Molecular Basis of Proteus Syndrome - Leslie Biesecker

Uploaded by GenomeTV on 12.10.2011

Leslie Biesecker: Good morning. This is a -- although I don't
know, Kim, this may be a tie. We'll never forget that incredible conference, workshop,
clinic we had in 1996, when a number of you were sitting up, I think it was in the ninth
floor of the outpatient center and all those moms and all those families together at the
NIH for the first time. That was pretty awesome, too, and that was an exciting time. I remember
very clearly Mike Cohen and I, you know, deciding at that time, gee, you know, this disorder
is more heterogeneous than we imagined it could be and thinking about that. I think
that was a really exciting time and capping it with this 15 years later.
So, I'm going to try and give you a summary of what we have done with the research, the
molecular research on Proteus syndrome so far and I will apologize in advance, this
will seem a bit like going back to biology and for those of you who liked biology in
school, you'll have a great time, and those of you who weren't so fond of it will want
to wring my neck, but that's okay. We'll get through this and explain the findings and
talk to you a little bit about what we've done, what it means, and how this can give
us a pathway forward.
Okay, so mosaicism, and we've been talking about mosiacism at these meetings for years
and we have been talking about it as a hypothesis or a model or a theory or whatever you wish
to call it, and we don't have to be quite so vague and theoretical anymore, because
now we know that that is what's going on. So, we should get a little into some depth
on that, because this is what this disorder is, and so we have to be really comfortable
with the concept. So, the word comes from the mosaic, just like the tiles, and the concept
is is that the bigger picture, the bigger thing, the person in the case of Proteus syndrome,
is a mix of elements that have different attributes. Okay? And so, like the tiles on the floor,
you make the bigger picture out of different-colored individual tiles and, as you'll see, what
we're talking about is a condition where the person is a mix of different kinds of cell
that have genetic differences.
Okay, so we use the term medically to mean that the individual is comprised or composed
of cells that have more than one, two or more, distinct genetic constitutions, and that's
different from inherited disease, and we've been talking for years about how we've thought
that Proteus isn't inherited, and these data that we'll talk about support that. Now, it
is important, though, to remember that most of the cells in the individual are, in fact,
identical to each other and most of the genes in those cells are identical in the affected
and the unaffected cells, but there is a key difference somewhere in there that is changing
the nature of the cells, and some have it and some don't.
So, from the abstract to the more concrete, what does mosaicism end up looking like? Okay,
so now again, back to biology from high school, right? So, eggs and sperm fertilization, et
cetera, and what we start with here. Do we have a pointer? Do we have a laser pointer?
Okay, if we don't I'll try and use the cursor. What we start out with is a gamete. Gamete
is the fertilized cell and that has the contribution of genes from the mother and the father and
the important thing to remember about that gamete is that gamete does not have the gene
alteration that we're talking about. Okay? So, it doesn't have the alteration, doesn't
have Proteus syndrome. So, that one cell fertilized gamete then divides and becomes two cells
and then it becomes four cells and then it divides again and it becomes eight cells.
And at some point after it is more than one cell, one of those cells undergoes a change,
an alteration, a gene alteration, but that single cell joins in the process that occurs
subsequently, such that that cell also undergoes division, and then comes to populate the embryo.
To make this a little more real, I brought with me -- I teach a class on embryology and
developmental biology, and what I have here is a movie so you can actually see. This is
a mouse. It's not a human, but this is a mouse embryo that's actually undergoing this division
process, and there's three frames to the movie. These two frames are photographed differently,
and this frame here actually has colored labeling on the DNA, the genes, within the embryo.
So here we're starting with a two-cell embryo. Here's one cell and here's the other one and
it's encased inside this sort of translucent shell here. And so what you will watch is
how an embryo actually divides in real time. So you can see again, you're starting with
two cells, three, four, see how they arrange themselves nicely. Then watch again, now we
change the color, five, six, seven, and then another burst, nine, 10. Now I can lose count,
because I can't count as fast as they're dividing.
Now the cells are aggregating into a mass. It's hard to see where the separate cells
are. And now it's making a big cavity, so this a big fluid-filled cavity and all the
cells, there's probably between 30 and 120 cells here and they're all packed down into
one little part of the embryo, and that's called the inner cell mass. Okay? That's how
an embryo, the cells in an embryo, actually divide along the lines that we're talking
Okay, so schematically what we're looking at is this. So, again, you have a one-cell
embryo that does not have the alteration. That one cell, oops, one cell divides into
two and we're just -- I'm making this up, because we don't actually know when the genetic
change occurs, but here we have two cells, neither of which have the alteration. And
when the two cells divide into four, one of those four undergoes a genetic change or an
alteration in a gene. All right? So at that stage, and this is a made-up scenario, but
this would be a four-cell embryo like you saw in the movie where one of the four cells
has the genetic alteration and then the key thing here is that when the four cells divide
into eight, then two out of the eight have the alteration, because these cells, these
two cells, both derive from this one, which has the alteration, and that alteration then
is passed down in those cells within that embryo such that the embryo is a mix and that's
shown schematically here where you have the one cell undergoing the alteration and then
there's two and then four and then eight cells with the alteration and then those altered
cells end up in different parts of the body and that where those cells land in the body
determines what parts of the body have the affects of the condition.
All right. So, it's actually a little more complicated than that.
It always is, isn't it? This is a schematic diagram of where things -- where the different
parts of our body are derived from and what tissues go into -- what cells going into making
what tissues. We start out as a multi-cellular structure I showed you in that movie, a cluster
of cells that are pretty much all the same as each other, and then the cells start specializing.
They specialize into lots of different general classes of cells, but you have different classes
of cells and end up going to form skin and the bowels and the brain and the heart and,
again, depending on which cells turn into which organs and which organ systems, and
where that altered cell is within this pathway, determines where the alteration ends up in
the mature individual and what parts of the body are going to be affected by the condition.
So that's the biology.
We've been talking about Professor Haapala [spelled phonetically] for years, again, at
these meetings and he came up with this concept of the mosaicism for Proteus syndrome and
he came up with that because he had to explain what we were recognizing clinically to be
part of the condition from evaluating patients. You know, it's not as though we come up with
these, we, the royal we, scientists come up with these ideas because we're so clever,
we come up with these ideas because we spend time with patients, like you and your families,
we make observations, and we say we have to explain what's going on. How can we possibly
explain it? Explaining it as an inherited disease disorder just didn't make sense. So,
Rudy Haapala came up with his mosaicism model and what he proposed is that the disorder
was caused by a somatic mutation and that's the scientific phrase for that process I just
described to you and that that can only occur in an embryo in a mixed state. That is, it
can only occur and it can only persist if the embryo, the developing gamete or embryo,
has some cells that have the alteration and some that don't. It's a mosaic, and that were
it to be non-mosaic, that is all of the cells were to be affected, that embryo would not
So, that's the theory or the hypothesis that he came up with to explain Proteus syndrome
and what it explain is, again, this is circular, because the model was based on these observations,
but it allows you to understand why the patients have mosaic or patchy affects in their bodies,
why there is an absence of uniformly-affected patients, that is patients who have the effects
of the condition evenly distributed throughout their body, why there are no examples of recurrences,
either siblings of affected children having the condition or for those who reach adulthood
and have children themselves, why we don't see that patients with Proteus syndrome have
children with Proteus syndrome, and why we have seen what we call discordant monozygotic
Monozygotic twins, more commonly known as identical twins, share essentially all of
their genes and they share those genes because monozygotic twins are actually created by
-- remember when I showed you in that embryo how it formed that little cavity and all the
cells were squished down into the bottom? What happens is that little clump of cells
in the bottom divides in half, so that's one embryo actually splits into two clumps of
cells and then each one of those clumps goes on to form a complete fetus, and so they are
genetically identical because they came from the same gamete. If you have that process
and you have some of the cells affected and some not, that allows you to understand how
you can have apparently identical twins where one has the condition and one does not.
