The p53 Tumor Suppressor Pathway

April 09, 2021

Information

Yale Cancer Center Grand Rounds | April 6, 2021

Guillermina (Gigi) Lozano, PhD

ID6404

To CiteDCA Citation Guide

00:00Of hosting today and we are incredibly
00:04thrilled and delighted that we
00:06were able to induce Doctor Gigi
00:09Lissonota to be our speaker today.
00:12Doctor Lozano graduated from Pan
00:14American University in Texas,
00:16did some work at Oakridge Laboratories,
00:19got her PhD at Rutgers,
00:22and then went to Princeton,
00:24where she was a postdoc with Arnie Levine.
00:28Immediately following that,
00:29she was recruited to MD Anderson,
00:32and in her long tenure there she has done.
00:38Path breaking work on P53.
00:41She's recognized for enormous contributions
00:43that include the recognition that P.
00:4653 works as a transcriptional activator.
00:49Many, many findings regarding the ways
00:52that MDM two and MDM three regulate P 53.
00:57Extensive use of transgenic models
00:59to understand these mechanisms.
01:01Better definition of specific P53
01:04loss and gain of function mutations
01:07and and their effects on P53's
01:10biology as well as studies of
01:14the P53 transcriptional program.
01:16She's a member of the National Academy
01:18of Sciences and the National Academy
01:20of Medicine and at MD Anderson.
01:23She is now chair of the Department
01:25of Genetics and Hubert L.
01:27Olive Stringer,
01:28distinguished chair in oncology.
01:30In the division of Basic Science
01:32Research at MD Anderson,
01:34so I don't want to talk for a long
01:36time because I'm really eager
01:38to hear what you have to say.
01:40So thank you very much for joining us
01:43and I will just say to the audience.
01:45You can use the chat function to
01:48enter your questions or we can.
01:49I believe we can unmute people
01:51at the end if we need to.
01:54So with that do welcome.
01:56Thank you, thank you very much,
01:59Barbara. For that introduction I
02:01hopefully now I'm sharing my screen.
02:07Can you see it?
02:10Weekend OK super Alright so
02:12anyway thank you Barbara.
02:14It's it's fun to always visit a place
02:17and now during covid we're doing these
02:19virtual seminars but hopefully at the
02:22end I'll be able to address some of
02:25the questions that people might have.
02:27What I thought I would do today is
02:30described some of the models that
02:32I think have to find some of the
02:35basic real basic understanding of
02:37the P53 tumor suppressor pathway.
02:39So I'll get started with my disclosures.
02:42I am on the Scientific Advisory
02:45Board for PMV Pharma.
02:47So the P 53 pathway.
02:49This is my myopic view of the pathway
02:51and it really highlights.
02:54I think some of the critical
02:56features of the pathway.
02:58First and foremost people decree is present
03:01in very low levels in normal cells,
03:03but any kind of abnormality
03:06that the South senses,
03:07hypoxia, DNA damage,
03:09inappropriate activation of an
03:11oncogene will stabilize that
03:12P53 protein and then P53 in turn
03:15functions as a transcription factor.
03:17To activate hundreds of genes
03:19and I'll show you some of those
03:22experiments in in a little bit,
03:24but some of the genes that P.
03:2653 is known to activate is P21 for example,
03:30which is a inhibitor of the South cycle.
03:33It also induces senescence program
03:35P53 can activate a slew of gene sick,
03:38initiate a pop ptosis and it also activates.
03:41Genes are involved in changing
03:43the metabolic functions of a cell.
03:47Now when P 53 is activated,
03:49if the cell is allowed to survive
03:52and proceed,
03:53P53 has to activate this protein
03:56called MDM 2,
03:57which is an E3 ubiquitin ligase
04:00that targets P53 for degradation
04:02and basically removes the Peachtree
04:04levels back down to normal.
04:07So if you just.
04:09Think about what people think we can do.
04:11We can do so much in getting arrested,
04:14phallic it, killed itself,
04:15and it can also induce its own
04:18inhibitor to allow the cell to survive.
04:21And even though there's so much
04:22work in the pizza tree killed,
04:25we still don't understand all of the
04:27cues that determine which of these
04:29pathways Peabody Creek and activate.
04:31But because it has all these functions,
04:34it is a critical tumor suppressor,
04:36and it is the most doctor
04:38mutated gene human cancers.
04:39So what I showed here is what the
04:42field called the Manhattan Plot,
04:45and it was developed by Magali Olivia.
04:47So across this axis are 125 genes
04:50that are commonly mutated and cancers
04:52and then across this axis here are
04:5536 different types of cancers,
04:57and there's some features that stand out.
05:00But this is the one I want to
05:03highlight here across the board.
05:05These are mutations in the P53
05:08tumor suppressor.
05:08So almost all cancers mutate.
05:1153 but P 53 pathway is inactivated by
05:15multiple mechanisms and I show here
05:18in some of the different cancers and
05:21how they inactivate piece of debris.
05:23So high grade,
05:25serious ovarian carcinomas,
05:26mutations in P53 are the most common.
05:29However,
05:29in liposarcomas it is upregulation of MDM 2,
05:33the P fifty 383 if it could be like a San.
05:37About 100% of these liposarcomas inactivate
05:40the pathway by overexpressing MD.
05:43In the new Asia glioblastoma,
05:45big Sweet come interested in glioblastomas.
05:48Recently, the P 53 gene is altered,
05:51deleted in approximately 1/3 of
05:53glioblastomas in the MDM to an Indian.
05:57For genes are upregulated in
05:59about it order and this is a
06:02mutually exclusive relationship.
06:04So if P53 is mutant,
06:06MDM 2 doesn't have to be upregulated.
06:10And the other thing I want to point
06:13out in glioblastomas is that we have
06:15about half of these tumors that
06:17have neither mutations and piece
06:19of D3 or upregulation of the MDM.
06:22Two and MDM four inhibitors of P53.
06:24And so since I really believe that
06:26the P 53 pathway has to be undermined
06:29in the development of all cancers,
06:31I think that there's a big hole here
06:34that we have to understand in more detail.
06:37So today's start is really going to
06:39concentrate on just a few proteins,
06:41and I'm not going to have time
06:43to show you a whole lot of data,
06:46but I thought I would use this slide
06:49to highlight some of the important.
