Metastasis as a Hereditary Disease Regulated by the Nervous System

Transcript

• 00:01Alright. I think that we
• 00:02will get
• 00:03started.
• 00:05So,
• 00:06it is my great pleasure,
• 00:09to be able to introduce,
• 00:11Doctor. Sohail Tavazoy
• 00:13from Rockefeller University who will
• 00:15be giving our distinguished lecture
• 00:16today.
• 00:18Thank you everyone for coming.
• 00:21So, as I said, Doctor.
• 00:23Tavazoy is currently at Rockefeller,
• 00:25where he is Leon Hess
• 00:27professor and also the head
• 00:28of the Meyer Laboratory of
• 00:30Systems Cancer Biology.
• 00:32And in addition, he directs
• 00:34the Black Center for Metastasis
• 00:36Research,
• 00:37at Rockefeller and
• 00:39still keeps a foot in
• 00:40the clinical world across the
• 00:41street, at Sloan Kettering.
• 00:44I think that
• 00:45what we're what we have
• 00:48really looked for in this
• 00:49particular series
• 00:51are,
• 00:52scientists and physician scientists
• 00:54who really bridge fundamental biology
• 00:57translation
• 00:58and
• 00:59are always keeping in mind
• 01:00the clinical needs and looking
• 01:02towards clinical impacts. And I
• 01:04think that
• 01:05that really,
• 01:06describes,
• 01:07the research and likely what
• 01:08we're going to hear about,
• 01:10today.
• 01:11I think,
• 01:13as a basic scientist, one
• 01:14of the things that I
• 01:15really appreciate,
• 01:17is,
• 01:19is is in particular the
• 01:21combination
• 01:21of approaches that doctor Tavazoa's
• 01:23lab has used,
• 01:25whether that's genetics,
• 01:26whether that's kind of systems
• 01:28of biology approaches,
• 01:30and really discovering
• 01:31fundamental mechanisms.
• 01:33I know that Joan is
• 01:34here. So if you're not
• 01:35familiar with the work, in
• 01:36particular work around,
• 01:38you know, kind of classic
• 01:39undergrad,
• 01:40biochemistry
• 01:41in terms of tRNA biology,
• 01:44really uncovering this entirely
• 01:46amazing world where tRNA is
• 01:48trying to be really important
• 01:49for affecting,
• 01:51a translation. Right? And so
• 01:52as much as we love
• 01:53our single cell genomics,
• 01:55right, a lot of things
• 01:56are actually governed in terms
• 01:57of how cells adapt
• 01:59at post transcriptional levels. And
• 02:01I think there's some really
• 02:02cool insights that have come
• 02:03out of the work.
• 02:05So I won't go on
• 02:06and on. I'll just say
• 02:07that, doctor Tavazoa has a
• 02:08recipient of many awards including
• 02:10the DP two, the innovator
• 02:12award from NIH, the Pershing
• 02:14Square
• 02:15Sun Cancer Prize, is an
• 02:16outstanding investigator of NCI,
• 02:19and a DOD Era of
• 02:20Hope Award. And he's been
• 02:21elected to the National Academy
• 02:22of Medicine and was the
• 02:23past president of ASCI. So
• 02:25really, again, covering all all
• 02:27the disciplines of basic, translational,
• 02:29and clinical sciences.
• 02:31Thank you so much for
• 02:32taking the time to visit
• 02:33us today, and we're looking
• 02:34forward to your presentation.
• 02:38Thank you so much, Megan.
• 02:39Of course.
• 02:41Yeah. I got you. Oh,
• 02:42okay.
• 02:46Thank you. Oh, thank you
• 02:47so much.
• 02:49I'm I thank you so
• 02:50much, Megan, for that, wonderful
• 02:51introduction. I'm really honored to
• 02:53be here and, to be,
• 02:55giving this lecture and to
• 02:56be,
• 02:57meeting,
• 02:59old friends and and meeting
• 03:00new friends and learning about
• 03:02the phenomenal science that you
• 03:03do,
• 03:04that's going on at Yale.
• 03:06And I know that for
• 03:07many of us, it's a
• 03:09difficult time,
• 03:11for science.
• 03:12And,
• 03:13I think,
• 03:15what we we need to
• 03:17realize the import as, you
• 03:18know, I was trained as
• 03:19a physician,
• 03:21and I think we all
• 03:22need to recognize the importance
• 03:23of basic science because my
• 03:25work would not be
• 03:27would be much, much less,
• 03:30relevant,
• 03:31if it wasn't for,
• 03:34the,
• 03:35PhD scientists and the basic
• 03:37science interactions,
• 03:38that I've, been been really
• 03:40privileged to experience. And
• 03:42and so,
• 03:44the other thing I wanna
• 03:44just say is that for
• 03:45the young trainees,
• 03:48you know,
• 03:49there have been when I
• 03:50first started at Rockefeller, I
• 03:51was in an elevator and
• 03:53I was complaining about getting
• 03:54my, you know,
• 03:56second or third r o
• 03:57one submission rejected in an
• 03:59elevator.
• 04:00And and, I was there
• 04:01was a scientist
• 04:03older scientist in the elevator,
• 04:04and he said, how's it
• 04:05going? I said,
• 04:06it sucks. It's so hard.
• 04:08You know? It's getting grants
• 04:09and doing this stuff. And
• 04:11he just turned to me,
• 04:11and he said,
• 04:13it it's always been hard.
• 04:15And he he was an
• 04:16he was a seventy eight
• 04:18year old scientist. His name
• 04:19is Emil Gottschliff. He's a
• 04:20Lasker
• 04:21prize winner. And what he
• 04:22did was,
• 04:24he he he developed
• 04:26meningococcal
• 04:27vaccine, so fifty years earlier.
• 04:29So I think it it
• 04:30it's also means that it's
• 04:31always been hard.
• 04:32Things get harder and easier,
• 04:34but there's increasing opportunities. So
• 04:36for the young trainees, I
• 04:38know it's challenging, but,
• 04:40it's incredible time to do
• 04:42science, to technologies that are
• 04:44available to you, the myriad
• 04:45of technologies and the opportunities
• 04:46to do things we could
• 04:47never imagine ten years ago.
• 04:50You need to kind of
• 04:51balance that with some of
• 04:52the,
• 04:53cynicism that's around with respect
• 04:55to science funding. So thank
• 04:56you for inviting me here.
• 04:59So our laboratory studies, the
• 05:01biology of metastatic progression.
• 05:04We're we're interested in how
• 05:05cells that have spread to
• 05:07distant organs
• 05:08are able to
• 05:09form metastatic colonies. These are
• 05:12rare cells that are are
• 05:13able to actually get to
• 05:15the distant organs, and we're
• 05:17interested in identifying the genes
• 05:19that are responsible for this
• 05:21critical transition,
• 05:22understand the mechanisms,
• 05:24and and the processes involved.
• 05:26And the goal is perhaps
• 05:28this understanding can give rise
• 05:29to
• 05:30therapies that might, be able
• 05:32to prevent this transition
• 05:34and thus potentially achieve cures.
• 05:37And,
• 05:38the
• 05:39the possibility exists that the
• 05:41pathways that these cells use
• 05:43to get to this next
• 05:44level, these cells may continue
• 05:46to be, addicted to these
• 05:48pathways. And perhaps if we
• 05:50can inhibit these pathways, we
• 05:51can actually get a reversion
• 05:53in the other direction, and
• 05:54that's something that we've been
• 05:56interested in,
• 05:57for for a long time.
• 05:58And how do we approach
• 05:59this problem? Well,
• 06:01traditionally, in terms of cancer
• 06:02biology, we approach problems by
• 06:04thinking about,
• 06:05cancer evolution and somatic mutations.
• 06:08Peter Noll in the seventies
• 06:09posited the clonal evolution hypothesis.
• 06:12The idea is on the
• 06:13left, you have a normal
• 06:14cell that you get transformation.
• 06:16You have a certain number
• 06:17of chromosomes. And as a
• 06:18result of diversification,
• 06:22at at the primary tumor
• 06:23side, but also selection from
• 06:25the immune system or chemotherapies
• 06:27or targeted therapies,
• 06:29you,
• 06:30select for a subpopulation of
• 06:32cells that evolve
• 06:33to have,
• 06:34traits that are able to
• 06:36overcome these barriers,
• 06:38by virtue of having,
• 06:40amino acid changes, mutations,
• 06:43that allow them to do
• 06:44this. All of this is
• 06:46proven to be correct, and
• 06:47this is a major basis
• 06:48and foundation for cancer biology
• 06:51and target,
• 06:53target,
• 06:54discovery. But one of the
• 06:55things that he posited at
• 06:57the end was that as
• 06:57a result of this process,
• 06:58at the end, there would
• 07:00be rare cells that are
• 07:01able to,
• 07:03achieve a metastatic state by
• 07:05having some special metastatic
• 07:07mutation,
• 07:09amino acid change that then
• 07:10in a special gene that
• 07:12would make those cells metastatic.
• 07:15And, this is one thing
• 07:16that has not yet been
• 07:18borne out,
• 07:19and this is with a
• 07:20quote from Barbolacine is still
• 07:22this is a case. Despite
• 07:24intensive efforts, consistent genetic alterations
• 07:26that distinguish cancers that metastasize
• 07:28from cancers that have not
• 07:29yet metastasized remain to be
• 07:31identified. So hundreds of million
• 07:32dollars sequencing metastases in primary
• 07:34tumors looking for mutations that
• 07:36emerge at the metastatic site
• 07:37but not the primary site
• 07:39that are causal
• 07:40and recurrent. This has not
• 07:42been found. And so how
• 07:43do we approach this problem
• 07:44if we don't have those
• 07:46mutations to, as guiding,
• 07:48lights? And what we do
• 07:50is we use in vivo
• 07:51selection. This is an approach,
• 07:53in terms of metastasis that
• 07:54was pioneered by Josh Fiddler
• 07:56back in the seventies. We
• 07:57start with cancerous cells. These
• 07:59are a heterogeneous population of
• 08:00cells, thus diversification.
