Pathology Grand Rounds, Sept. 18, 2025 - Joshua Warrick, MD

Transcript

• 00:02Good afternoon. Welcome to pathology
• 00:05ground runs.
• 00:06And,
• 00:08we have today's speaker, doctor
• 00:11Joshua
• 00:12Warwick.
• 00:14And,
• 00:15doctor Warwick graduated from
• 00:19the, West Wayne State University
• 00:21for her medical educations.
• 00:24After that, he he did
• 00:25his residency
• 00:27at initially, actually, it's a
• 00:29Lou,
• 00:32Saint Louis or Saint Louis
• 00:34University, Saint Litton,
• 00:36WashU,
• 00:37and complete his residency. And
• 00:39then which is followed by
• 00:41the fellowship training in Geo
• 00:43Passage
• 00:44as well as surgical massage
• 00:47at University of Michigan.
• 00:49So,
• 00:51he joined,
• 00:53Pennsylvania
• 00:54State University
• 00:56and, starting his career and,
• 00:58risk his ranking
• 01:00for initial assistant professor
• 01:02to full professor.
• 01:04So he,
• 01:06was the
• 01:08vice chair for clinical operation
• 01:12and director of
• 01:14the anatomic massage
• 01:16and also director of the
• 01:18GU Pathology Services
• 01:20there.
• 01:21We're fortunate
• 01:22to have him,
• 01:24you know, join the Yale
• 01:26and he just started his
• 01:27faculty
• 01:29here as director of the
• 01:31GU Passage.
• 01:32So
• 01:34Dodd Warrick did extensive
• 01:37research on
• 01:38the bladder cancers.
• 01:40His research funding was,
• 01:42supported by the NIH
• 01:45and cancer
• 01:46American Cancer Society and also
• 01:49DODs.
• 01:51He,
• 01:53has been invited to give
• 01:54many talks
• 01:56and, at national and international,
• 02:00conference.
• 02:01He path bridge extensively
• 02:03and has more than around
• 02:05the one hundred publications
• 02:08and peer reviewed,
• 02:10journal articles,
• 02:11book chapters.
• 02:13And,
• 02:14so today he gonna share
• 02:16his experience
• 02:17in the breast cancer with
• 02:19us. His topic the the
• 02:22title of his talk is
• 02:23the lineage praxisity
• 02:24of the blood cancer.
• 02:26So without further ado, I
• 02:28will hand over to doctor
• 02:30Warrick. Thank you.
• 02:35Okay. Thank you, Doctor. Kai.
• 02:39So, yeah, so we're going
• 02:39to talk about bladder cancer.
• 02:42So I have I have
• 02:43no conflicts of interest right
• 02:44now.
• 02:48So there's there's two themes
• 02:49to the talk today.
• 02:50The first is gonna be
• 02:51that, bladder cancer demonstrates remarkable
• 02:54lineage plasticity,
• 02:55even without selective pressure from
• 02:57treatment. A lot of cancer
• 02:58types, you know, lung cancer,
• 03:00prostate cancer, they undergo lineage
• 03:02plasticity, but they generally require
• 03:03treatment pressure like prostate cancer
• 03:04and androgen deprivation therapy. Whereas
• 03:06bladder cancer is just plastic.
• 03:08It just does it on
• 03:09its own.
• 03:11And this is key, in
• 03:12in part driven by key
• 03:14transcription factors and interferon gamma
• 03:16signal.
• 03:17So these two statements are
• 03:18are gonna really, you know,
• 03:19like, be the frame from
• 03:20which
• 03:21yeah.
• 03:21So these are the these
• 03:23will be the the ideas
• 03:24that we frame the entire,
• 03:25talk through.
• 03:29Can you guys see the
• 03:30mouse? Good.
• 03:31So
• 03:32transcription factors, like so what
• 03:34are transcription factors? Let's all
• 03:35get on the same page
• 03:36here. So
• 03:37transcription factors are proteins that
• 03:39regulate gene expression. So we
• 03:40have about fifteen hundred of
• 03:42these named, in the human
• 03:44genome. And what they do
• 03:45is they're small proteins or
• 03:46they're proteins that bind gene
• 03:48promoters
• 03:49or enhancers
• 03:50or silencers
• 03:51to alter gene expression.
• 03:53They they recruit, you know,
• 03:54different proteins and stuff. But
• 03:55by and large, what they
• 03:56do is they bind what
• 03:58we call cis regulatory elements
• 04:00to
• 04:01alter expression of a given
• 04:02gene.
• 04:04And these can be extremely
• 04:05powerful in in their So,
• 04:08this is Shinya,
• 04:10Yamanaka. He won the Nobel
• 04:11Prize in twenty twelve. He's
• 04:13kind of a hero to
• 04:14many of us.
• 04:15So he showed something really
• 04:17amazing,
• 04:18that he got the Nobel
• 04:19Prize for. He showed that
• 04:20you could take an adult
• 04:21dermal fibroblast,
• 04:22add four transcription factors, and
• 04:24turn it into an induced
• 04:25pluripotent stem cell.
• 04:27Those four transcription factors are
• 04:29SOX2, WAC4, KLF4,
• 04:31and and MYC.
• 04:33And so him and and
• 04:34others who worked in this
• 04:36space kinda bulldozer an old
• 04:37idea. The old idea was
• 04:39this this idea of Waddington's
• 04:40canals. This idea that that
• 04:42embryologic tissue was,
• 04:44was, destined to become one
• 04:46thing, and it couldn't really
• 04:46back up. It was like
• 04:47once you got to a
• 04:48certain degree of differentiation, you
• 04:50didn't reverse it. And he
• 04:51showed that four transcription factors
• 04:53can reverse it, which is
• 04:54kind of amazing. And that's
• 04:55why I won the Nobel
• 04:56Prize. And it it also
• 04:57goes to show the the
• 04:58incredible power
• 05:00of even a few,
• 05:02small proteins. A few transcription
• 05:03factors can drive tremendous phenotypic
• 05:06change,
• 05:07in in in a cell.
• 05:11And so one of the
• 05:12things that, makes some a
• 05:14subset of these transcription factors
• 05:16so powerful
• 05:18is they're called pioneer facts,
• 05:20pioneer transcription factors. And so
• 05:22as whereas many factors can
• 05:23only bind chromatin that's opened
• 05:25and has their their response
• 05:26elements or the areas they
• 05:27bind, open, pioneer factors can
• 05:30open close chromatin.
• 05:31And they can change the
• 05:33the the epigenomic landscape,
• 05:35truly
• 05:36of of, DNA that they're
• 05:37they're exposed to. And so
• 05:39three of these I've named
• 05:40because they're important in in
• 05:41bladder cancer. So one is
• 05:42Fox a one.
• 05:44That's a pioneer factor that's
• 05:45known to be important in
• 05:46in breast cancer, for example,
• 05:47in ER and AR binding.
• 05:49This one's particularly well studied.
• 05:51It it mimics a link
• 05:52or histone. That's how it's
• 05:53able to open up, close
• 05:55chromatin.
• 05:56Two other ones, GATA3 p,
• 05:57PAR gamma, these are probably
• 05:59also pioneer factors.
• 06:01So these these are,
• 06:02transcription factors, again, that can
• 06:04bind closed chromatin fundamentally on
• 06:06their own, change the the
• 06:08epigenomic landscape,
• 06:10is just a few a
• 06:11few of these these transcription
• 06:13factors.
• 06:16So let's talk about bladder
• 06:17cancer. So so bladder cancer
• 06:19has a very well well
• 06:21known progression.
• 06:23Those in the urology world
• 06:24and the the geopathology world
• 06:25know it well.
• 06:27Starts off as noninvasive disease,
• 06:29like like all cancers
• 06:30do. Either t, TIS or
• 06:32flat carcinoma in situ or
• 06:34TA, which is this noninvasive
• 06:36papillary tumor. It's kinda like
• 06:37a like a bush growing
• 06:39on the surface or like
• 06:39a cauliflower.
