Thank you for joining us. I'm Francis
Collins, the Director of the National Institutes
of Health and I'm pleased to be here today for
a conversation with Dr. Aviv Regev, who is the head of
Research and Early Development at Genentech but until
very recently was a Professor of Biology at MIT and the Broad Institute and
she is, I'm proud to say, the recipient of the 2020 Lurie Prize in Biomedical
Sciences, an award bestowed each year by the Foundation for the National
Institutes of Health. And I always get the privilege
of interviewing the prize winner and I didn't want to pass it up this year even
though we are doing this all virtually this time because of Covid-19.
and I've known Aviv for a while so this could be a fun conversation.
I will also say that we have some pride in her
particular contributions because she was also a Pioneer Award winner from the NIH.
You might also say that it's pretty unusual for somebody to win the Lurie Prize,
who's supposed to be a younger investigator, but who's already
been elected to the National Academy of Sciences, so
that'll tell you something about Aviv and the way in which she has absolutely
made amazing contributions to our understanding of biomedical research
particularly in this area of single-cell biology. So Aviv
welcome and let's get into it. So maybe start off -- what is so special
about single cell biology and single cell genomics?
What's the fuss? Hi Francis. I think the fuss about single cell genomics is because it allows us to see things that were either
impossible or very difficult for us to see before
about the cells that make up our body so until recently if we wanted to see
which genes different cells express or use
we had to analyze a lot of those cells together
basically looking at their average. One of our favorite metaphors is that if you
would think of each cell as a piece of fruit
then this would be very similar to making a fruit smoothie: What you would
see in the end is kind of the average of all of the fruits but not any one fruit
in particular so you wouldn't be able to say you know
if some fruit was very rare in our smoothie there were just a handful of
blueberries and many strawberries we might not be able to recognize
the blueberries existed even at all or if we did it would be just a smidge of
what their you know actual characteristics are.
And in the same way if you have a very complex piece of tissue, for example a
part of the brain or the lung or the heart or the gut,
then some of the cells that would be there would be very rare and if you
looked at all of them in aggregate and just checked their average you wouldn't
be able to recognize these cells at all. And so what happened in the last several
years is that we developed new lab methods, a new computational algorithms
that go along with them and they allow us to look at individual
cells. At first we could just look at a handful of individual cells and now we
can look at very large numbers of individual cells but each of them
separately from each of the others and so in a single go we get to see all
of the pieces of fruits in the salad and from this we can learn
a great deal of new biology. And this ability-- I'm, you
know, bespectacled, I have to use eyeglasses in order to
see and I still remember the first time I got my
glasses -- I got them as a teenager and before then the world all of a sudden
turned very very fuzzy for me and when I put the glasses on for the
first time everything came into this sharp relief.
In that sense, you know, single cell genomics is like a new kind of glasses
for us through which we can see biology in a much sharper way or a new kind of
microscope in which we can see cells in a much
sharper way and i think that's what all the fuss is about. Because they let us
see pretty much every single aspect of biology in a
sharper and as a result a better way and to derive much
faster insights as a result of that. I think that's pretty powerful and
appreciate all of the references to fruit
but also some have pointed out, I'm even fond of sort of
reflecting a bit on this myself, what would you say are sort of the biggest
discoveries of biology over the last many centuries and people would say, well,
probably that organisms are made up of cells, that's like,
right up there on the short list maybe along with evolution
and DNA. Okay let's take those three but for most of the time we've been
studying biology we haven't been able to study
cells one at a time. I mean you could study cells in a petri dish that were
all clonal and so you were kind of studying a cell
but you were really studying a lot of them
or you could study individual cells by maybe some antibody to stain it but
you couldn't really ask that cell what are you doing.
So how does single cell genomics allow you to sort of ask,
to query each cell what are you up to here? How does that work?
So what you do is that you try to look at the molecules that the cells
have, that's the genomic side where we try to
measure every possible molecule. So for example in the first
large scale successful technology in single cell genomics is known as single
cell RNA-seq and what we do there is that even though
all of the cells in our body have the same genome and they have the same genes
they don't use the genes to the same extent. So a liver cell
needs certain specialized enzyme brain cells and neurons need
certain neurotransmitters so they don't need the same genes even though they
have the same genes in their genome. The distinction is that they express
those genes and the way that they express
those genes is that from the code which is our DNA
cells express RNA nd this RNA is translated into proteins.
So by seeing which genes are expressed, which genes
are being used to make RNA from which protein
from which protein will be made we can look at
which RNAs are being made and which quantities.
