Tip:
Highlight text to annotate it
X
Hal Dietz: Good morning. It's really a pleasure to be
here. Thank you for that kind introduction. So I have a tall task today, and that is to
discuss treatment of genetic disorders in general. I'm going to approach this by presenting
many small vignettes that reveal both the power and pitfall of some new approaches that
are being used that have the potential to be very powerful.
So when we're first thinking about a disease -- I'm sorry -- often we feel stranded on
a desert island. We really have no compass and we can look in 360 degrees without clear
conviction about what direction to go in. Now genetics has received a lot of hype, appropriately
so, for its ability to allow you to make giant leaps in understanding without taking each
small derivative step along the way. You can learn about a new disease gene and suddenly
be firmly within a pathway that you never would have anticipated. And initially when
you get there, things look very appealing. But once you get your bearings, you often
learn that you're no better off than where you started and that it takes a lot of hard
work to reach the promised shore.
So initially when people think about treatment of genetic diseases, often gene therapy comes
to mind first. It's a fairly straightforward strategy: if the body is missing a gene product
because of a defective gene, you simply use a virus to reinsert the normal gene into the
cells of interest, which will then make the protein and restore function. Gene therapy
can also be used to treat patients' cells that have been removed from the body, again
giving them the normal gene and allowing them to go back into the patient to restore function.
There are many obstacles to gene therapy, including an immune response to viral proteins.
There've been a number of very catastrophic cases of failure of gene therapy due to this
effect. And also because of disruption of essential genes when the virus inserts the
gene into the host genome, for example, causing leukemia. But despite this there've been many
startling recent successes of gene therapy for conditions such as adenosine deaminase
deficiency, causing immunodeficiency, for Leber's congenital amaurosis, causing blindness,
and also for hemophilia B. So I'm not going to spend a lot of time on gene therapy. There
is a reference that is provided in the -- will be provided with the slide set that goes over
that topic in detail.
Rather, I want to cover the basic concept that it takes a village to really bring a
new concept to a patient. And then it really requires a confluence of and synergy between
both basic and clinical sciences to develop a full mechanistic understanding of a disease
process, and in that manner, to derive novel and rational treatment strategies. So the
first story I want to tell you concerns the so-called lysosomal storage diseases. A lot
of very well-known names here such as Hunter disease, or Hurler disease, or Pompe disease.
All of these conditions are unified by the toxic accumulation of lysosomal substrates
due to a deficiency of a lysosomal enzyme. So, real progress in this area actually derived
from an accident in a laboratory, in Liz Neufeld's laboratory.
For this particular example, I'm going to show a green cell as a normal cell that has
a full complement of enzymes and normal function and various flavors of red cells to show a
defective cell lacking a lysosomal enzyme. And you can see in this hypothetical example
that in this cell there's no Enzyme A, shown by a yellow circle, causing lysosomal storage
disease 1. In this cell there's no Enzyme B, causing a different lysosomal storage disease.
The accident came when a technician in the Neufeld lab accidentally mixed cells from
two different patients with two different lysosomal storage diseases. And the remarkable
result is that when they were mixed, all the cells adopted normal function; suddenly they
worked, everything worked. So this is called a complementation, where one cell will complement
the deficiency of another cell, and it really herald the concept of treatment for lysosomal
storage diseases simply by infusing the defective enzyme or the deficient enzyme. So in this
example, once again, this cell does not make the enzyme shown in yellow, this cell does
not make the enzyme shown in blue. What the Neufeld lab learned is that each cell secretes
the enzyme that it makes outside of the cell. And then the other cell can take up that enzyme
and in fact complement its deficiency by taking up the enzyme that its neighbor made. And
again this leads to cells that have both enzymes and normal function.
So everyone is understandably excited about the prospect of simply infusing enzymes, but
then there was a pretty startling and disturbing result that if you made the enzymes that are
defective in these patients, and simply put them in the culture medium of the cells, the
cells failed to take up that enzyme and there was no complementation. And the obvious question
is why, what's missing? So the answer came when the Neufeld lab studied a different disease.
This disease is called I-Cell Disease. And what they learned is that cells from patients
with I-Cell Disease make all the appropriate enzymes, they secrete all those enzymes outside
of the cell, but that other cells lack the ability to take up those enzymes that were
secreted. And what ultimately they learned is that in I-Cell Disease, there's deficiency
of a specific enzyme called N-acetylglucosamine-1-phosphotransferase that modifies all these enzymes by attaching
a chemical called Mannose-6-Phosphate. And it was learned that a recipient's cells, in
order to take up an enzyme, have to have a Mannose-6-Phosphate receptor that will then
bind the enzymes in its environment and facilitate cellular uptake. So now there's a complete
strategy for enzyme replacement therapy. You could synthesize these enzymes in the laboratory,
but you had to add this modification, the Mannose-6-Phosphate Group. So does it work?
The answer is yes.
So here's an example in treatment of Hurler Syndrome with recombinant enzyme alpha-l-Iduronidase.
You can see that all of these patients at the start of therapy had rather high New York
Heart Association classifications, but during the course of therapy, about 50 weeks of therapy,
all of them showed a clinical improvement and many of them showed a very striking clinical
improvement down to the best level. If you look at other markers such as liver size -- the
toxic substrates and lysosomal storage diseases often accumulate in the liver -- you can see
that liver size decreases dramatically during the course of infusion of the deficient enzyme.