• 06:41It invades the lamina propria.
• 06:42That's the superficial connective tissue
• 06:44underneath the urothelium,
• 06:46continues to progress into the
• 06:47the the muscular layer, the
• 06:49muscularis propria of the bladder
• 06:50extends outside of it, and
• 06:52then eventually extends into other
• 06:53organs. So we generally call
• 06:55this non invasive tumor. We
• 06:57call it non muscle invasive
• 06:58if it's up to a
• 06:59t one because it's not
• 07:00quite in muscle. And then
• 07:01we've got muscle invasive bladder
• 07:02cancer whenever it's beyond that.
• 07:04And this is used to
• 07:05to classify,
• 07:06you know, treatments and and
• 07:07and breakdown,
• 07:08you know, different prognostic tests
• 07:11and things. So let's keep
• 07:12this in mind as we
• 07:13as we keep chatting about
• 07:14the rest of the of
• 07:14the talk.
• 07:16And so this is the
• 07:17histology of of most bladder
• 07:18cancers. So this is urothelial
• 07:20carcinoma. This is the stage
• 07:21TA, the non invasive papillary
• 07:23tumor that we talked about.
• 07:24It's got these fibro vascular
• 07:25cores. It's got these this
• 07:27urothelial
• 07:27thickening. There's fusion. It's really
• 07:29just kinda confused cauliflower thing
• 07:31growing off the surface of
• 07:32the urothelium.
• 07:34You've got stage TIS or
• 07:35flat carcinoma in situ. It's
• 07:36these malignant cells sitting on
• 07:38the surface.
• 07:39Then you've got invasive urothelial
• 07:40carcinoma.
• 07:41Kinda looks like urothelial. It's
• 07:43not really as, you know,
• 07:44easy to describe as a
• 07:45lot of cancer types like
• 07:46colon cancer and stuff. But
• 07:48it's these jagged, you know,
• 07:49infiltrative nests of cancer that
• 07:50looks kinda like urothelial. And
• 07:52this is the astrology of
• 07:53t one. So the lamina
• 07:54appropriate invasive stuff, and then
• 07:56anything beyond that.
• 08:00So we can also classify
• 08:02bladder cancers into this, this
• 08:03luminal versus basal,
• 08:06dichotomy.
• 08:07And this got a lot
• 08:07of attention, about ten years
• 08:09ago. And so the luminal
• 08:11cancers tend to express, urothelial
• 08:13genes. So FOXA one, you
• 08:14know, get a three, PPAR
• 08:16gamma, uroplacants.
• 08:17They're enriched in FGFR three
• 08:18mutations. Then there's the basal
• 08:19cancers. Those express basal genes.
• 08:22They're enriched in high molecular
• 08:23weight keratins, TFAPs. They're also
• 08:25enriched in t p TB
• 08:26fifty three gene mutations.
• 08:28And so there was an
• 08:29idea about ten, eleven years
• 08:31ago when they this this,
• 08:32I think, first came out
• 08:33in bladder cancer, and this
• 08:34idea was that we have
• 08:35intrinsic subtypes. We've identified intrinsic
• 08:38molecular subtypes of bladder cancer,
• 08:39and we can start tailoring
• 08:41our treatments to these two
• 08:42different
• 08:43molecular subtypes. And a lot
• 08:44of attention, a lot of
• 08:45money, and a lot of
• 08:46effort went into coming up
• 08:48with ways to to treat
• 08:49these these different molecular subtypes.
• 08:53And I submit to you
• 08:54that that is not the
• 08:55case. They are not intrinsic.
• 08:57This is not a story
• 08:57of intrinsic molecular subtypes that
• 08:59are born basal or born
• 09:00luminal.
• 09:01The story of bladder cancer,
• 09:03the story of luminal and
• 09:04basal is a story of
• 09:05lineage plasticity.
• 09:07And
• 09:08the utility of the luminal
• 09:10and basal is really not
• 09:11in naming them a luminal
• 09:12cancer or basal cancer, but
• 09:14rather the utility is understanding
• 09:16the drivers of luminal differentiation,
• 09:18the drivers of basal differentiation,
• 09:20and how we can exploit
• 09:21those in the future
• 09:22to treat patients with bladder
• 09:24cancer.
• 09:28So so here's some relevant
• 09:29signatures whenever you break these
• 09:31these tumors down. So,
• 09:33what these data are are
• 09:34from the Cancer Genome Atlas,
• 09:36bladder cancer data, which is
• 09:37a nice repository if you
• 09:38wanna just look at stuff.
• 09:41And so what we did
• 09:42is we we took the
• 09:43cancer genome atlas,
• 09:45invasive bladder cancers, divided them
• 09:46into luminal versus basal based
• 09:48on, you know, just the
• 09:49standard, you know, dichotomous, you
• 09:51know, change. And then we
• 09:52looked at single sample gene
• 09:53set enrichment analysis scores,
• 09:55and then compared them, in
• 09:57these different scores. So as
• 09:58expected, luminal cancers tend to
• 10:00have, you know, high expression
• 10:01of luminal genes.
• 10:03Basal cancers tend to have
• 10:04higher expression of basal genes.
• 10:05This is obvious. But there's
• 10:07some things that that pop
• 10:08out that are
• 10:09also of of of value.
• 10:11One is cell cycle activity.
• 10:12It's a little bit higher
• 10:13in basal versus luminal. Another
• 10:15is inflammatory signature. So this
• 10:16is just, you know, our
• 10:17standard hallmark inflammation
• 10:19kinda signature.
• 10:20And it shows that the
• 10:21basal cancers are
• 10:23enriched largely in in an
• 10:25inflammatory signature.
• 10:27And perhaps more importantly,
• 10:29basal tumors are heavily enriched
• 10:31at interferon gamma activity.
• 10:33So we name them after
• 10:34the epithelial cell that they're
• 10:36differentiating toward more basal, more
• 10:37squamous.
• 10:38But whenever we break it
• 10:40down and look at the
• 10:41signatures, interferon gamma is extremely
• 10:43strong in the basal squamous
• 10:45or the basal type of
• 10:46bladder cancer.
• 10:51And molecular subtype also associates
• 10:53strongly with stage.
• 10:54And so to my mind,
• 10:56if these were intrinsic and
• 10:57they were either born basilar,
• 10:58they were born luminal,
• 11:00you would see precursors that
• 11:02are basal. You would see
• 11:04precursors that are luminal and
• 11:05they would be in roughly
• 11:06equal proportion, but that's not
• 11:08the case at all.
• 11:10It turns out that the
• 11:11vast majority of non invasive
• 11:12and even early stage invasive
• 11:14cancers are luminal.
• 11:15Whereas the basal nist, the
• 11:17basal phenotype doesn't really show
• 11:18up until you're in the
• 11:19muscle invasive stage.
• 11:21And so these data are,
• 11:23combined,
• 11:25cancer genome at least data
• 11:26for muscle invasive tumors. And
• 11:28then there's this large European
• 11:29study,
• 11:31Linsgrogg et al. They had
• 11:32the TA and t one
• 11:33cancers. So we the the
• 11:35analysis we're looking at here,
• 11:36was one of my group,
• 11:37we we lumped the cases,
• 11:38normalized them, and then used
• 11:40a UNC classifier that's published
• 11:41at University of North Carolina
• 11:42into either luminal or basal.
• 11:44And as expected,
• 11:45based on some prior studies,
• 11:46the TA were vast majority,
• 11:48ninety five percent plus for
• 11:49luminal.
• 11:51Most of the t ones,
• 11:52which is somewhat surprising, were
• 11:53also luminal, whereas the,
• 11:55the muscle invasive cancers, those
• 11:57that were in the muscular
• 11:58as appropriate or deeper were
• 11:59about half luminal, about half
• 12:00base. So this this to
• 12:02me even, like, looking at
• 12:03it says that this is
• 12:04a a story of these
• 12:05things changing as they develop
• 12:06higher stage disease.
