This program contains
graphic images and discussion of
medical procedures. Viewer discretion is advised. Welcome to a Closer
Look, I'm Robert Signer, Deputy Director of the Sanford Stem Cell Institute
Discovery Center, and associate professor of
Medicine at UC San Diego. Today, we are here to take a closer look at stem
cells and the human brain. Now, the brain controls everything: it
controls our movement, our touch, our breathing, our temperature, our hunger, our vision, our thoughts, our memories, and our emotions. But the human brain
does even more. It defines us as both a
species and as individuals. Our intelligence, our cognition, and our capacity for reasoning are the essence of
the human condition. The uniqueness of our
brains distinguishes each and every one of us to
make you uniquely you. As Emily Dickinson said, The brain is wider than the sky. Yet, there's probably no greater biological mystery
than the human brain. Together, this is what
makes diseases of the brain so terrifying
and so difficult to treat. They are an attack on our
essence and on our everything, and they challenge the limits
of science and medicine, but today you'll hear from two people that
are changing that. These are two of the world's
leading experts that are completely transforming the
way we see the human brain, as an organ, as an
interconnected network, and even deep within our
cells and molecules. Even better, they've teamed up to bring their
brains together, converging on stem cells as a key for keeping
our brains healthy. Our first guest today is
Dr. Alexander Khalessi. Dr. Khalessi is Chair of the Department of
Neurological Surgery and professor of
Neurological Surgery, Radiology, and Neurosciences
at UC San Diego. He's the president
of the Congress of Neurological Surgeons and sits on the board
of governors for the American College
of Surgeons. He serves as a key advisor to
numerous federal agencies, has published more than
150 peer-reviewed papers, and has spearheaded some of the greatest modern
advances in neurosurgery. He earned his medical degree at Johns Hopkins University and completed his
neurosurgical residency at the University of
Southern California. He obtained his bachelor's
and master's degrees from Stanford University
and also holds an MBA from MIT Sloan
School of Management. He also launched
a new brain tumor and neural restoration program here at UC San Diego with our second guest,
Dr. Frank Furnari. Dr. Furnari in addition to being co-director of the
aforementioned program is a professor of
medicine and head of the tumor biology
laboratory at UC San Diego. Dr. Furnari has made
seminal discoveries on the genetic and
molecular mechanisms that drive the emergence
of glioblastoma. His current research
is utilizing state-of-the-art
stem cell models to uncover why brain tumors resist treatment
and is working to develop breakthrough
therapies to treat both pediatric and
adult brain cancers. Dr. Furnari earned
his bachelor's degree from Hofstra University, his PhD from the University of North Carolina
on Chapel Hill, and completed
postdoctoral training at the Ludwig Institute
for Cancer Research. He has been recognized by prestigious Scholar Awards from the V Foundation, the
Kimmel Foundation, and the Goldhirsh Foundation, as well as major research awards from the Society
for Neuro-Oncology. Join us as we take a closer
look at stem cells and the human brain. Dr. Khalessi? Well, thank you, Rob, for that really generous
introduction. I think you really
beautifully framed, from a philosophical standpoint, the scale of the opportunity that we have when we actually think about the human brain. You've inspired me,
actually, at least, to diverge for just a moment
from my prepared slides because when we think about the human brain and the
opportunity to innovate, I think you really
captured for me, at least, why the future in
neurorestoration specifically is so
exciting because we've lived as neurosurgeons for a long time limited by our understanding of the
structure of the brain, but as you'll see in
some of what I'll share, we're moving into an
era where advances in computational power and
material science are really allowing
us to deconvolute the electrophysiology
of the brain in new and special ways. The reason that's important
is because there are whole families of
neurodevelopmental, neurodegenerative, and
psychiatric conditions where are likely a perturbation of the function of the brain. As we understand
those things better, they actually become
opportunities and targets for a
surgical intervention or the delivery of a
biological payload like neurorestorative
or stem cell therapies. I'd also say that
in sharing some of my own personal
background if we've had any measure of success
here in San Diego, it's been precisely
because I recognize that our progress can't be limited by my personal imagination
or efforts. I've been deeply
fortunate to have amazing partners that
have really allowed us to answer questions in
ways that other people's can't that are new and
special and different. The ecosystem that's
been developed by the Sanford Around Disease Teams is one interface and
one opportunity. I'll share some of the structure and the work
that we've actually done as a department that really
position us to actually understand some of
those opportunities a little bit better and a
little bit differently. The reason people tend to
focus and start with cancer just to take a half step
back is because, obviously, cancer is a situation
where the biology of the body has run astray. There are a lot of insights that come that actually provide immediate benefit to patients in terms of therapeutic targets, but as we actually understand the biology of cancer better, there's an opportunity
then to think a little bit more broadly about those
neurorestorative options. I share that framing
just to provide some context for some of
what I'm going to share. As I touched on earlier, I want to actually innovate, it really requires an
entire team of people. I'm quite fortunate to be part of a much
larger department that has outstanding expertise across the breadth and depth
of neurosurgery, and some of my department
partners are here today. We've developed an ecosystem here in San Diego
that actually really allows for meeting the standard of care at a very high level. A lot of what I'm going to
show you in later slides actually relies on quite a bit in terms of infrastructure. That once we actually
know the right thing to do in the care of
an individual patient, how do we make sure that we
hit that mark every time? That's actually been reflected in the broader partnerships we've developed
with the Sanford, with the School of
Engineering, with our sister Department of Neurosciences to really advance that