So this is a great model. It explained a lot of clinical observations, but that left us
struggling for nearly 15 years to figure out how can we find this alteration that Rudy
Haapala's model predicted is there. So the challenge was, take these observations, take
this understanding of the condition, now find the gene that's altered.
All right, so back to Biology 101. Cells, chromosomes, genomes, genes, and DNA, all
these words we throw around a lot need to be thought of in sort of an organized or hierarchical
way. So, our cells, our bodies are composed of a number of organs. Those you're familiar
with, brains, hearts, kidneys, et cetera. Organs are in turn built from tissues. Tissues
are a composition of different types of cells. Cells have within them, almost every cell
in the body has within it, a complete set of chromosomes. So now here we're on the diagram.
Here's the schematic cell that has within it the sort of X or sausage-shaped structures
and these are the chromosomes. The chromosomes are made of very long strands of DNA and within
the DNA are stretches or segments of the DNA that are the functional elements of chromosomes
and we call those genes. So, those are the genes that we inherit from our parents. And
then those genes are in turn composed of a series of molecules along two strands that
are called base pairs -- oh, wonderful. Thank you very much. All right, great.
And those base pairs are these things you see here, these four colored structures and
they have, we call them base pairs, because as you can see there are two of them at every
position along these strands and the strand is famously known as the double helix, for
obvious reasons, and that the order in which they occur along the strand is called the
sequence or the genome sequence, the DNA sequence, or the sequence of the base pairs, and that's
the structure from us down to our base pairs of our DNA. The important thing in the next
step is what the cell does with this. This is actually -- the DNA is actually information
and it's biological information that's passed from parents to child and that that DNA, its
actual job, is to make proteins. So, there's a whole bunch of machinery in the cell that
then converts these DNA sequences and the order of these base pairs determines what
the nature of this will be, but the order of these base pairs for each gene is then
converted into a protein.
Proteins, which are made of amino acids, are -- there's about 20,000 to 50,000 of them
in each cell, and the job of those proteins is actually to do the working business inside
the cell, to make certain chemicals, to break down the food we eat, to form structures in
the body, to tell the cell when to grow and when not to grow, et cetera, et cetera. So,
cell, chromosome, DNA, gene, base pairs, proteins, and then functions in the body, and we'll
come back to this concept of proteins a little bit later.
All right, so, here's the big picture and then you begin to ask the question, okay we've
got these chromosomes, we've got DNA, how much of this stuff is there and what is the
nature of this stuff? It turns out that there's a lot of it. So, this little diagram that
we've blown up here, like massively zooming in, into an incredibly tiny structure. When
you look at these letters here that you can read along here, you're seeing a tiny segment
of the DNA of the genes, because if you look at all of them in one cell, there are about
six billion bases of DNA in each cell. Six billion's a big number. Within that DNA is
20,000 genes, because we have, again, we inherit most genes, one copy from mom and one copy
from dad. We have two copies of most of them. Each gene is in turn composed of about 30,000
of those base pairs along the strand, but the important thing to recognize is that most
of our DNA actually isn't genes. Most of it performs other functions, but most genes code
for proteins. So, most DNA is not genes, most genes code for proteins.
If you take that DNA sequence and you change it, so changes in DNA lead to changes in the
proteins that the genes make and the changes in those proteins can lead to disease or other
medical conditions, inherited traits, things like that. But it's also important to remember
that most changes in the DNA that we have amongst us, every person differs in about
30 million base pairs of DNA within our genes are different, but most of those changes are
insignificant, just a tiny fraction of them are significant and are associated with disease.
Okay, so what we as geneticists want to do is understand the relationship between the
sequence of that DNA, the order of those letters along that double-stranded molecule, the relationship
between the sequence of those letters and health or disease. So, we do that by sequencing
instruments, all right? So, we've been sequencing for about 20 to 25 years now and sequencing
technology has advanced to the point such that in 2007, our institute supported four
sequencing centers, and these things, these places actually look more like factories than
they do like research laboratories, because their job was to massively sequence as much
DNA as they possibly could. So, this is a picture of a sequencing center. Again, you
can see lots of instruments scattered throughout. If you actually look closely you'll see a
number of these instruments are the same as each other. So, again, it's like a factory
producing a product, which is DNA sequence. We all were quite impressed with ourselves
back in 2007, because these sequencing centers could sequence a lot of DNA for that time,
but their total output of DNA sequencing, of four centers that were funded by our institute,
was about 200 billion base pairs of DNA per year. Remember, I said that one cell has six
billion. All right? So, yes, it was pretty awesome that we could sequence that much,
but that was actually only equivalent to the ability to sequence two human genomes per
year in all four of the sequencing centers that existed back in the old days, back in
2007. Now, that's old fashion, okay? This field changes so fast.
Now, in the past five or so years, a number of research laboratories and companies have
developed completely new technologies that have radically changed how we think about
this kind of science and this kind of sequencing. So, instead of having machines where each
machine is sequencing -- these instruments could sequence about 96 little pieces of DNA
at a time, that was, again, cool back in 2007. We have now developed instruments that can
sequence between half a million and a million pieces of DNA at one time. The way we do that
is by taking little pieces of DNA. So, if you imagine that double helix with those base
pairs on it, it's now this little gray bar here and then what we do is we take these
pieces of DNA and attach them to a glass slide and we attach about half a million pieces
or a million pieces of DNA to that glass slide and then sequence them all at one time, millions
of them at a time, such that what we do is by adding -- here's your base pairs again,
Your As, your Cs, your Gs and your Ts. You add one base pair at a time and every time
you add a base pair, depending on whether it's an A, a C, a G, or a T, what happens
on that slide is a little blip of a color light is released from that piece of DNA.
And so when you do that, you add one of these letters, A, C, G, or T, to all of these spots
of DNA on the slide.