• 12:12Okay.
• 12:13Did I break it? There
• 12:14we go. Okay.
• 12:15And so, there's another piece
• 12:17to the story in the
• 12:18histology that really starts to
• 12:20argue for lineage plasticity playing
• 12:22an important role in this
• 12:24phenotypic evolution. So,
• 12:26in
• 12:27bladder cancer world and the
• 12:28residents, we talked about these
• 12:29this morning and Tuesday morning,
• 12:31they're probably tired of me
• 12:32at this point.
• 12:35So the bladder cancer has
• 12:36many,
• 12:37histologic variants or
• 12:39divergent differentiation or histologic subtypes.
• 12:40There's a bunch of ways
• 12:41to describe these, but they're
• 12:43named histomorphologies
• 12:44that have specific,
• 12:46clinical correlations and complete and
• 12:48specific prognostic, you know, implications.
• 12:50You know, one of those
• 12:51is micropapillary. That's these tiny
• 12:53little, you know, micropapillary nest
• 12:55with attraction clefts. We've got
• 12:56plasmacytoid that's infiltrative single cells.
• 12:58We've got glandular, that means
• 12:59it's it's making glands.
• 13:01Nested, that means it's little
• 13:02tiny blend nest that are
• 13:03easy to to miss. Got
• 13:04small cell that looks like
• 13:05small cell lung cancer. You've
• 13:06got squamous carcinoma that's making
• 13:08keratin, not to be confused
• 13:10with with basal or basal
• 13:11squamous, which is usually what
• 13:12we use to describe, you
• 13:13know, the molecular subtypes. But
• 13:15rather, it's histologic squamous differentiation
• 13:17where it's making making keratin.
• 13:19You have lymphopodeal leoma like,
• 13:21that looks like nasopharyngeal carcinoma,
• 13:23which is an EBV mediated
• 13:25phenomenon with an intense inflammatory
• 13:26infiltrate.
• 13:27This actually isn't EBV mediated.
• 13:29It just looks like a
• 13:30nasopharyngeal tumor that, that is.
• 13:33Then there's sarcomatoid,
• 13:34which is relatively common, but
• 13:36it's differentiating towards a sarcoma.
• 13:38And so this is the
• 13:40the the diversity of, you
• 13:41know, of of bladder cancer
• 13:43for the large part.
• 13:46And these things tend to
• 13:47have, consistent
• 13:49molecular subtypes.
• 13:51Not perfect, but there's a
• 13:53strong tendency here. So, the
• 13:54plasmacytoid, the micropapillary, the nested,
• 13:56they tend to be luminal.
• 13:59The lymphopithelium
• 14:00alike, the squamous, histologically squamous
• 14:02tend to be basal.
• 14:03The sarcomatoid kinda do whatever
• 14:05they wanna do. They're turning
• 14:06into sarcomas, so they're a
• 14:07little wild.
• 14:09The glandular,
• 14:10you know, they make glands.
• 14:11They can be variable as
• 14:12well.
• 14:13You know, in neuroendocrine,
• 14:15the small cell, they have
• 14:16neuroendocrine phenotype, you know, they're
• 14:17not really getting, you know,
• 14:18you're the only one anymore.
• 14:20And so there's, you know,
• 14:22not only and my phone
• 14:23is just buzzing. Hold on.
• 14:25Spam. Do you guys get
• 14:26spammed constantly? It's making me
• 14:27crazy. It's, like, half ninety
• 14:29percent of my calls.
• 14:31Okay. So, so back to
• 14:32this. So,
• 14:35these histologic variants
• 14:37strongly associate with these specific
• 14:39molecular subtypes.
• 14:43And so, you know, back
• 14:45in twenty nineteen, we had
• 14:46a study published in European
• 14:48urology that got a lot
• 14:49of attention.
• 14:50Because
• 14:51before the study came out,
• 14:52there was this, like I
• 14:53said, there was this
• 14:54intense interest in intrinsicness
• 14:57of of these molecular subtypes.
• 14:59And there have been clinical
• 15:00trials that were being developed,
• 15:01putting people in these different
• 15:02subtypes and then treating them
• 15:03differently based on the the
• 15:05the classifications.
• 15:07And this study was one
• 15:07of the first to really
• 15:08kind of, you know, shoot
• 15:09a hole on that idea.
• 15:10And so what we did
• 15:11is we took about three
• 15:12hundred,
• 15:14consecutive,
• 15:15cystectomy cases. We mapped them
• 15:16out.
• 15:17We identified all the different
• 15:19areas of histologically distinct invasive
• 15:20carcinoma. We,
• 15:22named them. We identified areas
• 15:23of non invasive carcinoma,
• 15:25and we performed molecular subtyping
• 15:27on them using, the,
• 15:30a group or a a
• 15:31schema developed by the Lund
• 15:32University,
• 15:35in,
• 15:36in in Switzerland. No. Sweden.
• 15:38In Sweden. They would make
• 15:39they would be angry if
• 15:40I said they were Swiss.
• 15:41But they're, they developed this,
• 15:44system for for molecular classification.
• 15:47And so they have a,
• 15:48you know, this kind of
• 15:49type where they call them
• 15:50urothelial like and genomically unstable,
• 15:51but really practically speaking, those
• 15:53are those are luminal. They
• 15:54have a basal squamous equivalent
• 15:56to a basal on the
• 15:56other systems. They have mesenchymal
• 15:58like that is really rare,
• 15:59kind of like sarcoma, and
• 16:00then they have non type.
• 16:01What we did is we
• 16:02performed molecular subtyping on all
• 16:04the different areas from these
• 16:05bladders that we mapped out.
• 16:07And this is how we
• 16:08demonstrated those that had,
• 16:10you know, histologic diversity.
• 16:12And so each column is
• 16:14a histology, we've got conventional
• 16:15urothelial, squamous, micropapillary, glandular, etcetera.
• 16:18Each row is a patient,
• 16:20and then each color is
• 16:22a a molecular subtype.
• 16:24So what we we see
• 16:25here is that if we
• 16:27go along each row, we
• 16:28can see the different histologies
• 16:29each patient had. Gray means
• 16:31they didn't have it. Then
• 16:32we can see what molecular
• 16:33subtype they were assigned to.
• 16:35And what we can see
• 16:35here is that there's diversity.
• 16:37They're not all the same.
• 16:39They're not intrinsic.
• 16:40They differ. So for example,
• 16:42this first row of the
• 16:43conventional urothelial was a luminal
• 16:45subtype and associated histologically squamous
• 16:47was basal squamous.
• 16:49Same thing with these guys.
• 16:51Here, this was also luminal.
• 16:52These are basal squamous. The
• 16:53conventional was urothelial. The histologically
• 16:56squamous was a basal subtype.
• 16:58And then it gets even
• 16:58more interesting. Some have, you
• 17:00know, like
• 17:01a, you know, basal conventional
• 17:03urothelial, basal squamous, then there's
• 17:04a micropapillary in a different
• 17:05area that was luminal.
• 17:07So this
• 17:08really is one of the
• 17:09first indications that, like, no.
• 17:10These things are not just
• 17:11born one way and they
• 17:13stay that way. There's plasticity
• 17:14going on. These are changing
• 17:15as they evolve from different
• 17:17to different,
• 17:18histologies.
• 17:19And in fact, thirty nine
• 17:20percent of cases
• 17:22in which there was histologic
• 17:23diversity also demonstrated a difference
• 17:25in molecular subtype, and it's
• 17:27extremely common in those that
• 17:28had a basal squamous component.
• 17:30There was even one basal
• 17:31squamous component. Nearly eighty percent
• 17:33had another area that was
• 17:34luminal.
