Astex
Technology was launched in December 1999 in Cambridge, UK, to develop a
new approach to rational drug design, using high-throughput X-ray
crystallography. It started with seed financing of
800,000
($1.2m) provided by venture capitalists Abingworth Management and Oxford
Bioscience Partners.
Its founders, Professors Sir Tom Blundell and Chris Abell of the
University of Cambridge and Harren Jhoti, who had recently left
GlaxoWellcome, brought such a unique combination of scientific and
industrial experience that it was relatively easy for them to obtain that
initial exploratory funding.
"I think that clearly we were in the right spot, the time was
right, and, clearly, at the end of the day, the venture capitalists invest
primarily in people," explained Astex Chief Scientific Officer Harren
Jhoti. "Given that I brought the industrial experience, obviously
Prof. Blundell is an internationally renowned structural biologist, and
Prof. Abell is a very talented academic scientist - it's a good mix."
In the 1990s, Jhoti had come to recognize that research might be
performed more productively in a smaller outfit rather than in a big
company. "I'm not sure that research is scaleable to provide what
research scientists really need to keep that fire in their bellies, in the
same sense as marketing forces are scaleable. It's good to have a huge
marketing force," he said. "In fact, the data indicate that
there is a critical size, probably between 150 and 250 research
scientists, for drug discovery. That's the size with which you can
generate some very interesting drug leads."
In addition, Jhoti wanted to develop new technologies and to tackle
drug discovery in a slightly different way so as to reduce the attrition
that the drug industry has been experiencing, whereby some 90% of leads
failing to make it into the clinic. But, perhaps the most significant
factor was the sheer quality and the scientific pre-eminence of his fellow
founders. "The opportunity came about to work with Prof. Tom Blundell
and Prof. Chris Abell, and I don't think that you get too many of these
opportunities," he said. "The timing was right and the people
were right, so I decided to give it a go."
Not a University Spin-out
Although Astex carried out early scientific experiments in laboratories at
the University of Cambridge, it is not a university spin-out, he
maintained. "A company is usually spun out because some technology or
a breakthrough has been made in a university, and clearly the university
looks to generate value from that," he said. "This was a very
different scenario, where there was no technology that was taken from the
university. The concept, the ideas and the experiments were actually
devised after I left GlaxoWellcome and initially started the company. So
we are not a spin-out, although we have very strong links with the
university, and have access to a variety of different research programs
from Prof. Blundell's and Prof. Abell's labs."
The seed money that Abingworth and Oxford Bioscience provided was
enough to get Astex going and to arrange first-round financing. Then, in
2001, Advent International, Alta Partners and GIMV joined Abingworth and
Oxford Bioscience to provide
28.4m
($40.6m). Since then, Astex has gone from strength to strength. Currently,
it employs 92 people and is due to move into new laboratories comprising
36,000 square feet on the Cambridge Science Park in February 2003.
What's Different?
Astex is using X-ray crystallography in a number of ways: to analyze the
structure of biological targets; to select fragments of drug-like
molecules that are compatible with the shapes of active sites; and, to
screen these fragments to select initial "hits", which are
subsequently transformed into high-quality lead compounds using expertise
in computational and medicinal chemistry.
Astex combines a range of technologies in its Pyramid integrated
approach to producing hits, including:
- HTX - high-throughput X-ray crystallography - which automates
all stages of the gene-to-structure process: protein production (with
re-folding technology that increases the success rate);
crystallization; and, structure determination (with rapid X-ray data
collection using robotics, use of synchrotron radiation, and access to
the most advanced software from Global Phasing);
- GOLD - an Astex version of Cambridge Crystallographic Data
Centre (CCDC)'s GOLD program for predicting how flexible molecules
will bind to proteins - which uses the 3-dimensional structure of the
target protein to select fragments of drug-like compounds (the
so-called MolBlox compounds), which are then experimentally complexed
with the target protein; and,
- AutoSolve - software for the rapid generation of
3-dimensional views of protein/ligand complexes by automated
processing of X-ray data - allows Astex to bypass what has been a key
bottleneck in lead drug discovery.
