Finding New Cures in Existing Medicines
In the high-stakes race to develop new medicines, scientists are finding that the next breakthrough treatment might already be sitting on pharmacy shelves.
Imagine a world where a common diabetes medication could fight cancer, or a drug originally developed for hypertension could treat hair loss. This isn't science fiction—it's the exciting reality of drug repositioning, a revolutionary approach that's transforming how we discover new medicines. While traditional drug development often takes over a decade and costs billions of dollars, drug repositioning offers a faster, more efficient pathway to treatments by finding new therapeutic uses for existing drugs 1 5 .
At the heart of this revolution lies a powerful new strategy: integrating multiple types of disease data to systematically predict which old drugs might tackle new diseases. By combining different dimensions of disease information—from visible symptoms to molecular mechanisms—researchers are building comprehensive maps that reveal unexpected connections between drugs and diseases 2 .
The traditional drug development process is notoriously inefficient. Of thousands of compounds initially tested, only one typically reaches patients, after 12-17 years and investments exceeding $2 billion 3 5 . The attrition rate is staggering, with less than 10% of candidates progressing from preclinical stages to regulatory approval 3 .
12-17 years
3-12 years
Drug repositioning turns this model on its head. Since repositioned candidates have already undergone significant safety testing, they can bypass much of the early development timeline, potentially reaching patients in 3-12 years at about 60% of the cost of traditional development 5 6 . This accelerated path is particularly valuable for addressing urgent medical needs, as demonstrated during the COVID-19 pandemic when researchers screened existing drugs for potential antiviral effects 9 .
The economic and clinical impacts are substantial—approximately 30% of recent FDA approvals correspond to repurposed drugs, which currently generate about 25% of annual pharmaceutical revenue 6 .
Historically, drug repositioning occurred mostly by chance. The classic example is sildenafil, initially developed for hypertension but famously repurposed for erectile dysfunction when researchers noticed an unexpected side effect 1 6 . Similarly, minoxidil transformed from a blood pressure medication to a hair loss treatment when patients reported unexpected hair growth 1 6 .
Discovering that a drug interacts with unexpected biological targets, creating opportunities for entirely different therapeutic applications 1 .
The real power of modern repositioning emerges from integrating multiple data dimensions. Diseases can be similar in different ways—they might share observable symptoms (phenotypes), involve related biological pathways, or stem from common genetic factors. Each perspective offers unique insights into potential treatment opportunities 2 .
A groundbreaking 2025 study published in Scientific Reports exemplifies how integrating diverse disease data can enhance drug repositioning predictions 2 . The research team hypothesized that traditional methods relying on a single type of disease similarity (typically phenotypic data) were missing important connections.
Researchers constructed three distinct disease similarity networks, each capturing a different dimension of how diseases relate to each other:
Based on text mining the Online Mendelian Inheritance in Man (OMIM) database, which contains detailed descriptions of disease symptoms and characteristics 2 .
Using the Human Phenotype Ontology (HPO), which provides a standardized vocabulary and hierarchical structure for disease phenotypes, enabling more precise comparisons 2 .
Derived from disease-associated genes and their interactions in the HumanNet database, capturing similarities at the genetic level 2 .
These individual networks were integrated into a disease multiplex network and subsequently combined with drug similarity data to create a comprehensive multiplex-heterogeneous network. The researchers then applied a Random Walk with Restart (RWR) algorithm—a mathematical method that essentially simulates "hopping" across this complex network to identify potentially related drugs and diseases that aren't directly connected 2 .
The multi-network approach significantly outperformed single-network methods, correctly predicting 68 drug-disease associations supported by shared genes, 1,064 by shared pathways, and 84 by shared protein complexes 2 . Many of these predictions were subsequently validated against clinical trial data, confirming the method's practical utility.
| Network Type | Prediction Accuracy | Key Strengths |
|---|---|---|
| Phenotypic Network Only | Baseline | Captures clinically observable relationships |
| Molecular Network Only | Moderate improvement | Reveals shared genetic mechanisms |
| Multi-Network Integration | Significant improvement | Combines multiple dimensions for comprehensive mapping |
This experiment demonstrates that each type of disease similarity captures unique aspects of disease relationships. Phenotypic data might connect diseases with similar symptoms but different causes, while molecular data can reveal diseases with common genetic drivers despite different manifestations. By integrating these perspectives, researchers obtain a more complete picture of the complex therapeutic landscape 2 .
The growing success of systematic drug repositioning relies on an expanding collection of publicly available data resources and computational tools:
| Resource Type | Examples | Primary Function |
|---|---|---|
| Drug Databases | DrugBank, ChEMBL, PubChem | Detailed drug information, structures, and targets 3 |
| Target Databases | Open Targets, Pharos | Link drug targets to disease mechanisms 8 |
| Genomic Resources | Connectivity Map (CMap), LINCS | Gene expression profiles for drugs and diseases 3 9 |
| Disease Networks | OMIM, Human Phenotype Ontology | Standardized disease phenotypes and relationships 2 |
| Computational Tools | Mergeomics, ASGARD, iLINCS | Analyze multi-omics data for repositioning candidates 8 |
The practical impact of drug repositioning extends across numerous therapeutic areas, offering new hope for challenging diseases:
Research has identified several existing drugs with potential applications in Alzheimer's disease, including bexarotene (an anticancer drug) and liraglutide (a diabetes medication) 5 .
Repositioning is particularly valuable for rare conditions, where traditional drug development may be economically challenging 8 .
| Drug | Original Indication | Repositioned Indication | Repositioning Type |
|---|---|---|---|
| Sildenafil | Hypertension, Angina | Erectile Dysfunction, Pulmonary Hypertension | On-target 6 |
| Minoxidil | Hypertension | Hair Loss | On-target 1 6 |
| Thalidomide | Morning Sickness | Multiple Myeloma, Leprosy | Off-target 2 3 |
| Metformin | Type 2 Diabetes | Cancer (under investigation) | Off-target 1 7 |
Despite its promise, systematic drug repositioning faces several hurdles. Data heterogeneity—the challenge of integrating information from diverse sources and formats—remains a significant technical barrier 3 . Intellectual property issues can complicate development, particularly for generic drugs, potentially reducing commercial incentives . Additionally, computational predictions require costly experimental validation before clinical application 3 .
Looking ahead, the field is moving toward even more sophisticated approaches. Artificial intelligence and machine learning, particularly graph neural networks, are enabling more nuanced analysis of complex biological networks 3 . The integration of real-world evidence from electronic health records provides additional validation pathways 8 . Furthermore, researchers are increasingly focusing on multi-omics integration, combining genomic, proteomic, and metabolomic data for a systems-level understanding of drug effects 3 .
Drug repositioning represents a paradigm shift in therapeutic development, proving that the path to medical innovation doesn't always require creating new compounds from scratch. By systematically analyzing existing drugs through multiple lenses of disease similarity, researchers are building a more comprehensive map of the therapeutic landscape.
As computational methods continue to evolve and biological datasets expand, this integrative approach promises to accelerate the discovery of new treatments for diseases that currently have limited options. The medicine cabinet of the future might look surprisingly similar to today's—but with far more applications for each drug, unlocking greater value from existing medications and offering new hope to patients worldwide.
The next breakthrough treatment for a devastating disease might already be in your pharmacy—we just need to look at it from the right perspective.