understanding. Unlike in other areas of
science, if you can imagine, there's a phenomenology
that's unique to the human brain that I think Rob really touched on beautifully. Even to advance a
scientific question, what's really required is a
commercial-grade product. That's required us to be quite intentional about engaging
with industry partners that are unique to the
ecosystem that allow us to deliver on some of the therapies that
we'll talk about. We've been at the
leading edge of three stem cell trials that requires a proprietary
cell line, the ability to reliably deliver that vector in a way
that's atraumatic, and to be able to exert
a biological result. When you think about
all the steps that are required for any form of
neurorestorative therapy, there's a lot of
essential elements that we're fortunate
to have here in San Diego that allow
us to do things that other people
just frankly can't. That's required really
leveraging and tapping into this much broader
biotech VC ecosystem. Obviously, having support at the level of the chances
been quite important. In a group of
individuals who are really outstanding leading
scientists within their area, I've always been measured in the contributions
I could make actually as a practicing surgeon and some unique opportunities that actually come from
the surgical moment, or we really are
frontline observers to the physiology and
natural history of diseases, and actually, particularly the biological implications
of our interventions. I have this inset
picture of Halsted because I think there's a
belief among surgeons today, myself among them, that we're somehow much
slicker and technically have a much greater
command of anatomy than those who came before us, but the reality is, is the
converse is likely true. The advances that we've
made in surgery in the last 50 years and in the last 10 years,
specifically in neurosurgery, has really been driven
by advances in device, imaging and implants,
and radiology. We need to engineer
our procedures from artisanship to highly
reliable segmented tasks, and so you're going
to see that approach in some of the later
work that I show. We've actually
quite intentionally tried to develop in
this environment, a lot of the essential
infrastructure that's required to move
the insights that I hope that in this research
partnership that we'll develop as rapidly as
possible into patients. It's quite important
if you're dealing with a rare condition that there's an immediate actionable
result that comes from the insights that we get
from actually developing, for example, personalized
avatars around your cancer. We are much farther
along because of having done this
intentional work than other institutions in furthering that goal and developing that
network that's required. The best metaphor I can draw for people who
actually don't think about neurologic disease
every day is if we actually think about the
progression that we've seen in astronomy in the
last 10 years alone. The composition analysis of the James Webb
Space Telescope has absolutely revolutionized
our understanding of an object here, the Orion Nebula,
actually hasn't changed over the time
horizon of observation. What's allowed us to actually
really make progress is what we're able to see and distinguish in
the operating room. This is the courtesy
of one of my partners, Dr. Baumont, who's
in the audience, that you can actually
see here now in the operating room where prior to maybe five
or 10 years ago, it was very difficult
to distinguish between normal and abnormal
tissue when you're dealing with intrinsic
tumors of the brain. We now have navigation and aids to actually
outline in a heads up display normal
structures and actually use fluorescen agents to distinguish normal
from abnormal. We are also able because as
Dr. Furnari will touch on, we're dealing with
tumors that are quite heterogeneous in
their composition, we're able to actually select
cell populations quite specifically from
different regions of the tumor that we can surveil
and follow over time. This foundational work is
going to be quite important as a baseline for the progress
that we actually hope to make in our understanding
of these diseases. We also have the benefit
of infrastructure here, like interoperative MRI
that really allows us to apply energy and laser
ablation procedures, as you see in the
inset left, and again, the use of highly reliable
connectomics to actually map the insulated tracts of a
brain to actually allow us to pick a very precise
trajectory over the time. The reason that's important
is because we used to be a leading center
and still are in functional MRI
mapping with Magneto and cephalography that's now cited at the Qualcomm Institute, but this used to be the
level of our resolution. We would actually stimulate
someone in an MRI, and we'd be able to
map very basic things that are specific
to the patient, like arm and hand function. You can see quite dramatically how we now are in
a very different, both in mapping these insulated tracks in a way that's
special and different, being able to map those
structural findings to very specific biological
insights that you'll see at the
cellular level in some of what Dr. Furnari show, and we're actually
really leading the national conversation
in terms of actually creating much richer
tempor spatial images of the function or the
music of the brain. The reason that's important is it's really actually garnered much broader national attention
that when people actually think about what's next in
brain computer interface, the reason I'm personally excited about brain
computer interface just to be clear is so that we
can actually identify, as I said earlier, where are the structural and electrophysiologic abnormalities where there's a process that
needs to be interrupted. There's going to be a subset of those conditions
that the best way to interrupt that
process is going to be the delivery of a
biological payload. The best way to ensure in the long term that our neuro
restorative therapies are successful is that we select the right patients and know the right target
and have a clear, a traumatic path to get there. I hope you're taking
away from what I'm showing to date is
here in San Diego, we are in a very strong
position to do that, and that's precisely
why we've had the early opportunities
to lead some of these stem cell trials
because those are sources of failure,
sources of variance. Those are things
that we've actually eliminated or reduced
to the point that now we can focus on
the fun questions of what's the right stem cell line and how do
we actually match that to the right
clinical situation where the patient's natural history
is such that it could justify that risk and that early insight in a way
that really will over time, create real path to neurologic recovery
