TO PLAY TO the new management's strengths, Love wanted Hyseq to move a drug into clinical development quickly. Since none of the company's internal candidates was ready, he sought a licensing deal but was limited financially. Still, he managed to strike a 50-50 partnership deal with Amgen in January for a novel thrombolytic agent, alfimeprase, now in Phase I trials.

Whereas the advantages of partnership was clear for Hyseq, Love says that "the challenge became justifying their giving us half the drug and us not spending a dime." Amgen received warrants in Hyseq while Hyseq is paying future development costs to catch up with Amgen's investment to date.

Hyseq restructured in May to reduce costs and shift focus. The company and BASF Plant Science agreed to accelerate the completion of a collaboration, set up in 1999, to discover and develop agriculturally based genes. Hyseq is cutting about 38% of its workforce, or 79 jobs, mostly related to the BASF collaboration, by the end of this year.

Love also has been working to get a return on noncore assets. Last October, Hyseq moved its DNA chip and sequencing operations into a subsidiary called Callida Genomics. "There really was no synergy between the biopharmaceutical business and Callida's business," Love explains. And if the subsidiary succeeds with products on its own, Hyseq will have the opportunity in the future to sell its interest and raise money to support drug development.

Just recently, it sold Celera Diagnostics a license to a DNA sample and a clinical database related to cardiovascular disease. Built in collaboration with the University of California, San Francisco, the database was another noncore asset "basically sitting in our freezers," Love says. Since the deal is nonexclusive, Hyseq can replicate it with other licensees.

Hyseq anticipates having a dominant intellectual-property position based on a large number of filed patent applications on gene sequences. The company is working with Deltagen and Kirin Brewery's pharmaceutical division to conduct studies of gene function using transgenic mice. Deltagen produces "knock-out" mice, in which the function of the gene of interest is deleted, while Kirin produces "knock-in" mice with increased gene activity.

"We're going right into animal-based, not cell-based, studies to understand what secreted proteins are encoded by certain genes," Love says. Although the expected number of hits is low, the information on those that work is extraordinarily strong, he explains. The company intends to explore dozens of targets and identify at least two clinical candidates this year and have many more for internal development or partnering in the future.

Deltagen has had its own drug development program that was part of an initial plan to move from a technology-based to a drug discovery-based model. "We believed the way to actually translate the genome was by understanding the functions of the genes in vivo and thus understand therapeutic value," says William Matthews, Deltagen chairman and CEO. He and cofounder and chief scientific officer Mark W. Moore worked at Genentech before creating Deltagen in 1997.

Internally, Deltagen has focused on secreted proteins as drugs and has an alliance with Eli Lilly. The company also has the capability, gained largely through acquisitions, to generate small-molecule candidates for sale or licensing, and it offers tools and services for drug discovery. Glaxo, Pfizer, Merck, Schering-Plough, and Vertex Pharmaceuticals subscribe to its Deltabase gene function database, while smaller companies can access single targets through its pilot-phase DeltaOne program.

Just two weeks ago, Deltagen announced a cost savings and realignment plan. "I don't think anybody is willing to predict when the capital markets may again reappear. So cash management is very important to us," Matthews says. "We believe that our secreted protein work will provide long-term upside potential, and we seek to balance that with our small-molecule work, intending that to be a revenue generator for us."

Staff will be reduced by nearly 30%, or 130 employees, in the hope of lowering its cash burn rate by one-third from a current level of $85 million to $90 million. The company had just $49 million in cash as of Sept. 30, down from about $77 million at the end of June, owing to operating costs and outlays for new laboratories. It anticipates having $17 million in revenues and receiving $25 million in cash from customers this year.

Competitor Lexicon Genetics, a leading developer of gene knock-out technology, is on a more stable base with $135 million in the bank and $17 million in revenues for the first half of 2002. The company has drug discovery and functional-genomics collaborations with 12 major biotech and drug firms; custom-targeted gene knock-out deals with about 30 companies and universities; and about 50 corporate and academic subscribers to its gene-analysis and knock-out mice clones database, OmniBank.

