Gene therapy's obstacle: custom-made viruses

Eager to speed development of revolutionary treatments, the Food and Drug Administration recently announced that it would expedite approval of experimental gene therapies. But the regulatory process may not be the biggest obstacle here.

Biotech companies have exciting plans to introduce treatments that may be transformative, sometimes curing genetic diseases with a single treatment. And the firms are itching to test their products.

But they are struggling to obtain a critical component of the therapy: the disabled viruses used to slip good genes into cells that lack them.

This delivery system lies at the heart of many forms of gene therapy; without the disabled viruses, there is no treatment. But manufacturing them is costly and onerous.

The genes intended to fix a defect in the body are carried into each cell by a modified virus, usually a disabled version of an adenovirus or a lentivirus. These viruses must be custom-made in specialized facilities for each treatment.

Few gene-therapy companies have the factories or expertise to make the viruses for use in clinical trials, where standards are exacting and comprehensive. The firms that can do it are swamped with orders and requests.

The result is a logjam. Firms exploring new gene therapies may wait for years in line for bespoke viruses, said Dr. Jim Wilson, director of the gene therapy program at the University of Pennsylvania's Perelman School of Medicine.

"It's a real issue," said Udit Batra, chief executive of MilliporeSigma, which makes viruses under contract for drug companies.

MilliporeSigma and other such manufacturers, he added, are "oversubscribed, although companies like ourselves have doubled capacity to keep up with the demand."

One of the few big companies producing a gene therapy, Novartis, recently got approval from the FDA to market a treatment for a rare blood cancer.

But to get the viruses it needed, Novartis signed up years in advance with Oxford BioMedica, agreeing to three contracts starting in 2013 that, with incentives, add up to as much as $195.2 million and that included a provision to pay Oxford a share of the royalties when the drug was approved.

Only a few hundred patients a year might need Novartis' treatment, and the company is charging $475,000 for the one-time therapy.

Other gene therapy companies are not always able to afford the manufacturing costs or find a manufacturer. Some have taken to buying slots in virus production queues years in advance - like buying a nonrefundable airline ticket long before your vacation and hoping you can get away when the time comes.

Other firms are hedging their bets. Worried that production at one company will fail - as can happen with the finicky viruses - they buy places in line at two contract companies.

Still other biotechs have simply been shut out, unable to get their viruses made.

Then there is BioMarin, one of the larger and more successful biotech companies, which decided to spend several hundred million dollars to build its own virus-manufacturing plant. It does not plan to make viruses for anyone but itself.

"We don't want to be in a queue, that's for sure," said Robert Baffi, head of technical operations at BioMarin. The new facility also will give the company complete control over manufacturing, he added.

The process of developing a gene therapy usually starts with academic researchers who do the preliminary tests. For the viruses they need, they often turn to a few academic medical centers with expertise in the requirements for early clinical research.

But there, too, demand far exceeds capacity. At Indiana University, "we are backed up through  2018," said Dr. Kenneth Cornetta, a professor of molecular and medical genetics.

After a gene therapy gets through initial tests in an academic setting, researchers may license it to a biotech company or form their own small company. Then they have to find a manufacturer who will make their viruses according to the exacting standards required for treating patients.

Delays arise at every step. The contract virus-maker has to translate the small-scale production used for research purposes into a recipe for commercial production, where standards are extensive and documentation exhaustive. And the maker has to negotiate a contract to do all this.

Those two steps can easily take a year, said John Dawson, chief executive of Oxford BioMedica. When the contractor finally is ready to start making the viruses, it can be six months to a year before they are ready - assuming there are no glitches along the way.

Manufacturing custom-made viruses can cost biotech firms a third or more of their development budget, even for diseases so rare that they expect to treat only a dozen or so patients in their final study, Wilson said.

The gene therapy companies often have no drugs on the market and need money. But investors have become wary of companies that do not have a ready source of viruses.

"You've got to believe that every time someone gives a pitch to an investor, the investor will say, 'What are you doing about manufacturing?'" Wilson said.

The whole development enterprise has become nerve-racking, researchers said. "You don't know until the end that you have a product that is good enough to be used in a treatment," said Dr. David Williams of Harvard.

Or, as officials at Bluebird Bio can attest, whether you have any product at all.

The company was formed in 2010, hoping to show that gene therapy could work in adrenoleukodystrophy, a rare and fatal neurodegenerative disease that strikes boys. That was before the virus production logjam had begun, and all seemed well. Bluebird gave a virus manufacturer its recipe for making needed viruses.

Then, said Nick Leschly, the company's chief executive, he got bad news. Using Bluebird's recipe, the manufacturing company said it was going to cost Bluebird 1 million dollars to create enough viruses to treat one patient.

The company scurried to find ways to improve the efficiency of its recipe. Finally, they were ready to start anew. Manufacturing began, but months later there was nothing to show for it.

"We got no virus," Leschly said.

"It was an Apollo 13 moment," he added. "We put everyone in a room and said, 'We have to figure this out. Everything at the company is now stopped. Nothing can be done without virus.'"

They finally found the source of the problem - the acidity of the solution used to grow the viruses was slightly off, killing them.

While the recipe for making viruses can affect prices, the cost of a new treatment also depends on how many patients will take the drug and how many cells from each patient must be altered by a virus.

If a company wanted to deliver a gene therapy to the lung or liver, where the organ's "surface area is huge," the current price could be as much as $3 million per patient - "commercially unviable," said Dawson of Oxford BioMedica.

Oxford is improving its methods, he said, and should soon be able to cut that cost to approximately $300,000 per patient. Methods are improving, Dawson said, and his expectation is that it might cost a mere $30,000 for the viruses in the future.

The costs of testing the drug and marketing it are, of course, out of his hands.