in those patients. That also requires,
obviously having the infrastructure to try
all these approaches, and so we're quite
fortunate to be among very few places in
the world that have these high fidelity
clinical environments on campus that allow us to de risk these procedures and do the required animal work to actually validate some
of these approaches. I'd be remiss to also not emphasize that our advances are not just in
the operating room and that it may be that a
subset of these payloads are actually best delivered in
an intravascular route. We've done a lot of intentional
work to actually identify the subset of tumors that actually would benefit
from that approach, and we're actually
quite thoughtful in our current treatment plans of combined endovascular and
open surgical techniques in the appropriate setting. I know we're here to actually celebrate what to me is actually a quite exciting
research partnership for our entire department. But I do believe because
so much of this work actually needs to be defined at the level of the human brain, and the only place you can
access the living human brain is in the operating room
right across the street that we're really working hard as
a group of surgeons and with the right research partners
to make sure that we're making the most of every single one of those
opportunities, but I also think there's a responsibility for all
of us to think about when we think about
tumor cell biology or intrinsic brain cancer, what precipitates a metastatic
event when so much of metastatic cancer is driven by a terminal or debilitating
or defining CNS event, it's important for
us to actually understand how our efforts fit into a much broader
paradigm of care. When we think about what healthcare is going
to look like over the 20 or 30 or
50 years in which we all together are going to make progress in these areas, I actually think it's
quite important for us to contemplate the different
settings of care. Obviously, a lot of what
we're trying to do is an outpatient health
maintenance as we develop longitudinal data
collection on patients. There's going to be an
opportunity to really optimize some of these things in
a truly preventive way. Then, obviously, as we have longitudinal baseline
data on patients, we're going to see
that someone's going to potentially,
for example, start to drift in their
cognitive scores, and that may imply that
there's an early change in their baseline and an
opportunity to intervene. Obviously, there's
going to absolutely be the need for acute
care in the future, and will become more
seamless in concentrating critical care and imaging and procedural services in a way, that really organizes in a high fidelity way
around the patient. Then, we're going
to actually have much better longitudinal data on patients and the use
of wearable sensors and other things to make sure that we're avoiding complications
that we actually achieved our functional goal of getting patients
back to their lives. Then, most importantly, in the context of cancer, if you're intervening
in the acute way, what actually really
is required in the intermediate term getting a handle on the biology
of that disease. Surgery and radiation are by definition, local
therapies, immunotherapy, chemotherapy or
systemic therapies, putting that all together
in a way that provides a durable result for the
patient is obviously the goal. If we actually think about
that in terms of conditions, where you actually see
inset here in gold are the five most common causes of death and disability
in the United States. You can actually see
neurologic disease and cancer feature prominently
in two of the top five. If we actually think about, for example, individual
susceptibility to disease, you can imagine a young woman
who has a BRCA mutation, she's wearing accelerometers, she has a subtle asymmetry, she doesn't know that she
actually has the diagnosis, and you'd be shocked to
know how many patients with an occult diagnosis of breast cancer present for the first time with a solitary
metastasis to their brain. Let's say she has an
acute seizure event, the ambulance responds, they come to the hospital in the midst of an
electrical storm. You can imagine getting that high resolution imaging to deconstruct for their family, what's actually taking place. You're able to get
the required imaging to actually plan a trajectory for the removal of that mass. You take the mass out
for that patient, and then in the acute phase, you actually get a lot of information to make
sure that you've actually safely gotten that
patient back to their lives. Maybe you actually have them, climb the tallest
building in San Diego for breast cancer and brain cancer. Then, ultimately, this is the key part of what the
center is going to do, you actually get very
specific epigenomic data on that particular tumor, so that you can actually provide tailored therapies in the intermediate term
that will make sure that solitary metasis that was resected gives them a durable functional
life going forward. I really believe
the opportunities for this center are quite broad, they really, in my view, touch all of the major
aspects of human disease and disability that
we're going to be facing together over
the next 50 years. The reason I'm
particularly excited about this research
partnership is because it bears on a very devastating
form of brain cancer, obviously, intrinsic
brain tumors, which I know Frank expanded on. There's going to be implications
for metastatic cancer. As I mentioned, CNS
events in all forms of cancer are often defining in terms of the ultimate
prognosis for that disease. Then, I know we're
all concerned about neurodevelopmental psychiatric and neurodegenerative disease. I see some of the neuro restorative therapies
we're going to be thinking about as we develop a more sophisticated map of
the function of the brain, there's going to be real opportunities that
are going to allow us to reverse disease in ways that we can
only imagine today. I couldn't be prouder and more excited to be part
of this ecosystem, I can tell you that
it's really fun to surround yourself with
people who are much better, smarter and faster and
stronger than you. I'm just really excited to see what we're going
to do together and really looking forward to the conversation and
appreciate the opportunity. With that, I actually
think it's left to me to introduce my
partner in crime, and this, is that right, Rob? Let's bring up the good
Dr. Furnari to show us some of the science
that's going to lead to these advances. Thanks [APPLAUSE] Great. Thank you so much. Alex, that was just terrific. It's wonderful to be
a partner with you. I'm truly excited about what the future holds
for this program. I'm Frank Furnari.
