On October 9, 2014, Aastrom Biosciences, a homegrown biotech company, announced that it was changing its name to Vericel and moving its corporate headquarters from Ann Arbor to Cambridge, Massachusetts. The company has pursued stem cell therapies for twenty-five years but has never won approval for any. Long-term investors have lost almost everything. The company went public in 1997 at $7 a share; taking into account reverse stock splits, Wall Street now values those original shares at about 2c each. Aastrom has never made a profit, and it has burned through an average of more than $10 million in shareholder capital a year. Through December 2014, the company lost a cumulative $308 million.

It’s not unusual for biotech companies to hemorrhage cash for a decade or more paying for the clinical trials necessary to bring medical products to market. But Aastrom is an extreme case, reinventing itself multiple times. Now the company is once again pinning its hopes on a stem cell product that is undergoing human testing under the direction of the forty Vericel employees who still work at the company’s Domino’s Farms facility in Ann Arbor Township.

It’s employing basically the same technology pioneered in the late eighties and early nineties by U-M biologists Steve Emerson and Mike Clarke and bioengineer Bernhard Palsson, Aastrom’s scientific founders. The three were the first to demonstrate that cells capable of repopulating the circulatory and immune systems could be expanded in culture dishes. “We established a whole new field,” Emerson told the Observer in 1996.

Emerson now leads Columbia University’s comprehensive cancer center. “I wasn’t looking to start a company,” he recalls. “I was looking to do science that would result in therapies. And the fact that we ended up starting Aastrom was really an accident.” A chance conversation with a pharmaceutical company executive at a 1987 scientific seminar indirectly led a venture capital firm to finance the three scientists’ lab research. Together with the State of Michigan’s pension fund, venture capitalists then seeded the new company in 1989 for the development of an automated, self-contained system to grow cells for use in human patients. (The U-M also got shares in exchange for its patents on the technology.)

The potential seemed huge. Between 100,000 and 200,000 people a year in the U.S. needed to restore their immune systems after chemotherapy, and Aastrom’s “bioreactor” promised a faster and more complete recovery. The company’s first target was bone-marrow transplantation. Marrow cells taken from either the patient or a matched donor could be expanded in the bioreactor, minimizing the number that had to be extracted surgically while maximizing the number returned to the patient.

The company spent its first few years engineering and perfecting its bioreactor, a VCR-like device featuring disposable cassette units with interiors designed to closely mimic the natural bone marrow environment. Then Aastrom launched clinical trials for breast cancer and lymphoma. At a cancer meeting in 1997, researchers reported positive results in a small breast cancer trial, and Aastrom stock soon hit an all-time high of $9.94 a share.

It’s been almost all downhill since then. In 2000 and 2001 rigorous breast cancer studies found that high-dose chemotherapy with bone marrow transplantation didn’t work any better than standard chemotherapy. Almost overnight, thoracic oncologists abandoned marrow transplants. “The breast cancer autotransplant market evaporated,” recalls Emerson. Around the same time, the drug Neupogen corralled the rest of the bone marrow transplantation and chemotherapy support market. Neupogen caused stem cells to migrate from the bone marrow to the blood, where they could be easily and cheaply collected for transplantation. The drug also stimulated white-blood cell production after chemotherapy, making Aastrom’s bioreactor unnecessary there too. Fortuitously, the state had sold its entire position in Aastrom by March 2000, recovering roughly $4.5 million of the $4.6 million invested. (The U-M sold its shares in 2005.)

The culminating blow came from the federal Food and Drug Administration. For years its regulators signaled to Aastrom that they would review its bioreactor as a medical device–a relatively simple regulatory path. But in 2003, the FDA decided to regulate Aastrom’s cell products as drugs, vastly complicating the effort and expense that would be necessary to gain the agency’s approval. “They can change their mind–and they did,” recalls Susan Wyant, an Aastrom board member from 2002 to 2008. Aastrom terminated its bone marrow transplant trials, the stock dived, and in 2003 the company faced delisting from the NASDAQ stock exchange, which would have been fatal to its hopes of attracting new investors and probably would have resulted in the company’s liquidation.