• 08:02We can introduce these into
• 08:03the circulation of mice, expose
• 08:05these cells to similar microenvironmental
• 08:07pressures they're used to during
• 08:09human metastasis.
• 08:10And then we allow time
• 08:11to pass and we select
• 08:12for rare cells that are
• 08:14able to form macroscopic
• 08:16colonies such as in the
• 08:17mouse below, and we compare
• 08:19that to mice which do
• 08:22not have macroscopic
• 08:23colonies. Those cells do not
• 08:25progress. And we perform molecular
• 08:27comparisons
• 08:28to identify differentially expressed genes,
• 08:30and then we perform causal,
• 08:32testing.
• 08:33And over the last fifteen
• 08:35years, phenomenal scientists in the
• 08:37lab, including Claudio who's here,
• 08:39have,
• 08:39looked to identify use this
• 08:42approach to identify genes that
• 08:44are involved in this process.
• 08:45And what we have found
• 08:46is that
• 08:48a critical feature of cells
• 08:49that are able to,
• 08:51achieve these gene expression states,
• 08:53extreme gene expression states, and
• 08:55form macroscopic
• 08:56colonies
• 08:57is modulation of specific small
• 08:58RNAs. And you guys, Joan
• 09:00Stites has pioneered this field.
• 09:02And and what we find
• 09:03is that,
• 09:04specific tissue specific micro small
• 09:06RNAs such as microRNAs
• 09:08can become deregulated,
• 09:10either depressed,
• 09:11repressed in a certain cancer
• 09:12or over expressed in another
• 09:13one.
• 09:15And that can give rise
• 09:17to this gene extreme gene
• 09:18expression states. And more recently,
• 09:20we found that other other
• 09:22classes of,
• 09:23noncoding RNAs, specifically
• 09:25transfer RNAs, can also become
• 09:27modulated. So in some cases,
• 09:29these rare cells can overexpress
• 09:31specific transfer RNAs.
• 09:33And in those cells,
• 09:35what we observe is enhanced
• 09:37translation
• 09:38of transcripts
• 09:39that are enriched in codons,
• 09:41cognate
• 09:42codons to those tRNAs. So
• 09:44we can get specific tRNAs
• 09:46becoming overexpressed,
• 09:48and this can lead to
• 09:49enhanced translation of specific transcripts
• 09:51that can enhance the fitness
• 09:53of these cells as they
• 09:54metastasize.
• 09:55And also fragments of these
• 09:57tRNAs can also act as
• 09:58small RNAs. And what we
• 09:59use is we use these
• 10:00small RNAs as molecular probes
• 10:03to look for downstream genes
• 10:05that are modulated during this
• 10:07state, and we try to
• 10:09place these genes in pathways
• 10:11to uncover processes that are
• 10:12critical for this, metastatic colonization.
• 10:16And when we started the
• 10:17lab, we we decided to
• 10:18look at common cancers,
• 10:20and and,
• 10:22breast cancer, colorectal cancer, and
• 10:24melanoma. And the idea was
• 10:26we would look look at
• 10:27the use these cancers because
• 10:29they represent distinct embryonic lineages.
• 10:31So mesodermal,
• 10:33endodermal, and ectodermal. And the
• 10:34idea was perhaps with time,
• 10:36we might find common genes,
• 10:40that might be involved in
• 10:41this.
• 10:42That didn't end up being
• 10:43the case, but we actually
• 10:44I'll tell you about common
• 10:46processes that, are involved,
• 10:49in metastasis. And so I'll
• 10:51start off by a story
• 10:52in melanoma, and this is
• 10:55I'll at the end, I'll
• 10:56talk about, why this provides
• 10:58a solution to why metastasis
• 10:59driver mutations have not been
• 11:01found. Nora Panchova, a graduate
• 11:03student in the lab, performed
• 11:04this in vivo selection in
• 11:06using human melanoma cells injected
• 11:08into mice that lack adaptive
• 11:10immunity.
• 11:11She was able to derive
• 11:12four highly metastatic sublines shown
• 11:15in magenta,
• 11:16and she performed microRNA profiling
• 11:18looking at over eight hundred
• 11:19microRNAs. And she identified two
• 11:21microRNAs
• 11:22that are recurrently overexpressed in
• 11:24these highly metastatic sublines relative
• 11:26to the poorly metastatic sublines.
• 11:28Then she asked, what do
• 11:29these myocardies do? She places
• 11:31them
• 11:32into melanoma cells.
• 11:34If we take the parental
• 11:36poorly metastatic melanoma cells, inject
• 11:38them into the circulation,
• 11:39we see rare metastatic colonies
• 11:41in the lungs. This is
• 11:43a histology section of a
• 11:44mouse.
• 11:45But if she overexpresses these
• 11:47microRNAs, it's sufficient to substantially
• 11:49enhance the metastatic capacity of
• 11:52these cells. Conversely, if she
• 11:54takes highly metastatic melanoma cells
• 11:56that overexpress these microRNAs
• 11:58and she inhibits these two
• 11:59microRNAs, she sees substantial
• 12:01reduction in metastatic colonization. So
• 12:03these microRNAs are sufficient and
• 12:05required for enhanced metastatic colonization
• 12:07capacity.
• 12:08What do these microRNAs do?
• 12:10She places these microRNAs into
• 12:12the cells or inhibits their
• 12:13activity,
• 12:14and she found that each
• 12:15microRNA negatively regulates a number
• 12:18of transcripts
• 12:19stability when she looked at
• 12:20transcript,
• 12:22transcriptomic analysis.
• 12:24But what was interesting is
• 12:25that they had a convergent
• 12:26target. Both micron is negatively
• 12:28regulate a common target, and
• 12:30that target is APOE,
• 12:32a gene that's been canonically
• 12:33implicated in lipid transport, and
• 12:35variants of this have been
• 12:36associated with Alzheimer's.
• 12:38And using mutagenesis
• 12:39studies and epistasis experiments, she
• 12:41showed that these two microRNAs
• 12:43target
• 12:44APOE,
• 12:46both in the three prime
• 12:47UTR and the coding sequence.
• 12:49These microRNAs promote metastasis in
• 12:51the tumor compartment,
• 12:52and they negatively regulate APOE.
• 12:54And thus, APOE is a
• 12:56two metastasis suppressor
• 12:58in melanoma cells.
• 13:01But APOE is not only
• 13:02produced by the melanoma cells,
• 13:04it's produced by the host.
• 13:05All of us have APOE
• 13:06produced in the liver.
• 13:08Also macrophages produce APOE.
• 13:10This plays a role in
• 13:11in in lipid transport between
• 13:13organs. And so we wanted
• 13:14to ask about the role
• 13:15of host APOE. So we
• 13:17switched to an immunocompetent melanoma
• 13:19model,
• 13:20to do genetics. And what
• 13:21you can see is if
• 13:22we inject melanoma cells, b
• 13:23sixteen f ten cells into
• 13:25these mice, we see macroscopic
• 13:27meta melanoma
• 13:28or metastases.
• 13:30But if we inject the
• 13:30same melanoma cells into mice
• 13:32that are genetically inactivated for
• 13:34APOE,
• 13:34we found about a tenfold
• 13:36enhancement in metastatic colonization.
• 13:38So what that tells us
• 13:39a few things. One is
• 13:40that host APOE is also
• 13:42a negative regulator of metastatic
• 13:45colonization.
• 13:46It tells us that it's
• 13:47extracellular APOE. Right? Because it's
• 13:49the host is producing it.
• 13:50The melanoma cells are injected,
• 13:52and the host APOE is
• 13:54really acting on the melanoma
• 13:56cells.
• 13:57And APOE is a secreted
• 13:58protein. It also tells us
• 14:00that in our in the
• 14:01circulation of these mice, there's
• 14:02an endogenous metastasis suppressive mechanism
• 14:05that that's operating within the
• 14:07circulation
• 14:08that that impairs the ability
• 14:09of these cells to form
• 14:10metastatic colonies.
• 14:12So taking what I told
• 14:13you before and now,
• 14:15what we find is that
• 14:17in the tumor compartment
• 14:18and the host are producing
• 14:20APOE that's suppressive for metastatic
• 14:22colonization.
• 14:23In the tumor compartment, microRNAs
• 14:25are selected to silence this
• 14:26access,
• 14:27but the host or germline
• 14:28APOE is always
• 14:30present and can continue to
• 14:32impair this process.
• 14:34But why why is APOE
• 14:35so critical? The what we
• 14:37found is that the
• 14:38effect of APOE is is
• 14:40really high. We see very
• 14:41high magnitude effect of this
• 14:43gene. And over the years,
• 14:45trainees in the lab have
• 14:46really answered this question, and
• 14:47it's because it's pleiotropic.
• 14:49That means it mediates multiple,
• 14:51events,
• 14:52and it acts to suppress
• 14:54multiple
• 14:55promenostatic
• 14:56phenotypes. So we found that
• 14:57APOE acts on,
• 14:59an apoE receptor called LRP
• 15:01one on the melanoma cells
• 15:02to suppress their ability to
• 15:04invade through tissues.
