was barely a blip on the horizon when you started your research in the
1980s. What was it that motivated you?
I simply wanted to better understand how blood vessels grow. This
field was launched by Judah Folkman and his colleagues in the early
1970s, but what had been learned was mostly descriptive. People were
interested in the cellular mechanisms of the process, but very
little was known about the molecules involved. At that time, VEGF,
which stands for vascular endothelial growth factor, was one of the
most specific growth factors affecting the cells that actually line
the blood vessels, so that was a very promising candidate for being
a major player in angiogenesis. There was no really formal proof in
vivo, however, that this was an important molecule. One
way to achieve that at the time, and still today, was to make a
knockout mouse for VEGF.
VEGF always your primary target?
“Now it has become clear that VEGF is the major player in angiogenesis.”
Yes. When Napoleon Ferrara and my group made knockouts of VEGF
those were actually the first knockouts ever made in the
angiogenesis field. For that matter, the field of knockout mice was
very young, too. The first one made was in 1989. I happened to be at
the Whitehead Institute at the time, where there was a lot of
interest in this knockout technology. It wasn’t that easy at the
time. And for the VEGF knockout in particular, there was something
very funny and strange about it.
Well, there are two copies of every gene—two alleles. And if
you knock out one allele, you have one inactive allele; usually the
mouse is sterile. In mice missing both alleles, that’s when you
usually see the interesting phenotypes, some of which are lethal.
With VEGF, even the mice lacking a single allele died very early on.
That was unprecedented. And that made it a real challenge.
Originally we thought it wasn’t working. Napoleon Ferrara thought
did you get around the problem?
We realized that the embryos in these heterozygously deficient
mice were actually showing the phenotype expected. We characterized
the vascular defect. We also used special technology, pioneered by
Andre Nagy in Toronto, to actually generate homozygously-deficient
embryos in a single step. Usually you breed heterozygously deficient
embryos, and then cross-breed those. But since we couldn’t do
that, this was the only way to obtain mice that were homozygously
deficient for VEGF. They showed an even worse phenotype. Basically
that study was something of a landmark seminal study because it
showed that VEGF was really very critical for embryonic vascular
development. That got a lot of interest from the entire field. That
was the first sign. Later on, people started studying the role of
VEGF in pathological conditions. Now it has become clear that VEGF
is the major player in angiogenesis.
so the 1996 Nature paper ("Abnormal blood vessel
development and lethality in embryos lacking a single VEGF
allele," Nature 380: 435-9, 4 April 1996), which is
your most-cited paper, was reporting on the knockouts?
Yes. And it was published back-to-back with the similar study by
seems that that one knockout project took you years; why is that?
The VEGF gene was cloned in 1989. I started working on it in 1991
and then it was submitted for publication in December 1995. So it
was quite a battle. And the gene was very difficult to clone
initially, as well.
was the response to the paper on publication? Was it immediately
obvious to others how significant it was?
Yes, because the phenotype was so spectacular. So people
immediately started to believe that VEGF was an important molecule.
This was really making a true difference. It was an important
player. We showed also that blood vessels are the first organs
formed in the embryo. With the VEGF knockout mice, we realized these
embryos died very early on, which demonstrated that angiogenesis was
very critical for the embryo. So this was a breakthrough on several
you surprised by what you found?
Absolutely, everyone was. Especially this fact, this first
example, of what is called a haplo-insufficient phenotype—meaning
the lack of a single allele already results in such a severe
has the field changed in the eight years since your paper was
As I said, this was a very descriptive field through the early
1990s, and what was lacking was the identification of which
molecules were involved. This has now changed dramatically. Now we
know a lot of the molecules and whether they’re important or not.
And so now we also know how to inhibit these molecules—using what
are called rational inhibitors such as antagonists and antibodies—and
we now have data from three clinical trials showing that anti-VEGF
antibodies, for instance, are actually very effective in fighting
cancer. They’re now approved by the FDA in the US. So the field is
still young, but it has evolved with enormous rapidity, and we’ve
gone from a point where we knew almost nothing about the molecules
involved to having developed drugs that fight cancer.
the bandwagon that formed on VEGF affect the way you pursued your
At that time, and it’s still the case, the science was so
exciting, so young, and there were so many obvious questions to ask
that the entire field was very social and friendly. Very collegial.
This is quite different from fields that have been long established
and it actually gets more and more difficult to do top science. It
becomes more competitive, and then the industry gets involved, and
people don’t talk about their results and so on and so forth. But
there’s been none of that in this field. There’s always been
plenty of new science to step up and grab.
you still working mostly on VEGF?
We’re still doing quite a bit, but we actually have also
identified a novel function for VEGF. It’s not just about blood
vessels now. By making more subtle genetic deviations in the VEGF
genes, we actually made a mouse model of the motor neuron
degenerative disease, amyotrophic lateral sclerosis, or ALS, which
in the States, of course, is known as Lou Gehrig’s disease. It’s
a very dramatic incurable paralysis that’s caused by a slow
degeneration of the neurons that innervate the muscle. By chance we
found out that VEGF is also playing an important role in this
disease. These mice with a genetic mutation in the VEGF gene
actually develop all the symptoms and signs of ALS. We just
published another paper in Nature where we used VEGF gene
therapy for treatment of ALS in this mouse model. We’re still
evaluating whether we can use the VEGF protein for treatment of the
you think you can reverse the course of the disease?
The disease is so dramatic that no single treatment has shown
that you can reverse it. What you can do, at least in rat and mouse
models, is slow it down. You delay the disease onset and prolong
survival. It seems VEGF is not only having an effect on the blood
vessels directly but also on the neurons, particularly large motor
there other VEGF-like molecules that you’re working on?
We have been working for the last couple of years on better
understanding the role of the other members of the VEGF family. In
particular, what’s called PLGF, which stands for placental growth
factor. That molecule differs from VEGF in many different aspects,
and it also seems to be a player in angiogenesis, but only in cancer
and inflammation—not in the embryo. That makes it an attractive
molecule, because if you block it with an antibody you are not going
to affect the normal vasculature. This is different than anti-VEGF
antibodies, where they also affect normal vessels. In the long run,
those sorts of side effects will become a concern for
anti-angiogenesis gene treatments. So we need additional molecules
besides VEGF, and that’s why we’re exploring PLGF inhibition,
and why this will be part of our research for the coming years. We’re
also further exploring whether we can use PLGF—either gene
delivery or protein treatment—to stimulate vascularization in
tissues. That is still a major line of research. And now quite
recently, we’re also focusing on understanding how vessels are
guided to their target. You know nerves have to face enormous
challenges in finding their way from the brain and the spinal cord
to their final targets, and this new information about VEGF and
neurons tells us that there may be similar and common mechanisms and
molecules and guiding tools used by blood vessels, axons, and nerve
endings. That’s a very exciting development. So we are studying
that in much more detail.
Peter Carmeliet, M.D., Ph.D.
Center for Transgene Technology & Gene Therapy
Flanders Interuniversity Institute for Biotechnology
University of Leuven, Campus Gasthuisberg
most-cited paper with
1,078 cites to date:
Carmeliet P. et al., "Abnormal
blood vessel development and lethality in embryos lacking a
single VEGF allele," (Nature 380: 435-9, 4
April 1996). 1,078 cites.
Essential Science Indicators