And you can also look at skeletal performance. Patients typically would develop contracture
and you can see that there's increased movement at the shoulder, at the elbow, and at the
knee while these children are receiving infusion of the enzyme.
So I don't expect you to be able to read anything on this slide, but the basic concept is that
there are now many enzyme replacement therapies for many lysosomal storage diseases that are
either in late-phase clinical trial or that are FDA-approved to use for treatment of these
children. Is this the whole answer? Is this the cure for lysosomal storage diseases? Unfortunately,
there are some obstacles. In the example that I showed you, the enzyme with its Mannose-6-Phosphate
group is free to interact with cells in their Mannose-6-Phosphate receptor, leading to functional
recovery. But it was recognized that some patients over time develop antibodies that
attack the enzyme that's being replaced. And what distinguishes these two groups of patients,
those without and with antibodies, is the ones without antibodies make some small level
of the enzyme that's being replaced. It's not enough to achieve the function that is
necessary, but it's enough for the body to say this enzyme is okay, it's not foreign,
I'm not going to attack it. However, in patients that make no enzyme naturally, their body
has never seen that enzyme. So it's recognized as foreign and therefore attacked by the immune
system. Another obstacle is the so-called blood-brain barrier. The micro-vessels within
our brain contain a very solid barrier that prevents diffusion of multiple substances
from the circulation into brain tissue. And unfortunately, this blood-brain barrier is
impermeable to all the enzymes used in enzyme replacement therapy.
So currently, there's excellent utility of enzyme replacement therapy in conditions such
as Maroteaux-Lamy where there's no central nervous system manifestations, but there's
really very poor performance of enzyme replacement therapy in diseases such as Gaucher disease
type 2 or 3, where there's very severe brain manifestations and the enzyme simply can't
get there to help. There are some potential solutions. You know, just like children receive
allergy shots to teach their body to tolerate something that's recognized as foreign, similar
approaches are being used to induce tolerance to enzyme replacement therapy. There's alternative
targeting procedures that are being explored that might allow these molecules to get past
the blood-brain barrier to get into the tissue. And there's also intensive exploration of
complementary therapeutic regimens that use drugs rather than antibodies, and drugs have
a greater ability to diffuse into the brain.
There are many different classes of drugs that are being explored. In this scheme that
I'm showing you here, I show you a normal protein that's folded properly, it's transported
appropriately within the cell, it has its intended activity breaking down some substrate
into a metabolite. On the bottom I'm showing you a patient that has a mutation or a change
in that protein causing the protein to fold abnormally. Sometimes these abnormally-folded
proteins traffic properly to the right place in the cell, but they simply can't work. Other
times unfolded proteins are degraded within the cell. So one way to approach this would
be to just ignore the enzyme completely and come up with an alternative strategy to remove
the toxic substrate that's accumulating. So this is called substrate reduction therapy.
Another possibility is to simply block the toxic effects of this substrate that's accumulated
within the body, something that's called a pathogenic modulator.
There are other flavors of drugs that are being explored. For example, drugs that bind
the abnormal protein and allow it to fold properly, something called a chaperone drug.
There are also drugs that can cause proper folding and restore the activity of the enzyme.
And finally, there are stabilizer drugs that prevent the cell from breaking down abnormally-folded
proteins. And all of these classes of drugs can lead to some enzyme activity and therefore
a reduction in the amount of the toxic chemical. And once again, just for illustration, there
are many of these drugs that are in either clinical trial or are already FDA-approved
for some of these conditions.
So now I want to turn to a different disease, and that is cystic fibrosis. I'm sure you're
all familiar with this condition. These children have mutations in a protein called CFTR and
that leads to things like chronic lung disease, it leads to pancreatic insufficiency, it leads
to multiple problems with intestinal performance. So ultimately, these individuals typically
die of their condition. It used to be that they would die in early childhood. Now with
just supported therapy, many are living to mid-adult life. But again there's the question,
"Can we do better?" There are many possible problems that can occur from the -- if you
begin in the nucleus with the DNA that encodes this channel, the CFTR channel, all the way
to the protein being at the cell surface and having its function. You can have problems
producing the channel protein. You can have problems processing and trafficking the protein
within the cell. You can have problems regulating its activity; it's either inappropriately
on or off. Or you can simply have problems allowing chloride to move through the channel.
So this is perhaps the most exciting story, in my mind, in therapy for genetic disorders
that's come out recently. And it all stems from a public-corporate partnership between
the Cystic Fibrosis Foundation and a drug company called Vertex Pharmaceuticals. This
partnership set its sights high but focus narrow. So what they wanted to do was develop
a drug therapy for people with the class III mutation G551D. This means that the glycine
at position 551 in the protein is changed to an aspartic acid. And you might ask, why
so narrow? The chance of finding a drug that would address all the problems, all the potential
problems in making this channel, trafficking this channel, and regulating this channel
was really low. By definition, a drug that binds to the CFTR protein improves the function
of this particular mutant form might have potential to bind other mutant forms. So let's
start with this mutation, but then we can see whether this drug works for many different
people with many different mutations. And even at a minimum, a drug for this particular
mutation would address about 4 percent of all people with cystic fibrosis. This happens
to be a fairly common mutation.