I'm a professor in Department of Medicine within the Division of
Regenerative Medicine, and I'm co-director
with Alex for the Sanford Stem Cell
Institute Brain Tumor Program. What I want to tell
you about today, in addition to stem cells
in the human brain, I'm going to really
focus on how we could use stem cells to model brain cancer and what
lessons we can learn from this modeling system for
therapeutic opportunity. This slide here shows the incidence of a
variety of cancer types, notably brain, and other nervous systems circle
there with a red circle. Cancers of the brains and the nervous system locations
are rare as indicated here. But as indicated by
the number of deaths, they are very difficult to cure. Even in cancers that
are more clinically treatable when diagnosed
early like breast, lung, and colon, such as has
pointed out there, when they do metastasize
to the brain, they do tend to be deadly. Whether it's a tumor
that originates in the brain or metastasizes
to the brain, as Alex mentioned,
there's a lot we need to do to achieve
durable cures. The tumor I'd like
to tell you about, which is the most deadly for primary brain tumors
is glioblastoma. This accounts for
greater than 50% of all brain tumors in adults. There's about 13,000 cases
of GBM per year in the US. If you include the
European Union, that's about 25,000 cases. The five-year survival
for GBM patients is less than 10% and we believe these tumors arise
from glial cells, which are the cells that
hold your brain together. The risk factors for GBM are very similar to
other cancer types. Age is the number 1 risk factor. Males are slightly higher than females in terms of incidence, and Caucasians and Asians slightly higher than
other populations. Overall, the prognosis
is very poor, and what this shows here, this is the standard of
care therapy which is surgery followed by
radiation and chemotherapy. The chemotherapy here is temozolomide and prior
to temozolomide, the overall survival
was about 13 months. The inclusion in
temozolomide in 2005, you can see we had
an addition of two months of overall survival, so that's not good at all. More recently, we
have the inclusion of tumor-treating
fields as shown here. The overall survival
has really increased by about an additional five
months to 20.5 months, so we need to do better. But why are these cancers
so difficult to treat? I think this can be boiled
down to five main reasons. Number 1, they have limited response to traditional
and targeted therapy, as I alluded to on
the previous slide. They're highly invasive. You can see that in this slide. In this section,
there's a tumor mass in one hemisphere of the brain, and there's another mass on
the other side of the brain, so these cells have actually
crossed the corpus callosum. They're highly invasive, making surgery not capable of
curing these tumors. They're highly immunosuppressed
or they're cold tumors, which means that cancer
immunotherapies that have worked for other cancer
types such as melanoma, don't work for GBM, and there's no early detection, as we see for other
cancers such as breast, colon, and prostate. There's no detection that
I'm aware of yet that would say you have the formation
of a brain tumor. Above all these things,
as Alex mentioned, these are very
heterogeneous tumors. What is heterogeneity?
Heterogeneity, in a word, is diversity. We have two types of
heterogeneity that we think of when it comes
to brain cancer. We have intertumoral
heterogeneity. That's the difference in
tumors between patients, and we have intratumoral
heterogeneity. That is the
heterogeneity you see within an individual tumor. That could present itself by different gene expression
across the tumor landscape, different compositions of
microenvironment cells. A good example of this was done by one of my students,
Bret Taylor, where he used a technique called Murfish to look at
gene expression across the landscape of a tumor
that we had engrafted into a mouse and you can see there
the image that says GBM, that mosaic part of that, those are actually
different genes being expressed in the tumor, so there's no homogeneity
within this tumor, so that is itself the problem. Although we would like the
tumors to be like this, homogeneous, they're actually
very much like this. How do we address this issue? I think of cancer like
an evolutionary tree. How would I try to
chop down this tree? Well, I would take a saw to the trunk and
cut down the tree. However, we're faced
with drugs that don't really tackle all the
mutations in the tumor. For example, if you
have a drug that targets mutation D
shown in yellow, that would eliminate
part of the tree. If you have another drug that targets mutation C, likewise, you'd only eliminate
part of the tree, and the same for a drug
that targets mutation B. We really haven't gotten to the root of the problem or
the trunk of the problem, which I call truncal
mutations in these cancers. That is where we need to go to try to eradicate
these tumors. There's been a number
of groups that have addressed this question
to try to sequence tumors and figure out what are the early mutations
in the cancer that are conserved throughout the
landscape of the cancer. One very good paper that
came out early this year, in which they took
a whole series of GBM patients and took about 10 different
tumor cores from each patient and did deep sequencing of those
tumor pieces. They found that, by a large, there were many regions of the tumor that shared mutations. Some had very unique mutations, but what other regions shared. But they had one mutation
that came out to be shared among all the
sections of the tumor, and we'll call that mutation A. I'll come back to
that in a minute. But before I talk
about mutation A, I want to take a brief departure and talk about aging
and immortality, aspects of which have
been discussed by several of my colleagues at these
closer-look presentations. What I really want
you to focus on are these structures at the
end of the chromosome. This is a chromosome
here on the left. It looks like a pretzel,
but it's actually a chromosome and on the very tip there,
that's a telomere. With each cell division, these telomeres
become shorter and shorter because the enzyme that controls the maintenance of these telomerase which
is called telomerase, slowly turns off in
most of our cells in the body and this correlates
with genetic aging. Manifestation of aging
is wrinkled skin, graying hair, some of which you see looking right
at me right now, so what happens in our stem cells is that
telomerase is actually kept on, and this maintains the ends of the telomerase so
they do not sinus, and these cells maintain their proliferative capability
throughout our lifespan. Cancers try to find ways
to hijack telomerase. They acquire mutations that
activate telomerase and therefore keep the cells or