And that becomes basically the calling card of a cell's identity, you are what
you make if you're a cell. You have your genome, you
have the same set of instructions but your identity
is going to be defined by the genes that you express and the RNAs that you make.
So in single cell genomics what we do is that we actually profile
those RNAs and before what we had to do is take a lot of cells and look at their
RNA together. Now we look at each individual cell and
we look at its RNA profile and that gives us the
calling card of a cell so now if we have two different cells,
one cell and the second cell and we look at our RNA rofile and their
RNA profiles are similar then we can say oh they're cells of the
same kind. If we look at two different cells and
the RNAs are very dissimilar from each other we might say oh they're two
different types of cells. So this would be like my fruits you know
the strawberries versus the blueberries but you also have to recognize that even
cells of the same kind are not actually identical to each other
so right in the fruit analogy you have many strawberries but there's
the big strawberries and the small ones redder and less red ripe and less ripe
from each other and those things are meaningful as well.
And so the same is true for cells you can have all of the cells be one
kind of immune cell at some level they're all T-
cells for example, very important fighter of infecting viruses
but they're not actually all identical to each other this T-
cell might actually be promoting immunity and that T-cell might actually
be shutting things off and that's going to be reflected again
by those RNAs that they express from their genes. So we can find these finer distinctions between the cells
using these measurements. And now RNA is just one layer of what the cell has
there's methods to measure different types of proteins in the cells
there's methods to check which genes are on and off based on how the DNA is
organized increasingly we actually can look at all
of these things together in one cell or we can look at the cells not just
when there are pieces of fruit in a salad and every cell is separated from
each other cell but when they're organized nicely in the
tissue and we can tell which cell is next to each one.
And if you want to carry the fruit analogy all the way through we call
these the fruit tarts because the fruits are beautifully
organized on the top of the part and the same as the section
that you take across the tissue. Oh wow I'm glad we got to the tarts. Yes I can
see what you're saying. Well let's talk about some examples and
I have to start with one that's particularly
familiar to me because it was such a big deal when you and your team figured this
out. My team, way back in 1989, found the
gene for cystic fibrosis, a gene that nobody knew much about and
we tried to figure out what we could learn from it.
It's called CFTR and clearly it's the gene that must be really important
where cystic fibrosis is most apparent the lungs, the pancreas, the sweat glands.
But exactly how does it do what it does, we kind of figured,
okay it's in the lungs, it's in the airway, it probably isn't like every cell
in the airway. And then something happened.
So what did you guys come up with that just blew us all away? So we were of course not looking
for the cells that express CFTR because we like everyone else assumed we
actually knew which cells they were and that indeed
they were very abundant cells in the lung and airways and in other in
other tissues. We were interested, we're kind of curious
at what are the cells of the airways and we started actually with mice. This was
in the early days of the field it was easier for us to get tissue from mice. We took the trachea which is part of the airways, dissociated into single cells,
profiled those single cells. And at the at first we were using one of the
earliest techniques we had so you know doing a few hundreds or a few thousand
cells was a big deal. And out of several hundred cells that we
profiled if you remember I told you the cells have these RNA calling cards
and we group them together actually an algorithm does that for us
because the space is 20,000 dimensions and humans don't work that well
in that space but it identifies things that in high dimensions are close to
each other based on these profiles. And we get you know all the known
subsets of the cells there are cells called ciliated cells, they're relevant
for for you know CFTR and there's basal
cells and there's some cells that are more rare.
There were six kinds that we expected to find. The algorithm found
all six except that there were three more cells
and initially it looked like dirt on the computer screen
when every cell is is as adult because there were so few out of the
out of the many but they were completely different in the three were
and then when we looked at the genes that distinguish them from other cells
one of the top six genes was CFTR and we were like, that can't be.
It's supposed to be in the ciliated cells. And we go in the deep there
right in these little RNA sequences and there's no CFTR. All the CFTR is in these
cells. So first we gave them the very exciting name internally, we used to call
them the 'hot cells' because they seemed hot and cool.
And then we were like, we would never be able to drop this result i mean there's
literature from 30 years saying that they're in the
ciliated cells. We don't see them; it's a fault of the method.
