June 4 (Bloomberg) -- Genentech Inc., the world's largest maker of cancer-fighting drugs, said a charged-up version of its Herceptin drug for breast cancer shrank tumors in patients who failed to respond to other therapies.

Four of 10 patients had their tumors shrink during a study, with minimal side effects, when taking a version of Herceptin linked to chemical toxins, researchers said at the American Society of Clinical Oncology meeting in Chicago.

The combination is in the first of three phases generally needed for U.S. marketing approval. Genentech said its so-called `armed' antibody will advance into a second-stage trial. One analyst said the more-potent form could generate $1.5 billion on its own, if approved by U.S. regulators. The currently marketed product had $1.2 billion in sales in 2006.

``This is the most exciting thing to me that we are doing,'' said Pamela Klein, Genentech's vice president of clinical hematology and oncology, in an interview yesterday.

In April, when Genentech first announced it was testing the combination, analyst Geoffrey Porges of Sanford Bernstein & Co. in New York said the charged-up version may increase peak annual sales to $1.6 billion.

Genentech makes the combination with ImmunoGen Inc. of Cambridge, Massachusetts, a partner that provides linking technology for the drug. In April, when Genentech first announced it was testing the combination, analyst Geoffrey Porges of Sanford Bernstein & Co. in New York said the charged-up version may increase peak annual sales to $1.6 billion.

Shares

ImmunoGen's shares rose 2 cents to $5.94 at 4:00 p.m. New York time in Nasdaq Stock Market composite trading. The shares of South San Francisco based Genentech fell $1.55, or 2 percent, to $77.95 at 4:01 p.m. in New York Stock Exchange composite trading.

The study was designed to measure safety, and the drug was generally well-tolerated, Klein said. No serious heart damage was seen, the most worrisome potential side effect, Klein said.

The patients in the study were difficult to treat, and Klein said she didn't expect to see such a tumor response. She cautioned that the company needs to do further studies in a larger patient population, and has ``much work to do.''

The treatment binds tumor-killing chemicals to Herceptin, causing fewer side effects by attacking cancer cells and mostly leaving normal tissue alone.

Early Forms

The research, presented late yesterday, may give Genentech a way to sell its standard Herceptin for early forms of breast cancer, and then provide the more potent therapy after the disease progresses. The strategy may spur new treatment practices for a disease diagnosed in more than 178,000 U.S. women a year.

Herceptin had $1.2 billion in sales in 2006, making it Genentech's third-biggest drug behind cancer therapies Rituxan and Avastin. The drug also adds revenue for Basel, Switzerland- based drugmaker Roche AG, Genentech's largest shareholder.

The more potent version could also help Genentech fend off competition from GlaxoSmithKline Plc.'s Tykerb, analysts said. The U.S. Food and Drug Administration approved the rival drug, which also targets HER-2 proteins, in March.

About one-fourth of breast cancer patients have tumors that carry excess amounts of HER-2 protein, the protein that Herceptin, an antibody drug, is designed to target, Genentech says. The drug works by blocking HER-2 positive cancer cells from dividing and growing.

Different Approach

The new technology offers a different approach to earlier efforts that attached radiation to antibodies such as Herceptin, instead of toxic chemicals. Those approaches struggled commercially because oncologists must collaborate with nuclear medicine specialists, doctors say.

Included in the old approach is Biogen Idec Inc.'s radiation-antibody combination Zevalin, which recorded $16.4 million in sales in its fourth year on the market. The company is now seeking to divest that drug, according to the Cambridge, Massachusetts-based company's annual report with the Securities and Exchange Commission.

Past combinations have failed because the toxins either snapped off in the blood before reaching the tumors or hit healthy cells on the way, causing side effects. Seattle Genetics, another collaborator of Genentech, says it has developed new synthetic linkers designed to remain stable in the blood, yet let go of its toxic payload inside tumors.

ImmunoGen has worked on the linking technology since it was founded in 1981. The company doesn't yet have a marketed drug, or one in the final stage of testing, according to its Web site.

To contact the reporter on this story: Luke Timmerman in San Francisco at ltimmerman@bloomberg.net.

Last Updated: June 4, 2007 16:22 EDT