• 17:35It could have been the
• 17:36non invasive component, could have
• 17:37been another, but it really
• 17:38demonstrated that there's a a
• 17:39tremendous amount of,
• 17:41diversity in terms of the
• 17:42luminal versus basal,
• 17:44dichotomy in these things.
• 17:46And so that led to
• 17:47a different framework than they're
• 17:48born luminal or they're born
• 17:50basal, but rather they're mostly,
• 17:51if not all born, luminal.
• 17:53And certainly we can't be
• 17:54absolute here. This is
• 17:56this is cancer. It kind
• 17:56of does what it wants.
• 17:57We've seen squamous displays in
• 17:59the bladder, but by and
• 18:00large,
• 18:01most of these things start
• 18:02off, we think is not
• 18:03invasive. You're of the ileal
• 18:04carcinoma that's luminal. It invades
• 18:06like we see with the
• 18:07t one cancers that are
• 18:08luminal And early on, it's
• 18:10the invasive cancers also luminal.
• 18:11And then from there, it
• 18:13it undergoes lineage plasticity
• 18:15to a basal subtype.
• 18:16And those can turn into
• 18:17histologic,
• 18:19basal variance. And then the
• 18:20urothelial conventional can turn into
• 18:22invasive histologic luminal,
• 18:24variance.
• 18:25So this is the framework
• 18:26that we're working from now.
• 18:27And so,
• 18:29you know, the one of
• 18:30the questions as well, how
• 18:31do you know that it's
• 18:32not just, like, you know,
• 18:33collision tumors? Why aren't the
• 18:34why were the luminal micropapularies
• 18:36and the histologically squamous basal
• 18:38ones? Are they two separate
• 18:39tumors
• 18:40that just collided? And it's
• 18:42just a coincidence, and I
• 18:43think the answer is no.
• 18:44And we we answered that
• 18:46with, the paper, in nature
• 18:47communications a few years back.
• 18:49This was a collaboration with,
• 18:52Memorial Sloan Kettering, Hikma Del
• 18:53Amadi, and,
• 18:54Wenohu,
• 18:55were two of the main
• 18:56collaborators there. And so what
• 18:58we did in in this
• 18:59study is we,
• 19:01gathered
• 19:02a number of patients, twelve
• 19:04in total, who had invasive
• 19:06urothelial carcinomas with a clearly
• 19:08conventional urothelial carcinoma component
• 19:10and a clearly squamous invasive
• 19:12squamous carcinoma component. And we
• 19:14did,
• 19:16you know, comprehensive genomic evaluation
• 19:18on them or at least
• 19:19what was called comprehensive at
• 19:20the time. So we did,
• 19:21whole exome sequencing. We did
• 19:23RNA sequencing.
• 19:25We did this on a
• 19:25lot more than twelve, but
• 19:26the we only kept the
• 19:28twelve that had high quality
• 19:29RNA. It's difficult to get
• 19:30high quality RNA out of
• 19:31tissue blocks.
• 19:33So we we performed RNA
• 19:35sequencing,
• 19:35sequencing.
• 19:37And then we we, you
• 19:38know, saw the results. And
• 19:39this is the first thing
• 19:40we saw
• 19:42was the the the pair
• 19:43urothelial carcinomas and squamous carcinomas,
• 19:46were clonally related, but subclonally
• 19:48distinct.
• 19:49And so they were all,
• 19:50all twelve of them had
• 19:52multiple,
• 19:54cancer driver genes that were
• 19:55identical between the two different
• 19:56histologies. So for example, we've
• 19:58got, you know, we're demonstrating
• 19:59here, we've got the urothelial
• 20:00component, the squamous component
• 20:02Here, we've got,
• 20:05you know, like, the common
• 20:06precursor presumed common precursor.
• 20:08Here's the common mutations between
• 20:10the two. Here's common or
• 20:12here's mutations exclusive to the
• 20:13squamous component. Here's, mutations unique
• 20:16to the erythema component. And
• 20:17so these are just three
• 20:18example cases.
• 20:20So in this one, they're
• 20:22identical.
• 20:23You know, cancer driver mutations
• 20:25in Fgfr3, TP53, PIK3CA, and
• 20:27then thirty one more that
• 20:28were common to both. They
• 20:29were unique mutations in both
• 20:31the squamous and urothelial components.
• 20:33Same with this one, also
• 20:34f g f r three
• 20:35mutations. And remember, f g
• 20:36f r three is more
• 20:37associated with luminal cancer, and
• 20:38these developed histologically squamous disease.
• 20:40So
• 20:41this is really screaming lineage
• 20:43plasticity to my mind. So
• 20:44you've got f g f
• 20:45r three, you've got ATM,
• 20:47we've got PIK3CA.
• 20:48Again, there's some unique mutations
• 20:50in the squamous component including
• 20:51TP53.
• 20:53And then the same with
• 20:53this one, it's not an
• 20:54FGFR3 mutant cancer, but it's
• 20:57got,
• 20:58multiple
• 20:59known cancer driver genes between
• 21:01both the urothelial component and
• 21:03the paired histologically squamous component.
• 21:05Really, I think
• 21:07proving that these things iterize
• 21:09from a common precursor despite
• 21:10their distinct
• 21:11histology.
• 21:15So what about, your molecular
• 21:17subtype, this luminal versus bathel
• 21:20dichotomy? So we have the
• 21:21RNA Seq data. We only
• 21:22selected cases that had, you
• 21:24know, high quality RNA sequencing
• 21:25data.
• 21:26Anew did a few things.
• 21:28First, we put them into
• 21:29categorical,
• 21:30luminal versus basal subtypes
• 21:32based on the TCGA system
• 21:34and and centroid analysis. So
• 21:35that's shown here.
• 21:37Each column is a patient.
• 21:39Each row is a histology.
• 21:40So this is the urothelial
• 21:41part. This is the squamous
• 21:43part. The color is the
• 21:44subtype.
• 21:45So that's kinda orange salmon
• 21:47color is basal.
• 21:48Then the blue and the
• 21:49green are both luminol subtypes.
• 21:51There are a lot of
• 21:51different systems. The TCGA has
• 21:53one that's subclassified as liminal,
• 21:54but we'll pull them, for
• 21:56the sake of, you know,
• 21:56simplicity in the discussion.
• 21:58And four of the twelve,
• 22:01switched subtype,
• 22:02whenever we use the centroid
• 22:03analysis.
• 22:04So using the subtype as
• 22:06a categorical variable centroid analysis,
• 22:09we saw that, you know,
• 22:10a third of them were
• 22:11different between the two. This
• 22:12is consistent with what we
• 22:13found earlier.
• 22:15And and
• 22:16I think more interestingly, we
• 22:17found that it it it
• 22:19was went beyond that. So,
• 22:22we performed, you know, single
• 22:24sample gene set of enrichment
• 22:25analysis with gene lists of
• 22:26basal genes and and luminal
• 22:28genes to give a quantified
• 22:29or quantitative score of baseness
• 22:32and luminous.
• 22:34In every single case, the
• 22:35squamous component had a higher
• 22:37basal score than the luminal
• 22:39than the urothelial component.
• 22:41And in every single case,
• 22:42the squamous component had a
• 22:43lower luminal score than the
• 22:45urothelial component.
• 22:47Really showing that this it
• 22:49it's really, I think, putting
• 22:50more and more cracks in
• 22:51this idea of luminal versus
• 22:52basal dichotomy. That that we
• 22:54we have not only that
• 22:55a third of them change
• 22:57subtype,
• 22:58but that all of them
• 22:59are more basal, and all
• 23:00of them are less liminal.
• 23:02So this is looking more
• 23:03like a continuous variable than
• 23:04it is like a categorical
• 23:05variable.
• 23:08How about, you know, we
• 23:09talked about that,
• 23:11immunity,
• 23:12before.
• 23:13We talked about how they're
• 23:14the interferon gamma,
• 23:17and just looking at these
• 23:18subtypes is higher in the
• 23:19basal subtype, and it's pretty
• 23:20significantly higher. So what about
• 23:22immune
• 23:24subtype,
• 23:25between the urothelial and the
• 23:26histologically squamous components?