So how do they go about finding a drug for a patient with a specific mutation or change
in a specific protein? They used a process that's called small molecule screening. So
what you would do is you'd take a plate and you'd put patient cells into each of these
welds and you'd also introduce some sort of marker, some sort of, often a florescent marker
that will only glow when activated by chloride conductance when the performance of the channel
has been restored. And then you take a library of small molecules. These could include hundreds
of thousands of small molecules, often includes established FDA-approved drugs to see if they'll
have an effect. And in each weld of this plate you'd add a different drug. So you're basically
screening all the drugs within this library to see if they have any beneficial effect
on restoring how the channel works. You then read the marker, the fluorescent marker for
example, and you score the performance of different compounds within this library. You
look for the welds that glow the brightest, for example. You might have a reasonable compound,
something that's doing pretty well, but not a perfect compound. Then you can take one
of your hits from the screen, you can tweak that compound a little bit using medicinal
chemistry, and then put it back through this whole process to come up with the best drug
for this particular purpose. And this led to the identification of a drug for this particular
mutation causing cystic fibrosis. And again, when I read this paper, it was one of the
few times that I sat back in my chair and smiled. I mean, the results were just so remarkable.
So here we're looking at patients who were either not treated with this drug with CF
or people that were treated with this drug. And you can see that you're seeing changes
in the sweat chloride, the amount of chloride that's moving through the channel. So these
people that were taking the drug dropped their sweat chloride significantly. And in fact,
the absolute sweat chloride level went within the normal range. So it didn't only nudge
this in the right direction, it shoved it in the right direction. They then looked at
performance of these patients and they saw that by about two weeks, there was already
this dramatic elevation of FEV1, so a marker of lung function. There was a significant
drop in the number of serious events like pulmonary exacerbations that these patients
were having. Their respiratory score increased dramatically. And they showed abrupt and sustained
weight gain over a full year while taking this medication. The one question was, "Well,
do they cherry pick their patients?" Was it just the most mild patients that would respond?
And the answer was no. The response was dramatic, regardless of initial severity, regardless
of geographic location, regardless of gender, and regardless of age. So it really seemed
to work for the full spectrum of CF patients with this particular mutation. The drug was
also tolerated extremely well. There was a significant reduction of serious adverse events
when you compared the treatment group to the placebo group.
So the conclusions of this remarkable study was that this drug, now called ivacaftor,
was associated with improvement in lung function at two weeks and that this improvement was
sustained through 48 weeks. But there was substantial improvements also in the risk
of pulmonary exacerbations, patient-reported respiratory symptoms, weight, and concentration
of sweat chloride, and that ivacaftor was not associated with increased incidence of
adverse events when compared to placebo. This drug was FDA-approved on January 31 of 2012
and it represents the first and only drug that is approved for the treatment of cystic
fibrosis. Currently in children older than the age of 6, with this particular mutation,
but again, the value of this drug is now being explored for other patients. Yes?
Male Speaker: This is fantastic, but just for economics,
what is a ballpark figure of what it costs per year of this [inaudible]
Hal Dietz: Unfortunately, I don't know that answer, but
I can find it out.
So I'm going to turn to a different disease now, again, a very common condition known
as Duchenne muscular dystrophy. These individuals show loss or significant impairment of muscle
function usually with a diagnosis at around the age of 5. Usually these patients are wheelchair-dependent
by their teens, and death occurs due to respiratory insufficiency by young adult life. In contrast,
there's a different condition that's called Becker muscular dystrophy. And these people
are typically not diagnosed until their teenage years, are not wheelchair-dependent until
mid-adulthood, and if death occurs, it's generally delayed until the fourth to fifth decade of
life. The remarkable observation is that both of these conditions, very severe and mild,
are caused by mutations or changes in the same gene, the dystrophin gene. So, you know,
an obvious question is, "What's underlying this clinical difference?" If you look at
the muscle and you use an antibody that reacts with the dystrophin protein, in normal muscle
you see that every muscle fiber is surrounded by a mantle of dystrophin. In Duchenne muscular
dystrophy, generally the protein is absent. And you can see in Becker muscular dystrophy,
while not normal, there is at least some preservation of dystrophin expression.
What does dystrophin do? Well, it serves as a link between molecules deep in the cell
and a cluster of molecules that are at the cell's surface, you know, establishing a bridge
between what's called the cytoskeleton of the cell and the extracellular matrix outside
the cell. In normal people, not only dystrophin, but all the other proteins within this complex
are easily seen at the cell surface. If you simply take away dystrophin, you lose this
whole complex, the entire complex is destabilized. So you can see from this diagram that dystrophin
plays an important bridging role and you might infer from this that dystrophin needs its
head to attach to proteins within the cell; it needs its tail to attach to proteins at
the cell surface. But the obvious question is, does it need all of its middle? Is it
important exactly how long this middle segment is? So we're going to have to look at this
in a little bit more detail to understand what's going on.
