By Aaron J. Sender
Bridging the genome-to-drug gap: Infinity, Graffinity, Ambit, Amphora, and more
I was happy, to be honest, when I heard that there are only 35,000 genes,” says Dalia Cohen, global head of functional genomics at Novartis. Drugmakers are already drowning in new potential drug targets. But despite spending billions of dollars on genomic tools such as microarray, gene-knockout, overexpression, and antisense technologies to pinpoint a new target’s role in disease, drug development has become more expensive and less efficient.
In fact, “last year was the worst year in about 20 for FDA approvals,” says Michael Foley, a 10-year pharmaceutical industry veteran. Because even after all the expense, effort, and time spent on validating new genomic targets, pharmas and biotechs are not necessarily any closer to the legal tender of their trade: a drug.
“Somebody has to do something different. And we’re going to do that,” says Foley. Founded in July, his company, Infinity Pharmaceuticals of Boston, is one of at least a half-dozen new companies promising to bridge the gap between genetic breakthrough and medical advance by turning to a new genomic tool: the small molecule.
Instead of starting with a gene, the emerging approach — alternatively called chemical genetics, chemical genomics, chemogenomics, or chemical proteomics — starts with small drug-like molecules to screen protein targets and study their role in disease. The result: a validated gene function and a potential drug candidate all in one step.
True, it’s too early to tell whether this approach will really make finding drugs more efficient. Many targets may be intractable to current libraries of chemical compounds. And even if small molecules turn out to be effective in fishing for targets, most pharmaceutical companies are set up around therapeutic areas to develop and market drugs. Introducing a non-disease-oriented approach like chemical genomics may be too disruptive to organizations to be useful.
That’s probably why the new approach is crystallizing in the form of startups such as Infinity first. But there’s little doubt chemical genomics is creeping into the conscience of well-established pharmaceutical companies. “It’s very clear that this is the next big wave in drug discovery,” says Novartis’ Cohen.
If recent activity in the market is any sign, chemical genomics will catch on quickly. NeoGenesis of Cambridge, Mass., filed in November, in the midst of a bear market, for an IPO in which it plans to raise $115 million. Morphochem of Munich has raised 70 million in private equity since its launch in 1998 and plans to go public within the next 18 months. ComGenex, an already profitable high-throughput screening company in Budapest, has opened offices in New Jersey and San Francisco and has recently set its sights on chemical genomics. Graffinity of Heidelberg, founded in 1998 with just 125,000, completed one of the biggest financing rounds in Europe last year, raising 30.6 million. Protein-chip maker Caliper has just spun off Amphora Discovery to use its high-throughput technology to build a chemical gemomics database. San Diego-based Ambit Biosciences is just coming out of stealth mode after raising $10 million. And Aventis, Novartis, Merck, Schering Plough, Biogen, Roche, and Boehringer Ingelheim are among the biopharmaceutical companies that have signed deals with some of these chemical genomic startups to screen protein targets with small molecules.
To Infinity and beyond
Foley had spent 10 years as a medicinal chemist, first at Bristol-Myers Squibb and then at Glaxo Wellcome, with only a bachelor’s degree, when in 1995 Glaxo offered to sponsor his graduate training at Harvard in the laboratory of renowned chemist Stuart Schreiber.
At 26, Schreiber, an organic chemist, was the youngest full professor of chemistry in Yale’s history before returning to his Cambridge alma mater. Today he is widely considered the father of what he calls chemical genetics.
By the time Foley arrived, Schreiber had been synthesizing molecules that mimic compounds found in nature and using them as tools to study cell biology. At first his approach was not very popular. “Everyone was equally offended by this,” says Foley. “The biologists felt like he was an interloper in their world, and the organic chemists felt like he was abandoning their noble trade.”
But Schreiber saw the bridging of the two fields as a powerful new genetic tool. Since the early 20th century geneticists have been using radiation to cause random mutations in model organisms, such as flies, and then breed them over many generations to observe the effect of the altered gene — red eyes, a missing set of wings, death — and thus determine its function. Today researchers use more advanced methods such as knocking out the gene. Regardless, the genetic techniques all have the same underlying philosophy: If you want to understand something, perturb it.
Instead of using multi-step, complicated, and organism-specific genetic techniques to perturb a cell, thought Schreiber, why not use a small molecule instead? This approach had several advantages. The molecules bound and modulated the encoded proteins, which are closer to the action, rather than the gene. There was also no need to wait for the organism to complete its life cycle to study the effect.
In 1997, appreciating the potential of this interdisciplinary approach, Harvard, Merck, and the NCI funded the creation of the Institute of Chemistry and Cell Biology at Harvard for Schreiber to pursue his work.