What you do is you just take a picture of that slide, and so the picture of that slide
then is a massive array of colored spots on a piece of glass and it's a tiny piece of
glass and there is, as you can see here, a hundred to two hundred million clusters of
DNA on there. So, it's a hundred to two hundred million colored spots. Take a picture of that,
then you remove the color and you add the next pair -- the next base, which might be
a different color, but your DNA hasn't moved. It's in the same spot.
So, one spot in one cycle, as you can imagine, this is one DNA spot, the first picture that
was taken a T was added. Then the next picture that was taken a G was added, so that emits
a green light here. Then the next picture that was taken a C was added. That's an orange
light and then a T and then an A and then a C. When every one of those spots has a sequence
of colors, and what we do is feed this colored slide, remember this one shows, I don't know,
about 10 spots, but the actual slide has between 100 to 200 million of these spots, into a
computer. It looks at every spot and every picture from every cycle and then derives
from that what the sequence of the DNA was in any given spot.
So these instruments now look like this. So, this is the current sequencing factory that
we have and this is just down the road in Rockville, Maryland, the NIH Intramural Sequencing
Center, and these are what we call next generation sequencing instruments, and they have massively
increased our ability to sequence DNA and generate DNA from cells or blood samples or
whatever we put into it. So what this means is that the costs and the ability to sequence
DNA has gone through the roof and the costs are collapsing, which is great news for us.
So this is a graph that shows how things have changed. This is technology about 19 -- or
2007, and about every year or two since then we've changed generations of instruments and
the capacity to sequence has radically changed. So on the left here is how much sequence one
run, one cycle of one instrument, can create. And remember I said those entire four factories
were generating 200 billion -- sequencing 200 base pairs of DNA per year. It was 200
billion a year in four huge centers. Now what we're talking about is how much sequence one
instrument can generate in one run. Okay? And here on this axis of this graph, these
are how much DNA it can generate. KB is the abbreviation for a thousand bases. This is
for a million bases. This is for a billion bases and this is for a trillion bases. Okay?
This is a trillion base pairs of DNA and what you can see is with each generation of instrument
we jump up nearly half an order of magnitude so that we have gone from an instrument being
able to generate about 500,000 base pairs of DNA sequence in a single run to one instrument,
one of those blue instruments I showed on the previous page, can generate nearly a trillion
base pairs of DNA in one sequencing run. So that is then more sequence than four sequencing
factories in total could generate just four years ago. Then the good thing again about
this is that what that means is that with high production, costs have gone down.
So if you ask the question how much does it cost to sequence DNA, and again, it's just
crazy to actually have a graph like this, because what we're showing here is the cost
to sequence a gigabase, which is a billion base pairs of DNA. If you think back on what
we call the Human Genome Project, which was finished in 2003, it cost about $1 billion
to $2 billion to sequence the first human DNA. Okay? And so now what we're talking about
is how much it costs to sequence a gigabase, or 100 billion base pairs of DNA, and that
started out in the million dollar range, to sequence one gigabase of DNA, and now we're
down to about 300 bucks, so -- and it's dropping every couple of months further.
So what this actually looks like to us is, again, it cost about a billion dollars to
sequence a genome back in 2003. By 2007, when we had those factories running we could consider
sequencing a human genome for about half a million dollars. A couple of years later,
2009, it was down to $20,000, and that's when we started talking about thinking about sequencing
patients with Proteus syndrome, because now we're getting into the range of reality here.
This is not the range of reality for figuring out the cause for any disorder, but then this
is. And then by 2010, due to some shortcuts I'm going to talk about, the cost was 3,500
bucks. And again, the shortcut was, I mentioned earlier that most DNA actually doesn't code
for genes and if you think that Proteus syndrome is caused by an alteration in a gene, you
just sequence the genes, you don't have to sequence all of the DNA. So, if you just want
to sequence the genes, we call that exome sequencing. It's not a genome, it's an exome,
because it's composed of parts of genes called exons and then you can say, okay, now I want
to design an experiment to actually do this and use this incredible technology to answer
the question that I want to answer.
So, you have to ask the question, who should you sequence, and then you have to answer
the question, what samples from those patients is going to be sequenced? And that gets back
to what we did back in 1996, based on that first meeting where we met with all the patients
and we were trying to figure out who had Proteus syndrome and who had an overlapping disorder,
because if you sequence the wrong people, people with other conditions, it won't make
any sense and you won't be able to tell heads from tails, because again, all of use differ
in millions of base pairs of DNA and if you have a heterogeneous group of patients, you
can't untangle what's what. What you have to have is a uniform set of patients where
you're very confident everybody has the same condition.
So Mike Cohen and I developed these diagnostic criteria, which were pretty strict diagnostic
criteria, caused us more than a little grief with some of our colleagues who thought we
were being unreasonably strict, but turned out to have been a really essential thing
to do to make this research work. So we took patients who met those strict criteria and
we started to say okay, now what samples should we sequence? So we picked -- again, because
you think this is a mosaic disorder, you can begin to answer the question -- ask the question
of what's the difference in the sequence between the part of the patient who has the condition
and the part of the patient that doesn't have the condition, because there should be a difference
there. That difference should be the change that causes Proteus syndrome.
We had a couple of affected patients for whom we did not have a matched normal sample, including
the affected monozygotic twin. We had his match. We had a sample from his identical
twin who didn't have Proteus syndrome and we had five parents. So that adds up to 17
samples and at 2010 prices, that was a $59,500 experiment, which you, the Proteus Syndrome
Foundation, raised the money for. Okay? That was what you paid for with all those fundraisers
that you guys set up and the money that you donated to NIH to help do this work. And I'll
tell you there was more than one night that I didn't sleep very well worrying that spending
that money might not have worked, but I should have slept because it worked out fine.
All right. So, what exactly are you going to sequence? Another tricky question. So it
turns out, as many of you are more than well aware, we've been doing a number of surgical
procedures on patients with Proteus syndrome at the NIH. One of the reasons, one of two
reasons for doing that, the second reason, but it's important because it justifies us
doing it, is that we needed to do that surgery to collect affected tissue samples from the
patients and that, again, turned out to be a key part of this. And so we have a number
of samples from our patients where we did, I should say, Tom Darling mostly did, skin
biopsies and set up cell cultures from those skin biopsies and those turned out to be quite
useful, but most of the specimens actually came from the operating room. When one member
of our team would be in there, usually with Dr. Tosi, during the case while she's operating
on the patients and the site is open looking at the tissues saying, "I think this tissue
is affected with Proteus, I think this tissue is not," again, using our clinical judgment
to figure out which is which.
Now, a really important part of this is we did not use blood DNA. Most genetic research
is done by drawing blood from a patient, isolating DNA, because, again, in most disorders, the
DNA in all the cells is the same as each other, so it doesn't matter which cell you use, you
just take a blood cell, that's an easy one to get, and you sequence that. What we decided
based on what we knew about Proteus syndrome is that was the wrong thing to do, because
we could not recognize that the patients had any phenotype, anything abnormal with their
blood. So we weren't sure if those blood cells had Proteus syndrome or if they were unaffected
with Proteus syndrome, because we couldn't recognize any effect in the blood system from
the condition. So the best thing is to stay away from those.
