CONVERSATION · Curt Carlson of research hothouse SRI International on technology and accelerating drug development

By Kevin Davies

December 15, 2004 | Founded as Stanford Research Institute in 1946, nonprofit SRI International is a leading independent R&D laboratory, with nearly 1,400 employees and annual revenue of $220 million in 2003. Dubbed "spin-off city" and the "heart of Silicon Valley," SRI and its subsidiary, Sarnoff Corp. (formerly RCA Laboratories), have launched more than 20 companies, including Orchid BioSciences and Locus Discovery (see "Locus Focus," Dec. 2002 Bio·IT World, page 54).

President and CEO Curt Carlson was a computer scientist at RCA and Sarnoff, leading its HDTV team. Here, Carlson gives Kevin Davies a flavor of the "quality and integrity" at SRI, innovating drug discovery and development and much more besides.

Q: How did you reach your current position?
A: At Sarnoff Corporation, I worked with colleagues to create a dozen companies — Locus Pharmaceuticals was the last one I worked on. Locus is doing quite well. Currently, they are working hard to get products into the Phase I clinical stage. Locus was a fun yet challenging company for us — the initial skepticism was enormous. Their computational technology helps rapidly and accurately identify the biologically relevant active binding site(s) of a protein, and simultaneously designs de novo novel, small-molecule antagonist or agonists of the protein's activity. This approach enables faster identification of novel small-molecule drug candidates and on a larger scale than was previously possible. Unfortunately, we couldn't tell anybody how we did it, which made it more difficult to sell the vision for the company. In spite of this, we raised $50 million in two rounds in 1999-2000, so we did find a way to show investors the value of the work. Five-and-a-half years ago, I moved to California with my family to take my current position at SRI.

What is the structure of SRI?

"For all the money being spent, the therapies being developed are not getting into the clinic fast enough." -Curt Carlson, SRI

We work in three major areas: bio-, info-, and nanotechnology. We have five operating divisions: biosciences, engineering, information and computing sciences, policy, and physical sciences. Our policy group does amazing work, which ranges from researching addiction and sleep to economic development and innovation in education.

What are your current associations with Big Pharma?
With our partners, we have an end-to-end drug discovery and development organization. SRI is focused on discovery and pre-clinical work. We have a world-class pre-clinical group with a very strong track record. We're focused on high-quality work, and we do this by hiring the best people in the industry.

Over the years, we've developed our own pipeline of drugs, which would be the envy of most midsize pharmaceutical companies. For example, SRI came up with an effective treatment for malaria. We believe many lives have been saved as a result. We also have three drugs that have been licensed, a Phase III cancer drug, and six cancer drugs in various stages of development. Our strategy is to license these drugs to pharmaceutical companies to fulfill our ultimate goal of seeing patients benefit.

You talk about 'the valley of death' at the NIH? What do you mean?
We're trying to address the problem that [NIH] director Elias Zerhouni calls 'the valley of death' — the development gap between research and market readiness or manufacturing. This is a huge problem for Big Pharma. R&D spending is going up, yet drug approvals are going down, which means higher risk. The opportunities that the VC community used to obtain at universities and license to Big Pharma are drying up. This is also a problem for researchers, who are motivated by making a difference with their work.

Has the doubling of the NIH budget over the past five years had a chance to affect drug approvals?
It's still early, but drug discovery results have been going down for the past 10 years. They're still going down in spite of the increased spending and investment. If they were starting to go up, this would be a more persuasive argument. However, that's not what we're seeing. People give lots of reasons, such as the easy cures have been found, or it's harder to come up with a new target or new mechanism and come up with a cure. There must be a more systemic approach to improve our overall ability to create new drugs.

Does academia have any reason to get involved in drug discovery?
Academia plays a critical role. The problem is that for all the money being spent, the therapies being developed are not getting into the clinic fast enough.

To address this, we've formed a consortium called PharmaSTART — a collaboration between SRI, Stanford, UCSF, Berkeley, and UCSD, the four leading medical research universities in California. The leads coming out of a university are not always well qualified. Universities don't have the infrastructure to manage the qualification process. The further you take them into the pre-clinical phase, the higher the value. So, the thinking was to start a program at universities that helps sort through the leads they have, and create development plans. VCs will see something that has an action plan behind it, and we can assist researchers in obtaining NIH funding along the way. The goal is to create more value.

We'll do the first level of pre-clinical testing — toxicology, efficacy, find out if you can manufacture the substance, etc. ... For a modest amount of upfront effort, you can make a more informed decision on which [drugs] you should focus on and in what sequence. We've already had one funded drug lead go through this process. We're moderately funded right now.

What about the deployment of IT and in silico technologies?
We're doing a lot of work in this area. First, we're modeling pathways. SRI's Peter Karp, a pioneer in metabolic pathway modeling, has mapped at least 13 organisms so far. Karp is now working on HumanCyc — [an encyclopedia of] human metabolic pathways — so you'll be able to go from genome to pathway. He has another database, which is a compilation of all organisms, called MetaCyc, which is so large it reminds me of the Oxford English Dictionary. There are a lot of people working on signaling pathways, and they all have to be put together.

We're also working on a major program called BioSPICE — workstation software that will allow you to import these databases in an environment where you can manage complex computation, and export the results in a friendly way. It's an open platform. We are defining the APIs (application programming interfaces) so people can plug into this in a compatible way. The 'SPICE' comes from the IC community — it's the way they compute and design integrated circuits. The analogy is that we need an equivalent kind of workstation to analyze biological pathways to find the targets of opportunity.

The last piece is, once you've mapped these pathways you need to figure out what they mean. We've developed a powerful tool based on a form of mathematics called formal methods, which allows you to look at these pathways and ask questions about what's possible — for example, is it possible this pathway is an oscillator or a mux, or some other function? You still need to do the biology to prove it's an oscillator, but it tells you whether it's possible to compute that function. It's an example of the kind of tools that are needed if we're going to rationally understand if a SNP results in a different kind of pathway (signaling or metabolic). We see the functional behavior of that particular element and know where to go to target and fix that particular problem.

SRI is looking at several new biocomputational ventures. These techniques represent essential technology, but we have a great deal of work to do. For example, we don't have all the data we need, and we don't have a platform where we can collect it and synthesize it. These tools still need to be proven. The initial results are remarkable, but we still need to go back and do the biology, and that takes time.

What about some future directions?
SRI's Information and Computing Sciences group is well known. This is the group that invented the foundations of personal computing, from the computer mouse to hypertext, etc. ... Today, this group is working on next-generation mobile devices. Imagine a device that you can talk to, and that has natural language ability so it can truly make sense of what you're saying. Even more importantly, it has real intelligence and can navigate for you and learn about you.

We also have many projects in bio-detection and homeland security, such as a handheld pathogen detector. Last year, we were designated as the administrator for the Cyber Security R&D Center for the Department of Homeland Security. It is quite an honor to be given this very important responsibility. We believe that it is imperative to have security without compromising privacy and trust. Most people think it's an either/or — we don't. You must have both if you want a truly effective security infrastructure.

Finally, we are exploring with the University of Arizona and the FDA a new program to shorten the drug development cycle by combining research, education, and collaboration with the FDA to improve processes. That's a big deal. As a nonprofit organization, SRI's drug discovery and development organization is well positioned to team with Big Pharma, along with our partners, to make such an endeavor successful.