It was there that Foley and others helped turn the abstract idea into a systemized approach to studying biology. They synthesized a library of 2.2 million unique chemical compounds on small plastic beads — a feat named by Nature as one of the top 10 breakthroughs of 1997. They adapted high-throughput-screening instruments for panning the molecules against cells and multiple proteins. Robots select the beads, knock off the compounds, and array each into one of 96 or 384 wells on a plate. A robotic arm with a matching number of pins picks proteins or whole cells and reagents and drops them in to join the molecules. The researchers also print the compounds on glass slides, creating small molecule microarrays to probe potential protein targets.
As a second year grad student Foley, now 40, began hounding his mentor, who had already been involved in the founding of biotech companies Vertex and Ariad, to take the technology commercial. “I had a sense of what the pharmaceutical company issues were and I felt that this technology, this way of thinking about chemistry, could be applied to drug discovery,” says Foley.
It represented a revolutionary approach to drug discovery. Not all molecules are drugs. But all drugs are molecules. So instead of starting with a gene, exploring its relevance to disease, and then hoping to find a molecule to bind its encoding protein, why not start with the compound and let it do the heavy lifting of finding a disease indication?
The technology’s promise captured the imagination of some of the most prominent figures in genomics. Human genome project luminary and Whitehead Center for Genome Research director Eric Lander spent last summer in Foley and Schreiber’s labs and joined Infinity’s board. The board also includes former National Cancer Institute director Richard Klausner and Rockefeller University prez Arnold Levine. And dealmaker extraordinaire Steven Holtzman left his cushy position as chief business officer at Millennium to take on the mantle of Infinity’s CEO.
“Five years from now we will be a drug discovery company,” predicts Holtzman. “We will have molecules in clinical trials as well as a pipeline of molecules in pre-clinical.” What attracted Holtzman, a former Oxford University philosophy instructor and Rhodes scholar, was the unconventional strategy behind Infinity. “It opens up the realm of potential available drug targets,” he says. “Some people think that infinite means large. I think that infinite means without bounds. It comes from the concept of thinking without bounds.”
Chemical genomics business plans are as varied as the terms used to describe the approach. Ambit, founded in May 2000, for example, also plans on using small molecules to find proteins. But it uses a different technology — developed by David Austin’s lab at Yale (Austin did his postdoc in Schreiber’s lab) — and has a different goal: Ambit has no pretense of developing new drugs. It plans to screen drugs already on the market or well advanced in drug development that happen to work, but for which the target is yet unknown. “The idea would be to take that target and develop second-generation compounds against it,” says VP of corporate development Sanford Madigan. Or sell the info to the company already marketing the drug or to competitors interested in creating me-too drugs.
VC Kevin Kinsella, who was involved in the founding of more than 50 biotech companies including Vertex, and initially financed Ambit out of his own pocket, says he was blown away when he visited Austin’s lab. “It took him a couple of days to do what took us many months at Vertex” — finding a target for the immunosuppressant drug, FK-506.
The technology is a twist on conventional phage display, which uses viruses to display peptides used to screen against proteins. Ambit inserts an entire gene into each phage genome, which is then expressed on the phage surface as an intact folded protein or binding domain. “We have this ocean of 106 phage, theoretically representing the entire genome,” says Madigan. “We put a compound on a solid support and that’s our bait. And we fish with that to see what protein it pulls out.”
So far Ambit has raised $12 million in two rounds of financing. David Lockhart, who with Stephen Fodor invented Affymetrix’s GeneChip, is the company’s president and CSO.
Amphora, spun out of Caliper in September, takes yet another approach. Using Caliper’s high-throughput screening system and protein chips, Amphora is building a chemical genomics database. “We plan to scale up to 1,000 targets a year, which is at least 10 times what’s being done at any company now,” says co-founder and VP of operations Bill Janzen, who built the high-throughput screening facility for Eli Lilly’s Sphinx Laboratories. Amphora plans to sell access to the data on which drug-like compounds bind to which proteins, as well as to screen targets for pharmaceutical companies.
Like Janzen, Morphochem CEO Lutz Weber also left pharma because he was fed up with the old ways of finding new drugs. Weber joined the company from Hoffman-LaRoche. As a medicinal chemist he felt that it didn’t make sense that chemistry, the product of pharma, was being left out of the early drug discovery process.
“I was always kind of unhappy with the classical drug discovery process, where you have genetics at the beginning, then you have target finding, target validation, assay development, and high-throughput screening, and only then the chemist had the ability to look at the target,” he says. “The input of the chemists came just too late.”
Now at Morphochem, it’s payback time. Weber is making sure chemistry gets the upper hand. In fact, the company began as a chemistry supplier, synthesizing large libraries of diverse compounds using a technique called multi-component reaction, or MCR. Unlike conventional combinatorial chemistry, which creates new compounds by changing one functional group at a time, Morphochem varies the core skeleton of the molecule as well.
In 2000, the company decided it would get out of the compound market, a commodity business, and into chemical-genomics-based drug discovery. In March of that year it purchased Small Molecule Therapeutics of Princeton, NJ, for DM4 million for its screening assays and the following September created a subsidiary in Basel, Switzerland, to begin its own drug discovery program in infectious diseases.