All right, so if you sequence an entire genome, then again, you would expect to find about
30 million differences between any two people in this room, so 30 million sequence differences
in any two people. If you look just within the genes, that focuses you in on 300,000
or so variants between any two individuals and so you have a problem here. When you sequence
17 samples and any one of them, any pair of them has 300,000 differences, how do you figure
out what is the one you're looking for from the other 299,999 that are just background
normal variation? So again, you take a step back and you say, "What do I know about this
condition that allows me to filter out most of these other changes that I'm actually not
interested in?"
So we did a couple of things and this is very commonsense based on Rudy Haapala's model.
We can compare unaffected to affected DNA. So the gene alteration, again, should be present
in the affected cells or the affected tissues and not present in the unaffected cells or
tissues. A little problem here is we're using clinical judgment to determine whether a cell
tissue is affected when we remove it from a patient and that could be imperfect and
it's especially tricky for patients who have a progressive disorder, because we know full
well that there are occurrences where we see a patient on one visit and we look at a part
of the body and it is apparently unaffected and they come back a year or two later and
that part of the body has now become affected, because the disorder has advanced. What that
means is that in the previous visit, if we took a sample from an unaffected area, it
might have actually been affected and we just didn't know it yet, so you can be fooled by
We used the twin pair that was reported and published by Rudy Haapala and his colleagues
and we reasoned that we can be a little safer with unaffected cells or unaffected tissues
if we use for our normal sample the unaffected twin, because he didn't have the disorder
at all, so we didn't have to worry that his unaffected sample might be subtly affected,
because he had no signs of the condition. All right, but we have the same problem with
the affected twin. You're not sure if you're sampling an affected tissue and that actually
did turn out to be a problem. We also know that the disorder is not inherited, so the
alteration that's present in the child and their affected tissue should, in any case,
be absent from any of their unaffected parents. All right, now I've tortured you with biology
and now we have to do some math.
And, of course, it's just crazy. We just love this stuff.
Just love it, so you have to be tolerant of us. Okay. So mosaicism. I've told you that
cells and organisms can be a mix of affected and unaffected cells and so you have to think
about what the percentage is that you want to be looking for when you're doing the math
of this. So if you imagine a cell from a patient with Proteus syndrome where all the cells
in that sample have the alteration that's causing the disorder, and they have that alteration
in one of the two copies of the gene that causes Proteus syndrome, what you would predict
is that 50 percent of their DNA would have the alteration, right? Because remember, you
get one copy of the gene from mom and one from dad. That comes into the embryo. The
embryo has those two copies of that one gene and we're predicting that one of them is altered.
So, if it's not mosaic, it's 50 percent, so that's your ceiling.
If you imagine that instead of 100 percent of the cells having the alteration, let's
imagine that 40 percent of them have the alteration in one of two copies, so then your percent
of DNA would be 20 percent, if you had it in 12 percent, six, et cetera. So you can
see we're cutting -- the percentage is always cut in half. All right, so now back to this
figure. So you take the DNA, you collect the samples, the biopsies, the samples from the
operating room, the skin, the cell cultures, you isolate DNA from them. You prepare the
DNA. You put it in the sequencing instrument. The sequencing instrument takes those little
-- takes the DNA, cuts it into little teeny pieces, and sequences hundreds of millions
of those at a time. And then, as I said, you take that, which, it gives you that colored
spot picture and the instrument converts that colored spot picture into billions of little
short DNA reads and then those are fed into a computer that then takes those sequences,
those little short sequences, there are only 100 base pairs per sequence, and has to say,
okay, where do these belong in the genome, because that genome is huge.
Each piece of DNA could be from anywhere, so you can imagine taking something that's
100 letters and asking the question where does that line up in this universe of 300
billion? So, a lot of computation is done. When you do that, the output actually looks
like this. Is it possible we could get the lights down a bit? Because this is a very
dark slide. So, this is what the actual output on the computer looks like when the scientists
in the lab are analyzing DNA sequence. And this is the DNA sequence along one of those
X-shaped chromosomes and what that shows you, again, here is your As and your Cs, and Gs
and Ts, and these numbers here are the position. So we now know from the human genome project
what the position of each A, C, G, and T is supposed to be across the three [spelled phonetically]
billion nucleotide -- or base pairs of the genome.
And you can see here, it actually turns out we're on chromosome pair number 14 and the
position we're in is 104 million base pairs down along that chromosome. And what you can
see here are a stack of signals, or sequencing reads, and we have some abbreviations here.
All these abbreviations are color-coded, because there is so much information on this screen
it's hard to interpret. And it turns out some of the reads look at the DNA from one direction.
Some of the sequencing reads look at it from another direction. Depending on which direction
it is, it's either periods or commas. We use a comma or a period as an abbreviation to
tell us that this row of, I think that's commas, is one sequence read. Every place where you
see a comma, what that means is that the sequence of that read of DNA of those, you know, hundreds
of millions of reads, that sequence is the same as in our reference genome sequence,
same as in a typical person.
The different colors of the commas and the periods are a measure of the quality, the
likelihood that that sequence read is correct. And so white and yellow are good, blue are
bad, and you can see the sequencing quality is not so good at the end, and you can also
see then there's a few letters here and there on this screen. So, for example, here is a
G and that is not supposed to be a G in the normal sequence. But you can see if you look,
I don't know, maybe you can -- there's just one or two along there. Now here's a C and
there's just a single read here that has a C. So that's actually a sequencing instrument
error. These machines are not perfect, so every one of these base pairs that comes out
of these machines, there's a small probability that the machine might have read it incorrectly.
And so we have to separate out what is instrument error from what is a real difference. We do
that statistically and you can see right here that this position, the normal sequence is
-- well, perfect, thank you very much. That helps a lot. A normal sequence is a C and
you can see on a lot of the sequence reads here, there's either a comma or a period if
you look all the way down. But in a number of them there is a T and whether it is in
the forward or reverse direction, it's either a capital T or a lowercase T. Now there's
enough Ts here in this set of data to tell you that this is very unlikely to be a sequencing
error, because you're seeing it so many times. Okay?
This sequence read is actually the first screenshot when Marge Lindhurst [spelled phonetically],
who you'll meet, I think she's coming later today, actually viewed this screenshot when
she realized that this change from this C to a T in this sample was probably real and
turned out to actually be the change that was present in patients with Proteus syndrome,
but that is just one result in one patient and it's not valid until you do a whole bunch
of work to figure out if it's true or not. So, this is one sample in one patient. Now
I'll summarize for you here and what this is.
This is actually a summary of that previous screen. So if I were to take this screen and
say okay, how many reads are there? How many rows are there? How many of the rows are Cs
and how many of the rows are Ts? That would then be this. So there's about 28 rows there.
This was the number of Ts and this is the number of Cs, so you can see that's a mix.