• 15:05ApoE acts on another apoE
• 15:07receptor called LRP eight to
• 15:09suppress the ability of melanoma
• 15:11cells to recruit endothelial cells
• 15:13into the primary tumor or
• 15:15in the metastatic site, and
• 15:16it acts on myeloid LRP
• 15:18receptors to modulate antitumor immunity.
• 15:21At the primary site, there's
• 15:23a selection for cells that
• 15:25silence APOE,
• 15:27and cells that overexpress these
• 15:28micron is are able to
• 15:29silence APOE that derepresses these
• 15:32phenotypes
• 15:32and leads to enhanced metastatic
• 15:34colonization.
• 15:36Let me tell you about
• 15:37this last,
• 15:38part this this third, phenotype.
• 15:41And so all of you
• 15:42know,
• 15:43and Marcus Bosenberg is here.
• 15:44He's he's a leader in
• 15:46this area that t cells
• 15:47release
• 15:48factors such as interferon gamma
• 15:50and granzyme to mediate antitumor
• 15:52immunity. It's how they kill
• 15:53cancer cells. But over for
• 15:55over a hundred years, we've
• 15:56known that in tumors,
• 15:58there is the presence of
• 16:00myeloid cells that have gone
• 16:02by variety of names that,
• 16:04that can be suppressive more
• 16:06recently, suppress
• 16:07characterized to be suppressive for
• 16:09anti tumor immunity.
• 16:10And these cells,
• 16:12are people typically call them
• 16:14myeloid derived suppressor cells. They
• 16:15come with various in different
• 16:17varieties.
• 16:18But these myeloid derived suppressor
• 16:20cells can
• 16:21can express immunosuppressive
• 16:23factors such as arginase,
• 16:25which can degrade arginine, which
• 16:27is critical for t cell
• 16:28metabolism and function.
• 16:30These MDSCs can also express
• 16:31PD L1. That's a target
• 16:32for immune checkpoint inhibition that
• 16:34can also suppress t cell
• 16:36responses.
• 16:37And we really,
• 16:38don't have a good sense
• 16:40of why these
• 16:41myeloid cells accumulate in tumors.
• 16:43We do know that their
• 16:44abundance associates with reduced responsiveness
• 16:46to checkpoint therapies, targeted therapies,
• 16:49chemotherapies.
• 16:50And what we found is
• 16:51that these cells express the
• 16:53LRP eight receptor.
• 16:54And via genetic and pharmacologic
• 16:56approaches, we found that APOE
• 16:58acting on LRP eight receptors
• 16:59on these cells reduces their
• 17:01survival,
• 17:02reduces their abundance, and this
• 17:04can lead to t cell
• 17:05activation.
• 17:07Thus, a basis for
• 17:08silencing of this access to,
• 17:12give rise to,
• 17:14persistence of these myeloid cells
• 17:15in the tumor compartment.
• 17:17And so what's fascinating about
• 17:19APOE is that, within us,
• 17:21we have substantial germline genetic
• 17:23variation in this gene. Okay?
• 17:26So Ben Osendorf,
• 17:28in the lab, who now
• 17:28heads his own lab at
• 17:30the Sherite,
• 17:31became interested in this.
• 17:34So majority of us have
• 17:36a cysteine at position one
• 17:37twelve and an arginine at
• 17:38position one fifty eight. This
• 17:40is the model of APOE.
• 17:41It's seven alpha helical alpha
• 17:43helical
• 17:44protein.
• 17:46But but,
• 17:47some of us have two
• 17:48arginines at position one twelve
• 17:50and one fifty eight. That
• 17:51those are people who carry
• 17:52the APOE four allele. That's
• 17:54associated with increased risk for
• 17:55Alzheimer's.
• 17:56Some of us have two
• 17:57cysteines at those two positions.
• 17:59That's protective for Alzheimer's. Okay?
• 18:02Now
• 18:03these slight amino acid changes
• 18:05from cysteine to arginine have
• 18:06major structural implications
• 18:08for the ability of APOE
• 18:10to bind receptors because they
• 18:12alter salt bridge formation between
• 18:14these these the cysteine and
• 18:16arginine residues that are present
• 18:18there. That can impact the
• 18:19ability of these proteins to
• 18:20bind receptors. So for example,
• 18:22the APOE two form of
• 18:24APOE
• 18:25exhibits reduced receptor binding,
• 18:27and the APOE four form
• 18:29exhibits enhanced receptor binding and
• 18:31enhanced,
• 18:32signaling of this receptor.
• 18:35And so what Ben was
• 18:36interested in was asking, look,
• 18:37Nora has already shown that
• 18:39host derived a poe is
• 18:41critical. It's metastasis suppressive
• 18:43by acting on the melanoma
• 18:45cells. Host derived a poe
• 18:47is encoded by the germline,
• 18:49and we have substantial germline
• 18:51variation amongst
• 18:53us. Could germline
• 18:54differences
• 18:55impact
• 18:56differentially impact
• 18:58melanoma progression outcomes? And so
• 19:00the experiment that was done
• 19:01was to take mice where
• 19:03the mouse
• 19:04APOE is replaced by the
• 19:05human APOE
• 19:07gene, and you either replace
• 19:08it with APOE two, APOE
• 19:10three, or APOE four. These
• 19:11mice are identical.
• 19:13They differ from one another
• 19:14by either one amino acid
• 19:15or two amino acids. When
• 19:17you go from e two
• 19:17to e three, one amino
• 19:18acid, e two to e
• 19:19four, two amino acids.
• 19:21And what if we inject
• 19:22the same melanoma cells into
• 19:23these mice? And so this
• 19:25is the experiment.
• 19:26In gray, you can see
• 19:27these are the young, melanoma
• 19:29lines that, that were kindly
• 19:31provided by Marcus Bosenberg for
• 19:32the entire community.
• 19:34And so in gray, you
• 19:35can see that these melanoma
• 19:36tumors grow
• 19:37at a certain rate.
• 19:39But if you implant the
• 19:40same melanoma cells into mice
• 19:42that are the APOE two
• 19:43genotype, the same melanoma cells,
• 19:45the tumors grow more faster.
• 19:47But if you implant the
• 19:48melanoma cells into mice of
• 19:49the e four genotype, the
• 19:51tumors grow slower.
• 19:53And this should make sense
• 19:54because, as I mentioned to
• 19:55you, if apoE is suppressive
• 19:57for melanoma progression,
• 20:00apoE two is hypomorphic
• 20:02in binding these receptors,
• 20:04and we see enhanced tumor
• 20:05growth. It's less able to
• 20:06suppress melanoma progression. E four
• 20:09is hypomorphic in binding the
• 20:10receptor,
• 20:11so it actually is more
• 20:12effective at, repressing tumor growth.
• 20:15How about metastasis?
• 20:16If we take these two
• 20:17alleles that are hypomorphic or
• 20:19hypomorphic
• 20:20and we we use melanoma
• 20:22cells, the b sixteen f
• 20:23ten, that colonize the lung
• 20:24well,
• 20:25we see in in again,
• 20:27differential effects where the e
• 20:28two mice
• 20:30are exhibiting higher,
• 20:32metastatic burden
• 20:34relative to the e four
• 20:35mice. Now this was kind
• 20:36of a major claim, and,
• 20:38a a student in the
• 20:39lab, Noma Adak, who wanted
• 20:41to test this using entire
• 20:43genetic approach approaches. So what
• 20:45I showed you was genetic
• 20:47with transplantation of melanoma cells,
• 20:49but what we did was
• 20:50now use Marcus Rosenberg's BRAF
• 20:52p ten tyrosinase CRE model
• 20:54where you're inducing activation
• 20:57of the most common
• 20:58oncogene BRAF v six hundred
• 21:00e and tumor suppressor p
• 21:01ten in melanocytes
• 21:03via the tyrosinase
• 21:04CRE driver,
• 21:05you can get melanomas. But
• 21:07what Noma did was to
• 21:08cross these into e two,
• 21:10e three, or e four
• 21:11human,
• 21:14lines.
• 21:15And what was really cool
• 21:17was in the e two
• 21:18model, now you start seeing
• 21:19metastases,
• 21:20and you can see that
• 21:21these are melanotic dark pigmented
• 21:23lesions.
• 21:24And when she did the
• 21:25analysis, she found some it
• 21:27was even more impressive than
• 21:28what Ben had found, that
• 21:30the e two mice all
• 21:31have more metastases than the
• 21:32e four mice.
• 21:33Four of the e four
• 21:34mice have no met discernible
• 21:35metastases.
• 21:37And so,
• 21:38again, the germline genetics is
• 21:41dictating,
• 21:42melanoma metastatic outcomes in these
• 21:44mice.
• 21:45And so
• 21:47so this has all been
• 21:48studies in mice, but we've
• 21:50been studying the human germline
• 21:52variance. So what what Ben
• 21:54did was now he was
• 21:56tasked to go look at
• 21:57the Cancer Genome Atlas where
• 21:59where we have,
• 22:01exome sequencing data on patients
• 22:03with high risk melanomas.
• 22:05And the idea was this,
• 22:07typically, we look at the
• 22:08somatic compartment as, cancer biologists
• 22:10and oncologists,
• 22:11but the idea was, okay,
• 22:12go in there and ignore
• 22:13the somatic. Just look at
• 22:15the germline. Look at the
• 22:16normal tissue, the adjacent tissue
• 22:18spit, whatever.
