Back Home: Mechanisms of Neural Mimicry in Melanoma Brain Metastasis

March 15, 2023

Information

Yale Cancer Center Grand Rounds | March 14, 2023

Presentation by: Dr. Eva

ID9680

To CiteDCA Citation Guide

00:00Friday, and thank you so much Harriet
00:03and Marcus for your kind invitation.
00:05I promise I'll go back in person and
00:08hopefully with less technical issues.
00:11So as all of you know,
00:13probably even better than me,
00:15Melanoma is the fifth most common
00:17cancer in men and sixteen women.
00:19And can you please keep clicking?
00:21I'm going to go fast through this first
00:23slides because it's just background.
00:24As you know, the incidence of
00:26Melanoma continues to increase.
00:27It double s every 20 years
00:29and luckily this is like.
00:31Unfortunately,
00:31it's not a updated the mortality
00:34associated with Melanoma is starting to
00:37decrease and this is you definitely to
00:40the development of better therapies.
00:43For the past 15 years,
00:46Melanoma has become the poster
00:48child for many targeted therapies
00:50as well as immunotherapy space.
00:52Please keep clicking,
00:54but all of you know there is still
00:57despite the success of this.
01:01Uh,
01:01we're developing immunotherapy treatments.
01:04We still can keep clicking.
01:06We are still facing significant resistance
01:11to some of these treatments or a.
01:18Year um side effects.
01:20So approximately 40% of the
01:23metastatic Melanoma patients have
01:24not benefited from the advances
01:26that have been done over the past
01:28years in the treatment of Melanoma.
01:30So this takes me, can you please go on to my,
01:35the, the focus of my lab.
01:37Can you go back?
01:39Yeah,
01:39which evolves around these questions,
01:43right, the fact that we still need
01:45better markers for patient selection,
01:47understand.
01:48Where we check the patients that are
01:50going to uh go on to recur after the
01:52initial resection of their tumor,
01:54which ones are going to metastasize.
01:57And also in particular,
01:59as Harriet mentioned,
02:00we've developed an interest in
02:02understanding the biology of CNS metastases,
02:05particularly of brain metastasis
02:07because they show less level responses
02:10to immunotherapy and the incidence
02:12is increasing as patients are living
02:15longer and surviving from other systems.
02:18Astasis, so my lab is focused
02:21into aspects of metastasis,
02:23the initial steps of early dissemination,
02:25we now know that primary tumors,
02:28especially in Melanoma.
02:29Is that uh,
02:31disseminating start sharing cells that
02:33go into circulation and metastasize
02:36really early on in the process.
02:38So we want to understand,
02:39please click uh,
02:40which are the mechanisms that
02:43drive the metastatic behavior of
02:46certain melanomas and not others.
02:49We know that unity alterations
02:51don't explain this behavior because
02:53none of the well characterized
02:55genetic alterations of Melanoma
02:56or combinations of those genetic
02:58alterations can explain why.
03:00And primary tumors recurred,
03:02metastasized and some others don't.
03:05And also we want to understand
03:07what is the contribution of
03:09intratumoral heterogeneity.
03:10So we know now that despite the
03:13fact that primary tumors can have
03:15a relatively homogeneous genetic
03:17profile inside those tumors,
03:19we can recognize groups of cells with
03:22different transcriptional programs,
03:23what we call transcriptional states.
03:25And there is a recent search for what
03:28are the programs for their states.
03:30That drive these metastatic behavior
03:32and people have pointed to the neural
03:35Crest like or the EMT program on the
03:37other side of the equation.
03:39We are also trying to understand
03:40what happens at the end of the
03:42process once the cancer cells,
03:44in this case Melanoma cells have already
03:46extravasated into those distal organs.
03:48As you know Melanoma metastasis to the
03:50lung delivered in the brain and we are
03:53interested in understanding what are the
03:55site specific adaptations in those sites,
03:57where are the type of metabolic changes.
04:01The cell types of the cancer
04:03cells are interacting.
04:04And what is this crosstalk between the
04:06cancer cells and their environment?
04:07How is shaping the ability of animal cells
04:09to adapt and grow in those environments?
04:12And I will have to stories
04:15presenting today if time permits,
04:17one that focuses more in just the
04:20metastatic potential in general and how
04:22muscles cops some of the programs of the
04:24neural Crest cells from which they arise.
04:26And also in the second part of the talk,
04:30I will talk more specifically.
04:31About to bring metastasis and how Melanoma
04:35cells mimic some of the processes
04:38that happen during neurodegeneration,
04:40particularly to suppress new inflammation
04:42and be able to grow within the brain.
04:45So diving into the first story,
04:47these are the four areas in which my
04:49lab has been focused in recent years,
04:52epigenetic alterations, noncollinear,
04:54a post translational modifications
04:56particularly like oscillation
04:57and site specific adaptations.
04:59So moving on,
05:01can you please move these slides one more,
05:04the first story that has been
05:07recently accepted for publication
05:09is the work of three people.
05:12It evolve over time and we have.
05:15Uh,
05:15tightly muscles invoke and you like clear
05:18sticky genetic program during metastasis.
05:20So the genesis for this project was a very
05:24simple premise and it's the fact that,
05:26well,
05:27we all know that neural Crest cells are
05:29the cells of origin of melanocytes.
05:30And these are among the most totipotent
05:33and among the most invasive and migratory
05:36cells of our body and is well known
05:40now if you keep clicking that Melanoma
05:42cells adopt programs characteristic.