Now that was an affected sample from one patient with Proteus syndrome and we also had a paired
unaffected sample from that same patient and that gave this output. So it had just a couple
of Ts and a lot of Cs. So that's what we were looking for, right? That's the difference
between an affected and an unaffected tissue in one patient that should be genetically
identical, there's a difference.
Now let's look across a number of the other 17 samples that we did and you can see here
is one where it just didn't look exactly right. So, patient 60 [spelled phonetically] and
their affected tissues had a little bit of Ts, but mostly Cs, and their unaffected tissue,
they actually had more, proportionately, more Ts than in the affected, which is kind of
odd. Here's a patient for whom we have only two affected samples and you can see this
patient had a lot of Ts in one of the affected samples and fewer in the other. Here's a patient
where there were actually no Ts in this sample. This one has none. This one had just a couple.
But this was enough, using -- in those parental samples, none of them had any Ts. So this
led Marge to think that this change we're seeing across multiple patients, and even
though this relationship isn't perfect, this was just too suspicious and led us to do a
lot more work. So that's what that sequencing of 17 individuals got us was to take us -- and
I'll tell you also, I should mention screens like this, each patient's sample pair set
that Marge looked at had between 100 and 300 positions in the genome that kind of looked
this way a little bit, and she had to sort out which one of those was real out of the
hundreds of differences that she saw by doing that work.
And then we made another molecular test, because we couldn't afford to do exome sequencing,
we had almost 200 samples from our patients, so we obviously can't spend $3,500 on every
one of those samples. We had to design a cheaper molecular test, and so this is some technical
stuff that we'll skip, but that leads to an output then, it's a different kind of molecular
test that looks like this. So we can take a blood sample from a patient and that blood
sample, using this test, will generate two signals, a signal here and a signal here.
I'll start here and this peak of this skinny mountain here, it actually tells you how much
of the normal sequence the person has at that one position in DNA and this peak tells you
how much of the abnormal or affected sequence they have in their sample. Here's a sample
where it's about 50/50. These two peaks are equally high. Here's a sample where it's about
between 25 and 30 percent, because the abnormal peak is smaller than the normal peak. Here's
one. You see that little teeny bump there. That tells you that the DNA ratio is two percent
in this sample and here's one, which is a blood sample, that there is no peak there
whatsoever, showing that there's none of that in this patient's blood sample. Then we take
those and she did that for nearly 160 samples and that then looks like this, and you may
recognize this from the paper.
So here's a graph and every one of these spots, or every one of these spurs or petals on these
little symbols, represents one sample. The black-colored samples represent samples from
things, tissues, that we thought were affected, the red ones from tissues that we thought
were unaffected. The blue ones, there were some samples we took from patients where we
just weren't sure if it was affected or unaffected, and then there were some samples from patients
or persons who did not have Proteus syndrome at all, because we have to compare all these
things simultaneously. And this just shows you for each patient, each row is one of our
patients who was in the study for whom we had a sample, it shows what the percentage
of this DNA alteration was. And what you can see is, you just sort of glance at this, it's
a lot of spots and spiky spurs, but what you can see is that there seems to be more of
the black and blue spots out here and more of the red ones down here, telling you that
in patients with Proteus syndrome, this alteration is more commonly present in affected or unknown
tissues, and that the alteration is more often absent in unaffected tissues, which is exactly
what we were looking to find.
If you look at the samples of patients with Proteus syndrome, of their blood, you can
hardly see this. There's so many little spikes on this that it sort of blurs together, but
again, there were 29 here. In 29 out of 31 of the samples there was no detectable alteration
in the blood sample of these patients with Proteus syndrome; however, in two of them
we saw a little bit, so about 10 percent. We took blood samples from lots of parents
with Proteus syndrome over the years. All of them were uniformly zero for the alteration
and we also have cells and tissues taken from patients with other conditions and all of
those had zero percent of the alteration.
So when you turn that into English, what can you say? You can say that out of 31 patients,
and these are all patients who met those clinical diagnostic criteria that I described awhile
ago, 29 out of 31 of those patients had the identical one base pair change in their DNA
at this position in chromosome 14 and the DNA change is, again, exactly the same in
every single patient. Two patients that didn't have that are not clinically different, and
there's another explanation for why we're not finding that in these two patients. The
alteration is more common in affected tissues. The alteration is absent in most blood samples
of most patients with Proteus syndrome, so thank goodness we did not spend all of that
sequencing money on sequencing blood DNA, because we probably wouldn't have found it.
We have looked at DNA sequence data from lots and lots of individuals. We ran another research
study where we have done DNA sequencing on 572 adults who don't have anything like Proteus
syndrome. Not a one of them has this change in their DNA and there's also a sequence data
set out, which is random control individuals, and there's 30,000 sequence results from those
patients and one out of 30,000 has one alteration in it, which is probably a sequencing error.
So if you sequence 1,500 people you don't find this. Proteus syndrome you find it in
29 out of 31, that's a huge difference. Geneticists love to do what we call genotype/phenotype
correlations. We want to understand what is the relationship between the alteration and
the DNA in the person and where the disorder manifests in them, how severely it manifests,
how it's going to manifest in the future, because that will help us to take care of
and treat the patients.
Overall we can't say that there is a correlation of where those spots were on that graph with
how severely the patient is affected, and the two patients, again, who were negative
for the alteration did not seem in any way different; however, we do have some interesting
results since the paper was published and we can take a skin biopsy from a patient here,
so this is the lesion a lot of you know about, the linear verrucous epidermal nevus, or the
cerebriform connective tissue nevus. We take a skin biopsy from that and we can grow two
kinds of cells out of that skin biopsy. I should say we, this was actually done by Tom
Darling, and we can grow out the cells that are the superficial layer of the skin, the
layer of the skin that would be responsible for this kind of rough, brown discoloration,
or we can grow out the cells that grow deep within the skin that generate the cerebriform
connective tissue nevus. When we do that, it turns out when we culture the deep cells
from these lesions, we see lots of the altered, the genetic alteration in those cells. When
we culture out the superficial skin cells from this lesion, we don't see much, and that
makes sense because this is a deep skin lesion. This is a superficial skin lesion, so we get
the opposite result here. We culture the superficial cells. You see the alteration. When you culture
the deep cells you don't, so that begins to make sense with where, again, that mosaicism
model, so where the cells have ended up in the tissue then determines what kind of a
lesion will be present in that tissue. The mutation/alteration in the deep cells will
give you this. An alteration in the superficial cells will give you that.
All right, so all I've told you about is some DNA change, right? That doesn't mean much
by itself biologically. It tells you that that is associated with the disease that is
causing the disease, but how is it causing the disease? That's what you want to know.