• 22:20And and let's look at
• 22:21patients that have stage two
• 22:22three melanomas that are risk
• 22:24for metastasis,
• 22:25and let's ask how do
• 22:26they do depending on their,
• 22:28germline encoded APOE status.
• 22:31And the finding was really
• 22:32remarkable. Again, this is publicly
• 22:34available. You can all go
• 22:35check this out and validate
• 22:36that it's true. In gray,
• 22:38you can see the survival
• 22:39of e three e threes.
• 22:40That's the majority of us,
• 22:42sixty percent of us.
• 22:44In in red, you see
• 22:46the people who
• 22:47carry one or two alleles
• 22:48of APOE two, the hypomorphic
• 22:50variant,
• 22:51and you can see their
• 22:52survival is worse.
• 22:54E four are the ones
• 22:55that are carrying a variant
• 22:57in the same gene, but
• 22:59it's hypermorphic.
• 23:00Okay? And their survival is
• 23:02better.
• 23:04And and we at the
• 23:05time when we published this,
• 23:06we act another dataset became
• 23:08online. This is the MD
• 23:10Anderson high risk melanoma cohort.
• 23:12And when he did the
• 23:13analysis, he saw the same
• 23:14thing. So these variants are
• 23:18recapitulating
• 23:20what we see in the
• 23:21mice models,
• 23:23and, you might say, well,
• 23:25you already told us that
• 23:26APOEs
• 23:27plays a role in Alzheimer's,
• 23:28but I wanna emphasize that
• 23:29this is the opposite.
• 23:31Right? E four, those are
• 23:33the patients, the blue lines,
• 23:35those are individuals that are
• 23:36at risk for Alzheimer's mortality
• 23:38and morbidity.
• 23:39They're surviving better with high
• 23:40risk melanoma.
• 23:42E two are the individuals
• 23:43that are protected against Alzheimer's.
• 23:45They actually have a longevity
• 23:47benefit. They tend to live
• 23:48about four years longer than
• 23:49the rest of us. They
• 23:50act but when it comes
• 23:51to metastatic melanoma,
• 23:53their survival is worse.
• 23:55And so,
• 23:56what this shows us is
• 23:57that,
• 23:59you know, common genetic hereditary
• 24:01variants are predicting future outcome
• 24:04in a cancer, in this
• 24:05case, high risk melanoma.
• 24:08And so in summary, for
• 24:09this first part, I've shown
• 24:10you that APOI is a
• 24:11critical metastasis suppressor gene, and
• 24:13it and and it we
• 24:14find that it
• 24:15inhibits multiple
• 24:17prometastatic phenotypes. This is just
• 24:19what we find. We believe
• 24:20there are more. We we
• 24:21can only see what we
• 24:22can see.
• 24:23But and and cells that,
• 24:25at the primary set are
• 24:27selected to silence APOE,
• 24:29by overexpressing
• 24:31these microRNAs, they repress these
• 24:32phenotypes,
• 24:33and they can progress to
• 24:35metastasis.
• 24:36But the germ line is
• 24:37also producing APOE, and that's
• 24:39fixed by
• 24:40by your genetics.
• 24:42Individuals who are born with
• 24:43a hypomorphic form of APOE,
• 24:45APOE two,
• 24:47that's akin to silenced APOE.
• 24:50They tend to have higher,
• 24:52worse metastatic outcomes.
• 24:54Individuals
• 24:55that are carriers for e
• 24:57four, the hypermorphic
• 24:58variant,
• 24:59they have improved survival, and
• 25:01we've shown causality in mice
• 25:02that this variant is suppressive
• 25:04for metastatic,
• 25:05melanoma.
• 25:06So we'll come back to
• 25:07Bert Vogelstein's point and the
• 25:09fact that we don't have
• 25:10metastases driver mutations. And what
• 25:12I would posit is that
• 25:13at least in melanoma,
• 25:16APOE germline variants,
• 25:18the hereditary variants in the
• 25:20germline are actually, we believe,
• 25:22metastases is regulatory variance. And
• 25:24I think as a field,
• 25:24we've been ignoring the germ
• 25:26line when it comes to,
• 25:28cancer progression outcomes.
• 25:30And that the potent metastatic
• 25:32potential of a cancer, in
• 25:33this case, high risk melanoma,
• 25:35proceeds in the initiation of
• 25:36the cancer, and it's present
• 25:38at at birth and can
• 25:39be inherited.
• 25:40And so
• 25:41we've been interested in ways
• 25:43to perhaps exploit this therapeutically.
• 25:46So many years ago, we
• 25:47thought about ways to increase
• 25:48APOE levels in the circulation.
• 25:50APOE is transcribed by a
• 25:52nuclear hormone receptor pair, LXR
• 25:54and RXR.
• 25:55The pharmaceutical industry had been
• 25:57developing
• 25:58agonists for liver x receptor
• 26:00about twenty years ago to
• 26:02activate another target gene of
• 26:04this pair that's involved in
• 26:06lipid homeostasis.
• 26:08That approach did not work
• 26:09out, and it was abandoned.
• 26:11But we obtained these compounds,
• 26:13these LXR agonists,
• 26:14treat our melanoma cells. We
• 26:16can increase APOE, and we
• 26:17can increase APOE in the
• 26:19host as well.
• 26:20And and, we can treat
• 26:22our mice with, this compound,
• 26:23l x r agonist, and
• 26:25it can have metastasis suppressive
• 26:26effects on melanoma.
• 26:28And,
• 26:29via a small biotech that
• 26:30I founded,
• 26:32many years ago, we advanced
• 26:34this,
• 26:35another,
• 26:36more potent version of this
• 26:38compound
• 26:38into clinical testing.
• 26:40And and, this is,
• 26:43currently in phase one b
• 26:44studies,
• 26:45And this is, again, this
• 26:47is just an anecdote anecdote,
• 26:50but their the clinicians have
• 26:51observed more multiple examples of
• 26:53regression responses in a subset
• 26:55of patients as well as
• 26:57reductions in MDSC levels. For
• 26:59example, this is a sixty
• 27:00two year old woman with
• 27:01metastatic head and neck cancer.
• 27:03She's refractory to chemotherapy and
• 27:04anti PD one immunotherapy.
• 27:06She has she has a
• 27:07seven centimeter clavicular metastasis on
• 27:09her clavicle and also these
• 27:10two liver metastases.
• 27:12And you can see after
• 27:13eight weeks of treatment, there's
• 27:14been a shrinkage of these
• 27:16these lesions.
• 27:17And there are additional examples
• 27:18of this,
• 27:20and so this, I think,
• 27:21suggests that,
• 27:23at least provides proof of
• 27:24concept that,
• 27:25targeting these pathways
• 27:27could perhaps
• 27:28lead to regression responses. We
• 27:30think these pathways inhibition would
• 27:31be most optimal in a
• 27:33prevention setting, which is very
• 27:35hard and expensive, but we
• 27:36do believe that these cells
• 27:38are continue to exploit these
• 27:39pathways to progress.
• 27:41And so,
• 27:42you know, the question that
• 27:43you get is, what was
• 27:45the ancestral role for APOE
• 27:46variation? Right? It wasn't to
• 27:49protect you against melanoma metastasis
• 27:51because this is post reproductive.
• 27:54And
• 27:55with when it comes to
• 27:55Alzheimer's, it's the other way
• 27:57around. Right? E four increases
• 27:59risk for Alzheimer's.
• 28:00You get immune hyperactivation.
• 28:02E two,
• 28:04increases your risk for melanoma
• 28:05metastasis, protects you from Alzheimer's.
• 28:08Evolution wasn't selecting
• 28:10a two to protect you
• 28:11from Alzheimer's. Right? Again, post
• 28:13reproductive.
• 28:14And so,
• 28:15what is it? What does
• 28:16this access really do? What
• 28:18was it selected for? If
• 28:19we take the immune infiltrates
• 28:22of these metastases
• 28:23and perform gene expression analysis
• 28:26and we ask what's enriched
• 28:27in e four versus e
• 28:28two, the first the prom
• 28:30primary
• 28:31pathways we get are interferon
• 28:33alpha, interferon gamma, allograft rejection,
• 28:35inflammatory response.
• 28:37What are these pathways?
• 28:39Well, these are really antiviral
• 28:41pathways. Right? So antiviral pathways
• 28:43are emerging as the top
• 28:44pathways.
• 28:45And so for a long
• 28:46time, we've been interested in
• 28:48seeing if we can test
• 28:49the role of this access
• 28:51when it comes to antiviral
• 28:53responses and antiviral immunity.
• 28:55And,
• 28:56when Ben was in the
• 28:57lab, the COVID pandemic provided
• 28:58us an, a a way
• 29:00to do this because
• 29:01it allowed us, a way
• 29:03to access.
• 29:05We needed a large fraction
• 29:06of the population
• 29:07to be APOE genotype
• 29:09being infected with the same
• 29:11pathogen to really see if
• 29:12there's something here. And so
• 29:14we were able to obtain
• 29:15mouse adapted SARS CoV two
• 29:16from our Charlie Rice And
• 29:18Ben began infecting mice in
• 29:20BSL three,
• 29:22and large collections of APOE
• 29:24two, E three, and E
• 29:25four genotype mice. And what
• 29:26he found is that as
• 29:28we see ASAN in the
• 29:29patients,
• 29:30older mice did worse in
• 29:31terms of their survival.
• 29:33Male mice did worse in
• 29:35terms of their survival upon
• 29:36being infected with, SARS CoV
• 29:38two. And Ben and I
• 29:40had a bet. The question
• 29:41was, do we think
• 29:43e two mice were gonna
• 29:45do worse or e four
• 29:46mice were gonna do worse?