keep the tumors immortal. They have a full tank
of gas all the time. Telomerase mutations are
very frequent in cancer. Now, getting back
to what that gene A was that I showed you
on the previous slide, gene A actually turns out to be telomerase
transcriptase or TERT. This is very highly
mutated in GBM. It's an activating mutation. It keeps the
expression of TERT on, and this mutation is present in 85% of GBM samples
not only in GBM, but you can see here on
this chart that it's also present in many
other cancer types, so by mutation of TERT, we keep telomerase on, and GBM achieves immortality. Not only is TERT a vital
mutation in a number of cancers, but it's the third most
mutated gene in all of cancer. First is P53, and the second is KRAS. We have an hypothesis. Since TERT mutations
are early events in glioblastoma formation that lead to tumor cell immortality, we can develop drugs that
target this mutation, this would affect a rather a significant eradication
of these tumors. How should we do this? My lab has developed a new program called
Brain tumor Avatars. Hopefully, some of you have seen the movie many years ago. Avatars are actually, they're stand ins or
icons or figures that represent a particular person
or a particular thing. You can see here the image of
the Avatar from the movie, and on the right, is an
Avatar of yours truly. What we developed is a platform where we can make
brain tumors in a dish, as well as in the animals. We start with stem cells, induced pluripotent stem
cells into which we introduce gene mutations
by CRISPR-Cas9 editing, and these mutations
are specific for GBM. We then differentiate these
cells to a cell of origin, such as a neural
progenitor cell, which is a glial cell, and these will form
tumors in a dish. We could also engraft
these into mice, and then these will
form tumors in the mice somewhere
between 2-4 months. Then when we take
these tumors out and we look at them histologically, as shown here, they're actually indistinguishable
from human tumors. I've actually taken
some sections of our models to our
neuropathologist here on campus, and they actually thought these came out of an individual, but they're actually from our modeling system,
our avatar system. When you compare the avatar
to the patient GBM shown below, histologically,
they're indistinguishable. When we do genetic analysis,
mutational analysis, the mutations that are
acquired in the avatar are also indistinguishable
from the patient samples. This is what we refer to
as Phase 1 of our program. Phase 2 is actually to leverage these avatars to look for therapeutic targets or
therapeutic vulnerabilities. To date, my lab has generated a rather
extensive avatar catalog. Each avatar is designed to explore tumor specific
characteristics. For example, we've generated a model that allows
us to look at gene amplification
in the form of extra chromosomal
DNA as shown here, which has never been done in other animal systems
such as mouse. We've generated
pediatric avatars, such as atypical
teratoid rhabdoid tumor. It's a highly
aggressive tumor that occurs in children under
the age of one, typically. This particular model showed
us that there is a defect in the maturation of a stem cell in the brain of the animals. These cells actually get stuck in a certain phase
of differentiation, and we think that
we can leverage this model to find therapies that force the cells to differentiate into
mature neurons. We've also recently described or developed a model of
leptomeningeal spread. This is driven by a
bee raft mutation, and this is a very
interesting disease because it's reflective of
metastatic disease. We also see
leptomeningeal spread into the spinal cord in
this particular animal. Then, lastly, I will tell
you a little bit more about our model to try to address the issue of
heterogeneity and this once again has to do with the
mutation in the TERT gene. As proof of concept, we made this model with
the TERT mutation. We grew the cells in culture. These are cells growing in petri dish and you can
see short term culture, looking at these cells, there's absolutely no difference in how well they proliferate. They proliferate identically. However, if we age these cells in culture
for about three months, you'll note that the cells with no TERT promoter mutation
actually begin to senescence. They stop dividing
when compared to the cells with the
TERT mutation, those cells have
achieved immortality. This shows very clearly that the presence at that
mutation is utilized by the tumor cells to overcome this cellular
senescence in phenotype, to be a successful
tumor, if you will. What happens when we engraft
these cells into mice? These are primary tumors, so these are the original
cells injected into mice, and we looked at
survival and what you see here is that
both the mutant and the wildtype TERT
genes do not have any difference in the
survival of these animals. However, when you take
these tumors out and you re engraft them
into secondary host, we see a very big difference. Those tumors with
the wildtype TERT take much longer to
kill the animals, whereas the mutant
TERT still has that capability and you
can see on the right, when we look histologically, the tumors with the
wild type TERT are actually much smaller
in these animals. With all of this,
we're now at a phase where we can look for
genetic vulnerabilities. In other words, the Achilles
heal of these cancers, and so what we really
want to do is to find genes which are specifically needed in a case with
TERT promoter mutation. We did a genetic screen. I did this with our colleague
Gene Yeo in this building, and we found genes that
are specifically needed for the growth of tumors
with TERT mutation, and they're shown there
in the red circle. We have about 30 genes which are absolutely required if you
have a TERT promoter mutation. I'll show you just the
data of one of those. This is called METTL14. If we eliminate
this expression in a GBM cell line that
has a wild type TERT, there's absolutely no
effect on cell growth. However, if we eliminate it in these two different GBM lines
which have mutated TERT, it's very clear that the cell are beginning to senescence or slow down in proliferation. With this approach, I'm
fairly confident we can find genes which are entirely approachable for drug
therapy intervention. With that, I'm just going
to stop and thank my lab, who over the years, I've had fantastic lab members who have really developed
this platform. Alison, Tomo, Shun, and Daisuke are the originators of the Avatar
platform along with colleagues in Gene Yeo's lab
and currently Clark, Brett, Yohei, and Chris are taking
over the mantle pursuing these models for different types of vulnerabilities that we hope to report back to you in
not too distant future. With that, thank you all
for listening. [APPLAUSE Thank you so much, Dr.