So it turns out this alteration in the DNA is a known alteration in the DNA that has
been described before. It's a change in a gene that's named AKT1 and it predicts -- remember
how I said that the base pairs determine the amino acids or the sequence in the proteins,
it predicts a change in a protein that we name as this, P-glutamate 17 lysine [spelled
phonetically], so it changes one amino acid and one protein. Again, as I said, this alteration
has previously been known to occur in literature. Lucky for us, because this is a huge amount
of work that had already been done by the time we made the genetic discovery. This is
a database of genetic alterations that are found in tumors of patients that have various
kinds of cancer. What this shows here, this is a histogram or bar graph, that shows the
different alterations that can be present in this gene. If you imagine the gene is along
here and you put a bump along here wherever there is a sample that has an alteration in
this gene, what you see is there is a huge spike here, so there's 121 occurrences of
this DNA alteration in cancers and then there's probably five or six clustered right along
here and none here.
And it turns out that this alteration, the one that's 121, is exactly the same DNA change
that's present in patients with Proteus syndrome. Where is that coming from? It's coming from
a number of different kinds of tumors or cancers in these individuals and is most commonly
breast cancer. It's in -- about five percent of breast cancers have this mutation and then
a number of other cancers have it at a lower frequency. That is important, because that
finding previously, which was unrelated to Proteus syndrome, had been studied in great
detail in a paper in Nature Magazine by a former colleague of mine and he studied this
exact mutation because he was finding it in some cancers and thought that it could be
related to cancer. Long story short, what they found is that this alteration, this DNA
change, changes the structure of what this protein does and determines what it binds
to and what it doesn't bind to. The protein -- this alteration changes where this protein
is within the cell that it's made in. It causes this protein to be turned on all the time.
So proteins have -- some proteins have jobs what we call signaling. They turn things on
and off. They're like switches within the cell. What this DNA and protein change did
to this protein is altered it so that it was kind of turned on all the time.
That gave the cell some characteristics that they shouldn't have, like they tend to grow
too much and they don't stop growing when they reach a status where they're completely
filling the space they're in. They just keep growing. And if you put these cells with this
alteration in mice that have another gene mutation, they actually get a form of cancer
or leukemia. So, again, here we are now down at the protein level. We found the DNA change
that's related to the disorder. We found the protein that it's in and we understand what
the function or the dysfunction is of that protein in the cell. Now things get even more
complicated, all right? So here's our protein, AKT. This protein interacts with and talks
to, and that's sort of a metaphorical sense, talks to other proteins within the cell to
turn things on and off, and there's all these different proteins within the cell that it's
connected to and it says to turn things on and off. It turns on cell growth programs,
cell death programs. It turns on other genes. It turns off other genes, et cetera.
Then I want to draw your attention to the fact that one of the proteins that it's talking
to here is our old friend p10. So, AKT is near p10 in this protein talking network and
they talk to each other and turn each other on and off in a reciprocal fashion. Now we
can actually measure whether the protein is turned on or off by an assay that I won't
go into too much here, but when you see in this test, this is a test that tells you whether
the protein is turned on. Now we grow cells, believe it or not, the way you grow cells
is you put them in dishes with media that has sugar and salt, et cetera, in it, but
one of the things we also add to it is cow serum, believe it or not. You need cow serum
to keep cells alive, some kind of a serum, because it has growth factors in it. When
you grow Proteus syndrome cells, or any cell for that matter, in cow serum, this protein
is turned on really powerfully. And what that tells you is that the proteins in that cow
serum are telling the cell go, okay? Grow, expand, proliferate, et cetera, and that turns
on AKT1 big time.
And you can't see any difference between a cell that comes from a patient with Proteus
syndrome versus a cell from any other person when you grow them in that serum. But if you
take the serum away, you take away the stimulation to those cells, and what usually happens is
that, we call that serum starvation, the cells will slow down or stop growing and sometimes
even die. And when you take away the serum from cells from a patient who doesn't have
Proteus syndrome, you can see AKT completely turns off. It's off. It's gone. The protein
is essentially zero. But when you take away that serum from a patient with Proteus syndrome,
it doesn't shut off. Okay?
Furthermore, we did another experiment. Poor Marge, when you see her thank you for doing
this, she took cell cultures from patients with Proteus syndrome and these are dishes
of cells that have hundreds of millions of cells in them, and what she did was took those
cells out of the dishes, separated them from each other and then took one cell at a time
and tried to re-establish new cultures from just one cell. Okay? Tedious, tedious work.
But if you're thinking about mosaicism, this is exactly what you need to do, because you
know that your tissues from the patients are a mix. Some cells have the trait, some don't.
So what you want to do is separate out the ones that have it from the ones that don't
and she did that. These are called single-cell clones of patients from Proteus syndrome.
So what these two spots represent, this is one cell culture from one patient sample with
Proteus syndrome that was separated into two separate single-cell clones and we tested
the single-cell clones to find that one clone did not have the Proteus syndrome alteration
and the other one did. Okay? So these are, again, single-cell clones derived from one
sample from one patient. Again, when you take away the serum you see that cell, AKT, turns
off, but if it has the Proteus syndrome mutation it's turned on, and that is the data that
proves that Proteus syndrome is mosaic and is caused by this alteration that causes this
protein to be turned on when it's not supposed to be turned on. So it makes the switch from
an on/off switch to an on switch to a weakly on switch. It can never turn off. Okay.
You all know we've struggled for years with diagnostic confusion with Proteus syndrome
and these diagnostic criteria, et cetera. Proteus syndrome overlaps with a number of
disorders. This controversy about Proteus and p10 syndrome we can begin to now understand
better and explain. You know, patients with Proteus syndrome have been reported before
to have alterations in the p10 gene and what we now know is that p10 disorders are, of
course, somewhat similar to patients with Proteus syndrome, but we thought it was clinically
distinct, and what we now know is that they overlap because these two proteins talk to
each other and can have some overlapping effects, because one is telling the other, helping
the other turn on or turn off, and so it makes sense that the disorders are related.
How do we do the diagnosis? We have a lot of work to do, because as you know, the alteration
is present at low levels in the blood samples from our patients, so it won't be a simple
matter of just drawing some blood and doing a blood test to see if the person has Proteus
syndrome or not. And we're going to develop more sensitive detection problems, because
what we think is even though we saying this is zero percent now, this assay can only detect
down to one percent mutation levels, and we think that that alteration may be present
in those cells at lower than one percent level and we need new technologies to do that.
All right, so we know from years of experience what Proteus syndrome is. It's a progressive,
patchy, overgrowth disorder that can affect just about any tissue. Connective tissue and
bone obviously are the big ones that cause most of the overgrowth problems. Patients
can have vascular malformations. Certainly if severe it's progressive. There are some
patients with Proteus syndrome who have had tumors and now we know that this disorder
is caused by a single DNA alteration is this gene AKT1 and that explains a lot of things
about Proteus syndrome. It explains why it's so rare, because it can only be a single alteration
that can cause the disorder and it has to be mosaic. It explains the overgrowth in this
disorder. We know that this AKT1 pathway is a key regulator of growth in the developing
human. It explains why some of our patients have had tumors. AKT1 very mildly contributes
to tumor development and growth and so that gives the patients a subtle susceptibility
to tumors and explains that part of the disorder.