• 29:47I said, I think e
• 29:49two mice will do worse.
• 29:50It's immunosuppression
• 29:52that's what we see in
• 29:52cancer. Ben said, you're wrong.
• 29:54It's e four mice because
• 29:56they're hyperinflammatory,
• 29:57and the patients that are
• 29:58showing up in the ICU
• 29:59are are hyper inflammatory, and
• 30:01that's what they're dying of.
• 30:02And when we did the
• 30:03experiments, of course, Ben was
• 30:05right. The e four mice
• 30:07did worse. That's they get
• 30:09divorced survival, but I wasn't
• 30:10wrong.
• 30:11E two mice also did
• 30:13worse than e three, so
• 30:14they were, sort of in
• 30:15the middle. And so what's
• 30:17really cool here is now
• 30:18we have a disease context
• 30:19where
• 30:20e three is doing better,
• 30:22and e four and e
• 30:23two are are suboptimal.
• 30:25And, then what we did
• 30:27was then look at the
• 30:28UK Biobank,
• 30:29and when we look at
• 30:30the the human population, the
• 30:32UK Biobank, the large database
• 30:34that they had,
• 30:35over ten thousand patients were
• 30:37infected, you can see the
• 30:38same thing. The e four
• 30:40homozygotes
• 30:41did the worse.
• 30:42E two homozygotes did second
• 30:44worse, and the other, the
• 30:45heterozygotes or the e three
• 30:46homozygotes,
• 30:48did did better. And this
• 30:49is not a this is
• 30:50not an insignificant fraction of
• 30:52the human population.
• 30:53E two and e four
• 30:54homozygous accounts for three percent
• 30:56of the human population. That's
• 30:57about two hundred and fifty
• 30:58million people.
• 30:59And these findings were independently
• 31:01validated in the Finnish cohort
• 31:02where they show an association
• 31:04between,
• 31:05also show an association of
• 31:07APOE with survival in the
• 31:08large Finnish cohort. So
• 31:11so we believe that,
• 31:13when we think about
• 31:14selection,
• 31:15again, hypothesis,
• 31:18we believe that one selection
• 31:20for
• 31:21the emergence of e three
• 31:22as the predominant,
• 31:24variant. So the ancestral allele
• 31:27was e four,
• 31:28and that the and e
• 31:29three emerged, two hundred thousand
• 31:31years ago, and e two
• 31:33emerged eighty thousand years ago.
• 31:35So it's it's our hypothesis
• 31:36that,
• 31:37infection by viral pathogens trims
• 31:39at the edges, and that
• 31:41might contribute to e three
• 31:43being the most dominant,
• 31:45variant,
• 31:47allele.
• 31:48And and I I think
• 31:48we can under perhaps this
• 31:50also suggests that perhaps me
• 31:51to view
• 31:53Alzheimer's,
• 31:55as,
• 31:56something that the relationship between
• 31:58Alzheimer's and enhanced infectivity
• 32:01and, mortality with the viral
• 32:03infection as perhaps
• 32:05a a virally associated
• 32:07disease. And we're not saying
• 32:08it's SARS CoV two specific.
• 32:10We lots of viruses require
• 32:12this interferon pathway.
• 32:13And so
• 32:15I we've I've shown you
• 32:16that when we use molecular
• 32:18biology, Nora was able to
• 32:26test the hypothesis whether hereditary
• 32:28genetics could associate with a
• 32:31high risk melanoma.
• 32:33So we got very lucky
• 32:34there. And so when Binh
• 32:36Mei came to the lab
• 32:37and the idea was, could
• 32:38we actually use human genetics
• 32:41in another cancer to identify
• 32:42another variant that might play
• 32:44a similar role in affecting
• 32:46metastasis?
• 32:48And so what one bin
• 32:49did was to say, okay.
• 32:51You know, we don't have
• 32:52enough statistical power to do
• 32:53this because when we look
• 32:55at the datasets, we need
• 32:56a large dataset. So we
• 32:57started with breast cancer.
• 32:59That we that's we have
• 33:00it's a very it's a
• 33:01it's a highly prevalent disease,
• 33:03and we have a lot
• 33:03of data in terms of
• 33:04sequence data. But it's still
• 33:06we still have too much
• 33:07genetic variation.
• 33:08So we started to learn
• 33:09from ApoE. What is it
• 33:10about ApoE? ApoE is a
• 33:11secreted protein, thus it can
• 33:13be a source of communication
• 33:14between the host and the
• 33:15cancer compartment.
• 33:17ApoE has
• 33:18we have,
• 33:19substantial germline genetic variation in
• 33:21apoE, so you can we
• 33:22will focus on variants that
• 33:25had high,
• 33:26allele frequencies so we could
• 33:28have good statistics.
• 33:30And, also, it already is
• 33:31associated with a disease
• 33:34prior increasing the odds that
• 33:35it may also play a
• 33:36role in in another disease.
• 33:38So we looked
• 33:40we just limited our analysis
• 33:41to such variance in such
• 33:43genes that could perhaps play
• 33:45a role. And what Von
• 33:47Binh did was he took
• 33:48two breast cancer datasets, the
• 33:50TCGA
• 33:50and Bertoosti cohort, and asked
• 33:53in both datasets, are there
• 33:54any variants that associate with
• 33:57survival outcomes? And he identified
• 33:59eight variants in eight genes
• 34:01that associated again, this is
• 34:03pure association, no causality.
• 34:05And,
• 34:06what I'm gonna tell you
• 34:07about is one of those
• 34:07genes that our attention was
• 34:09drawn to, and that's PCSK
• 34:11nine. And that is
• 34:12a gene that's associated with
• 34:14cardiovascular
• 34:15cardiovascular
• 34:16disease, hypercholesterolemia,
• 34:18and a gene that, for
• 34:20which we have, antibodies
• 34:22that are used in patients
• 34:24that have extreme hypercholesterolemia.
• 34:26The variant in PCSK9 we
• 34:28identified is is has modest
• 34:29effects on hypercholesterolemia,
• 34:32and that's what we studied.
• 34:34So what you can see
• 34:35here is that we he
• 34:37identifies a pathogenic,
• 34:39SNP in in PCSK nine.
• 34:42On the upper left, that's
• 34:43the distribution of this allele.
• 34:46The ancestral allele is the
• 34:47SNP in blue. It looks
• 34:49like in the human population,
• 34:50the ancestral allele has been
• 34:52replaced
• 34:53largely with this pathogenic allele.
• 34:56And,
• 34:57and within the Caucasian population,
• 35:00seventy percent of of individuals
• 35:03are homozygous for the pathogenic
• 35:05allele. If we look at
• 35:06these datasets, the pathogenic allele
• 35:08shown in orange,
• 35:10survival in the TCGA breast
• 35:12cancer dataset based on germline,
• 35:15this germline STIP is worse.
• 35:17We see the same thing
• 35:18in Bertucci. And now we
• 35:20look at two independent datasets,
• 35:21the Dutch cohort. We can
• 35:23see worse survival outcomes. And
• 35:25in the Nix en Al
• 35:26cohort, this is a British
• 35:27cohort, we see worse survival
• 35:28outcomes. Majority of these datasets
• 35:30are in
• 35:32majority of the patients in
• 35:32these datasets are are European
• 35:35of European ancestry.
• 35:37So that's something to keep
• 35:38in mind. The other thing
• 35:39to keep in mind is
• 35:40that this allele is fixed
• 35:42in the Asian population. Therefore,
• 35:43we it's uninformative in the
• 35:45Asian population. We can't say
• 35:46anything about the Asian population.
• 35:46You can see that up
• 35:46there. And so the variant
• 35:46that we
• 35:56modestly
• 35:56increased cholesterol levels, very subtle
• 35:59in the in the human
• 36:00population. So this is hypermorphic.
• 36:02It's a hypermorphic allele. So
• 36:04what we did was to
• 36:05genetically
• 36:05inactivate PCSK nine using CRISPR
• 36:08and introduce breast cancer cells,
• 36:11different types of breast cancer
• 36:12cells into these animals, and
• 36:13we see that there's a
• 36:15substantial reduction in metastatic colonization
• 36:17when you genetically inactivate PCSK
• 36:19nine.
• 36:20On the lower left, we
• 36:21can implant these cells into
• 36:23the mammary gland and look
• 36:24at orthotopic metastasis.
• 36:26We see the same effect.
• 36:27PCSK nine is a promoter
• 36:29of breast cancer metastasis.
• 36:30We can use an entirely
• 36:32genetic model, the MMTB polyam
• 36:34and middle t model, where
• 36:35you're expressing the polyam and
• 36:37middle t,
• 36:39gene, which is oncogenic
• 36:41in the breasts.
• 36:42And we're either having wild
• 36:44type PCSK nine or genetically
• 36:46null PCSK nine, and we
• 36:47see the same effect.
• 36:49So what we next did
• 36:50was to,
• 36:52work with a, a CRO
• 36:55to replace the murine PCSK
• 36:57nine with a human PCSK
• 36:59nine
• 37:00and to make two variants,
• 37:02one nucleotide difference
• 37:05where we have homozygosity
• 37:06of the pathogenic allele or
• 37:08we have the nonpathogenic
• 37:10allele. The pathogenic allele, again,
• 37:11is shown in orange. When
• 37:13we compare
• 37:14metastasis
• 37:15in these two models, we
• 37:16see that the path mice
• 37:18that have the pathogenic
• 37:20allele have worse survival,
• 37:23higher metastasis than the nonpathogenic
• 37:25allele.