• 23:28So, we handle this in
• 23:30a few ways. One of
• 23:31them is we assign them
• 23:33what was called an immune
• 23:34subtype. So this group, you
• 23:35know, Thorson et al published
• 23:37a nice paper in immunity
• 23:38in twenty eighteen,
• 23:39that I found is really
• 23:40useful where they assigned six
• 23:41different inflammatory
• 23:43subtypes,
• 23:44to cancers, you know, based
• 23:46on on, you know, the
• 23:47TCGA data.
• 23:48They named them c one
• 23:49to c six.
• 23:51C two is an interferon
• 23:52gamma dominant
• 23:53in, inflammatory subtype. And then
• 23:56there's a number of other
• 23:56ones. C one's wound healing.
• 23:58C three is inflamed.
• 23:59They have a number of
• 24:00other named ones.
• 24:01And so we used our
• 24:03RNA sequencing data to put
• 24:04the the tumors into,
• 24:07these thorasin immune subtypes.
• 24:09And not surprisingly, we found
• 24:10all the histologically squamous areas,
• 24:13where the c two interferon
• 24:15gamma dominant
• 24:17subtype. We similarly found that,
• 24:20nine of the twelve,
• 24:21urothelials were also, you know,
• 24:23the c two interferon,
• 24:25dominant, but three of them
• 24:26were not. So there was
• 24:27some heterogeneity,
• 24:29in the immune subtype
• 24:31in in these tumors.
• 24:33PDL one also differed. And
• 24:34so,
• 24:36PDL one is a a
• 24:37target of interferon gamma. So
• 24:39interferon
• 24:40gamma signaling induces higher PDL
• 24:42one expression. This is pretty
• 24:43well well known. And the
• 24:44squamous component on average had
• 24:46higher PDL1 expression of the
• 24:47urothelial.
• 24:49So
• 24:50not only are they histologically
• 24:52different,
• 24:53and they're different in their
• 24:54immune sub their their molecular
• 24:55subject, their immune subtype seems
• 24:56to be different as well.
• 24:58So So this lineage plasticity
• 24:59is taking on a kind
• 25:00of a life of its
• 25:01own.
• 25:04And so what does it
• 25:05matter at this point? Right?
• 25:07That's always the question I
• 25:08like to ask myself. Why
• 25:09is this important?
• 25:10And so,
• 25:12Hikmat Elhamdi,
• 25:13from Sloan Kettering, he's a
• 25:14pathologist out there. So he
• 25:16was part of, one of
• 25:17the clinical trials that looked
• 25:19at at atezolizumab
• 25:20in metastatic and locally invasive
• 25:22bladder cancer.
• 25:23And so he was one
• 25:24of the guys who reviewed
• 25:25the slides and confirmed its
• 25:26bladder cancer and all that
• 25:27kind of stuff. And so
• 25:29he he he had an
• 25:30idea. He's like, alright. So
• 25:31we see this immune heterogeneity
• 25:33in our tumors. So I'm
• 25:35gonna
• 25:36grab the the tumors from
• 25:37Sloan Kettering that were part
• 25:38of this this clinical trial.
• 25:40I'm just gonna look at
• 25:41them blinded, and I'm gonna
• 25:42break down, is there histologic
• 25:44heterogeneity or not? Or is
• 25:46there morphologic heterogeneity?
• 25:47And he defined this precisely.
• 25:49He said, is it a
• 25:50named are there more than
• 25:51one named histology
• 25:53in this in this tumor?
• 25:55Yes or no? Just kinda,
• 25:56you know, made us hash
• 25:57marks. And then we did
• 25:58the statistics in terms of,
• 26:01response or no response.
• 26:03And it turned out those
• 26:04with,
• 26:05morphologic heterogeneity
• 26:08were enriched in the non
• 26:09responder group.
• 26:11Whereas those who
• 26:13lacked morphologic heterogeneity
• 26:15were enriched in the responder
• 26:17group. Certainly a small n,
• 26:18but it was significant and
• 26:19it was it was impressive.
• 26:21And so one of the
• 26:22thoughts here is that maybe
• 26:24we know there's this immune
• 26:26heterogeneity or we've seen this
• 26:27immune heterogeneity.
• 26:29Are these tumors with immune
• 26:30heterogeneity in spatially distinct areas
• 26:32better able to adapt
• 26:34to immune checkpoint inhibitor because
• 26:35they have
• 26:37this greater diversity to
• 26:39get around it. And so
• 26:41this is likely clinically important.
• 26:44So what about the experimental
• 26:46data?
• 26:46You know, this is all
• 26:47observational. You know, we've we're
• 26:49looking at human tumors. This
• 26:50is really pointing at,
• 26:52lineage plasticity as being being
• 26:53important here.
• 26:57Being the driver of, you
• 26:58know, the luminal versus basal
• 27:00dichotomy.
• 27:01What about experimental data? So
• 27:04I'll give it away, the
• 27:05experimental data is pretty good
• 27:07and we I think we've
• 27:08identified some main drivers
• 27:10of both luminal differentiation
• 27:12and of basal differentiation. And
• 27:14I think the luminal differentiation
• 27:16is driven by a few
• 27:17transcription factors and basal
• 27:19differentiation appears driven heavily by
• 27:21interferon gamma signaling. So I'll
• 27:23tell you the evidence we've
• 27:24got right now. So this
• 27:25is an older paper, we
• 27:26published twenty sixteen,
• 27:29where we,
• 27:31asked the question, can we
• 27:32change the molecular subtype
• 27:34of bladder cancer cell lines?
• 27:36So what we did is
• 27:37we,
• 27:39took the cancer cell line
• 27:40encyclopedia data, which is, you
• 27:41know, publicly available expression data
• 27:43for cell lines. We put
• 27:44them into two categories, either
• 27:45basal cell lines, luminous or
• 27:47luminal cell lines. We threw
• 27:48away a bunch of them
• 27:49because they didn't really, you
• 27:50know, have the expression pattern
• 27:51of either.
• 27:53Then we picked one of
• 27:53the basal cell lines, five
• 27:55six three seven. This is
• 27:57the CCLE one. This is
• 27:58ours.
• 27:59And we
• 28:01saw if we could push
• 28:02it from basal to luminal
• 28:03using transient transfection with,
• 28:07Foxy one, get a three,
• 28:08and then use rosaglitazone
• 28:10to is a PPAR gamma
• 28:12agonist to see if activation
• 28:13of these three transcription factors
• 28:15could push
• 28:16it. And and they did,
• 28:17and they impressively so. So
• 28:19here's our,
• 28:21five six three seven controls.
• 28:23This is FOXA one alone.
• 28:25This is got a three
• 28:25alone. This is FOXA one
• 28:26and got a three. Together
• 28:27you kinda get the picture.
• 28:29This is a centroid plot,
• 28:30so this is a basal
• 28:31centroid correlation.
• 28:33The higher it is, the
• 28:34more basal,
• 28:35the cell line is. Here's
• 28:37the luminal centroid correlation, the
• 28:38higher it is, the more
• 28:39luminal,
• 28:40the cell line is. And
• 28:42as we add more transcription
• 28:43factors, we we push them
• 28:45more in the liminal direction.
• 28:47To the point where when
• 28:47we've added all three transcription
• 28:49factors,
• 28:50we consistently, and at least
• 28:51two replicates, and this is
• 28:52we've seen beyond this, we
• 28:54push them to a a
• 28:55luminal phenotype.
• 28:56And
• 28:58so, go back to Yamanaka
• 28:59factors, we can turn a
• 29:00fibroblast
• 29:01into an induced pluripotent stem
• 29:03cell with four transcription factors.
• 29:05It looks like we can
• 29:06turn a bladder cancer cell
• 29:07line from a basal type
• 29:08to a luminal
• 29:10type with three.
