In this column, Gregory Weiss, MD, provides perspective on a recent study providing evidence in support of metformin as a potential therapeutic option for treatment of atrial fibrillation.
There are many benefits to collecting longitudinal data on as many patients as possible. Many of the most important medical discoveries came about either by accident or as a result of investigations into something completely different.
Developing new drugs for any condition is a costly and time-consuming process during which countless patients continue to suffer and, in many cases, die. In contrast, when a well-known drug is found to be useful for more than one indication the process may be dramatically accelerated.
Atrial fibrillation (AF) remains the most common cardiac rhythm disturbance in the world. As many as 6 million Americans suffer from AF, a disease that carries a significant burden of morbidity and mortality.1 Although many drugs are available to treat AF no significant pharmacologic advances have been made in over a decade.
Recently researchers at the Cleveland Clinic, utilizing cutting edge genomics-based technology, have found that many of the gene sequences that the diabetes drug metformin targets overlap significantly with genes known to be dysregulated in AF.1 AF has been a very difficult disease to treat for cardiologists. Most current drugs only regulate heart rate while many others produce unpleasant side effects.
“Finding drugs or procedures to treat atrial fibrillation is difficult because of potential serious side effects,” said lead investigator Mina Chung MD, of the Cleveland Clinic, on the difficulty of developing new drugs for AF.
Another senior investigator, Feixiong Cheng, PhD, a researcher at the Genomic Medicine Institute in Cleveland, utilized artificial intelligence technology to identify 30 genes metformin targets in common with AF and an additional 8 direct expression pathways they share.1 This approach to finding repurposing strategies for drugs is called network medicine. This paradigm considers the functional and topological organization of gene products as ‘neighborhoods’, so to speak, within the human protein-protein interactome.1 This novel approach allows researchers like Cheng to compare thousands of drugs based on their gene targets and using computer algorithms determine what drugs overlap and thus may be of use in other medical condition. The benefits of this approach are many.
“We can cut off 10+ years in the drug development pipeline. We already have the information there. We just have to test it in a very computationally efficient way, such as artificial intelligence technology,” Cheng said.
Utilizing this network-based approach, the authors were able to identify metformin as a high-confidence candidate for drug repurposing for AF and further, validate it, using gene expression analysis of drug treatments in human cell lines and large-scale pharmacoepidemiologics analysis.1 These are big new words with big implications. While these data were generated by complex mathematic algorithms the associations are grounded in known biochemical mechanisms. Metformin targets a gene sequence that regulates the metabolic stress response reducing oxidative stress and mitochondrial dysfunction which may protect cardiomyocytes from apoptosis under stress.1 In addition to having a protective effect on heart cells, metformin affects gene sequences known for suppressing the renin-aldosterone-angiotensin system further reducing oxidative stress and inflammatory markers that are ubiquitous in patients with AF.1
While this approach may be alien to the average clinician and even the traditional bench scientist it is clear that data analytics machine-based research is not only contributing to our field but also here to stay. The next step should be an observational study looking at patients taking metformin for diabetes or pre-diabetes and the prevalence of AF in this population. If an association is found clinically then prospective trials can begin. The authors have shown that new directions can be found without enrolling patients or wasting time pursuing an avenue that leads to a dead end. In essence, network medicine allows researchers to skip a step and go right to investigations with the greatest likelihood of uncovering novel treatments. This is an exciting prospect for clinicians and patients that suffer from AF and so many other conditions.