05:45You address the cells during the progression
05:48from primary to metastatic cancer.
05:50So we wanted to investigate which
05:52epigenetic changes happen during
05:54melanocyte differentiation from the
05:55neural Crest to the melanocyte that
05:58potentially could be reversed in the
06:00transition from primary to metastasis.
06:02And we did this by looking
06:05specifically at DNA methylation.
06:07There are of course many other
06:09mechanisms of epigenetic regulation
06:10being the enumeration one of them.
06:12So if you move to the next slide.
06:16This is what we did basically was
06:17a comparison of four data sets.
06:19So in one hand I cannot use my pointer,
06:22but you can see that we have 4 columns.
06:25We have neural Crest cells that were
06:27obtained from a collaborator in France.
06:29These are human cells.
06:31We also had the human melanocytes
06:33from different donors.
06:35And as you can see there is
06:36kind of a mirroring pattern.
06:38We were looking for CPG islands that
06:41were either hypomethylated in the
06:43conversion from neural Crest cells.
06:46To Milano sites that were progressively
06:48hypermethylated from primary to
06:50metastatic Melanoma or the converse,
06:52if you see at the bottom,
06:54we have some CPG islands.
06:57If you can click again,
06:58you will see there are some CPG
07:00islands around the gene called NR
07:0222 that are hypomethylated in the
07:04neural grass cells that become
07:07hypermethylated in melanocytes and
07:09progressively hypomethylated from
07:11primary to metastatic Melanoma.
07:12And this really call our attention
07:15because it really represented.
07:16Example of a potential gene or or
07:19candidate program to be modulated
07:22during neural Crest differentiation
07:24to melanocytes that was reversed
07:27during the progression from primary
07:29to metastatic Melanoma and what
07:31is first of all an R2 of two,
07:33so it's a nuclear receptor
07:35is also called cooked.
07:37TF2 is an orphan nuclear receptor.
07:39We don't know still what
07:40is the natural ligand,
07:41although retinoic acid can bind
07:44it at high concentrations.
07:46This is the motif that it binds,
07:47and as you can see the NRF 2 full
07:51isoform has a DNA binding domain.
07:54It has a ligand domain.
07:55So it's a conventional nuclear receptor
07:58that can form a **** or heterodimers.
08:01It is essential if you,
08:04yeah,
08:05it's essential for development and
08:07particularly for the formation of blood
08:09vessels and it has been shown to be
08:12important for neural test differentiation.
08:15Now there have been already some studies
08:17showing the role of NRF 2IN cancer,
08:20particularly pancreatic and
08:22prostate cancer metastasis,
08:24whereas in breast cancer it
08:25had a controversial effects.
08:27So you can say well you know this,
08:29this may already be known
08:30that this factor is important.
08:32Cancer, however,
08:32if you click one more slide,
08:35what really cool our attention is
08:37that the CPG islands that we found
08:40to be differential differentially
08:42methylated in neural Crest cells
08:45versus melanocytes and then later
08:47on in primary versus Melanoma,
08:49we're actually affecting exactly that
08:53region in the transcription start
08:56site that controls the expression
08:58of an alternative isoform isoform 2.
09:02Which is not the one that has
09:04been commonly characterized.
09:05So most studies have focused on
09:06the full length isoform, isoform,
09:08one that has the DNA binding domain,
09:11the like minded domain.
09:13But this methylated region or the
09:16methylated region is in atss that
09:18gives rise to a truncated isoform
09:20that lacks the DNA binding domain.
09:22So on the right you can see that as
09:25we had shown before the CPG islands
09:30controlling this this nuclear receptor.
09:32Networks to ISO two are hypomethylated and
09:36they appear in green in cells and stem cells,
09:40but completely hypermethylated
09:41in melanocytes,
09:42and this is also described here at the
09:45bottom where you can see the beta value.
09:47The beta value of 1 means completely
09:50methylated and a better value close to
09:530 represents a hypomethylated gene.
09:55This completely corresponds to gene
09:57expression because you can see that
10:00whereas isoform one is expressed in ESL.
10:021000 monocytes isoform 2 is expressed
10:04in your cells and neural Crest cells,
10:07but not expressed at all in melanocytes.
10:10OK.
10:10So in when you look at the human samples,
10:13you can see and that even though
10:15it's not completely
10:16black and white, what we see is that
10:18there is an increased percentage of
10:21hypomethylation of isotope in the metastatic
10:23cases compared to the primary cases.
10:26You can see there is from 30% approximately
10:29to more than 50% of the samples have
10:33hypomethylation for NR2ISO2 and if you
10:35keep clicking you will see that you can
10:38see there is a progressive increase.
10:40In the better value from Levi
10:43to primary and metastasis,
10:45so more hypomethylated and conversely
10:47increase expression from primary to
10:50metastasis with a very nice correlation
10:52between M RNA expression and methylation.
10:55Is exactly, if you look at
10:57the protein levels,
10:58we were able to develop actually an
11:00antibody specific for isoform 2,
11:02which was challenging because
11:03there are only 15 amino acid symbol
11:05amino terminal that that specific.
11:06But also two you can see that there is an
11:08increase in the expression of isoform 2.
11:10There are few cells that are positive
11:13in the primary cases and this
11:15population expands in the metastasis.
11:17This is something that we are actually
11:20trying to understand now whether
11:22there is a selection for those cells
11:25that expressed isoform 2 over time.