So we did two things. The first is we took a new technology that we were
developing at the time to do tens of thousands of cells and did many
more cells so that we would get to larger numbers and after we did that
the same results still held there was this group of cells that we have never
seen before that were expressing CFTR and there were
all the other cells that were supposed to express it and weren't there and by
that point we started believing the results but we still didn't believe
fully. We were like, maybe it's another part of the RNA, maybe
it's some other problem we will go back to the tissue with
antibodies. And we took the whole collection of antibodies
that detect protein in tissue and applied them to tissue from mice and we
couldn't find it sorry and we did find them in stain and
then we were like well there's a discrepancy. So we took tissue from a
knockout max that doesn't have the gene anymore and
we could still find the signal with the antibody and at that point we
were like probably the old reagents were not good enough and if you ask yourself
where do the reagents come from well it's very hard to know whether you're
actually measuring the thing you wanted to measure because the only
way to know if the thing is there is to use your
measuring reagents. It's one of these difficulties that we have when the
technology is just not there for us. It was for us instead so we had
themselves and then we realized that calling them the hot cells was probably
not a winning proposition and didn't say a lot about what so we started looking
at the other genes that they express and we found three that have not really
been studied in humans or in mice but they haven't studied in the cells of
fish and frog that are in you know fish gills and
in the skin of the frog which are again interfaces with the world just like our
lungs are. And in those organisms they are known as
ionocytes so we decided to call these the pulmonary ionosites and a new cell
type was born. Now the cell was always there doing its
business, being critical, actually, to the function including the
functions that we know are disrupted in cystic fibrosis,
but we just didn't know it was there. And that has real implications, Francis,
because if you think about cellular therapy cell
cell therapies for CF are something that people
develop. If you're targeting the wrong cell
you wouldn't get the therapy and so that was one of those mysteries just
just like--poof-- opened up in this way and it's actually
a big thing in rare disease, in CFTR, CFTR had a head start because
because of your beautiful work that people have been digging in it for many
years but in many initiatives, including the
Undiagnosed Disease Network, which is an initiative from the NIH,
we keep hunting for these rare genes and when you find your gene in the genome
not all cells in the body use it because the cells in the airway
use CFTR when the gene is defective you end up with an airway disease
like CF but if it is a gene used by muscle cells then of course the disease
would be a muscle disease and that's what happens in muscular dystrophy.
Well in some cases it's not in an obvious place
and so this information that we can now go after very rare things and find where
genes are expressed is extremely useful in many diseases
including in rare disease. That's such a great story and it totally did turn
upside down what we thought we knew about cystic fibrosis and along with
big implications as you said for cell therapy, for gene therapy, for drug
therapy, oh my. We had to sort of start over again.
And we never would have figured this out without the single cell approach, finding
those three cells in your experiment that just
didn't look like the rest. Do you have any other greatest hits, when
you look at the way in which single cell biology has like surprised
everybody by uncovering some rare cell that nobody knew was there and then all
of a sudden your understanding completely gets
turned upside down? Does that happen a lot?
It happens, it happens more than we think. I think there's two
two versions of this, in health and in disease. So maybe I'll
you know, the first study that we described that I described the ionocyte,
that was in healthy tissue first in mouse then we validated it in the human,
but, and a lot of such discoveries are happening for example in our brain
every day you can imagine we're finding a new cell but we kind of expected that
there there the number of neurons that we have
in the brain is huge and our expectation is that they're very
diverse, we just didn't know what they are. But there's also places
where things are hiding from us and really we we weren't knowing that we
were searching for them and that also happens for us in disease. So I'll give
you two examples from disease. One is from from cancer.
So we think of cancer as a very heterogeneous disease to begin with
and a lot of the variations between the both the tumor cells in the cancer
is the cancer cells in the tumor so the malignant cells but also non-malignant
cells, cells of the micro environment, the immune cells,
the connective tissue cells, they're not mutated they're not part of the
cancer but they're part of the tumor. They're inside the tumor doing their
business. Those cells are very diverse and they're
very different from each other but those differences are more subtle
they're not about a new cell type they're all cancer cells but some of
them are different from others. So one of our early discoveries around
the cancer cells was in melanomas, which used to be very deadly tumors, they
still are for many patients. And we found two types of programs
that cancer cells run that really impact the response to therapy,
one kind of program that characterizes cells even before
the tumor has ever seen in therapy before the patients have
been treated but it actually makes them more resistant to what we call targeted
therapies like those that target different mutated
genes in cancer and a second kind that we found was
programs that cancer cells activate and make them exclude T-cells out of the
tumor. Now one of the greatest advances that
has occurred in the last decade in in in in cancer treatment has actually been
the development of immunotherapies, so therapies that unleash the immune
system on to the tumor. They've been tremendously successful for
some patients but they have been completely
unsuccessful for others. These patients seem not to respond to
immunotherapy and we know now that some of the reason for that is that those
patient tumors do not allow T-cells in and we now found this program
in a rare subset of patient cells tumor cells
that actually excludes the T-cells out and once we find these programs, now we
have a new therapeutic target that we can go after.