06:51Functions that I of these proteins
06:54that I will discuss with you today.
06:56So MDM two is an inhibitor of P53.
06:59It's an E3 ubiquitin ligase and
07:02target speakers for degradation.
07:03MDM Four also inhibits P 53.
07:07It doesn't have any E3 ligase function,
07:10but it actually facilitates and makes
07:12MDM two or better yet riveting ligase,
07:15although it also has independent
07:17functions of MDM two and can
07:20actually bind and inhibit the pizza
07:23guy free transactivation domain.
07:25This relationship does MDM 2 New Four
07:28also conform hetero dimer and that
07:30header dimer is critical in embryo
07:33development to inhibit P 53 and then,
07:35as I indicated,
07:37and will discuss in some detail,
07:39P53 can activate the Indian two
07:42promoter so it can up regulate
07:44MDM two in inhibited so levels.
07:47Another important concept that
07:48I'll mention very briefly,
07:50maybe at the end is that MDM
07:53two can also inhibit.
07:55A mutant P53 protein and that's because
07:58these mutant proteins have mutations
08:00in the DNA binding domain but retain
08:02a transcriptional activation domain.
08:05But the important point that I want to
08:08make here is even though this mutant
08:12can be targeted by MD M2 and MD M4.
08:15It is mutant in so can no longer
08:18feed back it up.
08:20Regulate MDM two so with time in our
08:22in vivo models we find that these
08:24people period this can become stabilized
08:27'cause there's insufficient MDM two
08:29to down modulate the protein levels
08:31and in a few minutes you'll see how
08:35important it down modulation is.
08:37So the outline of my talk today is
08:39I'm going to talk about some of the
08:42models that showed us how exquisitely
08:44sensitive P 53 is to inhibition via video 2.
08:49To tell you what the molecular
08:51responses to people three activation
08:54in vivo and last but not least,
08:57I'm going to describe some of the new
09:00cancer models that were working with
09:03that expressed mutant P53 proteins.
09:05So let's first talk about the MDM
09:08proteins and their innovation P.
09:1053. So along time ago and now we.
09:15Attempted to make an Indian to know
09:17Mouse and it's just not possible
09:20and the reason is not possible is
09:22because an MDM two null embryo just
09:25prior to implantation is APOP totic.
09:28This embryo stained with the
09:30tunnel essay in every salad.
09:32This embryo is is dead.
09:35And at the time we knew that
09:38MDM two interacted with P.
09:3953, but we really didn't know how
09:43important interaction was and what
09:45we did was test the importance
09:47of P53 in this little embryo by
09:50crossing 2P53 miles and we completely
09:52rescue this phenotype.
09:54These Meister born and are
09:57perfectly normal except now because
09:59they lack P 53 they have it.
10:02So with with this experiment indicates
10:04is that what MDM two is doing in
10:08these embryos is upregulating P53,
10:10which is preventing the normal
10:12development of these embryos.
10:14MDM fours are related MDM,
10:16two protein that aren't Johansson discovered.
10:19And since MDM two has such a unique
10:23relationship with P53 we decided that
10:25we would make the MDM for knockout
10:28but weren't sure what to expect.
10:32But in MD for knockout is also embryo
10:35lethal a few days after the Indian
10:37to know but again that phenotype is
10:40rescued by deletion of P53 and we've
10:43also made mice that have no MDM to know
10:47Indian for Nokia 53 and these mice.
10:51The viable they have Pizza 3 two or
10:54phenotypes because they lack P 53.
10:56So at least physiologically,
10:58in the mouse,
10:59the two most important functions of
11:02these empty in proteins is to keep
11:05P 53 levels low during homeostasis.
11:07OK,
11:07so we have these MDM two heterozygous
11:10in these MDM for headers I customized.
11:12They have only one allele of
11:14each of these two genes.
11:16In these mice are perfectly
11:18normal and running around,
11:19but as I indicated in my
11:22in my introductory slide.
11:26P 53 is a DNA damage response protein,
11:29and so we wondered if there was
11:31any phenotypes in these mice.
11:33If we irradiated them,
11:35if we damaged their DNA, and sure enough,
11:38we saw a beautiful phenotype.
11:39So the MDM two in the MDM,
11:42two Ambien for headers agasse mice are
11:44sensitive to low dose ionizing radiation.
11:47So in this experiment what we've
11:49done is irradiated mice with six
11:51Gray and the black line above.
11:53Here is a normal mouse.
11:56That for 50 days just ignores.
12:00Six grade radiation, but the headers.
12:02I guess.
12:03Mice MDM four in Red and MDM two and
12:06blue are dead by three weeks of age.
12:10Importantly,
12:10if we now move P53 from this system,
12:14we completely rescue these phenotypes.
12:15So this is the rescue.
12:17The Indium MDM, two heterozygous mouse.
12:20And here's the rescue of the
12:22Indian for headers ecospace.
12:24So even though.
12:25The MDM two and MDM for heterozygotes
12:29mice have sufficient levels of the Zambian
12:32proteins to maintain homeostasis with damage.
12:36There's just not enough of these proteins
12:40to return degree back to normal levels.
12:44The next experiment that I want to
12:47tell you about is the importance
12:49of this feedback loop.
12:51So as I indicated,
12:52MDM two is regulated by P53.
12:55There are two peaks decree
12:57binding sites in the P2 promoter.
12:59There are people,
13:00three dependent,
13:01so people free byansi sequences
13:03and activates MDM 2.
13:05And So what we decided to do is to
13:07ask how important was this feedback
13:10loop in regulating P53 levers?
13:12OK,
13:12so we made point mutations and I show
13:15here the different point mutations
13:17because we didn't want to disrupt
13:19the architecture of the promoter.
13:21We just wanted to disrupt MD PhD degree
13:24binding to the end game to promote it.
13:28So these experiments in the bottom bar
13:30chip data that basically show that
13:33P53 cannot find this mutant promoter,
13:36which we call P2P2,
13:37and this is different promoter that
13:40shows our assays are working in P53 combined.
13:43The pull up remote.
13:45OK,
13:45so we made these point mutations
13:48and we were surprised that are
13:50most was perfectly normal.
13:52We really thought that this
13:54feedback loop is going to be
13:56critical for regulation of P53.
