A health-care start-up called Partner Therapeutics began last year with a single product: a leukemia medicine approved in 1991 that doctors rarely prescribe anymore. The drug, Leukine, made so little money that its previous owner did not even bother to disclose sales. It just dumped them on revenue reports under "other."

But now researchers in Colorado are testing Leukine's ability to regulate the immune system as a treatment for Alzheimer's disease.

The possible reinvention of a three-decade-old off-patent drug is among alternative approaches receiving fresh attention in Alzheimer's research after broad failures by major drug companies. Once-promising treatments that targeted the removal of amyloid plaques, an accumulation of debris on brain tissue that is a key sign of Alzheimer's, have consumed billions of dollars in vain.

As a result, big companies are retrenching. That mostly leaves start-ups and academic labs to test new hypotheses, including ways to get rogue inflammatory agents in an Alzheimer's patient's brain under control.

"A lot of folks are giving up because they've sunk a ton of money" into Alzheimer's research without success, said Robert Mulroy, chief executive at Partner Therapeutics, headquartered in leased medical offices just steps from the Revolutionary War's Lexington Battle Green, in Lexington, Massachusetts. "In the absence of huge companies, small companies come in, and they can continue the research and fill the gap. That's what's going on right now."

That puts at least a portion of the big companies' research money and development know-how on the sidelines for the time being. A few major drug firms are sponsoring clinical trials and supporting small companies with venture capital, but where the multinational firms will place their next big bets remains unclear.

Big Pharma's Alzheimer's disease failures have come one after another in the past two years. Biogen stopped a clinical trial in March that was the latest stunning collapse for experimental amyloid plaque drugs, sending the company's stock plunging more than 29 percent. Johnson & Johnson, Merck and a partnership of Eli Lilly and AstraZeneca experienced similar failures.

Last year, Pfizer ended in-house Alzheimer's research as it closed its neurology division, slashing its investment in the field and focusing on some smaller Alzheimer's-related partnerships. Pfizer produced statistical clues in 2015 that its anti-inflammatory drug Enbrel had potential as a preventive measure, but after some internal debate, the company decided it was just another dead end, The Washington Post reported last month.

As of last year, there were still more than 70 potential Alzheimer’s therapies in various stages of clinical trials, in addition to 22 remaining amyloid-targeting drugs, according to a review by the industry trade group Pharmaceutical Research and Manufacturers of America. That compares with 1,100 drugs in development for cancer, 445 for other neurological diseases, and 200 for heart disease and stroke.

Drugs targeting another protein accumulation in the brain called tau protein tangles, as well as anti-inflammation drugs, immune modulators, gene therapy, insulin and cannabis compounds, are among other avenues still being pursued.

One researcher likens a brain with Alzheimer's disease to an old car with multiple system failures. Treatment may require cocktails of drugs to treat different aspects. Drugs that influence the brain's immune cells, called microglia, are attracting renewed attention.

Researchers at the Rocky Mountain Alzheimer's disease Center are leading a trial of Leukine in 40 Alzheimer's patients. Partner Therapeutics bought the drug and a manufacturing facility in Lynnwood, Washington, from Sanofi last year for an undisclosed sum. Leukine is used to treat radiation poisoning and is being studied for cancer, as well as Parkinson's disease.

In mice with Alzheimer's disease, the same protein contained in Leukine cleared amyloid debris from the brain while reversing memory loss.

Given the many failures that have occurred, including drugs that showed promising results in mice, odds of a breakthrough in humans are long.

"This is really a completely different approach than anything that has been tried before," said Huntington Potter, director of the Rocky Mountain center, which is part of the University of Colorado. "It is one approach of many, and we're hopeful. But science will tell."

The role of the immune system and inflammation in Alzheimer's disease has been broadly known for years, but recent advances in human genome mapping have made it possible to more precisely target the role of specific genes. A trio of publicly traded California companies — Denali Therapeutics, Alector and INmuneBio — are capitalizing on the approach and being closely watched in the field. Their work raises the possibility that immune system advances in fighting cancer could be duplicated in brain disease.

Arnon Rosenthal, Alector's chief executive, said the industry spent too much time chasing down drugs to reduce amyloid plaques, despite abundant evidence that they were not working.

"It was irrational to do that. It was like the definition of insanity," he said. "People should not lose hope because one approach failed."

Rosenthal likened microglia to a police force that sometimes goes rogue in the brains of elderly people and not only is unable to stop crime but also starts indiscriminately shooting innocent victims. Excess inflammation and neuron damage are among the catastrophic results in the brain. Drugs that influence how microglia do their job hold great promise for treatment, he said.