11:27And again this is this is something
11:30we're coming time to investigate but
11:32if we can go on we we can see that
11:35isoform one the one that is the full
11:38length isoform is not modulated by
11:40methylation is completely hypomethylated
11:42in ES cells you know cells melanocytes
11:45and both in primary and metastatic
11:47Melanoma and there is no correlation
11:50between expression and methylation.
11:52So isoform one is always there,
11:54but is isoform 2 the one that.
11:56Is normally not expressed in melanocytes,
11:58but is increasingly demethylated
12:01and expressed from primary to
12:03metastatic Melanoma.
12:05Now of course the question is,
12:06is this isoform tool doing
12:08anything in Melanoma metastasis?
12:10And for that we had to move to cell
12:13lines and again we found this pattern
12:15in which melanocytes and some melanomas
12:17aliens had complete hypermethylation
12:19and lack of expression as you can see
12:21here on the right and you can also see.
12:27Hypomethylation in some of the
12:29cell lines and this corresponds
12:31to expression and the same can be
12:33seen at the level of protein.
12:36You can see isoform 2 expressed only in
12:37those cells that have hypomethylation.
12:39So this represents a good model to study
12:41loss of function and general function.
12:43But before I go into that you can see that.
12:49The methylation uh status can also
12:52be seen in short term culture.
12:54These are cells isolated from
12:56patients and again you have some,
12:59some of these short-term cultures
13:01have hypomethylation and some
13:03of them have hypermethylation.
13:05So this is not an artifact of invitro
13:08culture and it happens in cells derive
13:13very shortly from from patients.
13:16So if now we move into cells that.
13:20Have hypermethylation like mayor cells.
13:22If we treat them with editing agent
13:25like 5 ASA you can see that there is
13:28an induction of isoform 2 and this
13:31is seen also in short term cultures
13:34in which treatment with FAFSA results in
13:37induction of isophorone to expression
13:39with no change in isoform one.
13:41Now when we silence isoform to
13:43using the specific srnas against
13:45these isoform that don't have
13:47any impact on the other isoform.
13:49Initially we did not see any
13:52effect onto the proliferation,
13:54but we observed a clear decrease
13:56in the ability to form colonies
13:59in soft Agar or to form a sphere
14:02spheres upon single cells.
14:03So these are properties that
14:05are characteristic of metastatic
14:07cells and basically measure the
14:09ability of the cells to grow under
14:12very stressful conditions.
14:13Now if we overexpressed,
14:14sorry before we go into their expression.
14:17This is the experiment in in this case.
14:20We injected on the cancer cells
14:22intracardiac in a conventional
14:24in a model of metastasis that
14:25we use very often in the lab.
14:27And you can see that the silencing of
14:30isoform 2 between independent HR and
14:32I had a very significant decrease in
14:35metastatic potential in this model,
14:37which is quantified here on the
14:39right by bioluminescence as well
14:41as fluorescence intensity when
14:43we extract the organs.
14:44So it seems like isoform 2 silencing
14:48suppresses metastasis in these.
14:50A model and now we move into
14:52overexpression systems in which
14:54ectopic expression of isoform 2IN
14:56cells that have hypermethylation has
14:59no effect again into the culture,
15:01but it has a significant ability to
15:04increase the number of colonies and
15:06soft tagar as well as sphere formation.
15:08In vivo we injected these cells
15:10in the flank of the mice and then
15:13we did survival surgery.
15:14You can see that we cover two more
15:16area because there is always a lot
15:18of signal that comes even after
15:19you have resected.
15:20The majority of the tumor is the
15:23subcutaneous tumor and you can see
15:25that isotonic expression enhances
15:28metastasis both by bioluminescence
15:30as well as by histological analysis.
15:32We did this also by intracardiac
15:34injection and again we observed
15:36the same effect with an increase
15:38in metastasis overall.
15:39So it seems like ISO two,
15:42this particular truncated isoform of
15:44the orphan nuclear receptor and R2F2
15:47is able to promote metastasis and is.
15:50There's to be required in some
15:52models to in most of the models
15:53that we tried in lab and the paper,
15:55we have three or four different models.
15:58You can see a decrease in metastasis when
16:01you're silence or knockout this change.
16:03So now what are the programs
16:05that are modulated by isoform 2?
16:07We did RNA sequencing to get to this
16:10question and one of the pathways
16:12that was significantly modulated was
16:14EMT DPL to mesenchymal transition.
16:17Now we went on to demonstrate that
16:19a lot of the typical genes involved
16:22in the epithelial transition where
16:24silence in different cell lines
16:26when we deplete isoform 2.
16:29And here you can see a validation
16:32for snail which is reduced both
16:35transcriptionally and at the protein
16:37level when we silence ice form 2.
16:41Importantly, I as I mentioned
16:43at the beginning of my talk,
16:45now we have an understanding of
16:48intratumoral heterogeneity Melanoma
16:49and we have seen that most melanomas
16:51both in mouse and human display
16:54different transcriptional states,
16:55intermediate neural Crest
16:57like proliferation EMT or more
17:01melanocytic or differentiated.
17:04So we were curious to see whether
17:06NRF 2 and particularly the signature
17:08of genes modulated by isophorone.
17:11Who were particularly enriched in any
17:13of these transcriptional States and as
17:16you can see in this heat map, the EMT,
17:18in particular the transcriptional
17:20state that corresponds to an
17:23epithelial to mesenchymal transition
17:25seems to be the one that is more
17:29enriched in these two signature
17:31together with the neural signature.