• 37:27We were really fortunate
• 37:28to, this
• 37:30to be able to find
• 37:31a collaborator in in in
• 37:33in Sweden.
• 37:34Her name is Helena Jernstrom,
• 37:35and her focus is on
• 37:36breast cancer. Over the years,
• 37:38she's collected a very large
• 37:39data set of,
• 37:41early stage breast cancers,
• 37:43which are well annotated and
• 37:44for which she has,
• 37:46updated survival outcomes, and we
• 37:47can look at confounding variants
• 37:49and stuff. So we did
• 37:50this to do a blinded
• 37:52analysis. We gave her the
• 37:53snips, and we said, could
• 37:54you look in your dataset
• 37:56to see if these variants
• 37:58have, any association with metastatic,
• 38:01survival outcomes?
• 38:02And what she found that,
• 38:05indeed,
• 38:05the pathogenic variant after fifteen
• 38:07years, women with early stage
• 38:09breast cancers who are homozygous
• 38:11for the pathogenic variant, they
• 38:12had a twenty two percent
• 38:14likelihood of metastatic relapse at
• 38:15fifteen years.
• 38:17Women that were not homozygous
• 38:18for the pathogenic variant had
• 38:20a two percent risk of
• 38:21metastatic relapse.
• 38:22And so this is the
• 38:23the fifth cohort in which
• 38:24we've shown a role for
• 38:25this germline variant in survival
• 38:27outcomes.
• 38:28Because this is a secreted
• 38:30protein,
• 38:31and pharmaceutical companies have developed
• 38:33antibodies for it to treat
• 38:35recalcitrant
• 38:35hypercholesterolemia.
• 38:37So what we can do
• 38:38is we can in in
• 38:39inject these antibodies
• 38:41into these mice. These antibodies
• 38:42are now generic,
• 38:45and what we can do
• 38:46is we can inject these
• 38:47antibodies into mice,
• 38:48and we can see,
• 38:50similar,
• 38:51reduced effects on metastatic colonization
• 38:54using four t one model,
• 38:55e o seven seven one
• 38:57model, the spontaneous model, colonization
• 38:59model, and also the genetically
• 39:01engineered model.
• 39:03So what does it do?
• 39:04What is p c scan
• 39:04nine doing? We know that
• 39:06p c scan nine, the
• 39:07way it affects cholesterol is
• 39:08by binding to the LDL
• 39:09receptor
• 39:10and leading to its degradation.
• 39:13And so
• 39:14what we did was we
• 39:15did proteomic analysis in our
• 39:17breast cancer cells and looked
• 39:19for proteins whose abundance goes
• 39:21down.
• 39:22We saw LDL receptor is
• 39:23being degraded,
• 39:24but interestingly, we saw another
• 39:26receptor that's related to LDL
• 39:28receptor called the LDL receptor
• 39:30like protein
• 39:31LRP one that I had
• 39:33previously told you about our
• 39:34work in melanoma
• 39:35that was also degraded. And
• 39:37another's
• 39:38one group had shown that
• 39:40PCSK nine can also interact
• 39:42with LRP one.
• 39:44And indeed, we we found
• 39:45that PCSK nine
• 39:47interacts with LRP one
• 39:49on breast cancer cells. And
• 39:51what we find is that
• 39:52the way this system works
• 39:54is that if you look
• 39:55at the bottom,
• 39:56in the absence of PCSK
• 39:58nine,
• 39:59LRP one,
• 40:01the intracellular domain of LRP
• 40:03one we find,
• 40:04goes to the nucleus and
• 40:06represses
• 40:07transcription
• 40:07of certain prometastatic
• 40:09genes, USB eighteen and f
• 40:11one. We found that these
• 40:13genes promote the proliferative competence
• 40:15of breast cancer cells in
• 40:17the lung. They enhance their
• 40:18ability to initiate from single
• 40:20cells to macroscopic colonies.
• 40:22In the presence of PCSK
• 40:24nine, PCSK nine,
• 40:26degrades LRP one. You get
• 40:28the repression and activation,
• 40:30of these genes. The pathogenic
• 40:32variant is better able to
• 40:34bind LRP one, we find,
• 40:35and it's better able to
• 40:37degrade LRP one. And so
• 40:40this is kind of remarkable
• 40:41to me that
• 40:43total independent students working in
• 40:45two different cancers, one melanoma,
• 40:47and in this case,
• 40:49breast cancer, using independent approaches,
• 40:51one molecular biology, and in
• 40:53this case, germline genetics,
• 40:55have, converged on on a
• 40:57common pathway, LRP one. In
• 40:59one case, APOE is activating
• 41:01this receptor to act as
• 41:03a repressive,
• 41:04to mediate a repressive response.
• 41:06And in the other case,
• 41:07PCSK nine is is negatively
• 41:09regulating LRP one to promote
• 41:11metastatic progression.
• 41:13And so,
• 41:15in the last talk I'll
• 41:16tell you about is,
• 41:18so we think in these
• 41:19two breast cancers, we found
• 41:21examples for germline variation being
• 41:22important.
• 41:24We don't rule out somatic
• 41:25mutations being, relevant. People are
• 41:27searching for them. Of course,
• 41:28they could be important, but
• 41:29we believe hereditary differences are
• 41:31important in these cancers.
• 41:33But we've been thinking about
• 41:34any other commonalities.
• 41:36Our work with microRNAs in
• 41:38colorectal cancer,
• 41:40when we did those studies,
• 41:41what Xiaoming Liu found was
• 41:43that
• 41:44two microRNAs
• 41:45become silenced and this leads
• 41:46to the repression of a
• 41:47protein,
• 41:48called creatin kinase brain. It's
• 41:50a metabolic gene. What he
• 41:52found is that as these
• 41:53colon cancer cells arrive in
• 41:55the liver microenvironment, which is
• 41:57hypoxic
• 41:57and hypoglycemic,
• 41:59they're under metabolic stress and
• 42:00there's massive death.
• 42:02Less than one percent of
• 42:04the one in ten thousand
• 42:05cells survives this process.
• 42:07And those cells that survive
• 42:08are cells that overexpress CKB.
• 42:10They release it in the
• 42:11extracellularly
• 42:13both on the plasma membrane,
• 42:14but also release it into
• 42:16the microenvironment.
• 42:17And they can capture high
• 42:18energy high energy phosphate
• 42:20by phosphorylating
• 42:21creatinine, which is highly abundant
• 42:23in the circulation
• 42:24using ATP, which is released
• 42:26from dying cells because there's
• 42:28a massive amount of cell
• 42:29death in these metastases.
• 42:31They they produce phosphocreatine,
• 42:33which is imported via a
• 42:34creatinine transporter,
• 42:35SLC six a eight, then
• 42:37they use that high energy
• 42:38phosphate, which has about fifty
• 42:39percent more free energy than
• 42:40the gamma phosphate or ATP
• 42:42to replete ATP stores,
• 42:44and they can undergo this
• 42:45metabolic stress.
• 42:47And a large number of
• 42:48labs have shown that this
• 42:49mechanism,
• 42:51phosphocreatine,
• 42:52CKB, SLC six a eight,
• 42:54they've shown that these genes
• 42:55are promoters of cancer progression
• 42:57and metastasis
• 42:58in a number of other
• 42:59models recently. So these findings
• 43:01have been validated.
• 43:03So in in in colorectal
• 43:05cancer, creating kinase brain.
• 43:07In breast cancer, the work
• 43:08we did with microRNAs,
• 43:10led to another pathway. And
• 43:11that pathway, what the students
• 43:13in the lab found is
• 43:14that these breast cancer cells
• 43:16are inducing signals onto the
• 43:18vasculature
• 43:19and the vasculature is releasing
• 43:20a guidance molecule called SLIT
• 43:22two that acts on the
• 43:23cancer cells.
• 43:25It's a chemotactic guidance molecule
• 43:27and it enhances migration of
• 43:28breast cancer cells through the
• 43:30vasculature into circulation
• 43:31to then disseminate.
• 43:33And so what's really interesting
• 43:34to me is that
• 43:35these these approaches have have
• 43:37one commonality,
• 43:39in terms of the genes.
• 43:40These are nervous system genes.
• 43:42In the tumor compartment in
• 43:43colorectal cancer cells, treating kinase
• 43:45brain, a nervous system gene
• 43:46is being induced. In breast
• 43:48cancer cells, they're inducing this
• 43:49SLID two neural guidance molecule.
• 43:52APOE also is a gene
• 43:54implicated in neurodegeneration.
• 43:56So what we've what we
• 43:58find recurrently
• 43:59is that these metastatic cells
• 44:02are accessing neural genes, and
• 44:04these neural genes are having
• 44:05strong high magnitude effect size
• 44:07on metastasis.
• 44:08And so we started to
• 44:09think about what if we're
• 44:10sort of,
• 44:12myopic about this? What about
• 44:14the nervous system,
• 44:15with large? Like, what if
• 44:16we look beyond the the
• 44:18tumor compartment?
• 44:19Could something like SLID two,
• 44:20for example, be affecting innervation
• 44:23of these tumors? And so
• 44:24Veena Padmadabad, an outstanding,
• 44:27postdoc in the lab who
• 44:28was helped by Ethan and
• 44:29Isabel,
• 44:30started to ask this question.
• 44:32And so she started to
• 44:33look at innovation of breast
• 44:34tumors. And what she did
• 44:36was she used optical clearing
• 44:37approaches from our neuroscience colleagues
• 44:39to clear these breast tumors.