As with all other protein-encoding genes, the dystrophin gene is initially copied in
the cell to a molecule that's called precursor-messenger RNA. And this precursor-messenger RNA includes
blocks of sequence that encode the protein called exons and intervening junk sequence
that's called introns. And ultimately, all of these introns are removed by a process
that's called mRNA splicing. So you end up with all these coding blocks all in a row.
And there's a signal that tells the ribosome where to start making protein, that's called
the start signal. And there's also a signal that tells the ribosome to stop making protein,
that's called the termination codon. So in this normal example, a ribosome can come along
and it's going to read the triplet code of a messenger RNA. Every three base pairs encodes
a specific amino acid. So the ribosome will move along, you'll end up with a full-length
protein, and you'll end up with a normal muscle phenotype.
Now, in Duchenne muscular dystrophy, the most common cause is that one of the exon blocks
is skipped during splicing. And the important fact is that in Duchenne muscular dystrophy,
the block that is skipped is not an even multiple of three. So now when you splice these two
exons together that don't belong together, you're going to shift this triplet reading
frame. Everything downstream of that point is just going to be nonsense. And this invariably
leads to what we call a premature termination codon, something that will tell the ribosome
to stop too early. So again, the ribosome will latch on, it'll move along this RNA,
but it will stop early. And you might guess, "Well, that's going to cause a truncated protein
missing its tail." Now I've already told you the tail is very important. In fact, when
RNA has a premature stop signal, the most important consequence is that the cell simply
gets rid of that RNA for a process that's called nonsense-mediated mRNA decay. So you
don't have any potential to make protein from these prematurely terminated transcripts.
And the answer is you make no dystrophin and you get Duchenne muscular dystrophy.
So what's going on in the mild form, the Becker muscular dystrophy? Well, here you also commonly
have skipping of an exon, but in this circumstance, the blocks that are skipped are in even multiple
of three nucleotides. So your reading frame is going to remain preserved. You've got a
piece missing right here, but you don't have any premature stop signal, the code is all
intact. So again, the ribosome latches on, it reads through to the end, and it makes
a largely normal dystrophin protein that's missing a central block, but it's got both
its head and its tail to latch onto the appropriate protein complexes, and that's what causes
the more mild Becker muscular dystrophy. So the obvious question, is there anything that
we can do to turn this into this? That is a goal: let's make this more mild.
All right, so in order to explain how this happens, I have to tell you a little bit more
about splicing. And the only important point is that this is a very tightly orchestrated
process, it's not random. You have your precursor-messenger RNA that has different signals embedded within
it that tells protein complexes where to bind. And these protein complexes ultimately define
what's the beginning of the junk and what's the end of the junk, so it tells the cell
what to get rid of and what to splice together. And after this you end up with the appropriate
splicing event joining Exon 1 to Exon 2, and then this continues down the length of the
messenger RNA.
So what if you wanted to get rid of exon2? You thought that that might be a good thing.
So what can be done is you can introduce into cell a small little artificial piece of RNA
or DNA that we call an antisense oligonucleotide. You can make it so it attacks just this spot
at the beginning of Exon 2, for example. What would be the consequence of that? Well, some
of these protein complexes would continue to bind, but the protein complex that was
supposed to bind here is blocked. So this is no longer recognized as important sequence
and that would lead to skipping of Exon 2 and then you'd have the messenger RNA having
Exon 1 to Exon 3. How is that helpful? Why would that be a good thing?
I'm going to tell you now about a convention that I'm going to use for the next couple
of slides. So, these little exons that are shown as puzzle pieces show how the exons
are supposed to fit together to maintain the triplet code along the whole sequence. So
if the puzzle pieces fit together like this, everything is A-okay. You're, you've got a
good messenger RNA with an open reading frame for the ribosome, and you'll get your protein.
Now let's take an example where a patient has skipped Exon 45. You can see that when
Exon 44 splices to Exon 46, the puzzle pieces don't fit. You're going to get one of your
premature stop codons, you're not going to make dystrophin, and you'll get Duchenne muscular
dystrophy. Now in this same patient that skipped Exon 45, what would happen if you used antisense
oligonucleotides to tell the cell to also skip Exon 46? Now, suddenly 44 and 47 fit
together, the puzzle pieces attach. So you've preserved the reading frame for the ribosome,
you'll be missing a central chunk but you'll make a protein that still has both its head
and its tail, and therefore some function. Now, you might say, "Well, this is a nightmare.
Okay, I understand it might work, but you're going to have to come up with a different
method for every patient because everyone's going to skip a different exon. So you're
going to have to optimize this for hundreds and hundreds of different circumstances."
Well, that doesn't turn out to be the case because it's been recognized that about 60
percent of all muscular dystrophy mutations occur within this central block of exons,
from Exon 45 to 55. And someone came up with the brilliant idea of coming up with a method
of causing all of these exons to skip. So you'd use a pool of your antisense oligonucleotides,
targeting them all. Exon 44 fits together nicely with Exon 56, the puzzle works. So
this would be a potential cure for about two thirds of patients with Duchenne muscular
dystrophy, just a single cocktail of antisense oligonucleotides.
