The high therapeutic development failure rate means there are many existing therapeutic candidates that could be repurposed for use in a new disease indication.
Imagine a library of thousands of drugs that have already been tested for safety but failed to treat the conditions they were designed for. What if these same compounds held the key to treating completely different diseases?
This isn't science fiction—it's the promising reality of drug repurposing being advanced by the National Center for Advancing Translational Sciences (NCATS).
At NCATS, scientists are addressing a critical problem in medicine: the traditional drug development process takes approximately 14 years, costs over $1 billion per successful drug, and has a failure rate exceeding 95% 9 . In response, researchers have developed innovative approaches to find new uses for existing drugs, dramatically shortening the journey from laboratory discovery to patient treatment.
Drug repurposing (also called drug repositioning) is the strategy of identifying new therapeutic uses for approved or investigational drugs outside their original medical indication 1 . This approach bypasses much of the early safety testing since these compounds have already undergone significant research and development, including testing in humans 4 .
Repurposed drugs can reach patients in roughly half the time of new drugs
Development costs can be reduced by as much as 50-60%
Since safety profiles are already established, the risk of late-stage failure decreases
At NCATS, the Biomedical Data Translator program ("Translator") serves as a cornerstone of this repurposing effort. This multi-year initiative addresses a critical challenge in modern medicine: biomedical data exists in isolated silos across different locations, often in forms that aren't compatible with each other 3 . Translator integrates over 250 knowledge sources—including highly curated biomedical databases, ontologies, and NCATS-owned resources—creating a unified system that researchers can query to discover novel relationships between drugs and diseases 1 7 .
NCATS researchers leverage two powerful programs that provide complementary types of biomedical data:
The Toxicology in the 21st Century (Tox21) program established a library of approximately 10,000 compounds, including about 3,700 approved/investigational drugs and 5,200 environmental chemicals 1 . This comprehensive collection has been screened against more than 70 in-vitro assays covering targets and pathways related to nuclear receptor signaling, stress response, and cytotoxicity 1 .
The Biomedical Data Translator is an ambitious computational reasoning and knowledge exploration system that integrates disparate biomedical data sources 1 3 . Think of it as a "Google for biomedical relationships" that can answer complex queries like "predict treatments for disease Y" 3 . The system works through Autonomous Relay Agents that break queries into smaller tasks distributed to specialized knowledge bases, creating an iterative discovery process 3 .
| Category | Examples of Data Types |
|---|---|
| Clinical Data | Electronic health records, prescription data, health insurance claims, adverse event reports |
| Molecular & Cellular Data | Genes, proteins, proteomes, transcriptomes, epigenomes, molecular mechanisms |
| Disease Information | Diseases, signs of disease, symptoms, patient-reported outcomes, biomarkers |
| Intervention Data | Therapeutic interventions, intervention exposure, pharmacokinetics/pharmacodynamics |
A groundbreaking study published in PLOS ONE demonstrates how NCATS researchers systematically mine their resources for drug repurposing opportunities 1 . The methodology provides a template for how modern data-driven approaches can accelerate therapeutic discovery.
Researchers began with the Tox21 10K compound library, focusing on the 7,170 substances that had complete activity data across all Tox21 in-vitro bioassays 1 . Each compound's activity was measured using a curve rank metric (values from -9 to 9), where large positive values indicate strong activation and large negative values indicate strong inhibition of assay targets 1 .
The team applied a Self-Organizing Map (SOM) model—an artificial neural network that projects high-dimensional data onto a two-dimensional map while preserving topological relationships 1 . This technique grouped compounds with similar bioactivity profiles, operating on the hypothesis that chemicals with similar activity patterns may share similar targets or mechanisms of action 1 .
For each chemical cluster, the team identified significantly enriched gene targets by gathering known chemical-gene associations from comprehensive resources including Pharos and the Board Drug Repurposing Hub (BDRH) 1 . These databases provide carefully curated information about drug targets and drug-disease relationships.
| Resource | Function | Application in Drug Repurposing |
|---|---|---|
| Tox21 | Screening library of ~10,000 compounds against 70+ assays | Provides biological activity data for clustering compounds |
| Translator | Integrates 250+ biomedical knowledge sources | Supplies scientific evidence for drug-disease relationships |
| Pharos | Target Central Resource Database | Offers information on drug targets and understudied proteins |
| Drug Repurposing Hub | Manually curated collection of 4,704 compounds | Provides annotated chemical-gene and chemical-disease associations |
| GARD | Genetic and Rare Diseases Information Center | Offers rare disease information for validating associations |
The NCATS approach to drug repurposing relies on both wet-lab and computational resources that work in concert:
The foundation containing approximately 10,000 substances including drugs, pesticides, consumer products, and industrial chemicals 1 . This diversity enables discovery of unexpected therapeutic connections.
The 70+ in-vitro tests that measure compound activity across biological pathways including nuclear receptor signaling (55.90%), stress response (11.80%), and cytotoxicity (8.80%) 1 .
Self-Organizing Maps and hierarchical clustering techniques that identify patterns in complex biological activity data 1 .
Integrated networks including the NCATS GARD Knowledge Graph (NGKG) that connect rare diseases with various biological entities to validate repurposing hypotheses 1 .
The N3C Data Enclave provides a secure, harmonized health data resource containing information from over 22.9 million individuals, enabling validation of discoveries in real-world populations .
The implications of successful drug repurposing are profound, particularly for rare diseases that often lack treatments because developing new drugs is economically challenging for pharmaceutical companies 9 . NCATS has already funded projects to repurpose existing compounds for conditions including graft-versus-host disease, idiopathic pulmonary fibrosis, and chronic obstructive pulmonary disease 4 .
"The work happening at NCATS demonstrates that the future of medicine may depend less on discovering completely new compounds and more on looking at existing drugs through a new lens—one powered by data, computational reasoning, and a mission to get more treatments to more people more quickly."
The future of this field is increasingly data-driven and collaborative. NCATS partners with initiatives like the AIM-AHEAD Health Data Science Training Program to build workforce capacity in AI and machine learning applications for healthcare . These partnerships aim to create a representative AI/ML workforce capable of leveraging large, real-world health datasets to accelerate medical discoveries .
| Factor | Traditional Drug Development | Drug Repurposing |
|---|---|---|
| Timeline | ~14 years from discovery to approval | Significantly reduced (potentially 3-5 years) |
| Cost | $1 billion+ per successful drug | Substantially lower due to reduced early-stage testing |
| Failure Rate | >95% | Lower, especially for safety-related failures |
| Key Resources | Novel compound discovery | Existing compound libraries & biomedical data |
| Primary Challenge | High risk of failure | Identifying novel disease connections |
As NCATS Director Joni L. Rutter, Ph.D., emphasized in a recent message, easing the burden of rare diseases through translational science remains a key priority 2 . The systematic approach to drug repurposing—combining high-quality experimental data from Tox21 with the vast knowledge integrated by Translator—represents a powerful strategy to achieve this goal.
This article presents information about NCATS drug repurposing initiatives based on published research and publicly available data.