Scientists with Roche Holding AG, the parent company of Indianapolis-based Roche Diagnostics Corp., may have found a way to overcome a blood barrier that keeps drugs from directly entering the brain, potentially opening new pathways to attack Alzheimer’s disease.

The technique, tested in animals, makes use of a receptor that carries iron molecules across the barrier of blood, fluid and membranes that keeps bacteria and other substances, such as medicines, out of the brain, said Ryan Watts, a researcher at Roche’s Genentech unit in San Francisco. The scientists configured a protein called an antibody to hitch a ride on the receptor, he said.

Alzheimer’s impairs mental function in 18 million people globally, the World Health Organization says. Numerous drugmakers, including Indianapolis-based Eli Lilly and Co., are struggling to find a treatment.

Namenda, from New York-based Forest Laboratories Inc., and Aricept, made by Pfizer Inc. of New York and Tokyo-based Eisai Co., address symptoms without slowing or curing the disease. Efforts to alter its course with drugs have failed partly because of the barrier.

“It’s brilliant,” said Robert Vassar, a professor of cell and molecular biology at Northwestern University Medical School in Chicago. “They hijacked a mechanism that is a normal part of the blood-brain barrier.”

Roche, based in Basel, Switzerland, is Europe’s largest drugmaker by sales. The approach described Wednesday in the journal Science Translational Medicine may also work for Huntington’s and Parkinson’s, the researchers said.

It is “an elegant strategy” that provides “proof of principle” that this obstacle can be overcome, Steven Paul, a researcher at Weill Cornell Medical College in New York, said in a commentary published alongside the research.

The hallmark of Alzheimer’s is the formation of clumps of a protein called beta amyloid and tangles of another called tau. Scientists don’t know why they accumulate or become twisted, and there is debate as to whether they cause the illness or are an end-product of some different process.

Watts, the study leader, is developing a drug that blocks the action of an enzyme called BACE1 that’s involved in amyloid production. When his team tried to get their anti-BACE1 into the brains of mice and monkeys, they found that only a tiny fraction made it there.

“To do what we wanted to do in the brain, we had to dose like crazy, frequently and at high levels,” Watts said in an interview at Genentech’s campus. Such high dosages would be expensive and infeasible, Paul said.

“We needed a solution,” Watts said.

Watts turned to Mark Dennis, from Genentech’s department of antibody engineering, who took advantage of the fact that all cells, including brain cells, need iron. He engineered an antibody with two arms. One arm was the anti-BACE1 drug; the other docked with a receptor called transferrin that carries iron to brain cells, providing a ferry across the barrier.

The system allowed the researchers to deliver anti-BACE1 to the brains of mice, blunting the impact of the BACE1 enzyme and cutting in half the amount of amyloid in the brains of mice 48 hours after injection, according to the journal report.

More work is needed before the two-armed antibody can be tested in people, Watts said. A human version of the transferrin receptor antibody needs to be created and more safety testing must be done on large animals, he said.

“I think the prospects are quite strong” that the research could lead to a human therapy, said Vassar, who first identified and cloned the BACE1 enzyme in 1999.

Last August, Lilly released data showing that semagacestat, a drug directed against an enzyme involved in the production of amyloid, harmed patients instead of helping them. While that failure dampened enthusiasm for medicines that target amyloid, many researchers still see the plaques as being involved in development of the disease.

Pfizer, the world’s largest drugmaker, and Johnson & Johnson, of New Brunswick, N.J., for instance, are testing a drug aimed at amyloid called bapineuzumab, and Lilly is testing another, called solanezumab. Both are in late-stage trials that should be completed within two years, Weill Cornell’s Paul said in his commentary.