So how does this work? Well, I hear we're looking at local delivery of antisense oligonucleotides
by injecting them directly into a muscle group in the foot of Duchenne muscular dystrophy
patients. And what the investigators saw is that adding the antisense oligonucleotide
indeed causes the exons to skip. In each case, you get a smaller product. And suddenly, you
go from no dystrophin protein to lots of dystrophin protein in all of these patients. And it's
not just little tiny spots. If you look at the entire biopsy of the muscle, you can see
that there's widespread dystrophin expression. So this is looking really good. Of --
Male Speaker: How long does this take?
Hal Dietz: I'm sorry?
Male Speaker: How long does this take after treatment?
Hal Dietz: It takes somewhere between one to two weeks
to see robust expression.
Now, an important question is how would you need to deliver this to make a difference
to an individual? You can't just inject each muscle group. So instead, please have been
infusing this antisense oligonucleotides into the circulation. These are early days, but
in some patients, just putting it into the circulation causes significant improvement
in dystrophin expression. In other cases, there's perhaps a subtle effect. You know,
I think that this is still within the phase of optimization and the very important issues
are going to be to address how to best deliver these antisense oligonucleotides so they get
to all muscle groups and how do you improve the stability of these small, little antisense
oligonucleotides in the circulation? Are they simply just being degraded too quickly? But
I would say an extremely promising approach, for a very important condition.
Now I'm going to change gears and talk about a very rare, but very interesting condition
called Hutchinson-Gilford progeria, a premature aging syndrome that causes young children
to be -- to start an accelerated aging process in early childhood at around age 3 to 5 or
so. They rapidly show loss of hair, they show wrinkling of skin, atrophy of fat, osteoporosis,
coronary artery disease, and are basically dead of old age by the time they are about
15 to 20; really a devastating condition. So Francis Collins' group a few years ago
described the cause of Hutchinson-Gilford progeria, and it's due to a single point mutation
in a protein that's called lamin A. So here on top I'm showing you normal lamin A. In
patients with Hutchinson-Gilford progeria, there's an altered splicing event that causes
the skipping of a small chunk of lamin A shown in green. So the mutant lamin A protein in
these patients has both a head and a tail but it's missing a central bit. And it turns
out that this central bit that it's missing is really important. So normally what happens
is, when the cell makes this lamin A protein, that protein is modified by a process that's
called farnesylation. And this causes the protein to attach to the nuclear membrane.
But that little bit that's missing in progeria is so essential because ultimately enzymes
have to come along and cleave the protein right at that site to cause it to release
from the nuclear membrane. In patients with progeria, because they're missing that cleavage
site, this lamin A protein remains stuck to the nuclear membrane. It can't get off. And
so that led to the question, what would happen if you just prevented this farnesylation process
in the first place, using a class of drugs called farnesyl transferase inhibitors? If
it can't get on, it's not going to get stuck there, and it doesn't matter that it can't
get off. That's the logic.
So the, initially, there was a need for a marker to say are we doing something good
or not? And the marker that was selected is called nuclear blebbing. If you have lamin
A stuck to the nuclear membrane as patients with progeria have, the nucleus becomes very
misshapen. So that was initially used as the readout. So here are cells from patients with
Hutchinson-Gilford progeria in the absence of any treatment and you can see the obvious
altered contour of the nucleus. If you give a little bit of a farnesyl transferase inhibitor
that prevents the protein from getting stuck, you already see improvement. If you give a
little bit more, the improvement is striking; they look like normal nuclei. And here's that
quantified: in someone that doesn't have progeria, there's very little nuclear blebbing. In three
people that do have progeria, there's a lot of nuclear blebbing in the absence of drug.
Give a little bit, of drug it gets better; get a lot of drug, it gets even better. So
this, as a marker of progeria, this suggests that this treatment is working beautifully.
So now there has been, there have been trials of farnesyl transferase inhibitors in patients
and in mouse models with progeria. I'm going to show you the results from a mouse model
with the same type of lamin A mutation because it's much further along. So here we're looking
at both female and male mice. You can see that the patients, or mice that have progeria,
have very -- sorry, I'm having trouble seeing that. So their body weight is very low compared
to normal individuals. If you give them this farnesyl transferase inhibitor, there's a
significant increase in body weight. You can also see that mice with, that are treated
with this farnesyl transferase inhibitor, both female and male show a significant improvement
in grip strength of the muscle, that there is delay of death in mouse model of progeria,
and there's also a dramatic reduction in rib fractures.
So again, just by understanding where the mutation is, a little bit about the protein,
you can come up with a completely novel hypothesis. I mean, no one would have possibly imagined
that this class of drugs would have anything to do with treating progeria. What's really
interesting is that it's not only relevant to these children with this rare, devastating
disease. What's been shown is that as we get older, our bodies get a little bit sloppy
at how they make lamin A. And some of that lamin A looks exactly like the lamin A in
progeria. So the concept is that this accumulation over years and years and years of this mutant
progerin protein may be contributing to our aging, and that treatments under development
for this rare condition might be relevant to the broader population.
Now I'm going to end by telling you a little bit about my own work which has focused on
a condition called Marfan syndrome. So Marfan syndrome is a disorder of the body's connective
tissue: the material between the cells that give the tissues form and strength. It's a
very complex and variable condition, but the main features include dislocation of the lens
of the eye that shifts out of place, overgrowth of the bones and low fat stores, and also,
and most importantly, progressive dilatation of the root of the aorta just as it's leaving
the heart that will lead to aortic tear, rupture, and early death if left untreated.