It also provides us an avenue for early diagnosis in the future, which as you hear from Julie,
will be a key thing in going forward with future research and studies and the most exciting
thing is that AKT1, because it has already been known to be related to some cancers,
several pharmaceutical companies have been working on drug compounds that inhibit this
protein. Now remember, what is wrong in Proteus syndrome we know is that this protein is turned
on more than it's supposed to be, and from a treatment standpoint, that turns out to
be a very, very fortunate circumstance, because that means if it's turned on too much what
we need to find is a compound that turns it down. And it actually is easier to turn something
down than it is to turn it up. If I want to -- if I have a patient who has a disorder
that's caused by the loss of function of a protein, it's really hard to find the treatment
to turn that protein back on. But it actually turns out to be easier to turn things off,
so the fact that this is a constitutive or alteration that turns a protein on, is fortunate
for treatment. So our management is going to be mostly the same for the near future.
We have to continue these conservative surgical interventions, monitor our patients for the
deep vein thrombosis and embolism, make sure we don't get caught in what we call a surgical
catch-22, continue to monitor our patients for signs and symptoms of tumors, work hard
on the lung disease, which has severely affected a number of our patients, and now going forward,
begin to think about AKT1 treatment.
So what I want to do now is thank everyone, especially the foundations for raising the
money that allowed us to do this sequencing. Lots of co-authors were present on this paper
and I do want to draw your attention to, I'm just so tickled about this, I'm so annoyed
she's not in the room, two of the authors on this paper, you might notice, are here.
And poor Kim and Tracy [spelled phonetically] and we thought that they were appropriate
representatives of what the great things that you all have done to contribute to this effort,
and so we had to enlist them, I think was, was it in March? March or April? And what
was happening and the process of finishing up this paper and submitting it and having
it reviewed having them be co-authors, we swore them to secrecy, because scientific
publications have to be kept confidential until they are published. And so poor Kim
and Tracy had to keep this to themselves and it is completely my fault that she had to
do that, and so please forgive her for doing that, but we really thought it was essential
to acknowledge you all through these two contributors. So thank you very much, and I'm way over time.
Oh, I'm not so over time. I'm okay. I thought I was done at 10:00. All right, so we do -- ladies
we have time for a couple of questions? Is that what you want to do? Okay. Anybody have
a question or was that perfectly clear?
Yes, sir.
Male Speaker: Will this be the new [unintelligible] for
Proteus syndrome [unintelligible] these people have, you know, the classification of Proteus,
will this be the next test to verify exactly if you have it or [unintelligible] AKT protein?
Leslie Biesecker: Great questions. Okay, so what Brian's [spelled
phonetically] asking is how are we going to diagnose Proteus syndrome in the future? Will
it be based on the clinical criteria or this molecular DNA criterion that we have now found?
So that will be an interesting research target for us to work on in the next couple of years.
Now, what we've done here, if you think back, is we used those clinical diagnostic criteria
to very, very carefully strictly and rigorously define a subset of patients who we were really
as certain as we could be had Proteus syndrome. Okay? We took samples from those patients
and showed that this is present in nearly all of those patients. All right? So that
then proves that this alteration can cause this condition, but what it doesn't tell you
is if you ask the question of if I consider more mildly affected patients or patients
who present with signs and symptoms that are similar to, but not exactly the same as Proteus
syndrome by those strict criteria, I don't know what the alteration I'm going to find
is, because I haven't studied those patients. Right?
So you get into a little bit of a circular reasoning problem here and we have to be careful
we don't get fooled by that. So what we know is in patients who have what we define as
the strictly defined manifestations of Proteus syndrome, we find this alteration in almost
all of them, and that will be a reliable and robust test for people who present to us early
and that will be a great thing to allow us to make the diagnosis earlier, because some
of you have experienced this where we have evaluated you or your child over multiple
visits as the condition evolved before we were convinced that the person had Proteus
syndrome and that sort of prolonged torture of figuring out if it's going to be Proteus
or not, can actually be changed now, because we'll be able to do biopsies and if we find
this alteration in the biopsies, that will tell us that it is Proteus syndrome.
We also need to expand that and then think more broadly. So, if we don't apply such strict
criteria, how often do we find this alteration in patients with overlapping conditions? And
that's what we'll be working on in the future. How often do we find it in patients with CLOVE
syndrome and hemihypertrophy and these other conditions? We don't know the answer to that
yet and we're working on and collaborating with people to try and figure that out. So
for now what I think this will do is allow us to identify patients who have this condition
earlier on and be more certain about the diagnosis early in the process, and that's actually,
I think, Julie's going to bring that up when she begins to talk about treatment trials.
That'll be a really important thing when we start to think about who we're going to treat
and when we're going to treat them. So it's great. Did I answer -- you had two questions,
Male Speaker: That summed it up pretty much.
Leslie Biesecker: Okay, good. Yes, sir.
Female Speaker: I was just wondering, like, the gene you discovered
[unintelligible] is it in both CLOVE and Proteus?
Leslie Biesecker: It is not yet known to be altered in patients
with CLOVE syndrome.
Female Speaker: [unintelligible] it's so rare it's not like
you can have a mandate [unintelligible]?
Leslie Biesecker: Oh, that -- great question. Right, so what
you're asking is it's an ultra-rare disorder, so we're not going to be like doing this test
at a routine pediatric checkup.
Female Speaker: Right.
Leslie Biesecker: No, exactly, because the yield would be pretty
lousy on that. Right, but what it will do, and some of you will very, very clearly remember
this earlier in your experience before you came to the diagnosis is, what typically happens
is, you know, early in childhood some of the patients have subtle signs of the disorder
and you go to a number, usually a string, of physicians and say what is this and they
say I don't know, it's not too clear what it is, and you see a number of specialists.
We'll then short-circuit that process and we'll get to a diagnosis much faster and much
more clearly.
And now we have to work on figuring out what is the alteration in patients with things
like CLOVE syndrome, but this finding actually now really drives home the point. If you think
about the fact that the patients with the p10-related disorders that we've been struggling
with over the years, right, they overlapped a bit with Proteus syndrome, clinically how
they manifested their symptoms and manifestations. And those two proteins are talking to each
other, we know now, right? But I also showed you there wasn't just AKT and p10 on that
network, there was a whole bunch of other ones. So the most likely thing is that patients
who have other overlapping disorders, those overlapping disorders are probably caused
by an alteration in one of those other proteins that are on that pathway. All right?
So now we can begin to, instead of having to think that it could be any of 20,000 or
so genes, we're now down to maybe a couple of dozen, and so that gives us the target
of what to look for in other overgrowth disorders that also overlap with Proteus syndrome, because
like Proteus and p10, overlapping manifestations mean it's probably related in a similar pathway.
Female Speaker: Would you be able to apply like gene therapy
once you've concluded that the person has Proteus syndrome?
Leslie Biesecker: So, a question of gene therapy. So gene therapy
is a really hard thing to do and, in fact, is fairly dangerous, because it can cause
really severe side effects.