• 29:16And so what are the
• 29:17opposite direction?
• 29:18So
• 29:19the next thought is we
• 29:20FOXA1 seems like the best
• 29:22established. We knew more about
• 29:23FOXA1
• 29:24than any of these other
• 29:25factors.
• 29:27So we're like, what happens
• 29:28if we knock out FOXA1?
• 29:29So we knocked out FOXA1
• 29:31in a couple of luminal
• 29:31cell lines,
• 29:33and it doesn't just turn
• 29:34into basal.
• 29:35It doesn't just you know,
• 29:36you knock out FOXA1, they
• 29:37become basal. That doesn't happen
• 29:38at all. There's buffering in
• 29:40there.
• 29:41Just kinda like they're not
• 29:42even that much more more
• 29:43basal really when you knock
• 29:44out Fox a one. And
• 29:45we've done this with visa
• 29:46lines,
• 29:47but something else does happen.
• 29:49That's kinda telling, kinda fascinating
• 29:51and probably important.
• 29:53And I think it comes
• 29:54to, you know, what I
• 29:55was saying earlier about,
• 29:57molecular subtypes being not as
• 30:00important as a categorical things.
• 30:01We put things in and
• 30:02treat people differently, but giving
• 30:04us an understanding and insight
• 30:06into where the Achilles
• 30:07heels
• 30:08in bladder cancer. What are
• 30:10the super important
• 30:11little triggers that we can
• 30:12we can work with? And
• 30:13I think Fox a one
• 30:14is one of them. And
• 30:15I think
• 30:16that one of the pieces
• 30:18of data we have that
• 30:18is if you knock out
• 30:19Fox a one in a
• 30:21luminal cell line, it doesn't
• 30:22become basal,
• 30:24but it turns up interfering
• 30:26gamma signal,
• 30:28which is kinda weird. So
• 30:30like these are cell lines,
• 30:31these are not CD eight
• 30:32cells.
• 30:33These these are just epithelial
• 30:34cells,
• 30:36but the the basal cell
• 30:37lines tend to over express,
• 30:40interferon gammas even though they're
• 30:41just cell lines, whereas the
• 30:42luminal ones have lower expression.
• 30:44But if you take a
• 30:44luminal one, you knock out
• 30:46FOXA1, you turn up interferon
• 30:48gamma signaling, you don't do
• 30:49it subtly.
• 30:51So here's the here's the
• 30:52data for that. So,
• 30:53here we used the UMEC
• 30:55one,
• 30:56luminal cell line.
• 30:58There's this is our expression
• 30:59data from RNA sequencing,
• 31:01ROSA genes. These are some,
• 31:02what are called the interferon
• 31:04response genes.
• 31:05Blue is higher, red is
• 31:06lower. Wen Ho did that.
• 31:07Blame him. He's one of
• 31:08our collaborators. I wish we
• 31:09did on the opposite.
• 31:11But what it showed is
• 31:12that on our knockouts, we
• 31:13see this substantial increase in,
• 31:16expression of interferon
• 31:18responsive genes in our knockouts.
• 31:20And and we saw the
• 31:21same with interferon alpha signaling,
• 31:22interferon gamma signaling using, you
• 31:24know, standard GSEA.
• 31:26We also saw that when
• 31:27we knocked out FOXA1 PDL1
• 31:29expression went up, which we
• 31:30know is a a downstream,
• 31:31you know, product of interferon
• 31:33gamma signaling.
• 31:34So so this is, I
• 31:35I think, a potentially important
• 31:37thing, and we're digging into
• 31:38this a lot. We talked
• 31:38a little bit about this
• 31:39on on Tuesday at the,
• 31:41the the rip talk. But,
• 31:43this is, I think, an
• 31:44important thing
• 31:45in developing, you know, customized
• 31:46treatments for for, for bladder
• 31:49cancer. And we're we're continuing
• 31:50to work on on what
• 31:51this means.
• 31:54And it, you know, raises
• 31:55another question too. So, okay,
• 31:56we this is this is
• 31:57kind of strange. We knock
• 31:58out FOXA1
• 31:59in epithelial cancer cell lines.
• 32:02We drive interferon gamma signaling
• 32:04even though there's no inflammatory
• 32:05cells to be found.
• 32:07And we know that
• 32:08if through through some other
• 32:09data that, you know, basal
• 32:11squamous bladder cancers not only
• 32:12have higher interferon gamma signaling,
• 32:14but they're inflamed. There are
• 32:15more, you know, immune cells
• 32:16in them.
• 32:18How's it getting there? And
• 32:19the question arose, like, does
• 32:20what does interferon gamma do
• 32:23to luminal cell lines? So
• 32:25we know if we knock
• 32:25out FOXA1 in the luminal
• 32:27cell line, we turn into
• 32:28a basal cell line. What
• 32:29if we take a liminal
• 32:30cell line, we treat it
• 32:31with interferon gamma, what happens?
• 32:33And it turns out they
• 32:34turn basal
• 32:36or they get more basal
• 32:37at least. So what we
• 32:38did, here's three, liminal cell
• 32:40lines, r t one one
• 32:41two,
• 32:44SW seven eighty.
• 32:46These are our genes. These
• 32:47are our luminal genes. These
• 32:49are our basal genes.
• 32:51And so you can see
• 32:52here that, you know, with
• 32:53the knockouts, these are the
• 32:54knockouts here, these are the
• 32:55wild types. We see this
• 32:57increased expression
• 32:58in basal genes consistently. I
• 33:00mean, sorry. These are the,
• 33:01these are interferon gamma treated.
• 33:03I apologize. So control interferon
• 33:05gamma treated. You see this
• 33:06increased expression
• 33:08of the the basal genes,
• 33:10including the molecular the, you
• 33:12know, keratin six, you know,
• 33:13keratin fourteen, these high molecular
• 33:15weight keratin. It's not just
• 33:16the inflammatory things. It's the
• 33:17epithelial things as well that
• 33:19are going up.
• 33:21And so we also, you
• 33:23know, looked at centroid analysis
• 33:24similar to we did in
• 33:25that that,
• 33:26study where we turned luminal
• 33:28cell lines into basal cell
• 33:29lines.
• 33:30So here's control, here's interferon
• 33:32gamma treatment.
• 33:33This is showing the correlation
• 33:34of the basal centroid. So
• 33:36that means with centroid analysis,
• 33:38our sequencing data, the higher
• 33:39it is, the the more
• 33:41the more basalts becoming, the
• 33:42closer it's getting to that
• 33:43basal centroid.
• 33:46And so with each one
• 33:47of them, they got a
• 33:47little they got closer to
• 33:48that basal centroid and two
• 33:49of the threes flipped flipped.
• 33:51Meaning, they got closer to
• 33:53the basal centroid than the
• 33:54they were to the luminal
• 33:55centroid. And those two were,
• 33:57UMBC one and RT one
• 33:59one two, whereas the SW
• 34:00seven eighty cell line didn't
• 34:01flip, but it just got
• 34:02closer.
• 34:04And similarly,
• 34:06you know, the interferon dominant,
• 34:07we also put these in
• 34:08the inflammatory
• 34:09thoresen subtypes. And, you know,
• 34:10totally as you'd expect they'd
• 34:12be. They got much much
• 34:13more of this interferon gamma
• 34:14dominant,
• 34:15subtype.
• 34:16So that was interesting. So
• 34:17interferon gamma appeared to be
• 34:19driving the basal. It wasn't
• 34:20just that if you knock
• 34:21out FOXA1, it becomes more
• 34:22interferon gamma dominant. There's there's
• 34:24a loop here. There's some
• 34:25kind of a feedback mechanism
• 34:27going on.
• 34:29Then we thought, well, maybe
• 34:30we could do the opposite.
• 34:32So,
• 34:33what if we take basal
• 34:34cell lines and we inhibit
• 34:36interferon gamma signaling?