"If we can repair them and restrengthen them and get them to come back, we can convert a rogue police force to a good and healthy police force," Rosenthal said.

Denali is investigating similar immunotherapy approaches. The company's experimental molecules are engineered to pass through a membrane called the blood-brain barrier, which Ryan Watts, the company's chief executive, said will allow them to more directly influence the brain.

Other inflammation-regulating drugs are being studied.

IntelGenx, a company in Quebec, is performing a clinical trial of an anti-inflammatory asthma drug called montelukast, sold under the brand name Singulair, as an Alzheimer's treatment. It has been a generic drug in the United States since 2012. To provide a potential advantage over generics, IntelGenx has reformulated the drug as a dissolvable strip, similar to a Listerine breath strip, which the company says will make it easier for elderly patients to ingest.

In England, researchers are studying whether three existing drugs — known under the brand names Enbrel, Humira and Remicade — that target an anti-inflammation protein called TNF-a can slow the progression of cognitive decline in 360 elderly patients with rheumatoid arthritis. All three drugs are brand-name blockbusters for diseases such as rheumatoid arthritis, psoriasis and Crohn's but are facing generic competition. Enbrel showed signs of being helpful in Alzheimer's in a 2016 statistical analysis of patient records.

The drug, which is marketed by Amgen in North America and by Pfizer overseas, was shown in a 2015 internal, unpublished Pfizer analysis of insurance claims data to apparently reduce the risk of developing Alzheimer's by 64 percent. The Pfizer analysis showed that Humira, which is marketed by AbbVie, had a similar effect.

The results seemed to bolster the hypothesis that early use of anti-inflammatory drugs, before symptoms appear, can prevent the disease.

Pfizer opted against a long-term human trial to test whether its statistical findings would hold up in real practice, despite urging from some of its researchers, The Post reported. Pfizer said science does not support using Enbrel to prevent Alzheimer's, particularly because the molecule is too large to cross the blood-brain barrier.

The company also opted not to publish the data. The company said the statistical analysis lacked scientific rigor and would have raised "false hope."

Some in the industry have said that the insurance claims data appeared weak, and that Pfizer made the right choice in not sharing it with the broader scientific community. Others have said that sharing the findings with the public would enhance understanding, positive or negative, of anti-inflammatory drugs in Alzheimer's.

"It's important to share any information that could help the field make better and more informed decisions about future treatments, and that's what we've encouraged Pfizer to do," said Maria Carrillo, chief science officer of the Alzheimer's Association, a nonprofit organization that funds research, including experimental drugs focused on inflammation and immunology.

Several years before the analysis of insurance claims data, Pfizer provided Enbrel at no cost for a six-month clinical trial studying treatment in 41 patients with mild to moderate Alzheimer's (20 who received the drug, 21 who received a placebo).

Initial results showed some cognitive benefit, lead investigator Clive Holmes said, but after they were adjusted to reflect the number of tests, the benefits were deemed statistically insignificant. The trial lasted from 2011 to 2013 and was published in full in 2015.

"We had a good signal here that warranted further exploration," Holmes, a professor of biological psychiatry at the University of Southampton in England, wrote in an email.

Pfizer has declined The Post's requests to interview top executives about the subject. At a Pfizer-sponsored health-care forum held at The Post in June, Rod MacKenzie, the company's chief development officer and executive vice president, disagreed with Holmes's assessment of the trial.

"It did not validate the hypothesis. It was inconclusive at best, and there were no statistically different effects with the placebo and the active drug," MacKenzie said, "and that's a shame."

Pfizer was not the first company to opt against testing a blockbuster anti-inflammatory drug for Alzheimer's.

German researchers studying inflammation's effect on the disease in mice found that the molecule in Stelara, a drug sold by Johnson & Johnson's Janssen division, had benefit.

Johnson & Johnson confirmed that it looked at the data and decided not to move forward.

"We believe other options have a more viable scientific path forward in Alzheimer's disease," company spokeswoman Kellie McLaughlin said in an email. The company said it is investigating two Alzheimer's drugs, including a collaboration with a company called AC Immune that is developing a vaccine against tau protein tangles.

Frank Heppner, one of the German scientists who discovered the potential benefit of Stelara, said drug companies are wary of testing existing drugs in elderly people who are at higher risk of serious side effects.

More adverse events from an Alzheimer's clinical trial, Heppner said, "would give a negative impact for this drug, and would endanger the financial success of the drug in their primary market."

“Many companies burned their fingers stepping into clinical trials specifically in Alzheimer’s disease,” he added. “Even if they see a huge market, it’s a big fire wall they do not easily dare to step over.”