Khalessi and Furnari, we're happy to take
some questions. If you raise your hand, they'll bring a
mic around to you, but maybe I'll start
with the first one while people gear up. Dr. Khalessi, you showed us that image of the difference
in seeing a nebula. From you as a surgeon
in your experience, what's been the biggest change in terms of what you can see in the operating room maybe going back five or 10 years to now. What do you think is the
thing that will change the most in the next
five or 10 years? Sure. Thank you
for the question. I think it varies a little bit on what you're
intervening on. We obviously operate for a
lot of different indications. If I had to pick one thing, I would say having
the opportunity to map quite precisely
structurally, the deep tracks of
the brain has been enormously helpful
because it used to be, and you can imagine, if you had a lesion in the
substance of your brain, you'd look at a structural
atlas and you'd say, in general terms for
90 out of 100 people, this lives here, but
that doesn't really help you if you're one of the 10 people where it
lives somewhere else. Being able to with
high fidelity, pick a safe trajectory with
the use of navigation, I think, has been the largest
interoperative advance. The fluorescence I showed, as I mentioned, from
Dr. Baumont's case, has been a huge advance in the treatment of intrinsic
brain tumors specifically. If you have a glioma, having the availability of that fluorescent agent to aid your surgeon to maximize
your resection when we know volumetrically
achieving a higher resection while maintaining functional
outcome is really important. I think that's been a
quite important advance in glioma surgery specifically. When you think about
the next 5 or 10 years, it's one thing to
actually understand where the highways are for lack of
a better term in the brain, what is even more helpful when you think
about implications beyond cancer is actually
understanding how those different networks
communicate functionally. That's why I showed some of those very high density
electrophysiologic maps of the cortical
activity in the brain. Because I think once we actually
understand that better, that's really the first
best opportunity to modulate the nervous
system in new ways, either with stimulation, with the application of
laser energy or focused ultrasound or with a neurostorative
stem cell therapies. That, to me, is what the next
decade going to be about. Once you know how to
safely get somewhere, now you need to figure
out where you want to go. I think we're figuring out
where we want to go for a whole host of diseases
and beyond cancer. Thanks. While we get
the mic up here, maybe Dr. Furnari
you talked about how GBM cells hijack telomeras. Are there other
stem cell pathways that GBM cells will hijack or how else are stem cells specifically involved
in these glioblastomas? There's other ways
that GBM could actually hijack the
stem cell components. Another way that
they can maintain their telomeres is
through ATRX mutations. That's another telomere
specific mechanism. They also can self renew. We can look at the lineages of these tumors and we see
that there's a trajectory that represents very
similar aspects of development of the brain. A GBM, although it is a
tumor cell, is a tumor, if you look closely, it has aspects of glial differentiation,
neuronal differentiation. It utilizes the same pathways to maintain this ability
to self renew. If I could just add to
that really quickly. I think the real power
of Frank's work from my perspective as someone
who has the opportunity to treat people with brain
tumors every day is, you showed that beautiful slide of different colored M&Ms. Gosh, it'd be really
helpful for me to know which color M&Ms I
really need to worry about. That's helpful in terms of informing surveillance
because when you think about all the different treatment
decisions that are made over the course of a
cancer patient's journey, you're always weighing
competing risk and benefit. We think about it as surgeons in very simple terms as
volumetric residual, but not all volumes are equal. It may be that you are more targeted in your
radiation therapy. You have closer
follow-up surveillance. You can seed a modest deficit to actually achieve
a greater resection. All of those decisions are more informed as
we actually better understand the specific biology
of different components of tumors outside of actually being able to follow
them over time. I think it's really fascinating
how the advances in science aren't just informing the science and
drug development, but also the surgical
interventions as well and something that I think is a real breakthrough that
you guys are making. If I can follow up on that, what we're also learning is that these multi-colored bowls
of M&Ms can shape-shift. A cell that may be
green or red or yellow one day can be a
different color or next day. We're quickly learning which
cells within the tumor are actually sensitive to radiation and other types of therapy, but they don't stay
in that state. We need to figure
out how do we make the tumor cells more susceptible to our
standard of care therapy? How do we push them into
a corner where they're synchronous in terms of sensitivity to
radiation, let's say. That's something that
we're working on. We have drugs that look
like they may actually cause a tumor to be more
of a pro-neural state, which we know is more
sensitive to radiation. Take a question
from the audience? I have a question for Frank
