By Debbie Strickland

NEW YORK — Millennium Pharmaceuticals Inc. is launching a $100 million predictive medicine spin-off that could bring pharmacogenomics technology to the marketplace in as little as two years, according to the head of the new venture.

"We think that probably from the time we have a [pharmaceutical partnering] deal, it will take about two years to actually define and put some commercial format to that type of test," said Kenneth Conway, president of the newly incorporated Millennium Predictive Medicine Inc. (MPMx). "The molecular technology is there today, and maybe even more importantly, the informatics technology is there," said Conway. "We have the horsepower to look at tens of thousands of genes, in hundreds of patients."

Pharmacogenomics — the prescribing of a drug using a genetic or related protein expression analysis to help determine who will benefit and who won't — is one of two broad market opportunities for MPMx, Conway told investors at the 14th Annual BancAmerica Robertson Stephens Medical Conference, in New York.

The other broad area is a diagnostics system called Diagnomics: disease-state analysis using genomic and/or proteomic techniques. Diagnomics may be able to suggest the best course of treatment for a disease — surgery, drugs, other intervention or no action — by looking at genes, expressed proteins or antibodies.

For Millennium, creating a new company instead of a division offers greater financing flexibility along with a heightened ability to attract higher-level entrepreneurial managers, said Conway, himself a former senior vice president and general manager of immunodiagnostics at Emeryville, Calif.-based Chiron Corp.

In a similar fashion, Millennium in May spun off Millennium BioTherapeutics Inc., which is using its parent company's genomics and bioinformatics platform to develop biotechnology therapeutics, such as proteins, antibodies, and gene and antisense therapies. (See BioWorld Today, May 30, 1997, p. 1.)

Initial seed capital for MPMx will come from Millennium, which will house the new firm at the company's facilities in Cambridge, Mass. Millennium has transferred all rights to pharmacogenomic and diagnostic applications of its technology. Conway expects to finance the company based on a valuation of at least $100 million. Financing could come from Millennium, outside investors or a combination, with Millennium planning to retain a majority ownership stake.

Diseases likely to be targeted are cancer, cardiovascular diseases, and central nervous system disorders, especially depression. More broadly, the company will address, in order of importance, life-threatening diseases, diseases with a high degree of public awareness (such as breast, cervical and prostate cancer), and diseases that are very expensive to treat. The goal in the last category would be to reduce costs.

In the area of pharmacogenomics, MPMx's business plan calls for partnerships with pharmaceutical companies that will fund research and development, pay access fees and provide royalties.

MPMx may enter the drug development process at one of three levels:

* Pre-approval. Pharmacogenomics could aid drug development at the clinical-trial stage by targeting patient populations who will most benefit.

* Post-approval. The process could reposition "disappointing" approved drugs, again targeting patients most likely to improve.

* Drug rescue. The company hopes to revive drug candidates that failed in clinical trials by weeding out patients prone to side effects or non-response.

In the case of drug rescue, MPMx may seek to acquire or share ownership rights, fund clinical trials and pursue regulatory approvals.

Life-Threatening Disease Will Be A Priority

As for indications, life-threatening diseases such as cancer will be a major thrust, in part because such diseases offer a chance to verify effectiveness in the short term and bring a product to market quickly. Chemoprediction — improved targeting of available chemotherapies — is a likely project.

The company will use high-throughput technology to determine phenotypic differences and corresponding variations in gene expression levels.

"We're looking for the profile," Conway said. "What's the profile of a person with a positive response to a drug and the profile of a person with a negative response? We're looking for underlying biological variations of individuals."

Diagnomics will work similarly, but with the focus more broadly on diseases rather than specific drugs. The system will differ from existing diagnostic techniques by taking doctors and patients to a "decision point," Conway said.

For example, prostate cancer is typically diagnosed using a test for prostate-specific antigen, with a reading above four indicating a problem.

"From some work we have done with Lilly," said Conway, "[we learned] there are 100 genes implicated in prostate cancer, and if we can find the 20 that are most important — the ones that would allow us to distinguish between BPH (benign prostatic hyperplasia), very aggressive prostate cancer, and non-aggressive prostate cancer — patients could take different courses of treatment, depending also on who they are, how old they are and other factors."

Millennium's stock (NASDAQ:MLNM) closed Wednesday at $20, up $0.063. *