So that's an example in cancer there's examples like that also in non-cancer so
for example in ulcerative colitis which is a form of inflammatory bowel disease
again there's immunotherapies that we've been
very successful targeting immune molecules known as cytokines and
when you target them patients sometimes respond you know
miraculously well but again some patients don't respond and some
patients develop resistance. So they're treated with a drug and it benefits them
and then all of a sudden it doesn't benefit them anymore. So we found this
this cell type a form of connective tissue cell called the fibroblast
but they assume a new kind or a new type or a new program in patients that is
inflammatory we call them now inflammatory fibroblasts and they
actually feed the inflammation and they
characterize patients that don't respond to therapy and they're actually
predictive of the lack of response to immunotherapy.
And there's more of these very rare cells or cells that are sort of out of
place There's a super rare cell type called
the M cell, we usually see it only in the skin testing
but in patients with ulcerative colitis we see it in the colon
and it expresses genes that we know from genetic studies increase the risk
of developing colitis. So these cells that all of a sudden pop out
out of place or like you know you're supposed to have a strawberry salad
here's a blueberry that can be a sign of disease as well
there's many many stories like that. Different kinds of astrocytes which
are brain cell and microglia which are a brain cell
that all of a sudden pop up in the brains
in the context of Alzheimer's disease which is a neurodegenerative disease
and so on it's it's all over the place. Wow.
You know one of the projects that I'm working on with the Foundation for NIH
and supporting it is the Accelerating Medicines Partnership and one of the
things we're looking at is rheumatoid arthritis and actually getting biopsies
of the joints of people who have active disease versus
... disease couldn't help but notice you talked
about fibroblasts that have taken on inflammatory
behavior. That's exactly what we're seeing by doing single cell analysis
in these synovial tissues of rheumatoid arthritis. So maybe there's a common
theme here that we never would have been able to
recognize without the single cell approach. So
that's totally cool. So Aviv, not only have you done fantastic research
from your own lab with an amazing group of trainees that flock to you in great
numbers but you also have put a lot of your time
and effort encouraging the whole field and you have i think been the main sort
of push with a few other colleagues on
something called Human Cell Atlas to try to make the most of this
moment in scientific history where we can really look at single cells.
So say something about the Human Cell Atlas. Why did it come
to be, and what is it hoping to contribute?
So I'll start with a little bit of history. So around 2014
we kind of had many of these pieces so some of these stories I told you know by
the time there were a paper it was later but we've kind of had the results by
then, we've done the first study, you know first of immune cells and then
we've done a study looking at immune cells inside
mice in the context of autoimmune disease and we've had the
first tumor studied which was in glioblastoma. We started mapping cells
where they're located in in space which we did in the context of
zebrafish development. We started looking at how you might have new neurons being born in the brain
and along the way developed kind of an algorithm toolbox that lets you ask
questions of these very sophisticated data
and the experimental toolbox lets you profile more and more and more cells
more and more and more efficiently and ask a diversity of questions. So by
2014 we've seen enough of this. It would have
had answered many of your questions in 2014 they weren't as elaborate,
there weren't as many examples but there was at least one example
of almost every type of question that we might want to ask
so it became quite clear that with the technologies getting scaled we had
Drop-seq already kind of working in 2014. With the
technology scaled like that this is no longer something that you
know should just be one or another labs endeavor that we
could just go and build an atlas of the human body and you could say
that started in the 1600s when biologists-- when Hook-- first saw
cells under the microscope and people have tried doggedly to really chart the
cells of the body ever since. But it was always an endeavor, it was always
technology driven it could be the microscope or the stains or the fats,
there were many technologies but we never had this unifying
technology that we could just apply everywhere and it would work right away.
And so that became a really burning passion for me. I felt like we should
just do it, that the time has come and I started
evangelizing for it and actually the first time I evangelized for it was a
talk at the NAERI I was invited to give a challenge talk,
several of us were supposed to say what could the NIH,
what would you do if you had 50 million dollars in five years?
That was the way it was it was framed and I said why don't we make a human
cell atlas? I made a set of slides, I showed that it was technically feasible.
I even had like back of the envelope calculations and I said that would be
really beneficial with... it would build a map. And we know that
maps are extraordinarily beneficial for any human
endeavor independently for biomedical research.