• 34:38So we did, we took,
• 34:40the basal cell line stabber,
• 34:42which is a well established
• 34:43bladder cancer basal cell line,
• 34:45and we treated it with
• 34:46the the JAK inhibitor ruxolitinib.
• 34:48Ruxolitinib is it's a relatively
• 34:50new drug. It's used to
• 34:51treat myelofibrosis,
• 34:52a couple other heme diseases.
• 34:54It's a pan JAK inhibitor
• 34:56and it's JAK one and
• 34:56JAK two, and those are
• 34:58the receptor tyrosine kinases that
• 34:59carry out interfering gamma signaling.
• 35:02So we can treat it
• 35:03as an interfering gamma inhibitor.
• 35:05And so whenever we treated
• 35:06the basal cell lines GABA
• 35:07with the JAK inhibitor,
• 35:09it didn't flip all the
• 35:10way to luminol, but it
• 35:12became more luminol.
• 35:13So it it lowered expression
• 35:15of basal genes and it
• 35:16increased expression of of, luminal
• 35:18genes. So we've got here
• 35:19is our ruxolitinib treated group,
• 35:22our,
• 35:23control group. Here are our
• 35:25genes. Here are our basal
• 35:26genes. Here are our luminal
• 35:27genes. And you can see
• 35:29that whenever you're treated with
• 35:30ruxolitinib,
• 35:32the ex
• 35:33expression of the
• 35:35the the luminal genes went
• 35:36up pretty substantially.
• 35:38Not only to flip it
• 35:39because these SCABER cell lines,
• 35:41they're whenever you grow these
• 35:42things in xenografts, they're making
• 35:43keratin. They are very, very
• 35:45basal bladder cancers. Even those,
• 35:47we are able to to
• 35:48push more in in the
• 35:50luminal direction with ruxolitinib.
• 35:53And you'll notice among these
• 35:54luminal genes, FOXA1 went up,
• 35:57GATA3 went up, and PPAR
• 35:58gamma largely went up. And
• 36:00you'll recall those are the
• 36:01three that we used to
• 36:03drive the basal cell line
• 36:05to a liminal liminal type.
• 36:08And, you know, as expected,
• 36:09we we also looked at
• 36:11this, these these inflammatory
• 36:13subtypes, the thoras and ones,
• 36:14and we really diminished the
• 36:16the the, probability of a
• 36:18interferon gamma, you know, inflammatory
• 36:20subtype.
• 36:25So, so in summary,
• 36:27you know, I think that
• 36:29with the evidence we've collected,
• 36:30you know, bladder cancer is
• 36:32phenotypically diverse, we can know
• 36:33this, and it's likely a
• 36:35result of lineage plasticity. It's
• 36:36not that they're just born
• 36:37one way or the other.
• 36:38It's a plastic process that
• 36:39starts off probably from stuff
• 36:41that's differentiated toward urothelium.
• 36:44And then that lineage plasticity
• 36:45toward the basal phenotype appears
• 36:46driven by interferons
• 36:48or interferon gamma, whereas,
• 36:50differentiation toward the liminal phenotype
• 36:52appears to be driven by
• 36:53by these these key transcription
• 36:55factors.
• 36:57And that is all I
• 36:58have to say.
• 36:59Thank you. Happy to take
• 37:01questions.
• 37:09Yeah. Go ahead. I have
• 37:11two questions. Yes.
• 37:29Yes. No. So it was
• 37:31it's just kind of a
• 37:31theoretical,
• 37:32you know, precursor, so I
• 37:34can I can go back?
• 37:38Yeah. So it was like
• 37:41Yes. Yeah. So this one.
• 37:42So,
• 37:43you know, we we sequence
• 37:44both. These are part of
• 37:45the same tumor. So think
• 37:46of this as a tumor.
• 37:47You know, one area squamous,
• 37:48one area is epithelial.
• 37:50And so spatially distinct parts
• 37:51of the same physical mass,
• 37:54and we sequenced them. And
• 37:55so this dot here indicates,
• 37:58the the common precursor. You
• 38:00know, think of this like
• 38:01an evolutionary phylogeny.
• 38:03This would be the the
• 38:04the common ancestor of the
• 38:05two, and we can say
• 38:06it's the common ancestor because
• 38:07it shares the key driver
• 38:08mutations. And then this thing
• 38:10is, we just added that
• 38:11so you'd have, like, you
• 38:12know, some idea of, like,
• 38:13you know, the cell from
• 38:14which they both arose. It's
• 38:15more like eye candy. It
• 38:16really doesn't, you know, add
• 38:18much to the meaning of
• 38:19the figure.
• 38:33So this this was the
• 38:34biopsy. So those were, either
• 38:36metastatic or locally advanced, urothelial
• 38:39carcinoma that that we were
• 38:40looking at. So these would
• 38:41have been biopsies from
• 38:43it it they weren't super
• 38:45strict in terms of the
• 38:46histology that was required,
• 38:48but it was just some
• 38:49kind of biopsy demonstrating that
• 38:50was either locally advanced or
• 38:52or a metastatic disease.
• 38:54And you can't file with
• 38:55the GHAH and E even
• 38:57for our own Even from
• 38:58that. Yep.
• 39:04Anyone else?
• 39:06Excellent.
• 39:07Okay.
• 39:09So I have a question
• 39:09about your diagram.
• 39:19There is, but you you
• 39:20gotta draw the line somewhere.
• 39:23Yeah. So it's
• 39:24we we just pick these
• 39:25as kinda like to illustrate
• 39:26because their genes that that
• 39:27are important or are well
• 39:29known.
• 39:30But it was it's kind
• 39:31of amazing. Like, some of
• 39:32these cases, they were they
• 39:33were more different than they
• 39:34were similar.
• 39:36In some cases, they were
• 39:37more similar than they were
• 39:37different. But, yeah, we just
• 39:39kinda had to, you know,
• 39:39start somewhere, so that's why
• 39:41we we limited it. It's
• 39:42easy to find the current
• 39:44Mhmm. And then being on
• 39:45as one week.
• 39:51Yeah. Yeah. I mean and,
• 39:51well, I think, you know,
• 39:52it it
• 39:53possibly informing new biology.
• 39:56Because, you know, like, for
• 39:57example, this one, the TP
• 39:58fifty three, gene mutation was
• 40:00was private to the squamous
• 40:01part. And, you know, squamous
• 40:03is thought to be more
• 40:03aggressive, and I guess that
• 40:04makes sense.
• 40:06But, you know, in this
• 40:06one, the TP fifty three
• 40:08gene mutation and in this
• 40:09one, it was it was
• 40:10common to both. So it's
• 40:11it's really hard. I I
• 40:12think one of the things
• 40:13we've learned from the past
• 40:14ten years is that it's
• 40:15very hard to identify, you
• 40:17know, key driver mutations
• 40:19that are responsible
• 40:20or relate to squamous histology.
• 40:22It really seems to be
• 40:23more of a, you know,
• 40:24a transcriptional
• 40:25process than it is a
• 40:26mutation driven process, if you
• 40:28will.
• 40:31So I have a question.
• 40:33You know, why we we
• 40:35validate
• 40:36that
• 41:31Yeah. So it's it's a
• 41:32great question and much,
• 41:34there's been a lot of
• 41:34conversation about this in the
• 41:35bladder cancer world, a lot
• 41:37of lot of ink spilled
• 41:38or, you know, virtual ink
• 41:39spilled out on the topic.
• 41:40And so,
• 41:42one of the things that's
• 41:43popped out is that
• 41:45it's very difficult to create
• 41:47reliable
• 41:48categorical
• 41:49tests
• 41:50for basal versus luminal.
• 41:52And I think it relates
• 41:53to, you know, us showing
• 41:54this as a, a it's
• 41:56not a categorical variable. Truly,
• 41:57it's a it's a continuous
• 41:59variable. And so I think
• 42:00one of the the best
• 42:01stories well, not best stories,
• 42:02but one of the most
• 42:03telling stories,
• 42:04is with the DECIPHER bladder
• 42:06cancer test. So decipher bladder
• 42:08cancer test was something that
• 42:09was, marketed,
• 42:12I believe it was even
• 42:13FDA approved.