Back to the 1990s
To explain how Astex's lead in structure-based lead discovery, Jhoti
referred back to the 1990s. "When we sat in big pharma companies, we
were all in awe of combinatorial chemistry, massive parallel synthesis of
compounds, and this belief that the industry could make enough compounds,
and that with ultra-high-throughput screening we would find more and more
drugs" he said. "But, in the mid-90s, it became apparent, at
least to several of us, that it wasn't the magical solution, that the data
wasn't actually supporting it."
There were problems with combichem because the chemistries that could
be used for it were not necessarily those required, and because the
average molecular weights of the resulting compounds in corporate
collections were increasing, so the compounds were getting larger and
larger.
"What you were actually finding was that, when you did a
high-throughput screen against the target, there were, maybe, micromolar
affinities against the target," explained Jhoti. "The molecular
weight of the compounds was already 400-500, and, as we know, 500 is
approaching the upper end of a suitable drug candidate."
The challenge for the chemists was to strip back the bits that were not
important in the drug molecule and build in the bits that were needed, to
continue to hone the potency, the selectivity and the ADME/Tox
(Absorption, Distribution, Metabolism and Excretion/Toxicity) profile of
the drug. "People were struggling with this concept," said Jhoti.
" So we said, OK, if we stop looking at large molecules and look at
molecular fragments, take drugs that are 400-500 molecular weight and
split them up into fragments that are maybe 150 - very small entities -
then, if we can find some that look like interesting leads, we can add on
just what we want."
Building on Weak-binding Fragments
Another problem arose because chemists are so used to working with potency
that they rarely work with small, weak-binding fragments. However,
according to Jhoti, as long as you know what you can do to that fragment
to build it up, you should not be worrying about the potency. "You
need to have a technology that allows you to image the binding of that
fragment - the fundamental sciences are X-ray crystallography and NMR
(nuclear magnetic resonance)," he said.
"The whole concept of Astex's approach is to use high-throughput
X-ray analysis to look at the binding of these fragments into the active
sites of target proteins," he said. "The structural information
is useful to the chemists who then look to grow fragments in a particular
way, keeping the molecular weight of the compound down, to help it
downstream, in terms of attrition.
"It is also another way of getting far more novelty, because these
types of scaffold have actually been found by large pharma. It's a
fragment-based, lead-discovery approach. And, it's fundamentally different
from what you'd find if you were to walk into a typical large pharma
company. They know of this approach, but it's not generally in their
mainstream," said Jhoti.
Using its Pyramid approach, Astex has made significant progress in its
internal lead-discovery programs. It has identified novel lead compounds
against a range of kinases in the therapeutic areas of cancer and
inflammation, and these are now in the pre-clinical, lead-optimization
phase. It has also validated phosphatases and proteases implicated in
human disease.
Linking into Big Pharma
Astex can further differentiate itself from other drug-discovery start-ups
by pointing to its involvement with big pharma from the outset. Its
credibility with potential partners in big pharma derives from the
industry experience of many key staff, including Jhoti, who led the
Structural Biology and Bioinformatics groups at GlaxoWellcome in the
1990s, and Blundell, who advised pharma majors for 20 years.
"I think that it's been very important that many of us came from
that side. We have credibility. People know who we are, so, when we go
back and tell them what we're finding, there is a genuine trust. These are
the same guys that we used to work with," said Jhoti. Astex benefits
from its links into big pharma because the feedback that the company gets
helps it to hone its own technology to address the issues that big pharma
is interested in, such as ADME/Tox profiling.
Dealing with Attrition
As well as making excellent progress with its fragment-based,
lead-discovery approach, Astex has arguably made an even more significant
advance in dealing with the very high attrition rate of lead compounds.
Often such attrition is caused by the metabolic and toxicity problems
arising from their interactions with human cytochrome P450 enzymes.
"What we figured was that, if we could determine the crystal
structure of these human cytochrome enzymes, and try to determine the
principles of their recognition of small molecules, it might give us some
idea of how to design chemistries that were not affected too much by these
enzymes," explained Jhoti. "This was quite a new concept, a new
idea. Many people have been trying to solve these structures for years.
But we put a team together here and, at the end of last year (2001), we
announced that we'd solved the first crystal structure. More recently, we
solved the second, and also explained which isoforms of P450 we had solved
- 2C9 and 3A4."