In 1991 we were able to show that Marfan syndrome is caused by mutations in the gene that encodes
the connective tissue protein fibrillin-1. So what do we know about fibrillin-1? Well,
we knew that it aggregates outside of the cell to form these very complex structures
called microfibrils and that these microfibrils cluster around the maturing ends of an elastic
fiber during embryonic growth. So this simple spatial and temporal relationship led to the
absolute conclusion that you need a lattice of microfibrils to make an elastic fiber during
embryogenesis and that in people with Marfan syndrome this never happens; there's inadequate
elastic fibers. Game over. If you think about it, that really boded poorly for the development
of productive treatments. It suggested that children with Marfan syndrome are born with
inadequate elastic fibers and that there's nothing that you could do after birth to improve
the situation.
So I remember the day and even the moment that I walked into a patient room, saw these
exceptionally long fingers just like this, and thought to myself, "This just doesn't
make sense. Why would weakness of the tissues cause the bones to overgrow? For that matter,
why would weakness of the tissues cause the facial features of Marfan syndrome like downward
slanting eyes, flat cheekbones, and a small chin?" Each of these findings was more suggestive
of altered cellular behavior rather than simple tissue weakness. Now, to make a long story
short, we learned that microfibrils that are composed of fibrillin-1 serve a second important
function. They're not just glue, but rather they bind to the inactive complex of a growth
factor called TGF-beta. It's a molecule that tells cells how to behave. And what we learned
is that in Marfan syndrome, where you have inadequate microfibrils, you have failed matrix
sequestration of latent TGF-beta, and that leads to too much TGF-beta activation and
activity. They're over -- this molecule is now over-stimulating the cells. This sets
in motion a cascade of events inside the cell. One important event, and the only event that
I want you to notice here, is there's a molecule called phosphorylated SMAD2. That's going
to be our marker of how much TGF-beta activity is going on. And we were ultimately able to
show that this excess of TGF-beta stimulation of cells had consequences in Marfan syndrome
including emphysema, mitral valve prolapse, aortic aneurysm, and skeletal muscle myopathy.
The way that we proved that is we made Marfran mice that were deficient in fibrillin-1, still
had bad glue between the cells, and then inject them -- injected them with a TGF-beta-blocking
antibody, and we found that virtually all of these conditions were prevented by simply
blocking TGF-beta in Marfan mice.
So we then asked, "Well, is there a drug, and even better an FDA-approved drug, that
might mimic this protection?" And our attention got turned to a very common antihypertensive
medication called losartan that lowers blood pressure, something we think is good for people
with aneurism, but had also been shown to block TGF-beta in mouse models of kidney disease.
So we wondered whether this might be a magic bullet having two different effects. So we
know that angiotensin II, a molecule that regulates blood pressure, acts by working
through both a type 1 receptor and type 2 receptor, and it's the type 1 receptor that
simulates the TGF-beta pathway. So it's also known that angiotensin II, working through
its type 2 receptor, can suppress all the events that are caused by the type 1 receptor.
So in this view, AT1 is bad. That's the culprit. And AT2 might actually be protective. So we
reason that if you use an ACE inhibitor that stopped the production of angiotensin II,
you'd be limiting signaling through both the culprit and the potentially protective pathway.
But if you selectively block the AT1 cascade with a drug like losartan, you might actually
stimulate signaling through the protective pathway.
So we did a clinical trial in our mouse model of Marfan syndrome. Normal mice show very
slow rate of growth of their aortic root. Marfan mice treated with placebo showed this
very accelerated aortic growth. If we gave a drug such as propranolol that only lowered
blood pressure but did not address TGF-beta, the aortic growth was decreased to an extent.
But if we treated these mice with losartan, we saw that they showed absolutely normal
aortic growth for a full lifetime. And even more important, if you looked at the aorta
under the microscope, no observer could distinguish between the losartan-treated animal Marfan
mice from normal mice by any parameter.
So this, there's a large clinical trial that's now ongoing, but we felt compelled to treat
a subset of children with the most severe form of Marfan syndrome. These kids show unrelenting
growth of the aorta despite maximal treatment with beta blockers or ACE inhibitors reaching
death or surgical endpoints in early childhood. So here's the aortic growth curves for the
first two such children we treated with losartan, and you can see that there was a dramatic
plateau, no further growth of the aortic root once losartan was started. In this child,
this plateau has remained, now with eight years of follow-up. There still are some things
that require improvement. We found that some children show a relative response to losartan,
but then the aortic growth starts creeping up again. And what we've learned is that other
drugs in the class of angiotensin receptor blockers, such as irbesartan, with ultra-high
dosing can allow you to achieve a plateau in aortic growth even in these most severe
children.
So we again wanted to try to understand more about this pathway. I've told you about TGF-beta
activating the so-called SMAD pathway, but it's also known that TGF-beta can activate
other pathways. And the one that I want you to pay attention to is particularly this one:
the so-called ERK pathway. When we looked at the aorta of our Marfan mice, we saw that
the SMAD pathway was excessively activated compared to normal mice or what we call "wild-type
mice," but look at ERK activation. It's dramatic in the Marfan mice and almost non-existent
in the normal mice. So our attention turned to the significance of this ERK activation.
