Tata Consultancy Services said last week that it had signed Italian biotech Congenia as its first customer for a new in silico lead-optimization service built around a proprietary fragment-based drug design approach.
TCS, based in Mumbai, India, is one of Asia's largest IT firms, and employs more than 20,000 people worldwide. The company first entered the life science sector in 2001 when it began developing a bioinformatics package called BioSuite, which it launched last year [BioInform 06-21-04]. The company also sells a clinical trial management system, and offers genome-annotation services and software-development services. Now, TCS has added lead identification and optimization to its life science IT portfolio.
Sharmila Mande, head of the company's life science R&D division, said that TCS sees more commercial opportunity in lead discovery and identification than in target identification or validation. "We are not in target discovery at all," she said. "We have consciously decided that we will be in lead optimization."
Under the terms of its agreement with Congenia, the drug maker is providing the target — a protein called P66 that is involved in several age-related diseases — and the TCS life science R&D group will deliver three to five leads based on the protein in 18 months.
TCS will carry out all the computational aspects of the project, and it has sub-contracted Altiora Ventures, based in Israel, to perform all the wet lab work, which includes the initial crystallization of the P66 target protein, co-crystallization of the protein with the optimized leads, and final synthesis of the compounds. Congenia will validate the leads using cell-based assays.
"It is a tough project, but we had a long discussion about how to go about it, and we broke it up into different parts" to determine how long each step would take.
"Based on the structure, we'll be doing the fragment-based identification of lead molecules, and then we will optimize these in silico and also do in silico ADME prediction," Mande said. "Based on that, we will come up with an optimized set of leads which will again go into co-crystallization with the target molecule, which will be the structures that will be solved, and this iteration will go on."
Mande said that TCS is still developing a database of fragments that it will use for the lead-design step. So far, she said, the database includes more than 10,000 fragments.
Mande said that even though the protein has not yet been crystallized — she said the process could take "a couple of months" — the company expects to be able to stick to its 18-month timeframe. Around 10 people on the life science R&D team, out of a total of 35, will be dedicated to the project, she said.
"It is a tough project, but we had a long discussion about how to go about it, and we broke it up into different parts" to determine how long each step would take, she said.
Meeting the deadline will depend very much on the protein-crystallization phase, according to Henriëtte Willems, group leader of applied design at De Novo Pharmaceuticals, a Cambridge, UK-based firm that also provides fragment-based drug design services.
Willems, whose company is not involved in the TCS-Congenia alliance, told BioInform at the IBC Drug Discovery Technology Conference in Boston last week that protein crystallization is still "more of an art than a science" and that it "can take months or years" to complete. Once a protein structure is available, however, she said that her company can design — but not synthesize — a set of optimized leads in around three months.
The market for computational lead optimization services is still in its earliest stages, Willem said, so competition from lower-priced outsourcing shops such as TCS is not a great concern for companies like De Novo. Willem said that her firm has signed three contracts this year so far, but the business "has taken quite a bit of time to build."
Locus Pharmaceuticals of Blue Bell, Penn., has also developed a proprietary method for in silico drug design, but the company sees more opportunity for its technology in building its own compound pipeline than in offering it as a service, according to Robert Dickey, vice president of finance and CFO.
Locus builds a library of between 750 million and a billion fragments for each protein target that it studies, Dickey said. The company's software is able to sift quickly through that library to generate novel compounds with high binding affinity. Starting with the crystal structure, Locus can create a set of leads in about three months, Dickey said.
The company plans to file its first IND by the end of the year — an event that Dickey described as a milestone for the computational drug design sector. "We will be testing in humans a compound that has a strong computational origination to it," he said.
While computational approaches have long promised an efficient, cost-effective alternative to high-throughput screening and other experimental approaches, in silico drug design "has not had the impact up to now that it could have, because of the complexity of what we're doing," Dickey said
"The proof of the pudding is really going to be how things do in human testing," he said. "Once we have shown in humans what we can do with a computational approach, I think it really is going to change a lot of people's thinking about drug design and drug development.
— Bernadette Toner ([email protected])