13:59The mice are fine, but again,
14:01as in the previous case,
14:03their exquisitely sensitive to radiation,
14:05so this is the same experiment that I
14:09showed you before we rated it the we
14:13irradiated the mice with six grade.
14:16Normal mice.
14:18MDM two heterozygotes.
14:20Don't care about this dose,
14:22but you can see that the P2P2 homozygous
14:26mice that no longer have this feedback
14:29loop are dead for the most part.
14:32So what is the phenotype?
14:34These animals are actually
14:36dying because of the complete
14:38ablation of the ball mirror,
14:40so here's a heterozygous irradiated mice
14:43and you can see that at 12 days you have
14:47some disruption of bone marrow function,
14:50but it's still viable,
14:52whereas in the mice that lack
14:55the feedback loop,
14:56there's a complete ablation
14:58of Humana Pelisses,
14:59and this is again a P53 dependent.
15:03Process so if we take out
15:06just one illegal appeal 53.
15:08We completely rescued this phenotype so
15:11we can rescue this phenotype with complete
15:14deletion of P53 or header zygosity 5053.
15:17So with these experiments are beginning
15:20to tell us is that there's this.
15:24There is this important relationship
15:26between MDM two and P53,
15:28and then there's an important
15:30balance that has to be maintained
15:32for survival after DNA damage.
15:34The last experiment that I'll show
15:37you here is our attempts to try to
15:40understand which pathway downstream
15:42of P53 is important for this phenotype.
15:46So I've already told you that heterozygosity
15:49Peachtree rescues the phenotype.
15:51We also generated mice with deletion of P.
15:5421, which is the cell cycle inhibitor
15:57and that had no effect on the phenotype.
16:00Those mice are also very
16:02sensitive to radiation,
16:03and then we also deleted Puma,
16:06which is one of the APOP
16:08totic targets of P53.
16:10And here you can see that
16:12there was a complete rescue
16:13of this hematopoietic defects.
16:16So in this scenario it appears
16:18that it is the APOP totic function
16:21opekta degree that is killing.
16:23These hematopoietic stem cells.
16:27So those are just a couple of
16:30the numerous experiments being
16:32done to evaluate the relationship
16:35between MDM24 and P53,
16:37and it's just an exquisite relationship.
16:40You need sufficient MDM 2 for survival.
16:45And.
16:47And if you have too much and you
16:50too MDM for deletion of 53 deletion
16:53of downstream effectors of P53 can
16:55rescue those lethal phenotypes.
16:58Now come. Because.
17:00MDM two is so lethal early.
17:05During embryo Genesis,
17:06one of the experiments we wanted
17:09to do is really ask about an adult
17:12mouse and how important is MDM two in
17:16different tissues at different times,
17:19and we've used,
17:20we've generated this conditional
17:22allele of MDM two and this is using
17:26The LOX P system so these two lacks
17:29besides encompass two of the accents
17:32that code for the major finding.
17:35Region 2P53 so this conditional
17:37Leo then allows us to delete MDM
17:40two in any tissue that we want to.
17:42And we've generated number of
17:44tissues and experiments that way.
17:46But what I want to show you is
17:49what happens when you globally
17:51remove MDM two in the whole months.
17:55So as the title says,
17:56Indian too lost in the adult is
17:59results in a lethal phenotype.
18:01So what we've done here is we've
18:03used a Cree transgene that is
18:05tamoxifen inducible so this is
18:07this is a mouse that has one of
18:10the conditional alleles and it
18:12has the other legalism missing,
18:14so it's single recombination event is
18:16going to create an M2 normals or not sell in.
18:20All we do is inject tamoxifen and then
18:23we look at what happens to these.
18:25Phenotypes,
18:26and I think you can see from this
18:29graph here on the right that within 5
18:31days of treating the mice with tamoxifen,
18:34they're all dead.
18:36So,
18:36so losing MDM two and it's a
18:41Peachtree dependent process.
18:44Causes this enormous physiological response.
18:49These are some of the pathologies we
18:53see in these mice the hippocampus has.
18:57Less number of cells, the retina is is
19:00comprised of multiple beautiful layers
19:02and you can see that it all of these cases
19:06that it's it's a decreased cell numbers.
19:09There's some differences in in the liver,
19:12and it's actually if you
19:14measure liver function.
19:15Liver function is compromised
19:17when you have deletions of MDM.
19:19Two kidney has all these protein
19:22casts and dilated tubules,
19:23and then in the spleen we have
19:27complete absence of white.
19:30And as I indicated,
19:31these phenotypes are all P53 dependent.
19:34So it just.
19:35I mean, I think the important
19:37aspect of this slide is that in
19:40some of the previous slide is it
19:42highlights some of the pathologies
19:44that we might see when we use MDM.
19:47Two inhibitors in the clinic to it.
19:51To inhibit the tumors that have
19:54high levels of empathy into it.
19:56And the hematopoietic defense
19:58is actually observed in humans.
20:00It is treated with MDM two inhibitors.
20:04Uhm?
20:04But now I want to use this model system
20:07to understand what piece of degree
20:09is doing in these different tissues.
20:13In you know one reason for doing
20:15this is we would like to be
20:18able to reactivate P53 somehow,
20:20and it's kind of hard people.
20:22A lot of people are trying to
20:25to reactivate people to agree,
20:27but what we were hoping is that
20:29we might be able to identify
20:32downstream pathways to P53.
20:33There would be better targets
20:36for reactivation tours.
20:37So let me show you what we did.
20:41So again we used our MDM,
20:44two conditional mouse and we
20:46deleted MDM two in the adult mouse.
20:49But we did this acutely and we actually
20:52chose a 24 time our time point to
20:55ask what P53 targets are regulated
20:58in different issues that lead to
21:01these pathologies in the adults.
21:05OK,
21:06so this is now all the different
21:08issues that we initially looked
21:10at and what I'm showing you here
21:13is the percent recombination.
21:15So once we treat with tamoxifen,
21:17we induce recombination of the locus.
21:21And you can see that the pancreas,
21:23the heart being tested,
21:25had the highest level of recombination.
21:27Again,
21:27I will point out that we only have one MDM,
21:31two allele.
21:32The other allele is an allele,
21:34so single recombination event
21:36will activate P 53.
21:38And then on this axis we chose to
21:40look at P53 activation by measuring
21:42the expression of P 21 which is
21:45encodes a cell cycle inhibitor.