We went on to partner with the Therapeutics for Rare and Neglected Disorders Program at
the NIH and they were able to provide a compound called RDEA-119 that's a potent inhibitor
of ERK activation. When we treated our Marfan mice with this ERK inhibitor, not only did
we greatly diminish abnormal aortic growth, we actually caused regression in the size
of the aneurism over time. It got smaller over time, the first time that we've ever
seen that. So now ERK is firmly on the radar screen.
You know, having -- finding genes and making animal models allows you to find new therapies.
It also allows you to evaluate the performance of existing therapies. And currently calcium
channel blockers are considered the second-line treatment for patients with Marfan syndrome
that can't tolerate beta blocker medications. But there's really not a lot of evidence that
they work, and there's not even a lot of evidence that they're safe. So we decided to test this
on our mouse model. We were initially very optimistic because there was some work in
the literature showing that calcium channel blockers can blunt ERK activation in at least
some cell types. So we did a trial of amlodipine in our Marfan mice. Now, here I'm showing
you the heart and ascending heart of a normal or wild-type mouse. In a mouse with our Marfan
mutation we see that there is flaring at the base of the aorta. If you give a normal mouse
amlodipine you don't see much. Perhaps the aorta is looking slightly generous, but it's
subtle. Virtually 100 percent of Marfan mice given amlodipine develop these massive ascending
aortic aneurisms. And this occurs, the aortic size triples, within three weeks of starting
the medication. And the mice are dying due to aortic dissection within four weeks of
starting the medication.
If I show you this quantitatively, a placebo-treated Marfan mouse shows accelerated growth of the
aortic root. Look what happens when you add amlodipine. The rate of aortic growth, root
growth, doubles. If you look further up the aorta where Marfan mice normally don't have
an aneurism, you now see that amlodipine is causing this dramatic aortic growth. In this
Marfan mouse model shown in red, there is typically no death if you treat with placebo.
But if you treat the mice with amlodipine, death due to aortic dissection starts within
five weeks and then accelerates quickly. We are able to see that, contrary to hypothesis,
amlodipine is actually accentuating ERK activation rather than blunting ERK activation. And now
we found the identical results for all calcium channel blockers. This is not specific to
amlodipine, it's a class effect. Okay.
So we want to try to figure this out. And one thing that we did was ask, well, what
happens if you give amlodipine but also block ERK with our ERK-blocking drug? And the answer
is that you prevent all of this detrimental effect. So it is ERK activation that's responsible
and gives us greater confidence that this is a great therapeutic target. We also, in
our mice that are dying due to aortic dissection with amlodipine, if you give the ERK antagonist
you again see no death due to aortic dissection. So, not only do you suppress aortic growth,
but you prevent aortic tear. So I think it's pretty remarkable that seven years ago we
had the model that there's weak tissues in Marfan syndrome at birth and there's nothing
you can do about it. Currently we have over seven different medical treatment strategies
that have shown remarkable effectiveness in our mouse model and that we are moving forward
to people. Just as important as finding new effective drugs is learning about things that
might be detrimental, such as calcium channel blockers.
In the last two minutes I'm just going to tell you that what we've learned about Marfan
syndrome is not just relevant to that condition. It extends to many causes of aortic aneurism.
About six years ago now my colleague Bart Loeys and I recognized and described a new
aortic aneurism syndrome that has many features of Marfan syndrome like curvature of the spine,
long fingers, and aortic aneurism, but also many unique features like widely-spaced eyes,
a cleft palate, or a bifid uvula. Most importantly these patients don't just have aneurisms at
the root of the aorta but rather all through the arterial tree. So much more aggressive
condition. And these aneurisms rupture at young ages, as young as six months, and have
smaller dimensions when compared to Marfan syndrome.
So based upon what we had learned about Marfan syndrome, we bet that this condition that's
called Loeys-Dietz syndrome would also relate to TGF-beta. And in fact the very first two
genes we looked at are the two genes that encode the TGF-beta receptor, and we learned
that all of these patients have mutations in the receptor for TGF-beta. They also show
high TGF-beta signaling looking at our old friend phosphorylated SMAD2, in the nucleus
of cells in the aorta. We went on to make mouse models. We saw that these mice have
horrible-looking aortas. Their aortas grow really fast, but if we treat them with losartan,
aortic root growth returns to normal and aortic wall architecture returns to normal. We go
from this picture with all these fractured elastic fibers to a very normal-looking aorta.
Currently there is now a new class of conditions that are called the TGF-beta vasculopathies,
that are shown here, all aortic aneurism conditions that now have been associated with high TGF-beta.
So these data suggest that altered TGF-beta signaling is a common pathway to aneurism
and that treatments under development for Marfan syndrome may find broad application.
My last slide, the conclusions, are that the study of rare Mendelian disorders is both
an obligation and an opportunity. The obligation stems from the fact that while they are individually
rare, these conditions are personally burdensome and collectively common, and also that patients
with rare genetic disorders have really fueled progress in the field of molecular therapeutics.