17:33So this goes together with our findings
17:36that if two could be regulating the EMT.
17:41Um estate and in this additional
17:43analysis in which we we used scenic
17:46to understand the regulations the
17:49the basically the epigenetic and
17:51transcriptional factors that control each
17:54of these states melanocytic intermediate
17:56proliferative neural Crest like an EMT.
17:59You can see an I think it's more
18:01visible in the next slide that the
18:04top regulon controlling the event
18:06signature and this is a completely
18:08different analysis done in our mouse.
18:11Levels of Melanoma you can see that
18:14inner 2F2 is at the top is a top regulon.
18:18Controlling the EMT signature.
18:20So it seems like these transcriptional state
18:23of epithelium to mesenchymal transition
18:25seems to be mostly controlled by NF2 now.
18:29How this happens?
18:31Like I told you at the beginning that N2I2,
18:34another two ISO two is a truncated isoform.
18:38It lacks the DNA binding domain.
18:40So how is it possible that it has this
18:43capacity to control gene expression?
18:45Now when we started to work in this isoform,
18:48there were only two other papers that were
18:52studying alternative isoforms of two,
18:54all the papers that have been put out there.
18:58And study only the full length isoform.
19:01So there were two papers.
19:03One was suggesting that I2 acted as a
19:07dominant negative and was not removing
19:10or displacing isoform one from chromatin.
19:13With us another paper suggested the opposite
19:15that I saw two contributed to bind ISO,
19:18one to the chromatin and also what
19:20complicates more this interpretation
19:22of the results is the fact that
19:25isoform one has been or inner 2F2 has
19:28been described both as a receptor
19:30as a repressor of transcription as
19:32well as an activator.
19:33So we.
19:36It tested and the possibility that
19:38NRF 2 ISO two was interacting with
19:41isoform one and somehow modulating
19:43the ability of an artist to to bind
19:46chromatin and regulate gene expression.
19:47But for this to be true also from
19:50two had to have the capacity to
19:53bind to isoform one in the nucleus,
19:55and isoform 2 lacks the nuclear
19:59localization signal present in isoform one.
20:02It has an alternative nuclear localization
20:04signal that is much less potent.
20:06Thought we first show in a A
20:11fragmentation analysis that isoform
20:132 is able to reach the
20:16nucleus. You can see there not
20:19only the cytoplasm, right?
20:20So this suggests the possibility that
20:23it can interact with iPhone one.
20:26Then we also did IP analysis
20:28where we show that isoform one
20:30can pull down isoform two with two
20:32independent antibodies and the
20:34opposite is also true, we can use.
20:36Mile a GFP pulled down because
20:39they unfortunately the endogenous
20:41IP for I2 doesn't work well.
20:44So we use the exogenous construct
20:46that has a GP fusion and so we
20:49can pull down again isoform too
20:51with the isoform 1 suggesting
20:53that it's two isoforms interact.
20:55And when we look now at Chief of
20:59isoform or chromatin IP of isoform one,
21:03we can see that indeed when we silence
21:05isoform 2 there is a group of peaks.
21:07Of isoform one that now lose
21:10binding of isoform one,
21:11and this is more dramatic in this
21:14group of targets is a little bit
21:16less prevalent in this subset of of
21:19pigs that are bound by isoform one.
21:22But we can also observe the opposite.
21:24We can see that in a group of peaks there
21:26is an increased binding of isoform one.
21:28When I do is gone.
21:30So definitely the interaction between
21:32these two isoforms is complex.
21:34We know that the chip works because.
21:38Um and active two is their most
21:41significantly enriched transcription
21:42factor in these pics together with
21:45some other transcription factors
21:47which are we're also interested
21:49in in looking at the potential
21:51interaction of MR2 with those.
21:53So we can see that N22ISO2 modulates
21:57the binding capacity of the chromatin,
21:59but it not only in a way seems to be
22:03in a in a gene specific manner and
22:05when we integrate the cheap data.
22:08Our transcriptomic sequencing of cells
22:11in which we have depleted ISO two,
22:14we found that the majority of the genes are.
22:17A significant portion of the genes that
22:19are downregulated when we sell in size.
22:21Or two that are targets for an R2F2
22:24are genes that are involved in the EMT,
22:27snail, twist, Peter two, etc.
22:30So it seems like a lot of the NR2ISO1
22:34targets are actually in the direct.
22:38Same teachings.
22:39The opposite is that when we see
22:42look at the genes that are now
22:45upregulated by isoform 2,
22:46the majority of those genes are
22:49involving differentiation and
22:50pigmentation such as tyrosinase,
22:52DCT, etcetera.
22:53So with all of these,
22:55and still with many questions still open,
22:58we propose this modeling which
23:00in primary tumors isoform 2 is
23:02hypermethylated and silence.
23:04So the dimers of isoform 1 prevail over
23:07the heterodimers of isoform 2 and isoform 1,
23:10whereas in metastasis we see a displacement
23:13of the equilibrium towards the home,
23:16the heterodimer.
23:17Sorry,
23:17because now isoform 2 is present
23:19and is this heterodimer that allows
23:21the expression of neural Crest and.
23:24Empty genes such as twisted slag etcetera.
23:27So what we are now in the process
23:30of understanding is how these
23:32heterodimer is able to activate
23:34this EMT genes and we think that
23:37in part could be due
23:38to the interaction with third parties
23:41like additional transcription factors
23:43that are specifically attracted
23:45to the complex like isoform.