21:48So you can see in this experiment
21:50at the kidney,
21:51the pancreas in the intestine
21:54where the tissues that expressed
21:56the highest levels of P.
21:5721.
21:58And we were thinking the highest
22:02levels of of P53.
22:04So we've we've looked at these mice,
22:06so in 24 hours we see no.
22:09So let me back up just for one second,
22:12I hope I can do that,
22:14so that issues that we
22:15decided to look at where
22:17the kidney, the pancreas,
22:18the intestine in the heart.
22:21And the ovary in the ovary.
22:23Just because P 53 is mutant in
22:26ovarian carcinoma is at 95%.
22:3095% frequency and so we wondered
22:32if we might be able to to begin to
22:37understand that mutation frequency.
22:39So for the heart for the ovary,
22:42we saw absolutely no phenotypes after
22:4624 hours post deletion of MDM 2.
22:49In the intestine we saw a
22:52fascinating phenotype which
22:54is descript dropout phenotype.
22:57In yellow here,
22:58I've outlined the ***** of the
23:00intestine and one of the phenotypes is
23:03the complete absence of of the crypt.
23:06Sydney in the lab is quantified
23:08the number of ***** in these
23:11different animals and you can
23:13see the mice that have no MDM.
23:152 have about little more than
23:18little but half of the number of
23:20***** is in normal control mouse.
23:24The kidney also had some phenotypes
23:26at 24 hours and it had twice
23:29the number of protein casts,
23:31so you can see here.
23:32So this is an early phenotype in the kidney.
23:36Again,
23:36this is the normal kidney control experiment.
23:39And then the pancreas had to be a
23:43fascinating phenotype which will
23:44delve into a little bit more deeply.
23:47But we saw in the pancreas is this?
23:52Acinar to ductal metaplasia,
23:54so here's a normal pancreas in the top,
23:57and here's what the pancreas looks like
24:00in the in the animals that have no MDM.
24:042 Here we've stained with keratin 19,
24:07which is a marker for a ductal cell.
24:11And here we've measured the Metaplastic
24:14area and we also see an immune
24:17infiltration in these in these mice.
24:20So within 24 hours we saw this huge
24:23plasticity in the pancreas from you know,
24:27this acinar to ductal metaplasia.
24:30OK,
24:30so we've taken these five
24:32tissues an we've done.
24:34We've looked for expression of P53 targets,
24:37so on the left here I show you all of
24:40the genes that were regulated in these
24:43different issues in our RNA seek data.
24:46The dark region is the region that
24:49is upregulated in the lighter color,
24:51shows the regions that were downregulated
24:54in each of these five tissues.
24:57On the right here I show the percent
25:00of these these dysregulated genes
25:02that are actually P53 targets.
25:04They have a pizza degree binding
25:07site and we used data from the
25:10literature to identify these tools.
25:12With P53 binding sites.
25:14So in the intestine,
25:15for example,
25:16I think that number is 69% of the
25:20channels had pizza pre binding sites.
25:22So the the most.
25:24Most of the genes dysregulated
25:26it in the system by deletion
25:29of MDM 2RP53 targets the ****.
25:32** the other hand had a
25:35huge physiological response,
25:36600 for jeans that were dysregulated
25:39but only 16% repeated different targets.
25:42So what we're capturing here at
25:4524 hours is not just you know
25:48activation of P53 and P3 targets,
25:50But the downstream responds to
25:54that P53 activation.
25:56So this is now compilation of all
25:58those five different tissues to
26:00examine the overlap in P53 target genes.
26:03And as you can see from this figure
26:06on the left there were only 7 jewels
26:10that were commonly regulated.
26:12By MDM 2 lost that repeat 53 targets.
26:16So for example here in the pink is we
26:19have 206 genes, 135 of the P53 targets
26:23are specific to the pancreas and seven
26:26were shared with the other four tissues.
26:29So these seven jeans are MDM two cycling,
26:32G1 MDM two as we as I mentioned,
26:36the very beginning is regulated
26:38by P53 and we expect it well.
26:41We didn't expect them to be able to,
26:45but it's not a surprising result because
26:48the promoter is intact in MDM 2.
26:52Three of these genes segment you one GST,
26:55one piece art, one or cell cycle regulators
26:59to these jeans are transcription factors,
27:02and this gene EDA 2R herself directions.
27:05So these are the six Peachtree
27:08targets that have a common.
27:10They represent the common signature of of
27:13upregulated genes in in these three tissues.
27:16We wanted to validate the signature
27:19to make sure that they were.
27:22Truly, a P53 targets physiologically.
27:24And So what we did is we did our our DNA
27:28damage ionizing radiation experiment.
27:31We treated the whole animal
27:33with ionizing radiation,
27:34and here's the data for two of the jeans,
27:38and we've done it for all
27:41of seven cycling G1 E DA2R.
27:43Here's the wild type levels
27:45of expression of these genes.
27:47If we irradiate,
27:49you can see that these genes are upregulated.
27:52In in both cases,
27:55and if we irradiate a P53 null,
27:58you see you see no up regulation.
28:01So these are P3 target genes that are
28:04being upregulated following punishing
28:06radiation. So these experiments.
28:10Highlight this incredible repertoire.
28:13Transcriptional targets that P53
28:16physiologically regulates the vivo
28:18and I think it also suggests that
28:21maybe these specific targets can
28:23be used to understand in vivo.
28:25If you have.
28:27If you can reactivate piece of D3
28:30or convert mutant and wild type,
28:33these might be great markers to
28:37look at for activation 53.
28:40OK,
28:41I want to now just briefly discuss
28:44this this encrypted hypothesis.
28:47Acinar ductal hyperplasia that
28:49we see within 24 hours in the.
28:53In the pancreas.
28:55So one of the other experiments that
28:59Sidney Moyer in my lab did is we
29:02obtained these mice MST one missed one.
29:06Cree are OK,
29:07so missed one is an SNR
29:09specific transcription factor.
29:11An upstream of the missed one promoter.
29:17There is a criar transgene,
29:19which means that you can express create
29:21only in the *** and ourselves with the.
29:25Pancreas and so Sydney worked out the
29:29tamoxifen conditions that gave you
29:32a similar percent recombination as
29:35our previous experiments with MDM.