By itself, the achievement of Astex's scientists in solving the crystal
structures of human cytochrome P450s is a significant scientific
breakthrough, because no one has been able to do it before. But, because
drug metabolism and attrition is so important to big pharma, there is also
huge interest in this area. Isoform 3A4 particularly is regarded as the
Holy Grail in terms of seeking an understanding of form, as it is said to
affect half of all known drugs. "Quite often, when you have drugs
that have been withdrawn from the market, their adverse reactions are due
to these P450-related issues," he said.
Collaborations with Pharma Majors
Astex's pre-eminence in structural biology research is reflected in
collaborations with three pharma majors - AstraZeneca, Aventis and
Mitsubishi. "They recognize that we clearly are leading the field, so
they're working with us," claimed Jhoti. "They're giving us
compounds which they can't optimize and we're taking these compounds and
showing them how they can fit into the P450s, giving them an insight that
they didn't have before. That would be the basis by which they could
devise new chemistries to overcome these problems."
AstraZeneca was the first to sign a structural biology research
agreement, which Astex announced in May 2001, and it was focused on
solving human cytochrome P450 crystal structures, which were to be used to
optimize AstraZeneca compounds in order to improve its success rate in
drug development. After solving the crystal structure of CYP2C9 in
December 2001, Astex announced in March 2002 that it had received the
first milestone payment from AstraZeneca.
Also in March 2002, Astex announced that Aventis Pharmaceuticals had
signed the second such agreement, again focused on solving human
cytochrome P450 crystal structures, which were to be used to optimize
Aventis compounds in order to provide greater potential for their clinical
success. In September 2002, Astex announced that it had received from
Aventis the first milestone payment, which was related to the
determination of the 3-dimensional structure of CYP2C9.
In June 2002, Astex had announced a structural biology agreement with
Mitsubishi Pharma, under the terms of which Astex was to apply its
proprietary X-ray crystallography technologies to obtain crystal
structures of cytochrome P450s complexed to Mitsubishi compounds. Then, in
October 2002, Astex announced a drug-discovery collaboration with
Mitsubishi in the area of metabolic disorders. Using Astex's HTX and
Pyramid approach, this collaboration aims at the rapid identification of
novel drug candidates against a key protein target, and is then to be
extended to other targets.
"We are delighted that Mitsubishi has once again chosen to work
with Astex, subsequent to our first collaboration on cytochrome
P450s," said Astex Chief Executive Timothy Haines. "This
collaboration will initially focus on developing quality lead compounds
for an important Mitsubishi project, before expanding to additional
targets. As such, it provides further validation of Astex's significant
capabilities in structure-based drug discovery."
The Best Is Yet To Come
In many ways, the acid test of the strategy of a drug-discovery company
comes when its first quality lead compounds pass through preclinical
testing and go into the clinic. Astex still has some way to go, although
Jhoti said that he was certain that it would have a few candidates in
preclinical testing by next year (2003). He indicated that, depending on
progress, it would be beyond expectations if any were in the clinic by
2004.
Here, Jhoti displayed his big pharma experience, saying: "I'm very
nervous about trying to push hard to get a compound in the clinic so as to
have a compound in Phase I. I'd rather have a much more robust portfolio
of several compounds that we move forward, perhaps more slowly, but then
you have several shots on goal. The whole biotech model - get one
compound, get it in the clinic and then push forward - is such a fragile
model because, if it falls, what do you have left?"
He is also realistic about the taking Astex's own quality lead
compounds to market. "The business model we have is a hybrid one,
where we will be looking to partner very early on some of our chemistries
for big pharma to take forward, as we recognize that they have immense
power and expertise in that area," he said.
"But, as the company evolves and grows bigger, we may take our own
drugs to the market. If we truly believe that we have a discovery approach
that is coming up with novel chemistries, novel drugs, our investors will
rightly ask us: 'Are you sure that you want to be letting them go at this
stage?' There is this balance to be struck, whereby we take forward our
own programs as well," he said. "But we're very pragmatic and
realistic about where our strength lies as a company of just over 90
people. We'll grow to 130 or 140 in the next 12-18 months, but we're still
pretty small."
Contact Details
Astex Technology
250 Cambridge Science Park
Milton Road
Cambridge
United Kingdom
Phone: +44 (0)1223 226200
Fax: +44 (0)1223 226201
URL: www.astex-technology.com
