear now that we have this dynamic
31:03system of looking at the colonies.
31:06So the dogma that many people believe.
31:09Maybe not.
31:09The folks on this conference on this
31:12web and R is that an Emmy P might
31:15make its fate decision by in the
31:17presence of Thrombo poet and picking
31:19the megakaryocyte Lenny edge and
31:20in the presence of erythropoetin
31:22picking theorist Freud Lenny Edge.
31:24So just to blow that up here that
31:26if we were to grow the cells in
31:28the presence of Thrombo poet and
31:30in the absence of Ipoh we would get
31:33colonies that are only megakaryocyte.
31:34And if we left out the thrombi poet,
31:37and we'd get colonies that are
31:39just eryth roid.
31:41Those of you who know the literature
31:43more deeply might not predict that
31:45something just didn't show up here,
31:47which is the background on this
31:49information background here.
31:50So what I wanted to make sure to tell
31:52you is some of the background as to why.
31:56If you culture the cells in the
31:58presence of minus Devitte Bowen tipo,
32:00you might not predict that this
32:02is what that you would get.
32:04The only in MK only colonies.
32:07So initially what we did is we just
32:09perform static colony forming unit
32:11assays in the presence and absence of
32:14vivo en tipo and the background here
32:16is that we knew that if you don't
32:18have the Erythropoetin Receptor,
32:19you don't form normal Aris Freud Colonies,
32:22but you have some colonies,
32:23so it's not that they can't pick
32:25the fate decision.
32:26It really sounds like it really appears
32:29more that it's a survival signal,
32:31and similarly if you overexpress
32:32the thrombopoietin receptor,
32:33it doesn't cause the cells to just become
32:36megakaryocytes and not a rich way itself.
32:38So we thought there would probably
32:40be more to it and we decided to
32:42test that using this dynamic model.
32:44So the first thing that we did and
32:46this was work that was done by an
32:49undergrad in my lab several years
32:50ago is when we grew the colonies.
32:53In the absence of tipo,
32:54we saw no difference in the
32:56fate specification,
32:56just fewer colonies suggesting
32:58a survival detect.
32:59In contrast,
33:00when we leave out erythropoetin,
33:02we got absolutely no erythroid colonies,
33:04suggesting maybe that it was that
33:06are important is necessary for
33:08the erythroid colonies to grow,
33:10or it was necessary for the fate decision,
33:12but we couldn't tell the difference,
33:15but we've gotten a little smarter since then,
33:18and what we did is instead of just using
33:21CD-235-A as our marker for with Tripoli Sis,
33:24we're using an earlier
33:25marker which is CD71 CD 71,
33:27which is transparent.
33:28Receptor one goes up.
33:30Logarithmically as cells commit
33:31to the erythroid lineage and we
33:33found that if we staying for
33:35CD71 and CD41 instead of 2:35,
33:37but we could actually see that
33:39the cells were committing to
33:41be a Richard Lenny Edge even in
33:43the absence of a riffle poet.
33:45So these are now the data
33:47with this new marker,
33:48the CD 235 a what you're looking at is.
33:51This is a colony that formed both Mega.
33:53These are three different colonies that
33:55formed megakaryocytes and erythroid cells.
33:57Control colony.
33:58Here's one in the absence of Thrombopoietin,
34:00and here's one in the absence of A with
34:02report and see the colonies are much smaller,
34:05but they're still here.
34:06This is the and MK only colony,
34:09and here's a risk.
34:10Weighed only colony, so they all form.
34:12They're just much,
34:13much smaller in the absence of.
34:15Otherwise,
34:15report when we think in retrospect
34:17that this is that the standing
34:20for 235 was really the problem,
34:22but that neither equal nor tipo
34:24affects the fate decision of the MVP,
34:26so we wanted to look at to improve
34:29this further by making a video of
34:31timeless microscopy of colony forming
34:33in the absence of humble poet.
34:35And so this is a colony growing
34:38in the absence of crumble poet,
34:40and you saw that other colony earlier
34:42and it just kept growing and growing.
34:45And what you'll see here is as the cells.
34:48So start to proliferate and form a colony.
34:52Still start to die so you can
34:54see that we picked.
34:55Those red ones are the
34:56erythroid limited sells.
34:57The green ones are the Meg committed cells,
34:59but you're not seeing that same
35:01log Arhythmic expansion itself
35:03because they're starting to die.
35:05And what this is showing us is
35:06that are smaller colonies are not
35:08just due to decreased proliferation
35:10in the absence of from a potent,
35:12but actually do to cell death in
35:14the absence of Fraud Department,
35:15which we wouldn't want otherwise
35:17have been able to see.
35:21Mrs. Just showing you these
35:22colonies side-by-side,
35:23here's a normal colony with
35:24megakaryocytes in erythrocytes,
35:25and here's a colony with that has both
35:28megakaryocytes in a richer sites,
35:30but the vast majority of cells died so that
35:32you didn't get this huge colony expansion.
35:35So our data really have shown quite
35:37nicely that tipo versus EPO do
35:39not affect the MVP fake decision,
35:41so it's negative data,
35:43but it's negative data where we're
35:45starting to get a clue as to the
35:47fact that tipo is necessary for cell
35:49survival and Ipoh is necessary.
35:51For with Droid maturation.
35:54So of course we have a lot of on
35:56going studies looking at whether MVP,
35:59self renew and weather cell cycle speed
36:01predicts subsequent fate decisions,
36:02whether there are other characteristics
36:04that affect cell motility and really
36:06exciting new data that Vanessa is
36:08getting was on the role of cells,
36:10other cells within the micro environment,
36:11and how they affect MVP,
36:13fate specification 'cause all of
36:15our videos to Dayton colonies or
36:17with pure MVP and not with the
36:19other cells that they would be next
36:21to in the bone marrow environment.
36:24So to summarize what I told you today,
36:27single cell RNA seq reveals that
36:30MVP represented unique transitional
36:32state in both primary human cells
36:34and primary mooring cells.
36:36That these MVP are capable of self renewal.
36:39The single cell RNA seq also gave us
36:42a clue that cell cycle differences
36:45between MVP Meg progenitors inner
36:47it's right for janitors are.
36:50Probably playing a role in that
36:51fake decision where the slower cell
36:53cycle promotes a megakaryocyte fate,
36:55whereas a faster cell cycle promotes
36:58in Eryth Roid Fate.
36:59The in vivo timer mice supported
37:01the fact that the MVP or slower
37:04than Meg Progenitors,
37:05which are slower than originally projected.
37:08I showed you data that the iron
37:11content in MVP toggles the MK
37:13versus E fate decision via veg FERK
37:15or signaling and the time lapse
37:18imaging reveals that tipo Niko do
37:21not affect fate decisions per say.
37:24So I've not acknowledge people
37:26as we've gone along,
37:27so I hope that I called on everybody
37:30who was played a role in this work
37:32on may not have mentioned Lee Grimes,
37:35who works with Nathan,
37:37and Lee has also been incredibly
37:39helpful in us with us.
37:40Analyzing the time lapse images
37:42'cause he asks tremendously important
37:44questions that we're now beginning
37:46to address,
37:46which is how likely is it that an MVP
37:49will self renew versus undergoing mag
37:51versus inner. If word fate decision.
37:54Overtime in culture.
37:55I want to mention that my lab
37:57is looking for new postdocs,
37:59so please consider applying and
38:01that this work was also supported
38:03by the Yale Cooperative Center
38:06of excellence in Hematology.
38:08Thank you very much.