Old Drugs, New Uses: AI Lights a Beacon of Hope for Rare Disease Patients

Drug repurposing is more than just Viagra’s lucky accident. Learn how AI analyzes massive data to find new life for existing drugs, making rare disease treatment no longer an unattainable dream

Photo by Myriam Zilles on Unsplash

Danced with Death: David Fajgenbaum

I wrote this article because I read an exciting story. It all begins with a young doctor named David Fajgenbaum. In 2010, while passionately pursuing his medical degree, life dealt him a cruel twist of fate: he was diagnosed with Castleman disease, an extremely rare and deadly immune disorder. This wasn’t just an illness, it was a death sentence that led to multiple organ failure, severe abdominal pain, and swelling, bringing him to death’s door on five separate occasions.

Facing a dead end with no known cure, Fajgenbaum refused to surrender. He turned himself into the subject of his own experiment, using his blood samples and poring over mountains of medical literature day and night. This race finally led to a glimmer of hope: he discovered that

Sirolimus, a drug originally approved to prevent organ transplant rejection, worked wonders for his condition.

By repurposing this old drug, he successfully brought his disease into remission, where he has remained for over 11 years.

This near-death experience became his life’s mission. Convinced that countless other life-saving cures were hiding in plain sight, he co-founded the non-profit organization Every Cure. The NGO’s goal is to scale his personal journey, unlocking the hidden potential of existing medicines to offer hope to millions of patients still waiting for a cure.

Screenshot of homepage of Every Cure

What is Drug Repurposing? And Why Is It a Medical Goldmine?

Have you ever wondered if an ordinary pill could hold the secret to curing a completely different disease? That’s the core idea behind “drug repurposing” — finding new therapeutic uses for drugs that have already been approved and are on the market.

Experts point out that most drugs don’t just hit a single target; they often produce “side effects.” In the past, these were seen as nuisances to be avoided. Today, a growing number of specialists see them as “hidden talents” — potential solutions for other diseases. In fact, around 30% of all FDA-approved drugs already have at least one additional indication beyond their original purpose.

Drug repurposing is rapidly becoming a star player in future medicine for several compelling reasons:

• A Beacon of Hope for Rare Diseases: Globally, there are over 6,800 rare diseases affecting some 300 million people, the vast majority of whom have no treatment options. Traditional drug development is a long and expensive marathon, taking 10 to 15 years and billions of dollars, with a staggering 90% failure rate in human trials. Repurposed drugs, by contrast, have already passed extensive safety trials and have established data, making them a faster, cheaper, and more promising path forward.
• Better, Gentler Treatment Options: Beyond rare diseases, repurposing can identify alternative treatments for common ailments that come with fewer side effects, making the healing process less of an ordeal.
• Affordable, Accessible Medications: Many older drugs eventually lose their patent protection and become “generic drugs.” Once a patent expires, any pharmaceutical company can produce the drug, causing its price to plummet. This ensures that life-saving treatments don’t lead to financial ruin — just like the anti-rejection drug that saved Dr. Fajgenbaum’s life, which is now an affordable generic.

From Lucky Accidents to a Systematic Challenge

Before the age of AI, discovering new uses for old drugs was largely a matter of serendipity — or sheer luck. A classic example is Viagra (Sildenafil). Originally developed to treat angina, it was serendipitously found to be far more effective at dilating penile arteries and was repurposed for erectile dysfunction. Similarly, Minoxidil, a blood pressure medication, was discovered to boost circulation to the scalp, becoming the star ingredient in hair regrowth treatments.

But relying on lucky breaks is hardly a reliable strategy. It required a patient with Disease A to report a specific side effect that just happened to align with a potential treatment for Disease B, which would then kickstart years of research. The process was incredibly inefficient, like trying to find a needle in a haystack. As Dr. Fajgenbaum himself noted, while life-saving clues were buried in medical literature, the human brain could only compare one disease or one drug at a time, making it impossible to rapidly connect the dots.

Enter AI: How a Super-Brain Is Changing the Game

The bottleneck that had plagued medicine for decades began to dissolve with the rapid advancement of artificial intelligence. Dr. Fajgenbaum realized that this monumental task, too great for any human, was the perfect job for AI.

His organization, Every Cure, employs a powerful AI model that tirelessly reads and analyzes vast quantities of medical records and research papers. This AI platform acts like a super-researcher, having already analyzed 4,000 drugs to identify potential therapeutic links to nearly 20,000 diseases. It not only pinpoints which drugs (A, B, or C) might treat an illness but also ranks their probability of success, guiding researchers on where to focus their efforts. Backed by over $100 million in funding from initiatives like TED’s Audacious Project and ARPA-H, Every Cure is on a mission to unlock the full potential of every drug for every disease possible.

Another groundbreaking AI used for drug repurposing is TxGNN (also has a Web-based UI), developed by experts at Harvard University, including Marinka Zitnik. What makes TxGNN so powerful?

• It Explains Itself: TxGNN doesn’t just give you an answer; it explains the logical pathway it took to reach its conclusion, allowing physicians to understand the science behind the suggestion.
• It Predicts “Contraindications”: It excels at identifying which drugs are unsuitable for specific patient groups, with an accuracy 35% higher than other models, helping researchers avoid potential pitfalls in clinical trials.
• A Massive Knowledge Base: The model is pre-trained on a comprehensive knowledge graph containing 17,080 diseases and 7,957 therapeutic candidates.
• Zero-Shot Prediction: Once trained, TxGNN can make predictions for new diseases without any additional fine-tuning.
• Open and Accessible: Best of all, Harvard has made this powerful tool freely available, inviting doctors and researchers worldwide to join the hunt for new cures.

The Final Mile: The Funding Dilemma and the Bet on the Future

Despite the revolutionary potential of AI in drug repurposing, the field faces a very real and persistent obstacle: funding.

While AI is a powerful tool, it’s not infallible. Its conclusions still require rigorous evaluation and supervision by human physicians. More critically, major pharmaceutical companies have little financial incentive to fund this research. Since most repurposed drugs are old and off-patent, the profit potential is low. This leaves such studies heavily reliant on government funding.

Unfortunately, that funding is becoming increasingly precarious. The Trump administration, for example, significantly cut US federal research funding, criticizing some of it as “useless research.” The National Institutes of Health (NIH), a vital lifeline for over 300,000 researchers, saw its budget slashed, forcing labs to lay off staff and suspend promising projects.

Experts argue passionately that scientific funding should be shielded from politics, because what might seem like “useless research” today can lead to world-changing breakthroughs tomorrow. A prime example is the federally funded research in the 1980s on Gila monster saliva, which ultimately led to the development of the blockbuster weight-loss drug Ozempic.

As other global powers like China and the European Union (through its “Horizon Europe” program) pour money into scientific research, one question looms large: can the US afford to fall behind? The answer will not only determine a nation’s competitive edge but will also impact the lives of millions, like David Fajgenbaum, who are still waiting for their own “old drug, new use” miracle.