and then a question for Alex. For the Metal 14 that places N6 methyladenosine marks on RNA, is there any
indication that that's affecting turt
expression in any way? It does. When we eliminate
Metal 14 from the cells, there's a transcription
factor which regulates the mutation in
the turt promoter. That transcription factor, expression level goes down, which then leads to
a decrease in turt. We found the link to the transcription factor which
controls the mutated turt. Are there inhibitors
for Metal 14? There's an inhibitor
for Metal 14, which is not very good. There's one from Metal 3, which we've tried and that
seems to work somewhat, but I think this is
a great opportunity to actually develop inhibitors for Metal 14 that are much
superior to what's out there. They work as a Complex
3 and 14, yeah. Dr. Khalessi, this
was amazing what you conduct as yours and Frank,
you're always amazing. I was actually really
glad you asked that question of
Frank, the last one. But weirdly, they're
interrelated and now I'm going to
make you nervous. Now, when you show the
fluorescence image, my background, my PhD was on photodynamic
therapy using these photoactive molecules with a specific wavelength
and energy of light to activate signal
and oxygen production. It looks like you
could do that in the brain more readily
than I thought you could, considering the new
tools you have to image. Is that something that people have looked
at resurrecting? Yeah. There's actually
been a lot of work done in our neuroscience department here around optico genetics and the idea of actually providing a sensitizing
agent and then exposing them to
light as a way to deliver a therapeutic payload. There's been fewer applications to my knowledge for
primary brain cancer. It's been largely used with five ALA to distinguish
between normal and abnormal, but that absolutely would
be a potential opportunity. Most of the local therapies
that have been focused on has been around the polymer based
delivery of known agents, as you know, but I think
that field is wide open. Five amino leveulinic acid that five ALA is a very
rudimentary version. There are huge amounts
of things you can do, but the other thing I thought
when you were showing the fluorescent imaging is
we use fluorescent reporters of stem cell activity
and we always thought they were amminogenic
and we couldn't use them. Maybe you could
actually use them. You could use a turt
reporter with Frank. Well we have an ADR reporter. We've got a
proteostasis reporter. There's so many things
we could do now that we know you can image
in such a fine. Obviously, as I mentioned, the center for future surgery would be a great opportunity to try a lot of those
agents in animal models. You could deliver that vector and then we can actually see. The key thing is for
those things is to actually figure out how
specific they would be. I think that would be the
main because obviously, you always have this trade off, because a lot of what we try
to do is protect normal, obviously and we'd
want to make sure we're in a position to do that. I think we've got another
question in the audience. Thank you for all
the great talks. I have a question
for Dr. Furnari. Firstly, I would
like to see that really a lot of fantastic model, really great models that
you generated in your lab. I'm wondering when you
show the two the mutation, you show that the first round
injection transplantation, you don't really see the
prognosis difference, but then the second round, you do see the difference. The patients, they
don't really get transplantation so I'm wondering what will be the implication of
this for the patients? Also, do you see association between the mutation
and patient prognosis? Your second question first? TERT promoter mutation has definitely overall
worse prognosis. Now, in terms of is the mutation really needed for
these tumors to form? There has been some really
good evidence of late, that TERT promoter mutation are actually existing
in normal brain, so in autopsies, they've
done some deep sequencing, and these mutations do exist, but they don't go on
to become tumorigenic. Because they don't have
the other mutations are associated with GBM. I do think it's
definitely needed as to overcome the bottle
neck of cellular senescence. But unless you have the other mutations
that occur with it, it's going to be
inert in our brain. Just like P53 mutations on our skin don't all
develop into tumors. It's only the rare occasion where you have the
acquisition of additional mutations that
we get full blown cancers. Can I follow up with that and you say that the
mutations are inert, is there any chance that those mutations in normal brain have a positive effect at all? Well, that's an
interesting question. I would love to
explore that if we can maybe make a mouse where we introduce this mutation and look and see if
there's anything that might extend the pool of stem cells within the
subventricular zone. That's an excellent question. These mutations
are heterozygote. You never see a homozygote
mutation in TERT. They keep the wild type gene intact to operate under
normal situations. But in a tumor situation, it's the mutated gene which gets activated by your unique
transcription factor. There's some
interesting evidence in other tissues now that
mutations that are associated with cancer
may start coming originally on to help
with regeneration or maintenance of the
tissue and whether there's a positive side to that before
it turns to the dark side, essentially, I think
it is intriguing. Incredibly fascinating.
