The bioartificial kidney is one of the most promising examples to date of a bioengineered medical device. The innovative, external device passes blood through a cartridge of human kidney cells. In early clinical trials, it was shown to improve patient survival one month after treatment better than dialysis alone. But scientists now face a challenge that may be as great as designing the device itself: turning a successful academic invention into a mass-produced medical device.

“The question is, How do you turn 100 donated kidneys into 100,000 devices?” says David Humes, an internist at the University of Michigan, in Ann Arbor, and creator of the device. “You have to isolate the cells, expand them, and make sure they haven’t lost any potency in the process.”

Unlike small-molecule drugs or mechanical devices such as pacemakers, cells behave in unpredictable ways. That makes it difficult to develop reliable methods for growing massive quantities of a specific type of cell. Experts say this issue is proving to be a serious hurdle to the development of cell-based treatments. Despite huge demand for replacement organs, few companies have managed to produce tissue-engineered therapies for market. And the problem is likely to grow as more companies attempt to commercialize these therapies.

In the United States, 400,000 people have chronic kidney problems that require weekly dialysis, and 120,000 suffer acute renal failure, in which kidney function is knocked out by toxins or infection. Dialysis extends the lives of these patients, but it’s not a cure: life expectancy for most patients is just five years.

Traditional dialysis filters and discards metabolic waste from the blood, and then returns cleansed blood to the patient. Humes’s artificial kidney, also known as a renal assist device, adds an extra step to this process, passing blood and filtrate through a cartridge of human kidney cells. Humes theorizes that these cells perform some of the kidney’s noncleansing functions, such as regulating inflammation and metabolic processes, by secreting crucial chemicals into the blood.

To make bioartificial kidneys, scientists grow cells harvested from donor kidneys not suitable for transplant and then insert them into a specially developed filter tube. Because the finished product contains live cells, it is treated like an organ for transplant, flown to the receiving hospital by helicopter in a temperature-controlled case. Humes founded a company, now known as RenaMed, to commercialize the device, which has not yet been approved by the Food and Drug Administration.

Early clinical trials of the device show that it can dramatically improve the health of patients with acute renal failure. According to the results of a trial released last year, patients treated with the device showed a 70 percent improved survival rate 28 days after treatment. (Scientists at RenaMed are currently analyzing interim results from a subsequent trial.)

However, the devices used in these trials were made with a manufacturing process that is only appropriate for growing small batches of cells. To run the larger clinical trials required for approval by the FDA and to supply needy patients if the device is approved, RenaMed will need to find a way to make and deliver the device on a much larger scale.

When growing cells for therapies, scientists must create a robust protocol that reliably produces the target cell, as well as quality-control measures to keep the process on track. “People in academic labs develop techniques based on their own green thumbs at the bench,” says Michael Lysaght, a tissue engineer at Brown University, in Providence, RI. “But there’s a totally different culture when it comes to producing things for the FDA. Every step has to be understood and extremely well documented and capable of being done by anyone.” In addition, he says, the safety and reproducibility testing required by the FDA is much more rigorous than that needed for pilot clinical trials.

Both Humes and Lysaght liken the problem to that faced twenty years ago by researchers working with recombinant proteins, such as human insulin. Scientists could successfully make the proteins in the lab, but it took several years to figure out how to scale up that process for broad medical use. Lysaght says he’s confident the same will be true for tissue-engineered products once people recognize the extent of the problem.

If RenaMed can clear the hurdle, it may be able to lead the way for other bioengineered devices. “If there ever was a fair-haired child in tissue engineering, it was this device,” says Lysaght. “Everyone is hoping it will be a big success.”