Female Speaker: Right.
Leslie Biesecker: So gene therapy is currently being developed
for disorders where you have no other good treatment options, like for a drug or something
like that. Okay? It's a desperate situation. If you have a disorder that's caused by an
alteration in a gene and that gene codes for a protein that is a good drug target, you
are much, much better off working on a drug to treat that disorder than you are for gene
therapy, because it's much easier to control how you give and how you dose a drug than
gene therapy.
So what you want to do is find something that's as effective as it could possibly be and is
the least toxic to the individual. And so you could say I want to go for broke and try
and fix this by gene therapy, but you take a huge risk in trying to do that, whereas
you can take less risk with potentially good outcomes if you try and develop a drug, which
if it causes side effects, you just reduce the dose. You give it less frequently or you
stop giving it and you switch to another drug. That's a much safer approach to figuring out
a treatment. Great question. Yes.
Female Speaker: Where are they in finding a drug [unintelligible]?
Leslie Biesecker: Yeah.
Female Speaker: [unintelligible]
Leslie Biesecker: Right.
Female Speaker: There's a lot of money [unintelligible] --
Leslie Biesecker: So, where are we in finding the drug? That
is a superb question. And, again, this relates to the fact that this alteration when we landed
on that DNA change in Proteus syndrome and found that this exact alteration had been
known three or four years previously to be associated with cancer, was scientifically
-- I know these are all sort of mixed blessings, right, in the big picture -- but scientifically
was a wonderful thing, because the existence of that paper that I showed you basically
short-circuited probably five years' worth of work on our part, because we didn't have
to do the work to figure out what this DNA change did to the cells that have that DNA
change. That had already been done by a huge amount of work by another group and not only
that, because this DNA change is present in cancer, two or maybe even three, companies
have already been working on compounds that have gone through all of the chemistry that
you have to do to find drugs, through animal toxicity studies, and some of them have even
been used experimentally in some humans in what we call phase one trials, which I think
Julie is going to talk about what that means.
So that means that at least one of the compounds has already been, and probably two of them,
have been through phase one, which means they've been designed, they've been tested in animals,
they inhibit this protein in cells, they've been given to patients, and they don't cause,
so far, any severe side effects. So that right there, what I just described to you, is again,
between five and 10 years' worth of work that has already been done and we landed on that
the day we found this alteration. If we had landed in a part of the genome where there
was nothing known, we would have set out to start to do that work and it would've taken
us, again, years to get to that point. But because we landed on this particular alteration,
that's already been done for us and we are the beneficiaries of that other work that
has been done by those groups. So that has incredibly accelerated the progress.
Male Speaker: You might've opened a Pandora's Box with all
the new research and we probably have a hundred different [unintelligible] you couldn't [unintelligible]
in your upcoming research. What are your plans as far as continuing on? What would you be
interested in looking at, and how can the foundation continue to help you in funding
or otherwise to support that work?
Leslie Biesecker: Great questions. So the question centers on
the fact, as you know as a scientist, every scientific finding you make opens 10 or 100
new doors of things you could do and one of the key things about science is to stay focused
on what the big picture, big goal, is that you really want to solve. So number one for
us is to move toward the treatment for this disorder, and so we're actually meeting very
soon, I think it's in two weeks, with one of the companies that has developed one of
these compounds. You know, we've done a whole bunch of paperwork with our legal office to
allow us to have confidential data from that company, et cetera, and so we'll be meeting
with them and seeing if they will be interested in supporting a trial to actually start a
therapeutic treatment trial for this condition. So I'd say that's sort of job number one.
Number two is to understand, you know, we have the nucleus of a finding here. Again,
an alteration in typically-affected patients, so we need to figure out fast and soon if
we look at more mildly-affected patients, do they have the same alteration, or do more
mildly-affected patients have a different alteration in this gene? Or do mildly-affected
patients have an alteration in one of those other genes in this pathway? We've worked
with a number of families and affected patients over the years, who we have undiagnosed with
Proteus syndrome and we have not forgotten about them and we are going to start working
to see if we can figure out what are the causes of the phenotype in those patients. Again,
reasoning that it will be either a different change in this gene or a change in one of
the other genes that are in this pathway.
We also need to do some basic work on developing what we call animal models for this disease.
There isn't a perfect animal model for Proteus syndrome, because that's a tricky thing to
do, and we're going to work with some collaborating groups to try and make an animal model so
that if these first known treatments don't work, that we have a fall-back position where
we can have an animal model where we can test on not-yet-developed drugs in an animal to
move that toward human trials. So we have to sort of hedge our bets and divide our efforts
so that we have a number of pathways forward, both to bring in the other patients who have
different forms of Proteus and overgrowth syndromes, as well as all the infrastructure
we need to move toward effective treatment.
Male Speaker: The second part of my question was how can
we help in fundraising or otherwise for that work?
Leslie Biesecker: I actually -- can I defer that to -- yeah.
I'm going to defer that to Julie. That's a super question and I think after Julie's presentation,
ask that question again, and I think it'll be very well answered. Brian.
Male Speaker: In treatment of the AKT1 that [unintelligible]
chemotherapy be affecting that gene due to some of the protein that keeps growing, [unintelligible]
steroids for whatever procedure they're having or if they had another procedure [unintelligible]
chemotherapy for some kind of cancer, would that change [unintelligible] the growth factor
[unintelligible] or keep it growing [unintelligible]?
Leslie Biesecker: Great question. The question is would steroids
or chemotherapy have an effect on Proteus syndrome? I think reasoning that AKT1 we know
can be involved in cancer and sometimes you use steroids and chemotherapy to treat cancer,
so would that work with Proteus syndrome? Now one of the key benefits of doing this
kind of molecular genetic research is that it allows us to understand at a very precise
level what is wrong, and that allows the development of very, very focused, or what we call targeted
therapies that alter the very specific thing that is wrong.
Clinically in medicine for years we've used lots of therapies that are not so specific.
Steroids would be a good example. Chemotherapy is another example. Now what I like to analogize
to is a tool. So if you have, if you're working on a wood project, and it's got the tiniest
little brad that you need to drive into that wood to make that thing work right, if you
approach that brad with a sledgehammer --
It's not going to turn out well. Okay? And so if you use these coarse general therapies
to approach a very specific problem, they can work and, you know, steroids do work for
a number of medical conditions and they have a lot of side effects, and chemotherapy works
for cancer and it has a lot of side effects.
The promise of sort of the future of molecular medicine is not to be using sledgehammers,
but to go in there with a very specific hammer that's designed to hit that nail right on
the head and do exactly what you want without messing up all the other stuff. And so that's
what this gives us is one molecule, and that molecule is our target and that's what we
have to go after and the promise, the hope is, what I'm optimistic for, is that we can
specifically target this alteration that is present in the cells of our affected patients,
turn that protein back down just enough that things will either slow down or maybe even
stop if we can tune it just right.
So I'm past my time so I should probably stop now. We have lots of time for question --