They've given of themself, they've accepted risk, they've allowed us to learn, so there's
a real personal cost, despite a remote chance of personal advantage. The opportunity relates
to the single-gene basis of the defect. What we know when we find a gene for a Mendelian
disorder is that a defect in this pathway is sufficient to cause this condition, this
disease phenotype, and therefore it tells us that these pathways are inherently attractive
therapeutic targets. If you can nudge them in the right direction, they might make a
big difference. And such therapies can then be explored in more common conditions, like
emphysema rather than the lung disease in Marfan syndrome.
I'd like to end by acknowledging the truly remarkable young people that I have the privilege
of working with every day, my collaborators at other institutions, and my funding sources.
And I'd be happy to answer questions. Thank you for your attention.
[applause]
Yes, sir.
Male Speaker: [inaudible] Is there any evidence of a new
response then to the new protein?
Hal Dietz: So again it depends on whether someone normally
makes a little bit of dystrophin or whether they're completely na久e for dystrophin.
But so far, to my knowledge, that has not turned out to be a problem. And it seems that
everyone, even with the most severe form of DMD, is making enough protein to elicit tolerance
so the immune system doesn't react. Yes.
Male Speaker: It will take me weeks to absorb all this.
But I do treat osteoporosis, and you mentioned in progeria and Marfan that there was osteoporosis.
Is that just disuse or is it because the bone is much different from some of the other tissues?
Hal Dietz: Yeah, it is not a disuse situation. We've
been able to, at least in Marfan syndrome, tie the osteoporosis to the same TGF-beta
cascade. And our collaborator in New York, Francesco Ramirez, in culture systems, has
shown dramatic responses to some of the TGF-beta modulating agents. So I think it's going to
teach us something about common osteoporosis. Yes, sir?
Male Speaker: Are adults with Marfan all treatable?
Hal Dietz: Yeah, so that's a great question. The question
is, are adults with Marfan all treatable or is the window of opportunity to make a difference
over in childhood? At least in our mice, we can allow them to become mature adults. They're
sexually mature at about two months, by six months of age they are sort of mid-adult life,
and by a year of age they are old mice. And whether we start treatment right after birth,
in the middle of that sequence, or at the end, we see the same kind of benefits. So
we think that the window doesn't close, that there is an opportunity even later in life.
Male Speaker: Is there a test that can determine this embryonically?
Hal Dietz: Yes, so the question is, is there a test that
can determine these diagnoses as a fetus is developing? Yes, for all the conditions that
I discussed the gene is known so you would be able to do prenatal diagnosis and know
that a fetus has this predisposition.
Male Speaker: Sir, you didn't mention the so-called readthrough
drugs for in-frame premature termination codons. Do you have a comment on those?
Hal Dietz: Yes, I actually have slides for those, but
I just didn't have time to cover them. So there are drugs that are called readthrough
agents. Basically they make the ribosome sloppy. So the ribosome reaches a stop codon, but
it just marches through and inserts any amino acid at that position. There is a company
called PTC Therapeutics that has developed a drug called PTC124 that makes the ribosome
sloppy, so the whole idea is that the ribosome will march through an early stop codon and
make a full-length protein. There is a clinical trial, or was a clinical trial, for muscular
dystrophy with PTC124, but that trial was stopped because they did not reach endpoints.
There was some suggestion that there might have been a subtle benefit at a lower dose
as opposed to a higher dose.
So that's being explored, but there are many potential problems with the readthrough approach.
You know, one potential problem is that most of the RNAs are degraded by nonsense-mediated
decay. So there's not many transcripts left that are engaged by ribosomes where you could
have readthrough. Also the context of the premature termination codon defines the efficiency
of readthrough and there's only one context that's really potent that allows a lot of
readthrough to happen. We also know that inserting any amino acid where a stop codon was is often
not good enough. What's needed at that position is the intended amino acid rather than any
amino acid for the protein to have its function. So I think it's a really exciting idea. I
think it potentially could treat a broad spectrum of genetic diseases, but I think the idea
is in need of refinement. One potential refinement would be to both block this decay pathway,
this nonsense-mediated decay pathway, while also stimulating readthrough. I know that
a number of companies are looking at that closely. Yes?
Male Speaker: [inaudible]
Hal Dietz: Yeah, so the question is, what's actually
activating TGF-beta that doesn't bind to the matrix, for example in Marfan syndrome. There
are many, many activators of TGF-beta, including many proteases, including integrins, including
low pH. There is some evidence that in Marfan syndrome a molecule called the MMP9 or matrix
metallopeptidase 9, may be particularly important in activating TGF-beta. There are some trials
going on with different inhibitors of TGF-beta activators but there simply aren't results
yet to share. So it is a developed concept though to go after these activators.
Male Speaker: What do you think about the gene correction
using [inaudible] technology?
Hal Dietz: Yeah, so there are a number of very interesting
technologies that are being considered to actually correct the mutant sequence within
your gene, within the native gene. And some of them have shown promise in cell culture
systems. To my knowledge none have shown sufficient efficiency to suggest that they may work in
vivo. So I think it's again a very exciting concept. It would be curative for many genetic,
if not most genetic conditions, but I just don't think it's far enough along. There are
some still leaps in technology that will be needed to bring that to fruition.
All right, well, thank you again for your attention.
[applause]