• 44:41And then what you can
• 44:41do is you can do
• 44:42immunofluorescence
• 44:43in the entire tumor,
• 44:45and this can be a
• 44:46lot applied to a large
• 44:47variety of tumors. And when
• 44:48she's stained for nerves, in
• 44:50this case, sensory nerves, she
• 44:51found repeatedly
• 44:52that the highly metastatic tumors
• 44:54had more innervation than the
• 44:56poorly metastatic tumors. And this
• 44:58is not a one off.
• 44:59You look she's looked across
• 45:01a collection of different breast
• 45:03cancers,
• 45:04murine breast cancers, genetically initiated,
• 45:07PDX tumors from patients that
• 45:08are HER2 positive, ER positive,
• 45:10ER negative.
• 45:12We see sensory neuron innervation
• 45:14across,
• 45:15these breast tumors.
• 45:17And
• 45:18and, indeed, when you inactivate
• 45:20and the SLIT two in
• 45:22the vasculature,
• 45:23she finds that there's reduced
• 45:25innervation of these tumors.
• 45:27And if you look at
• 45:29breast cancer datasets and you
• 45:30use
• 45:31these neuronal,
• 45:33marker genes that are pan
• 45:34neuronal such as beta three
• 45:36tubulin or PGP nine point
• 45:38five, on the left, you're
• 45:39looking at mRNA levels for
• 45:41these neuronal markers. You see
• 45:43that women whose breast tumors
• 45:45have higher expression of these
• 45:47neural markers tend to have
• 45:48worse survive metastatic
• 45:50survival outcomes
• 45:52than those who don't. On
• 45:53the right, you have protein
• 45:55levels of beta three tubulin
• 45:57and perforin, and you can
• 45:58see the same effect. So
• 45:59these markers are associating these
• 46:01neural markers are associating.
• 46:02But what what what are
• 46:03neur what are neurons doing
• 46:05for the breast cancer cells?
• 46:06Could we break this down?
• 46:07And what Veena did was
• 46:08she established spheroid
• 46:10assays where she embedded these
• 46:12breast cancer,
• 46:13organoids into collagen. One, a
• 46:15major
• 46:16component of the extracellular matrix
• 46:18of the breast.
• 46:19On the up in the
• 46:20blue is this dorsal root
• 46:22ganglion sensory neuron. The innervation
• 46:24that is coming into the
• 46:25breast is from the dorsal
• 46:27root ganglia we find. We've
• 46:28done tracing analysis.
• 46:30In the center is a
• 46:31spheroid, a breast cancer spheroid.
• 46:33So what happened on the
• 46:34left, you just have the
• 46:35spheroid.
• 46:35In the middle, you have
• 46:36the spheroid
• 46:37with a dorsal root ganglion
• 46:39neuron in the inset. So
• 46:40when you put them in
• 46:41together, you guys can see
• 46:43that
• 46:44the the presence of the
• 46:46sensory neuron, the the breast
• 46:47cancer organs are much more
• 46:49invasive.
• 46:50And so it's not it's
• 46:51not subtle,
• 46:53and we've seen this across
• 46:54the collection of murine and
• 46:56human breast cancer models. We
• 46:59work with surgeons to obtain
• 47:00surgical biopsies of patients
• 47:02with breast cancer when we
• 47:04culture
• 47:05human breast cancer spheroids with
• 47:07dorsal root ganglia neurons, we
• 47:09see that this causes them
• 47:10to be more invasive.
• 47:12We also see that it
• 47:13causes them to be more
• 47:14proliferative. So this is k
• 47:15I sixty seven staining.
• 47:18And so are neurons
• 47:20doing anything?
• 47:21And so the way to
• 47:22the dirty way to address
• 47:24that is to inject
• 47:25capsaicin
• 47:27into
• 47:27the,
• 47:29into the breast, tumors, via
• 47:31intraductal injections. So your capsaicin
• 47:33is a TRP v one
• 47:34agonist. It'll cause denervation. It'll
• 47:37kill the neurons.
• 47:38And what she can do
• 47:39is she can follow a
• 47:40breast cancer growth and she
• 47:42sees reduced tumor growth and
• 47:43reduced metastasis. So this is
• 47:45a dirty way of doing
• 47:46this, loss of function experiment.
• 47:49So I'm gonna kinda tell
• 47:50you about studies that Veena
• 47:52did to get a load
• 47:52of this. What are what
• 47:53what are the neurons doing?
• 47:55So first, neurons light fire
• 47:57action potentials and they release
• 47:58neurotransmitters.
• 47:59So what she did was
• 48:00to load these neurons with
• 48:02a fluorescent indicator, difluor four,
• 48:04in the upper left.
• 48:06At the top, you can
• 48:07see if we just culture
• 48:08the DRG neurons alone,
• 48:11we see,
• 48:12flat not much activity in
• 48:13terms of calcium transient. But
• 48:15if we place the dorsal
• 48:16organoid neurons with the breast
• 48:18cancer
• 48:19organoids,
• 48:19we see spikes of activity.
• 48:21These are calcium transients.
• 48:23And then these are the
• 48:24roster plots, just the neurons
• 48:26themselves.
• 48:28No neurons plus the cancer
• 48:29cells. So the neurons
• 48:31the cancer cells are releasing
• 48:33some signal that we don't
• 48:34know what it is. It's
• 48:35enhancing calcium transients in these
• 48:38neurons. What do calcium transients
• 48:40do? They release
• 48:41they lead to release of
• 48:42neurotransmitters.
• 48:44For these sensory neurons and
• 48:45neurotransmitters
• 48:46are CGRPs,
• 48:47substance p, galanin,
• 48:49and and she looked at
• 48:51at these different neuropeptides
• 48:52and she found that one
• 48:53of them phenocopied
• 48:55what she saw, and that
• 48:56was substance p.
• 48:58So what you can see
• 48:59here is that this is
• 49:01the,
• 49:02breast cancer organoid
• 49:04alone.
• 49:05This is a breast cancer
• 49:06organoid when you add substance
• 49:07p, no neurons, and you
• 49:09can see that it's sufficient
• 49:10to enhance
• 49:11invasiveness.
• 49:13If she takes,
• 49:16dorsal root ganglia neurons from
• 49:17mice
• 49:18that are that are known
• 49:20for substance p, they have
• 49:22a deletion of tach one,
• 49:23which produces substance p.
• 49:25Now you no longer see
• 49:26this enhanced invasiveness.
• 49:28So substance p from the
• 49:31dorsal root ganglion neurons is
• 49:33required for inducing this invasion
• 49:36and proliferation
• 49:37phenotype.
• 49:38If we use mice that
• 49:40are genetically known for substance
• 49:42p, they cannot produce this
• 49:43neuropeptide. We implant the tumors.
• 49:46We see less tumor growth,
• 49:48less metastasis.
• 49:51When so what is substance
• 49:52p doing? We find that
• 49:53substance p derived from these
• 49:55neurons is acting on the
• 49:56cancer cells, and this is
• 49:57changing the condition media. The
• 49:59we can take the condition
• 50:00media and add it to
• 50:02breast tumor organoids, and it's
• 50:03sufficient to make
• 50:05the cells more invasive and
• 50:07more proliferative. So this is
• 50:08can just condition media from
• 50:09this neuron cancer co culture,
• 50:12and this is the proliferative
• 50:13index.
• 50:14When she treats this condition
• 50:16media as DNAs, the activity
• 50:17persists. Heat inactivation, the activity
• 50:19persists. When she treats it
• 50:21with ribonuclease,
• 50:22the activity goes away.
• 50:23So then the question was,
• 50:25is it single stranded or
• 50:26double stranded RNA?
• 50:27We don't think it's double
• 50:28stranded RNA because we when
• 50:29we treat with ribonuclease three,
• 50:32the activity persists.
• 50:33But if we treat with
• 50:35inhibitors of single strand
• 50:37of single stranded RNA,
• 50:39either RNase a, which is
• 50:40nonspecific, or RNase t one,
• 50:42which is specific, the activity
• 50:43goes away.
• 50:44So this suggests that,
• 50:46we're getting release of single
• 50:48stranded RNA that's acting to
• 50:50promote this phenotype.
• 50:52When we
• 50:53add,
• 50:55the when we take the
• 50:56condition media from the breast
• 50:58neuron co culture, we see
• 51:00more RNA in the condition
• 51:02media.
• 51:03If we take breast cancer
• 51:04cells and just add substance
• 51:06p in the absence of
• 51:08the neurons, we see more
• 51:09RNA in the conditioned media.
• 51:13And,
• 51:15and if we use a
• 51:16single stranded RNA mimetic
• 51:18here, s s r n
• 51:19a forty, and add it
• 51:21to the breast tumor organoids
• 51:22in the absence of neurons,
• 51:23it's sufficient
• 51:25to enhance their invasiveness
• 51:26and to enhance their proliferation.
• 51:29Okay? So single stranded RNA
• 51:31could be released by the
• 51:32cancer cells either via an
• 51:34active process
• 51:35or via the death of
• 51:37of cells that release all
• 51:38their contents and release some
• 51:39single stranded RNA. In support
• 51:41of this latter hypothesis,
• 51:44when we use a caspase
• 51:46reporter to in the tumor
• 51:48compartment
• 51:49to report on cell death,
• 51:51we find that when we
• 51:52add substance p, we see
• 51:54increased death of a very
• 51:56small fraction of the breast
• 51:57cancer cells, about less than
• 51:59one percent. And if we
• 52:00co culture the cancer cells
• 52:01with the DRG neurons, we
• 52:03see death of a subset
• 52:05of cancer cells.