Aastrom barely avoided delisting and survived. But with the original business plan in shambles, then-CEO Doug Armstrong “made a decision to go from being a device company to being a therapeutics company,” says Tim Mayleben, who joined the Aastrom board of directors in 2005. Aastrom’s technology was now a solution in search of a problem, a treatment in search of a disease.

In hindsight, the company made some bad calls. In the early 2000s, Aastrom developed cells for incorporation by other companies into therapeutic cancer “vaccines.” These treatments almost uniformly failed to work, and demand for the cells never materialized. Aastrom then began testing a proprietary cell mix to treat fractures that don’t heal properly and other orthopedic conditions. The company even launched a Phase 3 trial, the final step before FDA drug approval, in patients with osteonecrosis of the femoral head, a bone disease that causes the hip joint to fail from lack of blood supply.

But the bone trial shut down prematurely because, says Mayleben, it was taking too long to recruit patients. “The time required to get an answer to that clinical trial was … beyond any reasonable business model for a biotech company,” he says. “They needed to find applications for the technology, therapeutic areas, where the answers were going to be coming more quickly.” In any case, Aastrom was trying to compete with surgical treatments, including total hip replacement, that were already fairly effective. Orthopedics “really wasn’t a good market, even though the [Aastrom] product worked very well,” says former Aastrom chief scientific officer Ronnda Bartel.

So in 2008 the company reorganized again, this time developing cell products to treat cardiovascular disease. Aastrom’s ixmyelocel-T was a cell mixture heavy on mesenchymal stem cells–adult stem cells that form bone, muscle, and connective tissue–and certain immune cells that dampen inflammation. The idea was not that the stem cells would form new heart muscle or blood vessels, although this might happen to some extent. Rather, the stem cells would secrete proteins that would turn on the self-repair functions of existing cells. Meanwhile, the immune cells would dial down harmful inflammation. “We actually turned it into a pretty fantastic product,” says Bartel.

And it seemed to work. At a November 2011 meeting of the American Heart Association, researchers reported excellent trial results for ixmyelocel-T in critical limb ischemia (CLI), a severe blockage of the leg arteries that often leads to amputation. A year after entering the study, 67 percent of patients receiving a sham injection experienced gangrene, amputation, wound spreading, or death, compared to only 40 percent of the patients injected with ixmyelocel-T.

The data were solid, but much higher patient numbers would be needed to convince the FDA that the results were not a statistical fluke. In February 2012, Aastrom launched the Phase 3 REVIVE trial, which would treat 591 CLI patients at eighty-six sites across the country. The trial was financed by a $40 million stock sale to Eastern Capital Limited, an investment company owned by Ken Dart, a secretive billionaire investor whose family owns Dart Container Corporation in Mason, the world’s largest maker of foam cups.

But this trial, like the 2008 bone trial, shut down early with nothing to show. Aastrom’s managers seemed not to have learned their lesson, again overestimating their ability to quickly recruit patients. Only a trickle appeared, while the cost of providing professional support to eighty-six unproductive trial sites drained Aastrom’s coffers, including the Dart investment. “It was very difficult to recruit patients for that trial,” says current Vericel CFO Gerard Michel. “It was just getting prohibitively expensive, given the financial condition of the company, to continue.” Meanwhile, partnering talks with pharmaceutical companies, which could have extended a lifeline, failed to produce a final deal. In March 2013, Aastrom announced it was halting the REVIVE trial, cutting operating expenses by 50 percent, and laying off half its staff.

Terminating REVIVE stunned the entire stem cell research community, because it was the field’s great hope for a breakthrough. Stem cells, despite periodic media hype, have never delivered on their promise. Dozens of trials of mesenchymal stem cells have yet to yield an unambiguous success. (There have been only a few small trials undertaken with the ethically controversial embryonic stem cells.)