23:47So we are doing our time and other
23:49techniques to understand what are the
23:52complexes that are attracted in bound
23:54to the chromatin where we have the
23:56homodimers versus the heterodimers.
23:58Now moving on into the second
24:01part of my talk,
24:02I don't think I have to repeat the
24:04conclusions because that's basically
24:06the summary that I just have provided.
24:08So in the interest of time,
24:10I will move to the second part of the story
24:13in which we have focused on brain metastasis.
24:16This was the work mostly done by.
24:20I actually incorrectly stated PhD
24:21is an MD PhD from our laboratory,
24:24Kevin Kleffman that is now a.
24:28Accident at mass general.
24:31So we of course are interested in
24:34brain metastasis because it's an
24:36important and met clinical need.
24:38And although these tumors can
24:41respond to immunotherapy,
24:43we know that these responses are mostly
24:45seen in patients that are asymptomatic
24:47and in the symptomatic patients the
24:49responses are much poorer and these
24:51patients have overall very poor survival.
24:53So this remains very important
24:56clinical question and definitely.
24:58Fascinating biological, um.
25:00Uh,
25:01question.
25:01How Melanoma cells adapt to the
25:04brain microenvironment and why
25:05they have such a profound tropism
25:07for the brain is still something
25:10that we don't entirely understand.
25:12So uh in collaboration with the management,
25:15the director of the Melanoma program at NYU,
25:17we develop Melanoma short-term
25:19cultures that in some cases are
25:22derived from the same patients.
25:23So we can in some of the cases we
25:25were able to obtain a short-term
25:27cultures derived from a brain
25:29metastasis and extracranial metastasis
25:31from the same patient.
25:33And again this is a very
25:35difficult comparison to make,
25:37but we think it's very useful
25:39because it reduces some of the
25:41inter tumoral heterogeneity.
25:42As we observe what genetically
25:44and transcriptionally and what was
25:47really exciting to us is that when
25:49we labeled these cells with GFP
25:51luciferase and inject them back into
25:53mice with intracardiac injections,
25:55we observed that that metastatic the
25:59the the short-term culture that has
26:01been derived from the brain has in
26:05general more metastatic potential
26:07than the one that was derived
26:08from an extracranial metastasis.
26:10This is represented here on the right,
26:12but more specifically.
26:13The one that derives from the brain
26:16has more ability to metastasize to the
26:18brain and this is measured here as a
26:21ratio of brain to body luminescence.
26:23So it seems like this is a short term.
26:25Cultures retain some of the properties,
26:29some of the ability that they
26:31had gained in in people,
26:33in the patients of colonizing the brain
26:36and therefore could be a good model
26:38to study brain specific adaptations.
26:40So we went on to conduct.
26:44For the Omega analysis of these short-term
26:47cultures, in total we profile 25,
26:49approximately 12 and 13 brain
26:52metastasis versus second metastasis.
26:54Only a few of them, of course,
26:56were pair,
26:56the rest were unfair and the
26:58idea was to try to identify proteins that
27:00were differentially expressed in the brain.
27:02Metastasis input could be potential
27:05drivers of the adaptation.
27:07The first reply is that we found when
27:09analyzing the data is that the majority
27:11of the proteins found differentially
27:13expressed where proteins involved.
27:14Being neurodegenerative disorders
27:16such as Parkinson's, Alzheimer's,
27:18Oxfords and this was rewarding because
27:22our collaborator in at in the Anderson,
27:25Mike Davis had previously found that
27:27a lot of the proteins involved in
27:31different study transcript transcriptional
27:33profiling of brain metastases wouldn't
27:36reach in Oxford or proteins involved
27:39in the respiratory chain mitochondria
27:41and so this was confirmed in our.
27:45Plans.
27:45We found that short-term cultures of
27:48brain metastasis had elongated mitochondria,
27:51and they also had increased oxygen
27:54consumption rate in this seahorse analysis.
27:57But what we focus on was in the differential
28:00expression of proteins involved in
28:03Alzheimer's and Parkinson's disease.
28:05In particular,
28:06we landed for this study on AP.
28:09The amyloid processing protein there
28:11is a precursor for amyloid beta.
28:15It was induced in pre metastasis
28:17compared to metastasis,
28:18but not only AP itself,
28:20but the proteins that leave AP into
28:23amyloid beta like beta secretase
28:26or present presently.
28:28So we decided to modulate the loss
28:29of a P to see if it had an effect
28:32on brain metastasis.
28:33And initially we found that
28:37sorry went too fast.
28:39We found that supernatants
28:41of brain metastasis, Dr.
28:44short-term cultures,
28:44had higher secretion of family beta
28:47compared to the extracranial brain
28:49metastasis not only in our own hands,
28:51but also in short term cultures obtained
28:55from collaborators that we study institute.
28:57Silence.
28:58AP Again, we found no effect
29:00in proliferation in culture,
29:02but when we inject these cells
29:05intracardiac in immunodeficient mice,
29:06we observed this reduction of
29:09the brain to body ratio,
29:11suggesting that the loss of AP was
29:13particularly affecting brain metastasis.
29:15This was confirmed by histological analysis.
29:18This is entertaining of NUMA,
29:20which is a human marker,
29:22and therefore it can perfectly
29:25mark the cells that are.
29:27The deriving from the from the scenography,
29:30from the implant and you can see that
29:32there was a very significant reduction
29:35of brain and one positive cells but no
29:39effect on kidney or liver metastasis.