• 42:14I don't remember. I have
• 42:15to look that up, but
• 42:16it was marketed,
• 42:17to to assign molecular subtypes
• 42:19to muscle invasive bladder cancer
• 42:21to to decide if it's
• 42:23going to be,
• 42:24responsive to neoadjuvant cisplatin based
• 42:26chemotherapy, which is the standard.
• 42:28And they showed that,
• 42:31the their their basal subtype
• 42:33had better response than their
• 42:34non basal sub subtype. So
• 42:36they were they were advocating
• 42:37using this to make this
• 42:38key clinical decision, and that
• 42:40was the decipher group.
• 42:42Then the the Lund group
• 42:43I told you about came
• 42:44out with another project that
• 42:46was very large and very
• 42:47well done.
• 42:49It showed the opposite. They
• 42:50said that the basal group
• 42:52was resistant to neoadjuvant chemotherapy
• 42:54and their genomically unstable group
• 42:56was more sensitive.
• 42:58And so then there's another,
• 43:00then someone's like, what is
• 43:01going on here? So,
• 43:03forget what the what group
• 43:04it was, but they basically,
• 43:05they took some cancers and
• 43:06they sent them out for
• 43:07decipher and got them profiled.
• 43:09Then they did the lung
• 43:10classification,
• 43:11and they were all over
• 43:12the place. So there was
• 43:13there was an overlap.
• 43:15The basils from one didn't
• 43:16overlap with the the basils
• 43:17from the other.
• 43:19And so the I think
• 43:21that goes to the tell
• 43:22the story, like, whenever we
• 43:23say basilar, we say luminal,
• 43:24we're saying something that's somewhat
• 43:26arbitrary.
• 43:27Who's basal? Who's luminal?
• 43:29There's there's
• 43:31there really, I think that
• 43:32the way to do this
• 43:32isn't to say this is
• 43:33a luminal cancer or basal
• 43:34cancer, but rather, what are
• 43:36the processes?
• 43:37What are the signatures that
• 43:38are active and how can
• 43:39we use those to inform
• 43:41therapy?
• 43:42So at this point, I'm
• 43:43I really am not an
• 43:44advocate at all of the
• 43:45descending molecular subtypes,
• 43:47in the clinical setting. I
• 43:47think there's no we don't
• 43:49know how to do it,
• 43:50basically, and we don't know
• 43:51what to do with the
• 43:51information once we get it.
• 44:04Yeah.
• 44:06So with those
• 44:08with the base of our
• 44:23Yeah.
• 44:31So I don't know if
• 44:32anyone who's looked at the
• 44:33basal phenotype of noninvasive papillary
• 44:35cancers. I would assume they're
• 44:36high grade, but I I
• 44:37don't know that.
• 44:38But people have looked at
• 44:40the subtype of these, the
• 44:41flat carcinoma in situ, and
• 44:42there is a subset of
• 44:43those. There's a small subset
• 44:44of flat carcinoma in situ
• 44:45that has a basal phenotype,
• 44:47and that's more aggressive. That's
• 44:48associated with higher risk of
• 44:49progression to muscle invasion, etcetera.
• 44:51So, so it's not completely
• 44:53false that there's like this
• 44:54subset that start off basal.
• 44:55I just think it's a
• 44:56minority of them that do.
• 45:00Anyone else?
• 45:01Oh, yes. So the, a
• 45:03lot of us working on,
• 45:05cell lines. Yes. And and
• 45:06the,
• 45:07the prototypical aluminum cell lines
• 45:10and for physical
• 45:12cell lines,
• 45:22Yeah. They do. It it's
• 45:24kind of amazing, actually. Oh,
• 45:25did you have more? Go
• 45:26ahead. Go ahead. I can
• 45:27explain.
• 45:39So we haven't done the
• 45:40latter yet, but but the
• 45:42the former question, yes. So
• 45:43some of these
• 45:44do have similar histologies.
• 45:47It's it's difficult to to
• 45:48really compare them. And the
• 45:50way that we've done it
• 45:51historically is is with xenograft.
• 45:53So if you grow, like,
• 45:54the the cell in r
• 45:54t four, for example, which
• 45:55is a liminal cell line
• 45:57in a subcapsular,
• 45:59you know, kidney xenograft
• 46:01in skin mice. It looks
• 46:02like a papillary carcinoma, classic,
• 46:04you know, papillary or theeloid
• 46:06carcinoma. You grow a SCABER
• 46:07cell line. It's a basal
• 46:08one in the same model.
• 46:09Even a subcutaneous xenograft, it's
• 46:11keratinizing squamous cell cancer. So
• 46:12the the histology does mirror
• 46:14the molecular subtype very well.
• 46:17But they're they're hard to
• 46:18grow. Like, not every all
• 46:19all these cell lines grow
• 46:20easily in xenograft, some just
• 46:22don't.
• 46:23And so we've we've not
• 46:24done that experiment, but that's
• 46:25it's a
• 46:27yeah, that's it's something I
• 46:28thought about, but I've not
• 46:29done it. They don't have
• 46:30the state. It it's not
• 46:32that impressive. So in culture,
• 46:33yeah, like, the the r
• 46:34t fours are kinda like
• 46:35little little dots. They're little
• 46:36papillary looking things, but nothing
• 46:38is impressive as the xenografts.
• 46:39But it really doesn't change
• 46:41the just, like, the
• 46:42the, you know,
• 46:44the look of the cell
• 46:45lines. So it suggests that
• 46:46maybe some
• 46:47of
• 46:48the morphologic,
• 46:49histologic
• 46:50Mhmm. Is
• 46:51derived from interactions from the
• 46:53others. Yeah.
• 46:55Yes. Yeah.
• 46:56For sure.
• 46:59Yeah. Go ahead.
• 47:16Yes. So,
• 47:17we
• 47:18actually, I've I Andrew, I
• 47:19just analyzed your data he
• 47:20gave me, and so we
• 47:21actually have some of this
• 47:22we're looking at right now.
• 47:24They're they appear to be
• 47:25more basal in our hands
• 47:26at least.
• 47:27But in the hands of
• 47:28others who have published on
• 47:29this,
• 47:30it it's it's
• 47:32inconsistent. And it's it's a
• 47:33similar kind of story that
• 47:34if you look at spatially
• 47:35distinct parts of a tumor,
• 47:36you get different subtypes. If
• 47:38you also look at the
• 47:38lymph node metastasis or distant
• 47:40metastasis, it it can differ,
• 47:42from the primary as well.
• 47:43So it really it's what
• 47:44part of the main tumor,
• 47:45you know, gave rise to
• 47:46the MET.
• 47:52Alright. No one else?
• 47:54Oh oh, yeah. Go ahead.
• 47:56Just in general, you found
• 47:58lung classification that was reproducible.
• 48:01I I don't think any
• 48:02of them are that reproducible.
• 48:03I think that it just
• 48:04I mean, they're reproducible on
• 48:06their own.
• 48:07In the lung classification is
• 48:08nice because they have got
• 48:09some some biological,
• 48:11you know
• 48:13you know, consistency to it.
• 48:14So, you know, like, their
• 48:15lumenal types, there's two of
• 48:16them and they're they classify
• 48:17them based on the cell
• 48:18cycle gene that's been knocked
• 48:19out.
• 48:21But I don't I don't
• 48:22wanna use the word consistent.
• 48:23I think they're all consistent
• 48:23with themselves, but they're just
• 48:24not consistent with each other
• 48:25because it depends on the
• 48:26genes you pick to put
• 48:28them in the different groups,
• 48:29and they can be, a
• 48:30little inconsistent based on that.
• 48:35Alright. Well, thank you all.