Lexicon always believed its mammalian functional-genomics technology platform would feed directly into a drug discovery pipeline, and the seven-year-old company has gone through several steps to get there.

It spent its first few years scaling up its technology, explains Arthur T. Sands, president and CEO, which involved "knocking out the entire mammalian genome and then filtering through it to select the very best targets." The next phase has been the physiologic or phenotypic analysis of targets.

"WE CAN COVER 1,000 genes per year through our complete physiologic profile, which is essentially putting a gene through a hospital-like workup to determine the medical utility of the target," Sands explains. The goal of Lexicon's Genome 5000 program is to look at the in vivo function of 5,000 potential drug targets and therapeutic proteins over five years.

This second phase has already led to the discovery of 14 validated targets in the past 12 months. "These targets have significant market potential and applications in major areas of medicine," Sands says. The company will advance the targets--in oncology, cardiovascular disease, metabolism, neurology, and immunology--into drug discovery and development programs both internally and with collaborators.

To expand its capabilities, a year ago Lexicon acquired Coelacanth, a company founded in 1996 by Nobel Prize winner K. Barry Sharpless. The acquisition, which cost Lexicon only $32 million in stock, added experienced pharmaceutical industry managers and medicinal chemistry.

"We've incorporated that platform into our own drug discovery pipeline to develop small-molecule compounds addressing these 14 targets," Sands says. Although Coelacanth had been licensing its compound libraries to about 18 major drug and biotech firms, Lexicon is reviewing that business strategy.

Sands believes in vivo validation of targets and potential therapeutic proteins offers substantial advantages over competing approaches. Those beginning with gene sequences or bioinformatics information may "simply be too far away to provide a bridge all the way to drug discovery," he says.

"Drugs act by altering mammalian physiology, and they must act in the context of all of mammalian physiology running at the same time," Sands explains. "When we knock out a gene, we can see what the antagonism of that target does in the context of all of mammalian physiology, and that link--merging a gene sequence with its physiologic and medical applications--is critical for making decisions."

The sequencing of the human genome has put the same wealth of information into everyone's hands, and the race is on to find valuable targets and create drugs.

BY MOST ESTIMATES, the genome ended up containing only 30 to 40% of the number of genes anticipated. However, these 30,000 to 40,000 genes are expected to yield between 5,000 and 10,000 "druggable" genes, company executives say, and may encode for 10 times as many proteins. This still leaves a great deal of information to decipher and turn into drugs.

"As a person who was on the outside at the time, the excitement over the human genome looked a lot like hype and frenzy," Celera's Block says. "When bubbles burst, it can be painful, but, in fact, there is usually some substance there.

"And there is substance to the genome, but it needs to be mined, molded into knowledge, and from knowledge into drugs," he continues. "That will occur, but it's not a short process."

Beyond all other expectations, genomics' key contribution to drug discovery may be in providing validated targets for new drugs (see page 47). In the past few years alone, genomics firms have announced a few dozen disease targets. Eventually, most expect to have more targets than they can pursue, and they envision using licensing and partnership opportunities to tap into the resources of larger companies and bring in revenues as they try to build their own businesses.

Only 2 to 3% of druggable genes will offer extremely high value targets for drug discovery, Sands says. Still, that would provide at least another 100 to 150 novel targets, which will more than double the number against which the pharmaceutical industry currently has commercialized drugs. All commercially marketed drugs address only some 122 targets, Sands points out, while the top 100 selling products address only about 45.

"The human genome contains all the potential drug targets for all time," Sands says. "The whole key is finding which genes within the genome actually hold medical value for drug discovery. And if one can answer that question effectively, in a systematic, unbiased way, then the breakthroughs from the genome will be tremendous."

Chemical & Engineering News
Copyright © 2002 American Chemical Society