• 52:06And the cells that are
• 52:07dying tend to be the
• 52:08ones that have high expression
• 52:10of the receptor for substance
• 52:12p. That's TACR one because
• 52:14in the presence of substance
• 52:15p,
• 52:16half of them,
• 52:18are lost.
• 52:19So this is a little
• 52:20complicated. I'm telling you that
• 52:22substance p and these neurons
• 52:23are promoting cancer progression phenotypes,
• 52:26promoting invasiveness and proliferation.
• 52:29But the most proximal step,
• 52:31which was confusing to us,
• 52:33is they're actually killing a
• 52:35subset of the cancer cells.
• 52:37That leads to release of
• 52:39the contents of those cells.
• 52:41Within that content, there's single
• 52:42stranded RNA, and a single
• 52:44stranded RNA is doing something
• 52:45to the rest of the
• 52:46cancer cells.
• 52:48And so is this really
• 52:49true? If we inhibit
• 52:52apoptosis with the zVAT FMK
• 52:54compound,
• 52:55now
• 52:56you inhibit apoptosis,
• 52:57substance p is unable to
• 52:59enhance invasiveness,
• 53:00The DRG neurons are unable
• 53:02to enhance invasiveness.
• 53:04And finally, what what is
• 53:05the receptor for the single
• 53:07stranded RNA?
• 53:08TLR seven,
• 53:11is the receptor for single
• 53:12stranded RNA in in these
• 53:14in these cells. And so
• 53:15if we knock down TLR
• 53:17seven,
• 53:18now we see reduced tumor
• 53:20growth,
• 53:20and we see reduced metastasis.
• 53:23So the cells that are,
• 53:26the the other ninety nine
• 53:27percent of the population
• 53:29is co opting this response
• 53:31and benefiting from the single
• 53:32stranded RNA release. So I'm
• 53:34gonna summarize this this this
• 53:36complex story, and that's the
• 53:37end of my talk. We
• 53:38find that highly metastatic breast
• 53:40tumors
• 53:41have evolved
• 53:42to recruit
• 53:44in to enhance
• 53:45innervation.
• 53:46We see enhanced
• 53:48sensory neuron innervation
• 53:49from the dorsal root ganglia.
• 53:52We find that the vasculature
• 53:54is providing the signal.
• 53:56SLIT two is a ax
• 53:57neuronal action guidance molecule.
• 53:59We find that it it's
• 54:00enhancing innervation to the tumors.
• 54:03We find that that this
• 54:05this innervation is what it's
• 54:07what we find is that,
• 54:08substance p, which is a
• 54:09neuropeptide
• 54:11that's released from these neurons
• 54:12that's implicated in pain,
• 54:14is leading to death of
• 54:15a subset of breast cancer
• 54:16cells that overexpress TACR one.
• 54:18These are the TACR one
• 54:19high population.
• 54:21The death of these cells
• 54:23releases the contents of these
• 54:24cells, everything.
• 54:26Within this content, there is
• 54:28single stranded RNA,
• 54:30and
• 54:30the rest of the breast
• 54:31cancer cells,
• 54:32are expressing
• 54:34TLR seven.
• 54:36And this single stranded RNA
• 54:37binding to TLR seven is
• 54:39activating a proliferative invasive program.
• 54:42We do not find interferon
• 54:44activation.
• 54:45We find that this is
• 54:46my d eighty eight independent.
• 54:47So these breast cancer cells
• 54:48have sort of veered away
• 54:50from the inflammatory
• 54:53response induction downstream of t
• 54:55l r, and they're using
• 54:56a invasive
• 54:57proliferative,
• 54:59aspect of this of this
• 55:00program.
• 55:01And so,
• 55:03I'll just summarize
• 55:04what I just said. We're
• 55:05seeing this surprising
• 55:07thing where,
• 55:09the initial responses for the
• 55:11neurons to be killing a
• 55:13subset of cancer cells, but
• 55:14the population
• 55:15evolves to exploit this this
• 55:17response and to and it
• 55:19drives progression.
• 55:20And we think this this
• 55:21could be therapeutically targeted. For
• 55:23example,
• 55:24there are anti nausea
• 55:26medications
• 55:27that can inhibit this substance
• 55:29p access,
• 55:30that can inhibit this pathway
• 55:32and could perhaps be useful,
• 55:34in the clinic. I'll stop
• 55:36there. I mentioned the people
• 55:37who did all the work.
• 55:38I wanna thank you again,
• 55:39for inviting me. I'm super
• 55:41honored. Happy to take questions.
• 56:03Yeah.
• 56:08Yeah. No. That's fantastic. We
• 56:09have not done that. That
• 56:10would be a really interesting
• 56:11thing to do. I can
• 56:12just tell you that,
• 56:15people have found in certain
• 56:17contexts that when they've looked
• 56:19at cell death within tumors
• 56:22and thought that it should
• 56:23as as an oncologist, we
• 56:24we give chemo and and
• 56:25we we wanna kill that
• 56:27tumor, destroy that tumor.
• 56:29There are papers published,
• 56:31and there's one really good
• 56:32paper published where they really
• 56:33systematically
• 56:34looked at cell death in
• 56:35the tumor compartment. They find
• 56:37an inverse association. That means
• 56:38when there was more death
• 56:39in the tumor compartment from
• 56:41chemotherapy,
• 56:42the patients did worse. And
• 56:44so I think it'd be
• 56:45very interesting to think about,
• 56:48and it's possible that some
• 56:49chemotherapies are targeted agents. They
• 56:51may be effective, but part
• 56:53of their liability
• 56:54is enhancing death that can
• 56:56in that can drive these
• 56:58programs because the tumors will
• 56:59co opt it. But we
• 57:00haven't done that. It'd be
• 57:00really interesting to do that.
• 57:03We have a question online.
• 57:04Is that right?
• 57:08Oh, let me just see.
• 57:13Okay.
• 57:15Stop sharing? Or
• 57:18yeah.
• 57:20Uh-huh. Chat. Yeah.
• 57:24Oh, here we go.
• 57:26It's not
• 57:27metastasis broadly. However, what implications
• 57:28does this work have specifically
• 57:30for BRAIN and or leptomeningeal
• 57:32met metastases?
• 57:33So,
• 57:35we don't,
• 57:36we are not studying leptomeningeal
• 57:38or BRAIN metastases.
• 57:39I can tell you that,
• 57:41and we have not looked
• 57:43the PCSK9
• 57:44story, we have not looked
• 57:45at, brain metastases.
• 57:47The the colon cancer stuff,
• 57:49obviously, we haven't looked at
• 57:50it. I can tell you
• 57:51in melanoma,
• 57:52this APOE axis is general
• 57:55is is broadly repressive. So
• 57:57when Nora looked at brain
• 57:59metastasis,
• 58:00she found that APOE can
• 58:01reduce
• 58:02the colonization
• 58:03of the brain,
• 58:05and and we don't know
• 58:06what the mechanism that is,
• 58:07perhaps due to its effects
• 58:09on invasiveness.
• 58:10But, in general, we haven't
• 58:12looked beyond that.
• 58:14Marcus?
• 58:16TLR seven agonists, like, pretty
• 58:19widely used in pathology.
• 58:21And sometimes, you know, with
• 58:22individual patients, they might actually
• 58:23have
• 58:36Yeah. Yeah. So,
• 58:38so Vina has done in
• 58:39vitro experiments, and she she
• 58:41can phenocopy,
• 58:43is the in v invasion
• 58:45phenotype in vitro. We've not
• 58:46done anything in vivo. So
• 58:48it's we and, you know,
• 58:50done all the you know,
• 58:51done TLR seven knockdown and
• 58:52whatnot. So she believes that
• 58:54it really is agonism of
• 58:55TLR seven. We've not done
• 58:56anything in vivo.
• 58:58It'd be interesting to look
• 58:59at that.
• 59:00The other thing that's interesting
• 59:01is the possibility that, you
• 59:03know, with certain cancers that
• 59:05might express these
• 59:07TLRs,
• 59:08the possibility
• 59:09in patient populations
• 59:11when they're infected by RNA
• 59:13viruses,
• 59:13the possibility
• 59:14of
• 59:16metastatic recurrence events. And so
• 59:18we're interested in looking at
• 59:19that. The yeah. So you
• 59:20could imagine that,
• 59:23single stranded RNA viruses could
• 59:25perhaps,
• 59:26kindle,
• 59:27dormant cells,
• 59:28into the state.
• 59:30Yeah.
• 59:32Hi,
• 59:33Harry.
• 59:35Yeah.
• 59:43Yep.
• 59:47Yeah. There's yeah. I I
• 59:49have no idea. We have
• 59:50not looked at that. I
• 59:51mean, there's some association
• 59:52clearly with,
• 59:54APOE
• 59:55e four
• 59:57patients,
• 59:58tend to have some metabolic
• 01:00:00syndrome.
• 01:00:01So and so that would
• 01:00:03go in the right direction.
• 01:00:04So could this be
• 01:00:06contributing to that? It's it's
• 01:00:07it's very possible.
• 01:00:09Yeah. E twos tend to
• 01:00:10be lean. They tend to
• 01:00:11live to a hundred years
• 01:00:12of age. You know? They're
• 01:00:13very healthy, but
• 01:00:15just, you know, avoid, you
• 01:00:16know, large melanomas.
• 01:00:19Alright. Let's okay.
• 01:00:22Thank you very much.