Morphochem uses 1,536-well plates to screen its compounds against proteins. Software drives a robot to deposit a random selection of molecules in the wells and then collects the responses of each compound. Using this information the software directs the synthesis of a new set of compounds better suited for binding to the proteins. The cycle is repeated over and over again until the system “evolves” compounds that bind specifically to just a single target. Morphochem has two drug discovery deals with Aventis and one with Sequenom.
Perhaps the biggest chemical genomics player, however, is NeoGenesis. In the second half of 2001 the company racked up eight pharma and biotech collaborations — with Schering Plough, Aventis, Immusol, Mitsubishi Pharma, Oxford GlycoSciences, Celltech, Tularik, and Biogen — with potential for earning as much as $14.25 million for each drug candidate resulting from the screening of proteins. Its ultimate plan, though, is to become a fully integrated drug maker in its own right.
NeoGenesis has more than 10 million compounds in its library. The key to its technology is that instead of screening one molecule against one protein in each well, NeoGenesis screens a single protein with a mixture of 2,000 small molecules in one well. This is possible because each compound in a mixture has a unique mass, so that the ones that bind to the target protein can be identified with mass spectrometry.
Pharmas are not just catching the chemical genomics wave by proxy. In fact some pharmas are beginning to bring the approach in house. Pfizer, for example, as of last year was planning to use Aurora’s high-throughput screening system in its drug Discovery Technology Center in Cambridge, Mass., to screen every compound in its library against each target. And Novartis is in the midst of setting up a new CHF100 million-a-year HTS facility, dubbed the Drug Discovery Center, to deal with genomic targets. “A very significant part, about 20 percent” of the center will focus on chemical genomics, says Paul Herrling, head of Novartis global research. “It should help us find the most accessible targets and get access to targets that are today difficult to address,” he says. The facility is up and running but will not be completed until early 2004.
“We are going to supply them with genes … and they will produce the protein and look for a binder,” says Cohen.
Chemical genomics, the marriage of chemical compounds with genomic targets, just makes sense. “Because that is actually what you want to know: Can you modify the function of a protein with a chemical?” says Holger Ottleben, VP of scientific strategy at Graffinity, a chemical genomics company using small molecule arrays. “You can use other approaches like antisense technologies and knockouts to establish the function of a protein,” he says. “But you still don’t know whether you will ever be able to modify the function with a chemical. And that’s your ultimate goal. That’s what drug discovery is all about.”
“What are we trying to do here in the pharmaceutical industry anyway,” asks Kinsella, “but find a small molecule that has a therapeutic effect on a target?”
The promise of chemical genomics does not just affect target validation. It also has the potential of giving drug candidates a better chance of getting through clinical trials. Screening molecules against the proteome allows researchers to detect compounds that bind to several proteins, “so you see possible side effects early on,” says Morphochem’s Weber.
Perhaps, more than anything else, the increasing popularity of chemical genomics is a sign that genomics is growing up. It is no longer a separate discipline divorced from the rest of drug discovery and development. By linking it to chemistry, the raison d’être of the pharmaceutical industry, genomics is becoming part of the drug discovery family.
Infinity and the other startups intent on bridging the chemistry-genomic gap are hoping to pave a more efficient route to drug discovery. “These genome programs are so spectacular. The science is so incredible. And the chemistry road never rose to meet the data,” says Foley.
Spending all that money on genomics without addressing the chemistry is “like driving a Ferrari off a cliff,” he says.
“But we’re going to do something different, because everybody else has failed.”
Don’t really get the difference between chemical genetics, chemical genomics, chemogenomics, and chemical proteomics? Don’t worry; you’re not the only one. “Every person you talk to is going to give you a different answer,” says Michael Foley, VP of chemical technologies at Infinity Pharmaceuticals. Maybe it’s corporate branding, maybe it’s self-promotion in certain academic labs. Whatever the reason, says Foley, if the field is going to progress, “the world has to harmonize on what the heck it means.”
Chemical Genetics “The chemical analog of genetics,” according to Stuart Schreiber. The use of a small synthetic molecule, as opposed to, say, radiation to perturb a cell in order to elucidate the role of proteins or the role of the small molecule. Confusion factor: it’s not really genetics, just analogous to it. Some call this chemical genomics when performed systematically.
Chemical Genomics Most commonly used to refer to high-throughput screening with small molecules against proteins to do gene functional analysis and find drug candidates simultaneously. Foley uses this to mean the use of genomic tools in conjunction with chemical genetics.
Chemogenomics Vertex uses this to refer specifically to screening of molecules against target families, e.g. kinases. Iconix uses it to mean screening small molecules for toxicities against targets. To everybody else it’s the same as chemical genomics.
Chemical Proteomics Rarely used, except by those literal-minded folks who insist that the targets of chemical genomic screens are proteins, not genes.