29:41We did ex vivo imaging MRI to conduct
29:46volumetric analysis that show us that
29:49we're not only less perimeter studies
29:52but also smaller brain metastasis and
29:55of course we show that this effect.
29:57Happens, you know other models,
29:59this is not a brain Tropic Melanoma cell
30:01line type one and we use national approach,
30:04in this case a crisper cast 9,
30:06to show again a reduction in
30:09brain metastasis.
30:10But this of course open a lot of questions.
30:12The first one is which step offering
30:16metastasis is the one in which
30:18is particularly required for the
30:20adaptation and the the arrival
30:22of Melanoma cells to the brain.
30:24So as you know brain metas is
30:26a complex process.
30:27It involves multiple steps,
30:31the intravasation from the tumor
30:33into the first into the stroma,
30:36then the intravasation into the vasculature,
30:38survival in circulation,
30:39and when the cells arrive to the distal site,
30:41in this case the brain,
30:42they have to again extravasate.
30:44Many of these cells will undergo
30:46cell death or become dormant,
30:48but those that are able to proliferate
30:51and survive in this environment will
30:53form micro and macro metastasis.
30:55So when was abeta required for this?
30:58For this process,
31:00so Kevin embark himself in really
31:03a difficult task of monitoring the
31:06kinetics of cancer cells injected
31:08intracardiac in these mice.
31:10So he did brain slice immunofluorescence
31:13and a lot of confocal microscopy
31:15and was tracking this GFP positive
31:18cells in the brain over days.
31:20So you can see that just one day after.
31:24In the cardiac injection,
31:25these cells are stuck in the vasculature.
31:27They even have the shape
31:28of the blood vessels.
31:29This is a tomato, tomato,
31:31lectin marking the the blood
31:33vessels you can see there.
31:35At day three they start extravasation,
31:38they start getting out of the vasculature.
31:41Some of these cells die.
31:43A lot of these cells die in the
31:45blood vessels or outside when they
31:47are able to extravasate as sustained
31:49by Clifton Space Stream and you
31:51can see that later on they start
31:53crawling through the blood vessels.
31:55In these process called Vascular Co option,
31:58they can later on form micrometastasis
32:00that day 14 and finally micrometastasis.
32:03So when you compare the kinetics
32:06of control cells here in the
32:08black line to those that lack app,
32:10you can see that the first steps
32:13of the kinetics are really similar.
32:15There is this big crisis where most of
32:17the cells that are able to extravasate
32:20die either in the vessels or right
32:23after extravasation but then after they 7.
32:25When the control cells are able to start
32:28expanding and proliferating happily,
32:30the ones that lack APP can no longer
32:34grow after the first or second division
32:36and they eventually disappear.
32:38They are they are dead.
32:41So we wonder which effects were required
32:45for for this role of ebata in the brain.
32:49And remember that this is a very
32:51complex environment where there are,
32:53you know, the resident myeloid cells,
32:55the microglia.
32:55Macrophages in some cases will
32:57matter derived macrophages that
32:59attracted to the tumor and it excels
33:01interfiling for sites and astrocytes.
33:03So we first look at astrocytes
33:05as a potential.
33:09Still of interest, because of the
33:11literature that had shown previously
33:14that activated astrocytes can be Co
33:16opted by the cancer cells to support
33:18the growth in the brain environment.
33:20And this is indeed the case.
33:22Also in our models where we see that
33:24if you look at the left panels,
33:27you can see that over time,
33:29as the muscles arrive to the brain,
33:31these are the control cells,
33:32you can see an increased presence
33:35of GFP positive astrocytes.
33:36So there is.
33:38Some equipment of activated astrocytes
33:40we cannot distinguish if it's
33:42recruitment versus activation of
33:45the surrounding astrocytes to the
33:47point that they form this network
33:49of active astrocytes that is called
33:53active Astro cytosis supporting
33:57the Melanoma micrometastasis.
34:00So what we observe is that cells
34:03that lack APP are unable to
34:06trigger these reactive.
34:07Cytosis around them.
34:08When they arrive to the brain,
34:11there is a significant reduction
34:13of positive cells of GFP.
34:16Positive cells around the cells
34:18demand muscles that lack APP,
34:20suggesting that perhaps a beta is important
34:23in triggering these Astro cytosis.
34:26Now, in data that I don't
34:28have the time to explain,
34:30we also show that Amelia better not only has
34:33the capacity to activate the astrocytes,
34:36but also.
34:38Can suppress the phagocytosis
34:40coming from the microglia.
34:42So we can see that there is a
34:44reduction of neural inflammation
34:45in the presence of family beta.
34:48So we think that Ali beta secreted
34:50by cancer cells can have multiple
34:52effects in the brain metastasis,
34:54macular vironment, particularly on
34:56the astrocytes and the microglia,
34:59but also it can influence,
35:01as it has been reported in Alzheimer's,
35:03the interaction with the endothelial cells.
35:06So of course these open,
35:09these findings open some
35:11possibilities and therapeutic
35:13opportunities because of all the.
35:16Armamentarium of drugs that have
35:20been developed against America beta,
35:22some of them beta secretase inhibitors and
35:26more recently anti American antibodies,
35:29some of which have been developed
35:31for clinical use and in some
35:33cases even approved by the FDA.
35:35So these open the possibility of
35:38repurposing some of these drugs which
35:40are generally safe for brain metastasis
35:42and for proof of principle we've
35:44been testing some of these compounds.