Most of Aastrom’s competitors have since either shut down or been sold at fire sale prices. But somehow the company survived the carnage. “I’m amazed at the survival power of Aastrom,” says Susan Wyant. “They’ve gone to the well so many times but have always come up with enough cash to keep moving.”

Wyant gives Doug Armstrong, the CEO from 1991 to 2006, much of the credit. “He was a true visionary about the science,” she says. And, “he could sell ice to Eskimos.” Aastrom had to be doing something right to lose $308 million and survive, and even today ixmyelocel-T does have compelling selling points to knowledgeable investors. Its mixture of cells should pack more healing power than stem cells alone. Aastrom’s science has been rigorous, unlike many stem cell companies, which, says Wyant, took shortcuts on the science in order to generate quick but unreliable results. “Those companies don’t really exist anymore,” she says. Aastrom’s manufacturing process, thanks to clever engineering during the company’s early years, is much simpler and less prone to contamination than competing systems, and studies suggest it produces more powerful cells. But the bioreactor also has some limitations, mainly size. Because it was designed for bone marrow transplants, it is too small for some applications–for example, certain T cell therapies for cancer–and too big for others.

Emerson regrets the failure to produce a lab-scale desktop version of the bioreactor. “I would have loved to been able to have a small device you could do laboratory experiments in,” he says. More animal studies, he says, might have better informed the human trials, and perhaps helped avoid some of the company’s expensive missteps. “We scientists who pushed for this should have pushed even harder,” he says.

Meanwhile, the entire cell therapy field has become toxic to investors. “Nothing has come to market,” notes Wyant. “The promise has not been fulfilled.” And big pharma isn’t interested in acquiring cell therapy companies, so potential investors see no exit strategy. All this makes Aastrom’s fundraising accomplishments remarkable, but investor fatigue has hampered the company for years. Aastrom seemed to always generate enough cash to start an ambitious new initiative but never enough to finish.

The November name change and headquarters move are part of an aggressive reorganization under new CEO Nick Colangelo. Last June, Aastrom bought three commercial cell therapy products from the French pharmaceutical giant Sanofi for $6.5 million. These products, for cartilage repair and for severe burns, date back to the 1990s and were losing money for Sanofi. Vericel is now working hard to boost sales and cut costs, including laying off about fifty employees in Cambridge late last year, hiring sales staff, and streamlining production. “We’ve found quite a few ways to improve,” says Michel.

But Vericel is unlikely to wring enough profits from these old products to satisfy the growth expectations of Vericel’s large hedge fund investors. (Investors in biotech seek very high returns in exchange for high risk.) For that, Vericel is now finishing a 108-patient trial of ixmyelocel-T in ischemic dilated cardiomyopathy, a severe form of heart failure. In September, Aastrom raised $37.5 million in a public stock offering, and Michel says the company has enough money to see the trial through. (An earlier trial showed the treatment, which involves catheter delivery of ixmyelocel-T directly into the heart wall, was safe.) In January Vericel announced that the last patient had been treated, and the company should report results in early 2016. The treatment “would be a game-changer for the company even before it hit the market, if the current trial yielded robust results,” Michel says. If the results are strong, the trial might even be enough to win FDA approval, although that’s impossible to predict. For now, Vericel’s Ann Arbor jobs seem secure. “We’re very satisfied with having that group based there,” says Michel.

So, after a quarter century of frustration, the final verdict on Aastrom has yet to be written. Emerson still expects success. “The company was definitely ahead of its time,” he says. “It’s the right technology, right science.” Cell therapy, Emerson goes on to predict, “will be very big.” But stem cell therapy, whether Vericel’s version or someone else’s, needs to show it works, and soon, before investors finally lose patience. “People are only interested for so long,” says Wyant. “If it continues to be a promise with no basis, then people get disenchanted and move on to the next thing.”