And after I gave that talk besides evangelizing
in giving a lot of such talks I basically stuck a
a series of slides like that at the beginning of any seminar that I would
give. I would say like I'm going to talk about this but first let me tell you
about this Human Cell Atlas. In early 2016 Sarah Teichmann was a very
good colleague at the Sanger and now a
very good friend and I kind of got together by email
and she said, I know you're interested in trying
this idea of a Human Cell Atlas. I'm interested too. Why don't we talk about that and we did and we had a series of conversations
and then with a couple of colleagues and at first we were in the mindset of, we'll
just go and convince somebody to fund this
and then we were like we also need to have a real scientific plan. Let's do that and we invited 93 of our best
colleagues to to London to a meeting co-hosted with the Wellcome
andmany enthusiasts including great colleagues from the NIH with
Francis's help and we asked each other, Should we do
this? And the answer was yeah, and we'll figure out how. And we
spent a year which we called the planning process between
October 2016 and October 2017 actually mostly planning and launching
starting the launch of a data platform and some data collection efforts and so
on and but since 2017 we've been in in full-fledged
mode. And our mission is to create a
comprehensive reference map of all human cells
you know for diagnosing, monitoring, treating,
understanding biology, understanding disease, all of it in
in one fell swoop. And it's been going on great.
It's an international initiative, it's open to all:
anyone who wants to adhere to the principles is is welcome. We are very
committed to diversity both in the data we collect in the atlas
but also diversity in the scientists that make the atlas and it's been a
labor of love. So Aviv you've done
amazing things both for that kind of very large scale international
collaboration but your own lab's discoveries,
things that you've invented in terms of technologies and then applied them. But now you're winning the Lurie Prize. So tell me, what does that mean to you, getting recognized in this way? So i'll start by saying that
there's I think for all scientists there's
something particularly meaningful when your colleagues recognize work that
you've done. First for me, it's the recognition of
the work rather than the recognition of me
because that work as you pointed out is never the work of one individual.
First of all it's the work of a lab and i've been extraordinarily fortunate to
have a wonderful lab with wonderful grad students and
postdocs but also at the Broad, staff scientists, research associates,
colleagues who are computational and experimental clinical experts and
biologists and beyond my own lab that network of collaborators that works
with you through this kind of problems. So that recognition of the work is not just
a recognition of me personally, it's recognition of this community
that I belong to and that I'm very proud to have helped
generate and and mentor but it is a community , t's not just one person.
So it's not recognition of the work I think for my particular field
that means even a little bit extra -- I come from computational biology, a
field that strives to not just use computational tools to
understand biology better but actually use computational concepts
to understand biology better and a lot of the successes for single
cell genomics has been because of a computational
mindset. We devised biological experiments based
on computational ideas I think that's a little different than
some other fields in biology and the Lurie Prize is not a prize from
computational biology. It's a prize from the Foundation for the
NIH nd it is a prize in biomedical sciences
and I think for my community that means a great deal that the things we do
with our mindset of computational biology and genomics
are biomedical science, not with a qualifier. They're just biomedical science
and that means kind of an extra, an extra big deal not just to me but to me and
many many many colleagues, and it cannot be done by any field alone. It's
only in the convergence of these fields together
and I think that's a special moment in time for the field. And I
appreciate it myself. Very well said indeed and I completely resonate with
what you're saying about how computational biology has really turned
to be transformative for everything we're doing right now.
Aviv, what advice would you give a young student
who's thinking about a career in biomedical research?
Yeah so funnily enough I was actually asked that recently by an actual young
scientist virtually online when I gave a Mendel Lecture, European Society of Human Genetics, and so I
answered on the fly, and I actually liked my answer but I have some additions. So
I think my answer then was that they should follow their heart, they should go to the right place that lets them carry out what their
current mission is and they should keep a very flexible mind. And I'm going to
add to that three important things. They should also
follow their moral compass, they should be generous to others and to
themselves, times are tough these days, and I think
they should strive to do good in the world. If they do, everything else
will follow. That is a wonderful way to exhort a
young person to take a path that is going to lead to
great things and be true to themselves all the way along,
as you have been, Aviv. You're such a wonderful role model
to so many people who are listening to this conversation
and I am delighted having known you for a few years
that the Lurie Prize has been bestowed upon you. I can't think of a better
choice in 2020 than you, so thanks for your willingness
to engage in this little conversation. I wish you all the best
in your new role there in Menlo Park and once again congratulations on the
receipt of the Lurie Prize for 2020.
Thank you so much and thank you to the Foundation for the NIH for this honor.