35:46In collaboration with Eli Lilly,
35:48so we obtain beta secretase
35:51inhibitors in the diet of the mice.
35:54So initially we injected a cancer cells
35:57Melanoma cells in these mice and gave
36:00them a better second base inhibitor
36:02in the food or controlled diet.
36:05And you can see how this reduces the
36:07number of brain metastasis in this model.
36:10This is a short term culture but
36:12also in the five one Melanoma cells.
36:15Now of course this is more a prophylactic.
36:17Model because treatment starts
36:18at the time of injection,
36:20so we raise the bar a little bit
36:23by allowing the cells to establish
36:25metastasis first and then after 21
36:28days we gave doxycycline to the food
36:30and the in the water of the mice
36:33to activate docs inducible SH RNA.
36:36And again we saw that in this context,
36:38even when the treatment is initiated,
36:39once metastasis have been formed,
36:41we can see a reduction in the number
36:43of brain metastases and if we do
36:45the same thing with the Secretary.
36:47Keep in touch. We can also see again
36:50after initiating the treatment.
36:52Once the micrometer studies have been formed,
36:55we can see a reduction in brain
36:57metastasis we are currently trying to.
37:02Moving to well before I go into that,
37:05we are now testing the antibodies against
37:10Emily Beta either as a monotherapy
37:12or in combination with immunotherapy
37:14to see if we can recapitulate
37:17the same effects of surf here.
37:19So to summarize this part of the talk,
37:21we have shown the proteomic studies
37:24have revealed a novel connection
37:27between brain metastasis and
37:29neurodegenerative pathologies.
37:30This has now been confirmed by other studies.
37:32We collaborated with a group of men,
37:34iser, last year.
37:35We got a study done in single cell
37:39analysis of primate testis that also
37:43recapitulated these these finding
37:45that Melanoma cells mimic the.
37:49A neuronal pathways in another of the
37:53alterations that are seen in you know,
37:55degenerative disorders once
37:56they reach the brain.
37:58We see that Amelia Beta is particularly
38:01required for pre metastasis and
38:03not other sites of metastasis is
38:05acquired for steps that happen after
38:07extravasation and early survival
38:09in the brain parenchyma and among
38:12the multiple functions of family.
38:14But in this context we have seen that it
38:16triggers an anti-inflammatory response
38:18in the astrocytes and suppresses.
38:20Your inflammation, as I mentioned,
38:23we are also studying now whether
38:26these effects of family beta can
38:28be seen also in other models.
38:29So we have done a spatial
38:31transcriptomic analysis.
38:32In this case,
38:33it's not a Melanoma model,
38:34is the 41 model which is a breast
38:37cancer triple negative model
38:39that also colonizes the brain
38:41after intracardiac injection.
38:43These are on the top is a is a brain,
38:47is half of a brain of a sham
38:49mouse and this is the.
38:50Mouse,
38:51this is a mouse that was injected
38:53with the 41 cells and what I want
38:55to bring to your attention is that
38:58in this the next vision analysis we
39:01can see that the gfap positive cells
39:04around the areas where the tumors are,
39:07you have to probably take me you know
39:09take my my word these are the areas
39:11where the tumors are circle here
39:14and we see a special expression or
39:16increased expression of GFP around
39:18those tumor cells we also see.
39:20This one is 100 which has been seen
39:22it used in tumor cells in the brain
39:25environment in other cancer types.
39:27And interestingly we find these
39:29signature that we have that has
39:32been previously reported as the an
39:35Alzheimer's associated microglia
39:37signature that is a combination
39:39of 10 markers and is again seen
39:43particularly activated around the
39:46brain metastasis here in here.
39:49So it seems like perhaps these Alzheimer
39:52like response in the microglia genes
39:54around the brain metastasis cells
39:56could be a more general finding and
39:58not only characteristic of Melanoma.
39:59As I mentioned,
40:00we are now very excited by studying
40:03whether app genetic and pharmacological
40:06inhibition extends to other cancer
40:09types and whether the combination of
40:12beta secretase inhibitors or antibodies
40:14can work alone or in combination
40:17with checkpoint checkpoint locate.
40:19In immunocompetent models,
40:20I was hoping to tell you about a
40:24another very exciting story in the lab,
40:27but I I see the clock and we are reaching
40:29the 1:00 PM. So I will stop here and.
40:35This is this data I wanted to present
40:38but perhaps at the second occasion
40:40and just wanna thank all the members
40:43of the lab for their contributions.
40:46This work that I presented was
40:48mostly done by the first part,
40:50which are Veronica and Claudia in
40:53NRF 2 and Ali and Maya in CDP one.
40:57And Kevin Kleffman led the I'm a
41:02little better story and I want to thank
41:04of course all our funding sources.
41:06And thank you all for your attention
41:09and your patience with the technical
41:11issues at the beginning.
41:12I'll stop here and take any questions.
41:14Thank you.
41:17Alright, thank you so much, Eva.
41:19I'm going to ask folks to
41:21put questions in the chat.
41:22We only got a couple of minutes,
41:24but while people type in the questions,
41:26maybe I don't know if Marcus
41:28has any or I can ask one.
41:30Go ahead Marcus, I've got plenty.
41:34So either really Congrats also on the,
41:37you know the nature communications paper.
41:39And I was wondering with the NR2F2 story,
41:42you're probably aware of Chris,
41:44Marines work in Nature last fall on TCF 4.