29:38Two position in the entire pancreas.
29:41OK, so similar percentage of recombination
29:44and similar activation of P3 targets.
29:48So here we use two of our targets,
29:52EADE, DA2RG, STT, SC1.
29:55To measure people to the activation
29:58and you can see in both my sweet.
30:01MDM two deletion happens in the
30:03home pancreas or mice where it
30:05only happens in the acinar cells.
30:07You have similar activation of
30:09these three of these targets,
30:10so we felt we could do.
30:14We could actually compare deletion
30:16of MDM two in the whole pancreas.
30:18The deletion of MDM two jestoni
30:21Essen ourselves.
30:22And we have absolutely no phenotype.
30:25So these pancreas look completely normal.
30:29Here we're measuring just we're looking at.
30:32I mean,
30:33I ageny sections in here in the right.
30:37We're measuring the immune
30:40component and these these pancreas.
30:42These pancreatic perfectly normal. So.
30:46The take home message here is that.
30:49This esnard ductal hyperplasia that
30:52we see is a P53 specific hyperplasia.
30:56But it's it's arising from signals
30:59outside of the acinar cells.
31:02So to me,
31:03this is a fascinating experiment
31:07because no one's ever noted that.
31:11That the environment can can affect
31:14the pizza delivery response,
31:16and so we'll be delving into understanding
31:21this phenotype a little bit better.
31:25Pancreas is is one of the tumors
31:29with 7075% mutations in P53 an and
31:32it always has this very compromised
31:35stromal component and so maybe by
31:39understanding what P 53 is doing is
31:43physiologically important Organism,
31:45we might be able to impact our
31:49understanding of Peter mutations in
31:53pancreatic cancer. OK, so the let out.
31:56OK so I'm just going to check my
31:59Clock to see how much time I'm doing.
32:02Well, OK,
32:03so I've shown you a lot of data
32:05where we deleted MDM two and an
32:08I didn't show you data for MDM 4
32:11but you you see these people three
32:13dependent physiological phenotypes
32:14and that's all fine and good.
32:16It showed us how important this
32:18relationship is between these proteins but.
32:21What happens in human cancers is
32:23you've got high expressions of
32:25MDM two and this is just yes,
32:27mean Valentina Vega in the lab
32:29a number of years ago looked at
32:32Indian 2 levels in head and neck
32:34squamous carcinomas and these are
32:36some of her beautiful pictures.
32:38So here's MDM,
32:39two expressed a very highly in a small
32:42region of this squamous cell carcinoma here.
32:456 expressed almost across the
32:47entire tissue and then here is
32:49an interesting example of MDM.
32:51To be expressed in the cytoplasm,
32:53not the nucleus.
32:54So we really don't understand
32:56what it's doing in the cytoplasm,
32:59but not in all three of these experiments, P.
33:0253 is 1 type OK,
33:04so I think with this experiment
33:07in many others that
33:09people have done. Again.
33:13Again, show that what MDM two is doing in
33:17these tissues is inhibiting P53 activity.
33:20Now the I also don't want to leave
33:23you with the notion that MDM too.
33:27The P53 is the only MDM to target.
33:31Physiologically is the most
33:33relevant target because of the cell
33:36lethal phenotypes that we see,
33:38but in several experiments that my lab
33:42and Carol previous labs have done is,
33:45we've tried to overexpress MDM two in
33:48normal cells to understand what it's
33:51actually doing in with regards to
33:54transformation and tumor evolution.
33:56So here's what happens.
33:58So this is a normal control and the
34:02left these are mouse cells express a
34:05normal number of mouse chroma zones,
34:08and when we overexpressed MDM two
34:11we see this incredibly abnormal.
34:14Chromosome instability we can quantify
34:17the numbers of fusions here and
34:20we have a huge number of fusions.
34:22We also have a lot of fragments.
34:27So in data from multiple labs,
34:29if you overexpress MDM two in a
34:32normal cell the cell just dies.
34:35It can continue to grow.
34:37So and you know some experiments
34:39that are ongoing in the lab is OK,
34:42trying to understand.
34:44OK,
34:44so if we can't overexpress MDM
34:47two in a normal cell?
34:49Why do tumors have very high levels
34:51of ambient 2IN in one idea that we're
34:54working with is that there are some
34:57other rotation in those tumors that
34:59allows those tumors to survive with
35:02high levels of MDM two and so if we
35:05could understand what else MDM two is doing,
35:08we might be able to obtain a.
35:12A window of vulnerability to try to get the
35:16MDM two overexpressing cells to implode.
35:20But the screens that we're doing
35:23currently are are ongoing.
35:25OK then for the last few minutes
35:29of my lecture I want to.
35:32Move over to our understanding of P53
35:36mutations in breast cancer models.
35:39Didn't tell you earlier, but pizza degree.
35:42Why did tell you the people
35:45limitations for the most common,
35:47but really it speak into three
35:50missense mutations that are is the
35:53most common type of genetic lesion,
35:55and so my lab and that of Tyler left.
35:59Tyler Jacks is lab have made germline
36:02mutations in P53IN animal models and
36:05we show that these mice are tumor pro.
36:08But more importantly in contrast.
36:11Green or mice?
36:12These mice have a high metastatic capability,
36:15so this here is our data from the
36:18172 mutation corresponds to the
36:20origin 175 mutation,
36:21which is one of the hot spot
36:24mutations in human cancers.
36:26And here you can see a metastasis
36:28to the liver,
36:29and here stained with the P53 antibody,
36:32a metastasis to the brain.
36:35And this is in contrast to mice
36:38that have deletions of 353,
36:40so this really was the first example
36:43that suggested that expressing
36:44a mutant P53 was much more.
36:49Much more aggressive than not having
36:51people to create and and in the field.
36:53We call this a gain of function.
36:55Mutant P 53 is doing something in these
36:58cells to make them highly metastatic.
37:01So these are germline mice and what we
37:04wanted to do is to generate semantic
37:07models because the these germline models
37:10represent Lee from Many syndrome which is
37:14an inheritance of people to mutations.
37:16But that's a rare syndrome and we really
37:20wanted to understand this metastatic
37:23phenotype in a system where the the
37:26specific cell type has a Peach limitation
37:29and surrounding normal environment.
37:31To feel yourself to catch neutral
37:34goes to T cells are all wild type
37:37for P53 so that mouse did not exist.