It's something we can probably do in the mouse. Our next question
from the audience? Thank you. You mentioned early
on that the glioblastomas, and maybe this is a
question for you, Frank, are heterogeneous. But you also mentioned
that they can be metastatic from a
breast cancer patient. Was the example given early on. Just wondering what the ratio of metastatic derived
glioblastoma, versus primary glioblastoma,
I guess, in the brain. Just fundamentally how
different are those? Well I think this is a question maybe you can better address. Yeah, just to clarify. The example I was
given was a patient with primary breast cancer and
metastasized to the brain. But just to speak to
it more directly, the reason you're hearing a broader conversation around metastatic cancer to the brain, is this reflects a
more seismic shift in what neurosurgeons
treat because of our success in
treating systemic cancer. Ten years ago, 2/3 of
the craniotomies done in the United States were actually for primary
brain cancer, and the reason that was
is because if you add metastatic cancer to your brain, that was thought to be a
disabling or terminal event. In other words, there wasn't a meaningful benefit
to try and achieve technical intracranial
control because the systemic burn of
disease was too high. Now that ratio is flipped, so 2/3 of the craniotomies I do for tumor are for
metastatic lesions, as opposed to primary cancer. The reason that is is because our systemic treatment of
cancer has improved so dramatically that
if you can actually get someone good CNS control, that can buy them many
years of functional life. That is a credit
to the success and progress we've made in
other forms of cancer, that we've been able to focus so much more on the biology
of CNS metastases. Because there's this real
value that comes with that. To me, and to speak to the earlier question
that was asked about the application of some of the really exciting work
that Frank's doing, is if you actually think about a GBM and it's right
there in the name multiform, that there's all this distinct intertumoral heterogeneity. What becomes important is actually coming up with
meaningful subcategories. To me, the most powerful insight in some of the avatar
work that he showed, is we are starting to see
those reproducible lines, that actually are a first
step to then look at, okay, within this subset of tumors, what are the right
targets for further work? Where are the potential
therapeutic vulnerabilities? Because we used to, when we thought about the
biology of GBM, let's say, ten or 15 years ago, we were totally focused on much higher order
physiologic processes. We were worried about
tumor angiogenesis. A lot of the effort
was focused on reducing actually
the tumor's ability to recruit blood supply. Then we were saying, Well, the local tumor environment's
immunosuppressive, so how do we actually sensitize the immune system to the tumor? I think for the first time, as opposed to thinking about, the conventional
way the body either supports or responds to tissue, we are now actually saying. Well, what's the biology of the disease we're
trying to interrupt? Which is a fundamentally different question
because we now actually have the power to characterize that in a way that's
new and different, not just at the genetic level, but as Cat asked in
how those genetics manifest at different
time courses in the tumor's progression
in an epigenomic way. That, to me is
what's so exciting because the journey I take with cancer patients
is a longitudinal one. Having the right information
at the right moment. I think to me, is going to be the real
power that comes out of this partnership
in this work. We have a question here. Yes, question for Frank.
Great talk as always. Two things: one, you showed in your avatar you differentiate stem cells
along the glial lineage. I guess one question on that is, what is the cell of origin? How far have we
come or the cell, as identifying the cell
origin of glioblastoma, because as opposed to DIPG
and medulla where we have, fairly good idea,
which one it is. I guess for glioblastoma
is much more difficult. How do you address
that in your model? The other thing is
you showed that you did genome sequencing
of your avatar tumors. Have you also done
methylome illuminate 150_K? Have we done the methylome, you said? No, we haven't. We've done whole
genome sequencing, We've done obviously RNA Seq, single cell RNA seq, and those all perfectly aligned with what we see in
patient samples. The cell of origin question? We've routinely
pushed these towards a neuroprogenitor population to initiate our tumorogenesis. We've also pushed them
towards becoming astrocytes, and they're less tumorogenic. I think, because they have
more differentiation, such terminal differentiation
associated with it. At this present time, I mean, neural progenitor cells seem
to work for the GBM system, but that's not the same for the ATRT models
that we've built. That's completely different. The cell of origin there, is more likely a progenitor cell on the way to becoming a neuron. We're learning a lot, and we can map onto the
human disease to try to figure out trajectories
and really determine the cell of origin for probably all of our models. We've got time for
one final question. Great. Thank you very much. Those are really
fantastic talks, and I have a question
for Doctor Furnari. In your figure where you showed the example
of the heterogeneity, in the xenographs
you had four colors, and they were in patches
rather than a mix of cells. Then you mentioned
that even within that, the cells can also have
plasticity to change and shift. Now I was wondering,
how does the plasticity and the heterogeneity
of your avatars compare to the plasticity
and heterogeneity of patient derived xenograft tumors in the same kind of
injection models? Can you classify your
avatars into the different, PDXs and which ones
they're most alike? That's a great question.
Our first paper, we really wanted
to address that. Really, the only way to do that is by single
cell sequencing, and we saw complete alignment with what's been published out of institutions
like Mass General, where at the single cell level, you can describe a tumor
cell as pro neural, mesenchymal, OPC-like
or astrocytic-like. We were able to do that. It really was based
on the mutations. A proneural genetically
defined model was different than
our mesenchymal genetically defined model in
terms of the distribution of these what we call
cellular states now for GBM. We get exactly
what we see there. We see the same transitions between states that you see
also in patient samples. Was that your entire question? I can't remember. Thank you. Let me once again thank Doctors Khalessi and
Furnari for joining us today and also to thank them for devoting their lives to
something that is so important, and for working to
bring new light to what we said was one of
the greatest mysteries of human biology and human
medicine and I think are really making the difference
right now in people's lives. Thank you so much and thank
you-all for joining us.