41:49And I'm kind of wondering if you think
41:53your NR 2F2 is upstream of TCF four if
41:55you've seen any role there and there's
41:57probably going to be a subsequent
41:59story about resistance to therapies,
42:01so there's.
42:02Multiple parts to this question.
42:05A, the upstream downstream part,
42:06but then also the heterogeneity
42:08because you're talking about these
42:10things happening sort of uniformly.
42:12But I'm imagining that the epigenetic
42:14regulation is kind of cell by cell
42:16and that there's going to be a
42:17population of cells that have more
42:19or less of that along the way.
42:22All excellent questions and and
42:24these are all the questions that
42:25we are trying to address now.
42:26So,
42:26so we didn't see a total overlap
42:29between our population of N2F2
42:32with Chris Marines population.
42:35It seems like his population is smaller
42:39one inner two seems to be more broadly.
42:43I expressed in both EMT and also partially
42:47in the neural Crest like population.
42:49So his population seems to be a more,
42:51I wouldn't say minority,
42:52but it's a smaller population
42:54and not not directly overlap.
42:58You also ask the other challenge that
43:01we have is that a lot of the single
43:04cell analysis that have allowed us to.
43:06Distinguish this.
43:07Transcriptionist states don't have,
43:10don't allow us, don't don't have the
43:12death to look into isoforms, right?
43:14So you really have to have a different
43:17library preparation and pipeline to
43:19identify the different types of forms.
43:22So when people look at R2,
43:24they are just examining an R2 ISO one,
43:27the full length and the one that
43:29is really switching from the
43:32the primary to the metastatic,
43:34the one that is really triggering the EMT.
43:36Like program is only isoform
43:392 because of this.
43:41Balance between home and headliners.
43:45Yeah. So we we're definitely
43:46trying to understand what is the
43:48error key between these pathways,
43:49where is upstream, it seems to be
43:51one of the critical regulators.
43:54So again when we do this.
43:57Um, analysis is sending analysis,
43:59but still we don't know which
44:00one is upstream of which, right?
44:02Yeah, that's a very good question.
44:04And then you ask.
44:05The dynamics right of this
44:07of this population.
44:09So we are now in the process
44:11of labeling endogenously this
44:12isoform to be able to trace them.
44:14We want to understand whether
44:17these isoforms are, you know,
44:20dynamically regulated during the
44:22metastatic process and we have
44:24different labeling systems now
44:25that allow us to sell to monitor
44:28single cells during metastasis.
44:30So I hope we will have soon an
44:31answer for that, but we don't know.
44:33I I also dissipate that it would change.
44:36From the primary tumor as the
44:38cells switch from the more
44:40proliferative to the more invasive,
44:42and then back when they arrive
44:43to them at the static side.
44:45But still we don't have
44:46full evidence for that.
44:51So, but there were a couple of
44:53questions in the chat both relating to
44:55Alzheimer's disease and brain metastases.
44:58One is there an increased incidence
45:00in brain metastases amount Alzheimer's
45:02disease patients and the other one was
45:05whether a postmortem you see Alzheimer's
45:08clogs in patients with brain metastases.
45:11Good questions.
45:12So we haven't seen brain metastasis,
45:15we haven't seen plaques in brain metastasis.
45:18We look for for them,
45:19we don't think that they get to accumulate,
45:22so we don't, we don't think that the
45:25processing is wrong is to simply induced.
45:28So we see more soluble abeta being produced,
45:31but we don't see oligomers and we don't
45:33see plaques. We did that staining.
45:35That conversion is done by pathologists
45:37to to look at the plaques and we couldn't
45:39see either in our models nor in human.
45:41Examples and that's the reason why we
45:45think that the antibodies that we wanna
45:47try are antibodies that design against
45:49the soluble and libetta and not against
45:51the plaques which are the ones that
45:53have been now or the oligomers which
45:55have been now approved by the FDA.
45:57Now you ask the other question,
45:59the incidence in in Alzheimer's.
46:01So we'll look into that and we didn't see
46:06a epidemiological studies and association
46:09between Alzheimer's and brain metastasis.
46:13But normally remember that in the majority
46:15of the neurodegenerative disorders,
46:16there is an inverse correlation
46:19between cancer incidence and
46:21neurodegenerative disorders,
46:22particularly Alzheimer's and Parkinson's.
46:25And even though there is
46:26not a positive correlation,
46:27there is no negative association.
46:29Now there is a reported association
46:32between Parkinson's and and
46:34Melanoma brain metastases.
46:36And I mean I think Harry,
46:39this is skeptical or not, I see you.
46:43No, there's a slight increase,
46:45but anyway, I don't know that it's brain
46:47metastases specifically, it's just.
46:50Come on. I'm sorry. Sorry.
46:51Yeah, yeah, yeah. I misspoke.
46:52I I meant that. I know my kid.
46:53Right.
46:54Yeah.
46:55Yeah.
46:56No,
46:56but there's a really interesting
46:58observation there.
46:59So thank you and thank you so much
47:01for this amazing presentation.
47:02I have more questions,
47:04but I'm going to e-mail them to you.
47:06I don't believe there anymore in the chat.
47:08We appreciate your patience with all
47:09the technical challenges and thanks
47:11for virtually visiting next time.
47:12It'll be in person.
47:14Thanks for wonderful talk to
47:16really fascinating work.
47:17Thank you,
47:18Harriet.
47:18Looking forward to see you
47:20soon and thanks everybody.
47:22Bye.