37:39Tyler Jacks made a beautiful mouse that
37:42has been used extensively in the literature
37:44that basically is heterozygous for P53.
37:47So the entire mouse is missing 1P53 allele
37:50and Natalie all can be converted to a
37:52mutant P53 in a tissue specific fashion.
37:55So we didn't think that that was.
37:59Adequate enough to study the tumor,
38:01stroma tumor immune interactions because
38:03of heterozygosity of the P53 locus.
38:05So let me tell you a little bit
38:08about how we generate these mice and
38:12what our breast tumor phenotype is.
38:14OK, so here's how we generated these alleles.
38:18So we call these WM allele for
38:21wild type to mute P 53.
38:23So I'll show you in a minute this
38:25is a wild type allele normally and
38:28it's wild type because we earned
38:31started seeding a sequence upstream
38:33of the point mutation and this is
38:37this example is the Argentine,
38:39once it imitation which
38:40we generated previously.
38:42So what you have is a wild type P50.
38:45Three years of pollination site and
38:47then in a cream immediate fashion
38:50you can remove the wild type C DNA
38:53and basically reconstruct the locus.
38:55The arm and Cindy to Locust that we
38:58had studied forever in the germ line.
39:01With these mice and because it took a
39:04very long time to make these animals,
39:07we actually decided to make a second
39:09hotspot mutation and that's the
39:11Argentine 2.5 to double mutation,
39:13which corresponds to the 248
39:15hotspot mutation.
39:16Uhm?
39:19So this is just showing me wild type
39:21initially and then we committed fashion.
39:23You make them into protein and you can
39:25make it in akhri dependente manner.
39:28So this just shows you how normal
39:30those mice are.
39:32So here we're comparing wild type 2
39:35heterozygous mice with the 172 or the 2.5.
39:37We stabilized the mutant protein
39:39in response to DNA damage.
39:41And when we look at the activation
39:44of the three targets be 21 in Puma,
39:47they were activated to similar levels.
39:51We looked at the ability of these.
39:54Of DNA damage to induce labor ptosis?
39:57No difference between these
39:58two alleles in wild type mice.
40:00And again for the ability to rest the
40:03cycle in mouse embryo fibroblasts,
40:06there's no difference,
40:07so these mutant alleles really
40:09represent these condition alleles.
40:11That column you Tilly's really
40:13represent wild type allele.
40:15They can become a moot.
40:18And this is the last experiment I'll
40:20show you about the the actual alleles.
40:23What we've done is what we've
40:25done here is compared the 172
40:27heterozygous mice at 245 headers.
40:29I guess mice and wild type mice to
40:32each other over more than two years
40:34to look at the tumor phenotypes.
40:37Mice is the age.
40:38Just like people will sporadically
40:40get tumors, but what you can see is
40:42that there's absolutely no statistical
40:44difference between the three alleles.
40:46So for all practical purposes.
40:48This new allele that we generated
40:51expresses a wild type P53 protein.
40:54OK, so let me tell you bout two experiments.
40:58One is our semantic breast semantic
41:02model and what we did is we did
41:06use the R172H and the 245 alleles.
41:10We injected Adna virus Cree into the duct
41:14of the mammary gland and so and then.
41:17In addition, we use this TV tomato
41:20allele which is also create dependent.
41:22So when we inject adeno Korea not
41:25show you a picture in a minute.
41:28We basically make a mutant P
41:3153 and we label the cell red.
41:35So here is the.
41:38Control experiments to show that we
41:41do get recombination when we inject
41:44the cream expressing adeno virus here
41:46on the left is a low titer injection
41:49and you can see the red cells here.
41:521 to 5% of the ductal cells become infected.
41:57And then here on the right you can see.
42:00A high titer virus was used was
42:03injected into this gland in about 50 to
42:0570% of the ductal cells are checked.
42:09I also want to note that these
42:12mice are 50% balzi.
42:13Normally the P53 field is used to
42:16see 57 black, 6 string to study.
42:21Peabody create tumor phenotypes,
42:23but that strain is resistant to
42:26breast cancers for some reason,
42:28and the beltsy component brings in
42:31more sensitivity to breast cancers,
42:34and we really don't know the
42:37genetic reasons for this.
42:40OK, so here is the data.
42:42Let me go through it in detail.
42:45So this is the 172 mutation expressed
42:48only in a few mammary glands.
42:50Memory cells, low titer.
42:52We didn't see any tumors.
42:55High titer,
42:56we actually didn't see any tumors.
42:59This one tumor showed up.
43:01Post that to your end time
43:04point of the experiment.
43:07Because we weren't sure we were going
43:09to get any tumors by just making
43:12people scream Mutant in a few cells.
43:15We also.
43:17Irradiated with the sub lethal
43:19dose of radiation.
43:21This is not lethal to the mouse,
43:24but it does cause damage and if the
43:27damage is in the right or wrong genes
43:30that contributes to tumor phenotype.
43:33So with a low titer we begin to see tumors.
43:37We sell one at a higher titer.
43:40We now solve 4 tumors and
43:43one tumor metastasized.
43:45The 2.5 allele was a much stronger,
43:47had a much stronger tumor,
43:49phenotype with low titer.
43:51We saw four tumors and one of them was
43:55meta static with the high tier tighter.
43:58We saw nine tumors, so this is about
44:0175% and more than half were meta static.
44:04So let me kind of summarize all the data
44:07that we've done with these animals.
44:10So first let's just look at
44:13the tumor incidence. The R 172.
44:15Each allele all by itself in in.
44:17In these experiments, we only use
44:20one copy because we didn't want to.
44:24We didn't want to do LOH and
44:26we didn't want to.
44:28We just wanted to figure out what
44:30would happen with the minimal
44:32number of alterations.
44:34So if you compare the ones need two hitters
44:37Vegas to the 2.5 low and high titer,
44:40there's a huge number of tumors in the 248.
44:44Nice if we irradiated the ones only two,
44:47we got increased tumor incidence and
44:49then this is the experiment where
44:51we did mutate the other allele.
44:54So existing law supporters I got
44:57city and we can see.
44:59Increased tumor phenotype with the high dose.
45:02And this is a comparison on in the
45:04middle panel of the metastatic phenotype,
45:06and again the 245 documentation
45:08was the most metastatic.
45:10And then we looked at lots of header
45:14zygosity so the 245 mutation had.
45:18Variations in terms of LOH.
45:20About 50% of the mice showed LOH
45:23and then others retained some or
45:26all of the P3 alleles.
45:28The irradiated are once again 28.
45:31Showed 100%. LOH so to us.
45:35While we don't understand why we
45:38see 100% outrage with this allele,
45:41would it says?
45:42Is it that wild type allele is
45:44very strong at inhibiting tumors
45:47in this winsome need to background?
45:52These are the breast tumor subtypes that
45:55we saw the irradiated once in need two
45:58we saw mostly triple negative breast.
46:01There are 172 with the
46:03that lacks wild type P53.
46:05We sell mostly luminal B and then here
46:09with the 2.5 mutation we saw all three.
46:14Molecular subtypes to me this is fascinating
46:17experiment because we've made one mutation,
46:20we made a P53 mutation and yet here in this
46:24sample with the 2:45 we see off the tumor,
46:28molecular subtypes evolving and so one of
46:31the experiments that we're doing now is
46:35trying to understand with this 245 mutation,
46:38what are the triggers to
46:40these different subtypes?
46:41Triple negative breast cancer
46:43is very hard to treat.
46:46But for example,
46:47here to enrich tumors you can,
46:50you can treat with her two antibodies,
46:53so we're trying to understand
46:55basically the tumor evolution
46:56that initiates with this one, P.
46:5953 missense mutation.
47:02We've also wanted to so the data
47:05I just showed you says that
47:08the 248 mutation is much more.
47:11Dramatic.
47:12Then the 175 mutation.
47:14So what I showed here is a comparison
47:17of ovarian lung and breast tumors
47:20from people and just looking at the
47:23kind of mutation that they have and
47:26you can see in people that the 248
47:29mutation is has the worst outcomes.
47:33We couldn't do these data just
47:35for breast because the number of
47:38samples out there was not enough
47:40to give us significance.
47:43OK,
47:43so the one of the last experiments
47:46I'll show you here is just trying to
47:49understand tumor evolution because
47:52we made a semantic model that
47:54develop different kinds of breast
47:57cancers that were highly metastatic.
47:59And so I'm really interested in
48:01understanding the task sees in an in
48:05vivo physiologically relevant system.
48:07So we did.
48:08Here is we took the 2:45 mutant
48:11animals and we took.
48:13UH-22 memory tumors from these mice
48:15we sequenced them in three different
48:17regions trying to understand a
48:19little bit about the heterogeneity,
48:21and then we sequenced three
48:24metastases from each of these tumors.
48:26OK,
48:27and we sequence them to an
48:29incredible debt depth.
48:31So what we have here then is the
48:34comparison of the primary to the
48:37metastases and if we just look at mouse #4,
48:42there is some overlap in these these
48:45this overlap is considered an early
48:48gene signature and then you see this.
48:52Slate sequences that come up,
48:54which is how,
48:56which is what,
48:57where the metastases is now evolving
48:59when it gets into its metastatic site,
49:03which in this case was the lung.
49:06And so we can compare the early
49:09mutations of all three.
49:11The task season you see early
49:13mutations in both experiments,
49:15so these are the mutations that
49:17were acquired in the primary and
49:19metastatic lesion at the same time.
49:22But when you look at late mutations,
49:24here's all three mutations for mouse #4
49:27there's only one late mutation common.
49:30So with these sequencing data,
49:33indicate is that these matasa
49:35fees left the tumor very early
49:38during the metastatic process and
49:41then seated and had
49:43additional changes. So this was the
49:47first suggestion that maybe metastasis.
49:50Breast cancer metastasis driven
49:52by a new P53 is an early event.
49:56So to summarize, this model
49:58just briefly, we can make.
50:00P 3 point mutation in just a few
50:04cells that become a tumor that.
50:08Interessee migrate,
50:10proliferate and develop these metastases.
50:14Where we now have I called it a
50:16little factory but we just have
50:18these mice now developing tumors.
50:20We can isolate the circulating tumor cells,
50:22so we're trying to do is understand that
50:26the changes that occur for these cells
50:29to be able to survive in the blood.
50:32To home into an organ and then to develop.
50:35OK, if I have a few minutes which
50:37I have just a few minutes,
50:40I'm going to tell you about the
50:42other model that we made because we
50:45weren't sure that making a people
50:47communication in just a few cells was
50:49going to give us a tumor phenotype.
50:52So here we use K14 create,
50:54which expresses then a mutant P53 in all
50:56of the epithelial cells of the mammary gland,
51:00and this is a model that develops.
51:02Triple negative breast
51:04cancer 100% of the time.
51:06So again this.
51:07This suggests that you know South normal
51:10cell tumor cell interactions are altering
51:13the kinds of tumors that come up.
51:16And in this scenario we also had a
51:19cast 9 allele that is Creed dependent,
51:23so K14 CRV and allows us to make a
51:26mutant people degree in just a few
51:29in the epithelium of the mammary
51:32gland and to express castanon so we
51:35can use CRISPR technologies to to
51:39begin to address vulnerabilities.
51:41OK,
51:41so this vulnerability that we examine
51:44this in this model was whether these
51:48tumors were addicted to having immune to P53.
51:52OK,
51:53so here is the use that adnot
51:57associated virus that expresses
51:59a guide RNA that will delete P.
52:0253.
52:03So the tumor burden before treatment
52:06with a V in the control and the mute.
52:11Depleted tumors with similar,
52:13but you can see here in in the Purple line
52:16that those tumors that had depletion of
52:18you piece of degree survived much longer.
52:21This is just a picture of the tumor
52:24phenotypes that control you can see that.
52:27It's very.
52:28It's obviously a tumor.
52:30These mice die very quickly post
52:33identification of the tumors,
52:34and then you can see here with the
52:38depletion immunity theory this this
52:41this gland is looking more normal.
52:44OK,
52:44this is a whole bunch of data for the
52:48individual mice that this is a tumor volume.
52:52The controls and green here.
52:54Once we identify the tumor,
52:56they just keep growing the experimental
52:58cohort here in purple are animals
53:01that have recombined have basically
53:03deleted that mute people free protein
53:06in these mice live much longer.
53:08On the right is a tumor volume,
53:11so you can see that depletion of
53:14P53 affects the tumor volume.
53:17Greatly and then I'll just point
53:20out these two samples in Orange,
53:22which appeared not to respond
53:24to depletion of Mutant P
53:2753, and when we look at
53:29these two samples in detail,
53:31they did not express a stable mutant P.
53:3453, and so we think that
53:37these two tumors are actually.
53:40Since the P 53 isn't stable,
53:42the evolution that's occurring
53:44in these tumors is due to the
53:47absence of the P53 protein and
53:49not to a gain of function.
53:52OK, so I'm going to stop there
53:54and just I thrown a lot of data
53:57at you so I wanted to summarize
53:59briefly the first set of experiments
54:02really captured the exquisite
54:04sentence sensitivity of the MDM,
54:06two protein and P53 activity.
54:08We were able to.
54:12Identify this the molecular response
54:13to P53 activation in vivo identified
54:15numerous targets that are tissue specific.
54:18What are they all doing?
54:20I think we have our work cut out
54:22for us 'cause there's no way I can
54:25delete every one of these targets
54:27and see how important they are in
54:30vivo and then last but not least,
54:33I showed you the generation of of
54:36this novel mouse model that really
54:38allows us now to make a semantic
54:40point potential 53 in.
54:42Any cell of origin that we want
54:45to an in in the breast model.
54:49Specifically,
54:49we had a highly metastatic phenotype that
54:53we're trying to understand in more detail.
54:56OK,
54:56so my last slide is just the
54:59numerous people in the lab that
55:02have contributed to the studies.
55:05The Vinodh Pant did the feedback loop
55:08studies Roberto Navy MTM Tunicate Johnny DMD.
55:12Or not,
55:13God Tamera did all the studies
55:15with the radiation and with the
55:18node to look at the P3 response,
55:22Sidney generated the conditional mouse to
55:24look at the acute activation of people.
55:28Three targets you in terrific postdoc
55:30in the lab now has her own independent
55:34position generated that conditional
55:36mood P53 allele and Donata is the
55:40one who's studying the addiction.
55:42We're really wondering what the
55:44mechanisms is acquire these
55:46tumors imploding when they no
55:48longer have communities industry.
55:50So with that, I'll end in, oh,
55:53I'm glad to answer any questions.
55:56Thank
55:57you so much.
55:58That was a wonderful talk.
56:00I'm going to ask people to put
56:03questions in the in the chat,
56:05but I I wanted to ask you, sort of,
56:09from from the therapeutics perspective,
56:11people have been very interested in
56:13compounds like Prima and Cody that
56:16assist with re re folding of P53 with
56:18disruptive mutation and yet clinically
56:21those have been a little bit disappointing.
56:25Is is much known about whether or not
56:29those refolded P 53's are better,
56:32worse the same as as substrates for the MDM?
56:36Two MDM four? Yeah, so,
56:38so we've done a few studies using some of
56:42the drugs that are available, not many.
56:47My lab is focused on the genetics
56:49because if we take out him to an MP4 we
56:53see people three different phenotypes,
56:55but you know it's very different.
56:58Genetic told us mechanisms,
56:59but the drugs are really as you indicated.
57:03They're going to tell us
57:05whether they work or not.
57:07So I agree with you, I don't think
57:10that drugs are working very well.
57:13And you know,
57:15I don't know enough about those
57:17experiments to know you know how
57:19often you know how often the drug
57:21with the level of activation.
57:23I think that our.
57:26Our.
57:28That that that 7 gene signature that
57:31we identified would really help in
57:34those studies to try to understand
57:36what is the pika degree response?
57:39I also think that the people
57:42agree response required.
57:44Will vary in different tissues.
57:49We just know from our MDM two studies that
57:51some tissues are just much more sensitive
57:54to increase P 53 levels versus others.
57:57So I think that there's just a whole lot
58:00more work to be in to do in the clinic to
58:03be able to to understand that response.
58:07Yeah, yeah, I I think having a common set of.
58:11I mean I think people have just looked
58:13at P 21 over and over again and it's
58:16probably very inadequate, right?
58:18I see a couple questions and
58:20we have like a minute left.
58:22So first Jeff Townsend wants to know.
58:25Whether or not you've considered
58:28looking at sequencing of much larger
58:30cohorts of tumors and multi sample
58:33datasets to understand the temporal
58:36order of mutation appearances,
58:38that's exactly what
58:39we're doing right now I have a postdoc in
58:43a graduate student who just generated.
58:47A cohort of 100. Nice 'cause we want
58:49to understand the sequence of events.
58:51We want to understand the
58:52different events that lead to the
58:54different molecular subtypes.
58:55So we are in the midst of those
58:58experiments and we're going to do RNA
59:00seek and in DNA sequencing to understand
59:02both what happens if Arnie level.
59:04I think that it's critical we need
59:06to understand what's happening
59:07at the DNA level because I think
59:10that that's what gives rise to the
59:12different molecular subtypes.
59:13But I think it's the expression
59:15that's really going to tell us.
59:17What's happening to those cells once
59:20they reach home to deliver the line so?
59:24So we got all those are in progress.
59:28Then, Karen Anderson, who's my Co.
59:31Host for having invited you,
59:33wants to ask what your thoughts
59:35might be for a pro TEC directed
59:38against MDM two try to grade that
59:41as a therapeutic strategy. I
59:43think there's two Protex we
59:45should be thinking about.
59:46One is MDM 2. Although.
59:50Somehow I think for an MDM two
59:53inhibitor I think you've got to,
59:56you've got to target it
59:58to the to the tumor cell.
01:00:00Better just because of the
01:00:03hematopoetic toxic cities that have
01:00:05been seen with the MTM 2 inhibitors.
01:00:08But I also think we should
01:00:10start thinking about potentially
01:00:12doing degrading mutant P.
01:00:1353 Taking it out of the picture
01:00:16to see to see what happens.
01:00:20Particularly with those gain
01:00:22of function mutations,
01:00:24exactly exactly.
01:00:28We are over there more questions but
01:00:31we are over the time so I want to be
01:00:34respectful of very appreciative of of you,
01:00:36having joined us today and if
01:00:38anybody wants to ask me a question,
01:00:41email that they should feel free to I think
01:00:43allowed to answer additional questions.
01:00:48Super very much.